Is DER competitive with traditional utility investments, and if so, what are the costs and benefits?
The role of distributed energy resources (DER) in electric-utility planning continues to foster robust debate.1 Despite best efforts, utility planners and policy analysts often are stymied or unable to reach consensus in attempts to capture the range of values that DER might provide as an alternative to traditional utility solutions.
Driven by public-policy initiatives, utilities are responding to recent mandates to integrate nontraditional resources into their distribution systems. In several states, renewable portfolio standards (RPS), and incentives—tax and end-user direct—are driving photovoltaic (PV) and wind-system growth. California has a 33 percent RPS goal by 2020 and its related Solar Initiative has a PV goal of 3,000 MW by 2017; investor-owned utilities have a mandate to achieve 5 percent of peak supply via demand response by 2007. Utilities throughout the United States are making large investments over the next several years in distribution. To manage a budget that could reach several billion dollars within years, utilities will make hard tradeoffs regarding which investments offer greatest value. Specifically, utilities will need to use innovative methods to quantify DER value when integrated into the electric-utility distribution grid.
Role of DER in Electric Utility Planning
Utilities are being asked to view non-traditional technologies such as distributed generation (DG) and demand response (DR) from a fresh perspective, including redefining the role of DER when considering distribution capacity additions. In so doing, utility planners should consider the extent to which DER is an opportunity to meet load growth; i.e., the level of capacity that should be viewed as “firm.” Second, new analytics are needed to compare nontraditional resources (DG and DR) as viable alternatives, technically and economically, with traditional distribution investments.
The costs of traditional transmission and distribution (T&D) options include up-front investment costs, such as new feeders or substations. In contrast, DER costs could include incentive payments, typically to customers who install DG or participate in DR programs. DG also could represent a direct cost to the utility if it owns the device. DER typically is added in smaller increments, with costs accruing continuously over the planning interval.
Traditional T&D benefits include the avoided risk of outages that might otherwise occur absent the upgrade. The relative value of the T&D option versus other solutions can be derived by dividing the net present value benefit achieved by avoided outage risk by cost; projects with the highest benefit-to-cost ratio are deemed the best choices.
Traditional Capacity Planning
Utilities add distribution capacity when normal or post-contingency loadings reach capacity limits. Where backup support does not exist, such as a single transformer substation without feeder ties, capacity is added when loads are expected to exceed normal ratings.2 Beyond 10 to 20 percent overload, the potential that equipment emergency ratings will be exceeded is high.
Not all utilities add or upgrade equipment based solely on deterministic criteria; that is, adding capacity when normal ratings are exceeded. A joint probability analysis of outage exposure and equipment failure rates establishes the level of exposure caused by insufficient capacity. The result, when multiplied by the value of unserved energy, represents the risk avoided by adding capacity.
Most agree that the value of energy unserved exceeds retail electricity cost; some suggest at least by an order of magnitude. Some utilities apply value-of-service (VOS) methods to quantify the avoided risk associated with unserved energy. Surveys yield values of $5/kWh to $25/kWh that query residential, commercial, and industrial customers on the degree that outages impact production, inconvenience, productivity, company image, and other factors.
The savings achieved by T&D deferral should not be confused with the value of avoided risk—i.e., the cost of unserved energy. For the latter, a risk-based VOS approach is applied, where the annual value of avoided risk is derived by multiplying VOS times the amount of unserved energy due to insufficient capacity. A VOS calculation yields values that typically are far higher than the avoided cost of annual capacity deferral, and should be used to compare T&D to DER via comparable metrics.
Use of probability methods implies that the value of avoided risk is low for minor overloads with low exposure hours. These values increase commensurate with outage exposure and capacity deficits. As loads increase, the likelihood of equipment failure rises as well. Industry standards and professional organizations provide guidelines that predict transformer failure as a function of loading. Figure 1 illustrates how transformer failure rates increase exponentially as loads increase. Hence, the value proposition for new capacity rises significantly when loads extend beyond the knee of the curve.
A range of direct and indirect benefits may apply to DER. Table 1 groups these benefits by two categories: (1) direct versus indirect; and (2) degree of difficulty to quantify values. Direct benefits that are relatively straightforward to quantify appear in the upper right quadrant. Indirect or difficult-to-quantify benefits are part of the evaluation process.
Potential DER benefits include:
Deferred T&D Capacity. Benefits associated with deferred T&D upgrades or additions assume that “firm” DER, when combined with existing T&D capacity, is sufficiently reliable to meet feeder or substation peaks. DER may be a viable solution only for a limited number of years on high-growth circuits, after which the traditional T&D solution is necessary. Accrued benefits also must be restricted to those years in which DER is a solution.
Losses. Line and equipment loss reduction is in direct proportion to the square of the capacity of online DER. When loads are high, loss savings can be substantial. Savings include: (1) demand reductions at the time of the feeder or substation peak; and (2) energy losses over the entire period when DER is operating. For DR, operating hours are low, and savings mostly are demand-related. Loss credits are calculated on an incremental basis. This distinction is important as incremental losses tend to be higher than average; up to double for heavily loaded lines.
Voltage/Value at Risk (Var) Benefits. DER can provide steady state and dynamic power benefits. Dynamic benefits are provided mostly by synchronous DG, which can provide voltage support during system contingencies. Steady state Var and voltage support is provided by DG and DR by virtue of load reduction, which reduces Var demand on power lines and equipment.
Deferred Central Capacity. Capacity credits may accrue when DER is available at the time of the system peak. When “behind-the-meter” DG is online or DR is called, load that otherwise would be met by generation or imports is reduced. Capacity needed to meet reserve margins—typically about 15 to 20 percent of native load—is reduced proportionally.
The level of firm DER capacity is adjusted upward by 20 to 30 percent to capture reserve margin credits and peak-loss reduction. Incremental peak losses often are significant; 10 to 15 percent reductions are common.
Relying on DER to Defer Capacity
Utilities are understandably cautious when asked to rely on customers to provide firm distribution capacity. Utilities have an obligation to deliver power safely, reliably, and continuously—distribution availability that often exceeds 99.99 percent underscores utility efforts to maintain high reliability. Massachusetts, New York, and California have introduced performance-based rates (PBR), with penalties applied when reliability falls below minimum thresholds. If DER causes reliability to degrade, utilities quickly will reject it as a solution.
Figure 2 illustrates how DER “fits” into a feeder or substation load duration curve (LDC). The difference between peak load and rated capacity represents the amount of firm DER that must be available to enable capacity deferral. For most utilities, the highest loads occur over a small number of hours. Typically, the top 20 percent of load occurs during 10 percent or less of the total hours in the year.
Figure 2 provides a useful rule-of-thumb; namely, maximum load deferred by DER should not exceed 15 to 20 percent. Thereafter, the curve flattens—hundreds or thousands of hours of DER operation then would be needed. For DR, the maximum number of hours generally is limited to less than 200 hours, which usually corresponds to the upper 5 to 10 percent of peak hours.
Utilities need absolute assurance that DG will be available when needed. Physical assurance provides greater certainty that customers who agree to operate DG when needed will reduce load, either by running the generator or by interrupting an equivalent amount of load.3 Establishing net level of firm DER capacity recognizes that DG may be out of service due to forced outages, including not starting when called upon to operate. The cost of communications and controls can be significant, and it must be added to incentive payments to derive total DER cost. Use of control systems apply only to larger technologies such as combined heat and power (CHP), as costs for small DG and renewables would be prohibitive. Thus smaller DG is less likely to provide the same percentage of firm capacity as larger DG with physical assurance.
Moreover, some DG operates only at peak. Exceptions include CHP, whose process steam and hot water load often require continued operation to be economic. Generation that is not available or experiences output reductions at the time of the system peak, including fuel-burn restrictions, will realize limited or no capacity credits.
Renewable generation such as PV produces maximum output during mid-day hours. If the system (or feeder) peak does not coincide with the time of maximum PV output, firm PV capacity must be reduced.4 Figure 3 illustrates that the peak PV output does not necessarily coincide with the feeder peak.
Similarly, DR availability could be limited when the value of capacity is highest. Restructured DR programs would permit interruptions for T&D capacity purposes, with higher interruption hours and fewer restrictions. DR would continue to be used for emergencies or to meet system-capacity requirements as well.5
A common industry practice is to price capacity seasonally, where dollars-per-kilowatt demand is multiplied by statistically derived allocation factors.6 Intermittent DER resources, such as DR or PV, are assigned partial capacity credits based on the hours and seasons they are most likely to operate. However, derivation of “firm” DER must account for unscheduled outages (DG) or customer non-participation (DR).
DER Incentives, Options & Models
Successful DER implementation is based on a premise that incentives and marketing efforts will increase DER participation sufficiently to defer proposed capacity additions. Regulators typically will allow recovery of direct incentive payments and marketing costs for DER programs in electric rates when benefits exceed program costs. Utilities may find it useful to conduct an active marketing campaign, targeted to areas with projected capacity shortages.
The type of DG most suitable to customers is highly dependent on demographics (particularly for renewables), electric prices, and customer type (residential, commercial, or industrial), fuel (resource) availability, interconnection standards, incentive structure (state, federal, and utility), and technology status.
Customer interest and willingness to purchase DG depends on load characteristics and economic value. In several states, customer acceptance of renewables such as PV is robust, as a combination of federal and state incentives, ideal weather, and customer support of clean generation has spurred interest.
Screening methods (see Figure 4) can predict penetration limits for customer-owned DG. In Screen 1, customers whose load patterns, location, or usage preclude participation are culled from the list. Screen 2 identifies customers with suitable load and interest in DG. Screen 3 identifies the remaining subset of customers for whom DG is economically viable.
For example, studies may determine that customer suitability and interest for a specific type of DG is 50 percent and 30 percent, respectively, for Screens 1 and 2, leaving a net potential of 25 percent. The third screen is estimated using market adoption curves that predict DG penetration as a function of payback. If incentives are set to reduce customer payback to 3 years, predicted penetration is about 40 percent, with total net DG penetration about 10 percent.
Planners also must recognize that adoption rates are time-dependent; that is, penetration targets such as those cited may take many years to achieve. For example, a straight-line adoption curve of 10 years means that only 1 percent actually will install DG each year.
Factors that limit DG penetration include load mismatch, over- or under-sized technology, low thermal demand, and unfavorable economics. For example, small PV may be the only DG technology that is viable and economically attractive to residential customers (see Table 2).
The Role of Demand Response
The notion that DR provides value to utilities and their customers is supported by several ISOs/RTOs. Unlike DG, DR may provide value to utilities, but only for a limited number of hours. Traditionally, DR is called when a system emergency exists or when marginal prices are high. Restructuring DR programs for capacity deferral would lead to increases in participation targets and hours of interruption. Increased incentive payments also may be needed to achieve participation targets.
Customer interest in DR can be fickle. Additional customer incentives may be needed to sustain customer participation and acceptance. Complete shutdown over an extended number of hours usually should be avoided. Rather, cycling of all participating customers during the peak intervals hours may be needed to avoid customer opt-out (e.g., 20 minutes on; 40 minutes off each hour).
Successful DR programs share certain attributes.
Residential. Typically, DR consists of air-conditioning (AC) and hot water control (HW), either by direct shutoff via timers or electronic signal (e.g., radio). Figure 6 illustrates typical residential DR reduction levels achieved in the United States. To increase penetration, additional incentives are necessary. Restructuring existing programs is straightforward; timers or interruption signals are reset to correspond to T&D peaks. The hours of interruption likely will increase to ensure a sufficient amount of AC or HW is interrupted during feeder or system peak hours.
Commercial. Commercial DR applications range from small offices to large shopping complexes, often for cooling or heating systems (HVAC). Limited tolerance for extended interruptions suggests DR should be cycled as owners will bristle at the prospect of customer inconvenience.
Industrial. Given the criticality of industrial processes—primarily the economic impact of lost production—DR incentives should be structured to limit the maximum hours of interruption to 100 to 150 hours annually. For commercial and industrial DR, firm DR estimates should reflect customer non-participation or lower than expected load reductions caused by idled plant or facilities.
Utilities need absolute assurance that DR can be relied upon in lieu of new T&D capacity. For direct-controlled HW or AC, minimum firm capacity is assured if devices are utility controlled. C&I incentives and penalties must be set to provide greater assurance that customers will reduce load when requested by utilities. Further, once outage exposure exceeds 100 to 200 hours, additional DR cannot be relied upon to provide firm capacity. From operating hour limits, the maximum level of firm DR equals about 5 to 10 percent of the feeder or substation peak.
The article has described methods to evaluate DER comparably with traditional T&D planning solutions. An economic methodology that consistently predicts and analyzes benefits and costs over time and DER ensures that all valuation metrics are fully considered. A rigorous and unbiased approach is essential if DER is to gain widespread acceptance by utilities and customers. Defendable DER valuation methods capture not just deferred T&D investments, but also avoided central generation, improved efficiency via reduced losses, and reduced Var demand. DER may be a better choice when benefits are sufficiently high to offset incentive payments, and the benefit-to-cost ratio exceeds those of traditional investments. The likelihood DER can provide a long-term solution is low, but can be valuable over the short term, typically 3 to 5 years. Extending the analysis to include indirect and difficult-to-quantify societal benefits could extend DER value beyond those cited in this article.
1. DER options include distributed generation (DG) and demand response (DR), but could include other technologies such as energy storage.
2. Utilities often accept the greater outage exposure of radial facilities when loads are low or in rural areas where long lines or absence of feeder ties do not justify the cost of redundant facilities. Some utilities set a load threshold—for example, 20 to 25 MW—above which redundant or back-up facilities are installed.
3. Physical assurance is an arrangement whereby customers agree to interrupt load in an amount equivalent to the capacity rating of the generator as a condition of receiving capacity payments from the electric utility. Isolation is performed by utility distribution control center staff, if needed, when DG is needed but off line.
4. Statistical studies that utilize historic weather data can provide equivalent availability factors; 20 percent is often cited for PV. If variable tilt with solar tracking is employed, the equivalent PV availability can increase dramatically—up to 60 percent in some regions.
5. Often, the system peak is coincident with the day and time of the feeder or substation peak.
6. For example, probability of dispatch employs loss-of-load expectation methods to estimate the value of generation at any given hour.