内燃机特性曲线绘制的神经网络技术  

Study on Characteristic Curve Painting of Internal Combustion Engine by Neural Network Technology

作　　者：王兴海[1] 张谢恩[2] 宁智[2] 李玉珍[1] 马震[1] 

机构地区：[1]西北工业大学365研究所,陕西西安710072 [2]北京交通大学机电工程学院,北京100044

出　　处：《航空计算技术》2007年第5期1-4,共4页Aeronautical Computing Technique

基　　金：国家自然科学基金(40475048)

摘　　要：人工神经网络技术具有自动获取知识、联想记忆、自适应性、良好的容错性和推广的能力,处理内燃机特性曲线具有效果好、程序推广性强等特点,在内燃机台架性能试验自动化方面有着非常重要的作用。采用改进的神经网络BP算法,对内燃机万有特性曲线进行了训练和模拟,取得了良好的效果。Artificial neural networks have the abilities of obtaining knowledge automatically, associate anamnesis, self-accommodating, accommodating mistakes well and favorable popularizing. The technology of neural networks has the characteristics of precision and popularized easily in painting the characteristic curve of internal combustion engines, and has important effect in the aspect of bed-test automation of internal combustion engines. The modified arithmetic of BP neural networks is used for the training and painting of the characteristic curve of internal combustion engines this paper. The results showed that the method of painting characteristic curve of internal combustion engines by the technology of neural networks is feasible.

关 键 词：内燃机 特性曲线 神经网络 

分 类 号：V263.4[航空宇航科学与技术—航空宇航制造工程]