Model Predictive based Energy Efficient Control of an On-Grid PV Inverter

Abstract

Renewable energy is now one of the most interesting topics in the field of distributed energy generation. Among the renewable energy sources, the solar photovoltaic (PV) source has become the most popular and effective one. Generally, PV system is connected to an ac grid through an inverter. The inverter should be controlled in such a way that it can penetrate maximum power to the grid. However, designing a controller for the inverter is a difficult task due to the intermittent PV source. The traditional pulse width modulation (PWM) based inverters produce high total harmonic distortion (THD) in the output current. This causes a significant amount of power loss and thus less power penetration to the grid. It makes the on-grid PV systems inefficient. Therefore, model predictive control (MPC) based energy efficient conversion of PV power is proposed in this research work. In the proposed MPC, the control objectives (current and switching frequency) are predicted using a finite number of voltage vectors produced by the inverter. MPC selects an optimal control action (i.e. switching state) in every sampling instant for the inverter by minimizing a predefined cost function. The cost function includes current tracking error and number of switching transitions. These two control objectives are combined in the cost function with a weighting factor. The first control objective provides a smooth current tracking accuracy, which increases the r.m.s amplitude of the injected current and thus increases the power penetration to the grid. The second control objective reduces the average switching frequency, which actually reduces the switching loss. The value of weighting factor in the cost function is selected by making a tradeoff between the current THD and the average switching frequency. Simulation results show that the proposed controller tracks the reference current accurately with mean absolute error of 2.5% which is 30% for the PWM based controllers. The low current tracking error for MPC based inverter yields low current THD of 2.07%, whereas in traditional PWM based inverter the current THD is 7.26%. As a result, the proposed MPC based inverter penetrates 12.8% more active power to the grid than the PWM based inverter. The active power penetration is confirmed by load flow analysis using IEEE 13 bus test feeder. The energy efficient operation of the MPC based inverter is also verified by doing loss analysis. There are three types of loss considered in the research work: conduction, switching and harmonic losses. The conduction, switching, and harmonic losses of the inverter are reduced by 36.8%, 50%, and 91.9%, respectively, in comparison with the PWM control based inverter. The performance of the proposed controller is also analysed in terms of transient response, decoupling control, and fault tolerant ability. It is shown that the proposed MPC yields decoupled current control and fast transient response, and capable of handling the symmetrical and unsymmetrical faults in the grid. The research outcome from this comprehensive analysis proves that the proposed controller reduces the power loss to maximize the penetrated power and ensures the performance of the proposed controller as an energy efficient controller for an on-grid PV inverter.