基于贝叶斯网络与高斯混合聚类的配网设备数据分析模型设计

Design of distribution network equipment data analysis model based on Bayesian network and Gaussian mixture clustering

为了减少由配电网设备问题所造成的人力物力浪费,且在有限的品控投入下提高效率,设计了一种基于贝叶斯网络与高斯混合聚类算法的配网设备数据分析模型。该模型筛选出配电网设备中频繁出现的问题及影响较大的设备作为研究对象,并利用高斯混合模型对相似设备进行聚类分析。同时采用深度贝叶斯概率网络算法建立与设备运维成本相关的概率模型,再进行数据挖掘和分析建模。以两个典型县级供电局的配电网设备数据信息为对象,验证了所提模型的有效性,进而实现了基于数据的设备品质分析及针对设备效用性能的价值评估。

In order to reduce the waste of human and material resources caused by the distribution network equipment problems and improve the efficiency under the limited input of quality control,a distribution network equipment data analysis model based on Bayesian network and Gaussian mixture clustering algorithm is designed.The model selects the frequent problems in the distribution grid equipment and the equipment with large influence as the research object,uses the Gaussian mixture model to cluster analyze the similar equipment,and uses the deep Bayesian probability network algorithm to establish the probability model related to the equipment operation and maintenance cost,and carries out data mining and analysis modeling.Taking the distribution network equipment data information of two typical county-level power supply bureaus as the object,the effectiveness of the proposed distribution network evaluation model is verified,and the equipment quality analysis based on data and the value evaluation of equipment utility performance are realized.