3
\$\begingroup\$

I'm modelling the dynamic behavior of a weak power grid. I know the voltage level and short circuit capacity, and that's about it.

  • Short circuit max = 400A, 3-P
  • Voltage level = 66000 V

My problem is as follows:

I want to see the dynamic response if I start large motor, in order to determine if soft start / VSDs. For a normal strong grid, I would use settings such as:

  • Moment of inertia = High (99 s, which is maximum in my simulation tool)
  • Damping factor = Low (0, obviously minimum)
  • Direct-axis transient reactance = High (0.98 pu)
  • Short circuit capacity = High

The problem is, I can't model a weak grid with dynamic data saying it is strong. Not only does it not make sense physically, it also crashes the simulation tool (Paladin DesignBase 5). It works fine if I increase the SC-capacity, or decrease moment of inertia and reactance, but I don't know if it will make sense from a physical point of view.

Does anyone have any experience in modelling weak grids? Any way I can make this more realistic?

\$\endgroup\$
  • \$\begingroup\$ What mechanical data do you have for the motor and its driven load? \$\endgroup\$ – Li-aung Yip Feb 1 '14 at 4:14
  • \$\begingroup\$ None, I'm afraid. And it's not really where the problem is. My simulations fail regardless of load data. It's the representation of the grid that messes everyting up. \$\endgroup\$ – Stewie Griffin Feb 1 '14 at 9:43
  • \$\begingroup\$ I don't have any experience with Paladin DesignBase so I don't know how it models the grid. Personally, I would do some hand calculations as a first start. (Motor starting current for DOL = 6.5×FLC and see what voltage drop that produces at the motor terminals, and at the point of common coupling to the grid.) \$\endgroup\$ – Li-aung Yip Feb 1 '14 at 10:02
  • \$\begingroup\$ Thanks again for answering @Li-aungYip! I'm using 7xFLC (assuming high efficiency motors, thus high starting current). Intuition and simple hand calculations tell me it shouldn't be any problem. Do you have experience with modeling grids in other tools? I assume most of input parameters should be similar (the physics is the same). \$\endgroup\$ – Stewie Griffin Feb 5 '14 at 6:56
  • \$\begingroup\$ See below. Hopefully the below is useful and I haven't gone off on a tangent to what you actually wanted to know. \$\endgroup\$ – Li-aung Yip Feb 5 '14 at 12:23
1
\$\begingroup\$

The power systems analysis software packages I have used allow two ways to model the electricity supply to your network.

  1. As a "utility element" or "grid element" - a voltage source behind a Thevenin equivalent impedance. This is assumed to be a "large" grid with infinite inertia.

  2. As a generator element, with specified electrical parameters like rated kVA, direct/quadrature-axis subtransient, transient and steady-state reactances.

    The impedances are enough for basic load-flow/short circuit calculations. For dynamic/transient studies, the mechanical parameters such as moment of inertia, AVR/exciter control system, governor model, must also be specified.

As you have described it, Paladin DesignBase appears to use some strange hybrid of these approaches. It's not quite an "ideal utility element" (voltage behind a source), nor is it a detailed generator element with full electrical, mechanical, and control modelling.


At a guess, the correct answer to what you actually want to do, "can I start a large motor on this weak grid?", depends on the particulars of your supply.

  • Is the grid "weak" in the sense of being run from a local generator set?

    I.e. the site is being run off a portable generator-set. If this is the case, then the motor starting capability of the generator-set will be published on the alternator's data sheet. Usually this is expressed as a 'Motor starting kVA'. See below example from Stamford (Cummins) PE734C datasheet.

enter image description here

  • Is the grid "weak" in the sense of being connected to the national power grid, but at the end of a very long and skinny transmission line?

    That is, the inertia of the system is practically infinite, but the short-circuit capacity is low due to the high impedance of the feeder line.

    If this is the case, the motor starting time is irrelevant so long as you meet any technical rules imposed on you by the distribution network operator. In Australia this is usually a limit on voltage dip.

In both cases, the motor starting will only succeed if the motor terminal voltage is high enough to produce sufficient starting torque. If the motor terminal voltage is above 80% at motor starting, you are probably OK. If the motor terminal voltage is less than 70% at the max inrush current, there's a good chance your motor will stall. This would warrant closer examination of the motor torque/speed curve, load torque/speed curve, and motor thermal withstand curves.

Remember that torque is proportional to the square of voltage, so 70% voltage = 50% torque, and that excessively long start times may cause overheating of the motor windings.


The correct answer to the question as asked, "how do I model a weak grid?", is to fully model the generator(s). That means you need to model:

  • The electrical characteristics of the alternator.
  • The mechanical properties of the generator set.
  • The alternator AVR/Exciter block diagram and parameters.
  • The engine governor block diagram and parameters.

This information is only published for reasonably large units, i.e. 40MW aeroderivative gas turbine generators. For small portable generator-sets (1MW) you are expected to use the "motor starting kVA" data in the datasheet.

\$\endgroup\$
  • \$\begingroup\$ Comments: 1. It is on the end of a long skinny line connected to the main grid (should have specified it in the question). 2. Paladin offers a large number of possible ways to model the grid, some of which include moment of inertia, advanced AVR/exciter control system, governor models etc. 3. I would like to use the first approach you mention, but I don't know what line data I should use for the Thevenin-equivalent. If I use an equivalent that gives a SC-current at 400A, the voltage will be too low at full load. I could step it up, but then it would be to high at No load. \$\endgroup\$ – Stewie Griffin Feb 5 '14 at 12:50
  • \$\begingroup\$ 4. Thank you very much, again :) \$\endgroup\$ – Stewie Griffin Feb 5 '14 at 12:51
  • \$\begingroup\$ Voltage regulation is an issue on long, skinny lines in real life. That's what tapchanging transformers and other voltage regulators are for. :) \$\endgroup\$ – Li-aung Yip Feb 5 '14 at 12:51
  • \$\begingroup\$ Yes. But if the max SC-current is already reduced at the adjacent substation, but the voltage is (magically) held close to 1 pu, a Thevenin-equivalent reducing the current to 400A would be much too high. Or..? \$\endgroup\$ – Stewie Griffin Feb 5 '14 at 12:54
  • \$\begingroup\$ If you are connected to the main (national) power grid, then moment of inertia, AVR/governor controls, etc are all non-issues due to the practically infinite inertia of the main grid. Model your supply as a Thevenin-equivalent, with the impedance based on the short circuit capacity of 400 A. Solve the voltage regulation problem the same way that it's been solved on site - likely to be transformers with auto-tapchange. Note that the realistic volt drop on a transmission line at full load shouldn't exceed 5%-10% which is well within tapchanger range. \$\endgroup\$ – Li-aung Yip Feb 5 '14 at 12:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.