|
This thesis presents a MIMO-based optimisation framework for improving solar photovoltaic (PV) panel orientation under varying environmental and seasonal conditions. Instead of relying on hardware-based tracking systems with sensors and actuators, the study uses a dataset-driven MATLAB simulation to determine optimal tilt and azimuth angles.
The framework incorporates key parameters such as global, direct, and diffuse irradiance, ambient temperature, and wind speed to model PV performance more accurately. A Multi-Input Multi-Output (MIMO) approach analyses interactions among these variables and optimises panel orientation across annual, seasonal, and monthly timescales. The methodology combines solar geometry, irradiance modelling, thermal analysis using the NOCT model, and a two-stage optimisation algorithm. Practical constraints, including wind protection and thermal derating, are also considered.
Evaluation using meteorological data from Kathmandu, Pokhara, and Nepalgunj shows significant improvements in energy output. Dual-axis tracking increased annual generation by up to 50.1% in Kathmandu, with gains of 33.4% and 34.4% in Pokhara and Nepalgunj compared to fixed systems. Results also indicate that optimal tilt angles vary seasonally.
Overall, the study demonstrates that data-driven optimisation can enhance PV efficiency and offers a cost-effective alternative to complex tracking systems, with potential applications in smart grids and energy management. |
|
