基于改进层次分析法的电力营销状态评估及预警模型研究

Research on power marketing status assessment and warning model based on improved analytic hierarchy process

针对智慧电网建设全面展开、电力营销市场化改革新形势下传统电力营销模式呈现的复杂信息多维转换损耗较大、评估等级区间界限的软化度较低、状态评估预警因子体系拟合度较差等问题,提出了基于改进层次分析法的电力营销状态评估及预警模型。在层次分析法中引入模糊神经网络,建立电力营销状态评估及预警自适应模型并融入多目标决策机制建立面向电力营销的状态动态评估及预警模型,在Python 3.5.2环境下进行模型效能仿真验证,较好解决了多维应用背景下的电力营销的状态动态评估及预警问题,具有状态评估等级区间界限的软化、评估因子定性、定量表示的低损耗转换、状态评估预警因子体系精确拟合等优势。以我国南方电网某电力公司为效能评价载体,基于VS2016平台开发了验证环境并对模型进行了实证分析,分析结果表明,所提模型可以实现全方位的电力营销状态动态评估及预警,在状态评估分区界限的软化度、状态评估预警因子拟合度、复杂信息多维转换损耗等方面具有明显优势。

In view of the problems of the comprehensive development of smart grid construction and the reform of power marketing in the new situation,such as the large loss of complex information multi-dimensional transformation,the low softness of the evaluation grade interval boundaries and the poor fitness of the state assessment and early warning factor system,the power based on the improved analytic hierarchy process(AHP)is proposed.Marketing status evaluation and warning model.Introducing the fuzzy neural network into the analytic hierarchy process,establishing the self-adaptive model of power marketing condition evaluation and warning and integrating the multi-objective decision-making mechanism to establish the dynamic state evaluation and warning model for power marketing.The model effectiveness simulation is carried out under the environment of Python 3.5.2,which solves the power battalion under the background of multi-dimensional application better.The problem of dynamic state evaluation and early warning of sales has the advantages of softening the boundary of state evaluation grade interval,qualitative and quantitative expression of evaluation factors,low loss conversion,and accurate fitting of state evaluation and early warning factor system.Taking apower company of China Southern Power Grid as the carrier of efficiency evaluation,a verification environment is developed based on VS2016 platform and the model is analyzed empirically.The analysis results show that the model proposed in this paper can realize all-round dynamic evaluation and warning of power marketing status,the softening degree of the boundary of state evaluation zones,the fitting degree of state evaluation and warning factors,the multi-dimensional transformation loss of complex information,etc.It has obvious advantages.