Research On Design Of Monitoring Terminal And Load Forecast Of Distribution Transformer

As a hotspot in the new stage of power grid construction,the ubiquitous electric power Internet of Things aims to open up the information islands and build an application system to upgrade and expand the existing information systems with connecting all electric equipment,electric operators,power generation enterprises and customers.As an upgrade of the original distribution automation system,the emergence of the ubiquitous electric power Internet of Things has made it connected.At the same time,it also made it possible for advanced applications such as intelligent data collection,power distribution equipment failure prediction,and user power behavior recommendations.As we know,distribution transformer is an important device in the power grid.Its operating parameters can intuitively reflect the transformer status and power consumption.Therefore,monitoring the distribution transformer is of great significance.However,there are still some drawbacks in the grid system.Firstly,the development of the distribution network is relatively backward compared to the transmission grid.Secondly,the distribution network system lacks a complete monitoring and fault handling system.In addition,Existing system failing to resolve data communication between power distribution terminals and hardly involving load forecasting.This may cause some problems in the actual power system.Such as a large number of distribution transformers are still in the dead zone of detection because it requires expensive equipment and sensors.And lack of load forecasting capabilities.Based on the idea of a power distribution terminal in an automation system,this paper proposes a monitoring scheme that combines transformer data with environmental data after considering the influence of the environment on the working condition and load of the transformer.By combining the monitoring of the transformer body and environmental data,it can complete a series of functions such as data transmission,data storage,data display.In addition,it can accurately understand the temperature rise value and load factor of distribution transformers.It provides a basis for later equipment maintenance and load status analysis while improving economy and applicability.Thinking of the transformer can collect load data and to solve the poor prediction capability of the existing data collection system SCADA.After analyzing and summarizing a variety of prediction methods,this paper raises a prediction algorithm combining(PCA)and neural network.The algorithm is applicable to data forecasting of power distribution transformers and power grid load forecasting.First of all,it is necessary to process the massive load data,filter out the repetitive and extreme data,and then overcome the inherent shortcomings of artificial neural network by genetic algorithm.Accurate load forecasting results are helpful for the power generation department to formulate maintenance plans for power supply and distribution equipment.More importantly,it can provide an important basis for enterprise to formulate the plans.Finally,this paper uses Matlab simulation to obtain prediction results and compares them with real load values and unoptimized neural network algorithms.Experimental results show the superiority of the scheme.It can replace the original 20-day data volume with the screened power load values of 5 days.And the load prediction accuracy is better than the original algorithm.