The thing to remember is that voltage is electron pressure. That is literally what the "Volts" unit translates to.
If you have a water hose, you can use hose barbs and simple hose clamps on a garden hose because it's low pressure. If you try that with a hydraulic line, it blows right off. Similarly we have to do special "high pressure" terminations with medium voltage wiring. Creep, partial discharge, corona, whatever you call it is a serious problem that can shorten the life of equipment.
Contrary to popular belief, there is nothing inherently wrong with water and electricity. It's a combination of water, electricity, and some sort of contamination. I'm in maintenance so I get to see the end result of a lot of failed experiments like this. What typically happens is that at first you get condensation or leaks or any other kind of moisture building up and you get harmless drips of water where it falls down from a painted surface. It may carry dust with it. Eventually the water and dust corrodes the paint somewhere and now you are dripping a mix of water and metal oxides onto the surface which are all semiconductors. This causes corona/partial discharge that if left unchecked leads to a failure. Some equipment is much more susceptible than others...it depends on the brand and often whoever (person) that was building the equipment at the time.
Another major issue is something called "dry banding". As medium voltage equipment dries out, it tends to dry out in patches. The voltage across these returns to normal but the remaining wet areas are still relatively low resistance. When this happens, you get a MUCH HIGHER voltage flux across the dry areas, leading to burned or cracked insulators, which then provides more contamination to accelerate the process. You can google the key words "dry banding" and "insulators" for more information on this phenomena. Again the key thing here is medium voltage is high voltage PRESSURE.
"Partial discharge" does not exist below 2000 V to any appreciable degree. If we exceed a certain voltage flux (V/mm) above a certain level, we will get various electrical effects (glow discharge, streaming, eventually leading to a flashover) where current starts to flow. At low voltage except in special situations such as inside discharge lighting (neon tubes, fluorescent fixtures, metal halide, etc.), this effect simply doesn't exist. So we don't see insulation breaking down from these effects. But at medium voltage and higher we are working close to the limits of the materials. Hence we have to take great pains to maintain an even voltage distribution on the insulation. We use semiconductor coatings around larger diameter wire strands, shielded cabling, and carefully maintain large spacing away from both other phases and grounded surfaces within switchgear. This all works until we introduce contamination into the system. That's when breakdowns occur and things blow apart. Plenty of switchgear operates just fine in outdoor locations but design choices go into how we design equipment intended for different locations. There is a "pollution" rating on medium voltage gear. If you change the environment that it was designed for, you can jeopardize equipment not rated for it.
In some operations my company gets paid to go over all the equipment every year with vacuums, rags, and bottles of denatured alcohol to clean all the dirt and dust out of the equipment. It's a great way to do a full visual inspection and PM every inch of the equipment, but it's a huge expense and lost production time so best to just avoid it in the first place.
