To ensure reliable, efficient, and cost-effective operation of distribution networks as well as the continuous supply of power, there is a need to constantly monitor and maintain the networks’ power quality parameters under the desired levels in accordance to the relevant electrical standards [1].

The authors in [2] proposed a methodology to investigate power quality problems of a power distribution network and rank different load sites’ performance in order of mitigation measure requirement. A University campus grid of a school campus was modelled in MATLAB for power quality analysis using NP EN 50160; IES 61000-2-3 and IEEE Std 1159-2019 standards [3]. In [4], an energy audit was conducted for a university campus, it was discovered that Energy consumption was higher than normal due to voltage imbalance and low power factor. A modified IEEE 14 bus model was developed to simulate grid connected photovoltaic in a low voltage region of Liang, it was found that PV integration leads to positive impact on voltage distribution and reduced line losses to certain limit of PV integration [5]. Furthermore, the effect of decreasing power factor on electricity losses of rural low-voltage networks. It was discovered that as power factor decreases from 0.97 to 0.77 electricity losses increased tremendously [6].

The performance of different topologies of passive filters were evaluated for harmonic compensation, where double-tuned filter was identified as the suitable type for reduction of 5^th and 7^th harmonics [7]. In contrary, single-tuned passive filters proved efficient to mitigate individual harmonics and total harmonic distortion in distribution networks [8], as well as in a soap manufacturing industry that uses variable speed drives [9]. In [10], MATLAB was used to evaluate two fixed shunt capacitor banks (SCB) at a 33 kV busbar in a distribution substation for power factor correction, which showed substantial improvement on the bus.

Moreover, an optimization model was developed to analyze voltage quality improvement, several planning projects were ranked in certain area based on voltage improvement and results were promising [11]. In [12], a configuration method coordinated by series voltage regulator and reactive power compensation capacitor based on network admittance matrix was proposed for low voltage control of a rural distribution network.

In addition, a cost-effective algorithm was implemented to optimize the modulation parameters and minimize the total harmonics in an industrial power network [13]. A novel model was developed in [14] to reduce the number of N switching angles required to mitigate N-1 harmonics, and achieve both harmonic elimination and mitigation objectives. In addition, another model to overcome the poor accuracy and generality of traditional methods was also developed and tested [14]. In [15] further results were presented to improve estimation of the harmonic admittance.

In [16], authors used selective harmonic elimination technique was used to eliminate completely N number of harmonics with (N+1) number of switchings per quarter, harmonics up to 17^th harmonic were mitigated with two switchings per quarter. In addition, a smart, efficient, and accurate techniques were developed to estimate harmonics on a distribution network, artificial intelligence was developed and its performance compared with traditional harmonic estimation techniques [17].

Moreover, countermeasure methods; revised JIS C 61000-3-2 (based on IEC 61000-3-2 standard) to suppress the harmonic current outflow of AHPs and installation of LC filters was designed and developed, it was revealed that both can effectively suppress the voltage distortion [18].

The performance of different methods was evaluated for harmonic mitigation, these included Active Power Filter (APF) and Static Synchronous Compensator (STATCOM) with different models of nonlinear loads. It was shown that all higher-order harmonics have been maintained within the standard with STATCOM giving a comparatively better performance [19]. In [20], a study presented the results of a comprehensive power quality analysis of an open-pit mine’s electrical network utilizing the Electrical Transient Analyzer Program (ETAP), the findings indicated the presence of various power quality issues, including harmonic distortion, a power quality improvement plan was proposed.

Possible causes of harmonics in the University campus’ power network which has Solar PV plant of 100 kW include solar inverters, computers, welding machines and aircons. Just like any other electronic equipment, photovoltaic inverters inject harmonics into the electrical installation which leads to overheating. Moreover, personal computers draw non-sinusoidal current causing considerable distortion in voltage and current in the network. Distortion causes variation in voltages which might go beyond acceptable limits of ±10%, resulting in reduction in equipment lifespan and hence malfunctioning. These effects were noticed at the campus that necessitated further investigation.

Top priority of Utility companies is to ensure that power supplied to consumer is of good quality, but Northern Regions Distribution Company (NORED), a distribution company in northern Namibia, has been experiencing a low power factor (PF) averaging 0.2 lagging on its Okangwati distribution network. This falls below minimum acceptable limit of 0.85 Power factor lagging as set by the Electricity Control Board (ECB) of Namibia. Table 1 shows the PF data obtained from the Okangwati Auto-Recloser.

Preliminary investigations showed that- the network was lightly loaded with only 15%, resulting reactive power generation by shunt capacitance thus lowering the line’s efficiency. This also leads to higher receiving end voltage compared to sending end voltage a condition known as Ferranti effect.

The effect of Ferranti on distribution lines (80 km < l > 250 km) and can be explained in detail using the π (pi) model as represented in Figure 1 and eqs. (1-5) [21], [22]. In this paper, an investigation was carried out to determine the optimal loading of the line as will be discussed in IV. Using the pi model, the sending end voltage is determined using eqs. (1-5) [22]:

$V_s=\left(1+\frac{ZY}{2}\right)V_r+Z{I}_r$ (1)

Figure 1.

Pi model presentation of Ferranti effect

Eq. (1) shows the simplified KVL, however, at no load or lightly loaded, Ir is equivalent to zero reducing the equation to eq. (2):

$V_s=\left(1+\frac{ZY}{2}\right)V_r$ (2)

$V_s-V_r=\left(1+\frac{ZY}{2}\right)V_r-V_r$ (3)

$V_s-V_r=\left(1+\frac{ZY}{2}-1\right)V_r$ (4)

$V_s-V_r=\left(\frac{ZY}{2}\right)V_r$ (5)

where Z=(r+jωL) S and Y= jωC.

Resistance is neglected when modelling the Ferranti effect in medium and long transmission lines, hence equation is simplified as in eq. (6):

$V_s-V_r=-\frac{1}{2}\left(w^2{S}^2\right)lc{V}_r$ (6)

Eq. (6) is further sampled to:

$V_s-V_r=-\frac{1}{2}\left(w^2{S}^2\right)\left(\frac{1}{{(3\times {10}^8)}^2}\right)V_r$ (7)

Since, ω=2πf hence eq. (7) is simplified to:

$V_s-V_r=-\left(\frac{4\pi }{18}\times {10}^{-16}\right)\left(f^2{S}^2{V}_r\right)$ (8)

Eq. (8) confirms Ferranti effect, where sending end voltage is higher that receiving end voltage.

NORED is also grappling with elevated levels of power quality issues within its Eheke residential bulk supply network. In a network incorporating hammer mills with power ratings ranging from 15 to 40 kW equipped with variable speed drives (VSDs) and star/delta starters, several power quality challenges may emerge. The utilization of VSDs introduces harmonics into the electrical system due to their nonlinear characteristics, potentially causing increased heating in components and efficiency reduction.

In addition, star/delta starters, employed for motor starting, can lead to current peaks during transition phases, impacting electrical components and possibly inducing voltage sags. Furthermore, the variable speed drives may cause abrupt load changes, resulting in voltage fluctuations that could affect the performance of sensitive equipment. These power quality issues may collectively contribute to mechanical stress on the hammer mills, potentially reducing equipment lifespan leading to consistent failures. The nature of power quality issues caused by hammer mills on Eheke residential supply network and mitigation measures are discussed in IV.

For further analysis on University campus network, power quality analyzer was installed, where electric parameters such as power frequency, THD, individual harmonics, supply voltage limits, voltage unbalance, flicker severity, supply voltage compatibility, and instances of sags, swells, and interruptions were captured. Preliminary investigation was to determine if those parameters fell within acceptable ranges. Additionally, the University campus distribution system was modelled using the ETAP software as shown in Figure 2 for a more detailed analysis of power quality factors. The simulation results on bus 6 connecting Solar PV inverter to the grid is given in Figure 3. The waveform is distorted with high percentage of harmonics.

Figure 2.

The single line diagram for the Engineering campus

Figure 3.

Output voltage waveform at bus 6 before mitigation measures

With further analysis of the system in ETAP and using NRS 048-2:2 017 analyser, it was discovered 5^th harmonics and over voltages in the network as shown in Table 2 were the main cause of power quality problems. 5^th harmonics failed the test by overshooting to 0.46%. This is validated in Figure 4 where is clearly illustrated that 5^th harmonic order is the highest contributor to the harmonic content.

Figure 4.

Voltage harmonics for the Campus

Further analysis was on 3 harmonics. In Figure 5 it can be observed that all the recorded 3^rd harmonics were within the compatibility limits of 5% for the 3^rd week. Figure 6 clearly shows that all the recorded 4^th harmonics order data are within the compatibility limits of 1% for the 4^th week. From Figure 7 it can be observed that for some of the days, the recorded 5^th harmonic data have violated the compatibility limits of 6%, and thus failed the compatibility test.

Figure 5.

Voltage harmonics for the 3^rd week

Figure 6.

Voltage harmonics for the 4^th week

Figure 7.

Voltage harmonics for the 5^th week

Table 3 shows summary of power quality assessment from power analyzer, while Table 4 shows a sustained overvoltage for more than 3 seconds on the specified days. Most severe one was in August 16, and recorded an overvoltage of 209%, for 363,058 seconds on phase 1 and 2 upstream direction. A sustained overvoltage of 209% for 363.058 seconds can have severe and potentially catastrophic consequences on electrical equipment and systems. The prolonged exposure to a voltage twice the designed or rated level may lead to extensive damage to electronic components, therefore their causes should always be identified and appropriate measures taken to mitigate these kind of network problems.

Proposed quality mitigation measures for Network 1.

For mitigating power quality disturbances caused by harmonics, a single tuned filter was designed and installed at bus 6 as shown Figure 8.

Figure 8.

A single line diagram with an embedded single tuned harmonic filter at busbar 6

A single tuned filter consists of RLC components connected in series. To determine the values for these RLC components, eqs. (1)) - (4) were utilized, while also adhering to the IEEE 519-1992 harmonic standard to ensure compliance with the permissible limits [23]. The design of the filter was limited to 5^th order harmonic which was dominant.

Reactive power of the capacitor is calculated using eq. (9). The power factor of the network varies between 0.85 and 0.95 due to load changes over an 11-month period. A power quality assessment was conducted at the JEDS Campus in Ongwediva, Namibia. The lowest recorded power factor was 0.85. It is assumed that the power factor (pfn) can be improved from 0.85 to 0.95 using the filter:

$Q_c=pfn\left(tan(cos-1\left(pf1\right)-tan(cos-1(pf2\right))$ (9)

Eq. (10) is used in calculating capacitive reactance (X_c):

$X_c=V\times 2\times Q_c$ (10)

Using eq. (10) capacitance of the capacitor (C) is determined as given in eq. (11):

$C=\frac{1}{2\pi Q_c}$ (11)

At resonance, the quality factor of the single tuned passive filter is expected to be 100 according to the ETAP modelling tools. The inductance value was determined at resonance, as well as resistance of the filter as given in eq. (12). The values of the RLC filter components used are listed in Table 5. A fine-tuned sinusoidal waveform, Figure 9, shows that harmonics were significantly reduced.

$R=Xn$ (12)

Figure 9.

Output voltage waveform at bus 6 after mitigation measures

For Network 3, investigation of power quality issues on Eheke distribution network with numerous hammer mills utilizing star/delta starters and Variable Speed Drives (VSDs) was done using a portable Unipower UP-2210 logger, the data logger was strategically installed at key locations within the distribution network, particularly near the hammer mills, ensuring secure connections to relevant electrical panels. The logger was then configured to record both voltage and current waveforms over a period of 8 months in 2021.

From THD of individual harmonics, it can be shown that 3^rd, 4^th and 5^th harmonics were above threshold harmonic levels of 4%, 2.5% and 3% respectively, except for March when the THD was 3.900%. This is illustrated in Figure 10, Figure 11 and Figure 12 respectively. It is also noted that harmonic level was low in January, February and March, and this is attributed to the fact that hammer mills are hardly used during this period since it is rainy season and not yet harvesting season.

Figure 10.

Voltage total harmonic distortion

Figure 11.

Peak levels of the third harmonic order observed

Figure 12.

Peak levels of the fifth harmonic order observed

Figure 13 clearly indicates that 3^rd and 5^th harmonic are the highest contributors to THD. Figure 14 shows output voltage for two phases for October 2021, clearly highlighting two power quality issues: phase imbalance and harmonic. This imbalance in harmonic content can result to overload of neutral conductor. Spectral analysis showing harmonics content is as shown in Figure 15. It can be seen that VTHD compatibility level was violated throughout the month.

Figure 13.

The peak levels of the 1^st to 50^th harmonic orders observed in October

Figure 14.

Output voltages for two phases

Figure 15.

Daily VTHD levels for October 2021

Figure 16 shows daily records of 3^rd harmonics. It can be seen that for the whole month, 3^rd harmonics violated the power quality standard of 2.5% confirming Figure 10 which captures average values of 3^rd harmonics for 8 months. Similarly, in Figure 17, 5^th harmonic order levels recorded are above the compatibility level at 3%, confirming Figure 11 which captures average values of 5^th harmonics for 8 months.

Figure 16.

Daily records of 3^rd harmonics for October 2021

Figure 17.

Daily records of 5^th harmonics for October

Table 18 shows the summary for Ehenye power quality parameters and their status, clearly indicating power quality issues with individual harmonics, THD and unbalance test. In addition, there are 35 voltage sags recorded.