Example System

Figures 8.1 and 8.2 depict the plight of a fictional electric distribution utility, Metropolitan Power and Light, which represents the average experience of eight major metropolitan utilities on the east and west coasts and the Midwest United States.

In the period 1965 to 2000, Metropolitan Power and Light saw the peak demand in its central operating district go from 1380 MVA to 3058 MVA, an average annual growth rate of 2.3%. But while the peak load and demand for quality of service have increased greatly, the basic topology of its system did not keep pace. The original (1965) system, shown in Figure 8.1, consisted of 16 substations in the central operating district, with an average capacity of 128 MVA each, for a total installed substation capacity of 2048 MVA. This resulted in a utilization ratio of 67% (1380 MVA/2048 MVA). The majority of these substations were originally fed by 69 kV transmission loops (two, or in some cases, two sets of two), lines along the same right of way, providing a parallel and redundant (contingency backup) feed to each substations.

During the intervening 35 years, Metropolitan added one new substation site and expanded the size of four others. It also upgraded a number of 69 kV sub-transmission lines to 138 kV and completed one “loop” transmission route by adding a right of way to connect two substations (Figure 8.2). In addition, it increased the capacity of the 138 kV circuit lines in the backbone right of way through downtown by re-conductoring and bundling of conductor, and by upgrading part of the route to 345 kV.

Table 8.1 Comparison of Metropolitan Power and Light Downtown Systems in 1965 and 2000

Figure 8.1 In 1965, the central operating district of Metropolitan Power and Light used 16 substations averaging 128 MVA capacity each (total 2,048 MVA capacity) to serve a peak downtown load of 1,380 MVA. This diagram shows only the overall structure of the system. Many of the apparently radial lines feeding substations are in fact loops (two circuits) or double loops (four circuits).

Figure 8.2 (2000) Thirty-five years of slow but steady load growth result in a year-2000 peak demand of 3,058 MVA, more than double 1965’s. In that time, Metropolitan Power and Light has added only one substation site and expanded the area of four others, as well as upgraded various 69 kV lines to 138 kV lines. Total substation capacity now stands at 3,400 MVA, (200 MVA each, a 56% increase), for a utilization ratio of 90%.

However, this system growth did not keep pace with peak load growth. Year 2000 capacity is 3,400 MVA, an average of exactly 200 MVA per substation, giving a utilization ratio against the 3,058 MVA peak load of 90%. Utilization ratio of the transmission lines has similarly increased. Equally indicative are the facilities ratios, load/substation and load/transmission route, as shown in Table 8.1. Metropolitan Power and Light has a lot of eggs in very few baskets.

At first glance it might appear that the system in Figure 8.2 is fully as capable of providing high reliability of service as the system in Figure 8.1. In actuality, the system is more vulnerable to certain types of outages at the sub-transmission and sub-station levels. This increased vulnerability is due to the following factors: 1) higher utilization ratio; 2) “a lot of eggs in one basket and; 3) design choices dictated by limited options for facilities.

Why Did This Happen?

Before turning to a discussion of the impacts these changes had on the cost and reliability of electric service, it is worth looking at why Metropolitan got into this situation. Over a period of three and one half decades, as its peak demand grew, its system did not keep pace in some ways, out of both choice and constraint. This did not occur because the utility planners did not forecast the growth: for the most part they saw it coming. It happened because, incrementally, the utility allowed itself to be backed into the situation. The major reasons were:

Other Options

Every time that Metropolitan Power and Light needed a new site, there were always other options available to the utility. Among them were reinforcement re-design and re-configuration of existing facilities, and other means that could provide capacity and contingency capability. Many of these were only “stop-gap” measures to defer the need for new sites for a few years, but in every case they had a lower first cost and met the existing criteria. While a very long-term view would have shown that a few new sites were necessary, the paradigm at work led Metropolitan to take other routes in nearly every case.

A contributing factor to these decisions was the characteristics of the traditional N–1 planning criteria and tools Metropolitan’s planners used to apply it, as will be discussed in Chapter 8. The utility planners’ tools did not alert them to the weaknesses that were developing in their system. The limitations of the “outdated system structure” they were evolving were not completely visible to them. As a result they were expecting better performance from the system than it could actually deliver.

Incremental Mistakes and Boiled Frogs

Traditional anecdotal wisdom holds that if one throws a frog into a bowl of very hot water it will immediately hop out. But, if that same frog is placed in tepid water, which is then slowly heated over a fire, it will float in the water until it boils to death. The authors have not verified that this is true (it seems unusually cruel to the frog and the authors have no desire to have a boiled frog on their hands). However, it is a good analogy to the way many utilities, including Metropolitan, gradually backed into a situation that they never would have accepted had they foreseen it. Certainly the situation with regard to substation sites was very gradual, with no one incremental event ever so dramatic as to trigger concern.

Another example of this is the gradual erosion of feeder system strength seen by many utilities in the 1980s and 1990s. Section 8.4 will discuss in detail the impact that the period 1965 to 2000 saw with respect to Metropolitan Power and Light’s primary feeder system. As the load grew and the number of substation sites did not keep pace, the feeder system had a gradually increasing burden placed on it. From 1965 to 2000, the load it had to distribute out of every substation grew by 109% (86 MVA in 1965, 200 per substation in 2000). At times this necessitated adding new feeders – Table 8.1 shows that the number of feeders grew by 47%. But to a large extent the load growth was accommodated by allowing the loading (utilization rate on feeders) to escalate.

Table 8.1 showed that average feeder loading increased by 50% during the period; an annual compounded growth rate of 1%. Such a small annual increase is easy to lose among the annual fluctuations in loading due to weather and load transfers, and never enough to raise alarm. This gradual increase in utilization of feeders eventually degraded much of the feeder system’s potential to provide high quality customer service, as will be discussed in Section 8.4.

Metropolitan Power and Light’s inability to see fully the impact of increased feeder loading on its system reliability was, as in the cases cited earlier in this section, partly a failure of its planning tools and engineering paradigm. Traditional distribution planing tools focused on “feeder at a time” short range analysis, techniques which cannot fully assess the feeder system impacts that increasing loadings and reduced contingency reach have on the emergency-backup capabilities among feeders.¹ These tools fit the engineering and planning paradigm for distribution planning, which was to examine feeders in detail, but usually only on a one-at-at-time or small-area basis. Feeder system studies involving all the feeders in a region, modeled in company with the substations and sub-transmission, were never done. Therefore, Metropolitan's planners never realized that the degradation of feeder-level reliability, amounting to about a 25% increase in contribution to SAIDI, was occurring. Intuition and judgment alone did not alert the planners to the problem largely because of the very slow rate of change.

Poor Planning was a Major Contributor

The discussions above twice cited the limitations of traditional planning tools as a contributing factor in Metropolitan Power and Light’s getting into the outdated system structure situation without fully appreciating its implications with respect to reliability of service. Also cited were Metropolitan Power and Light’s lack of effectively applying long-range planning – of using too short-term a focus and not looking at the long-term implications and costs of a de facto policy of constant deferrals and opposition to any major projects.

A longer-range view of its system, even without modern planning tools, would have revealed that problems were going to occur. Even without the best planning tools, such a perspective would have drawn attention to the issue, at least qualitatively. In addition, a longer planning horizon would have permitted Metropolitan Power and Light to identify potential substation sites farther ahead and perhaps obtain them when they were affordable (or at least, less pricey).

Thus, ineffective planning – done with tools of limited capability, and on too short a timeframe – was a contributing factor in Metropolitan Power and Light getting into the situation that in found itself in as the new century began.

Site Congestion

Metropolitan Power and Light was unable to get either as many new substation sites as it needed, or to obtain expansion space it could have used at many of its existing substations. As such, many of the substations in the central operating district became congested, in the sense that space was completely used, and a lack of available space was the limiting factor on installed equipment at most sites. This lack of space caused two interrelated design compromises.

Higher Equipment Utilization

In spite of the efforts to often choose capacity over configuration, during the last third of the 20th century capacity growth did not keep pace with peak load growth in the Metropolitan Power and Light example given here, or in most large metropolitan systems in the United States. A primary effect of the changes in this case was the increase in utilization ratio, from a ratio of peak load to capacity for substations of 67% in 1965, to 90% in 2000. Again, as discussed in detail in Section 8.2, this occurred primarily because the utility could not get enough new sites, or enough room at existing sites, to provide all the capacity it might like. A contributing factor was management’s commitment to reduced capital spending, which favored increasing utilization of existing equipment anyway.

Regardless, an increase in utilization ratio has a tremendous negative impact on expected reliability of the system, one, which can be completely mitigated by a combination of artful engineering changes to the structure of the system (interconnection and configuration of switching) and operating policies. However, if not accommodated with such changes the increased utilization ratio will typically result in degraded reliability of service to customers.

In 1965, each substation transformer or line was loaded during peak periods to about 66% of its rating. This meant that if a substation transformer, or bus, or breaker fails, that 67% loading can be distributed over two neighbors (of equal capacity) who each take half, raising their loading during the contingency from the normal design peak of 67% to 100%. If brief overloads of equipment to 133% of rating are accepted during a contingency, the failed unit’s load can be fully supported by only one neighboring unit. Peak load occurs only a few hours per year (typically about 350) so this high level of stress is expected to occur rarely (only when peak load and a contingency occur). Off peak, the outage of even two units could often be supported by temporarily overloading only one remaining unit of equipment.

But in a power system with 90% loading, during peak load conditions a failed transformer or line now needs nine neighboring units, not two, to support its outage if contingency loading is limited to 100%. (The 90% loading on the unit has to be distributed among nine neighbors, who each have only 10% capacity margin to aid the situation.)

If equipment overloads are accepted during contingencies, the situation is still worse than before. At a 135% overload, two neighboring units are required to support the outage of the failed unit, not one as before. Table 8.2 shows required “contingency support count” for various combinations of utility utilization ratio (designed peak load/capability ratio) and contingency overload limit. It plots the number of units required for contingency support as a function of utilization ratio, for systems composed of identical capacity units.

Utilization Ratio Is Not the Culprit

High utilization ratio, per se, is not necessarily an undesirable element in power systems. Reliable and economical power systems with peak utilization ratios up to 100% can be designed and operated. However, the rise in utilization ratio must be accompanied by changes in the system’s ability to isolate outaged equipment and interconnect during emergencies to available contingency margin. The requirement for interconnection capability and flexibility rises along with system utilization ratio. Configuration must in some sense provide the strength that capacity margin does not. Increasing the utilization ratio without having sufficient configuration flexibility to accommodate it, is a mistake that will lead to deteriorated reliability of service.

Feeder Level Reliability

The primary voltage feeder system plays an essential role in the power system chain, taking the power from the substations to the vicinity of all customers. As such, its ability to perform its function in a reliable manner – to recover from outages of circuit elements and equipment – is essential to a good reliability record for the utility. Aging equipment, cables, and conductors in the feeder system all greatly impact reliability.

In almost all utility systems, the majority of customer service reliability problems is caused by problems on the distribution system, either the primary feeder system or the service-level equipment it feeds. Figure 8.5 shows typical operating results – about 45% of all customer outage minutes are due to problems that occurred on the primary feeder system, nearly twice as much as on the entire system-transmission-substation level.

The fact the feeder system is the major source of service interruptions in a power system is not the least surprising. In fact it seems inevitable. To begin, there is simply more of the primary distribution system than anything else in the power system. It consists of thousands of miles of line and tens of thousands of units of equipment.³ There is simply more equipment to fail. Additionally, in all but very rare cases, the primary feeder is radial in design, consisting of a single serial chain of lines and equipment leading to each customer. Any failure in this type of system will lead to an interruption of service to one or more customers.

³ The authors are counting service transformers as part of the primary feeder system in this example, because they are energized on the high side at the primary voltage level.

And finally, the feeder system is distributed – in fact it is the ultimate distributed resource belonging to any electric utility – a system of many, many units of equipment scattered about the system in roughly the same pattern as its customers. As such, locations of failures and repairs are scattered throughout the system, too. Unlike the substation level, the failed equipment will not be at a few well-identified sites. It often takes time – as much as an hour of searching in some cases – for repair crews to find the location of a failure.

Thus, the feeder system is the source of nearly half of a utility’s service reliability problems, because of its great extent, radial characteristics, and distributed nature. As such it is particularly sensitive to the increasing failure rates of aging equipment. The aging equipment trends (Chapter 6) mean that both “clear weather” and storm-related failure rates on the feeder system tend to increase rapidly with equipment age. Attention to the feeder level is critical just because so many failures occur there during the best of circumstances. So here is the potential for the increasing failure rates of aging equipment to cause very noticeable increases in SAIDI and SAIFI – even a small increase in such a large portion of the system will make a very perceivable difference.

Outdated System Structure Impacts on the Feeder System

Basically, many of the trends discussed in Section 8.2 lead directly or indirectly to changes on the feeder system that limit its ability to transfer loads, both to restore outages that occur in the feeder system, and to provide inter-substation transfer capability to mitigate substation-level outages.

Using the Feeder System to Mitigate Problems Caused by Substation-Level Outages

One of the most effective, if not the most effective, ways to support substation transformer outages in systems with high substation and sub-transmission utilization ratios is through the feeder system. When a transformer is loaded to 90% of its rating, as it is during peak in many utilities forced to high utilization rates by the issues covered in Section 8.2, avoiding lengthy customer outages when it fails means transferring a lot of load to other equipment, quickly. If the failed unit has only one neighbor that can interconnect to provide that support, that unit must go to 180% of rating – a level considered too high in all but extreme cases – or all of the load cannot be served. Concern over such high contingency loading levels is largely the reason why traditional utilization ratios for substation transformers were targeted at 66%. When one of two units at a substation failed while running at that level, the other had to go to no more than 133% of normal rating – a high but acceptable loading during a contingency.

One way to design a system so that any transformer has more “neighbors” that can support it during an outage, is to have more than two transformers at each substation. Some utilities (Houston) traditionally designed most of their distribution substations with four transformers. Loss of any one, even if loaded to 100% at peak, meant the load could be distributed among the remaining three with only a 133% overload of each. Transformer, bus, breaker, and similar contingencies could be contained within each substation, with no support needed for neighboring substations.

But in many cases, three- or four-transformer substations are not an efficient investment. In most cases, the economy of scale for transformers is greater than that for multi-transformer substation layout. Thus, for a given required capacity, two large transformers prove to be much more cost-effective than four smaller ones.⁵ As a result, as much as there is any “standard” substation layout used in the United States, it is the two-transformer substation, composed of two like units (same capacity). In a power system based on these types of substations, any outaged transformer-bus-breaker combination unit would have one remaining neighbor to provide contingency support (in the substation). The utility has three choices for a viable contingency plan for these outages.

1. Limit loading on the substation to what one transformer can take under contingency conditions. A few utilities have shifted to permitting 166% loading during contingencies (for brief periods), permitting up to 83% loading of the substation. Most, however, limit loading to 75% under normal conditions with a 150% contingency loading, still a very aggressive contingency loading level.

2. Provide support with a mobile transformer/substation. The plan being that a mobile spare could be brought to the site “within a few hours.” The utility accepts a brief outage during that time as the nominal result of a transformer failure. This is often not a palatable option because the average mobile substation-response time is actually 12-18 hours.

3. Provide support for transformer outages through the feeder system. The concept here is that the load of the failed transformer is split – the remaining transformer at the substation picks up part of the load, and the rest of the substation’s load is transferred to neighboring substations via re-switching of the feeder system.

This can be implemented via “programming” of the protection and breakers scheme in one of two ways.

1. Upon a failure of one unit, the remaining transformer picks up the entire load it can within its contingency rating. For example, if loading on both transformers was at 90%, the remaining unit goes to 135%, picking up half the failed unit’s load. The other half is dropped temporarily. Feeders are then switched to transfer load from the substation to neighboring substations, so as to pick up the load that was dropped. This may be done with remote switching, but is most often accomplished by manual switching, taking up to two hours.

⁵ Houston Lighting and Power (Reliant Energy) has a somewhat unusual situation in that its load density is much higher than average due to the intense amount of air-conditioning demand in the Houston area. Four transformer substations are therefore quite efficient in Houston’s case.

2. The remaining transformer immediately picks up the entire load. This means it would go to 180% of rating during peak, a level at which it is very quickly giving up lifetime to serve the load. As quickly as possible, feeders are now switched to transfer load to neighboring substations to bring loading down to a lower level for the duration of the contingency.

This scheme puts transformers (and buses and breakers associated with them) under high stress for brief periods, but avoids customer interruptions. If the switching is well managed, as for example with a computerized restoration system at the utility dispatch center, it can be done quickly enough to avoid large stress-induced loss of life on the one remaining operating transformer.

Figure 8.8 uses a simple example to illustrate the “transfer to neighbor through the feeder system” concept.⁶ At the left, a two-transformer substation with a hexagonal service area has boundaries with six (also hexagonal) neighboring substation areas. Transformers in this example are loaded to 90% of rating at peak, by design. Also shown are six feeder areas (three per transformer) alternating radially between each of the substation’s two transformers.⁷

In this example, when one transformer is out of service during peak conditions, the other picks up what load it can, going to 135% of its normal rating. The remaining load, corresponding to 45% of a transformer’s capacity, is transferred to neighboring substations through feeder re-switching. As shown, the concept is to transfer the peripheral load of the substation – the load closest to the neighboring substations. This assures the shortest electrical paths for the contingency-switching feeders (for voltage drop consideration). See Chapter 8, Section 3, for a fuller discussion of how switching capability like this is designed into a distribution system and what its benefits can provide. Support of transformer outages through the feeder system results in no, or very short duration, interruptions of service compared to the multi-hour or day-long outages if mobile transformers or other options have to be implemented.

⁶ See Chapters 13-16 in the Power Distribution Planning Reference Book – Second Edition (Willis, 2004).

⁷ This radial alternation scheme promotes feeder contingency support. See the Power Distribution Planning Reference Book – Second Edition, Chapter 14, particularly the discussion with respect to Figure 14.16, (Willis, 2004).

Figure 8.8 Left, hexagonal substation area with borders to six neighboring substation areas. This substation has two transformers that loaded to 90% of their rated capacity at time of peak demand. Under normal circumstances the two transformers, A and B, each serve half the load. When transformer A fails, unit B picks up half of A’s load, going to 135% of its rating, and 22.5% of each of the transformers’ loads is transferred to feeders from neighboring substations. In this way the utility can “get by” with loading transformers at two-transformer substations to 90% of rating at peak demand, and still have a viable contingency capability for the failure of either one. See text for details.

A Useful Concept: Feeder System “Strength”

In order for the feeder system to work well for contingency support in the example given, the feeder system must be able to transfer 22.5% of the substation’s load to neighboring substations (45% of one transformer’s load equals 22.5% of the substation’s). If the feeder system is not capable of transferring 22.5% of the load, the system dispatcher is forced to choose between dropping load (interrupting service to some customers) or loading the remaining transformer to an unacceptably high, potentially damaging, level.

On the other hand, if the feeder system can transfer more than 22.5%, then that capability can be used for support transformer contingencies even if design loading of the substation is raised beyond 90%. If it can transfer 35% of each substation’s peak load, then utilization ratio can be raised to 100% of transformer rating while still preserving contingency support capability.

The contingency strength of a feeder system is the portion of the substation load that it can transfer to neighboring substations while keeping voltages and circuit loading on the feeders within the utility’s standards for contingency operation.

Figure 8.9 Feeder system strength (inter-substation tie capacity) versus utilization ratio of transformers under normal conditions that can be tolerated while still preserving contingency capability for the outage of a transformer at a two-transformer substation. For example, the graph indicates that if a contingency loading of 166% is allowed by standards, then normal target loading must be limited to 83% if the feeder strength is zero, but can go to 100% if feeder strength is 17%.

Figure 8.9 shows the utilization ratio for a two-transformer substation that can be supported in contingencies versus feeder system strength – feeder system inter-substation transfer capability. The term “inter-substation transfers” refers to load transfers between substation service areas. The term “intra-substation transfers” refers to load transfers among feeders in the same substation area. Several different lines are plotted representing various levels of contingency loading limits. For example, the “133%” line represents the combination of utilization ratio and feeder system strengths needed to assure full contingency capability when contingency loading is limited to 133%. The Y-intercept values for each of the lines represent the maximum utilization ratio that a utility can use at a two-transformer substation given that no load will be transferred to neighboring substations in the event of a transformer outage. As can be seen, greater transfer capability through the feeder system means greater utilization ratios can be applied in the normal operation of the system.

Loss of Contingency Reach Due to Obsolete System Layout

“Reach” is the distance that a feeder system can move power while staying with the voltage standards applicable for its operating conditions. For example, a 12.47 kV, three-phase, overhead feeder of typical industry design, using 4/0 ACSR conductor, can move power corresponding to its thermal limit (the maximum amount of power possible within its current rating, 340 amps/phase, corresponding to 7.3 MVA) a distance of 1.8 miles before voltage drop on the line reaches 7.5% - the maximum permitted on the feeder system according to many electric utility voltage standards. This distance is 4/0’s thermal reach.

Substitution of another conductor will not materially change this distance: a 336 conductor will move its thermal limit (530 amps corresponding to 11.5 MVA) 1.8 miles; a 636 AA (770 amps, 16.6 MVA) will move its thermal limit 1.7 miles; a #2 ACSR (130 amps, 2.8 MVA), 1.8 miles. The larger conductor has lower impedance (less voltage drop per ampere) but a higher thermal limit. The two factors compensate for one another. Thus, the thermal reach of a 12.47 kV feeder system and the distance it can move power if operating at its thermal limits is 1.8 miles, regardless of conductor size.

Economic reach refers to the distance that power will travel on feeder segments before encountering the maximum allowed voltage drop, in situations conductor size and loading have been selected so that all conductor operates at or near its most economical loading level over the course of a year.⁸ The economic factors, load characteristics, and design standards that affect determination of economic loading vary from utility to utility. But again, the economic reach of each type of conductor turns out to be about the same. Under typical conditions economic reach works out to about 3.6 miles – about twice thermal reach (economical loads are usually about ½ the thermal limit).

Contingency reach refers to the distance that power can be moved on the under contingency conditions, when higher loading levels and more liberal voltage drop limits that are permitted. Again, standards vary from one utility to another. But as an example the ANSI C84.1-1989 standard specifies range A voltages as design criteria for normal conditions on the system, with a 7.5% maximum voltage drop on the primary feeder system permitted (9 volts on a 120 volt scale). It states that Range B voltage ranges apply in emergency situations, permitting up to 13 volts on a 120 volt scale (10.8% voltage drop).⁹

Thus, according to this standard, during a contingency a distribution utility is permitted to operate feeders in the affected area with up to 10.8/7.5 equals 144% more voltage drop than under normal conditions. This 44% additional voltage drop can be “spent” by either moving more power (more current equals more voltage drop) or by moving power over a longer distance (longer distance equals more voltage drop), or in some combination of more power moved a greater distance which is the normal case.

⁸ For a detailed explanation of how economic loading is determined for conductors and for a distribution system, see the Power Distribution Planning Reference Book – Second Edition, Chapters 11 and 12 (Willis, 2004).

⁹ See the IEEE Red Book - Recommended Practice for Electric Power Distribution for Industrial Plants (IEEE, 1990).

Generally, due to a bit of margin built into a feeder system for a variety of secondary and tertiary reasons, there will be a bit more voltage drop margin than minimum.¹⁰ Thus, under most circumstances it would be reasonable to expect about 50% – 60% to be the permissible additional margin, over and above normal peak load conditions, rather than just the 44%.

¹⁰ There are only a limited number of different conductors available. Often in “optimally sizing“ a particular segment engineers determine that one size is a bit too small, the next size a bit too large, to be exactly optimal. Necessity dictates they pick the larger size, which provides a bit lower voltage drop than technically necessary – additional margin for contingencies, among other things. In typical systems this adds about 10% to the available voltage drop margin.

Consider again Figure 8.8 and its example of transfers to other substations. Note that 22.5% of each feeder’s load is transferred to neighboring feeders. Roughly speaking, the additional distance for power flow is about 20% farther (assume for the moment that the other substation areas/feeder areas are the same size and shape as those in the drawing). Thus, additional loading times additional distance (122.5% x 120%) is very close to the 144% limit. If this feeder system is well designed and operated such that normal loadings are kept in the most economical range for all conductors (as it ideally would be) then no substantial reinforcement of its elements should be required to provide the required level of contingency support.

Little if any additional capacity needs to be installed to provide this contingency support. A well-designed feeder system with all conductors sized to be most economical, and with switching designed for inter-substation support, can be built at very little if any additional cost (beyond the extra planning and engineering work required) over a system that does not have these capabilities.

However, consider what happens when the loading on the feeder system is permitted to creep upwards, as occurred to Metropolitan Power and Light example case in Section 8.2. Over the course of three and one-half decades an increased loading on the feeder system was accepted as an inevitable result of changes the utility could not stop. Although reinforcement of the feeder system was done from time to time, substituting larger conductors where necessary, many portions of the feeder system ended up running at above their most economical loading ranges. As a result, they lose reach.

The net result is that a situation like that depicted in Figure 8.10 can occur during a contingency. There, two feeders for neighboring substations are shown (each a triangular area corresponding to the triangular feeder areas in Figure 8.8).

The rightmost feeder’s substation transformer has failed. Its outer 22.5% load has been un-switched from the substation (opening switch B), and the feeder tie switch at point A closed, so that the feeder at the left can pick up the outer most load from feeder B. Due to higher feeder loadings, the feeder system does not have the contingency reach to provide this capability. The dotted line in Figure 8.10 shows where voltage drops below unacceptable levels.

The outdated system structure issues discussed in Section 8.2 created two effects that led to this situation:

1. Loss of contingency capability. A combination of higher substation utilization ratios and higher loadings on the feeder system lead to a large shift in the ratio of capability available/capability needed. Higher substation ratios raise the amount of load that needs to be transferred over the feeder system during contingencies. Also, higher loadings on the feeder system reduce the voltage drop margin available for contingency reach. As a result, the number of situations and hours per year (near peak) when contingencies cannot be supported increases. Reliability of service suffers.

2. Contingency capability preempted by planning. In many systems, local pockets of load growth are accommodated without reinforcement of the feeder system in the area, via load transfers. For example, suppose that in Figure 8.9, the “contingency” that occurred was a planning contingency: the load in the feeder area on the right grew to be 10% more than that feeder could handle while staying within standards, but the feeder on the left had 10% margin. Utility planners can avoid any reinforcement cost by performing the switching operation shown, as a permanent solution to the “overloading” problem. This “fixes” the normal operating problem but leaves insufficient contingency capability in the system.

Feeder Problems Often Don’t Get the “Respect” They Are Due

While the feeder system is critical to any utility’s establishing a good record of reliability, failures there never cause the widespread, “headline grabbing” problems that failures at the substation and sub-transmission level can cause. The failure of an entire feeder might drop service to between 500 and 1,500 customers. By contrast a bad problem at a substation can drop twenty feeders at once, causing a widespread outage that garners serious attention from newspapers and community leaders. For this reason, efforts to fix aging infrastructure problems often focus on levels of the system other than the feeder system. Still, the feeder system is an important part of the system, and as shown can be an important resource in limiting the impact of substation – sub-transmission outages. Therefore, attention given to it is more important in the overall reliability equation than is often realized.

Figure 8.10 Here, the feeder on the right has experienced a failure at the substation (shown by an X). Switch B has been opened, switch A has been closed, so the feeder on the left can support both all of its load and the outlying 22.5% of the other feeder’s load (the rest is supported by switching the outaged feeder truck to another feeder near the substation). The feeder on the left is now serving 122.5% of its normal load, over a distance of transmission that is as much as 15% more. Dotted lines show the points where voltage drop reaches range B limit.

For this reason, efforts to fix aging infrastructure problems often focus on levels of the system other than the feeder system. Still, the feeder system is an important part of the system, and as shown can be an important resource in limiting the impact of substation – sub-transmission outages. Therefore, attention given to it is more important in the overall reliability equation than is often realized.

Outdated Structure Impact on the Feeder System: Reliability

Regardless of whether one or both of the situations discussed above develop the effect on reliability is the same. A growing portion of the equipment outages, which could potentially lead to widespread, lengthy interruptions (i.e. outages of major equipment at the substations) now have no viable contingency support.

In several large metropolitan systems outdated system structure issues (Section 8.2) eroded feeder contingency support strength due to these two reasons, to less than 40% of what was required to meet system needs, a loss of 60%. Figure 8.5 showed that substation – sub-transmission related problems cause about 25% of system SAIDI. Assuming again that Figure 8.5 is reasonably close to the distribution of frequency-of-outage causes in the system (see Footnote 6) and that degradation in results is proportional to degradation of capability (it is actually worse), this 60% reduction in capability to support substation-level problems would lead to something along the lines of a 15% increase (60% x 25).

This impact on reliability is roughly three times the impact on system reliability that issues related to feeder system reliability cause (that came to 5%). Overall then, outdated system layout problems (Section 8.2) lead to a 20% net increase in reliability problems, just due to their impacts on the feeder system alone.

Improved Communication with Community and Regulators

Electric utilities have to do a better job of identifying to those who approve their plans that new facilities and sites are needed if long-term reliability and service quality is to be kept at high levels. The quality of results that utilities obtain from their communication with regulators and community varies greatly. All utilities would be wise to emulate the methods employed by utilities that are particularly effective in this manner, such as PSE&G.

Compact Substation Designs

The compromises of capacity over configuration as illustrated by Figure 8.4 are simply unacceptable in a system where utilization rate is kept high and reliability is a priority. A better solution is to find a way to “cram” more capacity into existing sites. This can be done to two levels of improvement:

1. Compact designs. Various means exist to reduce the space required for air-insulated substation equipment. Usually this entails re-building the substation upward rather than outward. Among the items used are “low footprint” transformers, and high rise buswork designed to use little horizontal real estate (which as a consequence has a much higher profile). Compact designs can typically increase the capability of a substation (as measured by additional capacity added without any sacrifice in configuration) by up to 25 – 30%. Often this is not enough to provide all the improvement needed.

2. Gas-insulated switchgear. The solution of favor for crowded conditions in urban areas throughout Europe is the gas-insulated switchgear (GIS) substation, the so-called substation in a bottle. Slightly more than half the metropolitan substations throughout Western Europe utilize GIS technology. GIS substations have all of their busy-work and breakers contained inside a set of linked insulating capability. Voltage clearances between equipment can be much tighter than with air-insulated equipment, and the weather-protection rendered by the enclosures and gas means that much more compact designs can be used for all equipment. As a result GIS equipment requires far less space than its air-insulated equivalents.

GIS substations can increase the capability of a substation site over that obtainable from the tightest possible use of traditional air-insulated equipment, by a factor of at least three. If need be, an old 160 MVA site could be rebuilt with close to 500 MVA of capacity. Very often full use of this capability is not needed, but usually a doubling of capability, or potential for eventual growth, is implemented when GIS is retrofitted to a site. GIS also makes sense as the construction standard for new substations in aging infrastructure areas, because it enables the utility to have a much wider choice of available sites. Its compact nature and the fact that it can be put inside medium-sized buildings, means it will fit in many places that previously had to be rejected.

GIS substation equipment costs about 15% – 30% more than air-insulated alternatives. However it is slightly more reliable than, and requires much less maintenance cost than, air insulated designs of equivalent capability, basically more than canceling out this high initial cost if evaluated over ten or more years of operation. In addition, its minimization of real estate needs also contributes to lower cost for land, a major consideration and often more than enough to offset the higher initial equipment cost.

Reinforcement of Feeder Trunks and Switching

Given the typical effects of outdated system structures on the feeder system, providing the capability required for support of substation-level contingencies, will mean reinforcing feeder trunks with larger conductors, and installing additional switches for improved flexibility of operation. This is usually money well spent, and if planned well and coordinated with operation planning, feeder reinforcement for improved contingency capability is among the most effective ways to spend limited funds to obtain improved reliability of service.

Improved Planning

Improved planning can contribute significantly to a utility’s both avoiding the situation depicted in this chapter, and mitigating its effects if it is already in such a situation. Good planning provides foresight, rational consideration of options, and under the right conditions, innovation. The following recommendations should be taken as a whole: they provide the proper type of planning for aging infrastructure areas as well as for utilities who need to maximize both customer service and stockholder (financial) performance. Chapters 12 and 13 will discuss this in more detail.

Section 8.2 cited ineffective long-range planning as a contributing cause to many of the causes that ultimately lead to outdated structure and poor reliability. These problems have to be fixed before lasting solutions can be put in place. Chapter 8 covers the nature of several problems with traditional planning tools. Chapters 11 and 12 discuss how improved planning methods and tools appropriate for modern high utilization-ration systems can be effectively used.

But more is needed. Section 8.2 showed that a large measure of the problems utilities created for themselves were due to simply not looking far enough ahead or not using long-range planning capability that was there. Thus, in addition to changes in method and technique, utilities must simply pay attention to and use the longer-term perspective and evaluation of long-term consequences that flow out of long-range analysis of the consequences of growth and aging.