Intelligent detection for membrane damage based on electrical sensors

Abstract

Membrane module integrity monitoring is essential in the water treatment process. An intelligent detection method for membrane integrity based on array impedance measurement is proposed in the paper. The boundary voltage is collected in real time through the designed electrical sensor array. The Deep Neural Networks (DNN) is used for analyzing the degree of membrane damage based on collected voltage data. In our experiment, six kinds of membrane status are designed to verify the sensors and detection accuracy. Compared with traditional probe electrodes, electrical sensor arrays can obtain the overall condition of membrane inside the pipe without invasion. The membrane state recognition method based on DNN discards the empirical judgment in the traditional measurement method, and realizes intelligent detection by means of deep learning. For the samples obtained from the experiment, the recognition accuracy of this method reaches 94.1%. To illustrate the universality of this method, we performed a real-time detection experiment. As a result, the accuracy of network identification reaches 93.8%.

Keywords

• Membrane integrity
• Deep neural networks (DNN)
• Electrical sensor
• Real-time