基于免疫粒子群优化的不等时间间隔发动机性能综合指数组合预测  

Combined Prediction Research of Unequal-Interval Engine Performance Synthetical Index Based on Particle Swarm Optimization with Immunity

作　　者：李冬[1] 马力[2] 樊照远[2] 宋岩[3] 

机构地区：[1]海军航空工程学院研究生管理大队,山东烟台264001 [2]海军驻沈阳地区发动机专业军事代表室,辽宁沈阳110015 [3]海军航空工程学院基础实验部,山东烟台264001

出　　处：《燃气涡轮试验与研究》2013年第2期42-45,共4页Gas Turbine Experiment and Research

基　　金：国家自然科学基金青年基金(61102167)

摘　　要：首先利用奇异值分解滤波算法,对测量参数进行滤波处理,进而合成发动机性能综合指数。针对性能综合指数为不等时间间隔的情况,在改进灰色预测模型中引入新陈代谢思想,提出一种基于免疫粒子群优化权值的改进灰色模型与支持向量机相结合的性能指数预测方法。仿真实例表明:组合模型的预测精度明显高于改进灰色预测模型,略高于支持向量机模型;且对于大多数样本点,组合模型的预测结果变化更加平稳。Firstly the measured parameters were filtered by Singular Value Decomposition Filtering Algo- rithm, and engine performance index was synthesized. Aimed at the problem of different interval for perfor- mance synthetical index, metabolism was introduced in improved gray predicting model, and a performance index predicting method of improved gray model and support vector machine （SVM） combination based on particle swarm optimization with immunity was presented. Simulating example indicates that the predicting precision of this method is more superior to improved gray model, better than SVM. But predicting result change of combined model is more stable for most sample points..

关 键 词：不等间隔 发动机性能综合指数 改进灰色模型 支持向量机 免疫粒子群优化 组合预测 

分 类 号：V235.13[航空宇航科学与技术—航空宇航推进理论与工程]