COMPARISON OF CURRENT CONTROL SCHEMES FOR SINGLE-PHASE GRID-CONNECTED PV INVERTER WITH MPPT

Abstract

This project analyze and compares four current control schemes for a single-phase grid-connected PV inverter. The first two methods discussed are Synchronous Reference Frame (d-q) and Proportional Resonant (PR) based controllers. Both are linear regulators which uses Pulse Width Modulation (PWM) for the generation of the control signals. The PR controller provides high gain at a chosen frequency (also known as resonant frequency), due to which it is able to subdue the steady state error. Therefore, it can be successfully applied to the circuits where the controlled parameter is sinusoidal in nature with frequency of operation equal to that of resonant frequency. Whereas a PI controller suffers from steady state error and high sensitivity towards disturbance when used with sinusoidal input. To counteract this limitation of PI, the d-q transform is used to convert sinusoidal signals in stationary frame to DC signals in rotatory frame. This way the PI controller is able to bring the steady state error to zero. The other two methods described are Fuzzy logic control (FLC) and Model predictive control (MPC). These are advanced control techniques which takes into account the non-linearity of power converters. FLC also works on DC parameters and therefore the controller structure is similar to that of d-q controller except that instead of PI, a fuzzy inference system (FIS) is used to do the job. MPC on the other hand uses the state space model of the system for predicting the future value of the controlled variables. This predicted value is utilized by the controller to find the required switching state, which is in accordance with a predefined cost function. The four schemes designed are compared based on their mathematical modeling, working principle, dynamic response to disturbances, and the ease of implementation. v The modeled system consists of PV, single phase H-bridge inverter, LCL-filter and an AC grid. Compared to the L and LC filter, the LCL filter has higher harmonic attenuation at the switching frequency and better decoupling from the grid impedance. However, it suffers from inherent resonance which may introduce instability in the system. To overcome this problem, a proper damping method is required to suppress the oscillations and this is achieved using either passive damping or active damping. The often used passive damping method is to connect a resistance in series with the filter capacitor. This method has been used so as to keep the control system simple and reduce sensor circuitry. However, it reduces efficiency of the overall system. The PWM technique which has been used is Bipolar Sinusoidal Pulse Width Modulation (SPWM). This is because as compared to Unipolar SPWM, it reduces the ground leakage current and thus provide galvanic isolation to the PV system from the AC grid. The outer voltage control loop consists of a PI controller whose output acts as a reference for the inner current control loop. The outer loop regulates the PV voltage by comparing DC link capacitor voltage with the Perturb & Observe (P&O) MPPT algorithm to obtain maximum power from the modules. The effectiveness of the developed methods has been confirmed with the help of the simulation results. A 5 kVA grid-connected PV inverter is designed for this research on MATLAB/Simulink. A hardware implementation of a single phase H-bridge inverter with an LC filter is also presented. It consists of H-bridge IGBT inverter, driver circuit, level shifter circuit, auxiliary power supply, and a micro-controller. A basic open loop operation is performed and validated with the simulation results.