You need to review the whole circuit and mat again.
You have a 30mA RCD protected circuit a 16A MCB a 6mm² cable (single radial I assume) connected to a switch of some sort and then to a power backbox (connected correctly? the power back box should not be able to take 6mm² cables) then the controller is fitted on the back box (how is the controller set up? Are there jumper switches? what are the displays if any on the controller on start up?) the controller is designed to take no more than 16A.
From the power box there are three pairs of wires each pair going to one of three underfloor heating mats (in the same room?), there are some unused terminals for a floor sensor (possibly) and (possibly) a pilot wire terminal unused.
There is no floor sensor (is the controller temperature sensor clear of dust and debris)
The floor heating mats appear to be 2 x 2.2kW and 1 x 6.6kW(really?), it is not known if there is an earthing grid associated with the mats, make of the mats is unknown.
There does not appear to be an earth fault as the RCD does not trip.
The current anticipated would trip a 16A MCB in about 10 seconds. The measured current on start up is not known.
The controller has been working for 2 years satisfactorily but has (in the last ....days) stopped working and when powered up starts and then cuts out, then starts, then cuts out...... (possibly not calling for heat yet and so not tripping MCB, but MCB was holding when it was working, is the controller staged i.e. only powering the mats by PWM or such like)
The heating mats have individual continuity on each (24, 24 , 8 ohms).
The insulation resistance of the mats is unknown.
The insulation resistance of the circuit up to the power box with and without the switch is OK.
The controller and power box from another location, where they are working, have been used in place of the original and the fault recurs (so the fault is not in the controller or back box).
The instructions state that when using a contactor (not used in this instance?) a snubber should be fitted across the load terminals.
When each mat is connected individually the fault recurs (so not a current loading issue)(only a triple fault on every the mats or a cross connection between the mats could cause this if the fault is in the mats.)
[Right no faults on fixed wiring, controller or mats (except loading), your heating is fine madam, that will be £250, oh yes the resetting? it is typical of these units, just ignore it, that is what I am going to do!:nonod:]
Does the controller start up OK without the mats connected?
Does this controller work in the other location?
Do you have the specs / instructions for the controller, power box or mats that you can compare to current readings?
Did anything change just before the heating failed (any work DIY or Professional)?
Well I give up!