High-precision and short duration operating time dispersion in a fast mechanical switch driven by an ultrasonic motor: Modeling, prediction, and compensation

To mitigate electromagnetic transient impacts on the power grid during circuit opening and closing operations, precise phase control necessitates exceptional consistency in the operating time of fast mechanical switches driven by ultrasonic motors. However, mechanical clearance and collision introduce significant dispersion in operating time. Moreover, because the operating time lies on the millisecond scale, conventional closed-loop control methods face prohibitively complex implementation challenges. To address this issue, this study develops an analytical model of the fast mechanical switch driven by an ultrasonic motor and proposes a Support Vector Regression (SVR)-based prediction model to characterize and compensate for operating time dispersion via a delayed-start strategy. The analytical model provides the theoretical operating time and identifies key working parameters. The SVR-based prediction model then estimates the actual operating time. The required delayed-start time corresponds to the difference between the theoretical and predicted operating times. Unlike traditional closed-loop control, which demands high-frequency sampling and real-time computation of control laws, the proposed method only requires offline prediction of operating time based on observed data, followed by determination of the delayed-start interval. Experimental results demonstrate that this approach controls operating times within 23 ± 0.15 and 25 ± 0.15 ms, satisfying the phase control requirements for the fast mechanical switch. This performance surpasses the current operating time dispersion control of ±0.2 ms achieved in Thomson-coil-driven mechanical switches.