论文部分内容阅读
提出了一种基于人工神经网络的非线性组合建模与预测新方法,利用人工神经网络对复杂非线性系统的拟合能力,通过网络训练自适应地调整单一预测模型的权重,并应用MATLAB神经网络工具箱编制了非线性组合预测软件。将该方法应用于某发动机的故障预测,通过与自回归模型、灰色模型和线性组合模型的预测结果对比表明,该方法具有预测精度高,学习与泛化能力强的优点,在设备故障和可靠性预测中具有广泛的应用前景。
A new method of nonlinear combination modeling and prediction based on artificial neural network is proposed. By using artificial neural network (ANN) to fit complex nonlinear systems, the weight of single prediction model is adaptively adjusted by network training, and MATLAB neural Network Toolkit has developed a non-linear combination of forecasting software. The method is applied to the fault prediction of an engine. Comparing with the prediction results of autoregressive model, gray model and linear combination model, it shows that this method has the advantages of high prediction accuracy, strong learning and generalization ability, Sex prediction has a wide range of applications.