弓型折流板换热器动态响应神经网络预测

采用ARX动态模型设计了一种用于换热器动态特性辨识及预测的人工神经网络结构模型,基于有限的实验数据将人工神经网络技术应用于以水、油为换热工质的弓型折流板换热器动态特性的预测当中.采用Levenberg-Marquardt算法对网络进行训练和测试,辨识和预测了换热器单侧流量扰动情况下油侧出口温度的响应情况及换热器油侧进口发生温度扰动情况下油侧出口温度响应情况,并与数值计算进行了对比,神经网络的预测结果好于数值计算.分析了神经网络模型的泛化能力.计算结果表明,人工神经网络对于复杂系统的动态辨识及预测是相当成功的,与基于数理模型的数值预测相比,有更好的预测精度,且模型的泛化能力也很强.

Prediction of the dynamic behaviors of the heat exchangers with segmental baffles by artificial neural networks

A artificial neural networks (ANNs) model for the identification and prediction of the dynamic behavior of the heat exchangers is designed using ARX dynamic model.Based on finite experimental data,the dynamic behavior of the heat exchangers which use water and oil as wording media can be predicted by using the model.The oil side outlet temperature responses of the heat exchanger with the variation of its single-side flow-rate and the variation of the oil side inlet temperature of it are identified and predicted by means of Levenberg-Marquardt algorithm.The prediction results are compared with numerical results,it is shown that the neural network prediction results have higher precision and better generalization performance.At the same time,it is shown that ANN is competent for the dynamic identification and prediction of complicated systems.