Parallel Computing Of Reactive Power Optimization For Power System Based On OpenMP

People’s requirement for the accuracy of reactive power optimization is increasing, because reactive power optimization is related to the quality of voltage as well as the security and economy of power system operation. Thus the model of reactive power optimization is more and more complex, and the calculation quantity is larger and larger. The efficiency of traditional serial computing have been unable to meet the demand of online control of reactive power optimization, therefore, this thesis aims to improve the calculation speed of reactive power optimization by using parallel processing technology.Firstly, a reactive power optimization algorithm and its implementation process were introduced in this thesis. Meanwhile, its advantages and disadvantages in improving the optimization speed as well as the advantages in parallel computing were analyzed. Then the parallel processing technology based on OpenMP was selected to parallel the reactive power optimization process, according to the characteristics of the reactive power optimization algorithm and the parallel computing conditions.In this thesis, the related concepts of parallel computing and it’s applications in the field of power system were summarized briefly. Besides, the parallel processing technology based on OpenMP was introduced in detail. Then the parallel algorithm of reactive power optimization was designed. The ideas of the parallel algorithm may roughly fall into three categories: direct parallelization, "divide and conquer", adjust the method of programming before parallelization. Code segments corresponding to every category were also selected from the reactive power optimization program as examples. Next, the performance of reactive power optimization parallel computing based on OpenMP was analyzed.Finally, two examples were given to verify the effectiveness of the parallel processing technology in improving the speed of reactive power optimization. The parallel program was compiled with Fortran Compiler XE Visual 12.0 compiler in VS2010 operating environment of Windows Parallel Studio XE Intel 2011. The results showed that the parallel processing technology could significantly improve the calculation speed of reactive power optimization as well as get an accurate result at the same time. In addition, the work in this thesis still have a lot of problems and need to be further studied, so some important suggestions were given at the end of the thesis.