基于模糊神经网络的往复压缩机的性能预测研究

Study of Predication on Performance of Reciprocating Compressor Based on FNN Neutral Network

为了能够准确地预测出往复压缩机的性能参数值,利用了模糊神经网络(FNN)建立对其进行性能预测的神经网络模型,可以研究往复压缩机的流量、容积系数和效率之间的神经网络与预测关系。利用NGA遗传算法,依靠MATLBA软件实现了FNN神经计算,分别对往复压缩机的效率进行了性能预测,预测效果表明,FNN神经网络的计算模型可以提高预测效率和预测精度。

In order to predict the performance parameter value of the reciprocating compressor correctly, the neutral network model for predicting the performance of the reciprocating compressor was contracted based on the FNN neutral network and the neutral network and predicting model of the flow rate, volumetric coefficient and efficient could be studied. The FNN neutral calculation was achieved by MATLAB software and NGA genetic algorithm, the efficient of the reciprocating compressor was predicted. The results showed that the calculation model of the FNN neutral networks would improved the predicting efficient and predicting precision.