The Pennsylvania State University. The Graduate School. College of Engineering

Transcription

1 The Pennsylvania State University The Graduate School College of Engineering INTEGRATED FLIGHT CONTROL DESIGN AND HANDLING QUALITIES ANALYSIS FOR A TILTROTOR AIRCRAFT A Thesis in Aerospace Engineering by Thanan Yomchinda 29 Thanan Yomchinda Submitted in Partial Fulfillment of the Requirements for the Degree of Master of Science May 29

2 The thesis of Thanan Yomchinda was reviewed and approved* by the following: Joseph F. Horn Associate Professor of Aerospace Engineering Thesis Advisor Edward Smith Professor of Aerospace Engineering George A. Lesieutre Professor of Aerospace Engineering Head of the Department of Aerospace Engineering *Signatures are on file in the Graduate School

3 ABSTRACT iii Tiltrotor aircraft present a challenge for flight control designers and handling qualities engineers. Achieving consistent handling qualities throughout an operational flight envelope is difficult since the aircraft s flight dynamics change significantly at different operating conditions (e.g. speed, attitudes, etc.) and configurations (e.g. helicopter mode, conversion mode or airplane mode). The requirement to meet both helicopter and fixed-wing flying qualities specifications always results in substantial cost and time. Development of integrated methods for flight control design and handling qualities analysis could greatly enhance the future of tiltrotor aircraft. In 28, the University of Liverpool conducted a competition on flight control design for tiltrotor aircraft. The goal is to deliver level one handling qualities across the flight envelope for a predefined search and rescue (SAR) mission for the XV-15 tiltrotor aircraft. The competitors were provided linear models of XV-15 simulation model and specifications together with the following tasks to submit: The predicted bare-airframe handling qualities A flight control system achieving predicted level one handling qualities A full report of the flight control design and handling qualities results This thesis will present the integrated methodology of model following based flight control design and handling qualities analysis developed at the Penn State University and the test pilot evaluation conducted at the University of Liverpool. In this research, a linear model analysis was developed in MATLAB to evaluate predicted handling qualities based on assigned specifications and the dynamic model. A controller

4 iv was designed to meet level one handling qualities and also have maneuvering stability. The results of predicted handling qualities were determined and showed the control concept was effective. A few criteria showed degraded handling qualities below level one. These criteria were investigated and discussed. The pilot evaluation showed that the model following / inversion controller with airspeed-scheduled inversion model was effective in delivering level one handling qualities to the tiltrotor aircraft in helicopter configuration.

5 TABLE OF CONTENTS v LIST OF FIGURES...x LIST OF TABLES...xvi ACKNOWLEDGEMENTS...xviii Chapter 1 Introduction Motivation Research Objective XV-15 Tiltrotor Aircraft Aircraft Handling Qualities Literature Review Thesis Overview...1 Chapter 2 FXV-15 Tiltrotor Bare-Airframe Analysis FXV-15 Control Methodology FXV-15 Tiltrotor Bare-Airframe Model Aircraft state matrix Aircraft control matrix Aircraft output matrix Trim data Handling Qualities Criteria to be assessed Helicopter Mode Conversion Mode Airplane Mode Bare-Airframe Handling Qualities Evaluation Helicopter Mode Small-amplitude pitch and roll attitude changes - Short-term response to control inputs (bandwidth), (ADS-33E-PRF ) Small-amplitude pitch and roll attitude changes - Mid-term response to control inputs, (ADS-33E-PRF ) Moderate-amplitude pitch and roll attitude changes (attitude quickness), (ADS-33E-PRF 3.3.3) Large-amplitude pitch and roll attitude changes, (ADS-33E- PRF 3.3.4) Small-amplitude yaw attitude changes - Short-term response to control inputs (bandwidth), (ADS-33E-PRF ) Small-amplitude yaw attitude changes - Mid-term response to control inputs, (ADS-33E-PRF ) Moderate-amplitude heading changes (attitude quickness), (ADS-33E-PRF 3.3.6)...37

6 Short-term response to disturbance input - Yaw rate response to lateral gusts, (ADS-33E-PRF ) Large-amplitude heading attitude changes, (ADS-33E-PRF 3.3.8) Interaxis coupling - Yaw due to collective for Aggressive agility, (ADS-33E-PRF ) Interaxis coupling - Pitch due to roll and roll due to pitch for Aggressive agility, (ADS-33E-PRF ) Interaxis coupling - Pitch due to roll and roll due to pitch for Target Acquisition and Tracking, (ADS-33E-PRF ) Response to collective controller - Height Response Characteristics, (ADS-33E-PRF ) Response to collective controller Vertical control power, (ADS-33E-PRF ) Conversion Mode Pitch attitude response to longitudinal controller - Shortterm response (bandwidth), (ADS-33E-PRF ) Pitch attitude response to longitudinal controller - Mid-term response to control inputs, (ADS-33E-PRF ) Pitch control power, (ADS-33E-PRF 3.4.2) Flight path control - Flight path response to pitch attitude (frontside), (ADS-33E-PRF ) Flight path control - Flight path response to collective controller (backside), (ADS-33E-PRF ) Longitudinal static stability, (ADS-33E-PRF 3.4.4) Interaxis coupling - Pitch attitude due to collective control, (ADS-33E-PRF ) Interaxis coupling - Roll due to pitch coupling for Aggressive agility, (ADS-33E-PRF ) Interaxis coupling - Pitch due to roll and roll due to pitch coupling for Target Acquisition and Tracking, (ADS-33E-PRF ) Roll attitude response to lateral controller - Smallamplitude roll attitude response to control inputs (bandwidth), (ADS-33E-PRF ) Roll attitude response to lateral controller - Moderateamplitude attitude changes (attitude quickness), (ADS-33E-PRF ) Roll attitude response to lateral controller - Largeamplitude roll attitude changes, (ADS-33E-PRF ) Roll attitude response to lateral controller - Linearity of roll response, (ADS-33E-PRF ) Roll-sideslip coupling, (ADS-33E-PRF 3.4.7)...63 vi

7 Yaw response to yaw controller - Small-amplitude yaw response for Target Acquisition and Tracking (bandwidth), (ADS-33E-PRF ) Yaw response to yaw controller - Large-amplitude heading changes for Aggressive agility, (ADS-33E-PRF ) Yaw response to yaw controller - Linearity of directional response, (ADS-33E-PRF ) Lateral-directional stability - Lateral-directional oscillations, (ADS-33E-PRF ) Lateral-directional stability - Spiral stability, (ADS-33E- PRF ) Lateral-directional characteristics in steady sideslips - Yaw control in steady sideslips (directional stability), (ADS-33E- PRF ) Lateral-directional characteristics in steady sideslips - Bank angle in steady sideslips, (ADS-33E-PRF ) Lateral-directional characteristics in steady sideslips - Lateral control in steady sideslips, (ADS-33E-PRF ) Airplane Mode Longitudinal stability with respect to speed - Phugoid stability, (MIL-F-8785C ) Longitudinal stability with respect to speed - Flight-path stability, (MIL-F-8785C ) Longitudinal maneuvering characteristics - Short-period response, (MIL-F-8785C ) Longitudinal control - Longitudinal control in maneuvering flight, (MIL-F-8785C ) Lateral-directional mode characteristics - Lateraldirectional oscillations (Dutch-roll), (MIL-F-8785C ) Lateral-directional mode characteristics - Roll mode, (MIL- F-8785C ) Lateral-directional mode characteristics - Spiral stability, (MIL-F-8785C ) Lateral-directional mode characteristics - Coupled rollspiral oscillation, (MIL-F-8785C ) Lateral-directional dynamic response characteristics - Roll rate oscillations, (MIL-F-8785C ) Lateral-directional dynamic response characteristics - Bank angle oscillations, (MIL-F-8785C ) Pilot-induced oscillations, (MIL-F-8785C 3.3.3) Roll control effectiveness, (MIL-F-8785C 3.3.4) Roll control effectiveness - Linearity of roll response, (MIL- F-8785C ) Directional control characteristics, (MIL-F-8785C 3.3.5)...84 vii

8 Lateral-directional characteristics in steady sideslips - Yawing moments in steady sideslips, (MIL-F-8785C ) Lateral-directional characteristics in steady sideslips - Side forces in steady sideslips, (MIL-F-8785C ) Lateral-directional characteristics in steady sideslips - Rolling moments in steady sideslips, (MIL-F-8785C ) Lateral-directional characteristics in steady sideslips - Positive effective dihedral limit, (MIL-F-8785C ) Summary of Predicted Handling Qualities of FXV-15 Bare-airframe models Modification of the Bare-Airframe Model...91 Chapter 3 Model Inversion Controller Model Inversion Controller Design Overview Desired Response Inversion model Dynamics Compensator Euler Angle Conversion Turn Coordination Mode Switching The Complete Controller of FXV-15 models...17 Chapter 4 Predicted Handling Qualities of the Augmented Aircraft Results Helicopter Configuration Conversion Configuration Airplane Configuration Discussion Chapter 5 Test Pilot Handling Qualities Rating Results FLIGHLAB Penn State s Flight Simulation Facility The University of Liverpool s HELIFLIGHT Flight Simulation Facility Cooper-Harper Rating Results Hover Turn Hover Reposition Bob-up Pirouette Chapter 6 Conclusion Summary of Results viii

9 6.2 Recommendations for Future Work Bibliography...16 ix

10 LIST OF FIGURES x Figure 1-1: Illustration from 1974 Tiltrotor Research Project Plan [5]...3 Figure 1-2: Propeller responses to cockpit control inputs [5]...4 Figure 1-3: Control functions in helicopter mode...5 Figure 1-4: Airplane mode control functions...6 Figure 1-5: Conversion corridor of the XV-15 tiltrotor research aircraft. [5]...7 Figure 1-6: The Cooper-Harper pilot opinion rating scale [1]...9 Figure 2-1: Schematic of FXV-15 longitudinal control...15 Figure 2-2: The structure of the aircraft state matrix [ A ]...17 Figure 2-3: The structure of aircraft control matrix [ B ]...18 Figure 2-4: FXV-15 bare-airframe schematic...2 Figure 2-5: Eigenvalues of models of FXV-15 in helicopter mode...26 Figure 2-6: Eigenvalues of models of FXV-15 in helicopter mode in low frequency region...27 Figure 2-7: Definition of bandwidth and phase delay [2]...28 Figure 2-8: Bode diagram of pitch response to longitudinal control input...29 Figure 2-9: Bode diagram of roll response to lateral control input...3 Figure 2-1: Eigenvalues corresponded to lateral and longitudinal oscillation modes of FXV-15 bare frame modes compare to requirements in ADS-33E- PRF Figure 2-11: Requirements for moderate-amplitude pitch and roll attitude changes [2]...33 Figure 2-12: Bode diagram of yaw response to pedal control input...35 Figure 2-13: Handling Quality determination of yaw bandwidth criterion...36 Figure 2-14: Requirements for moderate-amplitude heading changes...37 Figure 2-15: Definition of derivatives corresponding to lateral velocity...39

11 Figure 2-16: Handling Quality determination of yaw rate response to lateral gusts criterion...4 Figure 2-17: Handling Quality determination of yaw due to collective for aggressive agility criterion [2]...41 Figure 2-18: Pitch attitude and bank angle responses following lateral or longitudinal control inputs...43 Figure 2-19: The step responses of the vertical rate response and its first-order fitting transfer function from the 4 knots flight speed model...45 Figure 2-2: Handling Quality determination for Height response characteristics...46 Figure 2-21: Eigenvalues of models of FXV-15 in conversion mode...47 Figure 2-22: Eigenvalues of models of FXV-15 in conversion mode in low frequency region...48 Figure 2-23: Handling Quality determination for the bandwidth of pitch response...49 Figure 2-24: Eigenvalues corresponded to longitudinal oscillation modes of FXV- 15 bare frame modes compare to requirements in ADS-33E-PRF Figure 2-25: Acceleration along z-axis on body frame following maximum pitch control input (pull) on the left and those following minimum control input (push) on the right...52 Figure 2-26: Bode diagram of vertical rate response (left) and pitch attitude response (right) to longitudinal input in the 8 knots flight speed model...54 Figure 2-27: Handling quality determination for the bandwidth of flight path response to pitch attitude (frontside)...55 Figure 2-28: Handling quality determination for flight path response to collective controller (backside)...56 Figure 2-29: Responses of pitch attitude change and normal acceleration change of the 8 knots flight speed model following large corrective input in down direction...58 Figure 2-3: Pitch attitude and bank angle responses following lateral or longitudinal control inputs...59 Figure 2-31: Handling quality determination for the pitch due to roll and roll due to pitch coupling for Target Acquisition and Tracking criterion...6 xi

12 Figure 2-32: Handling quality determination for the bandwidth of roll response...61 Figure 2-33: Handling Quality determination for the attitude quickness of roll response...62 Figure 2-34: Definition of roll sideslip parameters [2]...64 Figure 2-35: Handling quality evaluation for the roll sideslip coupling criterion...65 Figure 2-36: Handling quality evaluation for the bandwidth of yaw response...66 Figure 2-37: The solution of the lateral input and sideslip angle following induced pedal control inputs...69 Figure 2-38: The solution of the longitudinal input and bank angle following induced pedal control inputs...7 Figure 2-39: Models responses following input from the solution of Equation 2.1 and induced pedal input = 2.5 inches...71 Figure 2-4: Eigenvalues of models of FXV-15 in airplane mode...73 Figure 2-41: Eigenvalues of models of FXV-15 in airplane mode in low frequencies region...74 Figure 2-42: Short-period frequency requirements for Category A Flight Phases...76 Figure 2-43: Acceleration along z-axis on body frame following maximum pitch control input (pull on the left and push on the right)...77 Figure 2-44: Responses following a disturbance in bank angle...79 Figure 2-45: Roll rate response following a yaw-control-free step roll control command...8 Figure 2-46: Handling quality evaluation of bank angle oscillations criterion...81 Figure 2-47: Bode diagrams of bare-airframe models...82 Figure 2-48: Bank angle response following the maximum lateral input...83 Figure 2-49: The solution of responses following induced pedal control inputs...85 Figure 2-5: The solution of body velocity responses following pedal control inputs...87 xii

13 Figure 2-51: Models responses following input from the solution of Equation Figure 3-1: Model inversion control concept diagram Figure 3-2: Schematic of model inversion controller...95 Figure 3-3: Mode switching diagram...16 Figure 4-1: The controller in MATLAB / SIMULINK...19 Figure 4-2: The control mixing model in MATLAB / SIMULINK...11 Figure 4-3: The quasi-linear aircraft model in MATLAB / SIMULINK Figure 4-4: Rating on small-amplitude roll attitude change - short-term response to control inputs (bandwidth) Figure 4-5: Rating on small-amplitude pitch attitude change - short-term response to control inputs (bandwidth) Figure 4-6: Rating on small-amplitude pitch (roll) mid-term response to control inputs Figure 4-7: Rating on roll attitude quickness for all other MTEs Figure 4-8: Effect of rate limit to attitude quickness tested in the hover model Figure 4-9: Rating on pitch attitude quickness for other MTEs Figure 4-1: Rating on small-amplitude yaw attitude change short term response to yaw control input (bandwidth) Figure 4-11: Rating on moderate-amplitude heading changes (attitude quickness) for Target Acquisition and Tracking Figure 4-12: Rating on moderate-amplitude heading changes (attitude quickness) for all other MTEs Figure 4-13: Rating on short-term yaw response to disturbance inputs yaw response to lateral gusts...12 Figure 4-14: Rating on interaxis coupling pitch due to roll and roll due to pitch coupling for Target Acquisition and Tracking Figure 4-15: rating on response to collective controller height response characteristics xiii

14 Figure 4-16: Rating on pitch attitude short-term response (bandwidth) Figure 4-17: Rating on pitch attitude mid-term response Figure 4-18: Phase difference between vertical response and pitch attitude response Figure 4-19: Rating on flight path control flight path response to collective controller Figure 4-2: Rating on interaxis coupling pitch due to roll and roll due to pitch coupling for Target Acquisition and Tracking Figure 4-21: Rating on small-amplitude roll attitude response to control input (bandwidth)...13 Figure 4-22: Rating on moderate-amplitude roll attitude change (attitude quickness) for Target Acquisition and Tracking...13 Figure 4-23: Rating on moderate-amplitude roll attitude change (attitude quickness) for all other MTEs Figure 4-24: Rating on small-amplitude yaw response to yaw controller for Target Acquisition and Tracking (bandwidth) Figure 4-25: Eigenvalues of augmented system of FXV-15 in airplane configuration Figure 4-26: Rating on short-period frequency and acceleration sensitivity Figure 5-1: FLIGHTLAB s modeling user interface Figure 5-2: Flight simulation facility at the Penn State Vertical Lift Research Center of Excellence (VLRCOE) Figure 5-3: Simulation process diagram Figure 5-4: HELIFLIGHT (left) and its Interior (right) Figure 5-5: Aircraft positions and control inputs data in Hover Turn MTE test Figure 5-6: Aircraft attitudes and angular rates in Hover Turn MTE test Figure 5-7: Aircraft positions and control inputs data in Hover Reposition MTE test Figure 5-8: Aircraft attitudes and angular rates in Hover Reposition MTE test xiv

15 Figure 5-9: Aircraft positions and control inputs data in Vertical Reposition (Bobup) MTE test Figure 5-1: Aircraft attitudes and angular rates in Vertical reposition (Bob-up) MTE test Figure 5-11: Aircraft positions and control inputs data in Pirouette MTE test Figure 5-12: Aircraft attitudes and angular rates in Pirouette MTE test xv

16 LIST OF TABLES xvi Table 2-1: FXV-15 Cockpit controller ranges...13 Table 2-2: FXV-15 Control surface ranges...13 Table 2-3: Inceptor to control surface gearing ratios...14 Table 2-4: The aircraft state vector...16 Table 2-5: The aircraft control vector...17 Table 2-6: Output vector...19 Table 2-7: FXV-15 Trim position data...2 Table 2-8: ADS-33E-PRF handling qualities criteria required for helicopter mode...22 Table 2-9: ADS-33E-PRF handling qualities criteria required for conversion mode...23 Table 2-1: MIL-F-8785C Handling Qualities criteria required for airplane mode...24 Table 2-11: Handling Quality determination for the pitch attitude due to collective control...58 Table 3-1: Type Caption Here...17 Table 4-1: Predicted handling qualities of XV-15 simulation in helicopter mode Table 4-2: Rating on Large-amplitude pitch (roll) attitude change Table 4-3: Data required to rate the large-amplitude heading changes Table 4-4: Data to rate response to collective controller vertical axis control power Table 4-5: Predicted handling qualities of XV-15 simulation in conversion mode Table 4-6: Rating on interaxis coupling pitch attitude due to collective control Table 4-7: Rating on interaxis coupling roll due to pitch coupling for Aggressive agility Table 4-8: Achievable roll and roll rate responses following large lateral command...131

17 Table 4-9: Rating on large-amplitude heading changes for Aggressive agility Table 4-1: Rating on lateral-directional stability - lateral-directional oscillation Table 4-11: Predicted handling qualities of XV-15 simulation in airplane mode Table 4-12: Damping ratio of phugoid mode for phugoid stability criterion Table 4-13: Short-period damping ratio in airplane mode Table 4-14: Achievable load factor of models of airplane configuration Table 4-15: Percentage of the first minimum following the first peak in roll rate response Table 4-16: Time to achieve the assigned bank angle Table 5-1: Test pilot rating results xvii

18 ACKNOWLEDGEMENTS xviii First of all, I would like to express my thanks to my thesis advisor, Dr. Joseph F. Horn, for giving me an opportunity to work on this research and for his academic guidance, patience, support and encouragement throughout my graduate studies. Next, I would like to express my deepest gratitude to my mother, my father, and my family for their love, their support, and their attempt and eagerness to see me pursue my dream. Finally, I would like to thank my friends, in Thailand and United States of America, for their help and support.

19 Chapter 1 Introduction Handling qualities is defined as those qualities or characteristics of an aircraft that govern the ease and precision with which a pilot is able to perform the tasks required in support of an aircraft role [1]. Handling qualities requirements are intended to assure that there is no limitation on flight safety or on the capability to perform particular tasks resulting from excessive pilot workload. In order to meet those requirements, a suitable Flight Control System (FCS) is needed. Therefore, Flight Control design plays an important role to ensure the aircraft is compatible for a pilot to do particular missions [2, 3, 4]. 1.1 Motivation The tiltrotor handling qualities workshop, in association with the American Helicopter Society (AHS) / Royal Aeronautical Society (RAeS) Rotorcraft Handling Qualities Conference, offered a competition to develop a flight control system for the FLIGHTLAB XV-15 tiltrotor simulation model. The tasks are to analyze the handling qualities of the bare-airframe, and then design a controller that is suitable for a specific mission. The challenge in flight control design is to achieve desired handling qualities over the entire flight envelope. The tiltrotor aircraft has a wide range of configurations. Not only do the aircraft flight dynamics vary significantly for different operating point,

20 2 but also control characteristics change during the conversion from airplane-like configuration to helicopter-like configuration. In addition to the challenge, the research is motivated by the desire of developing a Flight Control System which makes an aircraft more accessible and easier to fly. This helps reduce pilot workload, increase safety of flight, and lower cost of pilot training. 1.2 Research Objective This research shares the main objective with the tiltrotor handling qualities workshop competition. The research aims to find a controller that deliver level one handling qualities across the flight envelope for a predefined Search and Rescue (SAR) mission for the XV-15 tiltrotor aircraft simulation model. The handling qualities criteria to be assessed are specified by the workshop. In addition to the main objective, controller should achieve closed loop stability over all normal flight conditions. Moreover, this research intends to investigate the design of Flight Control System for tiltrotor aircraft. The study aims to find advantages and disadvantages of tiltrotor aircraft in achieving handling qualities specified for rotorcraft (mainly helicopters) and fixed wing airplane. The achieved controller might also be used to be a template in designing Flight Control System for other tiltrotor aircraft in the future.

21 1.3 XV-15 Tiltrotor Aircraft 3 The XV-15 was a product of NASA-Army-Bell cooperation. Starting in early 197s, the XV-15 tiltrotor project was aimed to meet both civil and military needs [5]. This dual use capability is illustrated in Figure 1-1. Figure 1-1: Illustration from 1974 Tiltrotor Research Project Plan [5] The flight control system of the aircraft was designed for a crew of two. However, it was designed to permit a single pilot to perform any operation from either seat. Each pilot station had complete controls for pitch, roll, yaw, and thrust in all modes of flight. They consisted of control sticks, rudder pedals with brakes, and power levers (to control prop-rotor thrust and engine throttle functions). For throttles rpm governor, flap and landing gear controls, the middle console between seats was where a single set

22 4 of their controllers located. In the helicopter mode, the controls apply collective or cycle blade pitch changes to the rotors to produce control moments and forces. Fore and aft cyclic pitch produced by moving the center control stick fore and aft provides longitudinal control and, differential longitudinal cyclic pitch responding to rudder pedal motion produces directional control. Collective pitch which is indirectly commanded by collective lever is used for vertical control, and differential collective pitch resulting from sideward displacement of center stick is used for lateral control [5]. The collective lever is set to control throttle while a beta governor controls rotor collective pitch in order to keep constant rotor RPM. The control functions are illustrated in Figure 1-2 and Figure 1-3. Figure 1-2: Propeller responses to cockpit control inputs [5]

23 5 Figure 1-3: Control functions in helicopter mode. In airplane mode, when the nacelles are tilted fully forward, the controls apply commands to ailerons, elevators and rudders (fixed wing control surfaces) to produce flight path control moments and forces as a conventional airplane. The differential cyclic pitch and differential collective pitch of rotors are not used. Only collective pitch that is

24 indirectly commanded by collective lever is used to control forward thrust [5]. The control functions in airplane configuration are illustrated in Figure Figure 1-4: Airplane mode control functions.

25 7 In conversion mode, controls can be made suitable for a range of airspeeds, nacelle angles and fuselage attitudes. The rotor controls are designed to automatically phase out as the nacelles are tilted forward toward the airplane configuration. The purpose of the system is to minimize the need for control inputs during conversion. The phasing of the controls through conversion is smooth and not apparent to the pilot. All fixed wing control surfaces remain active in all flight configurations [5]. Figure 1-5 displays the conversion corridor of the aircraft. Figure 1-5: Conversion corridor of the XV-15 tiltrotor research aircraft. [5]

26 1.4 Aircraft Handling Qualities 8 There are two principle regimes in the science of aircraft flight test. One is the aircraft handling qualities. The other is the aircraft performance. Handling qualities, sometimes referred to as flying qualities, involves the study and evaluation of aircraft s stability and control characteristics that have a critical bearing on the safety of flight and on the ease of controlling an aircraft in steady flight and in maneuvers [4]. The study of airplane handling qualities started in the 19s and had been developed ever since. In 1954, the first specification of handling qualities that involves influence of task performance was clarified. In 1969, the quantification as measure of workload in the performance of a variety of tasks by pilot opinion rating was developed by Cooper and Harper [1]. Figure 1-6 shows the rating scale developed by Cooper and Harper in In the same year, thank to the large quantity of new research data available, an entirely new version of the US handling qualities specification, MIL-F- 8785B (ASG), was published. It was enhanced by pilot-in-the-loop research results and introduced the concept of handling quality levels related to pilot opinion ratings, aircraft states, flight envelopes and flight control augmentation effects. The further update, MIL-F8785C, had appeared in 198. It includes criteria accounting for the high order effects which appeared to be problems in many fly-by-wire control systems. Unfortunately, it did not perform its intention very successful as problems continued to happen. After more research to analyze the problems and publication of many new criteria, finally, a complete specification, MIL-STD-1797, was published in 1987 [4].

27 9 Figure 1-6: The Cooper-Harper pilot opinion rating scale [1]. 1.5 Literature Review Dynamic model inversion is a popular feedback linearization method for achieving consistent response characteristics [7]. It has been applied to several types of aircraft. Its capabilities have been demonstrated for both helicopter and fixed-wing aircraft [7, 8, 9]. In tiltrotor aircraft, there is little information about the flight control systems used in the real vehicle (XV-3, XV-15, V-22 Osprey, etc.) in the public domain. However, there were attempts on applying control laws to control a simulation model of tiltrotor

28 1 aircraft. Calise and Rysdyk did researches on adaptive flight control using model inversion control with neural networks for XV-15 simulation model [9, 1, 11, 12]. Their controller showed good results in tracking commanded attitudes for the tiltrotor model in helicopter mode. Mehra, Prasanth and Gopalaswamy approached the design of XV-15 Flight Control System by using model predictive control concept [13]. They claimed that the controller achieved performance and satisfied stringent robustness and stability requirements for XV-15 in airplane configuration. The approach of using H-infinity and μ-synthesis techniques in designing tiltrotor Flight Control System was done by Manimala, Padfield, Walker, Naddei, Verde, Ciniglio, Rollet and Sandri [14]. They used those techniques to design structural load alleviation control law for the FLIGHTLAB XV-15 simulation model. Walker and Voskuijl in University of Liverpool (UoL) also applied H-infinity to design a longitudinal axis augmentation system for XV-15 tiltrotor aircraft in airplane mode [15]. 1.6 Thesis Overview The remainder of this thesis is organized as follows: Chapter two will present more information regarding to FXV-15 model, handling qualities to be assessed and unaugmented handling qualities results. Chapter three will develop a flight control system based on the concept of the model following / model inversion control for the FXV-15 model. The controller will be demonstrated and evaluated for predicted handling qualities results in chapter four. Chapter five will present Cooper-Harper

29 handling qualities rating of the augmented FXV-15 simulation. The conclusion and recommendation for future will be in chapter six. 11

30 Chapter 2 FXV-15 Tiltrotor Bare-Airframe Analysis This chapter focuses on the bare-airframe dynamics for which we are to design a controller. The FXV-15 control methodology will be introduced. The linear models which are used in the FXV-15 Flight Control System development competition will be discussed and the handling qualities criteria to be assessed will be described. Afterwards, the handling qualities of bare-airframe models will be evaluated. Lastly, the rating results will be illustrated and discussed in each criterion. The overall handling qualities evaluation of bare-airframe models will be presented at the end of this chapter. 2.1 FXV-15 Control Methodology As previously mentioned in the chapter one, the control of the FXV-15 is designed to combine the helicopter control and fixed wing control together. The blending of these mode controls is critical in the conversion corridor. The cockpit controller ranges are shown in Table 2-1 and the control surface ranges are listed in Table 2-2. The connections between cockpit controller inputs and control surfaces vary as the aircraft transitions from helicopter mode to airplane configuration. In this research, the mechanical gearing ratios between inceptors and control surfaces for the configurations studied are specified in Table 2-3.

31 13 Table 2-1: FXV-15 Cockpit controller ranges Cockpit Control Range (inches) Lateral Stick ± 4. 8 Longitudinal Stick ± 4. 8 Collective Lever 12 Pedal range ± 2. 5 Table 2-2: FXV-15 Control surface ranges Control Surface Range (degree) Collective pitch angle 16.6 Differential collective pitch angle 6 Longitudinal cyclic pitch angle 2 Differential longitudinal cyclic pitch 15.4 angle Ailerons 37.8 Elevator 4 Rudder 4 The control surface ranges listed in Table 2-2 are the range of control available from the trim condition. For example, the rotor collective pitch angle which is defined as 16.6 degree might be added to a bias of up to 4 degree in airplane mode. The control mixing and power management system gives this bias. In this competition, all actuators of control surfaces are modeled as the transfer function defined in Equation 2.1. And all actuation rates are limited to be approximately 1% of their range per second. For example, collective pitch angle will operate at 16.6 degree per second. A = 1 1 s

32 14 Table 2-3: Inceptor to control surface gearing ratios Mechanical Gearing between cockpit controls and control Control surface surface (degrees per inch) 9 6 (Helicopter Mode) (Conversion Mode) (Airplane Mode) Collective pitch Differential collective pitch Longitudinal cyclic pitch Differential -1.6 <6kts <6kts longitudinal cyclic kts -.9 8kts pitch -.4 >1kts >1kts Ailerons Elevator Rudder Figure 2-1 displays a schematic of the longitudinal control architecture of the FXV-15 simulation model. An input from longitudinal stick is applied to both the elevator and the rotor longitudinal cyclic blade pitch. A similar architecture also applies to lateral control and pedal control. The rotor collective pitch of the FXV-15 is controlled by the collective lever and beta governor. Input signal from collective lever is adjusted by signal from beta governor in order to maintain constant revolutions per minute (RPM). To obtain the RPM during the simulation, an engine model is needed. However, no engine model is provided for competitors. Therefore, the simplified model is operated on the assumption that RPM of rotors is constant. Without an engine parameter to adjust the RPM, the collective control in this simplified model could use a control schematic similar to that is shown in Figure 2-1. The scheduled gain for collective control could be set to be 1.38 in helicopter

33 mode (9º nacelle angles) and scaled by the sine of the nacelle angles when the nacelle angles change as shown in Table Figure 2-1: Schematic of FXV-15 longitudinal control 2.2 FXV-15 Tiltrotor Bare-Airframe Model The information on the FXV-15 bare-airframe system was provided by University of Liverpool in the form of linear state space models. Equation 2.2 shows the standard form of a linear state space model. x& = A x + B u y = C x + D u 2.2 The derivative matrices for aircraft s state, control and z-axis acceleration output are provided in the following cases. For FXV-15 in helicopter mode, the cases of, 2 and 4 knots forward flight speed are provided. In conversion mode, the cases of 8, 1, 14 knots forward flight speed are provided. And the cases of 16, 2 and 24 knots forward flight speed are provided for airplane mode.

34 2.2.1 Aircraft state matrix 16 Table 2-4 shows the aircraft state vector, x, of the aircraft models. Table 2-4: The aircraft state vector State no. Parameters Symbols Units 1 Aircraft s bank angle φ rad 2 Aircraft s pitch attitude θ rad 3 Aircraft s heading ψ rad 4 Rotor2 gimbal pitch Rotor2 gimbal roll Rotor1 gimbal pitch Rotor1 gimbal roll Aircraft s body velocity on x-axis v xb ft/sec 9 Aircraft s body velocity on y-axis v yb ft/sec 1 Aircraft s body velocity on z-axis v zb ft/sec 11 Aircraft s roll rate p rad/sec 12 Aircraft s pitch rate q rad/sec 13 Aircraft s yaw rate r rad/sec The structure of dimensional derivatives in the aircraft state matrices [ A ] is shown in Figure 2-2.

35 17 Figure 2-2: The structure of the aircraft state matrix [ A ] Aircraft control matrix Table 2-5 shows the control vector, u, of the aircraft models. Table 2-5: The aircraft control vector Input no. Parameters Symbols Units 1 Rotors collective θ deg 2 Rotors differential collective θ d deg 3 Rotors longitudinal cyclic θ 1s deg 4 Rotors differential longitudinal cyclic θ 1sd deg 5 Rotors lateral cyclic θ 1c deg 6 Aileron ξ deg 7 Elevator η deg 8 Rudder ζ deg

36 The structure of dimensional derivatives in the aircraft control matrices [ B ] is shown in Figure Figure 2-3: The structure of aircraft control matrix [ B ] Aircraft output matrix Table 2-6 shows the output vector, y, of the aircraft models. Equation 2.3 is used to obtain output vector from the state space models.

37 Trim data Table 2-7 lists the trim inceptor positions. The required control surface displacements from each specific case can be calculated from these inceptor positions. All lateral stick and pedal cockpit controls are zero. Bank angle and heading are also zero. Table 2-6: Output vector Output no. Parameters Symbols Units 1 Body inertial z-acc z a ft/sec 2 2 Aircraft s bank angle φ rad 3 Aircraft s pitch attitude θ rad 4 Aircraft s heading ψ rad 5 Aircraft s body velocity on x-axis xb v ft/sec 6 Aircraft s body velocity on y-axis yb v ft/sec 7 Aircraft s body velocity on z-axis zb v ft/sec 8 Aircraft s roll rate p rad/sec 9 Aircraft s pitch rate q rad/sec 1 Aircraft s yaw rate r rad/sec x a a a a a a a a a a a a a y r q p w v u y G x G y G x G = ψ θ φ 2.3

38 Table 2-7: FXV-15 Trim position data Nacelle (deg) Speed (knots) Pitch attitude (rad) Longitudinal stick (inches) Collective lever (inches) v xb (ft/s) v zb (ft/s) Figure 2-4 shows the schematic of the FXV-15 bare-airframe model. The command input, δ C, of bare-airframe model consists of lateral stick input, δ lat, longitudinal stick input, δ long, collective lever input, δ col, and pedal input, δ ped. The control mixing takes command input, δ C, and return positions of aircraft control surfaces which are parameters of a control vector, u, in the linear state space models. Command Input δ C Control Mixing u x& = A x + B u y = C x + D u y Aircraft Responses Figure 2-4: FXV-15 bare-airframe schematic

39 2.3 Handling Qualities Criteria to be assessed 21 Due to the limited information on tiltrotor aircraft handling qualities exists in the public domain, an evaluation will be conducted using the criteria based upon ADS-33E- PRF [2] for low speed helicopter mode maneuvering and for conversion mode at 6 degree nacelle angle maneuvering. For airplane mode, an evaluation will be conducted using criteria from the military specification MIL-F-8785C [3]. These two documents provide handling qualities guidelines for piloted vehicles. The handling qualities criteria to be assessed for the FXV-15 were specified by the Tiltrotor Handling Qualities Workshop Helicopter Mode Table 2-8 lists the ADS-33E-PRF handling qualities criteria for the FXV-15 helicopter mode to be assessed.

40 Table 2-8: ADS-33E-PRF handling qualities criteria required for helicopter mode Short-term response to control inputs (bandwidth) pitch (roll) Mid-term response to control inputs pitch (roll) Moderate-amplitude pitch (roll) attitude changes (attitude quickness) Large-amplitude pitch (roll) attitude changes Short-term response to control inputs (bandwidth) yaw Mid-term response to control inputs yaw Moderate-amplitude heading changes (attitude quickness) Yaw rate response to lateral gusts Large-amplitude heading attitude changes Yaw due to collective for Aggressive agility Pitch due to roll and roll due to pitch for Aggressive agility Pitch due to roll and roll due to pitch for Target Acquisition and Tracking Height Response Characteristics Vertical axis control power Conversion Mode Table 2-9 lists the ADS-33E-PRF handling qualities criteria for the FXV-15 conversion mode to be assessed.

41 Table 2-9: ADS-33E-PRF handling qualities criteria required for conversion mode Short-term response (bandwidth) pitch Mid-term response to control inputs pitch Pitch control power Flight path response to pitch attitude (frontside) Flight path response to collective controller (backside) Longitudinal static stability Pitch attitude due to collective control Roll due to pitch coupling for Aggressive agility Pitch due to roll and roll due to pitch coupling for Target Acquisition and Tracking Small-amplitude roll attitude changes (bandwidth) Moderate-amplitude attitude changes (attitude quickness) Large-amplitude roll attitude changes Linearity of roll response Roll-sideslip coupling Small-amplitude yaw response for Target Acquisition and Tracking (bandwidth) Large-amplitude heading changes for Aggressive agility Linearity of directional response Lateral-directional oscillations Spiral stability Yaw control in steady sideslips (directional stability) Bank angle in steady sideslips Lateral control in steady sideslips Airplane Mode Table 2-1 lists the MIL-F-8785C handling qualities criteria for the FXV-15 airplane mode to be assessed. The FXV-15 is classified as a medium weight, low-tomedium maneuverability airplanes (class II) and operated in nonterminal flight phases that require rapid maneuvering, precision tracking, or precision flight-path control (category A).

42 24 Table 2-1: MIL-F-8785C Handling Qualities criteria required for airplane mode Phugoid stability Flight-path stability Short-period response Longitudinal control in maneuvering flight ( to 2.5g) Lateral-directional oscillations (Dutch-roll) Roll mode Spiral stability Coupled roll-spiral oscillation Roll rate oscillations Bank angle oscillations Pilot-induced oscillations Roll control effectiveness Linearity of roll response Directional control characteristics Yawing moments in steady sideslips Side forces in steady sideslips Rolling moments in steady sideslips Positive effective dihedral limit 2.4 Bare-Airframe Handling Qualities Evaluation The requirements listed in the previous section are used to determine the handling qualities of the FXV-15 based on the provided linearized models. The criteria generally categorize handling qualities into three levels (level one, two and three). A level one rating is the best handling quality of a criterion. A level three indicates bad handling qualities. If no requirement is satisfied (cannot satisfy level one, two and three), a degraded level will be announced for that criterion. If a criterion does not rate handling qualities into levels, a level one will be given if all requirements are achieved; otherwise, a degraded level will be given to that criterion.

43 25 Some criteria in ADS-33E-PRF have several requirements for different mission task elements (MTEs), the toughest requirement which is usually for Target Acquisition and Tracking MTE will be assessed. For those criteria, if a requirement of level one for Target Acquisition and Tracking MTE cannot be achieved, a requirement of level one for other MTEs will be considered Helicopter Mode Figure 2-5 shows all eigenvalues from the linearized models of the FXV-15 in helicopter mode. Eigenvalues corresponding to actuator modes are not shown. Figure 2-6 shows the same eigenvalues as Figure 2-5 but zooms in to the low frequency region. Shown in Figure 2-5 and 2-6, some eigenvalues corresponding to aircraft motion modes are in the right haft plane; therefore, the bare-airframe dynamics is unstable. The eigenvalue of roll subsidence is relatively small compared to typical rotorcraft. This may be attributed to the face that the XV-15 tiltrotor aircraft have high roll inertia due to mass of engines, located at a distance from the roll axis.

44 Rotors gimbals mode Imaginary axis Hover -4 2kts kts Real axis Figure 2-5: Eigenvalues of models of FXV-15 in helicopter mode

45 27 Roll mode Imaginary axis Short period mode Heave mode Dutch roll mode Phugoid in Hover Spiral mode -.4 Pitch mode Hover.8 Phugoid mode -.8 2kts kts Real axis Figure 2-6: Eigenvalues of models of FXV-15 in helicopter mode in low frequency region Small-amplitude pitch and roll attitude changes - Short-term response to control inputs (bandwidth), (ADS-33E-PRF ) This criterion has requirements on the bandwidth ( ω BW ) and phase delay ( τ p ) to prevent tendency of pilot induced oscillation (PIO). The definition of the bandwidth and phase delay is shown in Figure 2-7.

46 Figure 2-7: Definition of bandwidth and phase delay [2]. 28

47 29 5 Bode Diagram From: XB from Workspace (pt. 1) To: Model1/Demux (pt. 6) Magnitude (db) θ δ long (s) Phase (deg) Hover 2kts 4kts Frequency (rad/sec) Figure 2-8: Bode diagram of pitch response to longitudinal control input Figure 2-8 shows the bode diagram of the pitch response to longitudinal cockpit control inputs of bare-airframe models. It displays that only the phase of the 4 knots flight speed model reaches -135 degree which is the point that defines the bandwidth. Since the phase of other two models does not reach the point that defined the bandwidth, the hover and 2 knots flight speed models have undefined bandwidth.

48 3 5 Bode Diagram From: XA from Workspace (pt. 1) To: Model1/Demux (pt. 5) φ δ lat Magnitude (db) (s) Phase (deg) Hover 2kts 4kts Frequency (rad/sec) Figure 2-9: Bode diagram of roll response to lateral control input Figure 2-9 shows the bode diagram of the roll response to lateral cockpit control inputs of bare-airframe models. It displays that the bandwidth can be defined in 2 knots flight speed and 4 knots flight speed models. The model of the hover case has an undefined bandwidth. Because all the bandwidth was undefined, the rating of handling qualities in this criterion was predicted to be a level three. However, the criterion would be evaluated in closed-loop model analysis.

49 Small-amplitude pitch and roll attitude changes - Mid-term response to control inputs, (ADS-33E-PRF ) 31 The mid-term response requirements are specified for all frequencies below the bandwidth frequency obtain in previous criterion (ADS-33E-PRF ). It provides requirements in dampling ratio of longitudinal and lateral oscillation modes. Figure 2-1 shows the eigenvalues corresponded to lateral and longitudinal oscillation modes of the FXV-15 bare-airframe models. The eigenvalues corresponding to actuators and rotor gimbals are not shown because those modes are normally higher than the defined bandwidths. The figure shows that the bare-airframe models mid-term response is not good and satisfies a level three handling quality. This is expected from the unaugmented rotorcraft. 1.5 Hover 2kts 4kts Level1 Level2 Level3 1 Imaginary axis Real axis Figure 2-1: Eigenvalues corresponded to lateral and longitudinal oscillation modes of FXV-15 bare frame modes compare to requirements in ADS-33E-PRF

50 Moderate-amplitude pitch and roll attitude changes (attitude quickness), (ADS- 33E-PRF 3.3.3) 32 This criterion provides requirements on the ratio of peak pitch rate to peak pitch attitude change, q pk Δ θ and the ratio of peak roll rate to peak roll attitude change, pk p pk Δ φ. The requirements and definitions are shown in Figure pk In order to achieve high value of the ratio, the aircraft must be able to generate high pitch and roll rates in a very short time. Therefore, for this criterion the actuators time constant and rate limit have an effect on handling qualities. In this analysis, it was difficult to measure required parameters due to instability of the bare-airframe models. In order to obtain the required response as defined in ADS- 33E-PRF, it required significant reversals in the sign of the cockpit control input. This is in contradiction to the requirements. Moreover, an angular acceleration of roll responses of the aircraft might be insufficient to satisfy good handling qualities requirements due to high roll inertia. A degraded level of handling qualities was given.

51 Figure 2-11: Requirements for moderate-amplitude pitch and roll attitude changes [2] 33

52 Large-amplitude pitch and roll attitude changes, (ADS-33E-PRF 3.3.4) 34 This criterion provides requirements in achievable attitude change for attitude command / attitude hold (ACAH) type and achievable angular rate for rate command / attitude hold (RCAH) type. The achievable pitch and roll attitude change and angular rate are simply found by applying the maximum longitudinal and lateral input to the aircraft model. Since the XV-15 bare-airframe model was unstable, its responses were divergent. Its pitch and roll attitude change and angular rate change responses increased divergently as the time pass and were able to be higher than those specified in the requirements. The handling qualities level of the bare-airframe model was predicted to be a level one on this criterion Small-amplitude yaw attitude changes - Short-term response to control inputs (bandwidth), (ADS-33E-PRF ) This criterion provides requirements on the bandwidth ( ω BW ψ ( τ p ). The definition of the bandwidth and phase delay is shown in Figure 2-7. ψ ) and phase delay Figure 2-12 shows the bode diagram of the yaw response to pedal control inputs of bare-airframe models that was used to determine the bandwidth and phase delay. The parameters were able to be defined. Figure 2-13 shows the rating of the bare-airframe model in this criterion. A level three handling qualities is satisfied.

53 35 1 Bode Diagram From: XP from Workspace (pt. 1) To: Model1/Demux (pt. 7) 5 ψ δ ped Magnitude (db) (s) Phase (deg) Hover 2kts 4kts Frequency (rad/sec) Figure 2-12: Bode diagram of yaw response to pedal control input

54 Hover 2kts 4kts.3 Level3 τ p,ψ (sec) Level2 Level ω BW,ψ (rad/sec) Figure 2-13: Handling Quality determination of yaw bandwidth criterion Small-amplitude yaw attitude changes - Mid-term response to control inputs, (ADS-33E-PRF ) The mid-term response requirements are specified for all frequencies below the bandwidth frequency obtain in previous criterion (ADS-33E-PRF ). It provides a requirement on any oscillation modes to have damping ratio at least.35 ( ζ =. 35 ). The motion modes of the bare-airframe model are shown in Figure 2-6. From the figure, the oscillation mode that should be considered in this criterion would be the

55 37 dutch-roll mode in forward flight because it is the only oscillatory mode among lateraldirectional motion modes. Since the dutch-roll mode of the bare-airframe model does not meet the requirement for handling qualities level one but is good enough to satisfy level two. A level two handling quality is achieved in this criterion Moderate-amplitude heading changes (attitude quickness), (ADS-33E-PRF 3.3.6) This criterion has a requirement on the ratio of peak yaw rate to change in heading, rpk Δ ψ pk. It is required that the aircraft would generate the response as defined in Figure 2-11 without significant reversals in the sign of the cockpit control input. The requirement on the ratio is shown in Figure Figure 2-14: Requirements for moderate-amplitude heading changes In this analysis, it was difficult to measure required parameters due to instability of the bare-airframe models. In order to obtain responses as the definitions, significant reversals in the sign of the cockpit control input relative to trim position was required.

56 38 This failed to satisfy the requirement. Moreover, an angular acceleration of yaw responses of the aircraft might be insufficient to satisfy good handling qualities requirements due to high yaw inertia. A degraded level of handling qualities was predicted Short-term response to disturbance input - Yaw rate response to lateral gusts, (ADS-33E-PRF ) This criterion limits the peak yaw rate within the first three seconds when the aircraft encounters lateral wind gusts. It requires that the ratio of peak yaw rate to lateral wind gust velocity, r pk V g, is less than.3 degree-per-second per feet-per-second in lateral wind gust magnitudes of 1 to 25 knots for level one. In the analysis of aircraft in lateral wind gust, the aircraft model was needed to take lateral wind gust velocity, V g, as another input. Since a situation of aircraft in lateral wind was equivalent to a situation that aircraft have lateral velocity of the same magnitude as the wind, the aircraft model was modified to be able to set the value of body lateral velocity. In order to do that, the derivatives corresponding to lateral velocity, A v, from aircraft state matrix was used to simulate lateral velocity in that model. The derivatives corresponding to lateral velocity, A v, in an aircraft state model is shown in Figure Equation 2.4 shows the modification of simplified model to take lateral wind gust velocity, V g, as an input.

57 39 A v Figure 2-15: Definition of derivatives corresponding to lateral velocity x& = A x + B u + A v V g 2.4 Figure 2-16 shows the value of peak yaw rate which was measured after steady lateral wind gusts were presented. The graph shows that the ratio does not change with the magnitudes of the wind. This is expected from a linear model. The results also show that a level two of handling qualities is rated.

58 Level2 r pk /V g [(deg/sec)/(ft/sec)] Level1.5 Hover 2kts 4kts V g (knots) Figure 2-16: Handling Quality determination of yaw rate response to lateral gusts criterion Large-amplitude heading attitude changes, (ADS-33E-PRF 3.3.8) A requirement on the achievable yaw rate in hover is to be met in this criterion. Because of the instability of the models, the yaw rate response will increase exponentially from a disturbance (unstable spiral mode). Since there is no objection about instability in the requirements, level one handling quality is satisfied.

59 Interaxis coupling - Yaw due to collective for Aggressive agility, (ADS-33E- PRF ) 41 This criterion limits the yaw rate response to abrupt step collective inputs. Since a tiltrotor aircraft creates lift forces by two rotors at its left-and-right side symmetrically, torque moments following collective lever inputs from both rotors will cancel each other. There should not be excessive moment on the aircraft to change its yaw rate Level2 r 1 /V z (3) (deg/sec/ft/sec) Level1.2 Hover.1 2kts 4kts r 3 / V z (3) (deg/sec/ft/sec)) Figure 2-17: Handling Quality determination of yaw due to collective for aggressive agility criterion [2]

60 The results in Figure 2-17 confirm our idea mentioned above. Small values of the ratio of yaw rate, r, over vertical velocity, v z, implies that the aircraft has small coupling of yaw due to collective. A level one handling quality is satisfied Interaxis coupling - Pitch due to roll and roll due to pitch for Aggressive agility, (ADS-33E-PRF ) In this criterion, the ratio of peak pitch attitude within four seconds to bank angle at four seconds following an abrupt lateral control step input, Δθ pk Δφ4, and the ratio of peak bank angle within four seconds to pitch attitude at four seconds following an abrupt longitudinal control step input, Δφ pk Δθ 4, are limited. All values of responses are the magnitude deviating from their trim values. For an ideal tiltrotor aircraft, in straight wing-level flight case, aerodynamic force changes from longitudinal control inputs will be the same on the left and right sides. The roll moment will not affect the aircraft because roll moments from left and right sides perfectly cancel each other. The roll due to pitch coupling must be very close to zero. This idea is also held in the pitch due to roll coupling. Therefore, tiltrotor aircraft will have low cross-coupling. In Figure 2-18, the ratios of Δθ pk Δφ4 and Δθ pk Δφ4 are very small (less than.1) while the criterion allows at most.25 for level one. The handling quality level one is satisfied.

61 43 θ (deg) θ from XA δ lat δ lat φ (deg) φ from XA 8 Hover 6 2kts 4kts time time θ from XB δ φ from XB long δ long θ (deg) 4 φ (deg) time time Figure 2-18: Pitch attitude and bank angle responses following lateral or longitudinal control inputs Interaxis coupling - Pitch due to roll and roll due to pitch for Target Acquisition and Tracking, (ADS-33E-PRF ) This criterion provides another requirement for pitch due to roll and roll due to pitch coupling. In order to calculate parameters required in the criterion, the bandwidth parameters on pitch and roll, ω, θ ω, BW and BW φ, defined in Small-amplitude pitch (roll) attitude changes - Short-term response to control inputs (bandwidth) are needed.

62 44 Since the bandwidth of models cannot be defined, this criterion cannot be determined. However, because the cross coupling between roll and pitch is very small as shown in the previous criterion, the requirement of level one handling qualities is predicted Response to collective controller - Height Response Characteristics, (ADS- 33E-PRF ) This criterion provides a procedure to consider a vertical rate response, h &, following a step collective input. The response within five seconds after a step input is determined as a qualitative first-order appearance and the characteristics of that qualitative first-order appearance are used to rate the handling qualities. Equation 2.5 shows a first-order fit of the vertical rate response to a collective input. The parameters required to rate this criterion are defined in the following equation. h& δ col () s K exp = T h&, eq ( τ & s) h, eq s Figure 2-19 illustrates an example of the step responses of the vertical rate response and the first-order fitting transfer function from Equation 2.5.

63 45 12 Step Response 1 8 Amplitude Time (sec) Vertical rate 1st order fitting Figure 2-19: The step responses of the vertical rate response and its first-order fitting transfer function from the 4 knots flight speed model The result of calculation is shown in Figure 2-2. The requirement of a level two of handling qualities is satisfied.

64 46.35 ADS-33E.3.25 LEVEL 2 τ h dot eq (sec) LEVEL 1.5 Hover 2kts 4kts T h dot eq (sec) Figure 2-2: Handling Quality determination for Height response characteristics Response to collective controller Vertical control power, (ADS-33E-PRF ) This criterion provides limits on the achievable vertical rate at 1.5 seconds after initiation of a rapid displacement of collective control input. The bare-airframe models of FXV-15 was able to generate more than 1 ft/min from the maximum collective control input while the requirement requires at least 16 ft/min for level one. A level one handling quality was given.

65 2.4.2 Conversion Mode 47 Figure 2-21 shows all eigenvalues of the linearized models of FXV-15 in conversion mode. Eigenvalues corresponding to actuator modes are not shown. Figure 2-22 shows the same eigenvalues in Figure 2-21 but zooms in to low frequency region. From the graphs, all eigenvalues corresponding to aircraft motion modes are in left haft plane. Therefore, the bare-airframe dynamics are stable. Imaginary axis Rotors gimbals mode Dutch roll mode Roll mode kts 1kts 14kts Short period mode Real axis Figure 2-21: Eigenvalues of models of FXV-15 in conversion mode

66 48 Imaginary axis Phugoid mode Spiral mode kts 1kts kts Real axis Figure 2-22: Eigenvalues of models of FXV-15 in conversion mode in low frequency region Pitch attitude response to longitudinal controller - Short-term response (bandwidth), (ADS-33E-PRF ) This criterion provides requirements on the bandwidth, ω BW, θ, and phase delay, τ p,θ of pitch attitude response. This criterion is similar to the criterion in the small amplitude pitch and roll attitude changes - short-term response to control inputs (bandwidth).

67 Figure 2-23 shows the bandwidth, ω, θ τ, 49 BW, and phase delay, p θ, compared to the requirement. The requirement of level one handling quality is satisfied kts 1kts 14kts Level3 Level2 Level1.25 τ p,θ (sec) ω BW,θ (rad/sec) Figure 2-23: Handling Quality determination for the bandwidth of pitch response

68 Pitch attitude response to longitudinal controller - Mid-term response to control inputs, (ADS-33E-PRF ) 5 This criterion classifies the damping ratio of longitudinal oscillation modes. This requirement is similar to that in Small amplitude pitch and roll attitude changes - mid-term response to control inputs. Figure 2-24 shows eigenvalues corresponding to longitudinal oscillation modes and the requirement limit. A level one handling quality is satisfied Imaginary axis Level1 Level2 Level kts 1kts 14kts Real axis Figure 2-24: Eigenvalues corresponded to longitudinal oscillation modes of FXV-15 bare frame modes compare to requirements in ADS-33E-PRF

69 Pitch control power, (ADS-33E-PRF 3.4.2) 51 This criterion provides requirement on the service load factors. The service load factor is determined using the acceleration along z-axis on aircraft body frame that is developed by use of the pitch control alone. The load factor, n z, is defined in Equation 2.6. In this analysis, the service load factor was not specified for the aircraft in conversion mode but it was defined for the airplane mode. The competition defined service load factor of the aircraft in airplane configuration in a range of maximum of 2.5g to minimum of g. This service load factor was used for handling qualities evaluation in this criterion. Figure 2-25 shows time history of the load factor following maximum and minimum pitch control input respectively. The figure shows that load factor of the bareairframe model does not cover the service load factors. The requirement for level one handling qualities is not satisfied. One may notice that the load factor of 14 knots flight speed model is very high (almost 7g). The reason explaining this high load factor is because of the trim condition of pitch control input. The trim value of pitch control input is inches for level flight of the 14 knots flight speed model. Therefore, the pitch control input has a travel displacement of inches available to apply for load factor. The maximum load factor is then high as shown in the figure. az nz = g

70 Load factor, n z (g) 4 3 Load factor, n z (g) time (sec) -.2 8kts 1kts 14kts time (sec) Figure 2-25: Acceleration along z-axis on body frame following maximum pitch control input (pull) on the left and those following minimum control input (push) on the right In evaluation of the requirement for level two, the sufficient pitch control authority of the aircraft is checked. The pitch control input is required to be sufficient to accelerate flight speed from 45 knots up to maximum speed and decelerate from maximum speed to 45 knots. Since the pitch control input of the bare-airframe model was able to reduce flight speed down to 45 knots and increase flight speed up to approximately 18 knots level flight and the maximum speed was assumed to be 18 knots, the requirement for level two is achieved. A level two handling qualities was predicted.

71 Flight path control - Flight path response to pitch attitude (frontside), (ADS- 33E-PRF ) 53 This criterion has a requirement that the vertical rate response does not lag the pitch attitude response over the limit. The level one is given if the lag is no more than 45 degrees at frequencies below.4 radians per second. The bode diagram of vertical rate response and pitch attitude response are plot in Figure From the bode diagram, the phase of the vertical rate response and pitch angle response can be obtained. Equation 2.7 is used to determine phase lag of the vertical rate response. Results of the equation are shown in Figure Δphase θ δ long h& δ ( ω) = ( jω) ( jω) long 2.7

72 54 Bode Diagram Bode Diagram 35 h dot / δ long -5 θ / δ long 3-1 Magnitude (db) 25 2 Magnitude (db) Phase (deg) Phase (deg) Frequency (rad/sec) Frequency (rad/sec) Figure 2-26: Bode diagram of vertical rate response (left) and pitch attitude response (right) to longitudinal input in the 8 knots flight speed model Figure 2-27 shows that the phase difference (vertical rate lag pitch attitude) of the bare-airframe model is less than 45 degree in the region that satisfies the requirement of level one handling qualities. Handling qualities level one is achieved in this criterion.

73 kts 1kts 14kts Phase-difference limit Phase difference (deg) Level2 Level w (rad/sec) Figure 2-27: Handling quality determination for the bandwidth of flight path response to pitch attitude (frontside) Flight path control - Flight path response to collective controller (backside), (ADS-33E-PRF ) This criterion provides a requirement on vertical rate response following a step collective input. The response within five seconds of a step input is considered as a qualitative first-order appearance and the characteristics of that qualitative first-order appearance are used to rate the handling qualities. The determination method is similar

74 56 to criterion in Height Response Characteristics of Response to collective controller. The result of calculation is shown in Figure A degraded level in handling quality is given since the models cannot satisfy level two s requirement..35 ADS-33E.3.25 Level 2 τ h dot eq (sec) Level 1.5 8kts 1kts 14kts T h dot eq (sec) Figure 2-28: Handling quality determination for flight path response to collective controller (backside)

75 Longitudinal static stability, (ADS-33E-PRF 3.4.4) 57 This criterion has a requirement on the variation of longitudinal control stick force to change in speed. The force of longitudinal control stick can be treated as a proportion of its deflection because it normally varies as its displacement. The flight speed of the models increases (decrease) as the longitudinal control stick deviates in push (pull) direction from trim. The requirement of level one handling quality is satisfied Interaxis coupling - Pitch attitude due to collective control, (ADS-33E-PRF ) This criterion provides requirements on the ratio of peak pitch attitude change from trim, Δ θ peak, to peak of normal acceleration, Δ n z, peak, which occur within the first three seconds following specific step collective control inputs. Figure 2-29

76 58 2 θ (deg) Δ θ peak time (sec) Normal acceleration, n z (ft/s 2 ) Δ n z, peak time (sec) Figure 2-29: Responses of pitch attitude change and normal acceleration change of the 8 knots flight speed model following large corrective input in down direction Table 2-11 shows the ratio, Δθ Δ peak n z, peak, of the models following collective control inputs specified in the criterion. As shown it the table, the models are unable to satisfy the requirement. A degraded level is predicted on this criterion. Table 2-11: Handling Quality determination for the pitch attitude due to collective control Requirements for collective inputs Maximum limit of Model Δθ Δ 8kts 1kts 14kts peak n z, peak Small inputs Large inputs (up direction) Large inputs (down direction)

77 Interaxis coupling - Roll due to pitch coupling for Aggressive agility, (ADS- 33E-PRF ) 59 This criterion provides a requirement on the ratio of peak bank angle within four seconds to pitch attitude at four seconds following an abrupt longitudinal control step input, Δφ pk Δθ 4. As previously mentioned, this ratio is expected to be small. In Figure 2-3, one can see that the ratio, Δφ pk Δθ 4, is very small (less than.1) while the criterion allows at most.25 for level one. The Handling quality level one is achieved. 15 x θ from XA δ lat φ from XA δ lat 8 6 θ (deg) 5 φ (deg) 4 2 θ (deg) time φ (deg) time θ from XB δ φ from XB long δ long.1 8kts.5 1kts 14kts time time Figure 2-3: Pitch attitude and bank angle responses following lateral or longitudinal control inputs

78 Interaxis coupling - Pitch due to roll and roll due to pitch coupling for Target Acquisition and Tracking, (ADS-33E-PRF ) Figure 2-31 shows the results of the parameters calculated in this criterion comparing to the requirement limit. Handling qualities level one is achieved kts 1kts 14kts Level2 Level3 Average p/q (db) Level Average q/p (db) Figure 2-31: Handling quality determination for the pitch due to roll and roll due to pitch coupling for Target Acquisition and Tracking criterion

79 Roll attitude response to lateral controller - Small-amplitude roll attitude response to control inputs (bandwidth), (ADS-33E-PRF ) 61 This criterion provides requirements on the bandwidth ( ω BW, φ ) and phase delay ( τ p, φ ). Figure 2-32 shows that the models do not have a good bandwidth in roll attitude response. This is expected because the aircraft has high roll inertia. A Handling quality level three is predicted..25 8kts 1kts 14kts Level3.2 τ p,φ (sec).15 Level2.1.5 Level ω BW,φ (rad/sec) Figure 2-32: Handling quality determination for the bandwidth of roll response

80 Roll attitude response to lateral controller - Moderate-amplitude attitude changes (attitude quickness), (ADS-33E-PRF ) 62 This criterion provides requirements in the ratio of peak roll rate to change in roll attitude, p pk Δ φ pk. In order to achieve high value of this ratio aircraft must be able to generate high roll rate change in very short time. But actuators time constant and rate limit have an effect on the angular acceleration of roll response and cause degradation on handling qualities. Figure 2-33 shows result of handling qualities evaluation. A level three is satisfied Level1 8kts 1kts 14kts p pk /Δφ pk (1/sec) Level2 Level Minimum attitude change, Δφ min (deg) Figure 2-33: Handling Quality determination for the attitude quickness of roll response

81 Roll attitude response to lateral controller - Large-amplitude roll attitude changes, (ADS-33E-PRF ) 63 This criterion provides limits on the achievable roll rate and achievable bank angle to evaluate handling qualities for RCAH and ACAH control type respectively. The bare-airframe model was able to achieve maximum roll rate of 81.4 degree per second and generate roll attitude more than 9 degree. The values satisfied the requirement of level two handling qualities for target acquisition but the level one s requirement for other MTEs Roll attitude response to lateral controller - Linearity of roll response, (ADS- 33E-PRF ) This criterion requires no objectionable nonlinearities in the variation of roll response with roll control deflection. Since linear state space models of flight speed cases were used to represent the bare-airframe aircraft in this analysis, any requirements on linearity were automatically met. The nonlinear model needs to be analyzed to ensure that level one handling qualities are achieved Roll-sideslip coupling, (ADS-33E-PRF 3.4.7) This criterion provides a requirement on bank angle oscillation using of the parameter, φ / φ. The definitions of the parameters which are required in this OSC AV criterion are shown in Figure Figure 2-35 shows evaluation result of the models in this criterion. A level two is satisfied.

82 Figure 2-34: Definition of roll sideslip parameters [2] 64

83 kts 1kts 14kts φ OSC /φ AV LEVEL LEVEL ψ (deg) β Figure 2-35: Handling quality evaluation for the roll sideslip coupling criterion Yaw response to yaw controller - Small-amplitude yaw response for Target Acquisition and Tracking (bandwidth), (ADS-33E-PRF ) This criterion is mainly designed for an attitude command / attitude hold (ACAH) control type. It provides requirements on the bandwidth ( ω, ψ τ p, ) BW ) and phase delay ( ψ of the yaw response. Figure 2-36 shows the evaluation in this criterion. Handling quality level three is achieved.

84 kts 1kts 14kts.3 Level3 τ p,ψ (sec) Level2 Level ω BW,ψ (rad/sec) Figure 2-36: Handling quality evaluation for the bandwidth of yaw response Yaw response to yaw controller - Large-amplitude heading changes for Aggressive agility, (ADS-33E-PRF ) The heading change in one second following an abrupt maximum step displacement of pedal control that the models can achieve was degrees. The value does not satisfy the requirement for level three which requires four degrees in one second. A degraded level was predicted for the handling qualities in this criterion. This degradation might be a result of having high yaw inertia.

85 Yaw response to yaw controller - Linearity of directional response, (ADS-33E- PRF ) 67 This criterion requires no objectionable nonlinearities in the variation of directional response with yaw control deflection. As previously mentioned, linear state space models of flight speed cases were used to represent the bare-airframe aircraft in this analysis, any requirements on linearity were automatically met. The nonlinear model needs to be analyzed to ensure that level one handling qualities are achieved Lateral-directional stability - Lateral-directional oscillations, (ADS-33E-PRF ) This criterion limits the frequency and damping ratio of the dutch roll mode (lateral-directional oscillation) to evaluate the handling qualities. The positions of the dutch roll mode s eigenvalues are shown in Figure The oscillatory mode of the bare-airframe model is within the requirement of level three handling qualities for target acquisition and tracking. However, it satisfies level two if the requirement for all other MTEs is used Lateral-directional stability - Spiral stability, (ADS-33E-PRF ) This criterion provides requirements on time to double amplitude of the bank angle following a lateral pulse control inputs. The specified values in this criterion are supposed to apply to an exponential divergence and do not depend on size of control

86 inputs. Figure 2-21 shows spiral mode is stable. There will be no time to double amplitude. A level one handling qualities is predicted in this criterion Lateral-directional characteristics in steady sideslips - Yaw control in steady sideslips (directional stability), (ADS-33E-PRF ) In this criterion, the aircraft is required to perform yaw-control-induced steady zero-yaw rate sideslips with the rotorcraft trimmed for straight and level fight. In order to perform this flight condition in offline analysis, a control input vector, u v, that led the aircraft to the flight condition was needed to find. The solution was calculated from linear state space equation. Some parameters in the state equation could be found from the flight condition (e.g. steady flight indicates that &v v x =, zero-yaw rate indicates that r =, etc.). Since the criterion has no requirement on rotor states, the state space models were modified to have only aircraft motion states. The aircraft states, x v, could be separated into aircraft motion states, x v A, and rotors states, x v R as in Equation 2.8. v x&v A A x B A 1A 1R A v u x&v A A 1 = v + A R x R B 2.8 R Because of steady state flight condition, &v v x =, the rotor states, x v R, could be put into a form of aircraft motion states, x v A, and input, u v, as in Equation 2.9. v x R 1 v 1 v = A2 R A2 A x A A2 R B2 u 2.9

87 69 The state space models could be rewritten as the rotors state equations of Equation 2.9 was substituted into the state space equation (Equation 2.8). The new state space model with no rotor states became as Equation 2.1. x&v x&v x&v A A A = A 1 A v x + A = ( A1 A A1 R A v v = Aˆ x + Bˆ u A A 1R ( A 1 2 R A 1 2 R 2 A A v ) x 2 A A v 1 v v x A + A2 R B2 u ) + B1 u 1 v + ( A A B + B ) u 1R 2 R Plots in Figure 2-37 and 2-38 are the results from solving Equation 2.1. Time history plots of aircraft responses following one of the results are shown in Figure Lateral stick displacement (inches) Pedal input displacement (inches) β (deg) kts -15 1kts 14kts Pedal input displacement (inches) Figure 2-37: The solution of the lateral input and sideslip angle following induced pedal control inputs

88 7 Longitudinal stick displacement (inches) 1 x Pedal input displacement (inches) φ (deg) kts 1kts 14kts Pedal input displacement (inches) Figure 2-38: The solution of the longitudinal input and bank angle following induced pedal control inputs Figure 2-39 shows that the bare-airframe model can perform the specific flight condition following the calculated input. The heading change, ψ, and bank angle, φ, responses indicate straight steady zero-yaw-rate of the sideslip while the flight path, γ, indicates level flight. The sideslip flight condition is achieved.

89 71 φ (deg) α (deg) x kts 1kts 14kts time (sec) time (sec) ψ (deg) 4 2 β (deg) time (sec) time (sec) Figure 2-39: Models responses following input from the solution of Equation 2.1 and induced pedal input = 2.5 inches The plot in Figure 2-37 shows that the models satisfy the criterion s requirement. The right yaw control deflections ( δ > ) are required in left sideslips ( β < ). Handling quality level one is satisfied. ped Lateral-directional characteristics in steady sideslips - Bank angle in steady sideslips, (ADS-33E-PRF ) This criterion has a requirement for aircraft in yaw-control-induced steady zeroyaw rate sideslips. It requires that an increase in right bank angle accompanies an

90 72 increase in right sideslip and an increase in left bank angle accompanies an increase in left sideslip. Figure 2-37 and 2-38 show that the requirement on the relation between bank angle and sideslip angle is satisfied. Handling quality level one is predicted Lateral-directional characteristics in steady sideslips - Lateral control in steady sideslips, (ADS-33E-PRF ) This criterion has a requirement in the need of lateral control inputs for yawcontrol-induced steady zero-yaw rate sideslips. It is required that right sideslip shall not require left lateral control input and vice versa. Figure 2-37 shows that right sideslips requires left lateral control inputs and left sideslips requires right lateral control inputs in the models of 8 knots flight speed and 1 knots flight speed. Since the requirement is not satisfied, a degraded level in handling qualities is predicted in this criterion Airplane Mode Figure 2-4 shows all eigenvalues of linearization models. Eigenvalues corresponding to actuator modes are not shown. Figure 2-41 shows the eigenvalues as same as Figure 2-4 but zooms in to low frequency region. Since all eigenvalues corresponding to aircraft motion modes are in left haft plane, the bare-airframe dynamics is stable.

91 Short period mode Imaginary axis kts 2kts 24kts Dutch roll mode Rotors gimbal modes Real axis Figure 2-4: Eigenvalues of models of FXV-15 in airplane mode

92 74 Imaginary axis Roll mode kts Spiral mode Phugoid mode 2kts.96 24kts Real axis Figure 2-41: Eigenvalues of models of FXV-15 in airplane mode in low frequencies region Longitudinal stability with respect to speed - Phugoid stability, (MIL-F-8785C ) This criterion provides requirements on the long period oscillations when the aircraft seeks new stabilized state following a disturbance. Figure 2-41 shows that the bare-airframe models have good damp in phugoid mode. The requirement for level one handling quality is satisfied.

93 Longitudinal stability with respect to speed - Flight-path stability, (MIL-F- 8785C ) This criterion provides requirements for aircraft in the landing approach flight phase. The requirement is in terms of flight-path-angle change when the airspeed is changed following the pitch control. Since linearized models of high-speed, level flight cases were used to represent the bare-airframe model in airplane mode, there was no FXV-15 airplane mode model which corresponds to landing approach or low speed flight available. This criterion was not able to be determined Longitudinal maneuvering characteristics - Short-period response, (MIL-F- 8785C ) This criterion provides requirements on Short-period frequency and acceleration sensitivity, and on Short-period damping ratio. The short-period frequency, ω, and short-period damping ratio, ζ n /α, is calculated as in Equation SP, are found in Figure 2-4. The acceleration sensitivity, SP n = α 1 Z = m g α 1 Z 1 w = Z m w g α w 1 V g 2.11 Figure 2-42 shows that bare-airframe model satisfies level one handling quality on Short-period frequency and acceleration sensitivity requirement. But the Short-period damping ratios achieve only level two. Therefore, the handling qualities rating on this criterion is level two.

94 kts 2kts 24kts ω nsp (rad/sec) LEVEL 2 LEVEL 2 LEVEL 1 LEVEL 1 LEVEL 2 & 3 LEVEL n/α (g/rad) Figure 2-42: Short-period frequency requirements for Category A Flight Phases Longitudinal control - Longitudinal control in maneuvering flight, (MIL-F- 8785C ) This criterion provides requirement on the service load factors. The service load factor is determined using the acceleration along z-axis on aircraft body frame that develop by use of the pitch control alone. The load factor, n z, is defined in Equation 2.6. In this analysis, the service load factor for the airplane mode was defined in a range of maximum of 2.5g to minimum of g. This service load factor was used for handling qualities evaluation in this criterion.

95 load factor, n z (g) load factor, n z (g) kts 1kts 14kts time (sec) time (sec) Figure 2-43: Acceleration along z-axis on body frame following maximum pitch control input (pull on the left and push on the right) Figure 2-43 shows time history of the load factor. The bare-airframe model is able to achieve service load factors. The handling qualities rating on this criterion is level one Lateral-directional mode characteristics - Lateral-directional oscillations (Dutch-roll), (MIL-F-8785C ) This criterion provides requirements on the frequency and damping ratio of the lateral-directional oscillations, ω nd and ζ d. These parameters can be obtained from

96 Figure 2-4. The bare-airframe model satisfies the requirement of level one of handling qualities in this criterion Lateral-directional mode characteristics - Roll mode, (MIL-F-8785C ) This criterion rates handling quality by using the roll mode time constant, τ R. The time constant is calculated by using Equation where σ is the eigenvalue corresponding to the roll mode. The roll-mode eigenvalue can be found in Figure The time constants of roll mode required for level one handling qualities is 1.4 second. The bare-airframe model satisfies the requirement of level one. 1 τ R = where σ < and σ R 2.12 σ Lateral-directional mode characteristics - Spiral stability, (MIL-F-8785C ) This criterion rates Handling Quality by using time to double amplitude following a disturbance in bank of up to 2 degree. Figure 2-44 shows that responses following a disturbance in bank angle return to initial states. The Spiral mode is stable. The requirement for level one handling quality is satisfied.

97 79 bank angle (deg) time (sec) 5 16kts 2kts 24kts p (deg/s) ψ (deg) time (sec) time (sec) Figure 2-44: Responses following a disturbance in bank angle Lateral-directional mode characteristics - Coupled roll-spiral oscillation, (MIL- F-8785C ) This criterion requires that the airplane characteristics of aircraft in Category A do not exhibit a coupled roll-spiral mode in response to the pilot lateral control command. Figure 2-41 shows that roll mode and spiral mode do not couple because both have a real valued eigenvalue. The bare-airframe model satisfies a level one on this criterion.

98 Lateral-directional dynamic response characteristics - Roll rate oscillations, (MIL-F-8785C ) 8 This criterion provides requirement on the roll rate response, p, following a step lateral control command with yaw control free. The criterion imposes limits on the first the first peak rate. Figure 2-45 shows that roll rate oscillations is well damped. There is no minimum roll rate. The bare-airframe model satisfies handling quality level one on this criterion p (deg/s) kts 2 2kts 24kts time (sec) Figure 2-45: Roll rate response following a yaw-control-free step roll control command

99 Lateral-directional dynamic response characteristics - Bank angle oscillations, (MIL-F-8785C ) 81 This criterion provides the requirement on the parameter φ / φ following an abrupt impulse command in roll. The definitions of the parameters which are required in this criterion are shown in Figure Figure 2-46 shows the evaluation result of the bare-airframe model. The bare-airframe model meets handling quality level one for this criterion. OSC AV kts 2kts 24kts φ OSC /φ AV LEVEL 2 LEVEL ψ (deg) β Figure 2-46: Handling quality evaluation of bank angle oscillations criterion

100 Pilot-induced oscillations, (MIL-F-8785C 3.3.3) 82 This criterion demands no tendency of sustained or uncontrollable lateraldirectional oscillations resulting from the pilot s control efforts. Figure 2-47 displays the bode diagram of the bare-airframe models. Only bode diagrams of angular rate responses following corresponding control inputs are considered because a rate response is expected from an airplane control surface. Shown in Figure 2-47, pilot-induced oscillations may occur from yaw control efforts because the bandwidth is in human operation frequency. p () s Bode Diagram From: XA δ lat To: Roll rate -2 Bode Diagram From: XB δ long To: Pitch rate -2 q () s -2 r () s Bode Diagram From: XP δ ped To: Yaw rate Magnitude (db) Magnitude (db) -4-6 Magnitude (db) Phase (deg) -9 Phase (deg) -9 Phase (deg) Frequency (rad/sec) Frequency (rad/sec) Frequency (rad/sec) Figure 2-47: Bode diagrams of bare-airframe models

101 Roll control effectiveness, (MIL-F-8785C 3.3.4) 83 This criterion provides a requirement on the time to achieve a required bank angle change following an abrupt step lateral command. A category A aircraft is required to achieve 45 deg within 1.4, 1.9 and 2.8 seconds to satisfy handling quality level one, two and three respectively. Figure 2-48 shows that the bare-airframe model does not have good roll control effectiveness and satisfy only level three kts 2kts 24kts 12 Bank angle (deg) X: 1.47 Y: 45.7 X: Y: 45.1 X: 1.96 Y: time (sec) Figure 2-48: Bank angle response following the maximum lateral input.

102 Roll control effectiveness - Linearity of roll response, (MIL-F-8785C ) 84 This criterion requires no objectionable nonlinearities in the variation of roll response with roll control deflection or force. Control force is normally proportional to control deflection. It is reasonable to consider only a variation of roll response to roll control deflection for the requirement. Since the analysis used linearized XV-15 simulation model, a linear responses is certainly expected. Further analysis would need to be performed using a nonlinear model Directional control characteristics, (MIL-F-8785C 3.3.5) This is a subjective requirement based on pilot s opinion of yaw sensitivity to pedal force. It cannot be evaluated with off-line linear model simulations. The bareairframe model is not evaluated on this criterion Lateral-directional characteristics in steady sideslips - Yawing moments in steady sideslips, (MIL-F-8785C ) This requirement focuses on the characteristics in yaw-control induced steady zero-yaw-rate sideslips. The procedure to calculate those characteristics was previously introduced in Yaw control in steady sideslips (directional stability). Figure 2-49 to 2-51 show the results from solving the Equation 2.1 by applying the linear state space models of the FXV-15 in conversion mode. Figure 2-49 shows that the sideslip angle is limited by full pedal control deflection and the right yaw-control

103 85 deflections accompany left sideslip. The figure also demonstrates a need of left lateral control displacements to perform left sideslips. Equation 2.13 was used to calculate a sideslip angle, β. v yb β = arcsin vxb + v yb + vzb Lat. input displacement (inches) Pedal input displacement (inches) Long. input displacement (inches) Pedal input displacement (inches) 4 1 φ (deg) Pedal input displacement (inches) β (deg) -1 16kts -2 2kts 24kts Pedal input displacement (inches) Figure 2-49: The solution of responses following induced pedal control inputs Figures 2-49 and 2-5 show the linearity characteristics of linear state space model. The aircraft s states ( φ, v xb, v yb, vzb ) and control parameters ( δ lat, δ long ) linearly vary to the pedal input. The plot of steady flight path to the pedal control confirms the

104 86 level flight condition. The change of body velocity on x-axis is significant. In 24 knots flight speed model, the solution gave the body velocity on x-axis drops below zero in performing a steady sideslip with pedal input displacement about two inches. A negative value of body velocity on x-axis indicates an abnormal flight condition; therefore, the graphs do not show the data of these flight conditions. The nonlinearity in sideslip to pedal input can be explained by the significant change of the body velocity on x-axis to yaw control input in performing steady sideslips. In a right sideslip (negative pedals displacement), the aircraft simulation shows the increasing in body velocity on x-axis. However the velocity is shown to be decreased while a left sideslip is performed (positive pedals displacement). When this comes to a sideslip angle calculation using Equation 2.13, the denominator of the equation is higher in right sideslip while it is smaller in left sideslip. This result in nonlinearity in sideslip to pedals input (achieve higher sideslip angle in left sideslip than right sideslip). Considering a symmetric aircraft, the responses in longitudinal motion mode θ,v, v and q ) should be similar (or very close) whether left or right yaw-control ( xb zb perturbation is applied to the aircraft in trim straight level-flight. These linear models might not be appropriate to be utilized in the evaluation of handling qualities in sideslip criterion.

105 v xb (ft/s) 5 v yb (ft/s) Pedal input displacement (inches) Pedal input displacement (inches) v zb (ft/s) γ (deg) kts 2kts 24kts Pedal input displacement (inches) Pedal input displacement (inches) Figure 2-5: The solution of body velocity responses following pedal control inputs Figure 2-51 shows responses following yaw control inputs. The body velocity in x-axis dramatically decreases and becomes less than zero when a sideslip is performed by applying yaw control input of two inches.

106 v xb (ft/s) 5 θ (deg) 1 β (deg) time (sec) 2-2 γ (deg) δ ped =-2-1 δ 1 2 ped =-1 time (sec) 3 4 δ ped =1 1 δ ped = time (sec) time (sec) Figure 2-51: Models responses following input from the solution of Equation 2.1 The results of sideslip angles in both sides are questionable, especially in the case of 24 knots flight speed that body velocity on x-axis vary from negative values to approximately 9 feet per second. Even right sideslip angles seem to be more linear to left yaw control input. An aircraft should not fly a lot faster while performing a steady, level sideslip. Therefore, a nonlinear model is needed in verification of the results. In evaluation of this criterion, the plot in Figure 2-49 shows that the bare-airframe model satisfies the criterion s requirement. The right yaw control deflections ( δ > ) produce left sideslips ( β < ). The handling quality level one is achieved. ped

107 Lateral-directional characteristics in steady sideslips - Side forces in steady sideslips, (MIL-F-8785C ) 89 This criterion provides a requirement for aircraft in yaw-control-induced steady zero-yaw rate sideslips. It is required that an increase in right bank angle accompanies an increase in right sideslip and an increase in left bank angle accompanies an increase in left sideslip. Figure 2-49 shows that the requirement on the relation between bank angle and sideslip angle is satisfied. Level one handling qualities requirement is satisfied Lateral-directional characteristics in steady sideslips - Rolling moments in steady sideslips, (MIL-F-8785C ) This criterion defines requirements on the variation in lateral control input for yaw-control-induced steady zero-yaw rate sideslips. The requirement states that left roll control deflection shall accompany left sideslip and right roll control deflection shall accompany right sideslip. Figure 2-49 shows that the requirement for levels 1&2 is satisfied. The handling quality levels 1&2 is rated from this criterion Lateral-directional characteristics in steady sideslips - Positive effective dihedral limit, (MIL-F-8785C ) This criterion requires that the roll control power used to perform sideslip is never more than 75 percent of roll control power available to the pilot. In analysis of the simplified models, the roll control power that available to the pilot can be indicated by travel range of lateral control surface (aileron). The control surface is set to move

108 9 following the lateral stick displacement in linear manner and the gain in the control mixing is scaled so that the maximum displacement move the control surface to maximum position. Therefore, available roll control power in these simplified models could be measured by the range of the current position to the end of travel range of the lateral stick. Figure 2-49 shows that about 8 percents of lateral stick displacement is used in order to perform maximum sideslip angle. The requirements of levels 1&2 are not satisfied. A degraded level is predicted. 2.5 Summary of Predicted Handling Qualities of FXV-15 Bare-airframe models A major problem in helicopter mode is instability of the models. Many criteria give the handling qualities rating of degraded level (below level three). The aircraft has an advantage of being symmetric and therefore has low cross coupling, but instability will cause a lot of pilot work load and would likely result in poor ratings. In conversion mode, the strength of a symmetric aircraft is also displayed, however, the FXV-15 has disadvantages in criteria of large heading change and yaw rate. The models in airplane mode have stability and are able to achieve the best for handling quality requirement in many criteria. However, there are some criteria that the models get poor ratings. Thus, the overall all handling qualities of the FXV-15 bare-airframe model shall be level three.

109 Modification of the Bare-Airframe Model One can see from the results previously shown in this chapter that the aircraft is required to test in several operating conditions. Some of these operating points deviate from the equilibrium point of the provided model significantly. Therefore, the existing linear model is likely to fail in representing the aircraft dynamics for large amplitude maneuvers. Since the provided model is the only available source of the aircraft dynamics, the state space model (Equation 2.2) is modified. Nonlinear terms for the Euler attitude kinematics, gravitational terms, and Coriolis Effect terms were added to the equations of motion. The quasi-linear model could be represented by the Equation The Euler attitude kinematics, gravitational matrix, Vz Vy Vx G,,, added gravitational effects to aircraft velocities as defined in Equation 2.15 u D x C y x N x G u B x A x Vz Vy Vx + = = ) ( ) (,,,, ψ θ φ & 2.14 = cos cos cos sin sin,, θ φ θ φ φ g g g G Vz Vy Vx,where = r q p v v v R R R R x z y x ψ θ φ 2.15

110 The Coriolis Effect matrix, N φ, θ, ψ change. The new equations are calculated as in Equation , changed the equation of aircraft attitudes & φ = p + ( q sin( φ) + r cos( φ)) tan( θ ) & θ = q cos( φ) r sin( φ) q sin( φ) + r cos( φ) ψ& = cos( θ ) 2.16 The quasi-linear model was used mainly for the controller design and evaluation of FXV-15 after this point.

111 Chapter 3 Model Inversion Controller Designing a controller that achieves the most stringent (level one) handling qualities requirements specified in ADS-33E and MIL-F-8785C, and desired responses in all speed ranges presents a major challenge. The model following / model inversion control s design scheme is well suited to solve this problem. The controller is emphasized by its concept for achieving consistent response attributes [7]. The ideal response models or command filters are designed to meet the specified response characteristics (e.g. response type, small / moderate / large amplitude response). The feedback compensation can be tuned to meet disturbance rejection and cross-coupling requirements. The inverse model can be scheduled to account for changing aircraft dynamics, and thus the other components of the controller can be fixed or tailored to change response types for different operating points. 3.1 Model Inversion Controller The diagram of this control method is illustrated in Figure 3-1. The key element of this technique is to find an approximated transfer function, fˆ, of the physical system in order to use its inverse, f ˆ 1,to counteract the system itself. If the approximated transfer function, fˆ, perfectly describes the physical system, then f ˆ 1 f = Ι. Therefore, dynamics will behave as simple integrators by using the inverse. The only

112 thing needed to be designed is a simple controller to match desired handling performances. 94 Figure 3-1: Model inversion control concept diagram. In reality, the flight dynamics system of an aircraft significantly changes for each different operating point. It implies that the transfer function, f, representing the aircraft dynamics change. Therefore, the inverse transfer function, f ˆ 1, of each operating point is required in order to schedule as the aircraft changes its operating point. Besides the changing in flight dynamics, the constraints of aircraft actuators also need to be considered if this technique is applied. Assuming the aircraft structure can withstand any forces and moments that control surfaces create to get desired response, the aircraft dynamics are still limited by the control surfaces range and rate limits. These limits cause nonlinearity to the transfer function, f, at each operating point. These limits can also cause insufficient performance in tracking some of desired responses.

113 95 Figure 3-2: Schematic of model inversion controller research. Figure 3-2 illustrates the schematic of model inversion controller used in this 3.2 Design Overview One of the first things to be considered when designing a flight control system would be the response type. The response type can play an important role in aiding the pilot in performing the tasks. For the tiltrotor aircraft, suitable response characteristic for each of tiltrotor configuration are considered. The response characteristic of the aircraft is chosen from two common types for aircraft referred to as Attitude Command / Attitude Hold (ACAH) type and Rate Command / Attitude Hold (RCAH) type. The ACAH type controller is designed to change the aircraft attitude proportional to the cockpit input displacement and hold at that point while the RCAH type controller lets the pilot uses cockpit input displacement to control rate of attitude change and holds aircraft attitude when command input return to detent.

114 96 In helicopter mode and conversion mode, the attitude command / attitude hold (ACAH) type is chosen for lateral and longitudinal control to control bank and pitch angle respectively. The lateral stick displacement of -4.8 inches (left deflection) up to 4.8 inches (right deflection) commands the roll attitude from -9º (bank left) to 9º (bank right) linearly. While longitudinal input commands the pitch attitude -35º (pitch down) to 35º (pitch up) in the same manner. The rate command / attitude hold (RCAH) type is selected for pedal input to control yaw rate. The pedal input displacement of -2.5 inches (apply left pedals) up to 2.5 inches (apply right pedals) is designed to command yaw rate from -6 (left turn) to 6 (right turn) degree per second linearly. The collective input is decided to be an open loop control for rotor collective angle and engine power. Existing control as that of bare-airframe model is used. No controller is designed for collective control. This control characteristic is similar to that in conventional helicopters. In this research, it will be mention as a low speed control after this point. In airplane mode, the rate Command / Attitude Hold (RCAH) type controller is selected for lateral, longitudinal and pedal command inputs to control roll rate, pitch rate and yaw rate of the aircraft respectively. The roll rate is designed to be about 69 degree per second when the maximum lateral input is applied and to linearly vary with the input. For the pitch and yaw axes, the maximum achieved angular rate is determined for particular operating point by turn coordination mode. The Collective lever controls rotor collective angle and engine power. This control characteristic is similar to that in general fixed-wing airplane. It will be called the high speed control. The switching in the use of a low speed control and a high speed control is determined by using the aircraft nacelle angles. In this design, a low speed control will

115 97 be taken place by a high speed control if the nacelle angles reduced below 45º and once it become a high speed mode the switching back will happen if the nacelle angles increase over 55º. For the inversion model, since data of the dynamics of XV-15 is limited only to the model provided by the University of Liverpool, the aircraft is assumed to operate in -to-6-knots speed range with the helicopter mode only, in 6-to-15-knots speed range only with the conversion mode and 6º nacelle angles and fly with speed higher than 15 knots in the airplane configuration only. Turn coordination mode is designed to start activation when airspeed is over 45 knots and become full activation when airspeed is over 6 knots. When this mode activates, it allows pilot to perform a coordinated turn by using the lateral stick input only. The pedals, in this mode, are designed to control lateral acceleration (which is zero in coordinated turn when pedals are in detent). 3.3 Desired Response For helicopter and conversion configuration, roll and pitch axes were designed to hold the aircraft attitude in the position whereas the lateral and longitudinal inputs command. The bank angle and pitch angle were specified as command inputs. For this ACAH type control, a second-order command filter was used to obtain desired angular rates and their accelerations. ( φ φ ) = & 2 φ c + 2ζ ω & φ φφc + ωφ c cmd 3.1

116 The Equation 3.1 was used for the roll command filter where φ cmd 98 is the input, ωφ and ζ φ are desired natural frequency and corresponded damping ratio of the roll response, and & φ c, & φ c and φ c are the output of the filter. ( θ θ ) = & 2 θ 2ζ ω & θ + ω 3.2 c + θ θ c θ c cmd ωθ and The Equation 3.2 was used for the pitch command filter where θ cmd is the input, ζ θ are the desired natural frequency and damping ratio of the pitch response, and & θ c, & θ c and θ c are the output of the filter. For the airplane configuration, a rate command attitude hold type was selected for roll and pitch axes control. The roll rate, p, and pitch rate, q, were specified as command inputs. A first-order command filter was designed to obtain desired angular rate and their accelerations. ( p p ) = τ p& 3.3 p c + c cmd The Equation 3.3 was used for the roll command filter for the airplane configuration where p cmd is the input, τ p is the desired time constant of the roll rate response, and p& c and p c are the output of the filter. ( q q ) = τ q& 3.4 q c + c cmd The Equation 3.4 was used for the pitch command filter for the airplane configuration where q cmd response, and q& c and q c are the output of the filter. is the input, τ q is the desired time constant of the pitch rate

117 For yaw axis which was designed as rate command attitude hold (RCAH), yaw rate was specified as the command input. A first-order command filter was used to obtain desired angular rate and their accelerations. ( r r ) = τ r& 3.5 r c + c cmd Equation 3.5 was used for the yaw command filter where r cmd 99 is the input, τ r is the desired time constant of the yaw rate response, and r& c and r c are the output of the filter. 3.4 Inversion model An inversion model is used to cancel the aircraft dynamics and cause the aircraft dynamics to appear as a simple integrator to the controller. The inversion model varies for different operating points and is scheduled with airspeed. The aircraft models used for the inversion model in this study were the linear models provided by the University of Liverpool. Therefore, the inversion model was scheduled with airspeed of, 2, 4, 8, 1, 14, 16, 2 or 24 knots, with nacelle angles 9, 9, 9, 6, 6, 6,,, respectively. To find inverses of the given model, a state space model representing bareairframe dynamics was required. This state space model would represent control surface gearing ratios and actuators dynamics, and the rate and range limits in control mixing as well. However, the actuators dynamics, rate limits and range limits were treated as gains in order to simplify the state space model. The simplified version of the control mixing

118 1 for particular airspeed, B CM ( V T ), could be written in a gain matrix as shown in Equation 3.6. A gain matrix of control mixing for specific airspeed could be found from Table 2-3. Then the given state space model for particular airspeed in Equation 3.7 could be modified by adding information on control mixing in Equation 3.6 in order to take cockpit control inputs, δ C, as an input vector. The complete model of each flight speed cases could be written as in Equation 3.8. This model had 13 states in the state vector, x v, and four inputs in the cockpit input vector, v δc. After that, a 13 th -order model of the aircraft and its rotors was reduced to a 9 th - order model of only the aircraft states. Since the rotors state responses are much faster than the aircraft s rigid body states and are stable, they are assumed to reach equilibrium quickly, x& v =, while aircraft s states are changing. From, a model calculated in the r v u = B CM ( V T ) v δ C 3.6 x&v v v = A ( V ) x + B ( V ) u 3.7 UOL UOL Equation 3.8, its aircraft s and rotors states can separate as in the Equation 3.9. T UOL ( VT ) x + BUOL ( VT ) BCM ( VT ) C T x&v v = A v δ 3.8 x&v a A x&v = A r A A v xa B v + x r B 1 2 v δ C 3.9

119 11 Equations 3.1 and 3.11 show the calculation to find new aircraft state matrix, [ ] a A, and new aircraft control matrix, [ ] a B, by substituting = r x& v. Finally, a 9 th -order model of the aircraft was reduced to a simple third-order model of angular rate dynamics which had lateral, longitudinal and pedal command inputs as control inputs, and had only the states of roll rate, p, pitch rate, q, and yaw rate, r. The linear model coefficients corresponding to these states and controls were drawn out of the result from Equation 3.11 to build up a third-order model in Equation By neglecting terms related to attitudes and velocities, errors in desired responses are expected. An inverse of a reduced model were found by the Equation In this research, the inversion model used was that closest to the airspeed. Linear interpolation was used for smooth transition between two operating points. C a r B A x A A x δ v v v = C a a B A A B x A A A A x δ v v &v + = ) ( ) ( [ ] [ ] + = ped long lat B r q p A r q p δ δ δ & & & 3.12 [ ] [ ] = r q p A r q p B ped long lat & & & 1 δ δ δ 3.13

120 3.5 Dynamics Compensator 12 The dynamics compensator was used to track the difference between the commanded inputs, the desired responses ( φ c, θ c, p c, q c and r c ), and the current responses of the aircraft, the measured responses (φ, θ, p, q and r ),. The tracking errors for helicopter and conversion configuration were defined as Equation ~ φ = φc φ ~ θ = θc θ ~ r = r r c 3.14 The Equation 3.15 defines tracking errors for airplane configuration. ~ p = pc p q~ = qc q ~ r = r r c 3.15 For the ACAH type, a proportional derivative (PD) compensator was used to minimize the tracking errors. A proportional integral (PI) compensator was used for RCAH type control. && φ && K D φc = + && θ && D θ c K r& r D & c K Pφ Pθ P r ~ ~ & φ + K Dφφ ~ ~ & θ + K Dθθ ~ r + K ~ I rdt r 3.16 Then the pseudo-commands ( & φ D, & θ D and r& D ) representing desired angular acceleration can be computed for helicopter and conversion mode control as defined in Equation While the pseudo-commands ( p& D, q& D and r& D ) representing desired rate of change of angular rate for airplane mode control can be found as Equation 3.17.

121 p& q& r& D D D p& = q& r& c c c K + K K P p P q P r ~ p + q~ + ~ r + K K K I I q I r ~ pdt p qdt ~ ~ rdt The PD and PI compensator gains were initially chosen to match error dynamics and desired responses from command filter, and later on tuned for matching the best handling qualities of all criteria and maintaining stability of the model all the time. 3.6 Euler Angle Conversion The pseudo-command vector of the low speed controller given in Equation 3.16 represents the desired Euler angle accelerations. However, the inversion model requires the pseudo-commands to represent angular acceleration in body coordinate frame as in Equation An Euler angle conversion was used to obtain the desired angular rate as defined in Equations 3.18 and p & D && θ sin φ sinθ + & sin θ + ψθ & & cosφ D rd θφ cosφ sin θ ψφ sin φ sin θ cosθ φ + & & + & & = 3.18 && D cosφ cosθ (&& θ & sin θ ψφ & & D + rd + cos θ ) q & D = 3.19 cosφ In the high speed controller, no Euler angle conversion is required because the pseudo-command vector matches the input vector of the inversion model.

122 3.7 Turn Coordination 14 In hover and low speed flight (below 45 knots), the aircraft s roll attitude is controlled by lateral command input, which the pilot uses to control lateral velocity. The pilot can perform sideward flight or change lateral velocity through the control of roll attitude and use the pedal input to control heading independently. However, for an aircraft at higher speed, its lateral command input is expected to roll the aircraft in order to enter a coordinated turn in which the rate of heading change depends on the bank angle and airspeed. The controller was designed to perform a coordinated turn with an input of lateral command. In this mode, the pedal input is used to command the lateral acceleration, which is zero in a coordinated turn. In the case that the pilot would like to perform a sideslip flight or an uncoordinated turn then a pedal input could be applied; otherwise, if pedals are in detent the flight will be coordinated. This feature was added to the controller as an outer loop feedback control to calculate the yaw rate command needed to regulate the lateral acceleration as shown in Equation 3.2 [7, 16]. r cmd ( K a + w p + g sinφ cosθ ) u = 3.2 TCr y, cmd The turn coordination mode was implemented for airspeeds above 6 knots while the yaw rate command of low speed mode was preserved for airspeeds below 45 knots. In the transition, airspeeds between 45 and 6 knots, the yaw rate command was interpolated between the direct yaw rate command of the low speed mode and the turn coordination mode.

123 15 A turn coordination mode was also designed for the longitudinal axis to regulate the pitch rate to achieve a desired normal load factor in turns. With the turn coordination mode in pitch axis activated, the longitudinal input was changed to command vertical acceleration, a z, cmd, in the aircraft body frame. Equation 3.21 and 3.22 are used to compute necessary pitch rate to regulate the commanded vertical acceleration as an outer loop feedback control of pitch axis control. q cmd ( K a + g C ) u =, 3.21 TCq z cmd 1 C TC = 1+ where C TC cosφ As mentioned in the overview, the turn coordination mode in pitch axis was designed to activate when a low speed control is changed to a high speed control. The same method providing a smooth transition was used. TC 3.8 Mode Switching The mode switching between a low speed control and a high speed control was performed by a relay. A relay was set as a logic switch to give a one to indicate the use of a high speed control and give a zero when a low speed control is used. To avoid that pilots might encounter a sudden change of control mode, the transfer function defined in Equation 3.23 was used to delay effect of new control mode and get a smooth transition. C = 1 2 s MS x RELAY

124 16 During the transition, the effect of previous control mode is faded out smoothly. In the same time, the pilot will start to experience an effect of the new mode. Equation 3.24 was used to linear interpolate the commands between these two modes. r u D v u C v u = D, low (1 MS ) + D, high C MS 3.24 Figure 3-3 shows the diagram of the mode switching. The pseudo-command of yaw control, r& D, is not shown in the diagram because the mode of control for yaw rate control does not change (only the turn coordination mode which is in an outer loop does). Figure 3-3: Mode switching diagram

125 3.9 The Complete Controller of FXV-15 models 17 In the present work, all control parameters were selected and tuned so aircraft come as close as possible to level one handling qualities. Table 3-1 displays the final valued which satisfied level one handling qualities in most criteria. Table 3-1: Type Caption Here Parameter Designed Value Parameter Designed Value ζ.7 φ K Ip 4. ω 2. φ K Pθ 16. ζ.7 θ K Dθ 6.4 ω 2. θ K Pq 3.61 τ.25 p K Iq 3.61 τ.4 q K Pr 7.2 τ.25 r K Ir 16 K 4. Pφ K TCr.32g K 3.6 Dφ K TCq.521g K Pp 3.6

126 Chapter 4 Predicted Handling Qualities of the Augmented Aircraft In this chapter, the controller developed in the previous chapter will be evaluated using the quasi-linear model of the FXV-15 developed in chapter two. For handling qualities evaluation, the controller and aircraft model were built and implemented using MATLAB / SIMULINK. The criteria to be applied were the same as those specified in the chapter two. Only the handling qualities level result is shown for each criterion. The discussion is made at the end of the chapter. Figure 4-1 displays the overall controller, Figure 4-2 illustrates the control mixing model, and Figure 4-3 shows the quasi-linear aircraft model in MATLAB / SIMULINK.

127 Figure 4-1: The controller in MATLAB / SIMULINK 19

128 Figure 4-2: The control mixing model in MATLAB / SIMULINK 11

129 Figure 4-3: The quasi-linear aircraft model in MATLAB / SIMULINK. 111

130 4.1 Results 112 Tables 4-1, 4-5 and 4-11 display predicted handling qualities results of the XV-15 simulation model with the model inversion control laws in helicopter mode, conversion mode and airplane mode respectively. In these tables, the number shown in the Level column is the level that was predicted for a particular criterion. If the asterisk is shown, it indicates that that level was evaluated using a less stringent requirement, and otherwise would not achieve level one. For example, for an ADS-33E-PRF requirement, the criterion for target acquisition and tracking may not be achieved, so the criterion for all other MTE s is evaluated. The term PASS is used when a criterion does not evaluate into a level. The term DEGRADED is used when none of requirements in that criterion is achieved.

131 4.1.1 Helicopter Configuration 113 Table 4-1 shows results of predicted handling qualities of augmented FXV-15 in helicopter configuration. There are only two assigned criteria for which the controller was unable to deliver level one handling qualities. Table 4-1: Predicted handling qualities of XV-15 simulation in helicopter mode Handling Qualities Criteria Level Short-term response to control inputs (bandwidth) Mid-term response to control inputs Moderate-amplitude pitch(roll) attitude changes (attitude quickness) 2* Large-amplitude pitch(roll) attitude changes Small-amplitude yaw attitude changes Mid-term response to control inputs Moderate-amplitude heading changes (attitude quickness) Yaw rate response to lateral gusts Large-amplitude heading changes 1* Yaw due to collective for aggressive agility Pitch due to roll and roll due to pitch for aggressive agility Pitch due to roll and roll due to pitch for Target Acquisition and Tracking Height Response Characteristics Vertical axis control power 1

132 114 Figure 4-4 and 4-5 show handling qualities evaluation using bandwidth criterion for small-amplitude roll and pitch attitude change respectively..4 τ p,φ (sec) knot 2 knots 4 knots LV3 LV2.1.5 LV ω BW,φ (rad/sec) Figure 4-4: Rating on small-amplitude roll attitude change - short-term response to control inputs (bandwidth) knot 2 knots 4 knots LV3 LV2 LV1.25 τ p,θ (sec) ω BW,θ (rad/sec) Figure 4-5: Rating on small-amplitude pitch attitude change - short-term response to control inputs (bandwidth)

133 115 Figure 4-6 shows eigenvalues corresponded to mode of oscillation of the helicopter model. One can see that every oscillatory modes of the closed loop model has damping ratio greater than ζ =.35, the requirement for level one knot 2 knots 4 knots 3.5 ω n (1-ζ 2 ) 1/ LV1 LV2 LV ζω n Figure 4-6: Rating on small-amplitude pitch (roll) mid-term response to control inputs knot 2 knots 4 knots p pk /Δφ pk (1/sec) LV1 LV2 LV Minimum attitude change, Δφ min (deg) Figure 4-7: Rating on roll attitude quickness for all other MTEs

134 116 Shown in Figure 4-7, level one was not achieved for moderate-amplitude roll attitude changes. The rate limit of the differential collective actuator is the major reason of degradation. Increasing the rate of response through the command model did not improve the response. In Figure 4-8, the bottom graph shows the rate-limited response of the differential collective angle actuator. One can see that the rate limit of the actuator was reached and resulted in the bank angle response deviating from the desired value. This deviation greatly degraded the level of handling qualities. The ratio p pk Δ φ pk could be increased up to 15 percent if there were no rate limit and the response matched the command filter. p (deg/sec) φ (deg) real output desired output δ θd (deg) time (sec) Figure 4-8: Effect of rate limit to attitude quickness tested in the hover model.

135 117 Figure 4-9 display the evaluation on moderate-amplitude attitude changes of pitch response. The level one requirement was satisfied q pk /Δθ pk (1/sec) LV1.5 knot LV2 2 knots 4 knots Minimum attitude change, Δθ min (deg) Figure 4-9: Rating on pitch attitude quickness for other MTEs Table 4-2 shows achievable angular rate and angle attitude responses following maximum lateral and longitudinal step inputs. The requirement for level one handling quality requirements for target acquisition and track is satisfied. Table 4-2: Rating on Large-amplitude pitch (roll) attitude change The Controller Achievable angular rate (deg/sec) Achievable angle (deg) with a model Pitch Roll Pitch Roll At knot > 9 At 2 knots At 4 knots

136 118 Figure 4-1 shows handling qualities evaluation using bandwidth criterion for small-amplitude heading change..25 LV3.2 LV2 τ p,ψ (sec).15.1 LV1.5 knot 2 knots 4 knots ω BW,ψ (rad/sec) Figure 4-1: Rating on small-amplitude yaw attitude change short term response to yaw control input (bandwidth) Figure 4-11 and 4-12 show that level two was achieved on the yaw attitude quickness. This degradation was caused by the rate and range limit of the differential longitudinal cyclic actuators. Improvement on the response cannot be achieved by modifying the command filter or other controller components.

137 r pk /Δψ pk (1/sec) LV1 LV2.5 knot 2 knots 4 knots LV Minimum heading change, Δψ min (deg) Figure 4-11: Rating on moderate-amplitude heading changes (attitude quickness) for Target Acquisition and Tracking 2.5 r pk /Δψ pk (1/sec) LV1 knot 2 knots 4 knots.5 LV2 LV Minimum heading change, Δψ min (deg) Figure 4-12: Rating on moderate-amplitude heading changes (attitude quickness) for all other MTEs

138 12 Figure 4-13 shows that the yaw rate of FXV-15 model due to lateral wind gust is very small. This implies good performance of the controller in yaw attitude hold during disturbances LV2 knot 2 knots 4 knots r pk /V g [(deg/sec)/(ft/sec)] LV V g (knots) Figure 4-13: Rating on short-term yaw response to disturbance inputs yaw response to lateral gusts Shown in Table 4-3, level one handling qualities for large amplitude heading changes of target acquisition and tracking MTE was achievable in hover. The controller was designed to achieve exactly 6 degree-per-second yaw rate at the maximum yaw control input. For the 4 knots flight condition the controller achieved level two for aggressive agility and target acquisition and tracking MTE. However, level one was obtained for all other MTEs. Data in the table shows the maximum yaw rate following the maximum yaw control input. The range constraint on differential cyclic has a major effect on this criterion. At 4 knots, the inherent directional stability from the vertical stabilizers makes it difficult to achieve level one or even two.

139 121 Table 4-3: Data required to rate the large-amplitude heading changes The Controller with a model Achievable yaw rate (deg/sec) At knot At 2 knots At 4 knots The result from Figure 4-14 confirms that the coupling between pitch and roll is very small as mentioned in the chapter two. Average p/q (db) knot 2 knots 4 knots LV1 LV2 LV Average q/p (db) Figure 4-14: Rating on interaxis coupling pitch due to roll and roll due to pitch coupling for Target Acquisition and Tracking In Figure 4-15, step inputs of the collective control were used to find the time history of the altitude response. A least-squares-fitting algorithm was used to derive a first order approximation of the height response. The equivalent phase delay and time constant were found to be within level one requirement.

140 ADS-33E.3.25 LEVEL 2 τ h dot eq (sec) LEVEL 1.5 knot 2 knots 4 knots T h dot eq (sec) Figure 4-15: rating on response to collective controller height response characteristics Table 4-4 shows vertical rate following maximum travel range of collective lever. The FXV-15 has sufficient performance to achieve desired rate of climb. However, an accurate performance and atmosphere s model was not included in the linear model. The requirement should be checked with the nonlinear model. Table 4-4: Data to rate response to collective controller vertical axis control power The Controller with a model Achievable vertical rate (ft/min) Level At knot At 2 knots At 4 knots

141 4.1.2 Conversion Configuration 123 Table 4-5 shows the results of predicted handling qualities of the augmented FXV-15 in conversion configuration. The controller was unable to deliver the level one handling qualities on only one criterion. It was able to achieve level one on all others.

142 124 Table 4-5: Predicted handling qualities of XV-15 simulation in conversion mode Handling Qualities Criteria Level Short-term response (bandwidth) Mid-term response to control inputs Flight path response to pitch attitude (frontside) Flight path response to collective controller (backside) Longitudinal static stability PASS Pitch attitude due to collective control PASS Roll due to pitch coupling for Aggressive agility Pitch due to roll and roll due to pitch coupling for Target Acquisition and Tracking Small-amplitude roll attitude response to control inputs (bandwidth) Moderate amplitude attitude changes (attitude quickness) 1* Large-amplitude roll attitude changes 1* Linearity of roll response PASS Roll-sideslip coupling Small-amplitude yaw response for Target Acquisition and Tracking (bandwidth) Large-amplitude heading changes for Aggressive agility Linearity of directional response PASS Lateral-directional oscillations Spiral stability Yaw control in steady sideslips (directional stability) Bank angle in steady sideslips PASS Lateral control in steady sideslip PASS 1 1

143 125 Figure 4-16 and 4-17 show handling qualities evaluation on pitch attitude response to longitudinal controller. Figure 4-16 shows that the closed-loop aircraft has sufficient bandwidth. Figure 4-17 shows that all oscillatory modes are with level one requirement and well damped LV3 LV2 LV1 8 knot 1 knots 14 knots.25 τ p,θ (sec) ω BW,θ (rad/sec) Figure 4-16: Rating on pitch attitude short-term response (bandwidth)

144 LV1 LV2 ω n (1-ζ 2 ) 1/ knot 1 knots 14 knots ζω n Figure 4-17: Rating on pitch attitude mid-term response Figure 4-18 shows the phase of the vertical speed and pitch attitude response to a longitudinal input for each flight condition. The vertical rate response leads the pitch attitude response in the required frequency range in all cases. The requirement for level one is satisfied.

145 kn flight speed model 6 Vertical rate Pitch attitude 4 2 Phase ω 8 6 Vertical rate Pitch attitude 1kn flight speed model 4 2 Phase ω 8 6 Vertical rate Pitch attitude 14kn Flight speed model 4 Phase (degree) Figure 4-18: Phase difference between vertical response and pitch attitude response ω

146 128 The data shown in Figure 4-19 was obtained by simulating the vertical speed response to a full-collective control input. A least-squares-fitting algorithm was used to derive a first order approximation of the height response. The equivalent phase delay and time constant were found to be within level one requirement..35 ADS-33E.3.25 LEVEL 2 τ h dot eq (sec) LEVEL knot 1 knots 14 knots T h dot eq (sec) Figure 4-19: Rating on flight path control flight path response to collective controller Table 4-6 and 4-7 and Figure 4-2 show the handling qualities evaluations on interaxis coupling criteria. Requirements of level one were met by the controller. Results also show the advantage of tiltrotor aircraft which inherently has less interaxis coupling than a conventional helicopter. Table 4-6: Rating on interaxis coupling pitch attitude due to collective control Model Small Input 2 Large Input ( deg/ ft / sec ) 2 ( deg/ ft / sec ) Up direction Down direction Level 8kn flight speed kn flight speed kn flight speed

147 129 Table 4-7: Rating on interaxis coupling roll due to pitch coupling for Aggressive agility Model Δφ pk Δθ 4 Level 8kn flight speed e-4 1 1kn flight speed 8.575e kn flight speed 3.166e LV3 Average p/q (db) knot 1 knots 14 knots LV1 LV Average q/p (db) Figure 4-2: Rating on interaxis coupling pitch due to roll and roll due to pitch coupling for Target Acquisition and Tracking. Figure 4-21, 4-22 and 4-23 and Table 4-8 show handling qualities evaluation on roll attitude response to lateral controller criteria. In Figure 4-22, the moderate-amplitude requirements for target acquisition and tracking are used and requirement of level one is not satisfied. The effect of the control surface rate limits prevents the system from getting a higher ratio. Although, level three for the most strict MTE was obtained, the controller was good enough to achieve level one handling qualities for all other MTEs, shown in Figure 4-23.

148 τ p,φ (sec) LV3 LV2.1 LV1 8 knot.5 1 knots 14 knots ω BW,φ (rad/sec) Figure 4-21: Rating on small-amplitude roll attitude response to control input (bandwidth) LV1 p pk /Δφ pk (1/sec) LV2 LV3.5 8 knot 1 knots 14 knots Minimum attitude change, Δφ min (deg) Figure 4-22: Rating on moderate-amplitude roll attitude change (attitude quickness) for Target Acquisition and Tracking

149 knot 1 knots 14 knots LV1 p pk /Δφ pk (1/sec) LV2 LV Minimum attitude change, Δφ min (deg) Figure 4-23: Rating on moderate-amplitude roll attitude change (attitude quickness) for all other MTEs As shown in Table 4-8, even though the controller is designed to be an attitude command/attitude hold response type (ACAH), it also has a good achievable roll rate response in the conversion mode. Achievable roll rate and bank angle satisfy level one requirements for any MTEs except for Target Acquisition and Tracking MTE (level two handling qualities on target acquisition and tracking MTE). Increasing the gain for commanded roll attitude might increase the maximum roll rate but tends to degrade the level on other criteria. The achieved roll rates seem reasonable for this aircraft configuration. Table 4-8: Achievable roll and roll rate responses following large lateral command Model Achievable roll rate (deg/sec) Achievable bank angle (deg) 8kn flight speed > 9 1kn flight speed >9 14kn flight speed

150 132 Figure 4-24 and Table 4-9 show evaluations on handling qualities criteria of yaw response to yaw controller..25 LV3.2 LV2 τ p,ψ (sec).15.1 LV1.5 8 knot 1 knots 14 knots ω BW,ψ (rad/sec) Figure 4-24: Rating on small-amplitude yaw response to yaw controller for Target Acquisition and Tracking (bandwidth) The maximum heading change following a maximum step pedal input for the flight conditions are shown in Table 4-9. In this configuration, the turn coordination control mode might reduce effectiveness of yaw control. Combination of lateral and pedal input should be used in case larger heading changes are required. Table 4-9: Rating on large-amplitude heading changes for Aggressive agility Model Heading change in 1 second (degree) Level 8kn flight speed kn flight speed kn flight speed

151 133 Table 4-1 shows the evaluation on lateral-directional stability. The dutch-roll mode eigenvalues are well damped and within level one requirement for all MTEs. Table 4-1: Rating on lateral-directional stability - lateral-directional oscillation Model Eigenvalues Damping ratio Level 8kn flight speed ±.8964i kn flight speed ±.954i kn flight speed ± 1.114i

152 4.1.3 Airplane Configuration 134 Table 4-11 shows the results of predicted handling qualities of the augmented FXV-15 in airplane configuration. The controller was successful in delivering the best handling qualities of assigned criteria to the aircraft. Table 4-11: Predicted handling qualities of XV-15 simulation in airplane mode Handling Qualities Criteria Level Phugoid stability Short-period response Longitudinal control in maneuvering flight (to2.5g) 1& Lateral-directional oscillations (Dutch roll) Roll mode Spiral stability Coupled roll-spiral oscillation PASS Roll rate oscillation Bank angle oscillation Pilot-induced oscillations PASS Roll control effectiveness Linearity of roll response PASS Directional control characteristics PASS Yawing moments in steady sideslips 1& Side forces in steady sideslips PASS Rolling moments in steady sideslips 1& Positive effective dihedral limit 1&2

153 knots 2 knots 24 knots Root Locus Rotors gimbals mode Imaginary Axis Mode of actuators Short-period mode Real Axis Root Locus Dutch-roll mode Imaginary Axis τ p & τ r τ q Roll-spiral coupling Phugoid mode Real Axis Figure 4-25: Eigenvalues of augmented system of FXV-15 in airplane configuration.

154 136 Figure 4-25 shows root locus plot of eigenvalues from the three models. The controller adds six eigenvalues to the plots. Three eigenvalues are located at points corresponding to the time constant of the command filter eigenvalues. The other three are located near the origin, and correspond to integrators in the controller. Table 4-12 shows the damping ratio of the phugoid mode. The requirement for phugoid stability was met. Table 4-12: Damping ratio of phugoid mode for phugoid stability criterion Model Ph damping ratio 16kn flight speed kn flight speed kn flight speed.6617 Figure 4-26 and Table 4-13 show evaluation on longitudinal maneuver characteristics. The augmented system had an acceptable short-period oscillation mode. 1 2 ω nsp (rad/sec) LEVEL 2 LEVEL 3 16 knots 2 knots 24 knots LEVEL 2 LEVEL 1 LEVEL 1 LEVEL 2 & n/α (g/rad) Figure 4-26: Rating on short-period frequency and acceleration sensitivity

155 137 Table 4-13: Short-period damping ratio in airplane mode Model SP damping ratio 16kn flight speed kn flight speed kn flight speed.3635 The achievable load factors are shown in the Table 4-14, which shows the maximum load factor achieved from a longitudinal control input. The controller achieves a load factor maneuver between -1.2 to 2.6g even though the high speed controller was designed to achieve -.5 and 2.5 g load factor for the maximum control displacement. Table 4-14: Achievable load factor of models of airplane configuration Model Achieved load factor (g) XB> (Pull) XB< (Push) 16kn flight speed kn flight speed kn flight speed Table 4-15 shows the roll rate oscillations of each conditioned model. The lateral oscillations were within the level one requirement (over 6%). The controller satisfies level one. Table 4-15: Percentage of the first minimum following the first peak in roll rate response Model Percent (%) 16kn flight speed kn flight speed kn flight speed 61.82

156 138 Table 4-16 shows the capability of the controller to track roll commands well. The response satisfied level one requirement which requires the aircraft to be able to perform 45 degree bank angle within 1.4 seconds in every flight condition. Table 4-16: Time to achieve the assigned bank angle Model Time to achieve 45 degree bank angle (second) 16kn flight speed kn flight speed kn flight speed Discussion The advantages of tiltrotor aircraft over conventional single rotor helicopters were investigated. The tiltrotor aircraft not only have the capability to convert airplane mode for higher speed flight but also have impressively low inter-axis coupling. However, one of the disadvantages of the tiltrotor aircraft was found. The weakness of insufficient power in yaw axis was observed through degradation of handling qualities in large amplitude heading change in helicopter and conversion mode. Each tiltrotor aircraft has two engines placed in a distance away from its yaw axis; therefore, it has higher moment of inertia than a conventional helicopter which has engines close to the yaw axis. Because of larger moment of inertia, tiltrotor aircraft are not able to turn as fast as helicopters which have the same weight and the same applied yaw moment as the aircraft. Therefore, the aircraft have a difficulty to satisfy the large heading change requirements on ADS-33E-PRF which are mainly written for conventional helicopters.

157 139 Having engines far away from the longitudinal axis also causes tiltrotor aircraft high moment of inertia in its roll axis. Roll inertia also a factor in moderate amplitude roll attitude changes. High roll inertia causes aircraft a difficulty to response to the roll command quickly. The linear models that were used to test the controller seem to have problems in calculating straight steady sideslip flights. There was a big difference between right and left sideslip that occurred because changes in lateral axis did not cause logical changes in longitudinal axis (e.g. bank left cause pitch up while bank right cause pitch down, etc.). The criteria relating to sideslip flight should be evaluated further using the nonlinear model. The model following / inversion controller provides an effective design tool for flight control. The controller of each axis can be designed separately. Only inversion model needs to be schedules with operating condition. The concept of using command filter to generate desired response is very useful since they can be matched to handling qualities specifications. The designed controller seems to be very reliable in helicopter and airplane configurations. In addition to tracking the commanded response, closed loop stability was also achieved. However, for certain maneuvers, instability was observed in conversion mode. The constraints on rate and range of actuator motion are the suspected cause. In the conversion configuration, the nacelles are tilted at 6º. Commands to the lateral axis and yaw axis will be sent to a combination of control surfaces. If either actuator in roll axis or yaw axis hits its range or rate limit, there will be excessive moment from the other axis causing the attitudes to deviate from the command. The

158 14 instability occurs when the attitude error grows. Due to the use of an integral compensator which integrates amount of error that deviates from the command, the compensator emits a relatively large and abrupt signal (compare to the signal when the controller follows the command input). This signal will tend to exceed the actuators limits. As results, the instability will become more severe and the system may lose control. Further investigation of the problem is warranted in order to resolve this issue..

159 Chapter 5 Test Pilot Handling Qualities Rating Results In this chapter the FLIGHTLAB software that is used to simulate XV-15 real time dynamics is introduced. Next, a flight simulation facility at the Penn State Vertical Lift Research Center of Excellence (VLRCOE) used to tune the controller to real time flight is shown. Then, the University of Liverpool s HELIFLIGHT flight simulation facility where the controller was rated by test pilots is presented. Finally, the results of pilot rating are displayed and the discussions on the controller are made. 5.1 FLIGHLAB FLIGHTLAB is an aircraft modeling and analysis tool. The software provides user with abilities to produce models interactively, test and evaluate a model, and perform real-time simulations of aircraft models. The software provides a graphical user interface (GUI) which allows users to easily create models using the library of modeling components. The GUI automatically generates code needed for model evaluation and model simulation. Figure 5-1 displays the graphical user interface of FLIGHTLAB.

160 Figure 5-1: FLIGHTLAB s modeling user interface 142

161 5.2 Penn State s Flight Simulation Facility 143 At the Penn State Vertical Lift Research Center of Excellence, a new simulation facility has recently been built. The simulator, shown in Figure 5-2., is a PC-based and re-configurable type (model of aircraft can be changed) using FLIGHTLAB software. The simulator is a fixed-base platform using the XV-15 simulation cab which is donated by Bell Helicopter. The cab has four-axis control loading system that controls force-feel characteristic to four cockpit control inputs. Three-channel visual module is used to display image on a 18º, 15 feet diameter curved screen. Figure 5-2: Flight simulation facility at the Penn State Vertical Lift Research Center of Excellence (VLRCOE)

162 144 In the simulation process, the aircraft initial state is set from the operator station. The pilot command from the XV-15 cab is tracked by a cab interface computer, which sends the pilot control signal to FLIGTHLAB while sending control inceptor forces back to the control loading system in the cab. FLIGHTLAB runs to calculate aircraft responses the pilot control input and generates a visual image corresponding to current aircraft state. Figure 5-3 displays the process diagram of the simulator used for flight testing. Figure 5-3: Simulation process diagram. In this research, the simulator was used to observe response of the nonlinear aircraft model in real time simulation. The functions changing with airspeed and nacelle angles were checked. The aircraft responses due to the change were observed (e.g. responses in transition between low speed control and high speed control, etc.). Results from the real time simulation were used to fine tune the control laws. The XV-15 model

163 145 used at the Penn State University was developed by Advanced Rotorcraft Technologies (ART), and is slightly different from the model developed at the University of Liverpool. However, the model was sufficient to provide a functional check of the controller. 5.3 The University of Liverpool s HELIFLIGHT Flight Simulation Facility The HELIFLIGHT, shown in Figure 5-4, is a PC-based, re-configurable flight simulator developed to produce a relatively high-fidelity system [3]. The system consists of an inter-changeable flight dynamics modelling software (FLIGHTLAB) with a real time interface, a six degree of freedom platform, four axis dynamic control loading, a three channel collimated visual display and a computer generated instrument panel and head-up display. Figure 5-4: HELIFLIGHT (left) and its Interior (right)

164 5.4 Cooper-Harper Rating Results 146 The flight test maneuvers assigned for handling qualities assessment were selected from those documented in ADC-33E-PRF in order to cover all critical handling qualities needed in a Search and Rescue (SAR) mission. For the helicopter configuration, the mission task elements (MTEs) to be performed were Hover-turn, Hover reposition, Vertical maneuver (Bob-up) and Pirouette. For the other two configuration, the Roll-step and Heave-hop mission task elements (MTEs) were selected. The Cooper-Harper scale was used to rate aircraft in performing these missions. The rating result of helicopter configuration is shown in Table 5-1. The controller was shown to substantially improve handling qualities in the helicopter configuration. The designed model following controller (MFC) achieves Cooper-Harper rating of one to three which implies that it is satisfactory without any improvement and confirms level one handling qualities. Table 5-1: Test pilot rating results MTE Cooper-Harper Handling Qualities Rating Bare-airframe PSU MFC Hover Turn 6 3 Hover Reposition 6 1 Bob-up 6 3 Pirouette 8 1 However, the conversion and airplane configuration were not evaluated due to instability of the controller. In flight testing, instability occurred in forward flight causes the pilot a difficulty to finish the missions. Since the controller worked fine in the XV-15

165 simulation at the Penn State University, more testing in the XV-15 simulation at the University of Liverpool is required in order to investigate this problem Hover Turn Figure 5-5 and 5-6 present data gathered while the test pilot was performing the Hover Turn MTE. Figure 5-5 shows the key requirements of the MTE in top two graphs and time history of cockpit command inputs in the four below. In both graphs on the top, the bold line indicates the desired performance and the dash line represent the adequate performance in this requirement (the aircraft response shall not cross the line to satisfy the requirement). The results of bare-airframe testing showed high degree of instability. High pilot workload was observed from control efforts in lateral and longitudinal stick. For the augmented aircraft simulation, the top-right graph shows that the augmented aircraft has desired performance in maintaining altitude. The top-left graph shows that the augmented aircraft cannot maintain longitudinal / lateral position within 6 feet which does not satisfy the adequate performance. However, pilot s felt desired performance was achievable with better visual cues.

166 148 Figure 5-5: Aircraft positions and control inputs data in Hover Turn MTE test In the time history of cockpit commands and aircraft s angular responses, shown in Figures 5-5 and 5-6, most of the responses seem to be good. Small oscillation in roll response and lateral command are observed. They might come from an effort of the pilot to maintain the aircraft position or correct the heading.

167 149 Figure 5-6: Aircraft attitudes and angular rates in Hover Turn MTE test Cooper-Harper handling qualities rating of three was given to the augmented aircraft in this criterion. The comment of the test pilot implies good stability throughout the turn and high sensitivity of cyclic controls.

168 5.4.2 Hover Reposition 15 Figure 5-7 and 5-8 present data gathered while the test pilot was performing the Hover Reposition MTE. This task assesses the ability of the aircraft to capture a hover. Instability of bare-airframe model was observed and was severe in roll response in performing this task. High pilot workload was observed from control efforts in lateral stick. For the augmented aircraft results, in Figure 5-7, the time history of height response shows that the height response is briefly over limit of the desired performance (indicated by bold lines). However, this was a result of the collective lever input being slightly out of trim at the very beginning of the maneuver. However, pilot s felt desired performance was achievable.

169 151 Figure 5-7: Aircraft positions and control inputs data in Hover Reposition MTE test In Figure 5-8, slightly high roll oscillation is observed in the graphs of bank angle and roll rate responses. The other responses look good.

170 152 Figure 5-8: Aircraft attitudes and angular rates in Hover Reposition MTE test Cooper-Harper handling qualities rating of one was given to the augmented aircraft in this criterion. Slightly high control sensitivity was noticed by the test pilot however it was all satisfactory for him.

171 5.4.3 Bob-up 153 Figure 5-9 and 5-1 present data gathered while the test pilot was performing the Bob-up MTE. Figure 5-9 shows that the requirements on the desired performance are satisfied by the augmented aircraft. The augmented aircraft is able to maintain its position and the finish altitude very well. The time history plots of command inputs show that the vertical reposition is done by a use of collective lever only. Figure 5-9: Aircraft positions and control inputs data in Vertical Reposition (Bob-up) MTE test In Figure 5-1, all aircraft attitude and angular rate look very good. The collective control showed very small effects on them.

172 154 Figure 5-1: Aircraft attitudes and angular rates in Vertical reposition (Bob-up) MTE test Cooper-Harper handling qualities rating of three was given to the augmented aircraft in this criterion. The augmented aircraft was praised for its precisely heave control and accuracy in plan position.

173 5.4.4 Pirouette 155 Figure 5-11 and 5-12 present data gathered while the test pilot was performing the Pirouette MTE. The ability to accomplish precision control of aircraft simultaneously in roll, pitch, yaw and heave is checked in this task. The results of bare-airframe testing showed high degree of instability. A difficulty to perform the task was observed from control efforts in all input and implied very high pilot s workload. For the augmented aircraft, Figure 5-11 shows that the augmented aircraft is able to translate around the circle with no difficulty in the control. A single deviation of height response from the desired performance was observed, but pilot s felt desired performance was achievable with minimal compensation.

174 Figure 5-11: Aircraft positions and control inputs data in Pirouette MTE test 156

175 157 Figure 5-12: Aircraft attitudes and angular rates in Pirouette MTE test Cooper-Harper handling qualities rating of one was given to the augmented aircraft in this criterion. The responses of augmented aircraft impressed the test pilot with its zero coupling and predictable rates.