基于改进遗传算法优化BP神经网络的螺杆式制冷机组能效预测  

Energy Efficiency Prediction of Screw Chiller Based on Improved Genetic Algorithm and Optimization of BP Neural Network

作　　者：梅年丰 严良文[1] 钱峰峰 李威 於鹏 陈佳乐 Mei Nianfeng;Yan Liangwen;Qian Fengfeng;Li Wei;Yu Peng;Chen Jiale

机构地区：[1]上海大学机电工程与自动化学院,上海200072

出　　处：《计量与测试技术》2018年第5期13-16,共4页Metrology & Measurement Technique

摘　　要：为提高螺杆式冷水机组能效预测的准确性,本文提出了基于改进遗传算法(IGA)优化BP神经网络的制冷机组能效预测方法,利用供冷负荷、冷冻水流量、冷冻水出水温度、冷却水回水温度与COP 5个参数训练IGA-BP预测模型。本文以某酒店的螺杆式制冷机组为研究对象,利用该方法对机组的能效进行预测,并与标准遗传算法优化BP神经网络的预测结果进行比较,结果表明:改进后的算法计算能力有了很大的提高,SGA-BP预测模型的最大相对误差为6.73%,平均相对误差为2.96%;IGA-BP预测模型的最大相对误差为3.45%,平均相对误差为1.16%。IGA-BP的预测精度得到了较大改善。In order to improve the accuracy of the prediction of the energy efficiency of screw chillers, this paper proposes a BP neural network based on improved genetic algorithm （IGA） to predict the energy efficiency of refrigerating units, using the cooling load, cooling water flow, outlet temperature of chilled water, cooling water temperature and the 5 COP parameters training IGA - BP prediction model. This paper takes a hotel＇s screw refrigeration unit as the research object, and uses this method to predict the energy efficiency of the unit, and compares it with the prediction result of the standard genetic algorithm optimized BP neural network, the results show that the computational ability of the improved algorithm has been greatly improved. The maximum relative error of SGA - BP prediction model is 6. 73% , the average relative error is 2. 96% , the maximum relative error of IGA - BP prediction model is 3.45% , and the average relative error is 1.16%. The prediction accuracy of IGA - BP has been greatly improved.

关 键 词：制冷机组 能效预测 改进遗传算法 BP神经网络 

分 类 号：TB9[一般工业技术—计量学]