You will all be thankful to know that I have finally reached my last blog in this road test review of the Megger Previous blogs have covered bench reviews and tests a teardown and insulation testing from electrical panels on motors and air circuit breakers In this blog I will cover the testing of generator rotor windings from a 92MVA and a 145MVA turbogenerator.
This is the main function of my job and the reason why the road test was of great interest to me. I currently use a Fluke 1555C insulation tester, with a 250V to 10,000V output capability to carry out rotor and stator winding tests on turbogenerators. However, this instrument lacks a 100V output sometimes specified for insulated bearing tests and struggles with voltage compliance at the lower output levels. It also comes in a large padded case and takes up a quarter of the car boot space when travelling. Due to logistics and costs of testing generator stators,I test more rotors and taking the Fluke 1555C around with me is a bit of a burden.
Whilst the 1555C has some download capabilities it does not download a complete polarisation index test and I wanted to improve my reports so i opted for a Keysight that on paper seemed to meet most of my needs and would also provide me with meter functionality for RSO and AC Impedance tests Unfortunately the does not appear to have the capacity to test generator rotor windings so has been consigned to use as a multimeter most of the time.
This brings me to the road test of the Megger It is a smaller more managable unit than the 1555C and has a low enough output voltage for my requirements It does have some multimeter functionality that looks promising enough to meet my needs I am aware that this particular version does not have a results download capability but the next model up in the range the does have this capability So this was going to be a learning curve to see if further investment would be beneficial.
This will be the most onerous test conditions for the that I can offer and will test multiple aspects of its functionality
1) The variable insulation test range will be utilised to prevent over stressing of the rotor winding
2) The resistance range will be utilised to verify test connections during recurrent surge oscillograph tests
3) The voltmeter function will be used to measure AC volts during AC Impedance testing
4) The voltmeter function will be used to measure DC volts during winding resistance tests
Both the generator rotors tested are of a brushless design, with a rotating diode pack that must be disconnected to allow the insulation tests to be carried out. The rotor winding is, therefore, a DC rated component and considering that it would cost around £600,000 and take 3 months to rewind a rotor, care needs to be taken when applying test voltages to them, especially as the units age and become more contaminated with dirt.
Rotor Winding Insulation Test
For the 145MVA unit the rotor winding is designed to work at a maximum of 360V DC under short circuit conditions of the stator output. The has nominal test voltages of 250V and 500V. If I use 250V then I am not testing the rotor under its worst operating conditions. If I use the 500V range then potentially I am stressing an old winding beyond any voltage level it is likely to see in service. The 92MVA unit has a maximum rotor voltage of 143V. The offers nominal ranges of 100V and 250V around this value, with a similar scenario to the larger rotor.
The answer is to use the variable voltage function of the to set the output voltage to match the maximum in service conditions the rotor will see. This is the safest way I can test an old rotor winding, testing it to its full service voltage without over stressing it and risking damage. The variable voltage function defaults to 10V and is easily set to the required value using the settings function this also prevents inadvertent adjustment of the test voltage during use. Once set the range functions exactly like all the other insulation ranges with timed DAR and PI test setups and the ability to save any results obtained.
Throughout the tests on both rotor windings the performed without any issues, connections were easily made and the tests conducted. Unfortunately neither rotor was in too good a condition. The 92MVA rotor was tested on day 1 and had a low insulation value with no DAR. Overnight the heaters were turned on and the unit retested again in the morning. A doubling of the insulation resistance was seen but no DAR was obtained. The 145MVA generator was much worse and the insulation tester output voltage could not be maintained on the first test. No heating was available for this unit, so only a marginal improvement was seen the next day probably due to the lower ambient humidity. I hope to return to this rotor in the next two to three weeks after the heaters have been put back into service.
Rotor Winding DC Resistance
I do have a micro ohmmeter available to me to test winding resistance at 10A. The instrument however is not the best and will cut-out due to the time taken to allow the reading to stabilise due to the inductive nature to the winding. As an alternative, I have constructed a 100A variable DC supply and utilise a 4 wire methodology to measure the winding resistance at 5 different current values and then take an average as per the IEEE118 standard. The supply contains a 100A to 50mV current shunt that I connect a mV meter to, I then utilise the as a DC voltmeter connected directly across the rotor winding
Both rotors tested reasonably well. For comparison on the 92MVA rotor, I repeated the test with the standard multimeter usually used in place of the .
The produced a test result to -0.73% of the nominal DC winding resistance corrected to the specified temperature where as the usual test setup produced a result to -0.10% The specification is to be within 2 of the corrected value and so whilst the was not quite as accurate as the usual setup it still returned a result within tolerance
Recurrent Surge Oscillograph Test (RSO)
The RSO test is a specialist test on a generator rotor winding to detect inter-turn faults within the winding. It works on time domain reflectometry. Pulses are injected alternately from the two ends of the windings and the reflected waveforms across impedances connected to the windings are compared. If the two traces overlap, then no fault is present. A fault is seen as a deviation in one of the waveform traces.
Details of rotor reflectometers and the test can be obtained from Rowdtest Ltd who manufacturer the units and provide the diagram above of the principle of operation.
For this test I utilised the to measure the resistance of the test connections to the winding and to earth For an ideal RSO test there should be a maximum connection resistance of 1 Ohm across the winding and to earth This would usually be measured using the ohms function of a standard multimeter but I was curious to see if the automatic function of the resistance range on the would operate correctly across the inductive load of a large rotor winding As seen in the photo below the instrument function perfectly well recording a 0.26 Ohm resistance reading through the winding whilst sourcing a 200mA current.
Rotor Winding AC Impedance Test
The final test on the rotor winding is an AC Impedance test. This is an alternative test methodology to the RSO test for detecting inter-turn insulation faults. Where as the RSO test only applies 12V, in an AC impedance test, a voltage of up to 110V AC is applied across the winding, therefore applying more voltage across the inter-turn insulation. The test is carried out in the same manner as the DC winding resistance test, with the voltage across the winding and the current through it are measured at 10V steps and then the impedance calculated afterwards.
A good rotor winding will produce a plot of impedance against voltage with a slight upward trend. An inter-turn fault in a winding is seen as a step change down in the impedance along the plot. There is also the opportunity to carry out a long term historical impedance trend of the winding at a specific voltage, nominated as 50V during my tests.
As can be seen from the plots above for the 92MVA generator comparable test results were achieved using the in place of a standard multimeter. Again the instrument has proved that it can be utilised in place of a regular voltmeter for the kind of tests that I carry out increasing its value and flexibility.