oa 33-Level Single-Phase SubMultilevel Inverter for Grid-Connected Photo-Voltaic Systems Using Model Predictive Control

Green House Gases (GHG) that are currently being emitted from conventional energy production methods are of global concern; especially with the climate change that is occurring. The United Nations has taken the initiative by holding a global conference that aims combating climate change “United Nations Framework Convention on Climate Change (UNFCCC)”, which sets a common ground of regulations and agreements that aim at reducing polluting production methods. Qatar, as part of the UNFCCC, has signed and agreed to the Kyoto protocol. The first component of the protocol determines that countries should enhance energy efficiency in different sectors of their national economy.

In addition, “Qatar National Vision 2030”, states: “The rights of future generations would be threatened if the depletion of non-renewable resources were not compensated by the creation of new sources of renewable wealth”. With the increase in numbers of large-scale educational, industrial, commercial and residential buildings, Qatar needs to look into means of making them more energy efficient. Adopting solar energy is a priority for Qatar because of high abundance.

Photovoltaic (PV) systems transform sunlight into useful electricity with efficiencies that range from 12 to 20% without concentrators. PVs are designed with two layers, the n-layer (negative) and the p-layer (positive), which get charged when exposed to sunlight causing potential difference across the cell. PVs are static devices, eco-friendly, noise-less and require minimal maintenance. According to a study that compares unit generation costs of gas turbines and PVs, PVs are becoming a more viable alternative despite of the fact that they are still more expensive due to their initial installation costs. Another major disadvantage is the lower conversion efficiency. Power conditioning is also required in a solar PV system to convert DC power to AC power, regulate the load flow, offer different types of protection and keep the total harmonic distortion (THD) within the specified limit. Different types of inverters (DC to AC converter) have been employed for solar PV systems.

Dynamic behavior of solar energy resource entails the need of robust controllers that can converge to the maximum power point (MPP) to maximize energy harvest. The low conversion efficiency of PV systems is a significant hindrance to their growth, therefore Maximum Power Point Tracking (MPPT) is required to ensure the maximum available solar energy is harnessed from the solar panel. It is common in the industry that power conditioning is done at the array level. Less common but proven more efficient is to do conditioning at the module level. This guarantees maximum power is harnessed from one module regardless of the irradiance condition of the other neighboring modules. It would even be more efficient to do power conditioning at the cell level. In which case, each array of cells within a single module, has its own MPPT controller.

This project explores an improved Perturb and Observe (P&O) technique that combines a fixed step model predictive controller (MPC), to speed up the control loop, applied to a boost converter. The proposed MPC Maximum Power Point Tracking (MPPT) technique had proven higher efficacy and robustness over conventional MPPT. The main characteristic of MPC is predicting the future behavior of the desired control variables until a predefined step ahead in horizon of time. The predicted variables will be used to obtain the optimal switching state by minimizing a cost function.

A boost converter is used as a DC/DC converter prior to the sub multi-level inverter. P&O determines the reference current for the MPC, which determines the next switching state. This technique predicts the error of the next sampling time and based on optimization of the cost function g, the switching state will be determined. The inputs to the predictive controller are the PV system current and voltage, and the reference current. By deriving the discrete time set of equations, the behavior of control variables can be predicted at next sampling time k+1. The proposed methodology is based on the fact that the slope of the PV array power curve is zero at the predicted MPP, positive on the left and negative on the right of the predicted MPP.

The improved MPC-MPPT is tested for the first time on an MPC strategy of 33-level SubMultilevel Inverter (SMI) using 16 power arms cascaded with the H-bridge inverter. SMIs use less switches than conventional MIs and this is noticeable at larger number of levels. Such topology uses a switching frequency for the H-bridge that is equal to the grid frequency, which allows for lower switching losses when compared to conventional MIs. Such topology brings about many sizable benefits such as reduced number of power switches and their grate drivers when compared to the traditional multilevel inverter. MPC is also used as the control strategy for the SMI to eliminate complexities in the space vector pulse width modulation (SVPWM) and overcome the weaknesses of the inner control loop performance. This topology allows for each module to be divided into 16 parts each comprised of an array of cells. Each array of cells is connected to the MPC-MPPT boost converter. Then the 16 levels are fed to the input stage of the SMI. Finally, each module will have its own H-Bridge inverter that directly feeds the grid. Such topology provides the advantage of localized MPPT tracking, increased reliability of the whole power system as the failure of one cell won't fail the whole system, and lower components withstand ratings.

To verify the dynamic and steady-state performances of the proposed MPC scheme under various load and reference currents, simulation studies are performed with Matlab & Simulink software.