基于云理论的配电网空间负荷预测方法研究

提出了一种新颖的配电网空间负荷预测模型，该模型采用了基于云模型的知识表示，将定性概念所具有的模糊性、随机性有机地结合在一起，构成了定性和定量相互间的映射，克服了传统模糊理论的不彻底性。该文将云理论、面向对象的归纳方法以及粗糙集相结合，进行土地使用决策中空间属性信息定量定性转换、空间数据的离散化、决策规则的挖掘，克服了传统基于模糊集的空间负荷预测模型中模糊集参数及模糊系统规则选择的主观性，同时也利用云理论与粗糙集理论相互间的互补性,增强了知识发现的能力。运用基于云理论的不确定性推理计算小区对各用地类型适应性的评分，使推理结果更加合理而且贴近实际。给出了基于该模型的小区改造判据，改进了计及小区改造及经济性的用地分配的多目标规划模型，计算小区负荷。最后用实例说明了该方法的有效性。

A New Spatial Forecasting Method for Distribution Network Based on Cloud Theory

A new spatial load forecasting(SLF) model for distribution network is proposed. Cloud theory is introduced into SLF in this model, and the knowledge representation based on cloud model is used to integrate the fuzziness and randomness of qualitative concept, which constitutes the mapping between qualitative and quantitative analysis, conquers the halfway of conventional fuzzy theory. The cloud theory, Attribute Oriented Induction (AOI) and rough set are integrated to make transition of spatial attribute information between qualitative and quantitative analysis, spatial data discretisation and decision rules mining so that it conquers the subjectivity of fuzzy sets parameters and fuzzy system rules choice that based on conventional fuzzy sets spatial load forecasting .The complementary of cloud theory and rough set is also used to enhance the knowledge discovery ability. The uncertain reasoning based on cloud theory is used to calculate the grading to adaptability of every small area land-use style ,which make the reasoning result more reasonable and actual. The small area redevelopment criterion based on this model is provided, furthermore, the linear multi-objective programming is improved for the whole-optimal land distribution in which the economic effect and redevelopment is considered , and the load of the small area is calculated. Finally, this model is implemented and the analysis result for practical calculation example by this model is satisfied.