When you think about it, 30mA is quite a lot... 7.2W of leakage...enough to fully light one of those miniature incandescent nightlight bulbs.
Water heaters are a classic case of this especially when the copper element casing has split open and keeps working. Depending on how close the split is to the active end, the fuse may blow soon after or the increased current blows the element instead...or if it's at the neutral end it may keep going for years.
30mA is certainly not acceptable leakage for a dry heating element in my opinion. It could be lethal should the earth connection fail.
One way to locate the problem is to simply insert a 15W lamp in series with the active feed, with the neutral disconnected, and to turn on each element one by one. The brilliance of the lamp will indicate if it's just one element or a cumulative effect. The earth still has to be connected of course.
One other thing of course is what else the RCD is supplying.
A classic example of what I mean is in our computer laboratories... the problem has been active to earth current flowing via the RFI suppression capacitors in the computers and their peripherals. Individually, the leakage current is negligible, but with 16 computers it was sufficient to occasionally trip the RCD. Problem was solved by creating mutliple RCD protected power circuits.
I'm just wondering if the RCD supplies other things which are already contributing, say 20mA, live to earth leakage and the cooker's leakage is the the last straw so to speak...in this case, disconnecting everything else supplied by the RCD and only running the cooker would verify it.