摘要
研究水泵全特性曲线的神经网络预测模型。通过计算机实验,讨论样本、学习算法和网络结构等对神经 网络预测模型性能的影响及其改进措施。实验结果表明,神经网络预测模型具有较佳的学习能力和泛化能力。
ANN predication model for water pump complete characteristics curve has been studied. Through computer simulation, samples, BP algorithms and the influence of network structure neurula on model performance have been discussed as well as the improving measures. Experiment result shows that ANN prediction model is characterized by better learning capacity and generalization capacity.
出处
《炼油技术与工程》
CAS
北大核心
2004年第12期24-27,共4页
Petroleum Refinery Engineering
