Real-time Nonlinear Frequency-dependent Electromagnetic Transient Power Transformer Model Hardware Emulation

Résumé du livre

A transformer is a vital piece of transmission and distribution infrastructure in a power system which is used to transfer energy from one circuit to another. Real-time electromagnetic transient (EMT) simulation plays an important role in the design and testing of protection and control systems before they are commissioned in the field. An accurate power transformer model is paramount for accurately representing the host power system in real-time EMT simulation. However, due to excessive computational burden, transformer modeling has hitherto been limited to low-order, piece-wise linear, and low-frequency behavior. On one hand the real-time EMT simulation needs to finish all the calculations in a certain small time-step, and on the another hand the components need to be modeled with adequate complexity to reproduce highly accurate simulation results. The field programmable gate arrays (FPGA) provides a hardware-rich environment to implement power system component models in real-time with its paralleled architecture and reconfigurable logic resource. This thesis develops detailed nonlinear power transformer models in real-time which can accurately emulate realistic transformer behaviour under various operating conditions, such as saturation, ferroresonance, hysteresis, and frequency-dependent eddy-currents. To investigate the detailed flux distribution in the transformer core, a high resolution magnetic equivalent circuit is proposed. The models are fully parallelized and pipelined to achieve the lowest latency and smallest hardware resource consumption. Several power system test case studies are used to study the proposed FPGA-based real-time transformer models under various transient and power flow control conditions. The real-time results are fully validated using off-line EMT and finite-element modeling tools.