Research On Rapid Control Prototyping And Real-time Simulation For Aero-engine

Rapid control protyping and real-time simulation are key technology of model-based design foraero-engine digital control system.In rapid control prototyping, mechanism and application of real-time target of MATLAB werestudied. In MATLAB/xPC real-time target environment, recursive least squares method was used forparameter identification of a certain type of DC motor, and model validation was carried out. On thisbasis double-loop controller parameters were tuned to achieve double-loop control prototype for DCmotor in millisecond control step. Experiments showed that this method fully integrated the Simulinkgraphical programming with high real-time performance of xPC target, so algorithm development wasconvenient and efficient, and control step could be up to10microseconds.In real-time simulation, real-time system RTX was introduced to ensure real-time performance ofWindows system simuolation computer, and real-time software development based on RTX and MFCwas studied. Real-time mechanism was studied and real-time performance was tested between Win-dows and RTX. Integrated multi-threaded software model based on Windows MFC and RTX wasproposed, and model and controller software were developed on distributed real-time simulation plat-form based on RTX and reflective memory network. Experiments showed that this method combinedboth rich graphical interactive interface and easily design of software of Windows and high real-timeperformance of RTX, simulation step could be sub-millisecond.In controller and model algorithm development for rapid control prototyping and real-time si-mulation, rapid development of real-time software was studied based on MATLAB and its automaticcode generation. Mechanism of automatic code generation of MATLAB and the code structure wasstudied, aimed at platforms DSP and VxWorks real-time system which were commonly used inaero-engine control system, real-time code of typical algorithm LQR controller and Kalman filterwere developed separately using embedded MATLAB and Simulink/RTWEC. Key issues includeextract, modify and integration of the real-time code were solved, and seamless transition and migra-tion of M scripts and Simulink software to real-time software were achieved.Finally, MBD process was studied on the unified platform supporting development, tests and va-lidation platform MATLAB. The function of related tools was studied, such as requirement manage-ment tool, design verification tool, test case generation and management tool and so on. Problems ofestablish of custom librariy in Simulink, integration with existing C code and custom code generationwere solved. At last, using MBD method to develop control software for micro-turbine engine, elec- tronic controller in loop simulation showed that the use of MATLAB for model-based design was ef-ficient and iteratable.