In-service reliability assessment and top-down approach provides alternative reliability prediction method

The desire to develop a realistic Mean-Time-Between-Failure (MTBF), Mean-Time-To-Failure (MTTF), or failure rate prediction method, which doesn't assume that only piece part faults cause system failures, has led to a new methodology involving in-service reliability assessment and a top-down approach. This methodology reduces uncertainty in the reliability predictions by relying primarily on recent Line Replaceable Units (LRUs) field failure data history of inservice LRU designs with verified field performance that are similar in hardware and function to new LRU designs. Experience demonstrates that this new method as applied to hardware used in the commercial avionics industry reflects a high degree of accuracy on the MTBF prediction process applied to the new LRU designs. The Honeywell In-Service Reliability Assessment Program (HIRAP) method is an in-service reliability assessment process and top-down approach. It was developed as a viable alternative to MIL-HDBK-217 (Military Handbook, Reliability Prediction of Electronic Equipment) for MTBF predictions of new systems, end item, or LRU designs. It is noteworthy that MIL-HDBK-217's stated purpose is as a uniform method for predicting reliability for comparative purposes, not to predict "in-use" reliability. HIRAP was developed for the purpose of predicting or assessing "in-use" reliability as well as a circuit design comparison measure in design trade studies. The HIRAP methodology includes an assessment of end item field experience, a top-down prediction approach, a comparison of fielded systems to new designs, and an application of algorithms to ascertain the MTBF of a new design, when there is some degree of similarity to predecessor designs in-service. A comparison between the physical models, attributes and characteristics of an in-service system and a proposed design are sufficient to predict a new design's MTBF. In addition, the effort required to complete the new design's LRU level MTBF prediction is significantly less since LRU reliability predictions are calculated from predecessor LRU MTBFs. There is no need to calculate LRU level MTBFs from the bottom up unless no similarity is evident between predecessor designs and the new design. Two types of algorithms used in characterizing the failure rates of in-service designs are similarity and complexity algorithms. The similarity and complexity algorithms configure five categories of a physical model. These categories are active piece parts or components, passive piece parts or components, circuit interconnects, solder joints, and other end item, system, or LRU failure causes not related to the previous four categories. Various attributes and characteristics ate used to compare fielded and new designs, such as physical attributes, performance, technology maturity, component improvements, design process, and manufacturing effectiveness. Similarity analysis is used when the new product evolves from a predecessor. Similarity analysis starts by comparing functions, attributes, and the electronic hardware's bill-of-material (BOM) between the in-service design and the new design. The complexity algorithm is used when no similar LRU design exists in-service, but there are similar part types, circuits, or functions within fielded LRUs. The in-service design is compared to the new design at the Circuit Card Assembly (CCA) level and the piece part type level. LRU functions of the new design are identified in terms of percentages of the predecessor design. The output from a complexity algorithm could be a functional relationship based on percentage of field failure rates of LRU circuitry on several CCAs. The output could also be in the form of piece part type level failure rates based on piece part similarities or part type complexities that ate similar.