基于神经网络的热气机功率预测

Performance Prediction for Stirling Engine Using Neural Network

为了有效地分析热气机的输出功率与运行条件的关系 ,介绍了人工神经网络 BP模型、多元回归模型及基于神经网络的组合模型 ,并将各模型用于热气机的设计阶段预测其输出性能 ,由实例给出了各模型计算机仿真试验结果 ,比较了 BP算法和组合模型算法的收敛速度 ,结果表明基于BP网络的组合模型的预测精度最高。

In order to effectively predict the output power of a stirling engine under operating conditions in its design phase,three prediction models are compared:BP neural networks,multivariate regression model and neural network based combinatorial model.Computer si- mulated results indicate that in the models under comparison,neural network based combinatorial model contains minimum prediction error.