基于智能电厂大数据的关键参数目标值挖掘技术  

Target value mining technology for key parameters based on big data in smart power plant

作　　者：刘长良[1,2] 许涛 王梓齐 孙群丽[3] LIU Changliang;XU Tao;WANG Ziqi;SUN Qunli(State Key Laboratory ofNew Energy Power Systems,North China Electric Power University,Beijing 102206,China;Department of Automation,North China Electric Power University Baoding 071003,China;Science and Technology College,North China Electric Power University Baoding 071003,China)

机构地区：[1]华北电力大学新能源电力系统国家重点实验室,北京102206 [2]华北电力大学自动化系,河北保定071003 [3]华北电力大学科技学院,河北保定071003

出　　处：《热力发电》2019年第9期14-21,共8页Thermal Power Generation

基　　金：北京市自然科学基金资助项目(4182061);中央高校基本科研业务费专项基金资助(9161717007)~~

摘　　要：为响应能源结构调整、推进能源供给侧改革,火力发电行业迫切需要进行深刻的技术转型。关键参数目标值的计算是火电厂运行优化的重要环节,常规的计算和挖掘方法精度低、实时性差、难以投入闭环应用。本文探讨了基于智能电厂大数据的关键参数目标值挖掘技术的内涵和具体实现,涉及整体架构、专用数据挖掘技术等;并以国内某超超临界机组为例,介绍了针对超超临界机组的关键参数目标值挖掘系统,该系统利用在线迭代更新和离线挖掘修正相结合的策略,解决了数据体量大与时效性要求高的矛盾,有助于实现关键参数目标值的高效挖掘。In order to respond to the adjustment of energy structure and promote the reform of energy supply side, the thermal power generation industry needs a profound technological transformation urgently. The calculation of the target value of key parameters is an important part of the operation optimization of thermal power plants. The conventional calculation and mining methods have low precision, poor real-time performance and are difficult to put into closed-loop applications. To solve this problem, this paper discusses the connotation and concrete realization of key parameter target value mining technology based on big data of intelligent power plant, involving the whole architecture and special data mining technology. Moreover, taking an ultra-supercritical unit in China as an example, it introduces the key parameter target value mining system for ultra-supercritical units. By applying the combination of online iterative update and offline mining correction, this system solves the contradiction between large data volume and high timeliness requirements, and helps to achieve efficient mining of key parameter target values.

关 键 词：智能电厂 大数据 关键参数 目标值 数据挖掘 迭代更新 

分 类 号：TK32[动力工程及工程热物理—热能工程]