In an electromagnetic transients (EMT) study, each element is modeled with the level of detail the phenomenon demands. The transformer needs a saturation curve, remanence and core topology; circuit breakers need their operating times; inverters need the manufacturer's EMT model. An incomplete model gives a result that does not resemble reality.
What an EMT model is
An electromagnetic transients study solves the network in the time domain, with time steps on the order of microseconds. It uses detailed differential equations of the elements. With them it represents fast transients, sub-cycle oscillations, switching events and the interaction between equipment, controllers and the grid. How to specify and review a software-based study in general is covered in Power system studies with software: what to specify; system stability on a time scale of seconds is explained in Transient stability: what it is and what decides it.
That definition has a practical consequence: the result is valid only if each element is modeled with the detail the phenomenon requires. A model that is good for calculating a fault at power frequency is not good for energizing a transformer. That is why the first step of the study is to decide which phenomenon is simulated and, with that, which model each element gets.
| Element | What the model must represent | Data requested |
|---|---|---|
| Equivalent grid | The impedance seen from the point of connection | Short-circuit power and X/R ratio |
| Transformer | Saturation, remanence, connection and capacitances | Saturation curve, no-load current, short-circuit impedance |
| Line or cable | Length, geometry and capacitance | Conductor geometry or manufacturer parameters |
| Circuit breaker | Closing time, pole scatter, pre-insertion | Nameplate and manufacturer test data |
| Surge arrester | Voltage-current characteristic | Manufacturer curve |
| Inverter | Fast controls and current limits | Manufacturer EMT model |
The transformer: the element that weighs most
Inrush current depends on the transformer model more than on any other data. IEC 60076-8 (clause 4.10) explains the phenomenon: when energized at full voltage, the transient flux can reach the saturation limit of the steel and exceed it before the voltage changes sign. The amplitude of the magnetizing current can reach any peak value up to a maximum that exceeds the rated current and approaches the through-fault current of the transformer. It depends on the direction of the residual flux and on the point on the wave at which the voltage is applied.
From that description come the data the model needs:
- Saturation curve. Without it, the model cannot saturate and produces no inrush current.
- Residual flux. The standard notes that the extreme value of the current depends on the design flux density and on the winding arrangement: an outer winding has higher air-core inductance and draws less current from the grid. It also warns that measuring winding resistance with direct current can leave remanence in the core. If DC tests are done before energizing, the remanence assumption of the study must take that into account.
- Point on wave. Inrush current is random and appears at full magnitude only in some operations; that is why several closing instants are simulated.
- DC component. When the current is high, it is asymmetrical and has a DC component that decays in less than a second, more slowly in large transformers and in those with low-loss steel. That component and the harmonic content matter for protective relays.
Core topology. IEC 60076-8 distinguishes the three-limb core from the five-limb core, the shell-type core and banks of single-phase transformers. In the three-limb core, zero-sequence flux must close outside the excited winding, through a high reluctance. In the others, there are return paths through unwound limbs and the zero-sequence magnetizing impedance is very high, until those parts saturate. That difference changes the voltages in ground faults and in single-pole operations. The same text points out that the zero-sequence impedance depends on the winding connection.
Test data. IEC 60076-1 (clause 10) lists the tests that produce the data: no-load loss and current (10.5), harmonics of the no-load current (10.6), short-circuit impedance (10.4) and zero-sequence impedances (10.7). Among the special tests (10.1.3) are the determination of capacitances between windings and to earth and the transferred surge voltage characteristic. These last two are done only if the contract asks for them, so request them before the transformer is manufactured when the EMT study requires them.
Lines and cables
A line or a cable is not a single impedance. In a switching operation, what matters is its behavior at the frequencies of the transient: capacitance, length and conductor geometry. A long cable, for example, has a lot of capacitance and can resonate with the inductance of a transformer. That is why the model starts from the real geometry and the type of screen, not from the power-frequency impedance.
Circuit breakers and surge arresters
The circuit breaker is modeled with its closing time and the scatter between poles. IEC 62271-100 (IEEE C37.04) includes among the rated times the opening time, the closing time, the reclosing time and the pre-insertion time. If the breaker has pre-insertion resistors or controlled switching, those data go into the model, because they change the closing instant of each phase. The effect of the closing instant on inrush current is explained in its guide.
The surge arrester is modeled with the voltage-current characteristic supplied by its manufacturer; without it you cannot check whether the simulated overvoltage ends up limited.
Generators and inverters
A synchronous generator is modeled with its reactances and time constants, and with the voltage regulator if the study includes oscillations. For inverters, the fast controls weigh as much as the power electronics: that is why the model must be the manufacturer’s and not a generic one. Grid codes commonly require these models for solar and wind plants connected to the transmission system: a preliminary version well before connection, for the initial tests, and detailed models before operation, together with factory tests that compare the model with the real performance of the equipment. After commercial operation, the model must be validated and updated. The deadlines and the procedure are set by the utility or grid operator and are agreed in each project.
How an EMT model is validated
An EMT model is validated with real records. The simulated waveform of a switching operation or a fault is compared with the one recorded by the relay or the disturbance recorder. If they match in magnitude, frequency and damping, the model is reliable for extrapolating to other cases. To review a record you can use our COMTRADE viewer.
Common mistakes
- Using the short-circuit model without saturation. Inrush current does not appear.
- Assuming zero residual flux. It is the least demanding case, not the most likely.
- Simulating a single closing instant. The worst case is not found.
- Using a generic inverter model when the manufacturer’s exists.
- Not saving the model version. Every change in the equipment requires updating the study.
We perform your transient study
We build the model of your network from the nameplate and test data of your equipment, validate it with records and deliver the results with switching recommendations. See our studies for commissioning or write to us through contact. Studies are done remotely, worldwide.
Related guides
- Transformer trips on energization
- Transformer inrush current
- Ferroresonance in voltage transformers
- Load flow study
- Power system studies with software: what to specify
- Transient stability: what it is and what decides it
References
- IEC 60076-8, Power transformers, application guide: clauses 2.2, 2.3 and 4.10.
- IEC 60076-1, Power transformers, general: clauses 3.7.3, 10.1.3, 10.4 to 10.7.
- IEC 62271-100, High-voltage alternating-current circuit-breakers: rated times.
Frequently asked questions
Why is the short-circuit study model not enough?
Because the short-circuit model represents the elements by their impedance at power frequency. A transient depends on core saturation, capacitances and breaker operating times, data that this model does not include.
Which transformer data is usually left out?
The saturation curve and the residual flux assumption. Without them, the inrush current and ferroresonance cases are underestimated or do not appear.
Who provides the EMT model of a solar or wind plant?
The party that represents the plant. Grid codes for transmission-connected inverter-based plants usually require both RMS and EMT models, delivered well before commissioning, and then validated against measurements once the plant operates. The exact deadlines are set by the utility or grid operator.