On the Guidance and Control System of Epsilon Solid Rocket Launcher

On the Guidance and Control System of Epsilon Solid Rocket Launcher Y. Morita, M. Tamura, N. Ishii, T. Yamamoto Japan Aerospace Exploration Agency H. Ohtsuka, Y. Segawa and K. Tanaka IHI Aerospace Abstract The Epsilon solid rocket launcher proceeded to the full development phase in August 2010 and its first launch is officially declared to be conducted in 2013 to carry a space telescope satellite SPRINT-A. The concept of the vehicle, the next generation launch system, requires a simpler launch system and better user friendliness interface than its predecessor, M-V in order to provide small satellites with an efficient launch. Such innovation significantly affects the architecture of the guidance & control system. This paper describes the design of the guidance & control system of the Epsilon launcher. KEY WORDS Robust control, Guidance and control, Control theory, Attitude control. 1. Introduction The research on Japan s solid rockets started more than 50 years ago with a horizontal launch experiment of tiny pencil-size rockets and such endeavor was rewarded in 1970 when Japan s first satellite Ohsumi was launched and in 1985 Japan s first planetary spacecraft Sakigake was lifted toward Halley s comet. In mid-90 s, the M-V rocket became available to meet the strong demand for the full scale scientific missions including the world s first asteroid sample-return spacecraft HAYABUSA. Beyond this success, the Epsilon solid rocket launcher (Fig. 1) is now under development by JAXA to prepare for its first flight in 2013 to launch the space telescope satellite SPRINT-A. Now, take a look at the background of the Epsilon, that is, the achievement of the M-V. First of all, the M-V contributed to space science in almost all its fields from the space astronomy to even the planetary missions (Fig.2). Back in 2003, the M-V launched the world s first asteroid sample-return spacecraft HAYABUSA that returned home safely in 2010. In 2006, the M-V also launched the infra-red space telescope AKARI: at that moment, it became the world s best performance solid rocket launcher that can be utilized for both planetary and sun-synchronous missions. In addition, let s not forget that from the very beginning Japan s research on solid rockets has been conducted independently of foreign technologies: a source of Japanese pride. In this way, the space science communities of Japan have obtained series of world-leading achievements by using the M-V; however, they suffered from the relatively low launch frequency mainly due to the relatively high cost and the long development period of those full scale scientific missions: only 7 spacecraft in 10 years. The space science will not survive in this situation, thus they now focus on more satellites smaller in size, lower in cost and shorter in development period in order to increase the launch opportunities. Fig. 1 An image of the Epsilon rocket on launch pad.

Fig. 2 Achievements obtained by the M-V rocket. With this as background, the purpose of the Epsilon rocket is to provide small satellites with responsive launch, which means in the study we focus on a low cost, user friendliness and ultimately efficient launch system. The combined power of small satellites and the small launcher will increase the level of space activities. Throughout 50 years history of Japan s solid rockets, the research was conducted only to increase the rocket performance. Now, for the first time ever in the history, it requires optimization of the entire launch system, including the efficient launch operations and the compact ground facilities (Fig. 3). This attempt is equally applicable to the liquid rockets as well and paves the way to future space transportation systems. Then what is the largest revolution we aim at for the Epsilon? That is the innovative launch system to reform the old-fashioned launch system into an ultimately efficient one. The key to success is avionics systems: they are designed to be highly intelligent so that the vehicle performs check-outs autonomously and to be connected to the ground support facilities, through a high-speed network. Owing to this endeavor, lots of ground support equipments can be all eliminated and the associated set-up time and the number of personnel involved can be reduced. Until now, the launch control room contains tons of ground support equipments and lots of workers involved From now on, the intelligent check-out system and the secured high-speed network makes it possible to conduct the launch control anytime, anywhere in the world simply by using a single laptop computer, which is called a mobile launch control, a realization of the science fiction to the science fact (Fig. 4). The concept of the vehicle, the next generation launch system, requires simpler launch system and better user friendliness than its predecessor, the M-V in order to provide small satellites with an efficient launch. From the point of view of the astrodynamics, such innovation significantly affects the architecture of the guidance and control system. This paper describes the design of the guidance and control system of the Epsilon rocket launcher. 2. Hardware Architecture The configuration of the Epsilon rocket is a three-staged solid propellant vehicle, having a 1.2-ton payload capacity into a low earth orbit (LEO) while it is 92 ton in lift-off weight, 24 m in total length, and 2.5 m in maximum diameter (Table 1). Each of the first and the second stages has 3-axis attitude control capability (Fig. 5): The pitch and yaw in the powered flight can be controlled through a mobile nozzle thrust vector control (MNTVC) that are driven by a pair of electro-mechanical servo-motors. The first stage servo-motor is powered by a special high power thermal battery and the second through an integration of commercial lithium batteries. On the other hand, the roll in the same phase and the 3-axis attitude in the coasting phase can be stabilized by reaction jets. The first stage reaction jet, solid motor side jet (SMSJ), is generated by a solid propellant gas generator (GG) while the one for the second stage, reaction control system (RCS) by conventional hydrazine engines. Contrary to the M-V, the third stage is just Fig. 3 Concept of the Epsilon rocket launcher. Fig. 4 Artistic view of the mobile launch control.

Fig. 5 Configuration of the Epsilon solid rocket and its attitude control engines. spin-stabilized for more simplicity. To compensate for the residual orbit error caused by the spin-stabilized third stage, an optional tiny upper stage, post-boost stage (PBS) can be installed onboard, which will be propelled by conventional hydrazine engines (Fig. 6). This is for better orbital accuracy and maneuverability to increase the user friendliness. To further enhance the user-friendly characteristics, a special payload attach fitting (PAF) is under development to lower the level of high frequency vibration (around 50 Hz) that is caused by the combustion vibration of the first stage solid rocket booster (SRB-A). The mechanism consists of a multi-layered structure of rubbers and thin metals, having lower axial rigidity, to isolate the high frequency vibration (Fig. 7). The structure also causes a reduction in the lateral rigidity of PAF, resulting in lower bending frequency of the entire vehicle. This is absolutely a new challenge for the attitude control algorithm design because the rigid mode dynamics and the first order bending oscillation will be so close. Table 1 Specifications of the Epsilon rocket (EX) *E1 denotes the upgraded version of Epsilon that aims at lower cost character and is planned to be developed after the current version (EX). Configuration Length/diameter Lift-off mass Launch capacity Next-generation technologies Launch cost Development cost First launch Launch site 3-stage solid rocket launcher with optional PBS approximately 24 m/ 2.5 m approximately 92 ton LEO (250X500 km): 1.2 ton SSO (500 km): 450 kg Autonomous check-out Mobile launch control approximately 3.8 billion Yen (3.0 billion Yen for E1*) 20.9 billion Yen 2013 (2017 for E1*) Uchinoura Space Center(USC) 3. Guidance and Control Design 3.1 Guidance strategy 3.1.1 Powered flight guidance Until the M-V rocket (last launch in 2006), Japan s solid rocket launchers have utilized the radio Fig. 6 PBS stage configuration of the Epsilon. guidance strategy that were successful to launch various scientific satellites and even planetary missions including HAYABUSA. However, it requires high precision tracking radars and gigantic computer systems. To realize a simpler launch system that consists of efficient launch operations and compact ground facilities, it is decided that the inertial guidance will be applied to the Epsilon launch vehicle. This can be considered inevitable because even the range safety control is now planned to be autonomous in the next 5 years to eliminate the associated ground facilities. Despite the introduction of the inertial guidance, the guidance algorithm itself can remain virtually the same as utilized for the radio guidance, which is quite simple and makes the most use of the characteristics of solid rocket launchers. An example is the second stage guidance. First of all, and, the trajectory errors at burn-out of the first and second stages, respectively, can be connected by the second stage guidance, as: X 2= AX1+ BU 2. Then, the most important task is choosing the state vector to be guided out of the orbital elements, which may affect significantly the guidance performance for the mission. Note it is not possible to guide all of the orbital parameters using solid rockets. It can be well understood by considering sun-synchronous Fig. 7 Structure of payload attach fitting.

missions for which the essential is the relation between the orbital inclination and apogee height. Minimizing the quadratic form of : X 2 TJ X 2 yields the optimal guidance law: =-1T2U ( B B ) B A S X - T 1. Here is the sensitivity matrix that can be calculated prior to the launch. This simple but relatively accurate law was used for the M-V third stage guidance and demonstrated its effectiveness when launching HAYABUSA. On the other hand, for the second stage guidance, a much simpler algorithm was applied. It utilized a table format as indicated in Table 2. When we have the velocity of 2.90 km/sec and the azimuth angle of 182 degrees at apogee height after the second stage burn, for example, the optimal guidance will be obtained as the yaw attitude correction of 0.4 degree. The advantage of the method is the wide range of dispersions that can be corrected, even out of linearity. The disadvantage is the limited guidance accuracy that is restricted by the discretization of the Table 2 An example of the guidance table (PBS) can be installed, that is propelled by the conventional hydrazine engines and virtually identical to the M-V third stage attitude control systems, that is very small and simple in design. By using this option, a wide variety of orbits that small satellites will require can easily be reached and the orbit accuracy can be increased to as high as those of the liquid launchers, thus making the Epsilon the world s best performance solid rocket launcher in this category. From the point of view of guidance, the stage should be different from the solid rocket stages. First of all, the engine is able to be cut-off. Secondly, the acceleration produced by PBS is relatively small and the burn time is relatively long (more than 10 minutes), thus gravity term cannot be simplified and the J2 effect cannot be neglected. Thirdly, as the third stage is a spin-stabilized solid propulsion motor, the initial orbital dispersion can be significantly large. To tackle with those challenges, the LVIC (Long Velocity Increment Correction) guidance law will be utilized that are already applied to rendezvous of HTV (H2A Transfer Vehicle) with the Space Station. The current analysis indicates that the orbit accuracy can be increased to as high as that of the H2A (Table 3). It should be noted that it is not the guidance law but the inertial sensor that dominates the orbital accuracy. In this respect, the novel light and inexpensive integrated navigation system is now under study and it will cut half the orbital injection error. Table 3 Orbital injection accuracy of the Epsilon launch vehicle. table. The size of which is confined due to the capacity of the onboard computer. Still, this approach is acceptable for the second stage guidance because the stage has relatively large dispersions and its residual guidance error can be corrected by the third stage guidance. Note that the Epsilon s third stage is spin-stabilized as such the second stage guidance is much more important. Thus, the sensitivity matrix and the guidance table will be utilized together for the second stage, that is, relatively small orbital errors will be corrected by the accurate sensitivity matrix while larger dispersions will be guided by the wide range table. In this way, the established technologies obtained through the M-V will be made the most use of for Epsilon. 3.1.2 Guidance law for PBS flight The next topic is associated with the user friendliness that will be of the world leading level. In order to get a better orbital precision and maneuverability, an optional tiny post-boost stage Accuracy in SSO@500km Standard Configuration PBS onboard Perigee height (km) ±25 ±20 Apogee height (km) ±100 ±20 Inclination (degrees) ±0.6 ±0.2 3.1.3 Rhumb-line control The last topic in this section is quite new in launcher applications and it is the guidance of spin-stabilized rocket stage that is being paid attention to for the first time ever in 50 years history of Japan s solid rocket. This is prompted by the following fact: the direction of angular momentum of the spin-stabilized rocket stage is shifted by the thrust of the stage and the amount of the shift is dominated by the magnitude of initial thrust. Although the directional error remains as small as a couple of degrees, it significantly affects the orbital accuracy and increase the amount of fuel consumed by PBS in order to correct the orbital dispersion caused by the spin-stabilized third stage burn. To tackle this problem, the rhumb-line control

strategy will be introduced for the Epsilon. It utilizes a simple algorithm: the onboard computer fires a single-axis hydrazine thruster impulsively at constant direction as measured by the gyro to provide the precession moment in the expected direction so that the angular momentum is modified in a proper direction. The algorithm was already applied to two-stage sounding rockets but will be the first for the launchers. An example of the preliminary analysis indicates the effectiveness of the rhumb-line control, which is well appreciated in Table 4, showing the time required for the direction error to converge within 0.2 degree. Note that the fundamental error factors are taken to be fluctuations in firing timing of the thruster and the inertia moment of the entire stage. As represented in the table, for a satellite of 500 kg mass, the rhumb-line control can converge in 20 seconds. Considering the third stage burn lasting as long as 100 seconds, significant guidance efficiency can be expected, resulting in the PBS burn saving the fuel by more than 10 kg for the same payload mass. Table 4 Convergence time of Rhumb-line control. No. Fluctuations Delay (msec) Inertia Moment satellite mass (kg) 100 500 1200 1 30 0 13 7 20 2 80 0 11 10 25 3 130 0 17 15 43 4 30 +10 % 11 9 22 5 30-10 % 4 7 18 3.2 Attitude control algorithm The design algorithm of the attitude control applied to Japan s solid rocket launchers has evolved since mid 80s which is featured by its step-by-step enhancement of robustness character (Table 5). Note that despite the classical design methods are still adopted worldwide for the launcher control, the modern and post-modern design architectures were successfully applied to Japan s solid rockets to get better robust characteristics against variations in the plant dynamics: the entire system has a grade of uncertainty of system parameters. Especially, the H control theory was applied to the first three flights of the M-V (1997-2000) 1-5). Within the framework of a mixed sensitivity problem, the challenge is to achieve better tracking performance in a relatively low frequency region as well as preferable robust stability in the high frequency range: nominal performance and robust stability. Beyond this success, the further novel design was taken to renew the attitude control of the last four flights of M-V (2003-2006), which utilizes the more advanced -synthesis to enhance the system s tracking performance more directly 6-14). The -synthesis was applied for the first time ever in the world for the satellite launcher control beyond the reliable H control to get far better robust characteristics not only in stability but in tracking performance. In sharp contrast to the H design, the -synthesis can treat the robustness directly even in the performance. In this way, the robust control design of the M-V launch vehicle evolved since its first flight in 1997 using H theory through its last journey in 2006 applying the -synthesis. The same advanced approach is now under study to apply for the Epsilon rocket launcher. Table 5 Evolution of the Design Format Applied to the M-V launch Vehicle. Rocket Design Robust Robust Launch No. format stability performance M-V-1(1997) H Yes No M-V-3(1998) H Yes No M-V-5(2003) Yes Yes M-V-6(2005) Yes Yes M-V-8(2006) Yes Yes M-V-7(2006) Yes Yes As already mentioned, for the Epsilon launcher, a special vibration attenuator is installed at the satellite attachment to suppress the high frequency vibration (around 50 Hz) caused by the first stage combustion vibration. The mechanism has reduced axial rigidity to isolate the vibration and in turn it results in a lower bending frequency of the entire vehicle. This makes the control design absolutely difficult as the rigid body dynamics (0.5 Hz) and the first order bending oscillation (4Hz) will become much closer (6Hz for M-V). The same logic as M-V is tentatively applied to the Epsilon first stage at the time mark when the dynamic pressure reaches its maximum, in other words, most instability arises. The preliminary analysis indicates that the nominal as well as the minimum stability can be considered well within the scope of expectation (Table 6) 15). Note that the stability survey is conducted by changing the bending rigidities of the PAF and the entire vehicle as well as the structural damping coefficient of PAF. The sub-scale test of PAF shows that the damping of the first mode is more than 1 %. Table 6 survey. Entire body rigidity Nominal Minimum (-50%) Gain margin obtained by preliminary PAF rigidity Entire structural damping ratio 0.5% 1.0% 1.5% nominal -8.95 db -15.28 db -17.96 db -30% -6.62 db -12.65 db -15.13 db -50% -5.17 db -10.74 db -12.25 db nominal 0.11 db -6.85 db -10.71 db -30% 0.22 db -6.16 db -9.97 db

4. Concluding Remarks The purpose of the Epsilon solid rocket launcher is to establish a much simpler launch system than any other launchers of the world as well as a better user friendliness interface than any other solid rockets. This includes more accurate orbital injection and milder mechanical environment. Such innovative concept of the rocket significantly affects the architecture and algorithm design of the robust control system. A preliminary design, which utilizes the H control algorithm, indicates that the obtained robust stability can be considered well within the scope of expectation. The paper has described the current design of the robust attitude control of the Epsilon rocket launcher. References [1] Y. Morita, J. Kawaguchi and S. Goto. Software design of the M-V attitude control system. Proc. 6th Workshop on Astrodynamics and Flight Mechanics, Kanagawa, Japan, 1996b, pp. 78-83. [2] Y. Morita, J. Kawaguchi, S. Goto and H. Ohtsuka. Performance of the M-V attitude control algorithm during its first stage flight, Proc. 7th Workshop on Astrodynamics and Flight Mechanics, Kanagawa, Japan, 1997, pp. 304-309. [3] Y. Morita, J. Kawaguchi, S. Goto and H. Ohtsuka. Design of the M-V attitude control algorithm and its flight results. Proc. 48th Congress of the International Astronautical Federation, Turin, Italy, 1997, Paper No. IAF-97-A.2.02. [4] Y. Morita, J. Kawaguchi, S. Goto and H. Ohtsuka. General character of the M-V attitude control and its second flight. Proc. 8th Workshop on Astrodynamics and Flight Mechanics, Kanagawa, Japan, 1998, pp. 213-218. [5] Y. Morita and J. Kawaguchi, Attitude control design of the M-V rocket, Philosophical Transactions of the Royal Society, Series A, Royal Society, London, United Kingdom, Vol. 359, No. 1788, November 2001, pp. 2287-2303 [6] Y. Morita, Y. and S. Goto, On the μ-synthesis of the M-V rocket attitude control design. Proc. 12th Workshop on Astrodynamics and Flight Mechanics, Kanagawa, Japan, 2002, pp. 246-251. [7] Y. Morita and S. Goto, The μ-synthesis applied to a launcher control, Proc. of the 16th IFAC Symposium on Automatic Control in Aerospace, International Federation of Automatic Control, Sankt-Petersburg, Russia, June 2004. [8] Y. Morita, An Idea of applying the μ-synthesis to launcher attitude and vibration control design, J. of Vibration and Control, Vol. 10, 2004, pp. 1243-1254. [9] Y. Morita and S. Goto, Robustness enhancement using the μ-synthesis for launcher control, Proc. 7th IASTED International Conference on Control and Applications (CA2005), International Association of Science and Technology for Development, Paper No. 460-023, Cancun, Mexico. June 2005. [10] Y. Morita and S. Goto, Design for robustness using the μ-synthesis applied to launcher attitude and vibration control, Proc. 56th Congress of the International Astronautical Federation, Fukuoka, Japan, 2005, Paper No. IAC-05-C1.3.05, also to be published by Acta Astronautica. [11] Y. Morita and S. Goto, Robust control design using the μ-synthesis for launcher attitude and vibration control and its flight results, Proc. of the 5th International Conference on Aviation and Cosmonautics, Moscow, Russia, October 2006. [12] Y. Morita,, A success in robust control design for a satellite launcher, Proc. of The IASTED Conference on Modelling, Identification, and Control (MIC 2007), Innsbruck, Austria, February 2007, Paper No. 550-067, Acta Press. [13] Y. MORITA, An evolution of the robust control design for the M-V launch vehicle, Proc. 11th International Space Conference of Pacific Basin Societies (ISCOPS), Beijing, China, May 2007/ also published by the American Astronautical Society, Advances in the Astronautical Sciences. [14] Y. Morita and S. Goto, Design for robustness using the μ-synthesis applied to launcher attitude and vibration control, Acta Astronautica, Volume 62, Issue 1, January 2008, pp. 1-8. [15] Y. Morita, On the guidance and control system of Epsilon solid rocket launcher, Proc. of the 21st Workshop on Astrodynamics and Flight Mechanics, ISAS, 2011. ***