Research Of Multidisciplinary Design Optimization Methods Based On Design Space Decrease

The scientificity and frontier of modern instrument design methods have made a profound impact on the design works of instrument system,and greatly enriched implementation means to develop new and various instrument system products with different performances.Multidisciplinary Collaborative Optimization(CO)method to design a modern instrument system is deeply researched in this paper.COmethod could obtain the optimal design of instrument system synthetically considering different disciplinary parameters,such as the part sizes,performance indexes,and economy indexes and so on.COmethod could improve the design quality of complex instrument system and further develop the theories and practices of modern instrument design methods.As a result of the unique mathematical structure,the standard CO method sometimes encounters its predominant inherent drawbacks——the solving difficulties in system-level optimization problem and the inefficient calculation.To enhance the convergence ability,global optimization capability,and robustness of the standard CO method,the causes of the predominant drawbacks inherent in the standard CO have been theoretically analyzed in this paper.And Design Space Decrease Collaborative Optimization(DSDCO),an effective improved method,is presented.Its key idea is the design space decrease method,i.e.,the system-level optimization solutions are forced to close to the original feasible region as the system-level design space is gradually decreased,and the global optimal solution will be obtained when the system-level optimization solution moves into the original feasible region.The system-level constraints are reconstructed by replacing the consistency equality constraints in the system-level optimization formula of standard CO with variable boundaries in DSDCO.So DSDCO can effectively overcome the predominant drawbacks inherent in the standard CO method,enhance the abilities of CO method to solve the practical problems and further enrich the multidisciplinary collaborative optimization theories.The original design space can be intactly retained in the updated system-level design space when design space decrease method is used to solve a problem,which makes the system-level optimization solution close to even equal to the original global optimal solution.So,how to combine the design space decrease principles and the global optimization method to solve a nonlinear programming problem is deeply studied,and Increasingly Removing Infeasible Region(IRIR)global optimization method is presented.IRIR offers a new way to explore a global optimization algorithm with a better solving ability.To accelerate convergence,Enhanced Design Space Decrease Collaborative Optimization(EDSDCO)method is presented.EDSDCO reconstructs the formula of system-level and subsystem optimizations by introducing the set of auxiliary design points and the expression of KKT conditions for the original problem.Consequently,the searching range of DSDCO is decreased;the computational efficiency is effectively improved under the precondition of not increasing complexity of computing system-level design space.The coupling between the system-level optimization and subsystem optimizations is largely weakened,so that the robustness of DSDCO is increased.Taking a two-dimensional problem as an example,the solving principles of the above-mentioned new methods are geometrically analyzed.Mathematical computational formulas are deduced.After solving some typical test examples,contrastive analyses of their performances are done,their advantages and shortcomings are summarized,and their validity and effectiveness are verified.The results indicate that DSDCO,IRIR,and EDSDCO can be used to solve a optimization problem regardless of the convexities of constraints and positions of starting points,so they are very practical;DSDCO and EDSDCO can effectively overcome the predominant drawbacks inherent in the standard CO method,so they can effectively increase the practicability and reliability of CO to solve a complicated instrument or engineering system optimization design problem;moreover,EDSDCO is more robust and faster than DSDCO.Subject to some working requirements,the F-35 rudder Electro-Hydrostatic Actuator(EHA),a typical multidisciplinary instrument system,is designed using EDSDCO with the objective of the fastest moving speed.The optimized EHA's feasibility is verified,and its stability,rapidity,and accuracy are analyzed by simulation experiments.Results indicate:EDSDCO is effective,feasible,and practical for the multidisciplinary design optimization problem of EHA.The application of EDSDCO makes the design process of EHA more automated and scientific.So designing EHA or other complex instrument system using EDSDCO is a valid way to enhance the design qualities.In conclusion,aiming at overcoming the predominant drawbacks inherent in the standard CO,taking improving the practicability and reliability of the CO as core,centering on enhancing the convergence,global optimization capability,robustness,and convergence rate of a MDO method,DSDCO and EDSDCO are presented in this paper,and their correctness are assessed and verified by solving some typical test examples.A new idea is offered to enrich and develop the theories of CO and global optimization method.F-35 rudder EHA is designed using EDSDCO,which expands the application of CO theory and provides theoretical basis and practical example for the innovative designs of EHA and other complex instrument systems.