A Neural Network-Based Low-Cost Solar Irradiance Sensor

Measuring solar irradiance allows for direct maximizationof the efficiency in photovoltaic power plants. However,devices for solar irradiance sensing, such as pyranometers andpyrheliometers, are expensive and difficult to calibrate and thusseldom utilized in photovoltaic power plants. Indirect methodsare instead implemented in order to maximize efficiency.This paper proposes a novel approach for solar irradiancemeasurement based on neural networks, which may, in turn,be used to maximize efficiency directly. An initial estimatesuggests the cost of the sensor proposed herein may be pricecompetitive with other inexpensive solutions available in themarket, making the device a good candidate for large deploymentin photovoltaic power plants. The proposed sensor is implementedthrough a photovoltaic cell, a temperature sensor, and a low–cost microcontroller. The use of a microcontroller allows foreasy calibration, updates, and enhancement by simply addingcode libraries. Furthermore, it can be interfaced via standardcommunication means with other control devices; integrated intocontrol schemes; and remote–controlled through its embeddedweb server. The proposed approach is validated through experimentalprototyping and compared against a commercial device.