数据驱动的火电厂输灰系统状态感知与安全增效优化运行研究OA
Research on data-driven state perception and safety-enhanced efficiency-improving operation of ash discharge system in thermal power plants
以某电厂660 MW机组输灰系统为对象,提出数据驱动的状态感知与优化运行框架.基于主成分分析降维与最小二乘支持向量机构建飞灰生成量时序预测模型;以管网压力波动最小与总耗气量最小为目标,建立输灰管路智能调度模型;设计基于模型预测控制的空压机集群闭环协同策略,实现从灰量感知到气源控制的端到端智能决策闭环.结果表明,该预测模型在平均绝对误差、均方根误差等关键指标上均优于对比模型.测试期间,平均输灰周期由45 min动态调整为25~70 min,日均总输灰次数和单位时间输送气耗减少,压缩空气母管压力标准差由0.032 MPa降至0.013 MPa,灰斗高料位月均报警次数降至0次.该数据驱动优化框架实现了输灰系统对负荷与煤质变化的自适应,为燃煤电厂辅机系统的节能、安全与智能化改造提供了可推广的技术路径.
Taking the ash conveying system of a 660 MW unit in a certain power plant as the research object,a data-driven state perception and optimized operation framework is proposed.A time series prediction model for fly ash generation volume is constructed based on principal component analysis for dimension reduction and least squares support vector machine.An intelligent scheduling model for ash conveying pipelines is established with the goal of minimizing pipeline pressure fluctuations and total gas consumption.A closed-loop collaborative strategy for the air compressor cluster based on model predictive control is designed to achieve an end-to-end intelligent decision-making loop from ash quantity perception to gas source control.The results show that the prediction model outperforms the comparison model in key indicators such as average absolute error and root mean square error.During the test period,the average ash conveying cycle is dynamically adjusted from 45 min to 25~70 min,the average daily total ash conveying frequency and gas consumption per unit time decreased,the standard deviation of the main gas pipe pressure decreased from 0.032 MPa to 0.013 MPa,and the monthly average alarm times for high material level in the ash hopper dropped to 0.The data-driven optimization framework enables the ash conveying system to adapt to changes in load and coal quality,providing a scalable technical path for the energy-saving,safety,and intelligent transformation of auxiliary equipment systems in coal-fired power plants.
刘洪伟;刘永华;滕飞;田大伟
淮浙电力有限责任公司凤台发电分公司,安徽 淮南 232000淮浙电力有限责任公司凤台发电分公司,安徽 淮南 232000淮浙电力有限责任公司凤台发电分公司,安徽 淮南 232000淮浙煤电有限责任公司凤台发电分公司,安徽 淮南 232100
信息技术与安全科学
气力输灰数据驱动状态感知智能调度模型预测控制节能优化
pneumatic ash conveyingdata-drivenstate perceptionintelligent schedulingmodel predictive controlenergy-saving optimization
《节能》 2026 (7)
103-106,4
评论