Investigation Of Intelligent Damage-Mitigating Control Methods For Reusable Liquid-Propellant Rocket Engines

This dissertation integrates theoretical analysis with simulation computation to develop researches on methods of intelligent damage-mitigating control (DMC) for the reusable liquid-propellant rocket engines (LRE), a certain LOX/Kerosene rocket engine being selected as an example of the future reusable LRE according to the status and the trend of LRE in China. The goal of DMC of LRE is to achieve a desired trade-off level between the damage of critical components and the performance of LRE system so as to improve the reliability, the security, the structural durability of critical components, and to extend the service life of the LRE.Firstly, the system dynamics model of the LRE and the structural and damage models of the critical component are built. With these models the computations and analyses are carried out. Some results are obtained. For example, developed is the LRESim, a universal simulation module library for the LRE, which can be applied to investigate the related problems of the startup transient and the steady state process, or to demonstrate the different schemes in the phase of design of LRE. The finite element method (FEM) and the intensity calculating method (ICM) are used to analyze the structural dynamic characteristics, respectively. It is a kind of preferable method in analyzing real-time online structural dynamic characteristics at present. The ICM acts as a main online computing method, while the FEM provides offline correction for the computing results. The damage-calculating model based on the continuous time is built for the material constructed turbine blades. The simulation computations are carried out. The results indicate that the model is suit of this research.Secondly, the main innovative works are introduced. It is presented and clarified that the analysis and synthesis of the DMC law is based on the multi-objective optimization. The classical and intelligent optimization methods for the DMC are investigated for the reusable LRE, respectively. The theory and methodology of the analysis and synthesis on the DMC law based on the classical optimization, such as the main objective method (MOM) and the linear weighted sum method (LWSM), are clarified and discussed with the simulation computation. The merits and demerits of the MOM and the LWSM are pointed out. A general method and procedure for the analysis and synthesis of the DMC law by applying the genetic algorithm for the LRE is presented, and the Strength Pareto Evolutionary Algorithm (SPEA) is used to solve the DMC law. The results obtained with the simulation computation indicate that the genetic algorithm can overcome the demerits of the classical optimizations in solvingthe DMC law for the LRE. The fuzzy neural network (FNN) is employed to design a damage-mitigating controller (DMCer), the structure and the design procedure of the FNN DMCer is analyzed, and the simulation computation with the FNN DMCer in the LRE is implemented. The results show that the FNN DMCer can implement tradeoff between the system performance and the critical component damage for the LRE, and the expected goal of damage-mitigating is achieved.Finally, with the development of the intelligent damage-mitigating control technology, the results obtained will provide the theory and method of DMC for the reusable LRE in future, and also provide valuable reference for improving the design of existing and developing LRE.