張懷德,張建生,李蓓
(1. 河海大學(xué) 能源與電氣學(xué)院,江蘇 南京 210098;2. 常州工學(xué)院,江蘇 常州 213002)
基于多目標(biāo)協(xié)調(diào)內(nèi)點(diǎn)法的分布式電源配置
張懷德1,張建生1,李蓓2
(1. 河海大學(xué) 能源與電氣學(xué)院,江蘇 南京 210098;2. 常州工學(xué)院,江蘇 常州 213002)
分布式電源并網(wǎng),對(duì)系統(tǒng)的網(wǎng)絡(luò)損耗、可靠性等會(huì)帶來影響,且其影響程度與分布式電源的位置和容量密切相關(guān)。在求解電力系統(tǒng)優(yōu)化問題上,為充分利用內(nèi)點(diǎn)法的收斂快、精度高的優(yōu)點(diǎn),把傳統(tǒng)的選址定容模型,采用內(nèi)點(diǎn)法進(jìn)行連續(xù)求解。并提出一種多目標(biāo)函數(shù)歸一轉(zhuǎn)化成單目標(biāo)函數(shù),采用協(xié)調(diào)參數(shù)w使兩個(gè)子函數(shù)達(dá)到優(yōu)化效果,進(jìn)行選址定容。建立以網(wǎng)絡(luò)有功損耗和節(jié)點(diǎn)電壓水平最小為目標(biāo)函數(shù),對(duì)改進(jìn)IEEE 30節(jié)點(diǎn)的系統(tǒng)進(jìn)行測試,結(jié)果表明,基于多目標(biāo)協(xié)調(diào)內(nèi)點(diǎn)法是有效和實(shí)用的。
內(nèi)點(diǎn)法;選址定容;多目標(biāo)函數(shù);有功損耗;電壓水平
當(dāng)今,分布式電源在配電網(wǎng)得到廣泛的運(yùn)用,快速的發(fā)展。DG有利于減少用戶的電能花費(fèi),緩解電網(wǎng)的擁堵,在負(fù)荷集中點(diǎn)安裝環(huán)保能源,可以提高電壓穩(wěn)定性,減小網(wǎng)絡(luò)損耗,緩解儲(chǔ)備容量[1]。DGs一般指發(fā)電量在1kW到50MW之間,安裝在負(fù)荷集中區(qū)的發(fā)電電源[2]。
近年來,國內(nèi)外的大量的學(xué)者在這方面做了大量的研究。Sudipta Ghosh等采用牛頓拉夫遜求解網(wǎng)損和花費(fèi)最小的DGs的位置,獲得最大的經(jīng)濟(jì)效益[2]。Luis F等采用多期交流優(yōu)化潮流求解能耗最小的確定DG位置[3]。Isrsfil Hussain和Anjan采用DE方法以網(wǎng)損最小進(jìn)行選址定容[4]。M. F. Alhajri等采用FSQP方法以網(wǎng)損最小進(jìn)行選容[5]。但是對(duì)于多目標(biāo)選址定容,依然不能綜合考慮優(yōu)化。
內(nèi)點(diǎn)法已被證明是解決非線性規(guī)劃的一種強(qiáng)有力工具[6],表現(xiàn)出極好的收斂性和較高的精度[7-10],在電力系統(tǒng)領(lǐng)域上得到了廣泛的應(yīng)用[11-13],但在解決有離散問題上存在不足。在解決多目標(biāo)問題上,基于Pareto最優(yōu)意義的協(xié)調(diào)各目標(biāo)函數(shù)之間的關(guān)系[14];采用模糊理論適合描述不確定性及處理不同量綱及互相矛盾的多目標(biāo)優(yōu)化問題[15],把多目標(biāo)函數(shù)轉(zhuǎn)換成單目標(biāo)函數(shù),通過模糊選擇控制[16]實(shí)行。這些處理方法和人工智能算法具有很好的結(jié)合性,對(duì)于內(nèi)點(diǎn)法不能很好的實(shí)現(xiàn)?;诖?,論文從另一個(gè)角度考慮分布式電源的選址定容,充分利用內(nèi)點(diǎn)法收斂性好,精度高等優(yōu)勢。
1.1 目標(biāo)函數(shù)
1) 以有功網(wǎng)損最小:
(1)
2) 節(jié)點(diǎn)電壓水平:
(2)
(3)
式中:i=1,2;μ1,μ2對(duì)應(yīng)于系統(tǒng)網(wǎng)絡(luò)損耗和節(jié)點(diǎn)電壓水平的子目標(biāo)函數(shù)。
minf=(1-w)×μ1+w×μ2
(4)
式中:w協(xié)調(diào)因子。
1.2 約束條件
1.2.1 功率方程
(5)
i=1,2…n
(6)
i=1,2…NPQ
1.2.2 不等式約束
線性不等式約束:
(7)
?。篞DG=0.2PDG
非線性約束條件:
(8)
nbr為支路數(shù);nDG為DG臺(tái)數(shù);ng為發(fā)電機(jī)臺(tái)數(shù)。
把DG的選址定容的數(shù)學(xué)模型寫成標(biāo)準(zhǔn)形式:
minf(x)
(9)
連續(xù)部分采用內(nèi)點(diǎn)法如下,即把式(9)引入松弛變量Zm轉(zhuǎn)化為:
(10)
根據(jù)Karush-Kuhn-Tucker最優(yōu)一階必要條件得到:
(11)
對(duì)最優(yōu)化條件式(11)采用牛頓法求解得到式(12)。
(12)
(13)
(14)
牛頓跌代更新計(jì)算可以根據(jù)以下3步:
1) 根據(jù)式(14)計(jì)算Δx和Δλ;
2) 根據(jù)式(13)計(jì)算ΔZ;
3) 根據(jù)式(12)計(jì)算Δμ。
αp、αd分別為原變量和對(duì)偶變量步長,表達(dá)如式(15)和式(16):
(15)
(16)
變量更新如式(17):
(17)
論文采用Matpower4.1中的IEEE30節(jié)點(diǎn)作為測試模型(改進(jìn)IEEE30 bus)。以節(jié)點(diǎn)1作為平衡節(jié)點(diǎn),其基準(zhǔn)電壓為100MV。
4.1 協(xié)調(diào)因子w優(yōu)化
對(duì)于case 30bus以無DG網(wǎng)損為目標(biāo)函數(shù)ΔVmax=30.556,安裝2臺(tái)DG以節(jié)點(diǎn)電壓ΔVmin=29.205。以無DG節(jié)點(diǎn)電壓為目標(biāo)函數(shù)Ploss,max=0.025,2臺(tái)DG損耗為目標(biāo)函數(shù)Ploss,min=0.015以取case 30 busΔV區(qū)間[29.2056,30.5563],case30 busPloss區(qū)間[0.0148,0.0249]。建立多目標(biāo)函數(shù),w和Ploss,ΔV的關(guān)系如圖1所示。
圖1 w和Ploss,ΔV關(guān)系
為達(dá)到子目標(biāo)函數(shù)優(yōu)化效果(主要考慮網(wǎng)損最小),選取協(xié)調(diào)因子w=0.4。
4.2DG容量和最優(yōu)位置
如圖2所示,最優(yōu)位置是8節(jié)點(diǎn),P=0.236,Q=0.0472,S=0.2429,在IEEE30節(jié)點(diǎn)模型中,雖然節(jié)點(diǎn)7,8不是負(fù)荷集中區(qū),卻是負(fù)荷的最嚴(yán)重區(qū),所以DG安裝8節(jié)點(diǎn)合理性。故安裝2臺(tái)DG,安裝位置為8,11節(jié)點(diǎn),S8=0.2429,S11=0.1569。
圖2 PQ節(jié)點(diǎn)號(hào)和目標(biāo)函數(shù)的關(guān)系
4.3 優(yōu)化的效果
如圖3所示,對(duì)于安裝1臺(tái)DG建立的多目標(biāo)函數(shù)比沒有安裝DG以網(wǎng)損為目標(biāo)函數(shù)的電壓穩(wěn)定性總體上,有大幅度提高;同樣,相比較沒有安裝DG以節(jié)點(diǎn)電壓為目標(biāo)函數(shù),更趨于平穩(wěn)更靠近1點(diǎn)附近??瞻字伪硎景惭b有DG以多目標(biāo)函數(shù);斜線柱形表示沒有安裝DG以節(jié)點(diǎn)電壓水平為目標(biāo)函數(shù),交叉線柱形表示沒有安裝DG以網(wǎng)損為目標(biāo)函數(shù)。
圖3 節(jié)點(diǎn)和電壓幅值的關(guān)系
采用上述方法安裝1臺(tái)DG時(shí),Ploss=0.0166比較沒有安裝DG以網(wǎng)絡(luò)有功損耗為目標(biāo)函數(shù)Ploss=0.0227,網(wǎng)損減少了36.75%,安裝1臺(tái)DG以網(wǎng)絡(luò)有功損耗為目標(biāo)函數(shù)Ploss=0.0157,ΔVloss=30.4716,網(wǎng)損增加了5.42%;ΔVloss比相同條件下以節(jié)點(diǎn)電壓水平為目標(biāo)函數(shù)求得的ΔVu=29.0709,增加了4.8%,故滿足優(yōu)化要求。(一次連續(xù)迭代次數(shù)約31次,時(shí)間23.462s。)
論文構(gòu)建配電網(wǎng)選址定容的多目標(biāo)模型,把模型分解成離散和連續(xù)兩個(gè)部分,實(shí)現(xiàn)離散連續(xù)交替求解。對(duì)于多目標(biāo)函數(shù)采用協(xié)調(diào)因子w轉(zhuǎn)換成單目標(biāo)函數(shù),并闡述了內(nèi)點(diǎn)法實(shí)現(xiàn)原理和步驟。利用內(nèi)點(diǎn)法收斂性快、精度高的優(yōu)勢,對(duì)連續(xù)模型的優(yōu)化。算例仿真表明,在優(yōu)化的位置安裝合理的DG,負(fù)荷節(jié)點(diǎn)電壓水平得到大幅度提高并網(wǎng)損耗減少。
[1] S.Kamalinia,et al.A combination of MADM and genetic algorithm for optimal DG allocation in power systems[C]//Univ of Tehran.UPEC,2007 42ndInternational,4-6 Sept 2007,Brighton:1031-1035.
[2] Sudipta Ghosh,et al.Optimal sizing and placement of distributed generation in a network system[J].Electrical Power and Energy Sytems. 2010,32(8):849-856.
[3] Lius F.Ochoa,Gareth P,Harrison.Minimizing Eneray Losses:Optimal Accommodation and Smart Operation of Renewable[J].2011,26(1):198-205.
[4] Israfil Hussain,Anjan Kumar Roy.Optimal Size and Location of Distributed using Differential Evolution[J].CISP,2012:57-61.
[5] M.F.AlHajri,et al.Improved seqential quadratic programming approach for optimal distribution generation sizing in distribution networks[C]//Electr Eng Technol Dept Coll of Technol Studies.CCECE,2010 23rdCandadian Conference,2-5 May 2010,Calgary AB:1-7.
[6]黎靜華,韋化.基于內(nèi)點(diǎn)法的機(jī)組組合模型[J].電網(wǎng)技術(shù),2007,31(24):28-33..
[7] James A.Monoh,et al.The quadratic interior point method solving power system optimization problems[J].IEEE Transactions on Power System,1994,9(3):1327-1336.
[8] M.J.Rider,et al.Enhanced higher-order interior-point method to minimise active power losses in electric energy systems[J].Generation,Transmission and Distributed,IEE Proceedings,2004,151(4):517-525.
[9] Liang Xie,Hsiao-Dong Chiang.A Enhanced Multiple Predictor-Corrector Interior Point Method for Optimal Power Flow[C]//Sch.of Electron&Electr.Eng,Shanghai Jiao Tong Univ.Power and Energy Society General Metting 2010 IEEE,25-29 July 2010,Shanghai,China:1-8.
[10] C.Y.Chung,wei Yan,et al.Decomposed predictor-Corrector Interior Point Method for Dynamic Optimal Power Flow[J].IEEE TRANSCATION ON POWER SYSTEM,2011,26(3):1030-1039.
[11] 熊寧,張魏,等.基于約束松弛變量策略的中心校正內(nèi)點(diǎn)法[J].電力系統(tǒng)保護(hù)與控制,2012,40(14):20-25.
[12] 趙斌,王明渝,等.基于非線性內(nèi)點(diǎn)法的雙饋風(fēng)電場功率優(yōu)化分配控制策略[J].電力系統(tǒng)保護(hù)與控制,2012,40(13):24-30..
[13] 王林川,于奉振,等.基于奇異值分解和內(nèi)點(diǎn)法的交直流電力系統(tǒng)電壓穩(wěn)定性裕度研究[J].電力系統(tǒng)保護(hù)與控制,2011,39(20):89-92.
[14] 孫國強(qiáng),等.多目標(biāo)配電網(wǎng)故障的Pareto進(jìn)化算法[J].電力自動(dòng)化設(shè)備,2012,32(5):57-62..
[15] Wen Zhang,Yutian Liu,Multi-objective reactive power and voltage control based on fuzzy optimization strategy and fuzzy adaptive particle swarm[J].Electrical Power and Energy System.2008,20:525-532.
[16] Maghouli,P,et al.A multiobjective framework for transmission expansion planning in deregulated environments[J],IEEE Trans.Power Syst,2009,24,(24):1051-1061.
Distributed Generation Allocation Based on Multi-Objective Cooperative Interior Point Method
ZHANG Huai-de1, ZHANG jian-sheng1, LI Bei2
(1.College of Energy and Electrical Engineering,Hohai University Nanjing 210098,China;2.Changzhou Institute of Technology Changzhou 213002,China;)
Integration of distributed generation(DG) to the power grid,it has a great impact on power loss and reliability of distribution system. The degree of impact is closely related with the placement and sizing of DG. The interior point method is widely used to solve the different types of optimization problems in electric power domain. The authors use the traditional placement and sizing of distributed generation model during the solution of continuous variables and make full use of the advantages of convergence performance and high accuracy in IPM. It is proposed that a multi-objective optimization model is converted to mono-objective optimization model based on coordinate coefficient w to obtain its optimal placement and sizing, in order to minimize network power loss and voltage deviation. This approach is effective and practical.
interior point method;placement and sizing;multi-objection function;active loss;voltage level
張懷德(1985-),男,山東濰坊人,碩士研究生,研究方向:從事分布式電源的規(guī)劃研究。
TM02
B
1671-5276(2014)02-0167-03
2013-01-21