面向大型发电设备的数据高速存储与智能化分析技术研究  

Research on high speed data storage and intelligent analysis technology for large power generation equipment

作　　者：冯普锋 单涛 陈静 孙晋志 FENG Pufeng;SHAN Tao;CHEN Jing;SUN Jinzhi(Shiliquan Power Plant,Huadian International Power Co.,Ltd.,Zaozhuang 277103,China)

机构地区：[1]华电国际电力股份有限公司十里泉发电厂,山东枣庄277103

出　　处：《电子设计工程》2023年第3期127-131,共5页Electronic Design Engineering

基　　金：中国华电科技项目(CHDKJ20-02-130)。

摘　　要：针对双塔双循环脱硫系统耗电量高的问题,提出了基于数据高速存储与智能化分析的脱硫系统优化运行方法。该方法构建了面向大型发电设备的高速存储数据架构,包括数据采集、数据处理、数据存储、数据管理、数据诊断和数据展示共6个模块。其利用机组实时负荷、目标负荷与负荷变化率等数据,采用长短期记忆(LSTM)算法预测系统实时工况下的入口SO_(2)浓度和SO_(2)脱除量。同时,基于对机组负荷、浆泵组合、浆液pH值、SO_(2)入口浓度等历史数据的分析,通过层次凝聚聚类(HAC)算法构建样本数据库,并将系统实时工况数据与样本数据库进行精准匹配,以实现工况寻优。以某发电厂的脱硫系统进行数据仿真验证的结果表明,所提运行优化方法能够精准地调整浆泵组合运行方式,降低系统运行成本,并提高系统脱硫效率。Aiming at the problem of high power consumption of double tower and double cycle desulfurization system,an optimal operation method of desulfurization system based on high-speed data storage and intelligent analysis is proposed. A high-speed data storage architecture for large power generation equipment is constructed,including six modules: data acquisition,data processing,data storage,data management,data diagnosis and data display. Using the data of unit real-time load,target load and load change rate,the inlet SO_(2) concentration and SO_(2) removal amount of the system under real-time working conditions are predicted by using Long Short-Term Memory(LSTM) algorithm. Based on the analysis of historical data such as unit load,slurry pump combination,slurry pH value and SO_(2) inlet concentration,the sample database is constructed through Hierarchical Agglomeration Clustering(HAC) algorithm,and the real-time working condition data of the system is accurately matched with the sample database to realize working condition optimization. The results show that the proposed operation optimization method can accurately adjust the combined operation mode of slurry pump,reduce the operation cost and improve the desulfurization efficiency of the system.

关 键 词：脱硫系统 运行优化 聚类算法 数据分析 LSTM 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置] TN99[自动化与计算机技术—控制科学与工程]