发动机多目标优化的正交模糊神经网络方法研究
The Research on Engine Multi-objective Optimization with Orthogonal Fuzzy Neural Network
金昶明1
作者信息
- 1. 吉利汽车研究院(宁波)有限公司 浙江 宁波 315336
- 折叠
摘要
提出了一种基于正交优化方法、模糊函数和神经网络辨识系统三者相结合的多目标优化新方法-OFNN法(Orthogonal Fuzzy Neural Network).该方法综合了正交优化能够以最少的试验次数,达到与全面的试验等效结果的特性、模糊理论建立评判准则的特性以及神经网络自学习的智能特性,最终通过有限次数的仿真试验和模糊神经网络的自学习诊断,对发动机的油耗,NOx、碳烟和爆发压力找到了多目标优化合理的解决方案,该方法为多目标优化提供了一种新思路.
Abstract
Abs tract:The orthogonal optimization method,fuzzy function and neural network identification system were combined to create a new multi-objective optimal algorithm-OFNN method(Orthogonal Fuzzy Neural Network).This method combines the property of the orthogonal optimization method that equivalent results of comprehensive experiment could be obtained by the least number of experiments,the property of the fuzzy theory founding the evaluation criteria,and the property of the self-learning intelligence charac-teristics of the neural network.By limited tests with orthogonal table and fuzzy neural network self-learn-ing,the OFNN method was applied to optimize the multiple objects of fuel consumption,NOx,Soot,and cylinder peak pressure.It provides a new solution to multi-objective optimization.
关键词
发动机/多目标优化/正交/模糊/神经网络Key words
Engine/Multi-objective optimization/Orthogonal/Fuzzy/Neural network引用本文复制引用
出版年
2024
NETL
NSTL
万方数据