Adaptive Neuro-Fuzzy Inference System-Based DC Voltage Sensor for Solar Energy Measurement

Voltage measurements in a PV system are very important, especially in the development of an intelligent energy system. In an intelligent system, the high-voltage side of the PV array must be isolated from the sensitive circuitry of the monitoring device or microcontrollers. That is why an analog optocoupler is needed. The application of a non-linear device in measuring the voltage may affect the measurement. Inaccurate measurement of voltage results in wrong interpretation and making decisions in an intelligent system. The accurate measurement of voltages in a system is important to have an effective energy management system. This study proposed an ANFIS-based voltage sensor to eliminate the effect of the optocoupler and other components’ non-linearity in the measurement of voltage. Based on the result the root means square error (RMSE) is 0.49 and the mean absolute percentage error (MAPE) is 3.25% which shows its effectiveness.