The performance of a continuously variable transmission (CVT) system on a small-capacity automatic motorcycle is highly dependent on two adjustable components: roller mass and pulley angle. However, most previous studies have relied on a one-factor-at-a-time approach, adjusting either roller mass or pulley angle separately, thus their combined and interactive effects on engine performance have not been well quantified. This study addresses this gap by simultaneously optimizing roller mass and pulley angle on a 109.5 cc SOHC eSP automatic motorcycle to optimize engine power and torque. A Custom Design-based Response Surface Methodology (RSM) was applied, and thirteen experimental runs were tested on a dynamometer, with roller masses ranging from 11.4–15.6 g and pulley angles from 12.8°–14.2°. A Two-Factor Interaction (2FI) model was developed and evaluated using Analysis of Variance (ANOVA). Both models are highly significant (p 0.0001), with pulley angle identified as the dominant factor and the interaction (A × B) between roller mass and pulley angle statistically significant for both responses. Numerical optimization based on the desirability function yields an optimal configuration at a roller mass of 11.4 g and a pulley angle of 13.597°, which delivers an engine power of 4.777 W and an engine torque of 7.29 N•m with a desirability of 0.522. This validated predictive model quantifies the roller–pulley interaction and provides practical engineering guidelines for tuning automatic motorcycle CVTs to balance power and torque.
Copyrights © 2026