Reactance and Time constant tests for a generator 1

[NickParker]

How important is it to conduct tests for sub transient reactance, transient reactance, and time constant determination?
At what point — in terms of generator capacity and application — are these tests required to be performed by the manufacturer?
The case in point is of around 5 MVA diesel generator

 

[mparenteau]

That information certainly helps to model your capability curve.

Mike

 

[mfgenggear]

That information certainly helps to model your capability curve.

Mike

How would that be setup?
Program?

 

[mparenteau]

How would that be setup?
Program?

Programs like SKM, ETAP, etc.

 

[Gr8blu]

Most OEMs who manufacture "standard" ratings will have a very capable ability to calculate the expected synchronous, transient, and subtransient reactances for their generator internal geometries and "normal" ratings.

Proof of performance is accomplished during the factory acceptance testing: the specific method(s) used don't really vary all that much from one OEM to another. In a nutshell, the machine must undergo an open-circuit test, a short-circuit test, and a zero power factor test. From this data, reactances are determined ... at least for the UNSATURATED condition. Saturated reactances are another matter entirely, with most OEMs using a "rule of thumb" based on their experience with their designs. Does it get them sufficiently close? Yes.

The specifics of how and when to test are included in IEEE 115 "IEEE Guide for Test Procedures for Synchronous Machines - Parts I and II". There is a similar standard used in the IEC world - I just cannot recall what it is at the moment.

 

[NickParker]

Most OEMs who manufacture "standard" ratings will have a very capable ability to calculate the expected synchronous, transient, and subtransient reactances for their generator internal geometries and "normal" ratings.

Proof of performance is accomplished during the factory acceptance testing: the specific method(s) used don't really vary all that much from one OEM to another. In a nutshell, the machine must undergo an open-circuit test, a short-circuit test, and a zero power factor test. From this data, reactances are determined ... at least for the UNSATURATED condition. Saturated reactances are another matter entirely, with most OEMs using a "rule of thumb" based on their experience with their designs. Does it get them sufficiently close? Yes.

The specifics of how and when to test are included in IEEE 115 "IEEE Guide for Test Procedures for Synchronous Machines - Parts I and II". There is a similar standard used in the IEC world - I just cannot recall what it is at the moment.

one of the manufacturers tell me that, they only intend to perform tests for direct-axis reactances, but not for quadrature-axis reactances, zero-sequence and negative-sequence reactances, open circuit time constants. What could be the reason for this?

 

[Gr8blu]

The reason for "direct testing" of direct-axis reactances is that it IS pretty much a direct measurement from the data that can be physically tested for.

There is also some fairly well-known relationships between the direct axis and quadrature axis values, as there is between positive (direct axis) and zero- and negative-sequence values. The relationship may not be identical to the nth decimal point from all manufacturers, but WITHIN a manufacturer's capability, the correlation is very strong (due to time-tested empirical factors that describe their particular relationships between known and calculated values).

The time constants are calculated by knowing the various reactances and the specific conditions (e.g. open circuit, short circuit, etc.).

As the purchaser, you CAN specify that the manufacturer provides guaranteed (and/or as tested) values for performance parameters that are important to you - even if the manufacturer doesn't agree that they are important. This includes electrical characteristics such as reactances and time constants, as well as efficiencies ... and mechanical testing (like short circuit withstand), etc. Just remember that the more you ask for, the higher the price (to you) and the longer the time from order to delivery.

 

[wcaseyharman]

In larger machines originally WECC, and now NERC, require validation of those generator parameters to ensure accurate grid stability models. System protection also uses them to provide accurate fault currents in system fault programs such as ASPEN Oneliner.
However, NERC registration starts at 20 MVA. So the accuracy of your generator data is really how accurate do you need it for any necessary studies for whatever you are doing - arc flash, short circuit, etc.
For fault analysis, typically only the D axis data is used - the q axis only begins to play once load is on the machine, and during faults the assumption is the fault drags the voltage lower and the fault current dominates the current in the gen, and fault current in a generator is mostly reactive, so the q axis isn’t really needed. Most fault programs only take d axis data for this reason.