The Buchholz relays I am used to dealing with have two levels, a Gas Alarm (Volume) and Gas Trip (Surge). They have a nitrogen injection port on the tank side of the Buchholz to inject to test each of these contacts, and a bleed port from the top of the Buchholz which allows for the nitrogen or fault-gasses to be captured for laboratory analysis.
Obviously an open flame is a low-tech way of determining if there are short-chain hydrocarbons present in the gas in the Buchholz, but if it works it works. Short-chain hydrocarbons (methane, ethane, etc.) come from the breakdown of the transformer oil, which certainly occurs under fault, everything here in .au is done via laboratory testing of oil and gas samples.
Nuisance tripping of a Buchholz relay is common enough. If the level of oil in the conservator has been allowed drop beneath the outlet of the Buchholz relay, especially possible if the transformer has been off for so long unmonitored. How cold has the transformer been in that year, does the transformer leak from gaskets, if the crew who topped up the oil didn't bleed the air from the Buchholz then it will be in the alarm state, and possibly the trip state. in which case a trip time of 180 msec is a very long time for a simple auxiliary relay to pick-up... check the auxiliary relays pick-up time with a secondary injection test set or current-clamping power supply at 80% rated volts, station battery may have been low due to the very-recent energisation closing of the supplying circuit breaker. The timing of the circuit breaker contacts closing may also be impacting on your impression of the event, at what time did each contact close on the supplying circuit breaker? I would expect a perfectly functioning aux relay contact to operate in 10msec and a circuit breaker to operate in 100msec ... for a gas trip contact which is already made in the Buchholz. But depending on the physical plant, this may be faster, slower, or have other relays to make contact.
Start getting together the commissioning engineers and testing technicians, start testing the transformer for its winding resistances, insulation resistance (DC IR & AC DLA/DDF). Do the windings to each other and the core/tank. Start comparing them to previous results, if you don't have previous results then take another transformer like it out of service and test that too for a baseline figure. Have a second team looking at the protection system and compete a full functional from the protection relay wiring at the Buchholz and time it operating it's scheme all the way to the circuit breaker main contacts opening. Nothing beats a full functional test. Obviously this assumes that you have access to the transformer and it's circuit breaker now under fault. I mean, it's been out for a year, how seriously can the asset owner really need it? :)
I am not an engineer. I have some experience in the testing of primary plant, and some experience in the testing of secondary equipment. This kind of advice I would normally seek from the asset owner, and their engineering representatives. This information should only be used as a guide and certainly the numbers given may be completely wrong for the plant you are working with. Always consult the plant manufacturer, test instruments manufacturers and past results the asset owner holds before making your own decision which way to proceed.