基于邻域粗糙集和灰狼算法优化Elman的民航发动机滑油量预测

实时预测民航发动机滑油量对保障飞行安全具有重要意义.针对滑油量受发动机多个工作状态的多个参数影响,具有影响参数多,提取方法不确定等问题,提出了一种基于邻域粗糙集(neighborhood rough set,NRS)和灰狼优化(grey wolf op-timizer,GWO)-Elman相结合的方法预测滑油量.首先通过邻域粗糙集提取对滑油量重要度高的发动机工作阶段,将提取后的工作阶段有关参数作为特征向量输入到灰狼优化-Elman的网络模型中,灰狼算法通过计算和比较个体的适应度来优化El-man网络中的权值和阈值,保证Elman网络中的权值和阈值达到全局最优.预测结果表明,精度达到98.44％,满足工程应用的精度要求.研究结果为及时监测民航发动机滑油系统的健康状况提供理论依据.

Prediction of Aviation Engine Oil Quantity Based on NRS-GWO-ENN

The real-time prediction of engine oil consumption can find problems in time when the oil consumption exceeds the normal value. Because the oil volume is affected by multiple parameters of the engine''s multiple working conditions, there are many influencing parameters and the extraction method is uncertain, a method based on the combination of neighborhood rough set and grey wolf optimization Elman was proposed to predict the oil quantity. Firstly, the engine working stage with high importance to the oil quantity was extracted by neighborhood rough set, and the relevant parameters of the extracted working stage were input into the Elman network model of grey wolf optimization as feature vectors. The gray wolf algorithm optimized the weights and thresholds in Elman network by calculating and comparing individual fitness, so the algorithm can ensure that the weights and thresholds in Elman network reach the global optimization. The prediction results show that the accuracy reaches 98.44%, which meets the requirements of engineering application. This study provides a theoretical basis for timely monitoring the lubricating oil system health status of civil aviation engines.