变电站选址定容方法的研究与实现
摘要
变电站选址定容是城市电网规划的一个关键任务,天津大学开发的城市电网规划辅助决策软件CNP的变电站选址定容规划子系统能够在小区负荷预测的基础上,无需事先指定水平年候选变电站的位置和数量,通过大范围的自动寻优搜索,求得新建变电站的站址、站容和供电范围。本文针对该软件在工程应用中迫切需要解决的问题进行了功能完善开发和选址方法改进,主要工作包括:
(1)通过对交替分配定位的变电站选址定容方法进行研究,发现该方法的模型和算法都是建立在负荷值在负荷小区内整体大小差异不能过大这样一个假设条件下,否则将带来较大的误差。而在实际的工程项目中负荷分布预测结果往往无法满足这一条件,特别是在当今社会经济发展的条件下,城市迅猛发展,各种负荷在急剧膨胀,单点负荷值和小区负荷相差的程度不断加大,导致变电站选址结果偏离最优解,甚至出现新建变电站距离过近或重叠的情况。针对这种情况,本文提出了一种改进的考虑单点负荷值过大条件下的变电站选址定容方法。该方法借助负荷小区的地理信息在选址前对过大负荷点进行负荷分解预处理,从而达到满足原有选址模型算法假设条件的要求,最后结合交替分配定位的选址模型算法得出最优解,同时该方法也解决了单点负荷过大时供电范围计算结果不合理的问题。基于该方法编制的程序通过算例的验证,达到了预期目标。
(2)原有的交替定位法变电站选址定容规划方法只适用于目标年选址,本文提出了一种实用的可以同时考虑中间年的负荷分布情况与目标年变电站站址容量的中间年变电站规划方法。该方法在分支定界优化算法的基础上,引入“电源点—负荷点”间弧值的概念,以弧值作为判断依据,对变电站的容量及所带负荷的大小和数量进行优化得出负荷分配的结果。首先,列出所有负荷点与待选电源点之间两两相连的弧,并将其排序。然后,从弧值最小的选起,每选择一个弧就删除与相应负荷点相连接的所有弧,并将负荷值从相应待选电源点容量中扣除,再从剩下的弧中继续选取,直到所有负荷点均被分配,最后得到电源点优化选择方案。算例验证了该方法的可行性。
(3)本文将上述两个方法在城市电网规划计算机辅助决策系统(CNP4.0)中编程实现,并对原有系统中的变电站选址定容规划子系统进行了有针对性地改进升级,新开发了包括选址前校验和规避不可建站选址的新功能。
Substation locating and sizing is a key task for power network planning. TianjinUniversity has developed the substation locating and sizing planning subsystem ofdecision support software CNP for urban power network planning. On the basis ofresidence region load forecasting, this subsystem could get the new substation’s site,size and power supply area by optimal searching on a large range without designatingthe site and quantity of candidate substation in target year in advance. This paper hasperfected the function and improved the locating methods of this software. The mainworks include:
(1)By the study on alternative allocation location method for substation locatingand sizing, it’s shown that the model and algorithm of this method is under theassumption condition that the load value difference of the whole load point would notbe great or there would bring greater error. However, in the real projects, loaddistribution forecasting result is often failed to meet this condition. Especially on thecondition of current society and economic development, the cities are developingrapidly and various loads are expanding sharply which lead to the more and moregreater difference between a single-point load value and region load value so that thelocating result is not the optimal and there even appear that two substations are tooclose or overlapped. Whereas this issue, this paper proposed an improved model andmethod for substation locating and sizing in view of great single-point load value. Bymeans of geographic information of this load region, this method would take the greatload point load decomposing before substation locating to meet the assumptioncondition’s requirement of existing locating model method’s and then be combinedwith alternative allocation location method to work out the optimal solution. At thesame time, this method also resolved the issue that the calculation results of powersupply region were unreasonable when the single-point load is great. At last, theprogram based on this method has been verified by examples and we reached theexpected result.
(2)The existing alternative allocation location method is only suitable for targetyear locating and this paper proposed a mid-year substation planning method whichhas taken both mid-year’s load distribution and target year’s substation site location capacityinto consideration. On the basis of Branch and Bound Algorithm, this methodbrought the concept of“power source point– load point”arc value. Taking the arcvalue as judgment basis, it would optimize the capacity of substation and the load sizeand quantity to reach the result of load distribution. At first, it listed all the arcsbetween candidate power and all the load points and orders those arcs and then itwould choose the arcs starting from the arc with the least arc value. After eachchoice,it would take out all the arcs combined with the relevant load points and takeaway the load value from the capacityof relevant candidate power point, and continueto choose another one from the left arcs until all the load points would be distributed.At last it achieved the optimization selection plan. This method has been shownfeasible byexample.
(3)This paper has realized the above two methods in programming the computerassistant decision support system of urban power network planning, upgraded andimproved the substation planning subsystem of the existing system with a clear aim,and developed the new functions including verification before locating and avoidingsites not suitable for substations.
(1)通过对交替分配定位的变电站选址定容方法进行研究,发现该方法的模型和算法都是建立在负荷值在负荷小区内整体大小差异不能过大这样一个假设条件下,否则将带来较大的误差。而在实际的工程项目中负荷分布预测结果往往无法满足这一条件,特别是在当今社会经济发展的条件下,城市迅猛发展,各种负荷在急剧膨胀,单点负荷值和小区负荷相差的程度不断加大,导致变电站选址结果偏离最优解,甚至出现新建变电站距离过近或重叠的情况。针对这种情况,本文提出了一种改进的考虑单点负荷值过大条件下的变电站选址定容方法。该方法借助负荷小区的地理信息在选址前对过大负荷点进行负荷分解预处理,从而达到满足原有选址模型算法假设条件的要求,最后结合交替分配定位的选址模型算法得出最优解,同时该方法也解决了单点负荷过大时供电范围计算结果不合理的问题。基于该方法编制的程序通过算例的验证,达到了预期目标。
(2)原有的交替定位法变电站选址定容规划方法只适用于目标年选址,本文提出了一种实用的可以同时考虑中间年的负荷分布情况与目标年变电站站址容量的中间年变电站规划方法。该方法在分支定界优化算法的基础上,引入“电源点—负荷点”间弧值的概念,以弧值作为判断依据,对变电站的容量及所带负荷的大小和数量进行优化得出负荷分配的结果。首先,列出所有负荷点与待选电源点之间两两相连的弧,并将其排序。然后,从弧值最小的选起,每选择一个弧就删除与相应负荷点相连接的所有弧,并将负荷值从相应待选电源点容量中扣除,再从剩下的弧中继续选取,直到所有负荷点均被分配,最后得到电源点优化选择方案。算例验证了该方法的可行性。
(3)本文将上述两个方法在城市电网规划计算机辅助决策系统(CNP4.0)中编程实现,并对原有系统中的变电站选址定容规划子系统进行了有针对性地改进升级,新开发了包括选址前校验和规避不可建站选址的新功能。
Substation locating and sizing is a key task for power network planning. TianjinUniversity has developed the substation locating and sizing planning subsystem ofdecision support software CNP for urban power network planning. On the basis ofresidence region load forecasting, this subsystem could get the new substation’s site,size and power supply area by optimal searching on a large range without designatingthe site and quantity of candidate substation in target year in advance. This paper hasperfected the function and improved the locating methods of this software. The mainworks include:
(1)By the study on alternative allocation location method for substation locatingand sizing, it’s shown that the model and algorithm of this method is under theassumption condition that the load value difference of the whole load point would notbe great or there would bring greater error. However, in the real projects, loaddistribution forecasting result is often failed to meet this condition. Especially on thecondition of current society and economic development, the cities are developingrapidly and various loads are expanding sharply which lead to the more and moregreater difference between a single-point load value and region load value so that thelocating result is not the optimal and there even appear that two substations are tooclose or overlapped. Whereas this issue, this paper proposed an improved model andmethod for substation locating and sizing in view of great single-point load value. Bymeans of geographic information of this load region, this method would take the greatload point load decomposing before substation locating to meet the assumptioncondition’s requirement of existing locating model method’s and then be combinedwith alternative allocation location method to work out the optimal solution. At thesame time, this method also resolved the issue that the calculation results of powersupply region were unreasonable when the single-point load is great. At last, theprogram based on this method has been verified by examples and we reached theexpected result.
(2)The existing alternative allocation location method is only suitable for targetyear locating and this paper proposed a mid-year substation planning method whichhas taken both mid-year’s load distribution and target year’s substation site location capacityinto consideration. On the basis of Branch and Bound Algorithm, this methodbrought the concept of“power source point– load point”arc value. Taking the arcvalue as judgment basis, it would optimize the capacity of substation and the load sizeand quantity to reach the result of load distribution. At first, it listed all the arcsbetween candidate power and all the load points and orders those arcs and then itwould choose the arcs starting from the arc with the least arc value. After eachchoice,it would take out all the arcs combined with the relevant load points and takeaway the load value from the capacityof relevant candidate power point, and continueto choose another one from the left arcs until all the load points would be distributed.At last it achieved the optimization selection plan. This method has been shownfeasible byexample.
(3)This paper has realized the above two methods in programming the computerassistant decision support system of urban power network planning, upgraded andimproved the substation planning subsystem of the existing system with a clear aim,and developed the new functions including verification before locating and avoidingsites not suitable for substations.
引文
[1]余贻鑫,王成山,肖峻,等.城网规划计算机辅助决策系统[J],电力系统自动化,2000,24(15):59~62
[2]萧国泉,徐绳均,电力规划,北京:水利水电出版社,1993:50~52
[3] E.Masud, An Interactive Procedure for Sizing and Timing DistributionSubstation Using Optimal Techniques, IEEE Transactions on PowerApparatus and Systems, Vol.PAS-93, No.5, 1974, pp.1281~1286
[4] Dale M.Crawford, Stewart B.Holt, A Mathematical Optimzation Techniquefor Locating and Sizing Distribution Substations, and Deriving Their OptimalService Areas, IEEE Transactions on Power Apparatus and Systems,Vol.PAS-94, No.2, March/April, 1975, pp.230~235
[5] Wall D L, Tompson G L, Northcote-Green J E D. An optimization model forplanning radial distribution networks [J]. IEEE Trails. On Power Apparatusand Systems, 1979, 98(3):1061~1068
[6] G.L.Thompson, G.L.Wall, A Branch and Bound Model for ChoosingOptimal Substation Locations, IEEE Transactions on Power Apparatus andSystems,Vol.PAS-100, No.5,May,1981,pp.2683~2688
[7]施光燕,董加礼,最优化方法,北京:高等教育出版社,1999.9~10
[8] MASUD E. An Interactive Procedure for Sizing and Timing DistributionSubstations Using Optimal Techniques. IEEE Trans on Power Apparatus andSystems. 1974. 93(5):1281~1286.
[9] CRAWFORD D M, HOLT S B. AMathematical Optimization Technique forLocating and Sizing Distribution Substations, and Deriving Their OptimalService Areas. IEEE Trans on Power Apparatus and Systems, 1975,94(2):230~235.
[10]IOK L, NASHID I. Interactive Expert System for Optimal Design ofElectricity Distribution Systems. IEE Proceedings Generation, Transmissionand Distribution, 1996,143(2):151~156
[11]张崇见,余贻鑫,严雪飞.中压配电变压器优化规划.电力系统自动化,1999,23(1):27~30
[12]麦克维克斯,如何求解问题:现代启发式方法,曹宏庆,李艳,董红斌等译,北京:中国水利水电出版社,2003:1~6
[13]赵建中,多目标线性规划.运筹与管理,1993(3):14~23
[14]刘晓惠,景清泉,霍俊爽,基于0-1型整数规划的配送中心选址,中国市场,2009(2):146~148
[15]梅亚东,冯尚友,水电站水库系统死库容优选的非线性网络流模型,水电能源科学,1989(2):168~175
[16]王家耀,空间信息系统原理,北京:科学出版社,2001:289~292
[17]黄小铼,陈星莺,基于负荷转移的中压配电所综合规划,电力自动化设备,2005,25(4):62~64
[18]L.Cooper, Location-allocation problem, Ops Res, 1963,11:331~343.
[19]D.Hongwei, Y.Yixin, H.Chunhua, et al. Optimal Planning of DistributionSubstation Locations and Sizes-model and Algorithm. Electrical Power &EnergySystems,Vol.18, No.6,pp353~357
[20]余贻鑫,张永伍,严雪飞,高压配电变电站的优化规划,天津大学学报,2006,8(39):889~894
[21]汤红卫,王华,郭喜庆,一种基于地理信息系统的配电网规划方法,电网技术,2002,26(12):79~82
[22]Lo.K.L, Nashid.I, Interactive expert system for optimal design ofelectricitdistribution systems. Generation, Transmission and Distribution,IEEEProceedings, Volume 143,Issue 2, March 1996 Page(s):151~156
[23]余贻鑫,张崇见,城市电网中压配电变压器的模糊优化规划,中国电机工程学报,2000,20(4):23~28
[24]王成山,魏海洋,肖峻,变电站选址定容两阶段优化规划方法[J],电力系统自动化,2005,29 (4): 62~66
[25]李伟伟,罗滇生,姚建刚,等,基于受约束区优先处理的城市配网规划变电站选址方法,广东电力,2007,20(2):15~19
[26]葛少云,李慧,刘洪,基于加权Voronoi图的变电站优化规划,电力系统自动化,31(2):29~34
[27]岑文辉,赵庆,戴文祥,遗传算法及在电网规划中的应用,电力系统及其自动化学报,1995,7(2):1~9
[28]冯凯,李远德,侯云鹤,等.改进单亲遗传算法在电源规划中的应用[J].电网技术,2004,28(3):11~15
[29]赵登福,杨靖,等,基于改进遗传算法的配电网无功优化,西安交通大学学报,2001,32(12):1219~1222
[30]肖燕,奚杰,阳薇,遗传算法在电网规划中的研究与应用,华北电力技术,2004,6:48~50
[31]梁吉,东北华北互联电网频率及联络线交换功率控制方案及考核原则研究[硕士学位论文),北京:中国电力科学研究院,2000
[32]陈根军,李继洗,基于Tabu搜索的配电网络规划,电力系统自动化,2001,25(7):40~44
[33]高炜欣,罗先觉,基于Hopfield神经网络的多阶段配电变电站的规划优化,电工技术学报,2005,20(5):58~46
[34]李金超,李庚银,等,基于改进BP神经网络的配电变电站选址研究,华东电力,2007,35(3):10~12
[35]刘自发,张建华.基于改进多组织粒子群体优化算法的配电网络变电站选址定容[J].中国电机工程学报,2007,27(1):105-111
[36]俞欢军,许宁,张丽平.混合粒子群优化算法研究[J].信息与控制,2005,34(4):500-505
[37]路志英,葛少云,王成山.基于粒子群优化的加权伏罗诺伊图变电站规划[J].中国电机工程学报,2009,16(29):35~41
[38]于晗,高海波,张建华,等,基于改进粒子群算法的变电站选址定容方法,中国国际供电会议(CICED2006)
[39]徐珍霞,顾洁,离散粒子群优化算法在变电站选址中的应用,电气应用,2006,25(4):35~38
[40]吴育华,杜纲,管理科学基础,天津大学出版社,2004,修订版:40~48
[41]郭秀英,论运输问题表上作业法,科技与管理,2007,3:33~35
[42]龚强,论表上作业法与单纯形法的相关同一性,东北检测,1998,1:3
[43]杨甫云,格林公式的一个应用.高等数学研究,2004, 7(2): 51-52.
[44]张文忠,多边形重心计算释疑,数学通报,1996,1
[45]刑文训,谢金星.现代优化计算方法[M].北京:清华大学出版社,1999.248—282
[46]王笑一,计及地理信息的变电站选址模型研究:[硕士学位论文],天津;天津大学,2009
[2]萧国泉,徐绳均,电力规划,北京:水利水电出版社,1993:50~52
[3] E.Masud, An Interactive Procedure for Sizing and Timing DistributionSubstation Using Optimal Techniques, IEEE Transactions on PowerApparatus and Systems, Vol.PAS-93, No.5, 1974, pp.1281~1286
[4] Dale M.Crawford, Stewart B.Holt, A Mathematical Optimzation Techniquefor Locating and Sizing Distribution Substations, and Deriving Their OptimalService Areas, IEEE Transactions on Power Apparatus and Systems,Vol.PAS-94, No.2, March/April, 1975, pp.230~235
[5] Wall D L, Tompson G L, Northcote-Green J E D. An optimization model forplanning radial distribution networks [J]. IEEE Trails. On Power Apparatusand Systems, 1979, 98(3):1061~1068
[6] G.L.Thompson, G.L.Wall, A Branch and Bound Model for ChoosingOptimal Substation Locations, IEEE Transactions on Power Apparatus andSystems,Vol.PAS-100, No.5,May,1981,pp.2683~2688
[7]施光燕,董加礼,最优化方法,北京:高等教育出版社,1999.9~10
[8] MASUD E. An Interactive Procedure for Sizing and Timing DistributionSubstations Using Optimal Techniques. IEEE Trans on Power Apparatus andSystems. 1974. 93(5):1281~1286.
[9] CRAWFORD D M, HOLT S B. AMathematical Optimization Technique forLocating and Sizing Distribution Substations, and Deriving Their OptimalService Areas. IEEE Trans on Power Apparatus and Systems, 1975,94(2):230~235.
[10]IOK L, NASHID I. Interactive Expert System for Optimal Design ofElectricity Distribution Systems. IEE Proceedings Generation, Transmissionand Distribution, 1996,143(2):151~156
[11]张崇见,余贻鑫,严雪飞.中压配电变压器优化规划.电力系统自动化,1999,23(1):27~30
[12]麦克维克斯,如何求解问题:现代启发式方法,曹宏庆,李艳,董红斌等译,北京:中国水利水电出版社,2003:1~6
[13]赵建中,多目标线性规划.运筹与管理,1993(3):14~23
[14]刘晓惠,景清泉,霍俊爽,基于0-1型整数规划的配送中心选址,中国市场,2009(2):146~148
[15]梅亚东,冯尚友,水电站水库系统死库容优选的非线性网络流模型,水电能源科学,1989(2):168~175
[16]王家耀,空间信息系统原理,北京:科学出版社,2001:289~292
[17]黄小铼,陈星莺,基于负荷转移的中压配电所综合规划,电力自动化设备,2005,25(4):62~64
[18]L.Cooper, Location-allocation problem, Ops Res, 1963,11:331~343.
[19]D.Hongwei, Y.Yixin, H.Chunhua, et al. Optimal Planning of DistributionSubstation Locations and Sizes-model and Algorithm. Electrical Power &EnergySystems,Vol.18, No.6,pp353~357
[20]余贻鑫,张永伍,严雪飞,高压配电变电站的优化规划,天津大学学报,2006,8(39):889~894
[21]汤红卫,王华,郭喜庆,一种基于地理信息系统的配电网规划方法,电网技术,2002,26(12):79~82
[22]Lo.K.L, Nashid.I, Interactive expert system for optimal design ofelectricitdistribution systems. Generation, Transmission and Distribution,IEEEProceedings, Volume 143,Issue 2, March 1996 Page(s):151~156
[23]余贻鑫,张崇见,城市电网中压配电变压器的模糊优化规划,中国电机工程学报,2000,20(4):23~28
[24]王成山,魏海洋,肖峻,变电站选址定容两阶段优化规划方法[J],电力系统自动化,2005,29 (4): 62~66
[25]李伟伟,罗滇生,姚建刚,等,基于受约束区优先处理的城市配网规划变电站选址方法,广东电力,2007,20(2):15~19
[26]葛少云,李慧,刘洪,基于加权Voronoi图的变电站优化规划,电力系统自动化,31(2):29~34
[27]岑文辉,赵庆,戴文祥,遗传算法及在电网规划中的应用,电力系统及其自动化学报,1995,7(2):1~9
[28]冯凯,李远德,侯云鹤,等.改进单亲遗传算法在电源规划中的应用[J].电网技术,2004,28(3):11~15
[29]赵登福,杨靖,等,基于改进遗传算法的配电网无功优化,西安交通大学学报,2001,32(12):1219~1222
[30]肖燕,奚杰,阳薇,遗传算法在电网规划中的研究与应用,华北电力技术,2004,6:48~50
[31]梁吉,东北华北互联电网频率及联络线交换功率控制方案及考核原则研究[硕士学位论文),北京:中国电力科学研究院,2000
[32]陈根军,李继洗,基于Tabu搜索的配电网络规划,电力系统自动化,2001,25(7):40~44
[33]高炜欣,罗先觉,基于Hopfield神经网络的多阶段配电变电站的规划优化,电工技术学报,2005,20(5):58~46
[34]李金超,李庚银,等,基于改进BP神经网络的配电变电站选址研究,华东电力,2007,35(3):10~12
[35]刘自发,张建华.基于改进多组织粒子群体优化算法的配电网络变电站选址定容[J].中国电机工程学报,2007,27(1):105-111
[36]俞欢军,许宁,张丽平.混合粒子群优化算法研究[J].信息与控制,2005,34(4):500-505
[37]路志英,葛少云,王成山.基于粒子群优化的加权伏罗诺伊图变电站规划[J].中国电机工程学报,2009,16(29):35~41
[38]于晗,高海波,张建华,等,基于改进粒子群算法的变电站选址定容方法,中国国际供电会议(CICED2006)
[39]徐珍霞,顾洁,离散粒子群优化算法在变电站选址中的应用,电气应用,2006,25(4):35~38
[40]吴育华,杜纲,管理科学基础,天津大学出版社,2004,修订版:40~48
[41]郭秀英,论运输问题表上作业法,科技与管理,2007,3:33~35
[42]龚强,论表上作业法与单纯形法的相关同一性,东北检测,1998,1:3
[43]杨甫云,格林公式的一个应用.高等数学研究,2004, 7(2): 51-52.
[44]张文忠,多边形重心计算释疑,数学通报,1996,1
[45]刑文训,谢金星.现代优化计算方法[M].北京:清华大学出版社,1999.248—282
[46]王笑一,计及地理信息的变电站选址模型研究:[硕士学位论文],天津;天津大学,2009