15" Drivers in 18" Cabs

That can work but you may not be that impressed with the output you get. The Kilo can only handle about 1/2 it's rated power before exceeding xmax and the 15" version requires a 4.5cu ft(127L) cab with 42hz tuning to achieve 124dB, so depending how big those cabs are and how much output you want to achieve the Kilos may not be the best choice.

Also, what happens if the speaker is driven to its 1250 watt maximum? If it reaches xmax at 600 watts is there a risk of frying it at 1250 watts?

AJ113 wrote: Also, what happens if the speaker is driven to its 1250 watt maximum? If it reaches xmax at 600 watts is there a risk of frying it at 1250 watts?   Can you point me towards more info on the Kilo reaching xmax at half the power please?

Xmax is a mechanical limit, depends on cab size, depends on being BR/closed/hornloaded, tuning, etc... and obviously on the used speaker in the first place.
Power capacity is a thermal limit.
In BR usually the mech limit is reached sooner than the thermal limit, so you could feed more power than the point where xmax is reached and beyond, but when the voicecoil is not entirely emerged in the magnetic field anymore, efficiency drops hard, sound goes south with it and the excess power is only warming up amps and voicecoils, eventually until thermal destruction when stimulated hard enough.
Besides, power capacity is not a hard fixed value, but an average over time that the particular speaker can handle. Every half decent speaker can handle serious peaks as long as the average power over time does not exceed the advertised power capacity. Whether the same power is possible on a mechanical level, is a very different thing.

Reaching xmax is a function of what enclosure you put your driver in, what frequency it is driven at, and what power that frequency is supplied at.

Try looking up the bandwidth and filtering that the EIA A standard employs and that will give you an idea of why power ratings can be misleading depending on your application.

Ultimately that is the only way - the empirical method. The only other shot you have is to learn how to use a program such as HornResponse and model the output before cutting wood. But you still need to have an idea how those sims might sound in real life to decide whether to build it or not. 

Personally I would be tempted to design a new enclosure because bass reflex design is quite simple and the Kilomax are designed for them, and there is so much written about the process on the web. WinISD is quite a simple program to learn. And then once built it is easy-ish to verify your cabinet design with impedance and SPL measurements. Whatever you build you will need some way of high passing the signal below the cut off point of the cabinet to get the most efficient use of your drivers without shredding the cones.

Why would you lug cabinets to gigs bigger then you need! Crazy stuff. Recessing the speaker into the cab with a new baffle will affect the sound, although I wouldn't care to say exactly how. Bit like a planar wave horn in some ways - I would guesstimate a low pass filter effect.

I would flog on the 18" cabs, and look for some unloaded 15" of about the right dimensions. Sometimes its just worth trying to get it as right as possible rather than bodging.