In this paper, we will discuss the performance, evaluation, and optimization of pattern recognition techniques for applications in system diagnostics. One reason for measuring performance of a diagnostic technique is to clearly quantify it. Another is to compare its performance with that of competing designs. We discuss traditional dichotomous performance measures as well as extensions of these methods to handle multiple classes. We describe a MATLAB toolbox that we have designed to aid developers in rapid testing and optimization. The tool allows the user to select test features, design tests, determine optimal decision thresholds and improve diagnostic performance. The toolbox is demonstrated using modeled engine data. For illustrative purposes, the performances of Partial Least Squares, Principle Component Analysis, Support Vector Machine, and Probabilistic Neural Network data-driven classifiers are compared to that of a model-based classifier developed for a particular engine using modeled data.

Abstract-Overhaul and repair services are important segments of the remanufacturing industry, and are characterized by complicated disassembly, repair and assembly process plans, stochastic operations, and the usage of rotable inventory. In view of today's time-based competition, effectively scheduling such services and managing rotable inventory and uncertainties are becoming imperative to achieve on-time deliveries and low overall costs. In this paper, a novel formulation for overhaul and repair services is presented where key characteristics, such as uncertain asset arrivals and operation processing times, and rotable parts are abstracted to model an overhaul center and multiple repair shops in a distributed framework to reflect organizational structures. Interactions between the overhaul center and repair shops are described by sets of coupling constraints across the organizations. Rotable inventory dynamics is formulated in terms of repair operation completion times and asset assembly beginning times to facilitate minimization of inventory holding costs through scheduling. A solution methodology combining Lagrangian relaxation, stochastic dynamic programming, and heuristics is developed to schedule operations in a coordinated manner to minimize total tardiness, earliness, and inventory holding costs. Additionally, penalty terms associated with coupling constraint violations are introduced to the objective function to improve algorithm convergence and schedule quality, and a surrogate optimization framework is used to overcome the inseparability difficulty caused by the penalty terms. Numerical testing results show that the new approach is computationally effective to handle rotable inventory and uncertainties, and provides high quality schedules with low overall costs for stochastic remanufacturing systems.Note to Practitioners-Overhaul and repair services for jet engines, helicopters, airplanes, are important segments of the remanufacturing industry, and are characterized by complicated disassembly, repair and assembly process plans, stochastic operations, and the usage of rotable inventory. In view of today's highly competitive business climate, effectively scheduling such services and managing rotable inventory and uncertainties are becoming critical to achieve on-time deliveries and low overall costs. In this paper, a novel formulation for overhaul and repair services is presented where key characteristics, such as uncertain asset

No abstract

In this paper we describe how a condition-based maintenance (CBM) system might be operated for operators of assets such as aircraft engines. We show that merely having a monitoring and diagnostics system in place is not enough to derive the full, or even the majority of the, benefit from CBM. Our investigations show that to maximize the benefits from CBM for the enterprise, it is as important to focus on the aftermarket supply chain — i.e. the back-end of the process — as it is to develop better data gathering, diagnostics, and prognostics techniques. Via simulation, we show that optimizing the value chain results in lower costs and turn around times, and higher asset availability, spare part availability, and fill rates. For individual equipment operators the benefits could be substantial, depending on the size of their fleets and nature of their equipment usage. Providers of maintenance services, increasingly the original equipment manufacturers, stand to gain as well. As the demands on the aftermarket supply chain become more predictable, they can service the same asset base with lower spares inventory, while meeting increasingly aggressive turn-around-time targets, thus improving the bottom line and customer satisfaction.

A high fidelity system for estimating the remaining useful life (RUL) for Li-ion batteries for aerospace applications is presented. The system employs particle filtering coupled with outlier detection to predict RUL. Calculations of RUL are based on autonomous measurements of the battery state-ofhealth by onboard electronics. Predictions for RUL are fed into a maintenance advisor which allows operators to more effectively plan battery removal. The RUL algorithm has been exercised under stressful conditions to assert robustness.