TL;DR:

I own a home machine shop and from time to time it would be useful to be able to pump a few kilowatts of heat into a stuck fastener, pipe thread, raw stock for forging a one-off tool, etc.

Currently my go-to solution for this is a two pole carbon-arc torch. It works great.
...Sometimes a little too great if you catch my meaning. It is very easy to melt the surface of parts I am trying to heat that way.


So I'd like to better understand the various types of induction heater architecture out there and get a rough bill of materials to perform a cost-benefit analysis.

I'd like to use primarily chassis-mount components with lug-to-lug or flat bus connections. Available branch circuit is 240V 1θ, 50A @ 60 cycles.

I've got a bunch of surplus 100~400A, 1200V IGBT modules lying around and I'm wondering if those would have adequate performance characteristics to do the job. The bigger ones are half bridge (2 gates), while the smaller 100A modules are 3-phase bridge (6 gates). I'd pull part numbers for reference except I can't remember where I stashed the damn things at the moment. I want to say some of them are Mitsubishi CM400s?

Target power output is somewhere between 3 kVA and 12 kVA. The ability to vary the output and use flat & other unusually-shaped coils in conjunction with a remote wand would be ideal.

Is it financially worth pursuing a project like this, or am I better off just sticking with a simple and reliable carbon-arc flame?

 

The frequency results used in induction is HF when used, heats to shallow depth, (skin effect), for deep penetration with material in your pic, then a lower freq is used.
There are a few designs out there on the web, if you use a portable wand, how would you intend getting HF power to the coil?
A water cooled copper conductor is usually used in the form of 3/16" - 1/4" diam copper pipe.