Adaptive Neuro-fuzzy Control of Solar-Powered Building Integrated with Daylight-Artificial Light System

To meet the increase in energy consumption and to reduce global warming, power sources based on sustainable energy are in demand. Of the different sustainable power sources, solar energy is the most viable option. Artificial intelligence (AI) centered control strategies are a part of sustainable power source frameworks. This paper shows the effectiveness of an adaptive neuro-fuzzy inference system for maximum power point tracking (MPPT) of a photovoltaic system. A boost converter with adaptive neuro fuzzy-based incremental conductance MPPT algorithm and a lithium-ion battery-based bi-directional DC-DC converter controlled with a voltage-current controller for power balancing and DC bus voltage regulation is discussed. A Simulink model developed by taking the climate data as the standard input for solar photovoltaic (PV) module and daylight-artificial integrated scheme. The work involves system design towards the load side for building and designing of the PV system to achieve maximum power and efficiency for a fixed tilt using PVSYST. The performance of Fuzzy Logic Control based incremental conductance MPPT technique is compared with the adaptive neuro-fuzzy inference based control. Using the model energy availability and energy consumed is estimated. The models are validated using actual pyranometer readings.