A Research Of The Intelligent On-line Monitoring System Of Transformer

In power grid,the number of power transformer is of great large and the location is scattered,it has so many problems such as high maintenance costs,non-conducive to centralized monitoring and operating data and running status of transformer often appear unable to real-time monitor the problem,in the operation process of abnormal phenomenon is difficult to find in time,is of inconvenience to the user,not but conforming to the state's energy efficiency targets and the requirements of state grid corporation of fine management.This paper comprehensive comparative analysis of the current main transformer monitoring equipment focuses on monitoring parameters and function,on this basis,this paper proposes a multi-parameter fusion based intelligent transformer monitoring and diagnosis of early warning system.Intelligent monitoring and early warning of the technical requirements of transformer,according to the different parameters,the installation position of running monitoring and sensor types,a re-understanding of the Lord was proposed and designed for intelligent on-line monitoring system;At the same time,the characteristics of transformer fault type and aura of the classification research,get common multiple precursor of fault feature parameters,according to the basic monitoring data of direct access to the derivative can reflect the fault warning data calculation,determine the characteristic parameters of the fault early warning analysis,so as to set up a new type of intelligent monitoring and early warning system of transformer platform,has carried on the corresponding hardware and software design.Then,on the basis of the analysis compares several kinds of prediction algorithm,combined with the characteristics of transformer intelligent monitoring and early warning,the BP neural network is adopted to establish the mathematical model of the transformer intelligent online monitoring and early warning system,aiming at the design of the system platform of transformer run continuously collected data,select the five characteristic parametersof the BP neural network training,the characteristic parameters of the group can reflect the transformer fault is forecasted,and compared with the measured fault precursor data,the simulation results show that the proposed in this paper by using BP neural network for specific failure precursor prediction method of characteristic parameters in accordance with the basic of the failure data,so as to prove that the system can realize intelligent on-line fault of transformer monitoring and early warning.Through in this paper,the design of intelligent monitoring and early warning system,can not only achieve almost all in the process of transformer running state parameter of continuous measurement for a long time,but also on the basis of these data automatically precursor characteristics of transformer faults as well as search,judgment,and automatically send out warning signals;At the same time,a large database of transformer operation parameters can be set up for further analysis of the cause of transformer fault.