Multi Dimensional Data Quality Evaluation Method For Smart Distribution Network

With the development of information,automation and interaction of intelligent distribution network as well as the mutual penetration and integration on internet of things,electric power enterprises have accumulated a lot of data.Data quality problems such as anomaly,redundancy and omission have become increasingly prominent.As the basis and premise of improving data quality,data quality evaluation of smart distribution network is becoming more and more important.In this context,this paper studies the multi-dimensional data quality evaluation method of intelligent distribution network.The main work is as follows:Firstly,the multi-dimensional data quality evaluation system of intelligent distribution network is constructed.Based on the data characteristics and distribution characteristics of intelligent distribution network,aiming at the current situation that the basic measurement data of intelligent distribution network have many sources and the data quality is not high,the multi-dimensional data quality evaluation system of intelligent distribution network is constructed considering the importance,regionality,difference and real-time of data.This paper studies the idea of multi-dimensional data quality evaluation,chooses multi-dimensional data quality evaluation index,establishes a multi-dimensional data quality evaluation model,designs a multi-dimensional data quality evaluation process,and carries out multi-level,multi-directional and multi-angle analysis and mining on the data quality of intelligent distribution network,which ultimately enables power enterprises or users to understand the data quality of intelligent distribution network more comprehensively and systematically.Secondly,the realization method and weight calculation method of multi-dimensional evaluation index for intelligent distribution network are proposed.Aiming at the five evaluation dimensions of readiness,completeness,consistency,timeliness and redundancy,the index realization method and weight calculation method are studied respectively.Among them,the accuracy of intelligent distribution network data is judged by decision tree improved by association rules.Integrity is measured by whether there is record redundancy and non-existence of invalid data in the data set.Data redundancy of empty data realizes the judgment of data integrity of intelligent distribution network;consistency judges the data consistency of intelligent distribution network by means of least square method and linear regression analysis to check the reference consistency of the same data and the logical consistency of different data;timeliness judges the timeliness of data of intelligent distribution network through the time difference judgement of data updating;redundancy judges the timeliness of data of intelligent distribution network.Degree judges data redundancy of intelligent distribution network by detecting whether there is record redundancy and data redundancy in data set.Based on the needs of business applications and users' own needs,considering the importance of each evaluation index of intelligent distribution network,a weight calculation method combining subjective weight and objective weight is proposed.Through the above multi-dimensional evaluation index realization method and weight calculation method of intelligent distribution network,the final evaluation of the data of intelligent distribution network is completed.Lastly,the validity and applicability of the proposed method are verified by an example.The data of distribution transformer with 10 kV voltage grade and rated capacity of 200 kVA in a certain area are selected as the research object,and the data of intelligent distribution network in this area are evaluated and analyzed according to the evaluation and evaluation results of each dimension in the system,in order to verify the practicability of the evaluation system.