PQ solution costs are to be evaluated case by case, but it is possible to provide some useful statistical data based on surveys carried out in a large number of cases.
    Figure 18.15 charts the proportion of load per sector covered by different types of redundant or mitigating solutions.
    The analysis of these solutions produced some interesting conclusions. Many of the correlations between solutions (both investment and load coverage) and PQ cost, frequency of events or sensitivity to PQ problems, which were thought to have been significant, have not been proven. Table 18.11 presents all significant relations, where 0.05 is used as reference threshold between PQ consequences and PQ solutions as confirmed by surveys.
    Figure 18.16 shows a certain relation (although not proven by the linear regression model; R2 linear = 0036) between PQ investment and experienced PQ cost.


 Figure 18.15 PQ solutions – installation coverage in % [11] (Reproduced from the 2007 Leonardo Power Quality Initiative Survey, R. Targosz)
    Basically the slope angle suggests a positive relation between PQ investment and PQ unmitigated cost.



Table 18.11 PQ consequence/PQ solution correlations [11]

 
        Figure 18.16 PQ cost/PQ solution investment relation [11] (Reproduced from the 2007 Leonardo Power Quality Initiative Survey, R. Targosz)
Figure 18.17 Mitigated and unmitigated PQ cost per unmitigated (real) PQ cost ratio as a function of PQ solution investment [11] (Reproduced from the 2007 Leonardo Power Quality Initiative Survey, R. Targosz)
    Although there is no significant correlation between solutions and real cost, a strong correlation exists between investment in PQ solutions and the hypothetical to real cost ratio.
    This results in an indirect but clear link between solutions and (real) consequences. See Figure 18.17.
 Figure 18.18 Occurrence of equipment affected by PQ in annual % [11] (Reproduced from the 2007 Leonardo Power Quality Initiative Survey, R. Targosz)
    The following broad conclusions can be drawn:
    ? The increase in the ratio between hypothetical and real (mitigated/unmitigated) is very visible in the case of UPS.
    ? One side effect of UPS use is the increased cost of harmonics. This can be explained by suboptimal use of UPS systems that are based on diffused small units without active power wave modulation, which in turn generates significant input current distortion.
    ? There is a small but significant (positive) correlation between number of power lines and costs of short interruptions, whilst such a correlation is insignificant in so far as dips are concerned.
    The study provided a number of additional conclusions regarding the occurrence of PQ problems, their sources and the equipment affected by them.
    The occurrence of different equipment being affected by PQ is presented in Figure 18.18:
    ? Electronic equipment is most affected in the industry and service categories.
    ? Static converters and electric motors are the next most affected.
    ? All other equipment types are more evenly affected in the services category.
    Below are some additional findings from the survey.
    ? The perceived level of presence of different PQ disturbances for all sectors is presented in Figure 18.19 and varies quite noticeably.
    ? The semiconductor respondents did not specify experiencing long interruptions, though they did record very intensive occurrence of voltage dips and short interruptions.
    ? For all sectors, on average, the presence of short interruptions is perceived as being the most intensive and disruptive.
    ? The same differences in perception between the industry and services categories also apply to the consequences of poor PQ (see Figure 18.20) and amount to:
    – Loss of synchronization of processing equipment, which is very common for continuous manufacturing and caused industry considerable problems for its activity.
    – Lock-ups of computers and switching equipment tripping were the second most problematic.
    – As far as services were concerned, circuit-breakers tripping and data loss cause the greatest problems.
    – According to survey [11], respondents affirmed that electric shocks are not relevant to the PQ issues investigated.
    ? The main sources or causes of PQ problems, see Figure 18.21, are defined as follows:
    – Motor-driven systems and, in general, static converters are the main sources of PQ problems for industry.
    – Electronic equipment and components are the equivalent main source for the services category.
    ? Regarding PQ solutions, Figure 18.22 presents the preferences of the two categories, industry and services.




Figure 18.19 Presence (perceived) of PQ disturbances [11] (Reproduced from the 2007 Leonardo Power Quality Initiative Survey, R. Targosz)
    – Both specify UPS most frequently.
    – Backup generators, which prove to be most effective in the case of long interruptions, are dominant in services.
    – Harmonic mitigation through harmonic filters is reported at 45% to 65% of the frequency.
    – For industry, passive filters are almost three times more popular than active filters.
    – For services, active filters are more popular but the difference is small.
    – In general, services apply a higher frequency of different PQ solutions than found in industry.
    – Industry tends to favor less costly, less universal solutions whenever possible.




Figure 18.20 Frequency of PQ consequences as % of cases [11] (Reproduced from the 2007 Leonardo Power Quality Initiative Survey, R. Targosz)Figure 18.20 Frequency of PQ consequences as % of cases [11] (Reproduced from the 2007 Leonardo Power Quality Initiative Survey, R. Targosz)


    Figure 18.21 PQ problem source as % of cases [11] (Reproduced from the 2007 Leonardo Power Quality Initiative Survey, R. Targosz)



Figure 18.22 PQ solutions applied as % of cases [11] (Reproduced from the 2007 Leonardo Power Quality initiative Survey, R. Targosz)
    ? Looking at where the fault for poor PQ resides, in general the blame is usually placed at the foot of external causes. See Figure 18.23.
    ? Within that general statement, services more frequently admit that their installation could be the source.
    ? PQ measurement was of great concern because the survey [11] identified a different level of measurement of PQ parameters. Consequently the implication is that there exists an unequal level of understanding of and acceptance for the need for power-critical users to ensure consistently good PQ. Figure 18.24 presents the feedback to two questions – the ability to identify the sources of PQ events (the first four bars per category of the chart) and their frequency (the remaining eight bars per category) and the continuous monitoring of the key PQ parameters that further diagnose these issues:
     – For the identification data set, the average response across all information sources was 50% – a level which, to repeat, is significantly low for industrial sectors that depend on good PQ. Within that, rather surprisingly the services’ direct PQ measurement is much more frequent than that occurring in industry, where PQ data gathering is more reliant on the different PQ data acquisition components installed in its power systems.
 Figure 18.23 Poor PQ responsibility: 0, no; 4, high extreme [11] (Reproduced from the 2007 Leonardo Power Quality Initiative Survey, R. Targosz)
    – Concerning continuous monitoring of key PQ parameters, this is more prevalent in industry than in services. Both reactive power and flicker are subject to continuous measurement several times more frequently by industry than by services. In 70% of the industry cases, reactive power is subject to continuous measurement and this could be for financial reasons when reactive power is likely to be subject to separate accounting procedures.
    ? For the case of flicker, a high proportion of the companies interviewed in the survey (46 out of 62) agreed that flicker generates PQ costs in terms of losses generated in employee efficiency, which can amount to 10% of annual employment cost.
    These costs are related to vision problems with symptoms like fatigue and increased error rate.
    These consequences relate to reduced productivity/inefficiency in work and in extreme cases to employee compensation. These costs amount to E167m, which is equivalent to approximately 1.5% of all hypothetical (mitigated) and real (unmitigated) costs. As thisis an area of current and as yet inconclusive debate, and although respondents affirmed that their employees’ efficiency was reduced by the levels of flicker experienced, the flicker cases at this stage have all been treated as hypothetical.
    Finally, and in addition to the summary of these technical findings, as was stated earlier but merits repeating, it is astounding that industrial sectors, for which electric power is critical, are not fully aware of these issues.


   Figure 18.24 PQ monitoring: four left bars, source of PQ event information; remaining bars, measured PQ parameter [11] (Reproduced from the 2007 Leonardo Power Quality Initiative Survey, R. Targosz)
    The main conclusion, however, remains that PQ costs in Europe are responsible for a serious reduction in industrial performance with an economic impact exceeding E150 bn.