Evaluation and decision of task aircraft usage quality based on wavelet neural network

Abstract

The quality status of aircraft use involves multiple indicators, and has nonlinear correlations. The commonly used traditional evaluation methods often rely heavily on expert knowledge and cannot depict the nonlinear relationship of indicators, making it difficult to accurately and effectively support mission aircraft usage decisions. This article constructs one mission aircraft usage quality status indicator system that includes annual usage plan execution rate, matching degree of aircraft lifespan usage, variability of the remaining lifespan echelon difference, rate of satisfaction with aviation materials supply, and aircraft failure free rate. Wavelet Neural Network and BP Neural Network models are respectively used to predict and evaluate the quality status. Case validation shows that Wavelet Neural Network have faster data learning speed and higher prediction accuracy compared to BP Neural Network, which are basically consistent with actual results and can provide more accurate quantitative basis for mission aircraft usage decisionmaking.