Analysis On Recorded Data Processing And GPU High-performance Of Distribution Network In Chemical Park

Petrochemical industry is an important foundation of China’s economy.The production process 、 related equipment and related production technology of chemical enterprises have high requirements for power system.In recent years,with the large-scale power equipment used by chemical enterprises and the incorporation of various new energy generation into the power grid,the problems of deterioration,quality degradation and poor reliability of the power grid in chemical parks have become increasingly serious.In order to better guarantee the power safety of the chemical park power network and reduce the occurrence of power network faults,it is particularly important to strengthen the processing and analysis capabilities of the power network recording data,especially the fault recording data.In this paper,some methods are put forward for the data processing and analysis of wave recordings in Chemical Park power network.In the data processing of power network recordings,the standard of COMTRADE is used as the data format of unified power network recordings.Through the study of a large number of field and analog data,a new method is presented.The fault-tolerant processing and reading method of COMTRADE format files has been tested.The results show that this method can effectively improve the reading rate of various types of COMTRADE format files and improve the universality of power network recordings.In order to reduce the cost of storage and transmission of power network recorder data and improve the transmission efficiency of transmission power network recorder data,a compression and decompression algorithm for the recorder data is presented according to the characteristics of COMTRADE format file.This algorithm is fast and still has a good compression ratio when the memory of embedded power network recorder data collection terminal is small.In the analysis of power network recordings data,based on CUDA programming,the harmonic analysis of power network data is studied using GPU accelerated FFTand wavelet transformation methods,which effectively improves the analysis speed of power network recordings data.