Exploring parallel processing for wide area measurement data applications

In this paper the potential of parallelizing real-time monitoring applications using mainstream multi-core processors is explored. In particular, the focus is on implementing and assessing the oscillatory monitoring application. The objective is to evaluate the practicality of using parallel processing to achieve faster computational speed. In this paper, two monitoring techniques with different mathematical formulations are used. They are the block-processing Prony Analysis and recursive Extended Complex Kalman Filter. Both techniques are implemented to demonstrate different aspects of parallel processing. Prony Analysis is used to show the potential of operating multiple algorithms concurrently while ECKF demonstrates the possibility of parallelizing the decomposed subroutines. Comparisons with the corresponding sequential algorithms are carried out using synthetic signals in Visual C++. Simulation results show that parallel processing is able to reduce the computation time.