基于改进预测模型的水电机组状态趋势预测  

State Trend Prediction of Hydroelectric Units Based on Improved Prediction Model

作　　者：陈畅 张毅[1,2] 段炼达 王聪 CHEN Chang;ZHANG Yi;DUAN Lian-da;WANG Cong(China Institute of Water Resources and Hydropower Research,Buijing 100038,China;Beijing IWHR Technology Co.Ltd.,Beijing 100038,China;CREC Cloud Net Information Technology Co.Ltd.,Beijing 100039,China)

机构地区：[1]中国水利水电科学研究院,北京100038 [2]北京中水科水电科技开发有限公司,北京100038 [3]中铁云网信息科技有限公司,北京100036

出　　处：《水电能源科学》2022年第1期163-167,共5页Water Resources and Power

基　　金：中国水利水电科学研究院行业基础性、支撑性、应急性科研类项目(AU0145B482019)。

摘　　要：针对水电机组信号呈现的非线性和非平稳特性,提出了基于改进麻雀搜索算法与支持向量回归相结合的预测模型(ItSSA_VR)。将麻雀搜索算法进行自适应改进,首先在种群初始化方面,引入Sine混沌映射来影响算法的整个过程,使种群在搜索空间更加均匀的分布;其次采用自适应学习因子ω来提高算法的搜索能力,并通过Levy飞行算法这种随机游走策略来避免陷入局部最优的难题;最后,采用t分布对个体进行变异,提高算法的全局搜索性和局部搜索性的同时也提高了搜索速度。采用所提的改进麻雀搜索算法,建立SVR模型,基于iSMA2000一体化状态监测与趋势分析系统,对某水电站机组的上导X摆度值监测数据实例进行分析验证。结果表明,该预测模型具有更好的预测精度,且能拟合数据的波动状况,并根据实际工程应用需求,将此模型与HCON可视化软件结合,开发了组态化预测模块,实现了趋势预警模块的功能,对预测水电机组未来运行状态的发展趋势有一定的指导意义。In this paper,a prediction model(ITSSA_VR)based on improved sparrow search algorithm and support vector regression is proposed for the nonlinear and non-stationary characteristics of the signals presented by hydropower units.It adaptively improves the sparrow search algorithm.Firstly,Sine chaotic mapping is introduced in the population initialization to affect the whole process of the algorithm,which makes the population more evenly distributed in the search space.Secondly,it uses adaptive learning factor ω to improve the searching ability of the algorithm.The Levy flight algorithm,a random walk strategy,is used to avoid falling into the local optimal problem.Finally,the t distribution is adopted to implement individual mutation,which improves the global exploratory ability and local searching ability of the algorithm.The SVR model is established by using the improved sparrow search algorithm.Based on the iSMA2000 integrated condition monitoring and trend analysis system,the monitoring data of the top guide bearing x swing value for a certain hydropower unit is verified.The results show that the prediction model has better prediction accuracy and can fit the fluctuation of the data,which has a certain guiding significance to predict the development trend of the future operation state of hydropower units.According to the actual engineering application requirements,the model is combined with HCON visualization software to develop the function of the configuration prediction module and the trend warning module.

关 键 词：趋势预测 麻雀搜索 支持向量回归 HCON 预警模块 

分 类 号：TV742[水利工程—水利水电工程]