Abstract: Long-term operation of pump station led to increased resistance and weakened outflow capacity, thereby impacting urban drainage efficiency. In view of this, we have developed a flow capacity check and evaluation system for pump stations utilizing the Python programming language and digital image technology. Real-time data of the axial flow pump voltmeter and ammeter is collected using cameras, while ultrasonic Doppler profilers and radar water level gauges record the outlet flow and head of the pump station. To address the issue of flow data deviation caused by complex flow patterns and ambient noise interference, we have implemented a moving median filtering method and employed Bessel interpolation for data completion. These methods significantly improve the accuracy of flow data. Our system has been successfully applied to assessing the outflow capacity of pumping stations, uncovering that the measured outflow capacity falls below the theoretical value in conditions where power and head remain constant.

Key words: axial flow pump, flow capacity, ADCP, image recognition