Vertical-axis wind turbines have been considered for distributed generation, yet the performance of lift-driven Darrieus rotors has remained sensitive to rotor geometry and operating tip-speed ratio. This study screened the combined effects of blade height, chord length, and rotor radius on an S1046 Darrieus rotor developed from the Sultan Wind Turbine v4.5 numerical baseline. A three-factor, three-level L9 orthogonal array generated nine geometries, which were evaluated in QBlade Win v0.96.3 at seven sampled tip-speed ratios between 0.01 and 3.01. The maximum sampled power coefficient, the corresponding sampled tip-speed ratio, the trapezoidal area under the sampled power-coefficient curve, and descriptive factor-level differences were compared. The baseline produced a maximum sampled power coefficient of 0.294 at a tip-speed ratio of 2.01. Variant 7, defined by a blade height of 2.2 m, a chord length of 0.17 m, and a rotor radius of 0.32 m, produced the largest value, 0.544 at a tip-speed ratio of 2.51, representing an 85.2% increase relative to the baseline under the common simulation setup. Its curve-area index was 39.5% greater than the baseline value. Height and chord displayed the largest level ranges within the unreplicated matrix, whereas radius showed a smaller averaged main effect. The results identified Variant 7 as the strongest candidate among the nine simulated geometries, but not as a validated physical optimum. Independent reconstruction of the QBlade model, Reynolds-number-specific polar documentation, refined tip-speed-ratio sampling, uncertainty analysis, structural assessment, and experimental confirmation remained necessary before prototype adoption.