Research on intelligent calibration method of high precision digital multimeter based on supervised learning algorithm

Digital multimeter is a multi-purpose electronic measuring instrument based on high-precision data acquisition technology, generally including ammeter, voltmeter, ohmmeter. The traditional calibration method of high-precision digital multimeters does not consider some factors such as working environment, device aging, and working time, resulting in large changes in the final measurement results with the environment and poor stability. This article proposes a calibration algorithm based on supervised learning. Firstly, building an automatic calibration system, which includes a computer, a high-precision calibration source, a monitoring high-precision multimeter and a high-precision multimeter under test; then proposing a calibration model on the computer to control the test and monitoring multimeters. The system would acquire the output signal of the high-precision calibration source and collect the operating parameters of the multimeter under test at the same time; secondly, the computer processes the to-be-calibrated data, the reference data and the operating parameter data to obtain a training sample data set, and train the calibration model; finally, when the accuracy rate reaches the target, the trained calibration model would be used to calibrate the multimeter under test. Since the method comprehensively considers the signal data and the operating parameter data of the to-be-calibrated digital multimeter, the calibration speed and accuracy of the digital multimeter would be improved. This method has been applied in self-developed products. Under the premise of adapting to different working environments, the error of each measurement parameter can be controlled within 0.2%.