Peak electric field and average heating are different constraints. Define peak and average power, pulse width, duty, repetition rate, mismatch, fault duration and energized ports. Evaluate steps, gaps, sharp edges, windows, contamination, transitions and flange joints where field concentration can exceed the level in a uniform straight section. A family power label without waveform and geometry is not transferable evidence.
State gas or purge, pressure, altitude or vacuum, material, plating, surface condition, temperature and heat-removal path. Corona, arcing and multipactor depend on the actual geometry, field distribution, surface and pressure regime. Use configuration-specific analysis or susceptibility assessment and named margins; do not convert a generic chart or tool output into an unconditional component rating.
Before acceptance, inspect and clean the assembled path, verify low-power S-parameters, define interlocks and reflected-power limits, then ramp under monitored temperature, pressure, forward and reflected power. Hold the required duty and fault cases, inspect for damage and repeat the RF measurement. Record the exact assembly, instrumentation, calibration, uncertainty and pass criteria needed to reproduce the result.