基于改进鲸鱼算法的汽轮机主蒸汽流量预测模型  

Prediction Model of Main Steam Flow of Steam Turbine Based on Improved Whale Algorithm

作　　者：陈国峰 CHEN Guofeng(Tianji Power Plant of Huaihu Coal Power Co.,Ltd.,Huainan 232000,China)

机构地区：[1]淮沪电力公司田集电厂,安徽淮南232000

出　　处：《电工技术》2023年第14期26-28,共3页Electric Engineering

摘　　要：为了提高汽轮机主蒸汽流量预测结果的精准性,为火电机组运行的经济性优化策略提供可靠的数据基础,基于改进鲸鱼算法,设计了一种汽轮机主蒸汽流量预测模型。根据汽轮机的运行特征,采集并处理汽轮机现场运行数据,应用BP神经网络算法,建立汽轮机主蒸汽流量预测模型,并利用改进鲸鱼算法优化、修正预测模型的权值参数,提高模型的收敛精度与速度。检测结果显示,对于随机的200组测试样本数据,该方法对汽轮机主蒸汽流量的预测结果曲线与实际蒸汽流量曲线基本一致,预测结果的均方误差均值为4.3521,平均绝对百分比误差均值为2.18%,预测时间均值为2.6125 s。结果表明,改进鲸鱼算法的应用,提高了汽轮机主蒸汽流量预测模型的集成度与收敛精度,使模型具有高效性与精准性,为机组热损耗量的降低等优化工程提供了精准的数据基础。In order to improve the accuracy of main steam flow of steam turbine prediction results and provide a reliable data basis for the economic optimization strategy of thermal power unit operation,a prediction model of main steam flow of steam turbine is studied and designed based on the improved whale algorithm.According to the operation characteristics of steam turbine,according to the operation characteristics of the steam turbine,the on-site operation data of the steam turbine was collected and processed,and the BP neural network algorithm was applied to establish the steam flow prediction model of the steam turbine main.By improving the whale algorithm,the weight parameters of the prediction model are optimized and corrected,and the convergence accuracy and speed of the model are improved.The test results show that for 200 sets of random test sample data,the prediction curve of the main steam flow of the steam turbine is basically consistent with the actual steam flow curve,and the mean squared error of the prediction results is 4.3521,the average absolute percentage error is 2.18%,and the average prediction time is 2.6125 s.The results show that the application of the improved whale algorithm improves the integration and convergence accuracy of the steam flow prediction model of the steam turbine,makes the model efficient and accurate,and provides an accurate data basis for the optimization project such as the reduction of heat loss of the unit.

关 键 词：改进鲸鱼算法 汽轮机 主蒸汽流量预测 预测模型 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]