Probabilistic Analysis of Time-Series Production Simulation for Large-Scale Renewable Energy Bases in Desert-Gobi-Wasteland Regions Based on Dimension-Adaptive Sparse Grid Interpolation

• • The DASGI surrogate model reduces computation time by over 90% compared to the original time-series production simulation while maintaining a coefficient of determination (R²) exceeding 0.99 for renewable accommodation rate and transmission corridor utilization metrics, enabling rapid probabilistic planning studies. • • Monte Carlo sampling with 10,000 scenarios coupled with the DASGI surrogate achieves a 95% confidence interval width of less than 2% for the annual utilization hours of the HVDC corridor, providing statistically robust risk assessment for violation probabilities. • • The dual uncertainty framework—incorporating both stochastic renewable output and planning variables (installed capacity, peak load)—reveals that ignoring capacity uncertainty can underestimate the violation risk of renewable accommodation rate by up to 15%, underscoring the necessity of coupled probabilistic analysis. • • Increasing the number of DASGI configuration points from 50 to 200 improves the detection accuracy of potential violation risks from 85% to 98%, with a marginal increase in computational cost, demonstrating the method's scalability for high-dimensional uncertainty quantification.