Caractérisation et amélioration d'un système de mesures de consommation électrique résidentielle

Abstract

In this thesis, the description and characterization of a monitoring system for residential electricity consumption are presented as well as an explanation of improvements that were made to the system. The system consists of a network of wireless Hall effect sensors which measure the value of the current flowing in the wire to which they are attached. These sensors are installed directly in the electrical panel of the house at the output of each circuit breaker. The sensors sample the electrical signal at a frequency of 3,08 Hz. They then communicate the last 18 measurements to a central processing unit (CPU), located nearby, with the communication protocol IEEE 802.15.4. This CPU aggregates the received data, calculates their RMS value and transmits the data over the Internet. Informing the user of the energy consumption of each of his devices allows him to make changes to his lifestyle in order to reduce his consumption and his environmental impact. Due to significant variations on the measurements when changing the orientation of the sensors, a magnetic shield has been developed. This shield has reduced by more than 50% the variations in measurements during a rotation of the sensor. To characterize the sensors, the following tests were performed : evaluation of the standard deviation, impact of the wire gauge, possibility of a single calibration for all sensors, impact of the proximity of the circuit breaker, impact of a rotation of the sensor in the electrical panel, sensitivity to the proximity of another charged wire and reduction of the received power as a function of the distance between the sensors and the CPU. The following results were observed. Each sensor must be calibrated individually due to large differences between each sensor. The measurements of the sensors have an average standard deviation of 4,72%. Sensors calibrated for AWG 14 can be used for sizes 10 and 12, but for the wires of gauge 6 and 8, the calibration must be made directly on the wire. An increase of 5,8% of the readings was observed when the distance between the sensor and the circuit breaker changes from 1 to 4 cm. For a rotation of 180 ◦ around the wire, an average variation of 3,55% was measured. Finally, the range of the wireless communications between the sensors and the CPU was evaluated to a distance of 1 m when the sensors are installed in a closed metal housing.

Date	5 Sept 2013
Original language	French
Awarding Institution	École de technologie supérieure
Supervisor	Ghyslain Gagnon (Supervisor) & François Gagnon (Co-supervisor)