Real-time and faster-than-real-time simulation of Modular Multilevel Converters using standard multi-core CPU and FPGA chips

Abstract

Modular Multilevel Converter (MMC) has many advantages for high-voltage direct current (HVDC) transmission and to integrate renewable energy to the main grid. Recently, there are several MMC projects being planning or in development over the world. However, due to the fact that the MMC has a large number of sub-modules (SMs) and a sophisticated controller, there are high demands on simulation tools to validate its design. A digital real-time simulator can simulate the MMC system quickly and accurately, and is able to validate the real MMC controller through hardware-in-the-loop tests. The challenges are to have an accurate and efficient MMC model so the simulator can solve the system with large number of MMC SMs in a very small time step, below 25 μs. In this paper, two MMC models implemented in standard INTEL multi-core CPU and FPGA for a faster-than-real-time and real-time simulation platform are presented. The model performance and accuracy are studied through a back-to-back MMC HVDC system, and compared to a reference model made by SimPowerSystems blocks in Matlab. It is demonstrated that the both models have high fidelity and an MMC system with as high as 3000 SMs can be simulated in real time with a time step of 500 nanoseconds using FPGA chips.