Large-Scale Real-Time Electromagnetic Transient Simulation of Power Systems Using Hardware Emulation on FPGAs

Real-time electromagnetic transient (EMT) simulation plays an important role in the planning, design, and operation of modern power transmission systems with adequate security and reliability due to increased load growth, interconnectivity, and stressful operating conditions. Real-time EMT simulators are widely employed for such applications as testing of advanced protective schemes for lines and generators, testing closed loop control systems either for conventional power systems or for power electronic based applications such as HVDC and FACTS, and for the training of system operators under realistic scenarios.;Real-time EMT simulation of large transmission networks requires very high computational capability. To meet the stringent real-time step-size constraints, a compromise is usually made between the size of the system simulated and the complexity of the component models. Taking advantage of the inherent parallel architecture, high density, and high clock speed, field programmable gate array (FPGA) has gained increasing popularity in high performance computation for various computationally expensive applications.;This thesis describes how FPGAs can be used for realizing real-time electromagnetic transient simulation of large-scale power systems using digital hardware emulation. Detailed parallel hardware modules for various power system components are described, including linear lumped RLCG elements, supply sources, circuit breakers. Hardware modules for transmission lines include traveling wave (Bergeron) model, frequency-dependent line model (FDLM), and universal line model (ULM). Various rotating electric machines are modeled using universal machine (UM) model. The network solution exploits sparse matrix techniques for improved efficiency. A novel parallelled EMT solution algorithm is described that accommodates the parallel FPGA architecture. For inclusion of nonlinear elements in power system, a parallel iterative nonlinear network solver is described that uses Newton-Raphson method both continuous and piecewise. Multiple FPGAs are utilized for real-time EMT emulation of large-scale power systems. A novel functional decomposition method is introduced to allocate the model components to the available hardware emulation modules in the FPGAs. All hardware arithmetic units designed are deeply pipelined to achieve highest computation throughput. 32-bit floating-point number representation is used for high accuracy throughout the EMT simulation. The whole design is based on VHDL for portability and extensibility.;Various power system case studies are used to validate the proposed FPGA-based real-time EMT simulator. The captured real-time oscilloscope results demonstrate excellent accuracy and small simulation time-step of the simulator in comparison to the off-line simulation of the original systems in the ATP or EMTP-RV ® off-line EMT programs.