Awai Ujile gave a short talk on harmonics in power systems. She explained that linear loads, where the current drawn from the supply is proportional to the supply voltage, don’t inject harmonics into the supply whereas non-linear loads do. Examples of linear loads are heaters and incandescent lighting. The increasingly common use of power electronics has lead to a shift to a greater proportion of non-linear loads on the electricity distribution system.
Examples were given of non-linear loads, such as fluorescent lighting, battery chargers, solar panel inverters and HVDC (high-voltage direct current) conversion equipment. The later usually forming part of transmission systems, which were outside the scope of the talk.
A typical example of a repetitive non-sinusoidal current waveform was analysed and shown that it could be treated as the summation of harmonics of the supply frequency (f), i.e. 2f, 3f, 4f etc. Reference was made to the US IEEE standards, where for ranges of distribution voltages, there were specific limits for individual harmonics and to the UK’s Engineering Networks Association (ENA) Engineering Recommendation G5/4-1 that set limits in terms of THD (total harmonic distortion), that is the ratio of the root-mean-square summation of the harmonics and the fundamental.
Different strategies for estimating harmonics were discussed, classed as parametric, non-parametric and static and dynamic. Static methods used techniques such as FFT (Fast Fourier Transforms) while dynamic methods employed Kalman filters that use a series of measurements observed over time, containing statistical noise and other inaccuracies, and produces estimates of unknown variables that tend to be more precise than those based on a single measurement alone.
In a real distribution network the harmonics and their amplitudes can vary continually as different loads are switched on and off the network. Variations of the direct current component (imbalance between mean ‘positive’ and ‘negative’ cycle currents) and supply frequency variations also contribute to the variation of harmonics over time. The magnitude of harmonics varies across a distribution network, reducing towards the supply point.
The talk gave an overview of what supply harmonics are, how they are caused and the limits defined. A comparison was made of the results using different methods of design estimation. In the past harmonics were produced by non-linearity introduced into networks by such things as magnetic circuits, motor commutation and arc-rectifiers. Now it seems just about every load is fronted by power electronics that not only makes a non-linear demand on the supply but might even generate non-synchronous ‘harmonics’ of their own, so not just amplitude variant but time variant too. Perhaps the solution is in the problem - the control electronics of these complex loads actively extracting and injecting energy to counter the harmonic component of the load?