Interesting ... I suspect Jonno's explanation is spot on, although I do not know where written confirmation of the 250MVA figure might be. It might be specific by region so best contact the DNO to be sure.
Certainly, it could NOT mean that if the new generator fault level is less than 10% of the ACTUAL system fault level, that it would be a "high impedance" network, because that logic would be back-to-front. If the generator fault level was MORE than 10% of the actual system fault level, that might be used as a definition of a "high impedance" network, but thats not what's stated in G59/2. So, I suspect Jonnos explanation is the correct one. It would be helpful if the exact definition was included in that section or table, or at least with guidance on the maximum design fault level values.
Dr. Andrew Roscoehttp://personal.strath.ac.uk/andrew.j.roscoe