基于BP神经网络的轴流风机的性能预测

摘要

In view of the repeated workload and long period of the series of axial fan performance test and the non-linear advantage about neural network,BP neural network model for efficiency and total pressure of axial fan was established.Forty-three groups of experimental data were used as samples to train neural network and choose the better network model,then the five groups data except samples axial fan were selected to simulate and predict the BP neural network.The results show that the average relative error of predicted values and experimental data is not more than 5%.So it is possible to use the neural network to predict the performance of axial fan.%鉴于轴流风机系列化后试验重复性工作量大、任务繁琐、周期长且成本高,本文基于神经网络处理非线性系统映射的优越性,建立了系列化轴流风机的效率和全压的BP神经网络预测模型.用43组样本点试验数据对构建的网络进行多次训练,选取较好的网络模型,最后选取了5组样本之外的数据进行性能预测仿真.预测结果表明:预测值与试验值的平均相对误差不超过5%,验证了神经网络实现对轴流风机的性能预测是可行的.