Power quality disturbances can have significant financial consequences to different customers and the network operators. It is quite hard to estimate correct financial losses of poor PQ as many uncertainties are involved. Therefore field surveys, interviews and case studies are carried out to get an indication of the costs of poor PQ.

From literatures, many analyses are found on PQ costs for various types of customers. In contrast, very limited information is available on PQ cost for the network operators. As the cost evaluation of poor PQ is a complicated issue, the CIRED/CIGRE ‘Joint Working Group’- JWG C4.107 was formed to develop a systematic approach for estimating various costs related to PQ problems. This group proposed methodologies to determine PQ costs for the customers as well as the network operators.

Problem For Customers
From worldwide customer surveys on the electric supply, it is found that voltage dip is one of the PQ problems that causes large inconveniences and has significant financial impacts to various industrial process equipments. The actual financial losses are customer specific and depend mainly on customer category, type and nature of activities interrupted and the customer size.

Also, financial losses are event specific and different severity could incur different losses to various customers. It is noticed from different surveys that short interruptions and voltage dips are the major contributors to financial losses in terms of PQ related costs.

The European Power Quality survey report declared that PQ problems cause a financial loss of more than 150 billion Euros per year in the EU-25 countries [Targosz & Manson, 2007]. The survey was done over two years period during 2003-2004 among 62 companies from different industries and service sectors.

It was found that 90% of the total financial losses are accounted to the industries. Fig. 8 shows the percentage shares of total financial losses on various PQ aspects in the EU-25 countries. It shows that 56% of total financial loss in EU-25 is a result of voltage dips and interruptions, while 28% of the costs are due to transients and surges. Other financial losses (16%) are because of harmonics, flicker, earthing and EMC related problems.

As per the proposal of the CIRED/CIGRE JWG 4.107 group [Targosz & Manson, 2007], two distinct methods of measuring the economic impact of poor PQ have been identified.

• The first method is a direct method, which is an analytical approach to consider the probabilities and impacts of the events. This method leads to a precise answer, but mostly it is difficult to obtain correct input values.
• The second method is an indirect method, which considers historical data for analysis and the customer’s willingness to pay for solving PQ problems.

Total cost of a PQ disturbance for a production company consists of expenditures in various accounts as follows:

• Staff cost – this is the cost because of personnel rendered unproductive for disrupted work flow.

• Work in progress – this category includes the costs of raw material involved in production which is inevitably lost, labour costs involved in the production, extra labour needed to make up lost production etc.
• Equipment malfunctioning – if a device is affected, the consequences can be slow down of the production process, extra ‘idle’ time.
• Equipment damage – if an equipment is affected, consequences can be complete damage of the device, shortening of its life time, extra maintenance, need of stand-by equipment etc.
• Other costs – the costs paid for penalties due to non-delivery or late delivery, environmental fines, costs of personal injury (if any), increased insurance rate etc.
• Specific costs – this category includes extra energy bill due to harmonic pollutions produced by non-linear devices, fines for generating harmonic pollution in the network (if applicable). Reduction of personal working efficiency; related health problem due to flicker can also be included in this cost category.
• Savings – there are some savings in the production too. It includes saving from the unused materials, saving from the unpaid wages, savings on energy bill etc.

In a typical continuous manufacturing sector large financial losses are incurred by the lost work-in-progress (WIP) which is (most of the cases) about one third of total PQ costs. Also, the slowing down of processes and labour costs are quite significant in this sector.

In other sectors, the situation is not very clear with the labour cost and equipment related costs. In the public services like hotels and retail sectors, PQ impact is measured as slowing down their business activities, in terms of revenue lost.

In the industries the losses are mainly because of voltage dips, interruptions and transient surges. Fig. 9 shows the distribution of PQ costs in various accounts in industries and service sectors, as estimated in the LPQI survey for the EU-25 countries [Manson & Targosz, 2008].

PQ cost estimation survey was also performed by the EPRI and CEIDS consortium for the American industries in 2000. It was estimated that the US economy loses annually 119 billion dollars to 188 billion dollars due to voltage dips, short interruptions and other PQ problems [Lineweber & McNulty, 2001]. 

Digital economy and continuous manufacturing industries are found to be the most affected sectors. It can be remarked here that PQ cost data, obtained from different surveys, is quite difficult to compare as the references of representations in different surveys often vary. 

Hence, proper evaluation method of the analysis is required for correct interpretation of the cost data. Another report [McNulty et al., 2002] estimated the costs of momentary and 1 hour outages for various sectors in the USA. Similar type of survey was also conducted by UMIST, UK in 1992 [Kariuki & Allan, 1996] to estimate costs of outage to different customer groups. 

Table 3 compares both the findings of these surveys. It shows that outage costs in different sectors in UK and US vary significantly, except for the industrial customers suffering momentary outages.

It is quite difficult to make a general conclusion on financial losses in different industries as the PQ cost and the cost of outage due to interruption depend largely on the customer's installation characteristic and the devices involved.

Among the industries, there can be a wide range of variety in device usages and their sensitivity to PQ problems. The same is also applicable for the commercial sectors.



CBEMA and ITI Curves
CBEMA curve  is one of the most frequently employed displays of data to represent the
power quality. A portion of the curve adapted from IEEE Standard 4469 that we typically use in our analysis of power quality monitoring results is shown in below.

In the CBEMA Curve, the axes represent magnitude and duration of the event. Points below the envelope are presumed to cause the load to drop out due to lack of energy. Points above the envelope are presumed to cause other malfunctions such as insulation failure, overvoltage trip, and overexcitation.

The upper curve is actually defined down to 0.001 cycle where it has a value of about 375 percent voltage. We typically employ the curve only from 0.1 cycle and higher due to limitations in power quality monitoring instruments and differences in opinion over defining the magnitude values in the subcycle time frame.

Computer equipment sensitivity to sags and swells can be charted in curves of acceptable sag/swell amplitude versus event duration. This curve was originally developed by CBEMA (Computer Business Equipment Manufacturers Association) to describe the tolerance of mainframe computer equipment to the magnitude and duration of voltage variations on the power system.

While many modern computers have greater tolerance than this, the curve has become a standard design target for sensitive equipment to be applied on the power system and a common format for reporting power quality variation data.

In the 1970s, the (CBEMA) developed the curve [2.5] of Figure 2.5 utilizing historical data from mainframe computer operations, showing the range of acceptable power supply voltages for computer equipment.

The horizontal axis shows the duration of the sag or swell, and the vertical axis shows the percent change in line voltage.

In addition, the IEEE has addressed sag susceptibility and the economics of sag-induced events in IEEE Std. 1346–1998 [2.6]. This document includes measured power quality data taken from numerous sites.

The original CBEMA curve was originally developed by the Computers Business Equipment Manufacturers Association (CBEMA) and adopted by IEEE Standard 446. CBEMA has been renamed as the Information Technology Industry (ITI) Council, and a new curve as shown in Figure 1.1 has been developed to replace the original CBEMA curve.

The modified curve has been developed that specifically applies to common 120-V computer equipment. The concept is similar to the CBEMA curve. Although developed for 120-V computer equipment, the curve has been applied to general power quality evaluation like its predecessor curve.

Outside the bounded tolerance region, in the no-damage region, the applied voltages are very low, and sensitive computer equipment will not function properly; however, no damage occurs to the equipment.

In the prohibited region, sensitive computer equipment will be damaged due to the occurrence of severe voltage swells.

Both the CBEMA and ITI curves were specially developed for use in the 60-Hz 120-V distribution voltage system. The guidelines expect users in 50-Hz 240-V distribution systems to exercise their own judgment when applying the CBEMA and ITI curves.

Although there is no legal requirement to conform to these curves, most original equipment manufacturers build equipment that meet or exceed the limits set forth by these curves, with the occasional exception.

That is the value of CBEMA and ITIC Curves.



Poor power quality is usually identified in the “powering” part of the definition, namely in the deviations in the voltage waveform from the ideal. A set of waveforms for typical power disturbances is shown in Figure 1.5. These waveforms are either (a) observed, (b) calculated, or (c) generated by test equipment.

The following are some examples of poor power quality and descriptions of poor power-quality “events.” Throughout, we shall paraphrase the IEEE definitions.

■ A voltage sag (also called a “dip”9) is a brief decrease in the rms linevoltage of 10 to 90 percent of the nominal line-voltage. The duration of a sag is 0.5 cycle to 1 minute [1.44–1.50]. Common sources of sags are the starting of large induction motors and utility faults.

■ A voltage swell is the converse to the sag. A swell is a brief increase in the rms line-voltage of 110 to 180 percent of the nominal line-voltage for a duration of 0.5 cycle to 1 minute. Sources of voltage swells are line
faults and incorrect tap settings in tap changers in substations.

■ An impulsive transient is a brief, unidirectional variation in voltage, current, or both on a power line. The most common causes of impulsive transients are lightning strikes, switching of inductive loads, or switching in the power distribution system. These transients can result in equipment shutdown or damage if the disturbance level is high enough. The effects of transients can be mitigated by the use of transient voltage suppressors such as Zener diodes and MOVs (metal-oxide varistors).

■ An oscillatory transient is a brief, bidirectional variation in voltage, current, or both on a power line. These can occur due to the switching of power factor correction capacitors, or transformer ferroresonance.

■ An interruption is defined as a reduction in line-voltage or current to less than 10 percent of the nominal, not exceeding 60 seconds in length.

■ Another common power-quality event is “notching,” which can be created by rectifiers that have finite line inductance. The notches show up due to an effect known as “current commutation.”

■ Voltage fluctuations are relatively small (less than 5 percent) variations in the rms line-voltage. These variations can be caused by cycloconverters, arc furnaces, and other systems that draw current not in synchronization with the line frequency [1.51–1.61]. Such fluctuations can result in variations in the lighting intensity due to an effect known as “flicker” which is visible to the end user.

■ A voltage “imbalance” is a variation in the amplitudes of three-phase voltages, relative to one another.



According to rectifying or inverting operation of HVDC converters, reactive power is absorbed  from the bus in which the converter is connected. In either case of operation reactive power  compensation in AC side of converters is quite necessary. In addition to reactive power  compensation, due to nonlinear behavior of power electronics converters, considerable  characteristic and uncharacteristic harmonics are produced in both sides of links and often they are  filtered by passive and active filters. 

There has been a great growth in application of AC/DC links  and therefore the harmonic reduction and reactive power compensation method should be  improved but this increases the complexity in hardware and control strategies and also increases  the total cost of links. In this paper the new switching patterns of capacitance for achieving  continuously controlled compensation and reduction of harmonics produced by HVDC converters  are described. This method has a simpler structure and easier switching control strategies  compared to active filter configurations.  

The rapid development of power generated by increased demand for electric energy initially in industrialized countries and subsequently in developing countries led to different technical problems in the systems such as stability limitation and voltage problems.

However breaking advances in semiconductor technology then enabled the manufacture of powerful thyristors and later other elements such as the gate turn off thyristors and insulted gate bipolar transistors. High voltage DC transmission (HVDC) technology which is being considered as an alternative to long distance AC transmission is based on this development [1].  

Harmonics problem generated by nonlinear loads and thyristor converters becomes increasingly serious as 
they are widely used in industrial applications and transmission and/or distribution systems. Since the 
HVDC converters are large power converters, they have become important harmonic sources in power 
systems and without proper compensation the quality of power in system is deteriorate. 

So far, the shunt passive filters due to their low cost and high efficiency have hitherto been used to reduce harmonics in power systems. However shunt passive filters have many problems to discourage their applications. As shown in fig. 1, the filtering performance of passive filters is influenced by the ratio of equivalent impedance of AC source link side and passive power filter impedance. 

Since the source impedance is not accurately known and varies with the system configuration, strongly 
influences characteristic of shunt passive filter. Furthermore the passive filter may fall in series resonance or in parallel resonance with source impedance. 

Shunt active filters using PWM inverters have been developed as  the solution of preceding problems in passive shunt filters [3]. In the beginning, shunt active filters were proposed to suppress the harmonics generated by large rated thyristor converters used in HVDC transmission systems.