Aspects on measuring electrical current utilizing magnetic zero-flux technique

Licentiate thesis, 2018

The utilization of high accuracy measurements of electrical quantities is a prerequisite for the development of modern society. Generally, measurements serve different purposes and hence the criteria for the measurement equipment and method are different. For example, the demands are mild if the application is the continuous monitoring of the power grid, more finely tuned for measuring methods used in research and development, and often challenging in the case of certified measuring methods adequate for calibration and accreditation. RISE Research Institutes of Sweden (former SP Technical Research Institute of Sweden) is appointed the National Metrology Institute by the Swedish government for electrical quantities and continuously develops and provides measurement technology for the different needs governed by application.
The magnetic zero-flux technique is a non-contact measurement method for electrical AC and DC current and its design principle enables accurate measurements over a large current range. Its advantages come however with the price of a complex but sophisticated design. The zero-flux technique has been utilized for many decades, and there are a number of manufacturers providing commercial systems with somewhat different features.
This project is devoted to the further investigation and advancement of some metrological aspects of the magnetic zero-flux technique for AC. Practical laboratory tests on a state-of-the-art zero-flux system are used to create a picture of its properties at higher frequencies than its manufacturer has provided detailed specifications for. Focus is to determine how sensitive the measurement results are to practical arrangements and limitations of the measurement setup. A method and guide to how different configurations of the measurement setup affect the measured results in different frequency ranges is provided. Utilizing this characterization, practical set-ups can be made, as optimal as possible for the frequency range of interest, avoiding time-consuming focus on aspects not relevant for the specific application.
The identified aspects of interest are: (i) identifying the source of the measurement error in the zero-flux system’s design and, if possible, minimizing this error by design adjustments, (ii) measurement error and measurement uncertainty of a zero-flux system in presence of geometric asymmetry and disturbance from return or nearby conductors, (iii) simultaneous measurements of sinusoidal signals of different amplitudes, frequencies and phase angles, (iiii) detection of sub-synchronous events, and (v) non-steady state phenomena, like for example transients in the drive line of electrical vehicles. In this thesis, aspects (i) and (ii) above are in focus. Some conclusions can be drawn based on the performed study concerning aspect (iiii), whereas aspects (iii) and (v) remain out of its scope.
The initial step of this project was the choice of a generally applicable method for characterization and evaluation of a zero-flux system. The method chosen is the combination of a coaxial primary current path, or as near coaxial as was practically convenient, and a Digital Sampling Watt Meter (DSWM). The method can be utilized for the characterization and evaluation of other zero-flux systems. An investigation was performed to decide from which part of the construction the phase angle error stems. The performed characterization allowed compensating for the errors, making the measurement accuracy greatly improved. Two modifications to the circuitry of the zero-flux systems were introduced and evaluated, both of which yielded improvement of its high frequency characteristics up to 100 kHz. Also the accuracy within the low frequency range (from 10 – 50 Hz) was improved by one of the modifications.

The error of a zero-flux measuring system depends on the positioning of its sensor around the conductor carrying the measured current and the geometry of the primary current path. The total error increases with frequency, but which geometric factor that is the most important one varies with frequency. In this study, utilizing sinusoidal primary current, it was found that for 50 Hz, tilt of the sensor and positioning of the connection point for the measurement and zero-flux control cable (rotation) caused the largest effects on the scale factor. De-centring and the distances to different parts of the return conductor were less important in the 50 Hz case. For 25 kHz, de-centring and rotation were the main contributors to scale factor change, while tilt had the smallest measured effect. The total contribution from sensor positioning in the magnetic field to the expanded measurement uncertainty was estimated to 0.0024 % in the 50 Hz case, to 0.0040 % for 1 kHz, to 0.14 % for 10 kHz, and to 0.41 % for 25 kHz.