摘要
对注水系统输入功率、电流、电压、进口压力、出口压力等影响能效的主要因素进行分析和总结。在BP神经网络标准算法基础上,利用增加动量因子和可变学习速率改进算法,结合注水系统特性,建立符合注水能效预测模型,建立学习样本,对注水能效进行预测,并对该模型的准确度进行验证。结果表明:建立的BP神经网络模型能效预测与实测值基本吻合,满足能效预测要求。
The main influencing factors of energy efficiency of water injection system, such as input power, current, voltage, inlet pressure, outlet pressure were analyzed and summarized. Based on the standard BP neural network algorithm, using the momentum factor and variable learning rate algorithm, combing with the characteristics of water injection system, establishes energy efficiency prediction model, and learning samples to predic of water injection efficiency, and the accuracy of the model verification results show the predicted and measured the energy efficiency model of BP neural network that is consistent with the real value, which can meet the energy efficiency requirements
出处
《信息技术与标准化》
2018年第3期75-78,共4页
Information Technology & Standardization
关键词
注水系统
能效
电压
电流
BP神经网络
预测
动量因子
可变学习速率
water injection system
energy efficiency
voltage
current
BP neural network
prediction
momentum factor
variable learning rate