Electric springs (ES) are often known to be demand-side power management systems in decentralized grids with renewable energy sources, which, due to their nature, inject unreliable electricity power into the system. The second type of electrical spring (ES_2) produces an intelligent load by placing it next to a sensitive load, which, in addition to regulating the voltage level, will optimize various parameters of the power quality for the sensitive load. In imbalance grids with harmonic voltages and local non-linear load, where conventional controllers are used, the intelligent load cannot improve the power quality for a sensitive load. In this paper, a proportional resonance (PR) controller is used to regulate both the voltage level and the source voltage imbalances due to the least tracking error in the sinusoidal mode as well as a Repetitive controller (RC) is designed to reduce THD and improve the power factor thanks to its infinite poles on the imaginary axis. Finally, these two suggested controllers were tested simultaneously in three-phase and single-phase grids by mathematical and machine simulation. The grid considered in this paper has harmonics up to order 17 and voltage fluctuations in the range of 0. 954 to 1. 1 per unit, as well as a local non-linear load with 100% of the sensitive load size and, in the three-phase mode, the grid, has unbalanced three-phase load. The results of the MATLAB simulation for electrical spring, controlled by the PR RC controller, which is a combination of two proportional resonance controllers and a Repetitive controller in both three and single-phase mode Note that the sensitive load voltage is completely adjusted to its reference value, THD reduces by 80. 31% compared to the uncompensated condition and will be less than the value defined by the IEEE standard and the power factor rises up to near the unit.