The difference in EMT vs RMS simulation comes down to what is being solved. RMS simulation represents voltages and currents as 50 Hz phasors, which makes it fast enough to study the whole NEM. EMT simulation solves the actual waveforms at microsecond resolution, which is slow but captures fast inverter control. As the grid has filled with inverters and weakened in places, AEMO has come to need both.
Methods, Not Tools
PSS/E is an RMS (phasor-domain) tool and PSCAD is an EMT (electromagnetic transient) tool, but the physics behind AEMO's requirement is the same whichever software is used. For the tools themselves, and how AEMO accepts the models, see our PSS/E and PSCAD modelling overview. Here the question is narrower: what does each method calculate, what does it miss, and when does the RMS answer stop being trustworthy for inverter-based plant?
What RMS Simulation Captures and Misses
How it works
An RMS simulation assumes the network runs at, or very close to, 50 Hz. Every voltage and current is a phasor: a magnitude and an angle. The network is solved algebraically at each step, and only the slower dynamics (rotors, governors, exciters, plant controllers, the outer loops of inverters) are modelled as differential equations.
Ignoring the fast electromagnetic behaviour of lines and transformers allows a time step of a few milliseconds, which is what makes it practical to simulate thousands of buses for tens of seconds. Most connection-study RMS work is also positive-sequence: the three phases are assumed balanced, and unbalanced faults are approximated.
What it does well
- Wide-area stability: rotor angle, voltage and frequency stability across a region or the whole NEM
- Oscillations in the low-frequency range, roughly 0.1 to 2 Hz, between generators and areas
- Plant-level responses on the timescale of the performance standards: reactive power and voltage control, active power recovery after a fault, frequency response
- Large numbers of contingencies and dispatch scenarios, run quickly
It is also the method behind the steady-state work. Load flow, voltage and fault level studies, covered in our guide to steady-state vs dynamic studies, give the operating point that every dynamic simulation starts from.
What it misses
The phasor assumption breaks down exactly where modern plant is most active:
- Sub-cycle behaviour. Anything faster than about one cycle (20 ms) is averaged out. The first few milliseconds of an inverter's fault response are not represented.
- Fast inverter controls. Phase-locked loops, inner current control loops and fast protection act in milliseconds or less. RMS models represent them with simplified or idealised blocks.
- Unbalance and distortion. Phase-by-phase behaviour during unbalanced faults, harmonics and DC offsets are outside the method.
- Control interactions. Oscillations above a few hertz, caused by inverter controls interacting with each other or with the network impedance, can be invisible or misrepresented.
None of this mattered much when large synchronous machines held the waveform close to an ideal sine wave. It matters a great deal now.
What EMT Simulation Captures
How it works
An EMT simulation solves the instantaneous voltage and current on each phase. Lines, cables and transformers carry their inductance and capacitance, so resonances and saturation appear naturally. Inverters run their actual control code, usually a compiled or encrypted manufacturer model. Time steps of roughly 1 to 50 microseconds, about a thousand times finer than RMS, are why EMT is slow and why the surrounding network is usually reduced.
Fast controls
The plant's response in the first one or two cycles after a fault is decided by the inverter's inner controls and protection. EMT shows whether the phase-locked loop stays locked when the voltage collapses and jumps in angle, whether the current controller produces the reactive current the performance standards require, and whether any protection function trips the plant when it should ride through.
Weak grids
This is the main reason EMT has become routine in Australia. Where system strength is low (short-circuit ratio is the usual first indicator), the connection point voltage moves with the inverter's own current. Controls tuned for a stiff grid can then interact with the network and neighbouring inverters, producing oscillations or loss of synchronism that an RMS model often will not predict. The system strength framework requires connecting inverter-based plant to have its system strength impact assessed, then either pay the system strength charge or self-remediate, and that assessment relies heavily on EMT. Our guide to the system strength impact assessment explains the study.
Harmonic and resonance interaction
Inverter controls have a frequency-dependent impedance. Combined with cable capacitance, filters and network impedance, it can create resonances from hundreds of hertz into the low kilohertz range. EMT can show control-driven instability that a harmonic study treating the inverter as a fixed current source may miss.
It also solves each phase separately, so single-phase and two-phase faults, far more common than balanced three-phase faults, are represented directly.
EMT vs RMS Compared
| Attribute | RMS (phasor) simulation | EMT simulation |
|---|---|---|
| Typical tool in AEMO connections | PSS/E | PSCAD |
| Quantity solved | 50 Hz phasors, usually positive-sequence | Instantaneous waveforms on each phase |
| Typical time step | A few milliseconds | About 1 to 50 microseconds |
| Network representation | Large: region or whole NEM | Smaller area around the plant, wider-area where required |
| Inverter model | Generic or simplified vendor model of the outer controls | Vendor control code, including inner loops and protection |
| Run time per case | Seconds to minutes | Minutes to hours, depending on network size |
| Good for | Wide-area stability, voltage and frequency control, many contingencies | Fault ride-through detail, weak grid stability, control interaction, unbalanced faults |
| Blind spots | Sub-cycle response, PLL behaviour, fast oscillations, unbalance | Impractical for very large networks or very long simulations |
| Typical studies | Load flow based dynamics, transient stability, performance standard compliance | System strength assessment, fault ride-through cases, model acceptance tests, oscillation investigation |
The two are complementary. RMS shows how the plant behaves within the whole system. EMT shows whether it will actually do what its RMS model claims, where that is in doubt.
When AEMO and NSPs Require EMT
For every registered generator
AEMO's model requirements call for both PSS/E and PSCAD models of a registered generating system. Both go through acceptance testing, and the two must benchmark against each other: for the same disturbance, key responses should agree within tolerance. Benchmarking is a common source of rework, because it exposes simplifications in the RMS model and initialisation problems in the EMT model.
The Generator Performance Standards are then demonstrated across both. Fault ride-through and the multiple fault requirements lean heavily on EMT. Voltage control, active power control and wide-area interactions are typically demonstrated in RMS, with EMT confirming where the network is weak.
Where system strength is low
In weak parts of the network, studies shift further towards EMT. The NSP and AEMO will usually want the plant studied with nearby inverter-based generators, because the risk is interaction between plants. That means more models, larger EMT networks and longer run times.
Wide-area EMT
Single-plant EMT studies, with the network reduced a few buses away, cannot show interactions across a region. AEMO and some NSPs now run EMT over much larger portions of the network where inverter-based plant is concentrated. Your PSCAD model then becomes part of a shared network model, so its quality matters beyond your own study.
At DNSP level
Smaller embedded generators on a distribution network usually need only steady-state and RMS studies. EMT appears when the plant is large relative to the local fault level, sits near other inverter-based plant, or will register with AEMO. Ask the DNSP early: adding EMT late is a common cause of slippage, alongside the others in our article on GPS application pitfalls.
What this means for the program
The EMT package, not the RMS one, usually sets the critical path of the dynamic studies. It depends on when the OEM's PSCAD model for the procured firmware is available, how many cases are needed, how large the EMT network must be, and how many rounds it takes to settle benchmarking differences. A PSCAD model that initialises cleanly and matches its PSS/E counterpart early saves more time than any other single step.
Get EMT and RMS Studies Done Right
ElectraGlobe's power system studies team runs both sides: PSS/E load flow and RMS dynamics, and PSCAD EMT studies for fault ride-through, weak grid performance and model acceptance. We benchmark the two models before they reach AEMO, where most avoidable review rounds are lost.
If you have a project heading into connection studies, or an EMT result you are not sure about, contact our team and we can tell you where the risk sits.
Frequently Asked Questions
What is the difference between EMT and RMS simulation?
RMS simulation represents voltages and currents as 50 Hz phasors and uses time steps of a few milliseconds, so it can study large networks quickly. EMT simulation solves the instantaneous waveform on each phase with microsecond time steps, capturing fast inverter controls, unbalance and resonance. RMS is used for wide-area stability, while EMT is used where fast behaviour decides the result.
Why does AEMO need PSCAD models?
Inverter-based plant behaves in ways that phasor models cannot fully represent, particularly in the first cycles after a fault and in weak parts of the network. A PSCAD model runs the manufacturer's actual control code, so AEMO and the NSP can check fault ride-through, control stability and interactions with nearby plant. It also lets them verify that the simpler PSS/E model is a fair representation.
Is PSS/E an RMS tool?
Yes. In Australian connection work PSS/E is used for load flow and positive-sequence phasor (RMS) dynamic simulation. That makes it fast enough for wide-area NEM studies, but it does not resolve sub-cycle waveforms, which is why AEMO pairs it with PSCAD.
How long do EMT studies take compared with RMS?
A single RMS case on a large network typically runs in seconds to minutes. A single EMT case can take minutes to hours, depending on how much of the network is modelled and how long the event is. Across a full package with hundreds of cases, EMT usually takes considerably longer to run, check and report, and it often sets the critical path for the dynamic studies.