The evaluation of the long-term drift of electrical capacitance measurement standards by two methods

Measurements of electrical capacitance are essential in various fields of electrical engineering, electronics and other fields. Capacitance measurements are necessary for correct design of electrical circuits and devices. Such measurements help ensure the stability and reliability of electrical systems such as power supplies, filters, capacitors, etc. In some applications, such as radio transmitters, filters, and other electronic devices, it is needed to maintain certain frequency characteristics. It is critical and appropriate to consider various factors related to the drift of the measuring instrument or measurement standard when performing measurements to ensure the required accuracy and reliability of the measurements. The study of the time drift of the measurement standard is mandatory when carrying out comparisons of national measurement standards. The types of the drift and main methods of its evaluation for measuring instruments and measurement standards between their calibrations were analysed. A conventional method of the long-term drift analysis involves the use of regression models followed by their detailed analysis. Such models are specific mathematical functions that describe theoretical values that best represent the underlying bias in the time series for the long-term drift. Exponentially Weighted Moving Average (EWMA) charts reduces the lag inherent in conventional moving averages by giving more weight to recent observations. The results of the evaluation of the long-term drift of measurement standards of electrical capacitance for high-precision calibration of measurement standards by polynomial regression diagrams and EWMA charts are given. Polynomials of the 2nd degree were sufficient to approximate the drift of electrical capacitance measurement standards under consideration. The application of EWMA charts showed greater sensitivity to drift changes over the past few years of observations compared to the regression analysis. Consistent results were obtained.