Research On Modeling And Optimization Of Manned Lunar Exploration System Using Resuable Cislunar Transfer Spacecraft

Manned lunar exploration is of great significance to explore the historical origin of the Moon and solar system,to expand and enhance the space science and space application technology,to develop and utilize the space resources of the moon and to build the cislunar economic.Considering the long cycle,high cost and high risk characteristics of manned lunar exploration mission,it is necessary to carry out the modeling and optimization of manned lunar exploration system beforehand.It will clarify the rocket carrying capacity and optimize the type and scale of the spacecraft module which would reduce research and development difficulties,shorten the mission cycle and reduce the mission risk.As a high value component in manned lunar exploration system,the reusable research of manned spacecraft has been paid lots of attention.In this paper,the modeling method and the influencing factor analysis method of manned lunar exploration system is studied.The variable neighborhood radius negative selection particle swarm optimization algorithm is designed and validated.The reuse scheme of reusable cislunar transfer spacecraft is analyzed and the parking velocity increment of spacecraft based on LEO space station is modeled and optimized.The mass of manned lunar exploration system is then modeled and optimized.The description model,behavior model and interaction model of agent are established after analyzing the structure and function of manned lunar exploration system.The agent model of manned spacecraft,propulsion vehicle and rocket are established as well as the Markov decision process model for solving the mass scale of manned lunar exploration system.Considering the multi-factors coupling characteristics of the multi-agent system,the influencing factors analysis approach which combined computation experiment and data analysis method is proposed and validated.In order to improve the global convergence of standard particle swarm optimization(PSO)algorithm,the variable neighborhood negative selection PSO algorithm is proposed which combined variable neighborhood radius mechanism and negative selection mechanism from Bioimmunology.The design parameters and performance of the modified PSO are analyzed and validated based on standard test function.The reuse scheme of manned lunar exploration system is analyzed.The Earth reentry and orbit maneuvering model of cislunar transfer spacecraft is established.The path constraints such as dynamic pressure,overload and heat flux density of the reentry process are analyzed.The influencing factors of parking velocity increment and its influence rule are analyzed.The modified PSO is used to optimize the reentry orbit parameters,design parameters and the control parameters of spacecraft.The parking velocity increment can be optimized to about 100m/s with the parameters range in the examples.The change rule of parking velocity increment is discussed under the situation part of influence factors are limited as well as the times of atmospheric decelerations changes.The launch dynamic model of rocket and the path constraint model are analyzed.A two-layer optimization model for rocket mass which composes of improved PSO algorithm and Lagrangian equation is proposed.The calculation model of propulsion vehicle is established which taken the dry weight and the propellant evaporation into consideration.The influence factors of the initial mass in low Earth orbit(IMLEO)of the manned lunar exploration system are analyzed.The multi-agent reinforcement learning algorithm is used to optimize the selection of propellant type and propulsion task assignment.The flight mode analysis of the reusable manned lunar exploration system shows that the ideal docking orbit of earth-moon transfer vehicle and cislunar transfer spacecraft is the orbit below the safety radius of LEO space station orbit.A hierarchical hybrid optimization model based on modified PSO and reinforcement learning algorithm is established to optimize the Earth departure mass of manned lunar exploration system.In this paper,Agent-based modeling method is employed to realize the modeling of the individual and interactive behavior among the manned lunar exploration system.The method of computing experiment and data analysis is used to analyze the influencing factors of the system.Considering the multi-factor coupling characteristic in the manned lunar exploration system,the variable neighborhood radius negative selection PSO algorithm is proposed and validated with standard test function and application examples.The parking velocity increment in low Earth orbit of spacecraft is optimized using the proposed modified PSO algorithm.The parking velocity increment is optimized to about 100 m/s which demonstrate the feasibility of reusable scheme from the aspect of velocity increment.The mass scale of launch vehicle,IMLEO and Earth departure mass of manned lunar exploration system are optimized based on modified PSO algorithm,multi-agent reinforcement learning algorithm and hierarchical hybrid optimization algorithm respectively.It will provide a reference for the analysis and demonstration of manned lunar exploration system.