The liquid dummy load is an intriguing solution, both literally and figuratively, to a common problem
Like most engineers, I have used dummy loads spanning DC through RF, and from a few watts to hundreds of watts. Most of these loads were built up from basic resistors, and in some cases special heat-sinking was needed to keep things under control. I fondly remember building a 1-kW, non-inductive RF load where the dissipative element was housed in a one-gallon paint can filled with mineral oil, to keep it from overheating as it would have in free air (yes, active forced-air cooling was another option, but that approach brought a new set of problems).
But sometimes, when you think you have seen it all, you haven't. I came across a fascinating article n the January 2010 of Power Electronics Technology, entitled "Testing Power Converters Using a Liquid-Rheostat Dummy Load". This article is not academic-theoretical or merely speculative: it has full analysis, rationale, construction details, performance graphs, and reference information on how to build and use an electrochemical cell as a dummy load. This is clearly a case where A solution is THE solution, so to speak. Apparently, this type of dummy load is not new at all (and how often do we even see the word "rheostat" in our electronics world?).
The other thing I found interesting was that the load in the story was not for some extreme power raring. Instead, it was for a 3.3 V, 40 W (continuous) output, which is a fairly modest power level. The conventional alternatives, according to the author, were too costly for his modest budget, and also would have required a complicated switching arrangement to set different test-loading levels, which would add undesired inductance and additional cost.
I'm not saying the liquid rheostat is the answer to you problems, But it's an interesting alternative, and interesting in itself as a creative solution (there's that word again!) to a specific test challenge.
Note: the two modest equations in the article did not show up in the online version–I saw the article in that ancient medium of print, where they did appear–so here they are:
R = (ρ × L)/A
R = d/(δ × A)
Enjoy getting out of your dummy "comfort" zone!♦