10kW for 10 secs is a puny amount of energy for a brake to get rid of. Even a small motorcycle disc brake will handle this amount of energy, provided you don't keep the load on for minutes at a time.
If you use a quick 'rule of thumb' comparison, and make some crude assumptions, then if a motorcycle takes the same time to slow down under braking as it does to accelerate, you can fairly assume that roughly the same amount of power was being absorbed by the brakes as was being generated by the engine. If the motorcycle can slow down faster than it will accelerate (which is usually the case) then the brakes will absorb even more power.
90% + of the braking will come from the front wheel, so if your motorcycle wheel/disc is from a machine with more than about 10 to 15kW then it should be fine, although it'll get a bit warm.
The key is that you only need intermittent use, for no more than maybe 10 secs at a time, which should give the disc time to cool down.
You might want to think about a conventional band brake dyno. Many years ago I used one of these at college to measure the torque of a big (multi hp) electric motor as a lab experiment. It had a steel (or maybe cast iron) drum around a foot or so in diameter, with a flexible friction material band half wrapped around the outside. This band was attached to a tension adjuster and a spring balance, which was calibrated to directly read the torque applied to the drum. The drum was hollow, with a lip around the outer edge. It was partially filled with maybe 4 or 5 litres of water. Cooling was simply by evaporation through the central filling hole. IIRC this thing would run for several minutes absorbing quite a few hp without boiling the water. Not really surprising when you think how long it takes an electric kettle, with maybe 2.5kW of power, to boil a litre or so of water.
Jeremy