基于集成学习的热电联供机组标煤耗量实时估计算法研究与应用  

作　　者：姜志锋 潘生华 顾宝 傅骏伟 宋歌 郭庆 叶欣楠 王豆 郭鼎 张震伟 

机构地区：[1]浙江浙能数字科技有限公司,浙江杭州311121 [2]浙江浙能绍兴滨海热电有限责任公司,浙江绍兴312000 [3]浙江浙能技术研究院有限公司,浙江杭州311121 [4]浙江省能源集团,浙江杭州311121

出　　处：《工业控制计算机》2023年第9期148-150,共3页Industrial Control Computer

摘　　要：为了降低电厂的运行成本,目前大多数燃煤电厂普遍采用混煤掺烧技术进行发电生产。由于煤质成分无法实时测定,电厂运行人员缺少对机组标煤耗量的精准评估,从而导致运行人员无法有效地针对关键发电设备进行优化调度。为了解决上述问题,提出了一种基于集成学习的热电联供机组标煤耗量实时估计算法。该算法采用梯度提升树对热电联供机组进行标煤耗量回归分析,得到基于特定工况的标煤耗量回归曲线。经实验验证,该算法能够快速估计实时标煤耗量并协助优化调度。As the coal composition cannot be measured in real time,the operators of the power plant lack of accurate assessment of the coal consumption of the unit,which leads to the operators being unable to effectively optimize the scheduling of key power generation equipment.In order to solve the above problems,this paper proposes a real-time coal consumption estimation algorithm based on integrated learning for co-generation units.The algorithm uses gradient lifting tree to carry out regression analysis on coal consumption of co-generation units,and obtains the regression curve of coal consumption based on specific working conditions.The experimental results show that the algorithm can quickly estimate real-time coal consumption and assist in optimal scheduling.

关 键 词：热电联供 集成学习 标煤耗量估计 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程] TM621[自动化与计算机技术—控制科学与工程]