The Statistics And Analysis System Of Region Electric Outage Information Based On Energy Management System

The normal dispatching business is heavy and complex,whereas the information submission revolving is required to be timely and accurate.To complete the power failure information report submission work in the early stage,it is generally necessary to arrange a special person to spend one or two hours to solve the data collection,summary and submission work on the same day,which tremendously affects the daily work.Nowadays,there are more and more dispatching automation means,however,most of the power outage information reports still rely on manual submission,which is not convenient for the provincial dispatching organization to grasp the power outage data in real time upon the whole province.Moreover,due to the unskilled business of the dispatchers and the high pressure of conventional work,there are often problems about false and missing reports,which do not satisfy the requirements of dispatching reports for information accuracy.Based on the investigation upon the accuracy rate of information submission across all counties in the whole province,it is found that the error rate of the information collection revolving to the four links of power outage information report submission is relatively high.After summarizing the possible reasons,it is found that the main reason for the low accuracy rate of power outage information report submission lie in the lack of automatic information collection submission module together with single information submission mode.To settle these problems,in addition to strengthening the management of information submission process,it is still necessary to improve the efficiency of information submission methods.With the continuous development of machine learning application field,the previous work,which is difficult to be completed by exhaustive attack method,can be solved by using machine learning related algorithms.The decision tree algorithm in machine learning has good application effect for data classification and other work.The Decision Tree algorithm in machine learning has a good application effect upon data classification and other work.In the study of Decision Tree algorithm,it is found that C4.5 algorithm has gained more and more recognition for the processing of discrete data and continuous data due to its wide application scenarios and good classification effect in recent years.However,there are a lot of associated data for power outage information classification work which C4.5 algorithm can complete the free matching classification of massive data.After the construction of Decision Tree.the weak correlation and invalid branches can be eliminated by pruning algorithm to ensure the overall structure of decision tree is succinct,and the effect ion of such classification has stronger universality.To verify whether the power outage information classification model constructed by applying the Decision Tree algorithm satisfies the requirements,it is also necessary to evaluate its classification effect with relevant evaluation indexes,that is,only when it meets the use requirements can it be included in the power outage information statistical submission business.The main innovations of this paper are as follows:1.The overall process of power outage information reporting at the provincial company level is systematically analyzed for the first time.Each link is disassembled and analyzed one by one using statistics related knowledge,thus sorting out the reporting time of power outage information at each stage,and solutions are proposed for the long-term link.2.A learning model based on Decision Tree Algorithm is proposed.The variables affecting outage data are reduced and generalized.The rules of power failure data classification are established,and the Decision Tree Algorithm is applied to the basic data classification link of power failure information for the first time,which provides important data support for the construction of the statistical analysis system of power failure information.3.Combined with the existing OMS-related function module,the mechanism of time-sharing reporting for power outage information is realized,which greatly improves the accuracy of the reporting for the county-level dispatch report,and the whole process has a low degree of manual participation and reduces the work pressure of the personnel on duty.4.It is the first time to propose and develop the power outage information extraction and display based on EMS system,which realizes the full-automatic closed-loop management of automatic information collection,automatic type judgment,automatic filtering for invalid information,and automatic statistics of lost load,which is helpful for regulators to timely insight into abnormal load reduction.