Hydro-quebec

Alexandre Bouffard

François Léonard

Arnaud Zinflou

Mathieu Viau

Francois Leonard

François LÉONARD

The disclosed method detects anomalies in an electrical network. An ohmic matrix model of the network is initially generated from consumption measurements produced by meters connected to a same transformer. The model has currents carried by the meters as input, relative voltage drops of the meters referenced to a voltage of a reference node located on the network as output, and resistive quantities initially determined by currents and voltages based on the consumption measurements as matrix terms, the relative voltage being a voltage difference between a voltage determined for a meter and an average of voltages determined for a set of meters of the network. The model is iteratively modified according to detected anomalies, and a diagnosis of the network characterizing the detected anomalies is provided.

Detection of anomalies in an electrical network

The disclosed method corrects consumption measurements provided by meters presumed to be connected to a same network. The measurements taken by the meters at time intervals are collected in a form of samples used to determine electrical quantities in relation with currents and voltages relating to the meters. A selection of a group of retained samples among the samples considered to be valid, corresponding to a load range or a range of load variation ratio carried by the network based on the electrical quantities is performed. Correction functions of the consumption measurements are defined from the group of retained samples, and the consumption measurements are adapted according to the correction functions. An anomaly corresponding to an electrical non-compliance of a meter can be detected by the method.

Growing community of inventors

CA

Alexandre Bouffard