The imbalances experienced by low voltage distribution networks following the individualization of energy consumption among domestic subscribers as well as the presence of non-linear loads in the network, reduce the quality of electrical energy and are accompanied by consequences which can on the one hand reduce the lifespan of certain equipment and on the other hand, lead to a resizing of equipment and cables in the network. However, the objective pursued in this work was to carry out a study that could eliminate imbalances and reduce the rates of harmonic distortions in a low voltage network. Indeed, the processed data were obtained following measurements carried out at the secondary level of transformer number 13 of 160kVA of the distribution network of the Gécamines city (MAMPALA district). These different measurements were carried out over a period of one week for a sampling time of 20 min using an MS2205 three-phase digital clamp multimeter. Analysis of different data has proven that this network is out of balance with imbalance rates of up to 26%. The presence of non-linear loads in this network generates harmonic currents with harmonic distortion rates of up to 13.8 or 14% which leads to the presence of distorting power in the network. The latter reached a peak of around 15kVAD. Different techniques for decontaminating low voltage networks are possible, within the framework of this thesis we have opted for the modern depollution technique, that is to say active filtering. We used a four-arm parallel active filter. The reference currents were identified using the synchronous reference method, the DC bus voltage was controlled by a neuro-fuzzy regulator and the filter was controlled by modulated hysteresis. The results obtained after simulation on Matlab-Simulink proved that connecting the four-arm parallel active filter to the low voltage network eliminated the imbalance and reduced the current harmonic distortion rate from 13.8 to 1.59%.