Reports

The wave of concern provoked by the summer of 2003’s massive power blackout has largely receded, but in its wake deeply-rooted transmission investment dilemmas remain. With deregulation and expansion of electricity markets, the operation and planning of transmission have increasingly become separated from ownership of generation, and from the retailing function. Besides complicating transmission planning, this increases the diversity of stakeholder interests, and accentuates the “who benefits?” question.   

Like highways and other infrastructure, electric transmission is critical to efficient and reliable delivery of services, but would likely be inadequately funded if left to individual market participants.[1] One problem is that potential beneficiaries of transmission investments are numerous and diverse, and may not be located directly on a particular “route” being funded. Also, benefits will be long-term, influenced by uncertain future developments including high-consequence outcomes that are individually unlikely.  Finally, it may be inefficient, undesirable and even impossible to control distribution of benefits.

It is thus essential to value transmission investments from a network- and market-wide “big picture” perspective of sufficient scope to capture major consequences and uncertainties. At the same time, assessment of benefits must be sufficiently detailed and disaggregated to provide an explicit (if probabilistic) view of how different stakeholders would be affected. This provides a basis not only for evaluating and negotiating transmission investments, but also for verifying the credibility of projected benefits.   

The “big picture” generally involves “societal value,” reflecting combined impacts on consumers, producers and recipients of congestion revenues (discussed below). However, the appropriate measure of value can be ambiguous or contentious. How far a field should we look geographically? Should we exclude monopoly (noncompetitive) rents? Should we only consider consumer benefits, especially where transmission costs are ultimately rolled into overall rates? We could measure value in terms of reduction in production costs due to more efficient generator utilization, but this ignores how much consumers pay above production costs and who receives the payments. How benefits are distributed among stakeholders may be of considerable importance, especially if some stakeholders experience large benefits or disbenefits.

In any event, valuation of transmission investments must consider the fundamental dynamics and integration of transmission networks with electricity markets. A realistic and sufficiently broad-scoped assessment can capture benefits that might otherwise be missed or understated, and can facilitate recognition of stakeholder impacts. 

Besides economic efficiency benefits it is necessary to consider reliability benefits, which requires addressing network characteristics and operations with sufficient realism and detail. Reliability may be treated as a threshold issue (it must be good enough, anything better has no extra value), converted to a monetary value, or left in physical terms such as loss of load probability. Comprehensiveness and consistency of the transmission investment valuation process are enhanced by analyzing reliability and economic benefits in a single integrated methodology. Reliability considerations can restrict economic operation of generators, while on the other hand, generator operating and investment strategies can impact reliability.

Reliability and economic benefits sought through transmission investment may be achieved by alternative means, including new generation of the right type and location, better use of existing generation, or demand-side measures. For example, better use of synchronized reserves could help address reliability problems exemplified by the massive 2003 blackout.[2]  Often the appropriate alternative against which to value transmission investment is not doing nothing, but rather investing in generation or demand side measures.[3] Thus, the transmission investment valuation process must ultimately be able to address, within a consistent framework, both transmission and non-transmission alternatives. This requires analytic tools that are sufficiently realistic and comprehensive in treating the interaction of transmission, generation and loads.

In the following example, LCG Consulting’s UPLAN Network Power Model was used to simulate hourly optimal commitment and dispatch of a simple 6-generator system serving load over a 7-line transmission network (Figure 1), before and after upgrading the transfer capability of the line identified as “E2D” from 200 MW to 400 MW. Value was calculated over one 8760-hour year of operation, from the perspective of different stakeholder classes, and as well as individual market participants within classes. The example demonstrates how the interdependent nature of a simulated AC network obeying electrical principles gives rise to outcomes that may be counterintuitive, but in which consumers broadly benefit from network upgrades, especially if the supply system may become stressed in the future.  Insights and value assessment might be less clear or detailed if using analyses that treat network power flows and constraints more simplistically, or that do not integrate market and network dynamics.    

In this example (Figure 1), loads are located at four of the six busses. Zone 2 (busses D and F) is a “load pocket” having limited transmission access to low-cost generation. The characteristics of the generators, loads and transmission lines are summarized in Tables 1 and 2. Loads cycle daily, peaking in the summer and having an annual load factor of about 60%. The load profile is slightly different for Zones 1 and 2.   

Figure 1. Example 7-Bus System, Indicating Base Case Flows and Congestion Prior to Transmission Upgrade

Key UPLAN simulation results examined included locational electricity prices, generator revenues and net incomes, bus-specific consumer costs, bus-specific unserved energy, and the hourly power flows, congestion and congestion revenue rights (CRR) payments for each line. UPLAN provides the capability to incorporate numerous N-1 transmission contingencies in generator commitment and dispatch, as well as to dynamically integrate energy and multiple ancillary services markets.[4] While important for full-scale studies, these complexities were omitted from (and are largely inappropriate for) this simple example. Noncompetitive bidding could be examined, but was not.  

Table 1. Generators and Loads (Base Case)

 

Table 2. Seven-Line Transmission Network

The example system was simulated for one 8760-hour year under three different generation-load cases: Base Case, Add Generation case and Add Load case (described below). The impact of upgrading line E2D’s transmission capability from 200 MW to 400 MW was examined under each generation-load case.  This represents a snapshot of one year, whereas actual studies would likely consider several time horizons.   

As expected, power flow direction over 8760 hours of the year is primarily towards the load buses and especially towards the “load pocket” (Zone 2, busses D and F in Figure 1). The most frequently binding constraints are on line B2E (reverse direction) and, especially on line E2D, both carrying power out of generation-rich areas.  Competitive marginal cost-based generator bidding combined with transmission constraints produces locational marginal prices (LMP) that vary substantially from bus to bus (Table 1), with load pocket consumers experiencing high prices. As a consequence, the lowest-cost generator in the load pocket (Generator 5) achieves a high net income.

Outside of the load pocket, high-cost Generator 1 operates intermittently, but is sustained financially by occasional high LMP at its location (Table 1), due to the way that transmission constraints affect network-wide generator dispatch (discussed below). The lowest-cost generators 2 and 4 (buses A and E) operate virtually all of the time, and their energy market income reflects their infra-marginal costs. Higher cost Generator 3 has the misfortune of being located in a “generation pocket” with limited access to loads, thus being on the margin when it operates.  To remain viable it might have to achieve significant capacity or ancillary services revenues, and/or benefit from generators bidding above marginal costs, none of which was considered in this example.[5]

Table 1 illustrates how LMP can exceed the bid of the highest-bidding generator, in this case $50/MWh. This occurs often in the load pocket and also affects Generator 1 at bus C, which runs when there are high prices and congestion. The explanation is that the LMP at a bus represents the incremental system cost for serving an increment of load at the bus. With network constraints this may require not only ramping up some generators but also ramping down others, so that the incremental cost can exceed the cost of simply ramping up the marginal generator. Such interdependencies are addressed by realistic power flow models, and will generally be most pronounced where the network is substantially non-radial and experiencing congestion. Transmission upgrades can then have widespread and sometimes unexpected effects on economic efficiency.  

The transmission investment consisting of upgrading line E2D from 200 MW to 400 MW capacity was valued based on four components of value:   

• consumer surplus, based on reduction in annual electric energy payments by consumers overall and at each bus,[6]
• reliability of service to loads overall and at each bus,
• producer surplus based on generator net income overall and for each generator[7]
• congestion payments received by transmission owners and/or holders of congestion revenue rights (CRRs), overall and by transmission line.

Reliability of service can be valued in different ways. For a short time horizon with no warning or opportunity to mitigate, the value of reliability may be very high, having typically been estimated on the order of thousands of dollars per MWh of unserved energy.[8] However, our example assumes a longer “investment” time horizon, and calculates the reliability value of a transmission upgrade based on the avoided cost of alternative measures for maintaining reliability. The cost of such alternative measures is assumed to equal a capacity cost of $80 per kW-year per peak kW of unserved demand eliminated, plus an energy cost of $50 per MWh of unserved energy eliminated. This is very roughly equivalent to the cost of a new, advanced combustion turbine.

Under a locational pricing system incorporating congestion revenue rights (CRRs), congestion payments associated with any transmission line represent the difference between what consumers pay for electricity imported over that line, versus what the exporting producers receive. This difference occurs because consumers pay based on their locational price while producers are paid based on their locational price, and these locational prices differ when congestion restricts optimal power flows.[9] Thus, when transmission constraints lead to inefficient generator dispatch and elevated consumer payments, some of these elevated payments go to generators, and some go to holders of CRRs (who in some instances may be the transmission owners).[10]

Not surprisingly, in the Base Case, consumer energy cost savings dominate the transmission upgrade benefit calculations, so that the $56.2 million reduction in annual consumer energy payments greatly exceeds in magnitude the income losses for generators and CRR holders (Table 3). The upgrade eliminates unserved energy, but the consumer energy cost benefit is substantially greater than the reliability benefit. However, including the reliability benefit significantly increases the net benefit (after subtracting producer and transmission owner disbenefits) by about 50%.

The overall result suggests a stakeholder class distributional issue in that consumers gain and both generators and CRR holders lose. Where load serving entities own the generation and CRRs, this distributional issue may recede, if generator net incomes and congestion revenues are flowed back to consumers to offset revenue obligations.  

Table 3. Breakdown of Transmission Upgrade Benefits Under Three Generation-Load Cases

We now consider how transmission investment benefits are distributed among stakeholders, particularly the two load zones and six generators. Of course, some generators could be affiliated with other generators or with load serving entities. On the other hand, there could be more complex stakeholder interests such as based on contracts that hedge and reallocate risks and benefits. In any event, detailed and realistic modeling of the network, the market and their interaction is required to distinguish impacts on different stakeholders, especially in more complex real-world situations.   

Under the Base Case, all consumers (all load buses) benefit from the upgrade, although those in the load pocket benefit the most, with their energy payments dropping by more than 17%, versus a 5.5% for other consumers. Realistic network simulation captures the strong interrelationship among network elements and generators, including the way that the transmission upgrade on line E2D provides greater flexibility for operating the overall system, not merely for moving power into the load pocket. This explains how consumers outside of the load pocket can benefit from the upgrade. Before (but not after)  the upgrade, overall network constraints caused the LMP at bus C outside the load pocket to occasionally reach high levels above $50/MWh, due to sometimes having to adjust dispatch across the network in order to serve the last increment of load at bus C.[11]

For generators, impacts of the transmission upgrade are more dramatic and varied (Figure 2). Due to more flexible and less constrained system operation after the upgrade, there is a large reduction in prices in the load pocket, shrinking generator 5’s formerly generous operating margins, and driving generator 6’s energy market margin essentially to zero (perhaps requiring capacity, ancillary services or RMR payments to sustain the generator). Generator 1 experiences a similar effect, since after the upgrade it is able to operate more hours but at reduced prices, because the upgrade eliminates power flow constraints that occasionally produced high prices at Generator 1’s location.

Figure 2: Changes in Generator Net Income Due to Transmission Upgrade

These generator outcomes illustrate the substantial income loss faced by some generators due to transmission upgrades in competitive markets having locational pricing for generators. This raises questions of whether such generators are entitled to any of their original profit margin, whether capacity markets, RMR or other arrangements could keep them viable, or finally, whether it is simply acceptable and economic for them to be replaced. These generators’ owners would likely resist the transmission upgrade, and the assessment process might need to consider stranded or retired generation scenarios.  

As indicated in Table 3, change in overall congestion revenues represents a relatively minor outcome of the transmission upgrade. The disaggregated results are more interesting, in that the upgrade reduces congestion and congestion revenues for line E2D and to a lesser extent for the other Zone 2 import line C2D, but actually increases congestion on lines B2E and C2D. This illustrates an important phenomenon in transmission planning:  investing to relax a binding constraint somewhere on an interconnected network will facilitate economic power flows up to the point that some other constraint(s) become binding. One recent transmission planning study using UPLAN concluded that “although there are often hidden layers that could potentially be binding constraints, they have no impact on the generation dispatch until the outermost layer [the initial constraint] is removed…. a power flow analysis is not capable of determining which of several contingency/overload pairs is the most binding constraint; however, UPLAN is capable of identifying which one is most binding (costly)…”[12]

In our example, the upgrade actually increased projected total line-hours of congestion (one line congested for one hour = one line-hour), but reduced the cost of congestion, as the upgrade was intentionally focused on the economically most burdensome constraint.  Since the investment decreased congestion revenues for the upgraded line (by reducing price differentials across the line), receipt of congestion revenue rights (CRR) for the line might not be a good motivator for transmission investment.  

UPLAN’s network-market simulations can be used to evaluate numerous uncertainties, including automated volatility analysis that converts input probability distributions for key drivers into probability distributions for such results as prices and asset values. Here, we simply examine two discrete alternative scenarios regarding generation and load, “Add Generation” with a new, low-cost 125 MW generator added in the load pocket at bus D, and “Add Load” with the load at bus D increased by 10%.  Actual transmission investment valuation would typically consider a much wider range of uncertainties.

Under “Add Generation”, net societal benefit of the upgrade is about $22 million/year (over 70%) less than in the Base Case, due to lower prices and reliability problems in the load pocket, to start with (Table 3). However, under “Add Load”, the societal value of the upgrade is about $15 million per year (about 45%) higher than under the Base Case, due to starting with more constrained supply and higher prices.   

The net benefits can be disaggregated. Under “Add Generation” the load pocket experiences lower energy costs and reliability problems to start with, so that the transmission upgrade produces $28 million/year less consumer energy cost benefit and $10 million/year less consumer reliability benefit, compared to benefits under the Base Case (Table 3). However, “Add Generation” also starts with lower generator incomes and congestion revenues, so that the upgrade has a $15 million/year lower impact in reducing producer income and a $1 million/year lower impact in reducing CRR revenues, compared to the Base Case.

Under “Add Load” the opposite occurs. Starting from higher levels of congestion-driven prices and producer surplus, the transmission upgrade has greater impacts. It produces $64 million/year greater consumer energy cost benefit than it did under the Base Case, and $5 million/year greater reliability benefit. However, these increased benefits are mostly offset by a producer surplus decline (due to the upgrade) $47 million/year higher than the decline projected under the Base Case, and a CRR revenue decline $7 million/year higher.  It could (and likely would) be debated how loss of high congestion-driven revenues should be treated in the “societal benefit” calculation. In any event, such revenue effects illustrate how even competitive (non-strategic, not exercising market power) generator bidding can produce high returns when there are substantial transmission constraints. 

Finally, note that under the Add Load case’s constrained supply conditions, even consumers outside of the load pocket benefit considerably from the transmission upgrade, since pre-upgrade network operation was so severely restricted as to even elevate some consumer prices outside of the load pocket. However, under the Add Generation case’s more relaxed supply conditions, consumers outside of the load pocket are very slightly harmed by the upgrade, because the main impact of the upgrade is simply to permit export of additional lower-cost energy out of Zone 1 into the load pocket (Zone 2).  

Among the generators, impacts of the transmission upgrade fall upon the three generators originally benefiting the most from transmission constraints (Figure 3). Tighter supply conditions under “Add Load” produce the greatest network constraints and the greatest pre-upgrade revenues for these generators, as well as the greatest upgrade-caused income declines. Without the upgrade, price run-ups resulting from transmission constraints clearly aid these generators’ financial survival. What is less clear is how to value (and what to do about) their possibly fatal economic losses if the upgrade is implemented. The other generators see little economic impact from the upgrade.

The wide distribution, longevity and uncertainty of transmission benefits help make it unwise to rely on “the market” for transmission investment. For valuing and selecting such investments, a realistic, integrated network- and market-wide “big picture” assessment is very helpful, in order to capture broad system-wide benefits while also revealing the distribution of stakeholder impacts.  Why praise strategic transmission planning unless we are willing to apply it in actual investment decisions?

This example illustrates how the value of transmission upgrades can be counter intuitively large and widespread, using electrically realistic integration of network power flows and electricity market dynamics, even without the further complexities of security-constrained commitment and dispatch, or strategic bidding. Furthermore, consumers will generally be the main beneficiaries of network upgrades, especially if stressed supply conditions might occur, and even consumers physically removed from the upgrade itself can benefit. Furthermore, transmission investment may show especially high value to consumers under low-probability, high-impact events such as major outages or disruptions, which should be considered in transmission valuation. On the other hand, there may be some losers, especially among generators previously benefiting from congestion. An integrated, realistic and sufficiently detailed analysis provides the best basis for characterizing, evaluating and discussing the potential stakeholder gains and losses. 

Dr. Rajat Deb is the founding president of LCG Consulting, Los Altos, California and chief architect of the pioneering UPLAN program. He has more than 30 years experience in energy industry and academia and has contributed extensively to both the theory and practice in the energy field. Dr. Keith D. White has consulted on energy matters for 25 years, focusing on technical, economic, environmental and policy aspects of electricity supply, as well as transmission planning and market design. Other staff members of LCG have contributed to the model simulation and provided valuable comments and suggestions for this paper.  Dr. Deb can be reached at deb@energyonline.com or 650-962-9670.