基于聚类分析和决策树的“一库多级”水电站日调度方法
|
摘要
一库多级"流域梯级因下游电站对上游水库发电放水的高敏感性使得梯级短期发电计划安排面临很大难度。该文从实用性出发,利用知识发现技术,提出耦合聚类分析和决策树的梯级水电站群短期发电调度方法。采用线性回归确定上下游电站间的电量匹配关系,以此为基础,从海量实际数据中聚类得到各电站的典型出力曲线,并将影响发电调度的日计划电量、库水位、电网特性等因子,与典型出力过程进行分类训练,构建梯级发电调度决策库,以便于采用决策树方法快速确定上下游水电站的联合调度方案,最后引入约束修正策略进行出力微调,以保证结果可行性。通过红水河干流天生桥"一库两级"电站日发电调度验证,结果表明该方法能够根据发电边界条件快速得到梯级水电站发电调度计划,并能够与实际出力过程保持较好的一致性。
Short-term operation of cascaded hydropower plants with one reservoir is very difficult because the generation of downstream hydropower plant is highly sensitive to the release of upstream reservoir. This paper developed a novel practical method for short-term operation based on knowledge rule technology. This method integrates the cluster analyst and decision tree algorithm. First, energy production relationship among cascaded hydropower plants was determined by the linear regression method. Second, typical generation curves of all hydropower plants were identified using cluster analysis method based on massive practical data. These curves were classified into several generation decision processes by considering daily energy production demand, forebay water level and grid characteristic. Thus, the decision library of generation operation can be built. Besides, several strategies of solving complex operation constraints were presented to guarantee the feasibility and practicality of optimal results. This method was verified to determine the day-ahead generation schedule for Tianshengqiao cascaded hydropower plants in Hongshui River. Results show that the method can quickly obtain generation schedule according to the given conditions and constraint boundaries. Moreover, the obtained generation profile is consistent with history records.
Short-term operation of cascaded hydropower plants with one reservoir is very difficult because the generation of downstream hydropower plant is highly sensitive to the release of upstream reservoir. This paper developed a novel practical method for short-term operation based on knowledge rule technology. This method integrates the cluster analyst and decision tree algorithm. First, energy production relationship among cascaded hydropower plants was determined by the linear regression method. Second, typical generation curves of all hydropower plants were identified using cluster analysis method based on massive practical data. These curves were classified into several generation decision processes by considering daily energy production demand, forebay water level and grid characteristic. Thus, the decision library of generation operation can be built. Besides, several strategies of solving complex operation constraints were presented to guarantee the feasibility and practicality of optimal results. This method was verified to determine the day-ahead generation schedule for Tianshengqiao cascaded hydropower plants in Hongshui River. Results show that the method can quickly obtain generation schedule according to the given conditions and constraint boundaries. Moreover, the obtained generation profile is consistent with history records.
引文
[1]Teegavarapu R S V,Simonovic S P.Short-term operation model for coupled hydropower reservoirs[J].Journal of Water Resources Planning and Management,2000,126(2):98-106.
[2]Zhou Jianzhong,Zhang Yongchuan,Zhang Rui,et al.Integrated optimization of hydroelectric energy in the upper and middle Yangtze River[J].Renewable and Sustainable Energy Reviews,2015,45:481-512.
[3]Cheng Chuntian,Wang Jiayang,Wu Xinyu.Hydro unit commitment with a head-sensitive reservoir and multiple vibration zones using MILP[J].IEEE Transactions on Power Systems,2016,31(6):4842-4852.
[4]Cheng Chuntian,Shen Jianjian,Wu Xinyu.Short-term scheduling for large-scale cascaded hydropower systems with multivibration zones of high head[J].Journal of Water Resources Planning and Management,2012,138(3),257-267.
[5]Cheng Chuntian,Shen Jianjian,Wu Xinyu,et al.Operation challenges for fast-growing China’s hydropower systems and respondence to energy saving and emission reduction[J].Renewable and Sustainable Energy Reviews,2012,16(5):2386-2393.
[6]Catal?o J P S,Mariano S J P S,Mendes V M F,et al.Nonlinear optimization method for short‐term hydro scheduling considering head‐dependency[J].International Transactions on Electrical Energy Systems,2010,20(2):172-183.
[7]Catalao J P S,Mariano S J P S,Mendes V M F,et al.Scheduling of head-sensitive cascaded hydro systems:a nonlinear approach[J].IEEE Transactions on Power Systems,2009,24(1):337-346.
[8]李辉,申建建,廖胜利,等.基于模糊聚类分析的水电站日初水位估算[J].中国电机工程学报,2010,30(35):118-124.Li Hui,Shen Jianjian,Liao Shengli,et al.Estimating the initial water level of the next day for day-ahead hydropower generation dispatch based on fuzzy cluster analysis[J].Proceedings of the CSEE,2010,30(35):118-124(in Chinese).
[9]武新宇,程春田,申建建,等.大规模水电站群短期优化调度方法Ⅲ:多电网调峰问题[J].水利学报,2012,43(1):31-42.Wu Xinyu,Cheng Chuntian,Shen Jianjian,et al.Short-term optimal operation methods of large-scale hydropower plantsⅢ.solution method of peak load regulation for multiple power grid[J].Journal of Hydraulic Engineering,2012,43(1):31-42(in Chinese).
[10]Mariano S J P S,Catal?o J P S,Mendes V M F,et al.Optimising power generation efficiency for headsensitive cascaded reservoirs in a competitive electricity market[J].International Journal of Electrical Power&Energy Systems,2008,30(2):125-133.
[11]唐海华,陈森林,赵宇.三峡梯级电站短期优化调度模型及算法[J].水电能源科学,2008,26(3):133-136.Tang Haihua,Chen Senlin,Zhao Yu.Model and algorithm of short-term optimal scheduling of three gorges cascaded hydropower stations[J].Water Resources and Power,2008,26(3):133-136(in Chinese).
[12]马超.耦合航运要求的三峡-葛洲坝梯级水电站短期调度快速优化决策[J].系统工程理论与实践,2013,33(5):1345-1350.Ma Chao.Fast optimal decision of short-term dispatch of three Gorges and Gezhouba cascade hydropower stations with navigation demand considered[J].Systems Engineering--Theory&Practice,2013,33(5):1345-1350(in Chinese).
[13]尹正杰,王小林,胡铁松,等.基于数据挖掘的水库供水调度规则提取[J].系统工程理论与实践,2006(8):129-135.Yin Zhengjie,Wang Xiaolin,Hu Tiesong,et al.Water supply reservoir operating rules extraction based on data mining[J].Systems Engineering-Theory&Practice,2006(8):129-135(in Chinese).
[14]王嘉阳,程春田,廖胜利,等.复杂约束限制下的梯级水电站群实时优化调度方法及调整策略[J].中国电机工程学报,2015,35(17):4326-4334.Wang Jiayang,Cheng Chuntian,Liao Shengli,et al.Real-time scheduling method and strategy for hydro cascaded system with complex constraints[J].Proceedings of the CSEE,2015,35(17):4326-4334(in Chinese).
[15]Bessler F T,Savic D A,Walters G A.Water reservoir control with data mining[J].Journal of Water Resources Planning and Management,2003,129(1):26-34.
[16]张双虎,张忠波,徐卫红,等.基于决策树技术的多年调节水库年末消落水位研究[J].水力发电学报,2012,31(6):44-48,49.Zhang Shuanghu,Zhang Zhongbo,Xu Weihong,et al.Study on year-end level of carryover storage reservoir based on decision tree[J].Journal of Hydroelectric Engineering,2012,31(6):44-48,49(in Chinese).
[17]Han J W,Kamber M.Data mining:concepts and techniques[M].Beijing:Mechanical Industry Press,2001.
[18]Salzberg S L.Book review:c4.5:programs for machine learning by J.Ross Quinlan.Morgan Kaufmann Publishers,Inc.,1993[J].Machine Learning,1994,16(3):235-240.
[19]牛文静,申建建,程春田,等.耦合调峰和通航需求的梯级水电站多目标优化调度混合搜索方法[J].中国电机工程学报,2016,36(9):2331-2341.Niu Wenjing,Shen Jianjian,Cheng Chuntian,et al.Ahybrid search method for multi-objective optimization operations of cascaded hydropower plants with peak load regulation and navigation demands[J].Proceedings of the CSEE,2016,36(9):2331-2341(in Chinese).
[20]李树山,廖胜利,申建建,等.厂网协调模式下梯级AGC控制策略[J].中国电机工程学报,2014,34(7):1113-1123.Li Shushan,Liao Shengli,Shen Jianjian,et al.Automatic generation control strategies of cascaded hydropower plants oriented to the coordination of power plants and power grids[J].Proceedings of the CSEE,2014,34(7):1113-1123(in Chinese).
[21]牛文静,申建建,冯仲恺,等.基于数据挖掘的梯级水电站群指令调度优化方法[J].电力系统自动化,2017,41(15):66-73.Niu Wenjing,Shen Jianjian,Feng Zhongkai,et al.Data mining based optimization method for instruction dispatching of cascade hydropower station group[J].Automation of Electric Power Systems,2017,41(15):66-73(in Chinese).
[2]Zhou Jianzhong,Zhang Yongchuan,Zhang Rui,et al.Integrated optimization of hydroelectric energy in the upper and middle Yangtze River[J].Renewable and Sustainable Energy Reviews,2015,45:481-512.
[3]Cheng Chuntian,Wang Jiayang,Wu Xinyu.Hydro unit commitment with a head-sensitive reservoir and multiple vibration zones using MILP[J].IEEE Transactions on Power Systems,2016,31(6):4842-4852.
[4]Cheng Chuntian,Shen Jianjian,Wu Xinyu.Short-term scheduling for large-scale cascaded hydropower systems with multivibration zones of high head[J].Journal of Water Resources Planning and Management,2012,138(3),257-267.
[5]Cheng Chuntian,Shen Jianjian,Wu Xinyu,et al.Operation challenges for fast-growing China’s hydropower systems and respondence to energy saving and emission reduction[J].Renewable and Sustainable Energy Reviews,2012,16(5):2386-2393.
[6]Catal?o J P S,Mariano S J P S,Mendes V M F,et al.Nonlinear optimization method for short‐term hydro scheduling considering head‐dependency[J].International Transactions on Electrical Energy Systems,2010,20(2):172-183.
[7]Catalao J P S,Mariano S J P S,Mendes V M F,et al.Scheduling of head-sensitive cascaded hydro systems:a nonlinear approach[J].IEEE Transactions on Power Systems,2009,24(1):337-346.
[8]李辉,申建建,廖胜利,等.基于模糊聚类分析的水电站日初水位估算[J].中国电机工程学报,2010,30(35):118-124.Li Hui,Shen Jianjian,Liao Shengli,et al.Estimating the initial water level of the next day for day-ahead hydropower generation dispatch based on fuzzy cluster analysis[J].Proceedings of the CSEE,2010,30(35):118-124(in Chinese).
[9]武新宇,程春田,申建建,等.大规模水电站群短期优化调度方法Ⅲ:多电网调峰问题[J].水利学报,2012,43(1):31-42.Wu Xinyu,Cheng Chuntian,Shen Jianjian,et al.Short-term optimal operation methods of large-scale hydropower plantsⅢ.solution method of peak load regulation for multiple power grid[J].Journal of Hydraulic Engineering,2012,43(1):31-42(in Chinese).
[10]Mariano S J P S,Catal?o J P S,Mendes V M F,et al.Optimising power generation efficiency for headsensitive cascaded reservoirs in a competitive electricity market[J].International Journal of Electrical Power&Energy Systems,2008,30(2):125-133.
[11]唐海华,陈森林,赵宇.三峡梯级电站短期优化调度模型及算法[J].水电能源科学,2008,26(3):133-136.Tang Haihua,Chen Senlin,Zhao Yu.Model and algorithm of short-term optimal scheduling of three gorges cascaded hydropower stations[J].Water Resources and Power,2008,26(3):133-136(in Chinese).
[12]马超.耦合航运要求的三峡-葛洲坝梯级水电站短期调度快速优化决策[J].系统工程理论与实践,2013,33(5):1345-1350.Ma Chao.Fast optimal decision of short-term dispatch of three Gorges and Gezhouba cascade hydropower stations with navigation demand considered[J].Systems Engineering--Theory&Practice,2013,33(5):1345-1350(in Chinese).
[13]尹正杰,王小林,胡铁松,等.基于数据挖掘的水库供水调度规则提取[J].系统工程理论与实践,2006(8):129-135.Yin Zhengjie,Wang Xiaolin,Hu Tiesong,et al.Water supply reservoir operating rules extraction based on data mining[J].Systems Engineering-Theory&Practice,2006(8):129-135(in Chinese).
[14]王嘉阳,程春田,廖胜利,等.复杂约束限制下的梯级水电站群实时优化调度方法及调整策略[J].中国电机工程学报,2015,35(17):4326-4334.Wang Jiayang,Cheng Chuntian,Liao Shengli,et al.Real-time scheduling method and strategy for hydro cascaded system with complex constraints[J].Proceedings of the CSEE,2015,35(17):4326-4334(in Chinese).
[15]Bessler F T,Savic D A,Walters G A.Water reservoir control with data mining[J].Journal of Water Resources Planning and Management,2003,129(1):26-34.
[16]张双虎,张忠波,徐卫红,等.基于决策树技术的多年调节水库年末消落水位研究[J].水力发电学报,2012,31(6):44-48,49.Zhang Shuanghu,Zhang Zhongbo,Xu Weihong,et al.Study on year-end level of carryover storage reservoir based on decision tree[J].Journal of Hydroelectric Engineering,2012,31(6):44-48,49(in Chinese).
[17]Han J W,Kamber M.Data mining:concepts and techniques[M].Beijing:Mechanical Industry Press,2001.
[18]Salzberg S L.Book review:c4.5:programs for machine learning by J.Ross Quinlan.Morgan Kaufmann Publishers,Inc.,1993[J].Machine Learning,1994,16(3):235-240.
[19]牛文静,申建建,程春田,等.耦合调峰和通航需求的梯级水电站多目标优化调度混合搜索方法[J].中国电机工程学报,2016,36(9):2331-2341.Niu Wenjing,Shen Jianjian,Cheng Chuntian,et al.Ahybrid search method for multi-objective optimization operations of cascaded hydropower plants with peak load regulation and navigation demands[J].Proceedings of the CSEE,2016,36(9):2331-2341(in Chinese).
[20]李树山,廖胜利,申建建,等.厂网协调模式下梯级AGC控制策略[J].中国电机工程学报,2014,34(7):1113-1123.Li Shushan,Liao Shengli,Shen Jianjian,et al.Automatic generation control strategies of cascaded hydropower plants oriented to the coordination of power plants and power grids[J].Proceedings of the CSEE,2014,34(7):1113-1123(in Chinese).
[21]牛文静,申建建,冯仲恺,等.基于数据挖掘的梯级水电站群指令调度优化方法[J].电力系统自动化,2017,41(15):66-73.Niu Wenjing,Shen Jianjian,Feng Zhongkai,et al.Data mining based optimization method for instruction dispatching of cascade hydropower station group[J].Automation of Electric Power Systems,2017,41(15):66-73(in Chinese).