Algorithmic development for power quality monitoring by a network of wireless sensors

Abstract

Human activity is more than ever dependent on electronic systems. These devices are fed through the electrical grid, which in turn supplies multiple devices simultaneously. The normal operation of an equipment can produce quality events in the power bus that can lead others to fail. This justifies the need for a wireless network of sensors (WSN) spread through the electrical installation. In addition, a wireless solution results in easiness of installation and maintenance. The main task of a WSN will be to monitor in real time the occurrence of power quality events. These goals will demand the development of efficient computational algorithms to detect and report these quality events through the network. In this paper two different solutions are described and compared. In a first approach the wavelet transformation is used for event detection. This solution is afterwards compared with a less computational demanding algorithm based on segmentation and labelling which also compacts the information to be transmitted.