摘要
灵活调峰状态下,火电机组回热系统各加热器的水位存在调节偏差过大的问题。提出了基于压缩最近邻和遗传算法的前馈控制逻辑,并将其应用于某火电机组2号高压加热器水位的运行控制。利用历史运行数据,通过聚类数据约简得到负荷均匀分布的样本数据,采用最小二乘拟合方法得到一组模型参数,以此为初始值采用遗传优化算法进行局部优化,最终获得改进的高加水位前馈控制系统的多变量回归模型。现场试验结果表明:高加水位在负荷变化时,水位波动保持在±5 mm以内,在设定值阶跃扰动试验中保持在±2 mm以内,显著提升了高加水位的控制效果。该方法得出的前馈参数有效抑制了负荷变化带来的水位波动,使得高加水位几乎不受负荷变化影响。
In the state of flexible peak regulation,the water level of each heater of the thermal power unit’s regenerative system has the problem of excessive adjustment deviation.This paper proposes a feed-forward control logic based on compressed nearest neighbors and genetic algorithm,and applies it to the operation control of the water level of the 2# high pressure heater of a thermal power unit.This method uses historical operating data to obtain sample data with uniform load distribution through clustering data reduction,and uses least squares fitting method to obtain a set of model parameters,and then uses this as the initial value for partial optimization via genetic optimization algorithm,and finally obtains the improved multi-variable regression model of feedforward control system for the high heater water level.The field test results show that the water level fluctuation is kept within ±5 mm when the load changes in the high water level,and within ±2 mm in the set value step disturbance test,which significantly improves the control effect of the high water level.The feed-forward parameters obtained by this method effectively suppress the fluctuation of the water level caused by the load change,making the high water level hardly affected by the load change.
作者
黄青岭
曹越
庞海宇
司风琪
HUANG Qing-ling;CAO Yue;PANG Hai-yu;SI Feng-qi(Key Laboratory of Energy Thermal Conversion and Control of Ministiy of Education,Southeast University,Nanjing,China,Post Code:210096;Inner Mongolia Daihai Electric Power Generation Co.Ltd.,Beijing Energy Investment Holding Co.Ltd.,Wulanchabu,China,Post Code:013700)
出处
《热能动力工程》
CAS
CSCD
北大核心
2020年第12期121-127,共7页
Journal of Engineering for Thermal Energy and Power
关键词
火电机组回热系统
灵活调峰
数据约简
遗传算法
控制策略
Thermal power unit regenerative system
flexible peak regulation
data reduction
genetic algorithm
control strategy
