Energy Quality Improvement of PV Grid Connected System Associated with Active Power Filter Based on Fuzzy-Predictive Direct Power Control Sabir ouchen Electrical engineering department LGEB laboratory Biskra Biskra , Algeria Ouchen_sabir@yahoo.fr Achour Betka Electrical engineering department LGEB laboratory Biskra Biskra , Algeria betkaachour@gmail.com Jean Paul GAUBERT Laboratory of Computer Science and Automatic Control for Systems (LIAS- ENSIP), University of Poitiers France jean.paul.gaubert@univ-poitiers.fr Sabrina Abdeddaim Electrical engineering department LGEB laboratory Biskra Biskra , Algeria s_abdeddaim@yahoo.fr Mehdi sellali Electrical engineering department LGEB laboratory Biskra Biskra , Algeria sellalimehdi7@gmail.com Abstract— Predictive control is a very large control category. It has found a fairly current application in the mastery and the control of power converters. This paper presents a simulation study which includes a control strategy to improve the performance of a double stage photovoltaic system connected to the network associate with shunt active power filter. A predictive direct control power is proposed to compensate the reactive power and to inject the active power into the network. A fuzzy logic algorithm is also proposed to track the maximum power point (MPP) of photovoltaic generator. The results show that the control strategies applied on the proposed system provide fast and high performances under different irradiation conditions. Furthermore, the controllers have presented a flexible settlement of amounts of active power exchanged between the PV system and the grid under a unite power factor and a low total harmonic distortion that meets to the IEEE Standard 519-1992. Keywords— PV Grid connected; fuzzy logic control; predictive direct power control; maximum power point. I. INTRODUCTION Nowadays, at the current rate of consumption of energy resources, the deposits of fossil fuels (coal, oil and gas) and fissile (uranium) represent only a few decades. In these times, when the energy becomes a major challenge, both economically and ecologically, the situation is even more alarming that energy demand is growing [1]. Renewable energies offer undeniable benefits for the environment and security of supply. They emit little greenhouse gases compared to fossil fuels. Among these source, photovoltaics is positioned as a solution to the depletion of fossil energy resources[2]. From the moment that the exposure of the photovoltaic panels of a nonlinear P-V characteristic curve which is variable with the temperature and irradiance conditions. The methods to attain the surveillance of maximum power tracking Point Tracking (MPPT) is a very significant technology [pso1]. Recently, numerous MPPT techniques have been suggested, the majority of these techniques can be grouped into two types: off-line and on-line approaches. Offline or indirect technique such as open-circuit voltage (OCV) [3], and short-circuit current [4] . It’s called by this name because the PV panel must be disconnected from the system to measure the parameters such as short-circuit the open or circuit voltage. Online or direct technique such as perturb and observe (P&O)[5], incremental conductance (IC) [6], because they use the instant output current or/and voltage of the PV panel during its operation. Due to drawbacks of conventional MPPT algorithms, many algorithms and artificial intelligence techniques have been proposed to cover the old flaws. These new MPPT algorithms which their principle is taken from biological structure and nature have been developed to magnify output power from PV array. As an example, one can cite: artificial neural network(ANN) , genetic algorithm (GA)[7] and particle swarm optimization (PSO) [8]. In addition to that, the fuzzy logic controller (FLC) [9] that is used in the present work, which is widely used recently.