Design And Implementation Of Electric Transmission And Transformation Equipment Fault Question Answer System

Electric transmission and transformation equipment is not only the component of the power system,but also the key to ensure the safe and reliable operation of the power system.However,the equipment is prone to produce a variety of fault problems in the process of long-term operation.Once the equipment failure will inevitably affect the safe and reliable operation of the power system,and due to many factors leading to equipment failure,maintenance personnel can not deal with all kinds of equipment failure,often encounter many problems in the process of equipment maintenance.To solve the relevant problems in the process of troubleshooting can help the maintenance personnel to better carry out equipment maintenance,improve the operation stability and economic benefits of the equipment,so as to ensure the safety and reliability of the power system.Based on this,this paper designs and implements a power grid power transmission and transformation equipment fault question answering system based on natural language processing technology.The system can answer the user’s input power grid power transmission and transformation equipment fault questions,and also can analyze the equipment fault phenomenon contained in the questions.Firstly,this paper studies the design of efficient text matching algorithm.Text matching algorithm based on deep learning has achieved many excellent research results,but most text classification models based on neural network are mainly realized by semantic interaction of word granularity.Moreover,due to the multi-layer network structure,the low-level features of the model can not be effectively used,and the multi-layer network structure makes the model easy to appear The problem of gradient vanishing.Based on this,this paper designs and implements a text matching algorithm based on multi-level semantic interaction.The algorithm uses multi-level semantic interaction.On the basis of using word granularity semantic interaction,combined with attention mechanism,this paper proposes a semantic interaction method that can carry out sentence granularity interaction,so that the model can fully carry out the interaction task between texts.In addition,the model introduces the enhanced residual connection structure to realize the effective use of the underlying features.In the model design,attention mechanism is used to improve the encoder to help the model better extract information from the text.Then,this paper studies the design of efficient text classification algorithm.Methods based on deep learning are widely used in the field of text classification,including convolutional neural network,LSTM and so on.Because CNN and RNN can capture local features and sequence features,these methods can capture semantic and syntactic information well in the current text,but these methods will ignore the discontinuous global word co-occurrence information and long-distance semantic information in the corpus.Based on this,this paper designs and implements a text classification model based on graph volume product word vector,which can better fuse the global word co-occurrence information and long-distance semantic information with the local sequence information of the text,so that the model has better performance.At the same time,the model can still maintain good performance in the case of less annotation samples.Finally,based on the above algorithm,this paper designs and implements the power transmission equipment fault question answering prototype system.The system is based on three different types of data:power grid fault text,power grid fault Q&a pair data and power grid standard data,which can make full use of relevant text knowledge.This paper uses relevant tests to verify the function of the system.