As part of the repair and restoration theme for project14, I am taking a look at an obsolete UBX117 protection relay, that was removed during routine testing of a generator protection panel, due to faulty operation. I will go through the process of determining the schematic of the board, carrying out the repairs and demonstrating the final test of the repaired relay.
Stator Earth Fault Scheme
A generator stator winding will have a number of protection schemes to prevent or limit damage to the winding during operation. One such scheme is earth fault protection for detecting a fault current flowing through the stator insulation into the generator frame. There are a number of ways for this protection to be configured, with one method depicted below that utilises and earthing transformer at the star point of the stator winding that feeds to a voltage driven protection relay. The objective of the earthing transformer is to provide an impedance for an earth return fault current and limit it to some extent.
The manufacturer of this panel has labelled the relay as '64ST' which is unusual and does not really match the ANSI standard. It may be because the panel was British built and the overall scheme has the generator hard wired to the generator step-up transformer with no synchronising breaker between them. The usual ANSI label for this relay would be 64S.
In this kind of scheme, the relay is set to 5% of the maximum earth fault current at the rated generator voltage, thus it will provide protection for 95% of the stator winding. The 5% of the stator winding down to the neutral either remains unprotected or special relays measuring the third harmonic voltage generated across the earthing impedance due to the flow of third harmonics through the shunt capacitance of the stator winding can be employed. The decision to leave the 5% unprotected is a risk assessment as the voltage to earth of this 5% of the winding is very low and unlikely to lead to an earth fault condition. However, if it does, the damage can be significant as a lot of burning is generated with faults in this region.
UBX117 Protection Relay
The UBX117 is an old voltage driven protection relay from BBC, the company that merged with ASEA in 1988 to form ABB, that I found to be faulty during the testing of a generator protection panel.
The UBX117 only has one setting for the voltage trip level which is expressed in terms of %UN. For this particular relay, UN is set to 100 Volts, therefore the trip level can be set between 4 and 12 Volts.
An LED gives a visual indication if the relay has operated. The relay is manually reset using one of the pushbuttons on the front panel and the final pushbutton allows for a manual test of the tripping delay, which is nominally 500ms for this relay, and is not adjustable.
These relays are now obsolete, so the option is to try and find a second hand relay or repair the faulty unit. As these are relatively simple relays, the repair option seems to be a good approach.
Initial Test Results of Relay in Panel
When initially inspected and tested, the UBX117 relay was found to be set at 9 Volts, but the final settings documentation for the protection panel specified a setting of 10 Volts. The initial test showed the relay was operating correctly for its particular setting as shown in this extract of the protocol sheet.
Whilst the relay was functioning correctly, its setting was wrong and therefore the site engineer requested that the relay be rest to the correct setting to match the documentation. The relay was turned up to the 10 Volt setting and then retested to verify the new tripping level.
With the relay set to match the documentation, the trip level was determined as 9.9 Volts, but the trip time had dropped to 212 milliseconds.
The trip level was just out of tolerance, but for such an old relay, it would be deemed as acceptable. The trip time at more than 50% from the expected value, was not acceptable. A couple more tests were made at the new setting to verify the bad reading. The trip level was consistent, the trip time had an upward trend but was still significantly less than the expected value.
A relay from a decommissioned protection panel was used as a replacement but produced high readings for both the pickup voltage and the trip delay.
This relay is totally unacceptable for service.
Luckily a new protection relay was procured from storage at another site, that was tested and left in service to allow the site to operate. They only have the one spare relay though, so the repair of either of the other relays becomes more desirable, as to upgrade to a new relay would require extensive work and costs.
The test results for the new relay provides a baseline to repair the original UBX117 relays to. A new relay is tested through a range of settings before being set to the required setting.
As can be seen from the test results above, the new relay has reasonable pickup voltage tolerances, but the trip time results do have quite a wide range to them. This may just be a characteristic of the relay and the way it operates. It is still more accurate than the relay it replaced.
The following is a video on the bench tests of the second relay.
As can be seen in the video and the picture below, the test setup is not fantastic with crocodile clips being used to make the connections. This is something that I will aim to improve as the repair of the relays is progressed.
The investigation into the timing fault with the relay begins with a visual inspection of the components and board, looking for physical damage, bad solder joints, and overheating of components. Nothing was really evident in this inspection, although it did highlight a number of reasonably sized polyester film and tantalum capacitors. Given that this relay is in the region of 20 years old, it may be that some of these have drifted out of tolerance and may be the cause.
A couple of the polyester film capacitors were found installed underneath the edge of the on board transformer, there is a chance that these may have been under more thermal stress than the other capacitors. Other than this observation, there were no problems found with the components.
The track side of the PCB was also found to be in good condition, with no signs of damage or overheating. All of the solder joints were good.
Since there is nothing obviously wrong with the relay that would cause the timing fault, I will need to move onto tracing out the circuit and identifying the components. This will form the basis for the next blog.