Sounds like your switching the mains after the inverter... this is fine, and it will carry any reasonable load on start up... I have hit mine with 8kw loads or more on turn on....no problems
It is driving the ac input to the inverter that is the problem. It syncs in fine ( although they still say to turn off, I don't think it is necesary, as it gently ramps over.... but it is turning off the mains to the inverter ( acting as charger at that point) that dices with death. It is instantly goes to 50hz, which depending where on the curve it is, will determine the surge current into the transformer.... it may also be that your single transformer is not big enough to cause the fault.... mine is huge, so kills it very effectively.
This is taken from ebay advert for the 3kw unit
PLEASE NOTE: for battery charge mode: before charge the battery, please turn off the inverter first, connect the AC power cord with home AC source and inverter, then turn on the inverter to start the function. When finished, need to turn off inverter first, then disconnect the AC power cord. Otherwise the PCB board might be damage due to high voltage spark.
Interestingly they also claim it is a UPS unit....go figure.
Your taiwan kipoint inverter may not use torroids (judging from the low efficiency I think this is the case (80%))... and non-torroids do not have the surge of torroids... can be 60 times the run current. The electronics was designed for EI transformers from the W7 style of inverter, and they have no problem.. even with two huge transformers in them./... torroids are a class of their own.... they are very tight magnetically. If they actually had control of the software, they could simply slowly move back to default frequency, and the big problem would be solved.... but I don't think they have control of the software in the chip for many functions.
They are putting in ferrites now, and I believe that has solved the original problem.... have not bought a new one for years now ( yes they still are running fine)
.............oztules