Analyzing the Implementation of the Newton Raphson Based Power Flow Formulation in CPU+GPU Computing Environment

Power flow analysis is a key to operate and design an electric power system. However, analyzing a complex electrical power system is a computationally time-intensive activity. This paper assesses the implementation of Newton Raphson based power flow formulation in the heterogeneous computing environment (GPU+CPU) for speeding up the computation process. To this end, the computationally intensive Newton Raphson algorithm involved in power flow calculations is discussed along with the modalities in their implementation on sequential as well as parallel processing environments. The performance comparison of computation using sequential and parallel processing is then presented with the purpose to identify the bottleneck in speeding up calculation using GPU. The results show that solving the system of linear equations becomes the main bottleneck for speeding up the Newton Raphson based Power flow method in CPU+GPU environment.