Reliability analysis of servo actuators installed in an unmanned air vehicle

This study is a summary of the reliability concept and the "why and how" to make a reliability analysis for the aviation industry framework according to the regulations, advised applications, and handbooks are put forward. Airworthiness regulations of both FAA and EASA states that reliability analysis is mandatory. Documents like aircraft recommended practices ARP-4754, ARP-4761 and military standards MIL-STD-1629A describe reliability analysis, including FHA (Functional Hazard Assessment), PSSA (Preliminary System Safety Assessment), FTA (Fault Tree Analysis), FMECA (Failure Mode, Effects and Criticality Analysis). According to the related documents, a civil aircraft's catastrophic failure, which means many casualties and even the loss of the aircraft, happens one in every million flights. Ten percent of these failures are due to sub-systems. It is assumed that there are one hundred of different sub-system originated catastrophic failures. Hence, one sub-system's catastrophic failure rate shall not be over 10-9/flight hour. For lower severity level failure conditions, probability of failure might be higher. For hazardous failures, which results with severe damage and few casualties shall have a failure probability of 10-7/flight hour or less. For major class failures, which means significant loss of functionality and small injuries, failure probability must be less than 10-5/flight hour. Similarly, 10-3/flight hour failure rate is the maximum limit for minor failure conditions, which covers slight functionality loss and one or few small injuries. Other accidents which may cause lower level safety reductions do not have a probability target. First, for a reliability analysis, a Functional Hazard Analysis (FHA) must be done. In this analysis, main airplane functions are determined and possible effects of loss of these functions are evaluated. Failure conditions are categorized, and probability limits are indicated for failures that cause function losses. This is airplane level FHA. Second, system level FHA is performed. Sub systems' effects on function losses are analysed. After FHA analysis, PSSA analysis (Preliminary System Safety Assessment) should be performed. One or more sub-system or part shall fail for a higher-level function loss. For sub-system, another lower level sub-system or part's failure is required. This top-down analysis shall be done from a function down to each part that may have a higher-level effect. Fault Tree Analysis (FTA), Dependence Diagram (DA), or Markov Analysis (MA) is suggested for this purpose. FTA, DA or MA is an important part of PSSA analysis Airworthiness regulations also demand that no single failure shall cause a catastrophic failure, even it has a failure rate lower than 10-9/flight hours. FTA, DA or MA analysis is also used in order to investigate and remove single failure conditions. A redundant or back-up part may be added, or this part's failure effect shall be reduced from catastrophic to a lower level. All top-down analysis are based on single parts' failure rates and failure modes which indicates the probability of a part's failure and possible higher-level effect of this failure. One part's failure mode, which means type of the failure, would have totally different higher-level effects. For instance, on a 4-engine passenger jet, shut-down and fire of a single engine would have very different higher-level effects. Similarly, if an actuator of the cover for landing gear is stuck open, the result is minor effect on safety with little performance loss. However, if the same actuator is stuck, the result is a bell-landing, which is a hazardous failure. Therefore, not only failure rates of the parts, but also failure mode probabilities are critical. Every single part's failure rates and failure modes shall be calculated, then these failure modes' higher-level effects shall be put forward. This part of reliability analysis has a bottom-up form. Best way to find out failure rates and modes is test results and field result. However, it's a very expensive and time-consuming method. For new designs; although tests of prototypes are possible, it is even more expensive and time-consuming and not practical at design phases. Instead historical data and empirical formulas are used as an alternative to test results. Upper system's design criteria, working conditions and assumptions are inputs for empirical formulas. For electric and electronic components, many references can be used. However, for mechanical parts' failure rates, the only available data is NSWC (Naval Surface Warfare Center)'s Handbook of Reliability Prediction Procedures for Mechanical Equipment. [27]. Failure rates shall be calculated according to the NSWC formulas. Since all similar parts' failure rates like gears or seals will be calculated from one single formula, results will not be very accurate for the particular part. Still, formulas include different environmental effects, load variances, and similar effects on part failure rates are considered Failure modes of mechanical parts are derived from historical failure modes of existing similar parts. RIAC's Failure Mode Mechanism Distribution 2013 is the reference that includes necessary information for this study. Whenever these assumptions are considered inaccurate, test results shall be used. Top-down analysis like FHA and FTA analysis and bottom-up analysis FMECA (Failure Modes, Effects, and Criticality Analysis) establish reliability analysis. This study includes an example for bottom-up part of the reliability analysis. An electric servo actuator is analysed. Failure rates of mechanical parts are calculated and failure rates for electrical parts are taken from similar parts. FHA and PSSA (FTA) analysis are not included in this thesis study. One electrical servo actuator mounted on an unmanned air vehicle's elevator and rudder is examined. For security reasons, actuator is simplified, and sizes are changed. There are 31 sub components like gears, bearing, seals, sensors etc. Main mechanical parts' failure rates are calculated according to the NSWC hand book. Simple parts like covers or joints are not calculated but estimated. Similarly, failure rates of electronic devices are not calculated and rough numbers close to real servo's electronic parts are used. For calculations, all variables like environmental conditions, working conditions, design features, speeds and forces are used on the related formulas. After failure rates, failure modes of these parts are evaluated. RIAC's failure modes are taken as an input. Unrelated failure modes, according to the design are omitted and modified failure modes are found. Undesired higher-level effects are estimated and as a result, total probabilities of higher-level effects are summarized. As a conclusion, benefits of reliability analysis and further studies are given. Most important point is that all existing documents are based on manned aircraft and safety of occupants is the main focus for these references. For an unmanned aircraft, failure rate targets must be put forward. A total reliability analysis including FHA, and PSSA is meaningful only after this.