WITH THE PROLIFERATION OF ELECTRONIC EQUIPMENT IN MANUFACTURING INDUSTRIES, utilities have to face the risk that permissible quality-of-supply standards could be violated on their networks. This has brought the quality of supply to the top of utilities' strategies in a more structured way than in the past. With the unbundling of the vertically integrated utilities in Egypt into independent companies of generation, transmission and distribution, there is a clear ownership of the origin of the power-quality problems; and as such, there is a need for commercial arrangements to address propagation of power-quality problems from one system to another.

The Alexandria Electricity Distribution Co. (AEDC) supplies the Alexandria, Egypt, area via a 220-kV transmission system, with the city being supplied via 220-kV/66-kV grid substations. The city's 66-kV underground cable network area is mainly direct buried, supplying a number of 66-kV/11-kV indoor substations. The 11-kV distribution system is designed to operate on an open-ring basis, with each 11-kV feeder capable of supplying 8 to 10 medium-/low-voltage (MV/LV) indoor or outdoor packaged substations equipped with 11-kV/0.38-kV transformers that have capacities ranging from 300 kVA to 1000 kVA. Because each MV/LV substation is ring-connected, all supplies can be maintained in the event of a fault in any cable section. The single-phase supply voltage for residential consumers is 220 V.

The underground cables installed, comprising of aluminum conductors and cross-linked polyethylene (XLPE) insulation, were manufactured in Egypt with the following characteristics:

  • Three-core, 11-kV solid-dielectric insulated armored cables with cross-sections up to 240 mm2 (0.37 in2). Larger-sized cables are single core, shielded with solid-dielectric insulation.

  • Armored three-phase, four-core with reduced neutral cables are used for LV.

  • Four-core, XLPE-insulated quadraplex conductors are erected on the LV overhead network.

AEDC currently supplies some 1.5 million consumers — of which 85% are residential, 6% commercial and 0.6% industrial consumers — with the remainder being agricultural and public bodies. In terms of the total annual energy consumption 4.64 million kWh, residential usage is 44%, commercial 9% and industrial 27%.

The needs and expectations of all customers generally focus on:

  • Constancy or reliability of supply and, thus, the absence of interruptions

  • Voltage that only fluctuates within standard limits

  • The absence of voltage dips.

Moreover, the distribution company must ensure that customers' equipment does not generate abnormal disturbances, such as harmonics and voltage flicker, that could be transmitted to other customers via the public distribution network.


AEDC conducted a large-scale power-quality monitoring program to obtain a better understanding of the problems, and to formulate ways to reduce or eliminate their impacts. The most important task for the project, which started in 1999 and continued for three years, was the development of a power-quality database and analysis system. Several aspects of the planned monitoring scheme were considered, such as site selection criteria, measuring equipment specification, database and analysis function.

The measurements were made in compliance with a sustained program, mainly at the points of common coupling in substations and distribution points at the MV level. The program included 15 high-/medium-voltage (HV/MV) substations and 20 MV distribution points. The selected sites had to be a statistically valid set of monitoring locations, so that in the future, any nonmonitored sites' characteristics could be determined from the information recorded at the selected sites.

Power-quality monitors had to be designed to record and process the full range of parameters, unlike the less versatile instruments that solely record flicker or harmonics data. Each power-quality monitor had the following capabilities: a high sampling rate, simultaneous eight-channel monitoring, data capture without the standard threshold setting and adequate memory storage.

These parameters were measured:

  • Voltage magnitude

  • Frequency

  • Voltage sags and swells

  • Flicker

  • Harmonics (total harmonic distortions as well as individual harmonics)

  • Symmetry of the three-phase voltage

  • Short and long interruptions.

To summarize the data (time-series of numbers) collected from each site, histograms were created for each monitored parameter. An important computation made from the distribution is the cumulative frequency value, which is represented as a curve in Fig. 1, where the time-series variation of the voltage flicker at one site is presented as an example. Using a statistical analysis program, the 95th percentile value was then calculated and a histogram of all the computed 95th percentile values was created.

Table 1 represents the computed summary statistics of all data sets, the maximum, minimum, average and 95% values for each parameter. The acceptable level for the Egyptian networks is also included in Table 1.

The results show that the overall average computed voltage deviation is 4.3% of the nominal voltage, 10.5 kV. The maximum and minimum registered voltage did not exceed the 10% Un. The flicker level is analyzed on the basis of both the short-term flicker (Pst) and long-term flicker (Plt). The values of Plt exceed the limit in about 57% of all data sets. The results of voltage unbalance show that the limit of 2% ratio is not exceeded in all of the data sets. Finally, the average voltage THD is temporally distributed for each site, but at 19% of the monitoring sites, the standard limit of 5% was exceeded.

A recently computed summary of power-quality standards showed that the average values were almost identical to those listed as the average percentage shown in the fourth column in Table 1.


In Alexandria, about 70% of power-quality complaints are related to voltage sags, which disrupt customers' activities, leading to production breakdowns, equipment deterioration and lost production. For industrial and commercial customers, voltage sags adversely affect motor-controllers protection and computers in control processes. The frequency of occurrence, recovery times and resulting undervoltages are functions of several factors: the topology of the system, transformer connections, fault rates of adjacent transmission lines, and type and location of the fault.

Variations in the supply voltage waveform, even for a short period, that previously were not a concern can become very expensive in terms of process shutdown and equipment malfunctions. Voltage sags and interruptions are a matter of concern to both customers and utilities. End users are affected by the high economic consequences of these disturbances, as they can interrupt productive processes in the industry and service sectors.

The measurement collection and characterization process of the monitoring project resulted in the development of a voltage-sag-events database and analysis system. To understand the significance of voltage sags, the annual frequency of them is also important when assessing the annual cost of voltage-sag-oriented problems.

Measurements have provided a classification in terms of length and amplitude of the voltage sags and short interruptions observed in the MV network. A large difference in the number of recorded sags exists from site to site; the majority of voltage sags were recorded in the 90% to 85% Un voltage range.

Voltage sag coordination charts, or sag frequency profiles, were produced. Figure 2 shows the cumulative sag frequency for one of the sites, and Fig. 3 illustrates the results for sag and interruption magnitude rates for the events recorded during the project period. The results include all project monitoring sites. Fig. 3 indicates that the frequency of severe sags is minor compared with the frequency of sags whose remaining voltage is greater than 80%.

The voltage sags may not affect customers if the duration is extremely short; however, as the duration of the event increases, the cost due to the lost capacity also increases. Deep voltage sags can cause tripping or malfunction to most types of load, while minor voltage sags do not cause problems for the majority of the customers. The less severe sags are dominated by faults on the transmission system, but all actual interruptions affecting equipment are caused by distribution faults. The study included all voltage sags less than 50% of the remaining voltage and of duration higher than 100 ms. Analysis of the survey results confirmed that the annual rate of sag events whose value is less than 50% Un and duration greater than 100 ms was found to be 3.4 sags. (The annual rate of sags in 2004 was 3.3.) The cost attributable to voltage sag varies among different customer categories and for different industrial processes.

For industry, the following costs are normally included:

  • Product-related costs, such as loss of product, loss of material, disposal of lower-quality product.

  • Labor-related costs, such as idle employees, equipment clean up, overtime cost, repair and restarting.

  • Additional cost due to substitution of damaged equipment, shipping-delay penalties and customer dissatisfaction.

Assessing the economic value of voltage sag for AEDC included the three customer categories: residential, commercial and industrial. The precise evaluation of a customer category sag price is a difficult task, so for the study, the cost figures for short interruptions and sags were considered identical. The approach, called “direct costing,” is widely used by utilities to develop estimates of event costs. The survey requested customers in each category to estimate the costs they would incur from each event. The estimated cost values for voltage sags obtained from the survey are shown in Table 2.

Table 2 gives average values, which are based on combining information provided by the customers from each category. Table 3 shows the total number of customers fed from the networks of the monitored sites for the three categories of loads that were studied.

The annual cost of voltage sags of each customer type was calculated based on:

  • Customer numbers

  • Event price

  • Frequency of events recorded.

Figure 4 shows the comparable results determined from the product of these three parameters. The results show that voltage sags and short interruptions in the surveyed part of the network cost more than 50 million LE per year (US$10 million per year). The bulk of this estimated loss, 27.8 million LE (US$5.6 million), is concentrated in the industrial sector, which is particularly vulnerable to equipment damage. The commercial sector loses an estimated 21 million LE (US$4.2 million) to voltage sags annually, primarily from lost productivity and idled labor. The costs related to the residential sector are relatively insignificant.


AEDC needs to develop a program that is able to balance the economical cost of improving the power quality at source and limit the pollution caused by disturbing industrial loads. The technology of MV compensating devices has proven to be successful in compensating voltage dips. However, the solution is expensive because of switchgear cost; it is therefore expensive to protect the entire plant using such devices. It may be that only selected critical loads require protection and in some instances — as the available solution may be cost prohibitive — the answer may be to do nothing.

AEDC has started taking corrective action in the industrial sector, including the installation of power conditioners, surge suppressors, uninterruptible power supplies and tuned filters at selected loads in different facilities. The financial savings associated with the installation of these mitigation devices are directly related to the total cost of power-quality events avoided. The devices have the capability to mitigate a percentage of these events, which otherwise would cause a partial or total shutdown. Therefore, the number of avoided shutdowns multiplied by the estimated cost of a production interruption gives the estimated annual savings. If the device installed also performs other functions, besides sag/interruption compensation, which provide economic benefits to the plant (e.g. power factor improvement, harmonic reduction), additional savings have to be considered. The costs associated with the mitigation devices are termed “initial” and “operational.”

Initial costs comprise the cost of the device, installation cost and cost of power system improvement, if necessary, such as the addition of a redundant feeder for installation of a static transfer switch. Operational costs comprise losses and maintenance. Thus, each individual solution has been evaluated in terms of the cost and performance improvement.


AEDC gained valuable experience in establishing power-quality-monitoring strategies, and shows the utility's commitment to improve power-system quality for the benefit of its customers, suppliers and equipment manufacturers. The primary goal of the project was to develop baseline data and analysis technique describing quality levels in the systems. AEDC commissioned the project to obtain a definitive estimate of the direct costs associated with power disturbances and obtained these data by surveying customers in key sectors on the costs resulting from voltage sags and short interruptions. Data was also captured on the number and duration of disturbance events experienced. Examination of the system reliability in terms of interruption frequency and duration indicates an improving system performance as shown in Table 4.

Disturbance costs vary with the magnitude and duration of the event, and AEDC's study revealed the high annual cost attributable to system disturbances, the largest losses being experienced by the industrial consumer group. This project has given AEDC the key cost data it needed to take into consideration when seeking compliance with power-quality standards, which always requires an economical balance between the utility's cost to improve the power-system quality and the consumer's cost to limit the pollution created by the disturbing load connected to the power system.


This article is based on a paper presented at the CIRED 2003 17th International Conference & Exhibition on Electricity Distribution, held in Barcelona, Spain, on May 12-15, 2003, and published in the AIM Conference Proceedings.

Nazineh G. Eassa graduated from Alexandria University Faculty of Engineering in 1975, and received her MS degree from the Electrical Power Department of Alexandria University Faculty of Engineering in 1980. Her career experience includes the modernization of Alexandria's distribution network in cooperation with the USAID projects, the Alexandria Electricity Distribution Co. — Research and Quality of Power department. Her special fields of interest include monitoring and observations of power-quality problems on high-, medium- and low-voltage distribution networks. nazineh99@yahoo.com

Dr. Ahmed Abouelseoud graduated from Alexandria University Faculty of Engineering in 1993, and earned his MS degree and Ph.D. from the Electrical Power Department of Alexandria University Faculty of Engineering in 1999 and 2004, respectively. His employment experience includes responsibilities in the Electricity Distribution Co. - Research and Quality of Power Department. His special fields of interest include the development of new research work for enhancing the performance of power-quality supply in medium-voltage distribution networks. abouelseoud@yahoo.com

Table 1. Overall Values for all Monitoring Sites (Minimum, Maximum, Average and CP95)
Power-Quality Issue Minimum (%) Maximum (%) Average (%) CP-95 (%) Standard Limits (%)
Voltage deviation 1.6 8.8 4.3 7.7 10
Voltage unbalance 0.08 0.8 0.33 0.65 2
Frequency deviation 0.05 0.29 0.13 0.22 1
Short-term flicker (Pst) 0.135 2.7 1.27 2.1 1
Long-term flicker (Plt) 0.14 1.4 0.71 1.2 0.8
Voltage THD 0.7 5.8 3.09 4.7 5
Table 2. Estimated Cost of Sag Event
Type of Customer Cost/Event (LE)
Residential 1
Commercial 84
Industrial 1510
Table 3. Load Category and Surveyed Number of Customers
Residential 390,734
Commercial 73,661
Industrial 5417
Table 4. System Interruption Statistics
Year Interruption Time/Consumer Per Year (hr) Number of Interruptions/Consumer Per Year
2002 6.1 5.7
2003 5.8 5.0
2004 5.3 4.7