Neural Network-Based Active Cooling System With IoT Monitoring and Control for LCPV Silicon Solar Cells

One of the methods of increasing output power of solar cells is increasing the concentration ratio using lenses and mirrors. Due to the temperature increase of silicon solar cells with concentrating lenses the system must be provided with an active cooling system. PV systems cooled with active or passive cooling methods using air, water or other substances are not taking into account the relationship between the power of solar radiation and operating power of the pump used in the cooling system, which is sufficiently complicated and little studied. This paper proposes a new active energy-efficient cooling system for silicon LCPV (Low Concentrating Photovoltaic). The idea of the proposed cooling system is to make the pump operate as efficiently as possible relative to the power of solar radiation and initial temperature of the heated solar cell. Performed experiments show that at fixed temperature and fixed solar radiation there is the most optimal pump operating power. Neural network is used to find this optimal pump power. Using neural networks requires large computing resources and development software which is not often available on local control computing devices nearby solar power plants. IoT (Internet of Things) technologies allow not only remote monitoring and control, but also forecasting the consumption of the cooling system depending on the current temperature and solar radiation power. Thus, energy efficiency of the cooling system is improved using neural networks and IoT technologies. The simulation of the cooling system operation was performed using various algorithms of the cooling system operation. The proposed active cooling system for silicon LCPV using IoT technologies allows the power consumption to be reduced by 60% compared to algorithms based on the threshold value of temperature.