• • The TRD expands in a stepwise manner with regulation cost investment and exhibits diminishing marginal returns; identifying the stepwise growth point provides a reliable reference for cost optimization, enabling operators to avoid overinvestment beyond the point where additional cost yields negligible domain expansion.
• • As renewable penetration increases, the TRD follows a steep-rise, plateau, and sharp-drop pattern; the critical penetration threshold identified via this evolution can guide renewable construction to avoid triggering system security violations, with the sharp drop indicating a hard limit on inverter-based resource integration.
• • The prior-constraint-guided deep neural network achieves efficient and accurate boundary characterization for high-dimensional TRDs, overcoming the curse of dimensionality that causes exponential growth in computational complexity for point-search methods such as point-wise simulation and vertex search.
• • The TRD model incorporates intertemporal coupling constraints across multi-period, multi-type resources, enabling global assessment of regulation capability that point-based evaluation cannot provide, thereby supporting scheduling and resource allocation decisions with quantified feasible space.