Brian K. Johnson
Electromagnetic transient (EMT) simulation has moved from a tool used for a few specialist applications, such as insulation coordination, to becoming a common tool for interconnection studies for inverter-based resources (IBRs). Historically, many utilities had few, if any, engineers with a background to perform EMT studies. Today, the proliferation of inverter-based generation and storage resources makes EMT simulation a critical tool for protection and planning studies.
In early power systems, engineers performed transient analysis using hand calculations to solve differential equations. This approach was limited to small-scale problems. Applications included evaluating transient overvoltages due to switching transients such as capacitor switching or opening breakers. Some utilities applied transient network analyzers, which were scale model power systems that represented transmission or distribution lines as a series of modules with shunt capacitors and series resistor/inductor pairs. Adding more model segments improved the transient transmission line model to come closer to the real system. The transient network analyzers took a long time to set up but always ran in real time. Some utilities and equipment vendors used transient network analyzers for hardware-in-the-loop (HIL) testing of protective relays and of controls for high-voltage direct current transmission and static volt-amperes reactive compensators. Transient network analyzers continued to see use into the 1990s but were complicated to set up and had limits to scalability.
Hermann W. Dommel and W. Scott Meyer developed an EMT program for the computer simulation of EMTs at the Bonneville Power Administration in the late 1960s. The underlying algorithm is used in many of the EMT simulation tools in common use today. Initially, engineers simulated switching transients and performed insulation studies. Due to limitations in computer capabilities and difficulties in building and verifying models, EMT simulations were limited to cases with a limited number of nodes. A small number of engineers used transient simulation, and many utilities often had few, if any, engineers familiar with EMT simulation.
Protection engineers were among the earlier adopters of EMT simulation tools, especially for classes of protection studies not amenable to using phasor-based programs. One significant example is protection studies for series compensated lines, where transient low-frequency oscillations in the voltage and current pass through the digital filtering and cause relay overreach. Relay engineers started out using offline EMT studies with simulation results played to relays using amplifiers or directly to relays as low-level analog outputs to test protection settings.
The growing importance of power electronic controls in power systems led to another area where EMT simulation had a growing importance over the years. Simple power electronic switch models and control system models provided the ability to model the Pacific High Voltage dc (HVdc) intertie in the Bonneville Power Administration EMT program. The Manitoba HVdc Research Centre developed an EMT program based on Dommel’s paper to simulate the Nelson River HVdc system. An early EMT program application was to identify the impact of a sensor’s time constant on the control response.
EMT simulation soon moved on to the modeling of other power converter applications, including power quality studies. Researchers started modeling voltage source converters for photovoltaic inverters, flexible ac transmission systems (FACTS) devices, and many other applications. One of the challenges with modeling power converter systems is access to accurate models of the converter controls. Unlike generator exciters and governors, standard models are not available. Accurate modeling of protective relays faces a similar challenge. It is possible to build generic models, but obtaining sufficiently accurate models of a specific vendor’s control is an ongoing challenge.
Real-time EMT simulation grew out of the need to replicate the transient network analyzer capabilities to test the closed response of commercial protection and control devices for power system conditions. These simulators run Dommel’s EMT simulation algorithm in real time with high-speed input and output to external devices, enabling HIL simulation. A key challenge was dividing the system into small parts solvable in parallel without compromising the simulation accuracy. The solution takes advantage of the propagation delay for transients moving over transmission lines to divide the problem into pieces that can be solved in parallel. The approach requires that the propagation delay is longer than the signal delay for exchanging information between processors. There are other options to divide the problem as well. Protection engineers were early adopters of real-time EMT simulations, with commercial relays connected to simulators to test relay responses in an HIL environment. HIL testing is commonly used in the design and testing of controls for HVdc and FACTS systems. Asset owners are increasingly purchasing replica copies of their converter controls to use in simulator studies, both to test changes to controls and to test firmware updates from vendors.
The capabilities of real-time simulators have evolved as processing power has improved. User requirements for real-time simulation have increased as well. The increased use of automation schemes and the advent of digital substations introduce an arena where HIL simulation supports design and testing of automation schemes, protection schemes, and the underlying communication systems. Consulting firms, utilities, and equipment vendors are designing and testing their automation schemes, including remedial action schemes, through HIL simulation. These applications are moving beyond power engineering to support industrial control systems in other industries. HIL simulation is also a growing platform for some areas of cybersecurity research.
The capabilities of conventional offline EMT simulation tools and available models have evolved to meet the demands of users, taking advantage of improvements in computing platforms. Improved graphical interfaces for EMT programs eliminate a significant source of errors in earlier programs. EMT programs can now simulate much larger systems. However, building and validating a model of a truly large system is challenging. Several simulation programs support black box models from equipment vendors, such as controls for HVdc converters and IBRs, allowing engineers to perform studies with more accurate models of the device response.
Modern grids with large-scale installations of IBRs, such as wind turbines, photovoltaic generation, and energy storage systems, are increasing the need for applying EMT simulation in planning studies and in protection studies. The impacts of fast controls on the fault current responses of voltage source converter-based IBRs, HVdc converters, and FACTS devices is difficult to capture in phasor-based fault programs. In addition, aspects of the dynamic responses of these resources are difficult to capture in other standard tools. EMT simulation is a valuable tool for a growing number of types of system studies.
There are challenges with the increased use of EMT simulation. Acquiring adequate model data remains a challenge, and verifying models is an increasingly significant challenge, especially for systems with high levels of IBR generation from different vendors. Utilities build these models through a combination of in-house expertise, external consultants, or equipment suppliers. There is a need to help engineers learn to perform all the steps of the EMT simulation process, requiring knowledge of the applications as well as the specific simulation tools they will use. While user interfaces have made EMT simulation much easier to set up, there is still a need for specialized training.
EMT simulation is now an indispensable tool for power engineers to ensure resilient operation of modern power systems.
Brian K. Johnson (bjohnson@uidaho.edu) is with the University of Idaho, Moscow, ID 83844 USA.
Digital Object Identifier 10.1109/MELE.2023.3320482
2325-5897/23©2023IEEE