Harmonic emission at any point of connection (PoC) depends on the loads connected there and the background voltage distortion at that point. Under balanced supply voltage and load, the harmonics adhere to the rule that third harmonics are zero sequence, fifth harmonic are negative sequence, etc. However, under unbalance in either load or supply voltage, this assumption is no longer true. A phasor analysis of the voltage and current measured at different locations in a city is presented in this paper. Individual harmonics are decomposed into their sequence components, and the results are presented in terms of polar plots. The components are further quantified in terms of balanced and unbalanced THD, which is an indicator of the impact of unbalance. The harmonics do not follow the defined norm of sequence component in presence of unbalance in the system. The triplen harmonics were shown to have higher unbalanced sequence components than non-triplen harmonics.

This paper presents harmonic voltage measurements for a part of a distribution network including low-voltage (400 V) and medium-voltage (10 kV). Measurements were performed during 2017 and 2018 in the north of Sweden. This paper shows correlations between different harmonic orders in low voltage and the correlation between low and medium-voltage harmonics. This information can be used for harmonic propagation studies in distribution networks. The results showed significant correlation between the 5 ^th and 7 ^th harmonics for low-voltage side and very high correlation between 5 ^th harmonic order in low-voltage and medium-voltage. It is also observed that 5 ^th harmonic has almost the same values for all transformers connected to the same 40 kV network.

The ability of the distribution network to host electric vehicle (EV) charging might be limited by the harmonic voltages due to their harmonic emission. Network harmonic impedance seen from the point of connection of the charging station plays an important role in harmonic voltage calculation. In this paper, the hosting capacity is estimated for a fast charging station close to a distribution transformer considering different scenarios in terms of network modelling. Data from a typical Swedish distribution network is used, together with a one-month measurement of the emission from state-of-the-art EV charging. The hosting capacity using harmonic voltage limits is compared with the hosting capacity using the transformer rating. A stochastic approach is used for both. This paper shows that harmonic hosting capacity studies are needed; it shows that details of the distribution network must be included to get an accurate estimation of the harmonic hosting capacity; it also shows that a stochastic approach is needed for estimating the harmonic hosting capacity.

This paper presents results of harmonic voltage measurements at multiple locations in a low voltage network. Measurements were performed in different residential, industrial and commercial urban areas in the north of Sweden. This paper shows existing level of individual harmonics in these locations as a sample of an urban load and compares the harmonic variation between these points. The results showed during this period, characteristic harmonics are well below the standard limits. Correlation matrix is used for investigating possible correlation between individual harmonic orders in this urban area. It is observed that there is a significant correlation between 5th, 7th and 11th harmonic levels of low voltage customers connected to a common 10-kV network.

This paper presents the impact of load impedance uncertainty on harmonic source and transfer impedance and consequently on harmonic propagation in distribution networks. A generic MV/LV model of the network and a more detailed model of an existing network are used. Both models include MV network and remote LV customers. Utilizing a Monte Carlo simulation method, it is shown that variation in load impedance has a big impact on the main resonant frequency of the local network but only limited influence on resonant frequency caused by remote loads connected to other LV networks.

This paper presents the role of the medium-voltage network and low-voltage loads in harmonic voltages with low-voltage customers and harmonic propagation. A general model, as well as a detailed transfer function-based model, are used. By applying them to an existing network, it is shown that remote low-voltage loads have a significant impact on the source and transfer impedance. The main impact occurs for a specific range of frequencies below the resonant frequency of the low-voltage network. The general model is able to estimate this frequency range with an acceptable level of accuracy. Furthermore, by utilizing the concept of overall transfer impedance, it is shown that voltage harmonic levels for harmonic orders around this frequency range are determined mainly by aggregated remote emission rather than local emission.

This Paper proposes an alternative method for defining and aggregating the complex harmonic components (magnitude and phase-angle) in a three-phase system. Instead of using the DFT to define the harmonic components, the Park or dq-transform is used. This results in a natural method for aggregation, for both magnitude and phase angle. The method is evaluated utilizing generated signals and two sets of measured signals as normal case and extreme case.  Some aspects of the method, including a set of indicators, are discussed in the paper.

The harmonic emission from an electric vehicle fast charger depends on factors like charger topology, EV type, initial state of charge of EV battery, as well as supply voltage and background distortion. This paper presents the results from harmonic current measurement of a fast charger for a period of one month in Sweden that has charged a variety of EVs from different brands under different state of charge and background distortion. Besides the common harmonic emission pattern, a high level of variation in emission is observed that can affect the aggregation of the emission from multiple chargers. To include such uncertainties, the harmonic hosting capacity is obtained for a fast EV charging station in a stochastic way. A new method, based on Bayesian statistics and the correlation between harmonic magnitude and fundamental magnitude, is proposed for the generation of stochastic samples. It is shown that the proposed method, to a high extent, can model the stochastic behavior of harmonic emission from a fast charger. Furthermore, the results show that neglecting the correlation between harmonic magnitude and fundamental magnitude can underestimate the harmonic hosting capacity.