摘要
火电机组运行智慧化是“双碳”目标约束下机组灵活运行的关键技术,基于大数据云计算平台的机组数字化架构是其重要组成部分。但目前现役机组的云平台改造建设尚缺乏深入技术探索及报道。以2台350 MW机组为例,系统开展了火电大数据云平台改造建设的实践研究。首先,发展了利旧的私有云平台搭建技术,实现了整合厂内原有6台服务器资源的低成本部署,同时保障建设期内机组正常运行;其次,针对火电机组万亿级的年数据量存储及电厂应用场景下数据读取需求,发展了云平台数据库数据同步技术,改进了数据库的测点管理方式及数据存储方式,开发了数据接口服务,使云平台数据库实现10^(5)量级测点数据的秒级写入及3 s内全年数据的单点检索。在此基础上,在云平台上部署了依托2.45×10^(9)条数据的大数据寻优算法,实现了机组能耗的可视化运行指导;搭建了针对高噪声测量数据的预处理模型,完成了对风量等数据的趋势提取;开发了选择性催化还原(SCR)入口NO_(x)浓度的神经网络预测模型,并通过构建负荷变化因子对预测结果进行反馈调节,可捕捉不同幅度的喷氨量变化需求。上述模型框架均可在云平台实现灵活调整与拓展,所提技术可为火电企业及新型低碳/零碳燃烧能源设备的智慧云平台建设提供参考。
Smart coal-fired power generation plays a pivotal role in facilitating flexible unit operations aimed at achieving carbon neutrality.The digital unit architecture based on cloud platforms is an essential component.However,an in-depth exploration concerning the construction of cloud platforms for the existing units is scarce.The research and practice of the critical technologies involved in the construction of big-data cloud platform for two 350 MW units were performed.A cost-effective deployment technology for cloud platforms integrating six existing servers in the plant was developed,which ensured the normal operation of units throughout the construction phase.Considering the requirements of trillions of data storage annually and data retrieval in the application scenarios of power plant,cloud platform database synchronization technology,database management and storage methodologies,and data interface services were developed.These innovations allow the cloud platform database to achieve second-level write capabilities for 10^(5)-level measurement data and facilitate single-point retrieval within 3 s across an entire year's data.Based on these foundations,a large-scale optimization algorithm utilizing 2.45×10^(9) data points was implemented,offering visualized operational guidance for unit energy consumption.A preprocessing model for high-noise measurement data was established,effectively extracting trends from variables such as wind speed.Furthermore,a neural network prediction model for inlet NO_(x) concentration of SCR was developed by integrating a feedback adjustment mechanism that accounts for load variations to address varying demands for ammonia injections.The proposed modeling framework allows for flexible adjustment and expansion,which can be implemented on the cloud platform.The proposed technical solutions offer a reference for the construction of intelligent cloud platforms for thermal power enterprises and novel low-carbon/zero-carbon combustion systems.
作者
王志敏
黄骞
王可轩
陈树宽
李敏
王海
李水清
WANG Zhimin;HUANG Qian;WANG Kexuan;CHEN Shukuan;LI Min;WANG Hai;LI Shuiqing(Key Laboratory of Thermal Science and Power Engineering of Ministry of Education,Department of Energy and Power Engineering,Tsinghua University,Beijing 100084,China;Yan'an Thermal Power Plant,Datang Shaanxi Power Generation Co.,Ltd.,Yan'an 716004,Shaanxi Province,China)
出处
《动力工程学报》
北大核心
2025年第2期282-291,共10页
Journal of Chinese Society of Power Engineering
基金
国家重点研发计划资助项目(2022YFB4003900)。
关键词
燃煤火电机组
智慧电厂
云平台
大数据
煤耗
coal-fired power generation unit
smart power plant
cloud platform
big data
coal consumption
作者简介
王志敏(1998-),女,内蒙古乌海人,硕士研究生,研究方向为低碳智能发电,E-mail:wangzm21@mail.tsinghua.edu.cn。
