Calibration of Zmpt101b Voltage Sensor Module Using Polynomial Regression for Accurate Load Monitoring

Smart Electricity is quickly developing as the results of advancements in sensor technology. The accuracy of a sensing device is the backbone of every measurement and the fundamental of every electrical quantity measurement is the voltage and current sensing. The sensor calibration in the context of this research means the marking or scaling of the voltage sensor so that it can present accurate sampled voltage from the ADC output using appropriate algorithm. The peakpeak input voltage (measured with a standard FLUKE 115 meter) to the sensor is correlated with the peak-peak ADC output of the sensor using 1 to 5th order polynomial regression, in order to determine the best fitting relationship between them. The arduino microcontroller is used to receive the ADC conversion and is also programmed to calculate the root mean square value of the supply voltage. The analysis of the polynomials shows that the third order polynomial gives the best relationship between the analog input and ADC output. The accuracy of the algorithm is tested in measuring the root mean square values of the supply voltage using instantaneous voltage calculation and peak-peak voltage methods. The error in the measurement is less than 1% in the peak-peak method and less than 2.5% in the instantaneous method for voltage measurements above 50V AC, which is very good for measurements in utility. Therefore, the proposed calibration method will facilitate more accurate voltage and power computing for researchers and designers especially in load monitoring where the applied voltage is 240V or 120V ranges.