AVIONICS Challenges in Reliability Prediction of Aircraft Subsystems Raghuram R HCL Technologies, India. D e c e m b e r 2 0 0 8
TABLE OF CONTENTS Abstract 3 Introduction 3 Reliability Prediction Improvement 4 Standby Redundancies 5 Duty Cycle Considerations 5 Sensitivity Measures 6 Limitations of MIL-217 8 References 9
Abstract As the new generation Aircraft Subsystems has more features and design advancements compared to predecessor systems, it will be increasingly complex to perform the Reliability Prediction of subsystems and maintain the compliance towards safety requirements. This paper discusses the various challenges in performing Reliability Prediction of Aircraft Subsystems during the system design phase and provides option on revising the existing reliability equations and empirical formulae based on the past experience. The Operational Reliability data obtained from previous generation Aircraft subsystems provide the first level results of reliability prediction. These results will help to identify the weak parts\ elements and provide insight into the areas, which can be improved to enhance the inherent reliability of new generation system. This paper also explores the relationship between Aircraft subsystems reliability and individual component reliability estimates. The sensitivity and redundancy studies are performed to analyze the system-level impact of component-level reliability data. Finally, this paper presents useful information to assess subsystem reliability improvement strategies Introduction The new generation Aircraft Subsystems will be designed and developed using concurrent engineering processes to reduce the aircraft development cycle time and accelerate the time to market initiatives. In order to assess the Operational and Safety performance of the aircraft, it is imperative to predict the reliability at the early stages of product development process so that customer requirements can be taken care in the product design. A new generation system will likely have some subsystems and components that have been used previously so that there
lot of reliability prediction data that can be reused. However, these systems can also have newer-technology components (e.g. 64-bit processors, Memory devices with >1 M bits) with very little reliability data available. For these components, reliability estimation can be based on extrapolated accelerated life testing or an analytic physics-offailure model. There are limitations within each of these sources. These limitations need to be analyzed to decide how limited resources should be allocated to improve reliability prediction Reliability Prediction for aircraft subsystems is often performed iteratively as the design evolves. For some systems, a preliminary reliability prediction is determined using previously used data to initially compute reliability predictions at the assembly or subsystem-level. Then, as more detailed design information becomes available, the assembly-level predictions are updated or replaced with reliability predictions based on the components within the assemblies. For other system-reliability predictions, a parts-count reliability prediction is initially determined, once the parts within the system have been identified, but prior to determining the specific stresses acting on the part. Then, when more detailed design information has been determined, the system-reliability prediction is revised, and hopefully improved, based on a parts-stress reliability prediction and part-level models. When reliability prediction is conducted in stages, it is advisable to prioritize reliability-prediction improvement activities. The uncertainty of reliability predictions for some components might have negligible effect relative to the system-reliability prediction. In these cases, it might not be efficient to devote any additional effort to improving the prediction by decreasing the estimation variance. This requires additional and more detailed analyzes or testing that might yield unimportant system-level improvements. Alternatively, the reliability predictions for some components have a relatively large effect on the variance associated with the system-reliability estimate. Then, additional analyzes or testing is warranted. Reliability Prediction Improvement The Reliability Block Diagram (RBD) is popularly used to model the complex aircraft subsystems that cannot be computed using the normal reliability prediction analysis. A reliability prediction analysis can provide important calculated values like failure rate, MTBF, reliability, and availability. These calculations are based on established reliability prediction models. As new generation aircraft subsystem configurations become more complex, more complex calculation methods are required to calculate values like failure rate, MTBF, reliability, and availability. The RBD method provides the ability to perform these complex calculations. The Reliability Block diagrams are also used to concise visual shorthand the various series-parallel block combinations (paths) that result in item success
Standby Redundancies The Standby redundancy techniques example is demonstrated using RBD modeling of secondary power subsystem of an aircraft. Standby redundancy offers the ability to protect a system through the use of cold, warm, or hot standby units and junctions. In order to improve the Reliability of subsystem, the redundancy analysis is carried out by adding redundant element to each of the existing elements of the subsystem to see which element yield better improvement in overall system reliability. It has been found that by adding redundant element has negligible impact for the most of the cases except for igniter module, where substantial improvement in the System Reliability is observed. Using this RBD analysis data, initiatives are taken to include additional igniter element at the design stage. Following figure shows the RBD model of the subsystem Controller Module Pressure Sensor Temperature Sensor Speed Sensor Inlet Valve Outlet Valve Igniter Engine Redundant Igniter Figure 1: RBD Model The following table shows the reliability percentage with redundant elements: Subsystem Configuration Reliability % Configuration (without redundant element) 99.99906% Redundant Controller Module 99.99906% Redundant Pressure Sensor 99.99908% Redundant Temperature Sensor 99.99906% Redundant Speed Sensor 99.99906% Redundant Inlet Valve 99.99906% Redundant Outlet Valve 99.99906% Redundant Generator 99.99906% Redundant Igniter 99.99944% Table1: Reliability data Duty Cycle Considerations Duty Cycle Considerations In order to correctly assess the overall failure rate of the system and improve the MTBF value, we have computed the operating failure rate values and non-operating (dormant) failure rate values of the parts and used the following equation to arrive at the overall part failure rate. For most of the cases the non-operating (dormant) failure rates of the parts are negligible, however it cannot be ruled out.
λ Part = λ operating * Duty Cycle + λ non-operating * (100 Duty Cycle) In secondary power subsystem the igniter module operates only during the start sequence of the engine. Hence the duty cycle formula is applied to compute the effective failure rate of the part. Using the duty cycle methods for all the pulsed operations the overall MTBF value of the subsystem is improved by 15%. Sensitivity Measures Sensitivity measures are used to prioritize the most important or critical components within a system. These measures indicate which components contribute most to the variance of the system-reliability estimate, and thus, which components might require additional testing or analysis to improve the component-reliability estimate. The following figure provides the details on failure rate distribution values of electronic parts for some of the important aircraft subsystems. It is apparent to note from the below pie chart that Capacitors are major contributors of electronics reliability of the subsystems. This is arrived by carefully analyzing the Reliability prediction reports and averaging the failure rate data of different aircraft systems. (The subsystem names and their failure rate values are not provided in this paper) Failure Rate Distribution 45% 30% Capacitors ICs 25% Others (Resistors, Inductors, Transistors, Diodes, Connectors, PCB) Figure 2: Failure Rate Distribution At the outset, it is important to reduce the failure rate values of the capacitors in order to reduce overall failure rate and improve the system MTBF. Most common reliability prediction techniques by the aircraft industry is empirically driven MILHDBK-217. Though it is popular and widely accepted in industry, they do not provide correct values in a numerous of situations. Therefore, during the last few years this method s popularity has been gradually declining, mostly due to proliferation of new electronic packaging technologies, continuous improvement in quality and reliability, and thus subsequent inability of MIL-HDBK-217 to make accurate failure rate predictions.
However, most of the mathematical models in MIL-HDBK217 along with the relevant principles of physics remain largely valid. This paper presents an attempt to further enhance the process of reliability prediction of capacitors and hence the overall system reliability by adjusting the existing empirical equations of Capacitor s Reliability prediction. The Failure rate values of capacitors will be improved lot by selecting established reliability and highest screened parts, however this will increase the cost and it is not a prudent solution unless the design/application warranted. The MIL-217 empirical equation for failure rate computation of a capacitor is given by = λ b π T π C π V π SR π E failures per 10 6 hours Nomenclature description: Part Failure rate λ b Base Failure rate π T Temperature Factor π C Capacitance Factor π V Voltage Stress Factor π SR Series Resistance Factor Quality Factor π E Environment Factor Based on field data, test data and related experience data, the Change in Value failure mode of the capacitor has very minimal or no impact in the system operation or performance, especially when the capacitor is used for decoupling purposes, hence there is a need to update the existing MIL217 empirical formula to derive the failure rate based on specific application of Capacitors. The existing MIL-217 formula is updated as (Modified) = λ b π T π C π V π SR π E * πa Failures per million hours i.e. (Modified) = (Existing) * π A π A is Adjustment Factor. π A = 1; for non-decoupling application π A = 0.65; for ceramic capacitors used for decoupling application Following are other important points that need to be considered to improve the reliability of capacitors: a) Select the Derating factors for the parts well below the standard derating guideline value (60%) b) Minimize the number parts with the help of appropriate design tradeoffs
Limitations of MIL-217 The MIL-HDBK-217F standard will not support the failure rate computation for some of the important parts that are used in the new generation product design. These include, but not limited to: 1 Memory devices with size more than 1 mega bits 2 64 bit processors Failure rate computation for complex memory devices using MIL-217: The failure rate equation for the memory devices as per MIL217 handbook is given as = (C1π T + C2π E + λ cyc ) π L failures per 10 6 hours Nomenclature description: C1 C 2 π T π E λ cyc π L Part Failure rate Die complex Failure rate Package Failure rate Temperature Factor Quality Factor Environment Factor Read/write cycle induced Failure rate Learning Factor In the above equation, the Die complex Failure Rate (C1) values for the memory devices having storage bits size more than 1 M bits is not provided in the MIL- 217 handbook. The C1 values for memory devices with storage size more than 1 M bits are derived by linearly extrapolating the existing values provided in the MIL-217 handbook. These values are shown in the below table. Memory Size, B (Bits) ROM MOS PROM, UVEPROM EEPROM EAPROM DRAM SRAM (MOS & BiMOS) Table 2: Die Complexity Failure Rate for Memories C1 ROM PROM Bipolar SRAM Up to 16K 0.00065 0.00085 0.0013 0.0078 0.0094 0.0052 16K < B < 64K 0.0013 0.0017 0.0025 0.016 0.019 0.011 64K < B < 256K 0.0026 0.0034 0.0050 0.031 0.038 0.021 256 K<B<1 M 0.0052 0.0068 0.010 0.062 0.075 0.042 1M< B< 4 M 0.00104 0.00136 0.020 0.124 0.150 0.084 4M<B<16M 0.00208 0.00272 0.040 0.248 0.300 0.168
Failure rate computation for 64-bit processors using MIL 217: The failure rate equation for the Microprocessor devices as per MIL-217 handbook is given as = (C1π T + C2π E ) π L failures per 10 6 hours The MIL-217 handbook provides the C1 values up to 32-bit processors. The C1 values for 64-bit Microprocessors are derived by linearly extrapolating the existing values provided in the MIL-217 handbook. These values are shown in the below table. No of bits Bipolar MOS C1 C1 Up to 8 bits 0.060 0.14 Up to 16 bits 0.12 0.28 Up to 32 bits 0.24 0.56 Up to 64 bits 0.48 0.108 Table 3: Die Complexity Failure Rate for Processors C1 The failure rate values computed using this approach is almost inline with manufacturer s warranty data for both memory devices and processors References [1] Military handbook -Reliability prediction of electronic equipment. MIL- HDBK-217F-Notice 2 [2] Military handbook Electronics Reliability Design Handbook. MIL-HDBK- 338B [3] Failure Mode Distribution -91 Handbook CONTACT Raghuram R HCL Technologies LTD 64-66, SP, 2 nd Main Road Ambattur Industrial estate Ambattur Chennai-600 058 E mail:raghuramr@hcl.in Web: http://www.hcltech.com
CUSTOM APPLICATION SERVICES ENGINEERING AND R&D SERVICES ENTERPRISE APPLICATION SERVICES ENTERPRISE TRANSFORMATION SERVICES IT INFRASTRUCTURE MANAGEMENT BUSINESS PROCESS OUTSOURCING About HCL About HCL Enterprise HCL Enterprise is a $5 billion leading Global Technology and IT Enterprise that comprises two companies listed in India - HCL Technologies & HCL Infosystems. The 3-decade-old enterprise, founded in 1976, is one of India s original IT garage start-ups. Its range of offerings spans Product Engineering, Custom & Package Applications, BPO, IT Infrastructure Services, IT Hardware, Systems Integration, and distribution of ICT products. The HCL team comprises over 58,000 professionals of diverse nationalities, who operate from 20 countries including 360 points of presence in India. HCL has global partnerships with several leading Fortune 1000 firms, including leading IT and Technology firms. For more information, please visit www.hcl.in. About HCL Technologies HCL Technologies is a leading global IT services company, working with clients in the areas that impact and redefine the core of their businesses. Since its inception into the global landscape after its IPO in 1999, HCL focuses on transformational outsourcing, underlined by innovation and value creation, and offers integrated portfolio of services including software-led IT solutions, remote infrastructure management, engineering and R&D services and BPO. HCL leverages its extensive global offshore infrastructure and network of offices in 19 countries to provide holistic, multi-service delivery in key industry verticals including Financial Services, Manufacturing, Aerospace & Defense, Telecom, Retail & CPG, Life Sciences & Healthcare, Media & Entertainment, Travel, Transportation & Logistics, Automotive, Government and Energies & Utilities. HCL takes pride in its philosophy of Employee First which empowers our 52,714 transformers to create a real value for the customers. HCL Technologies, along with its subsidiaries, had consolidated revenues of US$ 2.0 billion, as on 30th September 2008. For more information, please visit www.hcltech.com