Sensor Selection and Optimization for Aerospace System Health Management under Uncertainty Testing

概要

論文の詳細を見る
Prognostics and health management (PHM) has an important part in aerospace systems. Information sensing and testing are the bases of PHM, and design for testability (DFT) developed concurrently with system design is considered a fundamental way to improve PHM performance. The traditional DFT, which is only based on the requirements of fault detection and isolation, is not suitable for sensor design and optimization for PHM. Aiming to solve this problem, the intrinsic requirements of PHM for testability are firstly analyzed qualitatively and the corresponding testability indexes are defined quantitatively. Then, a sensor selection/optimization process for PHM is presented. Fault detection uncertainty is also analyzed systematically from the view of fault attributes, sensor attributes and fault-sensor matching attributes, respectively. Based on the requirements and process, the object and constraint models of sensor optimization selection problem are studied in great detail. For aerospace system health management, a sensor optimization selection model is constructed that treats sensor total cost as the objective function and the proposed testability indexes under uncertainty test as constraint conditions. Due to the NP-hard property of the model, a generic algorithm (GA) is introduced to obtain the optimal solution. The application examples show that the proposed model and algorithm are effective and feasible.