1. 吉林大学 计算机科学与技术学院, 吉林 长春 130012;
2. 陆军炮兵防空兵学院, 辽宁 沈阳 100867
收稿日期:2016-10-12
基金项目:国家自然科学基金资助项目(61472161, 61402195, 61502198)。
作者简介:王生生(1974-), 男, 吉林长春人, 吉林大学教授, 博士生导师。
摘要:研究了用于解决微网优化调度问题的群智能算法.针对微网优化调度问题的多目标、多约束条件等特点, 对微网优化调度问题建模; 提出了改进的花朵授粉算法, 并将其应用到微网优化调度问题.在初始化时, 采用对立点方法增加种群多样性和优化搜索空间; 局部更新时, 使用一种新的局部更新算子提高算法收敛速度; 此外, 为了减少计算量和避免陷入局部最优, 定义了是否使用遗传操作的判断条件.仿真结果表明, 该算法性能优于原始花朵授粉算法和遗传算法等其他算法.
关键词:微网优化调度群智能花朵授粉算法多目标
Modified Flower Pollination Algorithm and Applications on Optimization Dispatch of Microgrid
WANG Sheng-sheng1, DU Peng1, DONG Ru-yi1, LI Yong-he2
1. College of Computer Science and Technology, Jilin University, Changchun 130012, China;
2. College of Army Artillery and Air Defense Crops, Shenyang 100867, China
Corresponding author: DU Peng, E-mail: 1584055676@qq.com
Abstract: The swarm intelligence algorithm for solving the problem of optimization dispatch of microgrid was investigated. A model for optimization dispatch of microgrid was established under the consideration of the process characteristics, such as multi-objective and multi-constraints. A modified flower pollination algorithm (MFPA) was proposed and applied to the optimization dispatch of microgrid. During initialization, opposition method was utilized to improve the diversity of the population as well as fully explore the space. During local updating, the new operation can accelerate the convergence. In addition, the condition for using genetic operations was defined in order to reduce the calculation and avoid the local optimal solution. Simulation results demonstrated that the performance of MFPA was better than those of FPA, GA and several other algorithms.
Key Words: microgridoptimization dispatchswarm intelligenceflower pollination algorithm(FPA)multi-objective
微网是由多种分布式电源构成的微电网, 是解决传统电网存在的远距离传输损耗大, 能源使用效率低, 环境污染等问题的有效途径.微网优化调度是电力系统的一类重要问题[1].其特点是多目标、多约束条件、非线性化, 所以传统数学优化方法难以满足其要求[2].由于群智能算法有不依赖于求解问题本身数学性质的优点[3], 所以此类算法作为微网系统的优化工具得到了广泛地使用[4-5].但寻找更适合解决微网优化调度问题的算法仍是研究难点之一.
受自然界中显花植物花朵授粉过程的启发, 2012年Yang提出了花朵授粉算法(flower pollination algorithm, FPA)[6].FPA提出后, 在相关领域得到了广泛的应用[7-8].
实验发现FPA相比其他群智能算法更适合解决微网优化调度问题.然而FPA在解决高维问题时, 存在寻优精度低, 容易陷入局部最优等缺点.本文在初始化和搜索阶段对FPA进行改进; 此外, 又增加了对遗传操作时机的判断, 提出改进的花朵授粉算法(MFPA).实验表明, MFPA比FPA拥有更高的全局搜索能力和更快的收敛速度.将MFPA应用于微网优化调度问题, 对算例的测试验证了MFPA的寻优能力和收敛速度均优于FPA和几种其他群智能算法.
1 改进的花朵授粉算法FPA所解决的优化问题一般形式如下:
(1) |
1.1 对立点初始化相关学者证明了对立点搜索的有效性[9].Xi的对立点定义为
(2) |
(3) |
1.2 “按需采纳”局部更新FPA的局部搜索是按式(4)进行的,
(4) |
定义1??个体与最优个体距离差值:
(5) |
(6) |
1.3 “趋同值”触发的交叉变异多数的群智能算法在生成新解之后就进行交叉、变异, 其缺点是过于盲目而带来不必要的计算量.本文引入“趋同值”作为判断是否需要对种群进行交叉、变异的依据.
定义2??个体“位置趋同值”:
(7) |
(8) |
(9) |
(10) |
(11) |
2 微网优化调度问题的数学模型2.1 微网的组成微网有两种运行模式:一种是孤立模式[10], 另外一种是联网模式[11].本文讨论的是在联网模式下只考虑微网从大电网购电的情况.
本文所优化的微网系统由光伏阵列(PV)、风力涡轮机(WT)、储蓄电池(BT)、微燃机(MT)和燃料电池(FC)组成.由于太阳能和风能发电的间断性和不稳定性, 所以把储蓄电池与它们联合使用, 以确保满足负载用电需求和减少化学能源使用.
2.2 目标函数微网优化调度是一个多目标优化问题, 本文提出的目标函数综合考虑经济成本、环境效益、健康影响等多个指标.采用线性加权法将多目标问题转化为单目标问题.其优化目标函数为
(12) |
(13) |
(14) |
(15) |
(16) |
2.3 约束条件微网优化调度问题是一类约束满足问题, 有运行电压约束、容量约束、传输约束、旋转备用约束等.本文所提出的调度模型中也考虑了这些约束, 如式(17)~式(19)分别是容量约束、平衡约束、交换约束.
(17) |
(18) |
(19) |
令P=[PPV, PWT, PFC, PMT, PBT, Pbuy], 则微网优化调度问题具有如下形式:
(20) |
在解决微网优化调度问题之前, 首先对高维函数进行测试, 分别测试了遗传算法(GA)[16]、萤火虫算法(FA)[17]、蝙蝠算法(BA)[18]、花朵授粉算法(FPA)[6]和改进的花朵授粉算法(MFPA).表 1表明MFPA的平均优化结果明显好于FPA等算法; 测试结果说明MFPA对解决高维优化问题有一定优势.
表 1(Table 1)
表 1 高维测试函数优化结果Table 1 Optimization results of higher-dimensions benchmark function
| 表 1 高维测试函数优化结果 Table 1 Optimization results of higher-dimensions benchmark function |
将MFPA应用到微网优化调度问题上得到如图 1所示的各微源最优调度.由图 1可知, PV, WT, FC和MT在用电高峰期出力较大以满足用户负载(242 kW)需求; BT在用电高峰期放电, 低谷期充电, 起到了削峰填谷的作用; Buy的走向说明微网在电价高的峰时段(11~15h, 17~19 h)购电量少, 电价低的谷时段购电量大, 满足经济性要求.
图 1(Fig. 1)
图 1 微网最优调度结果Fig.1 Best result of microgrid dispatch |
图 2体现了5种算法的运行效果, 可知当迭代次数较少时FPA和MFPA的效果相差不多, 随着迭代次数增加MFPA的收敛性要优于FPA和GA等其他算法, 并且MFPA得到的最优值要优于其他算法.可见MFPA能够获得更优的结果和更快的收敛性.
图 2(Fig. 2)
图 2 5种算法迭代过程Fig.2 Iteration process of the five algorithms |
表 2为算法GA, FA, BA, FPA, MFPA的100次运行所产生的最小值和平均值.从运行结果可以得出结论:MFPA的寻优能力和搜索精度要优于FPA和其他算法, 改善了原始算法的迭代效率.
表 2(Table 2)
表 2 5种算法运行结果比较Table 2 Results comparison of the five algorithms
| 表 2 5种算法运行结果比较 Table 2 Results comparison of the five algorithms |
4 结语对微网优化调度问题建模, 针对花朵授粉算法存在的问题, 提出了改进的花朵授粉算法并将其应用到微网优化调度问题上.通过对高维度测试函数的测试, 验证了算法性能优于原始算法.通过算例实验, 验证了改进的花朵授粉算法能够较好地解决微网优化调度问题, 提高了算法的性能.
参考文献
[1] | Wu Z, Tazvinga H, Xia X. Demand side management of photovoltaic-battery hybrid system[J].Applied Energy, 2015, 148: 294–304.DOI:10.1016/j.apenergy.2015.03.109 |
[2] | 李鹏, 徐伟娜, 周泽远, 等. 基于改进万有引力搜索算法的微网优化运行[J].中国电机工程学报, 2014, 34(19): 3073–3079. ( Li Peng, Xu Wei-na, Zhou Ze-yuan, et al. Optimal operation of microgrid based on improved gravitational search algorithm[J].Proceedings of the CSEE, 2014, 34(19): 3073–3079.) |
[3] | Othman M M, El-khattam W, Hegazy Y G, et al. Optimal placement and sizing of voltage controlled distributed generators in unbalanced distribution networks using supervised firefly algorithm[J]. International Journal of Electrical Power & Energy Systems, 2016, 82: 105–113. |
[4] | Tharakeshwar T K, Seetharamu K N, Prasad B D. Multi-objective optimization using bat algorithm for shell and tube heat exchangers[J].Applied Thermal Engineering, 2017, 110: 1029–1038.DOI:10.1016/j.applthermaleng.2016.09.031 |
[5] | Reddy S S, Rathnam C S. Optimal power flow using glowworm swarm optimization[J].International Journal of Electrical Power & Energy Systems, 2016, 80: 128–139. |
[6] | Yang X S. Flower pollination algorithm for global optimization[C]//International Conference on Unconventional Computation and Natural Computation. Berlin: Springer-Verlag, 2012: 240-249. |
[7] | Dubey H M, Pandit M, Panigrahi B K. A biologically inspired modified pollination algorithm for solving economic dispatch problems in modern power systems[J].Cognitive Computation, 2015, 7(5): 1–15. |
[8] | Abdelaziz A Y, Ali E S, Elazim S M A. Implementation of flower pollination algorithm for solving economic load dispatch and combined economic emission dispatch problems in power systems[J].Energy, 2016, 101: 506–518.DOI:10.1016/j.energy.2016.02.041 |
[9] | Rahnamayan S, Tizhoosh H R, Salama M M A. Opposition versus randomness in soft computing techniques[J].Applied Soft Computing, 2008, 8(2): 906–918.DOI:10.1016/j.asoc.2007.07.010 |
[10] | Abedini M, Moradi M H, Hosseinian S M, et al. Optimal management of microgrids including renewable energy scources using GPSO-GM algorithm[J].Renewable Energy, 2016, 90: 430–439.DOI:10.1016/j.renene.2016.01.014 |
[11] | Malysz P, Sirouspour S, Emadi A. An optimal energy storage control strategy for grid-connected microgrids[J].IEEE Transactions on Smart Grid, 2014, 5(5): 1785–1796. |
[12] | 钱科军, 袁越, 石晓丹, 等. 分布式发电的环境效益分析[J].中国电机工程学报, 2008, 28(29): 11–15. ( Qian Ke-jun, Yuan Yue, Shi Xiao-dan, et al. Environmental benefits analysis of distributed generation[J].Proceedings of the CSEE, 2008, 28(29): 11–15.DOI:10.3321/j.issn:0258-8013.2008.29.003) |
[13] | 丁明, 张颖媛, 茆美琴, 等. 包含钠硫电池储能的微网系统经济运行优化[J].中国电机工程学报, 2011, 31(4): 7–14. ( Ding Ming, Zhang Ying-yuan, Mao Mei-qin, et al. Economic operation optimization for microgrids including Na/S battery storage[J].Proceedings of the CSEE, 2011, 31(4): 7–14.) |
[14] | 陈洁, 杨秀, 朱兰, 等. 微网多目标经济调度优化[J].中国电机工程学报, 2013, 33(19): 57–66. ( Chen Jie, Yang Xiu, Zhu Lan, et al. Microgrid multi-objective economic dispatch optimization[J].Proceedings of the CSEE, 2013, 33(19): 57–66.) |
[15] | 毛晓明, 陈深, 吴杰康, 等. 分时电价机制下含蓄电池微网的优化调度[J].电网技术, 2015, 39(5): 1192–1197. ( Mao Xiao-ming, Chen Shen, Wu Jie-kang, et al. Optimal dispatching of microgrid containing battery under time-of-use price mechanism[J].Power System Technology, 2015, 39(5): 1192–1197.) |
[16] | Holland J. Adaptation in natural and artificial systems[J].The Quarterly Review of Biology, 1975, 6(2): 126–137. |
[17] | Yang X S. Firefly algorithms for multimodal optimization[C]//Proceeding of the 5th International Symposium on Stochastic Algorithms: Foundations and Applications. Berlin: Springer-Verlag, 2009: 169-178. |
[18] | Gandomi A H, Yang X S, Alavi A H, et al. Bat algorithm for constrained optimization tasks[J].Neural Computing and Applications, 2013, 22(6): 1239–1255.DOI:10.1007/s00521-012-1028-9 |