Research On Health Prediction Method Of Gun Control System Based On Machine Learning

As the core component of tank weapon,gun control system is mainly responsible for controlling the firing direction and accuracy of artillery,and its performance is the key factor to ensure the tank’s fighting force and survival ability in the battlefield.However,because of its complex control system structure and harsh combat environment,it is very important to study its health prediction.With the deepening implementation of our military’s operational concept of "enhanced firepower attack",the powerful and reliable gun control system has gradually become a key link to ensure the combat effectiveness of tanks.Therefore,on the background of the actual project,this paper proposes a health prediction method for the gun control system of a certain type of tank,and designs the corresponding software platform of the prediction system.This method can effectively reduce the probability of possible failure of the gun control system.The details are as follows:Firstly,through the in-depth analysis of the structure and working principle of the gun control system,combined with the diversity,complexity and suddenness of its fault mode,a health assessment method based on fuzzy comprehensive evaluation method and improved combination weighting method is proposed.According to the multi-index evaluation characteristics of the gun control system,the concept of relative deterioration was introduced to unify the numerical range.The gun control system is divided into five health status levels by fuzzy comprehensive evaluation method,and the membership function matrix of the five health status levels is constructed.The improved combination weighting method is used to weight the indicators to calculate the health degree of the gun control system.The results show that the health status evaluation results of the gun control system are consistent with the actual situation,which proves the accuracy of the proposed method,and provides reliable data support and decision-making basis for the subsequent establishment of health prediction model.Secondly,a combined prediction method based on improved grey Wolf optimization algorithm was proposed.The convergence factor of the grey Wolf optimization algorithm was adjusted to improve the global search ability and avoid falling into the local optimum value.At the same time,autoregressive moving average model,radial basis function neural network,least squares support vector machine and other models with good health prediction effect were selected to search the optimal weight coefficient and establish a combined prediction model.And the gun control system data are respectively brought into the single model,the improved grey Wolf optimization algorithm combination model and other comparison combination models for health prediction.Through the evaluation analysis of the prediction result error and the comparison of the remaining service life error,the results show that the improved grey Wolf optimization algorithm is better than the traditional grey Wolf optimization algorithm in searching the optimal weight coefficient of the combination model,and the prediction accuracy is better than other combination models,and the prediction effect is more accurate and stable.Finally,based on QT Creator development environment and C++ programming language,SQL database is selected for data storage,and an intuitive and fast gun control system health prediction system is developed,and its various functional modules are displayed and explained.