Development of Diagnostics and Prognostics Methodology for Electronic System Health Monitoring
CALCE Team: Sachin Kumar and Prof. Michael Pecht
Objectives: Develop a Methodology for Baseline Characterization of Electronic System. Perform experiment to establish baseline for electronic system under various environmental and usages conditions. Perform same set of experiment on a different system of same model to verify the repeatability of experiments conducted at earlier stage of work. Perform set of field experiments on same system model to validate the baseline established earlier. Perform set of field experiments on different system model to determine the applicability of baseline from previous study and characterize new system model based on earlier system model.
Introduction: The desire and need for real predictive prognostics capabilities have been around for as long as man has operated complex and expensive systems. Real predictive prognostic capabilities are just one element among many interrelated and complementary functions in the playing field of Prognostic and Health Management (PHM). Any industry would like to have maximum operation availability, low warranty returns, very low Returned Tested OK (RTOK), low number of spares required and small logistics footprint, maximum life usage and no false alarms etc. One would like to be able to have the ability to accurately predict future health status, anticipate problems and should prepared ahead of being surprised by “hard downing (failure or recall)” events. Prognostic capabilities would let planner see ahead of both unplanned and even some necessary preventive events. Several challenges are involved with developing and implementing prognostic capabilities with new performance based logistic support concepts. Similarly, several technical issues are involved are consideration of dynamic operating environment, unknown material characteristics, unseen events, absence of unique mathematical technique which can address all kinds of fault and failure in various system and issues of uncertainty. Electronics as a product and as an integral part of many system provides functionality and control for mechanical or electrical systems. Electronic products with a long life cycle ensure customer satisfaction. Increased warranties and severe liability with product failures forces manufacturers to analyze systems performance during their field operations and to determine their operational availability. This means electronic systems should be considered equal in comparison with any other systems for prognostics and health management. Assessment of electronic systems performance will be more informed by considering information of its in-service use, operation, and environmental conditions. This raised interest in in-situ monitoring system health to predict failures and provide warnings in advance of failure. It can be argued that decision-making goals of diagnostics and prognostics are different but studying them separately is not feasible. Diagnostic results are utilized for maintenance purposes whereas prognostic results are used to prevent or avoid risk in terms of cost or life and to maximize the remaining life of replaceable components while minimizing operational risks. Combination of information from both the field provides better understanding of system that helps in improving system design that will not only improve system reliability but also reduce maintenance time. The goal is to develop a methodology for baseline characterization of any given electronic systems with reasonable amount of experiments and data collection. Develop a diagnostic methodology that utilizes system baseline information and identify variations in system performance due to presence of any anomalies or system failure. Development of prognostic methodology based on features extracted from the data or critical parameters identified.