Real-Time Digital Simulation of Dynamic Systems Using Parallelism.

In recent years, there has been a growing need for the real-time digital simulation of dynamic systems. This technique can help engineers not only understand a complex physical system to support the development of sophisticated digital controls but also evaluate the validation of digital control hardware and software developed. For a complex system, more computing power is required. However, the consistent escalation of component speed characterizing earlier decades did not occur in the last decade, we must seek greater computing speed through the parallel use of conventional components. Parallelism can be achieved through either the development of parallel integration algorithms or the decomposition of the dynamic model. There are three main results in this thesis. (1) New parallel algorithms which cut down the computation time per step are developed. (2) Their characteristics on the parameter plane are investigated with respect to accuracy and stability range. (3) An algorithm to distribute those time-consuming function evaluations between processors in a distributed system is formed. Two aircraft models are studied to illustrate the idea. The hardware and software problems in the implementation of the system by using microprocessors are also investigated. The simulation results and computation time savings strongly support the developments in the real -time digital simulation area.