Abstract
Traditional rotational speed measurement methods, such as invasive sensors and visual recognition technologies, are often constrained by physical wear and environmental limitations. This paper introduces a non-invasive rotational speed measurement approach based on vibration signal frequency spectrum analysis. The proposed method utilizes the Zoom-FFT algorithm to process vibration signals collected during pump operation, enabling the precise identification of the dominant frequency and its correlation with the pump shaft frequency for accurate speed calculation. The experimental results obtained from a centrifugal pump under varying operating conditions demonstrate the following: At a constant rotational speed, flow variations have a minimal impact on the measurement accuracy, with errors ≤0.04%. Under constant flow conditions, the speed calculation accuracy achieves an error rate of 0.27% across different speeds. Compared to traditional methods, the proposed approach exhibits superior reliability and accuracy. This non-invasive method minimizes physical wear and reduces dependency on environmental factors, offering an effective solution for mechanical equipment monitoring and fault diagnosis.