| The development of fault diagnosis systems has received more and more attention from countries and regions,especially the fault diagnosis of spacecraft.Because spacecraft is expensive,and many components are unique.In order to make the spacecraft more autonomous and intelligent,its fault diagnosis technology has evolved from a single to a complex integrated health management system.The research background of this paper is based on the spacecraft’s health management system.The complete health management project is divided into three modules: the user end,the parsing end and the back end algorithm part.This paper focuses on the core algorithm in the health management system,which is mainly implemented in the Common lisp language.In the aspect of fault diagnosis search,the traditional fault diagnosis methods mostly use knowledge-based methods based on expert systems.This article will use the improved A* algorithm to troubleshoot the spacecraft.The improved A* algorithm greatly improves the speed of searching for fault points and reduces search time.Reducing the time of fault search is to avoid doing useless work when looking for the optimal solution.Its key point is decision making.Many decision problems are now redefined as optimization problems,involving searching on discrete spaces and obtaining a set of optimal search schemes that satisfy constraints.The A* algorithm is most commonly used to find the optimal path,such as navigation or game maps.However,the A* algorithm has been successfully applied to Livingstone,the NASA fault diagnosis system in the United States.It turns out that the A* algorithm is completely feasible for fault diagnosis.In this paper,the fault diagnosis search problem is also compared to a relationship of finding a path.An inference search tree is established.By finding the fault point in the inference search tree,a constraint condition-cost function is added to expand the branch in the search tree.At the time,the cost function is used to judge and filter,so that it can reduce many unnecessary searches and search for the optimal solution with the least amount of time.In this paper,the traditional A* algorithm is improved.By adding constraints,that is,the cost function,based on the traditional A* algorithm and adding additional structure,especially considering the state in the variable assignment,and the constraint set is Decomposed into a set of constraints.By using mutual priority independence(MPI)to solve the optimal constraint satisfiability problem(OCSP),the optimal constraint satisfiability problem is constructed as a state space search problem,and each search state corresponds to the partial allocation of decision variables.Combine the cost of each allocation to calculate the invalidity,and achieve the purpose of fast fault search. |
