| 引用本文: | 王 辉,尚金成,文福拴.联合采用下偏差和信息熵描述风险的动态购电策略[J].电力自动化设备,2013,33(12): |
| WANG Hui,SHANG Jincheng,WEN Fushuan.Dynamic electricity purchasing strategy with risk characterized by lower deviation and information entropy[J].Electric Power Automation Equipment,2013,33(12): |
|
| 摘要: |
| 考虑到购电风险的根源在于不确定性，提出基于下偏差-信息熵的联合风险度量指标，其中下偏差衡量损失发生的可能性和损失的大小，信息熵衡量损失的不确定性。在机会约束规划的框架下，构造供电公司在多个市场购电的动态组合优化模型，将实际收益不小于给定目标收益约束在一定置信水平下，这样在风险最小化的同时把收益约束在可接受的水平。采用粒子群优化算法求解所发展的优化问题。仿真结果表明，所提出的方法能够更加全面和准确地度量购电组合风险，供电公司的目标收益和多个市场电价的不确定性程度都会影响其购电决策。 |
| 关键词: 电力市场 动态购电策略 风险 下偏差 信息熵 优化 模型 |
| DOI： |
| 分类号: |
| 基金项目:河南省电力公司科研项目（Ｈ２０１１３５４１） |
|
| Dynamic electricity purchasing strategy with risk characterized by lower deviation and information entropy |
|
WANG Hui1, SHANG Jincheng2, WEN Fushuan1
|
|
1.College of Electrical Engineering，Zhejiang University，Hangzhou 310027，China;2.Henan Grid Power Exchange Center，Zhengzhou 450052，China
|
| Abstract: |
| Due to the risks originated from the uncertainties，a risk measurement index based on the combination of lower deviation and information entropy is proposed，of which，the lower deviation is used to measure the probability and amount of loss while the information entropy the uncertainty of loss. A dynamic combined optimization model of electricity purchase by the power-supply company in multiple electricity markets is built under the framework of chance-constrained programming，which constrains the actual profit not less than the given target profit at a certain confidence level，resulting in the acceptable profit with the minimum risk. The particle swarm optimization algorithm is adopted to solve the developed optimization problem. Simulative results show that，the proposed method can more comprehensively and accurately measure the combined risk of electricity purchase and the electricity purchase strategy of a power supply company is dependent on its target profit and the uncertainty degree of electricity prices in different electricity markets. |
| Key words: electricity market dynamic electricity purchasing strategy risks lower deviation information entropy optimization models |
