Application Of Data Mining Technology In Alarm System Of Power Distribution Network

In dispatching automation system of distribution power grid, there is a large amount ofinformation which usually contains a lot of noisy information. The traditional alarm systemsimply uploads alarm information, so the untreated and fast changing mass alarminformation will result in ignoring really important information for dispatcher, and it willalso delay the processing of the fault and endanger the safety of the power grid.Aming at the problems above, this paper introduces data mining technology to thenoise processing of alarm information in power grid，using the improved ID3algorithm toconstruct noise filtering classifier dealing with alarm information in power grid. Theclassifer constructed in this paper can divide the alarm information into two kinds: noisyinformation and normal alarm information, and it can resolve the defect that the alarminformation is uploaded to dispatcher without recognition. The main work done by thispaper is as follows:(1) This paper studies the data mining technology deeply, mainly discusses decisiontree algorithm in detail; This paper sums up the characteristics, strengths and weaknesses ofthe follow algorithm: ID3algorithm, C4.5algorithm, CART algorithm, SLIQ algorithm,and SPRINT algorithm.(2) This paper discusses two shortcomings of the classical ID3algorithm: toocomplicated in calculating and leaning on the attribute which have more value whileselecting the optimum attribute, aiming at these two aspects, this paper puts forwardcorresponding improved method. ID3algorithm which bases on information entropy theoryhas multiple logarithm operation in choosing the splitting attribute when building thedecision tree, so the computing speed will slows down especially when the data is large,aiming at this problem, this paper puts forward corresponding simplified method.Furthermore, the ID3algorithm applicates the information gain as the criteria of seletingsplitting attribute, so it often leans on the attribute which has more value, aiming at thisproblem, we introduce a divisor to the simplified ID3algorithm to overcome the bias onattribute.(3) Through the analysis of examples, the superiority of improved ID3algorithm isverified and this paper carries out the construction of decision tree by MATLAB language. (4) Based on the study of the alarm information’s characteristics, this paper realizesthe preprocessing and relationship induction of the alarm information. The alarminformation has complex forms, so it’s necessary to preprocess it and summarizerelationship from it before data mining.(5) Constructing noise information recognition classifier used in alarm information ofdistribution power grid with improved ID3algorithm, this classifier can recognize noiseinformation from alarm information.