抽水蓄能电站技术供水系统智能调节方法研究  

Research of intelligent adjustment method for technical water supply system of pumped storage power station

作　　者：袁静 杭晨阳 方杰 张新 郑晓楠 杨杰[2] YUAN Jing;HANG Chenyang;FANG Jie;ZHANG Xin;ZHENG Xiaonan;YANG Jie(PowerChina Huadong Engineering Corporation,Hangzhou 310000,Zhejiang,China;School of Electrical and Power Engineering,Hohai University,Nanjing 211100,Jiangsu,China)

机构地区：[1]中国电建集团华东勘测设计研究院有限公司,浙江杭州310000 [2]河海大学电气与动力工程学院,江苏南京211100

出　　处：《水利水电技术（中英文）》2024年第10期148-155,共8页Water Resources and Hydropower Engineering

基　　金：国家自然科学基金重点项目(52339006);国网新源集团(控股)有限公司科技项目(SGXYTB00YWJS2100090)。

摘　　要：【目的】在抽水蓄能电站技术供水系统的实际运行过程中,对于冷却水量的调节较为粗放,其历史数据并不能直接用于研究,因此考虑将机组温度作为中间变量,通过研究工况-温度-流量的映射关系,实现技术供水系统的智能调控。【方法】针对该问题,首先通过建立改进的长短期神经网络(LSTM)技术供水系统温度预测模型,将负荷等工况参数作为输入量预测技术供水对象的温度,随后通过Fluent仿真建立温度与流量的映射关系,最后根据拉依达准则计算得到的温度阈值判断是否报警,并给出推荐流量。【结果】实例计算结果显示:所提改进LSTM方法相较于其他方法,预测精度至少提高58.45%,同时仿真所得推荐流量基本符合电站实际运行数据,在满足安全运行的前提下节水量至少达到2.33%。【结论】结果表明:所提方法可以对目标温度做到较为精准地预测,并在此基础上通过内置仿真模型,实现更合理地分配用水量,对抽水蓄能电站技术供水系统智能调节具有一定指导意义。[Objective]During the actual operation of the pumped storage power station′s water supply system,the regulation of cooling water volume is relatively crude.Historical data cannot be directly used for research.Therefore,considering the unit temperature as an intermediate variable,the intelligent regulation of the technical water supply system is achieved by studying the mapping relationship between operating conditions,temperature,and flow rate.[Methods]To address this issue,an improved Long Short-Term Memory(LSTM)model for predicting the temperature of the technical water supply system is first established,with load and other operating condition parameters as input to predict the temperature of the technical water supply object.Subsequently,the fluent simulation is used to establish the mapping relationship between temperature and flow rate.Finally,the temperature threshold obtained by the Pauta criterion is used to determine whether an alarm should be raised and to provide the recommended flow rate.[Results]The result of the case study show that the proposed improved LSTM method improves the prediction accuracy by at least 58.45%compared to other method.At the same time,the recommended flow rate obtained from the simulation basically matches the actual operating data of the power station,achieving a water saving of at least 2.33%while ensuring safe operation.[Conclusion]The result indicate that the proposed method can achieve a relatively accurate prediction of the target temperature and,based on this,realize a more rational allocation of water through the built-in simulation model.This has a certain guiding significance for the intelligent regulation of the pumped storage power station′s technical water supply system.

关 键 词：抽水蓄能电站 技术供水系统 改进长短期神经网络LSTM Fluent仿真 阈值报警 

分 类 号：TV743[水利工程—水利水电工程]