Health Condition Monitoring of Aero-engine with Known Clustering Number Based on Ant Colony Algorithm

摘要

An algorithm based on ant colony algorithm for health condition monitoring of aero-engine was put forward. The algorithm conversed the health status classification of aero-engme into solving the clustering-based optimization problem with constrain. Ant colony algorithm based on colony collaboration and learning could solve this clustering problem. The proposed algorithm after being optimized by BP neural network was applied to monitor health condition of aero-. engine. The emulation result shows that the algorithm has the merits of simple realization, fast convergence, strong' parallelism and robustness, high identification accuracy and high reliability, and is fit for health condition monitoring of aero-engine with low demands on fault samples and with known clustering number.