Application Of Swarm Intelligence Algorithm In Rocket Engine Health Management System

Space technology has become a new focus for major countries to compete for military advantages.The development of space technology is of great significance to ensure national defense security.Rocket propulsion technology is the core of aerospace technology.It has important value for fault detection and diagnosis of rocket power system,and is the guarantee of successful launch of spacecraft.In this paper,two different swarm intelligence optimization algorithms combined with neural network and machine learning are selected to establish the fault detection and diagnosis model of liquid rocket engine,and the existing data are used to simulate the fault detection and diagnosis of liquid rocket power system.The main research contents are as follows:(1)Neural networks are widely used in the field of fault detection.In view of the shortcomings of BP neural networks,such as slow convergence speed,inability to guarantee convergence to the global minimum,and uncertain network structure,this paper first uses Tent chaotic mapping to improve the basic atomic search algorithm,and then uses the improved atomic search algorithm(IASO)to optimize the weights and thresholds of BP neural networks,Thus,the IASO-BP model is constructed and applied to the fault detection of liquid rocket power system.The simulation results show that the BP neural network optimized by the improved atomic search algorithm overcomes the defects of slow convergence speed and can not guarantee convergence to the global minimum point,and has a high accuracy for fault detection of liquid rocket power system,which can be used for fault detection of liquid rocket power system.(2)Support Vector Machine(SVM)is often used in classification problems,among which the optimization of kernel function parameters and penalty factors is the key technology,which has a great impact on the classification results.To solve the above problems,this paper uses Squirrel Search Algorithm(SSA),which has good optimization ability and real-time performance,to optimize the kernel function parameters and penalty factors of SVM,and constructs SSA-SVM model for fault diagnosis of liquid rocket power system.The SSA-SVM model is simulated by 846 groups of existing historical data,and compared with the unoptimized SVM model and the particle swarm optimization SVM model(PSO-SVM).The results show that the SSA SVM model has a high classification accuracy and can be applied to the fault diagnosis of liquid rocket power system after the SSA algorithm optimizes the kernel function parameters and penalty factors of SVM.(3)In order to make the results easy to observe,a stand-alone version of the rocket health management system is constructed.The system is composed of four modules: algorithm module,data conversion,algorithm operation,and management module.It can realize the functions of data conversion,algorithm import,simulation results display,and has the advantages of good stability,security,and cross platform use.