CIESC Journal ›› 2024, Vol. 75 ›› Issue (3): 924-935.DOI: 10.11949/0438-1157.20231345

• Process system engineering • Previous Articles     Next Articles

LiDAR measurement based on model predictive control for boiler combustion optimization

Yibin DONG1(), Jingchao XIONG2, Jingyu WANG1, Shoukang WANG1, Yafei WANG1, Qunxing HUANG1()   

  1. 1.State Key Laboratory of Clean Energy Utilization, Zhejiang University, Hangzhou 310027, Zhejiang, China
    2.China City Environment Protection Engineering Limited Company, Wuhan 430205, Hubei, China
  • Received:2023-12-18 Revised:2024-02-23 Online:2024-05-11 Published:2024-03-25
  • Contact: Qunxing HUANG

融合激光雷达料位测算的锅炉燃烧优化模型预测控制

董益斌1(), 熊敬超2, 王敬宇1, 汪守康1, 王亚飞1, 黄群星1()   

  1. 1.浙江大学能源清洁利用国家重点实验室,浙江 杭州 310027
    2.中冶南方都市环保工程技术股份有限公司,湖北 武汉 430205
  • 通讯作者: 黄群星
  • 作者简介:董益斌(1999—),男,硕士研究生, 22160492@zju.edu.cn
  • 基金资助:
    浙江大学科研项目(XY2022015)

Abstract:

A data-driven predictive control method with laser radar material level measurement was proposed for the combustion process of a power station boiler. This control strategy uses laser radar to monitor the amount of biomass fed into the furnace online in real time, and uses the maximum mutual information coefficient (MIC) method to analyze its effect on key parameters such as main steam pressure, combustion chamber temperature, and outlet flue gas oxygen content. Combined with the characteristic parameters of distributed control system (DCS) after screening, the model data set was constructed. Based on the model data set, an auto-regressive with extra inputs model optimized by particle swarm optimization algorithm was established. Based on the boiler key parameter model, the proposed control method attempts to minimize the flue gas oxygen content deviation under the constraints of main steam pressure and combustion chamber temperature. Taking 700 t/d biomass combustion power generation boiler as the test object, the experimental results based on the actual production data, as well as the comparative analysis demonstrate: (1) The predictive model can accurately predict the boiler key parameters and meet the demands of boiler combustion process control and optimization; (2) Compared with PID control and fuzzy control, the model predictive control algorithm shows higher control performance, and the average deviation of oxygen content from its set value in the simulation results can be controlled within ±25%. The proposed model predictive control method has good practical significance in theory, and can provide reference for the optimization and transformation of biomass boiler combustion in the future.

Key words: biomass boiler, laser radar, dynamic modeling, model-predictive control, algorithm

摘要:

提出了一种融合激光雷达料位测算的锅炉燃烧优化控制策略。该控制策略通过激光雷达对生物质的入炉给料量进行实时在线监测,并采用最大互信息系数(MIC)方法分析其对于主汽压力、燃烧室温度、出口烟气氧量等关键参数的时延长度,结合经过特征筛选的锅炉特征参数,构成模型样本集;通过粒子群优化(PSO)算法进行参数优化,获得燃烧过程外源自回归(ARX)模型,并采用模型预测控制的方法优化锅炉燃烧。以700 t/d生物质燃烧发电锅炉为测试对象,该控制策略在主汽压力与燃烧室温度约束下实现了烟气氧量优化。实验结果与对比分析表明:(1)该预测模型具有较高的精度与泛用性,能满足预测控制的要求;(2)相较于PID与模糊控制,该控制算法拥有更高的控制性能,在仿真结果中氧量与其设定值的平均偏移能控制在±25%以内。该控制策略理论上具有较好的实践意义,可为生物质发电锅炉的智能化稳定燃烧提供解决方案。

关键词: 生物质锅炉, 激光雷达, 动态建模, 模型预测控制, 算法

CLC Number: