Demand Response and Energy Efficiency For Silicon Valley Power
Final Report Volume 3 of 4: Demand Response and Distributed Generation Potential Estimates Rocky Mountain Institute
May 2007 Section 1: Introduction ________________________________________________ 2
SVP's existing DR and DG resources__________________________________________ 2 Section 2: Distributed Generation _______________________________________ 3
Background on Distributed Generation ________________________________________ 3 Distributed Generation economics ____________________________________________ 4
a. Site characteristics and timing_____________________________________________________ 4 b. Interaction with the grid __________________________________________________________ 4 c. Fuel cost (spark) spreads ________________________________________________________ 4 d. Ability to capture waste heat ______________________________________________________ 5 Implications for SVP ________________________________________________________ 5
DG economics for SVP ____________________________________________________________ 6 DG considerations ________________________________________________________________ 7 Section 3: Demand Response___________________________________________ 9
Key Findings ______________________________________________________________ 9 Summary: Why DR may not be very attractive to SVP ____________________________ 9 Brief Description of Demand Response_______________________________________ 10 Understanding SVP's Load Profile ___________________________________________ 11 Estimate of DR Potential and Best Candidates for DR Program Implementation_____ 13
Residential and Small Commercial Customers _________________________________________ 13 Industrial Customers _____________________________________________________________ 14 Large Commercial Customers ______________________________________________________ 15 Potential Savings for SVP from DR __________________________________________________ 20 Cost-effectiveness of DR Programs _________________________________________________ 23 Summary of Findings_____________________________________________________________ 23 Section 4: Conclusion ________________________________________________ 24 Appendix A: Data and Analysis ________________________________________ 26 Appendix B: ________________________________________________________ 30 Additional Information ________________________________________________ 30
Types of DR Programs ___________________________________________________________ 30 Cost-Effectiveness of DR Programs _________________________________________________ 31 Program Design_________________________________________________________________ 33 Designing DR program to meet customer and facility needs ______________________________ 34 DR Synergies and Considerations __________________________________________________ 35 1 Section 1: Introduction
Silicon Valley Power (SVP), a municipal utility in Santa Clara, California, has retained Rocky Mountain Institute (RMI) to obtain a good estimate of the potential for demand-side management--including energy efficiency, demand response, and distributed generation--in its service area. This project encompasses the following sections: 1. Volume 1: Discussion and review of SVP's system, existing strategic plan, and Operating Study, as well as a discussion of RMI's approach to the project. 2. Volume 2: Analysis results of SVP's cost-effective, achievable energy efficiency potential. 3. Volume 3: Analysis results of SVP's cost-effective distributed generation potential and demand response. 4. Volume 4: Implementation strategies for capturing cost-effective demand-side resource potential. Volume 1 of this report established the overall approach to the analysis included here. Volume 2 described the analysis of cost-effective achievable energy efficiency potential for SVP. This report is Volume 3, and includes a summary of RMI's estimate of SVP's demand response (DR) and distributed generation (DG) potential. Implementation issues including strategies for capturing cost-effective DR and DG potential will be discussed in Volume 4. SVP's existing DG and DR resources
SVP does not currently own any DG capacity but the utility has some limited DR resources. While some of SVP's customers do have back-up generators on-site, SVP neither owns nor operates the equipment. With respect to DR capacity, SVP provides an interruptible rate for one customer. SVP also provides some information on their website requesting that customers reduce their load during statewide power alerts. 2 Section 2: Distributed Generation
Background on Distributed Generation
Distributed generation (DG) generally refers to small-scale electricity generation close to the end user, and can take a number of forms, including: · · · Back-up generation--many of SVP's large customers have back-up generation systems that do not run in parallel with the grid, but can provide emergency power in the case of a grid outage. Combined heat & power --combined heat & power (CHP) utilizes a prime mover to generate power and waste heat, which is captured and used to displace some portion of the end user's heating load. CHP generally runs in parallel with the grid and trips in the event of a grid outage. Combined cooling, heat & power --combined cooling, heat & power (CCHP) utilizes a prime mover to generate power and waste heat, which is captured and used to displace some portion of the end user's heating or cooling load. Cooling load is displaced through the use of an absorption chiller. Like CHP, CCHP generally runs in parallel with the grid and trips in the event of a grid outage. It appears that the largest potential cogeneration system at a single customer site could be sized at approximately 5000 kW in terms of electric generation. For reasons of practical and economic limitations, the minimum size is generally 5000 kW, although there may be significant potential also in the 250-500 kW range. In the 500-5000 kW range, there are three primary generation technologies: combustion turbines, reciprocating engines, and microturbines. The basic performance and cost assumptions about each are shown in Table 1. Table 1. Cogeneration Performance and Cost Assumptions Parameter Capacity, kW Heat rate, Btu/kWh Heat rate with cogen, Btu/kWh Capital cost, $/kW O&M cost, $/MWh Absorption chiller cost, $/ton Absorption chiller COP Combustion Turbine 4500 10,000 5,000-8,000 1200 5 650 1.1 Reciprocating Engine 800-3000 10,500-11,000 5,000-8,000 1000-1100 7.5 750-850 1.1 Microturbine 480-720 14,000 7,000-10,000 1900 10 950 1.1 Fuel Cell 500-2000 8000-9500 4,500-7,000 4000-5000 25 850-950 1.1 Sources: RMI estimates and EPRI, Economic Costs and Benefits of Distributed Energy Resources, EPRI Technical Update 1011305, November 2004. The choice of prime mover technology largely depends on the end user's capacity requirements, as well as a number of economic factors, as discussed in the following section. 3 Distributed Generation economics
Meeting electricity demand through DG resources such as cogeneration can have many potential benefits for both the utility and the customer. 1 Realizing the potential value of DG, however, is by no means a guarantee. The following sections discuss key factors in determining whether DG is cost-effective or not. a. Site characteristics and timing
The cost-effectiveness of DG is highly site specific and depends on the project timing. DG is most cost-effective when it can be sited in capacity-constrained areas or locations where new, likely expensive, distribution capacity is planned and could potentially be avoided by careful siting of DG resources. Because DG is sited closer to--and sometimes literally on--customers' premises, it often must operate quietly and have low emissions of air pollutants. For this reason, natural gas is often the preferred choice of fuel for on-site DG since it produces low emissions compared to other fossil fuels. Nevertheless, obtaining clean, cost-effective DG technology for on site use can be challenging, and installing additional emissions and noise controls and permits on existing, conventional DG technology can sometimes be cost prohibitive for the investor. DG value is highest when load growth is high or uncertain, and helps defer capacity expansion only if the resource is available at times of area peak load. b. Interaction with the grid
In some locations, DG can contribute to system reliability by alleviating congestion and providing smaller, more diverse sources of power close to load. In other locations, DG can pose a challenge to system operation and reliability due to frequency variations, transmission capacity demand in case of failure, and other factors. Furthermore, customers in particular often lack the resources to routinely maintain on-site DG technology for optimal operations and grid compatibility. These issues pose significant concern to SVP, since its high system reliability is a major attraction to its commercial and industrial customers. c. Fuel cost (spark) spreads
Generally, the greater the spread between electricity and gas prices, the more cost-effective the investment in on-site CHP or CCHP. This is because DG reduces the quantity of electricity purchased from the grid, but results in a net increase in natural gas purchases to fuel the DG (despite a potential reduction in gas purchased for heating). Therefore, avoiding high electricity 1 For the utility, distributed scale generation (DG) is cheaper and faster to build, starts generating revenues faster, and match loads more closely than traditional utility-scale power plants. DG can serve to increase system reliability and extend the lifetimes of existing utility-scale power plants. For the customer, on-site DG with combined heat and power generation (CHP) can improve both the reliability and quality of electricity service, offset additional natural gas purchases for facility heating needs and/or offset electricity use for space cooling, and potentially earns additional revenue from selling excess electricity back to the grid. 4 prices in exchange for purchasing low-cost natural gas can make DG an attractive investment. 2 On the other hand, avoiding low electricity prices in exchange for purchasing high-cost natural gas is not generally cost-effective. The ratio of the purchased electricity price ($/kWh) to the purchased natural gas price ($/MMBtu) is also called the "spark spread", and can be thought of as the net heat rate (Btu/kWh) of the DG unit at which the customer is indifferent between buying gas and generating electricity. d. Ability to capture waste heat
For CHP and CCHP, cost-effectiveness depends strongly on the ability to capture and use all of the waste heat generated by the prime mover to offset costs of other purchased energy (either natural gas for heat or electricity for cooling3). The ability of CCHP to supply both heating and cooling loads makes the thermal load larger and less variable seasonally, which tends to increase both the sizing and efficiency of the cogeneration system. Because CHP and CCHP provide power and thermal energy at the customer site, the sizing and performance are typically limited by the customers' demand for electricity and heating or cooling, as indicated by their electric and natural gas billing data. The ability to capture all of the waste heat created by the DG prime mover is key to the sizing of the DG unit, and it is therefore important that the customer have a relatively flat thermal load over the course of a year. That is, if a customer has a high heating load in the winter, and no thermal load in the summer, it will not be possible to utilize the DG waste heat in the summer, and that energy will be lost. The sizing of the prime mover is typically either 50-70 percent of the electric demand or 70-80 percent of thermal load. In both cases, and given the technology availability in Table 1, the ideal customer for cogeneration is typically a large commercial or industrial customer larger than 1000 kW in demand with relatively constant thermal loads year-round. Implications for SVP
Given that SVP is an electricity provider only, no customer-specific natural gas consumption or price data was available. Therefore, it is not possible at this time to identify good candidates for CHP or CCHP, or whether there is significant cost-effective potential in SVP's service area. However, several observations can be made that will help inform future analysis and decisions regarding the viability of DG for SVP. 2 Because it is not a gas utility, SVP may not be able to take advantage of wholesale prices to arbitrage the spark spread. Nonetheless, the option of purchasing low-cost natural gas can still be valuable to SVP customers and therefore SVP should be sure to explain this advantage to any customers considering DG. 3 In combined heating, cooling, and power generation (CCHP), some of the waste heat energy recovered from the electric generator is used to power and absorption chiller for space cooling. An absorption chiller uses a thermal compressor (consisting of an absorber, a generator, a pump, and a throttling device) rather than an electricity-driven, mechanical vapor compressor. 5 DG economics for SVP
Typical accounts with significant and constant thermal loads in SVP service area include research laboratories, manufacturing, hospitals, data centers,4 and traditional industrial. These types of customers comprise approximately 60 percent of SVP's service territory load, with most of these accounts with electricity demand larger than 1000 kW. Figure 1. Load profiles of representative SVP large customers Examining SVP's current electricity rates for large commercial customers and commodity prices for natural gas as reported by PG&E for large commercial and industrial customers,5,6 the spark spread is large enough such that CHP/CCHP could be economically feasible for large commercial customers. In absence of detailed customer data, RMI estimated the economics of stand-alone DG for a generic customer with operational characteristics shown in Table 2, assuming the technology cost and performance parameters in Table 1. RMI estimated that economics of stand-alone DG without waste heat capture is not attractive under the majority of cases, assuming the technology cost and performance parameters in Table 1. Allowing for waste heat capture, SVP-owned cogeneration at the largest customer sites would cost about $0.07/kWh for gas turbines and engines and about $0.09/kWh for microturbines, assuming a natural gas fuel cost of $8/MMBtu. 7 This conservatively assumes a net heat rate of 7,000 Btu/kWh and cogeneration capacity factor of 65 percent. This compares favorably with SVP's retail electric rates of $0.087/kWh for large customers and its avoided energy costs of $0.07/kWh off-peak and $0.13/kWh on-peak. 4 5 Where the thermal energy is used to drive a chiller and help manage the air conditioning load. PG&E current and historical natural gas rates for large commercial customers G-NR2 (Jan 2006 ­ Present). Retrieved April 7, 2007 from http://www.pge.com/rates/tariffs/GRF.SHTML#GNR2. 6 PG&E tariff Gas Transportation to Electric Generation (Apr 2004 ­ Present). Retrieved April 7, 2007 from http://www.pge.com/rates/tariffs/GRF.SHTML#GEG 7 Over the past year, retail gas rates have varied between $6/MMBtu and $12/MMBtu. From: PG&E current and historical natural gas rates for large commercial customers G-NR2 (Jan 2006 ­ Present). Retrieved April 7, 2007 from http://www.pge.com/rates/tariffs/GRF.SHTML#GNR2. 6 Table 2. Levelized Cost of Energy for Generic SVP Commercial Customer, Assuming Utility Ownership of DG Fuel Cost, $/MMBtu Parameter 8 No Cogen With Cogen No Cogen 12 With Cogen 65 Capacity 30 65 30 Factor, % Levelized Combustion 0.12 0.075 0.17 0.11 Cost of Turbine Energy, Reciprocating 0.13 0.075 0.18 0.11 $/kWh Turbine Microturbine 0.19 0.077 0.25 0.12 Fuel cell 0.17 0.090 0.21 0.14 *All calculations assumed a weighted average cost of capital (WACC) of 5% assuming equity cost of 5% and debt cost of 3% Assuming waste heat capture for cooling and heating increases the capital cost of the project, but also decreases the net heat rate of the system. We estimate the net effect will be a marginal increase of about $0.01-$0.02/kWh. Detailed data on natural gas consumption and cooling loads are needed, however to develop a more precise answer. Again, distributed generation economics are very site specific, and there will be sites for which the economics will be even more favorable. DG considerations
Customers eligible for cogeneration can either own and operate the equipment themselves or contract with a third party. Many facility managers find the contracting model desirable because they lack the resources to navigate the details of building, owning, and operating distributed generation. Customer ownership of DG is also more difficult to justify economically, as their rate of return requirements are much higher than utilities and public entities. While SVP does not provide natural gas service to its customers, this does not necessarily preclude them from owning DG at its customer sites. Under utility ownership, there is more operational flexibility and therefore the types of benefits that can be counted economically. The lower discount rate helps, but the key is that the utility sees the full marginal avoided costs of electric supply, and use of waste heat to meet the facility's thermal loads, including cooling. Particularly if the use of waste heat to meet thermal loads allows the project to be sized such that excess electricity can be fed to the grid, it can provide additional benefits in areas of rapid load growth or constrained capacity. In Santa Clara, the default natural gas provider is Pacific Gas & Electric (PG&E). The industry is deregulated, however, and most large customers enter into supply contracts with third party natural gas suppliers to be delivered through PG&E pipelines. It is conceivable for SVP to also enter into such contracts with the existing or competing gas suppliers at lower cost, offer to own and operate cogeneration at the customer site, and sell both the electricity generated and waste heat at a cost that is lower than the customer is currently paying. SVP in turn receives the 7 increased distribution reliability along with avoided energy and/or capacity benefits. SVP currently owns gas turbine plants and thus has experience contracting with natural gas suppliers. However, separate infrastructure may be needed to provide scheduling and balancing services for these large retail customers. 8 Section 3: Demand Response
This deliverable analyzes potential and opportunities for developing demand response (DR) programs in the SVP territory. Our assessment is based primarily on SVP's customer data and will inform RMI's recommendations regarding demand-side management programs appropriate for the system. Key Findings
· Given SVP's high load factor (about 70 percent for 2006) and the fact that the utility is not capacity constrained, SVP does not appear to have much to gain from implementing DR programs. RMI analysis shows that, if SVP was to implement a DR program, targeting their large commercial customers could provide an estimated technical potential of 21 MW of DR. Effectively calling on this DR capacity for seven events during July of 2006 could have saved SVP a maximum of $102,000 in energy purchases.8 According to RMI analysis based on DR potential by end-uses, "lodging" accounts could realistically provide 7 MW of DR potential. This is the largest share of DR potential in the "large commercial" sector. Because there are relatively few lodging customers, 9 it may be easier and more cost-effective to enroll customers from this sector in DR. Based on these facts, lodging might be an attractive sector to involve in SVP's DR programs. There is likely additional potential from SVP's residential, industrial and small commercial customers but further analysis would be needed to estimate how much. Whether it is cost-effective for SVP to implement a DR program largely depends on the cost of acquiring and maintaining DR capacity, and any technology cost required to implement the program. SVP should also consider evaluating whether DR can help relieve distribution congestion within SVP's territory. As SVP's load and the California energy market change, DR could potentially become more valuable and SVP should continuously re-evaluate potential benefits from DR. · · · · · · Why DR May Not Currently Be Very Attractive to SVP
SVP has a relatively flat demand curve on both a daily and annual scale. Although hot summer afternoons increase demand within SVP's territory and drive-up spot-market energy prices significantly, RMI analysis indicates that the benefits from DR for SVP are not large enough to justify implementing large-scale programs. RMI came to this conclusion by first estimating the technical potential of demand reductions that SVP could potentially achieve and found that SVP could realistically achieve 21 MW of DR
8 These savings are for calling only those seven events in July 2006, there may have been other occasions that would warrant calling DR events during 2006 and doing so could have provided additional savings. 9 Out of a total of the 1160 large commercial customers with demand data, 30 of these were lodging. Note, however, that RMI does not have access to data on "willingness to participate" from this sector. 9 capacity in the large commercial sector. In order to generate an estimate of monetary savings to SVP, RMI then used real energy market data from 2006 to calculate the potential savings that SVP could have experienced from reducing demand by 21 MW. RMI found that during July 2006, 21 MW of DR reductions would only save the utility about $102,000 for that month. This was even under the fairly generous assumption that SVP would be able to capture most of the benefits of reducing energy purchased and increasing energy sold during hours of high spotmarket prices. Although current analysis indicates that there is not much incentive for SVP to implement largescale DR programs, it is possible that additional factors or future events could make DR more valuable. For instance, if there are areas on SVP's distribution that get particularly overloaded at a certain points of the day, targeted DR programs could potentially help relieve congestion. Future scenarios, such as increased demand within SVP's territory or increased peak energy prices in the California energy market, could also make DR programs more financially enticing for SVP. It is also worth noting that the main scope of the analysis presented here is with respect to SVP's large commercial customers. Due to insufficient data, DR potential for the small commercial, residential and industrial sectors are only addressed qualitatively in this report. Quantifying SVP's full DR potential and savings would require further data and analysis of these sectors. Brief Description of Demand Response
Demand Response (DR) refers to demand-side management programs that offer customers incentives to reduce their demand for electricity during periods of either critical system conditions or high market power prices. Utilities can sometimes purchase reductions in load demand from their customers for lower rates than it would cost them to serve the customer's load. Reducing peak loads lowers the utility system's peak demand and can help reduce peak wholesale power market prices, depending on the price exposure that the utility faces. Additionally, customers can also be asked to reduce load during non-peak periods to help maintain grid reliability and to defer or eliminate expansion of both generation capacity and transmission/distribution. DR is not an efficiency measure and is not intended to reduce total energy consumption in a utility's territory. Instead, DR programs enable utilities to shift energy consumption from times when energy is scarce and expensive to times when cheaper energy is available. DR can prove particularly valuable in situations where the system peak is much higher than base-load demand. Utilities like SVP that purchase from or sell into energy markets with price peaks for spot energy can also benefit from DR programs. In such cases, utilities can save money by using DR to reduce the amount of expensive energy purchased during short periods of elevated prices (more in-depth discussion of types of DR programs is included in Appendix D). 10 Understanding SVP's Load Profile
DR technical potential depends on the characteristics of the utility system and the energy market in which it participates. Following is a discussion of SVP's system trends and characteristics. On average, the peak period remains centered around 11:00 -20:00 hours. In 2006, the peak load was 486 MW, which occurred on July 25 ( Figure 2). Meanwhile, SVP's average load for that same year was 329 MW. According to this information, SVP's load factor in 2006 was approximately 70 percent. This is a very high load factor, and likely indicates that SVP's cost of producing power is not significantly higher during the peak period. That said, NP15, the market in which SVP participates, does experience high peak prices. Therefore, DR could help avoid purchasing energy during such instances of elevated prices (this topic is discussed later in this report). Figure 2. Hourly Load for Day with Highest Peak in 2006 As shown in Figure 3 by the load duration curve10 for SVP (Figure 3), in 2006 there were only a few hours of very high demand during the year. Specifically, the load was greater than 450 MW for only 50 hours of the year and above 400 MW for about 410 hours. Figure 4 is a more granular chart showing the 1000 hours of highest demand for SVP. 10 The load duration curve shows the number of hours that system was above a certain power demand. It is similar to a histogram that represents every hour of system demand ordered from the hour with the highest power demand to the hour of lowest power demand. 11 Figure 3. SVP Load Duration Curve, 2006 Figure 4. Hours of Highest Demand, 2006 The 50 hours above 450 MW occurred over 10 days during the summer and occurrences lasted between two and eight continuous hours (see Appendix A for a list of the time and date of the 50 hours with the highest demand for the SVP territory as a whole in 2006). 12 Estimate of DR Potential and Best Candidates for DR Program Implementation
Residential and Small Commercial Customers
Because SVP only records demand data for large business customers (namely, customers that consume more than 8000 kWh for three consecutive months), it is difficult to estimate demand savings potential for its smaller customers. Small commercial and residential customers constitute a relatively small share of energy consumption in SVP's territory (the residential sector consumed approximately 10% of the energy sold by SVP in 2006, while small commercial customers 11 only accounted for about 2% of SVP's energy consumption) and therefore probably do not offer a huge amount of DR potential. Furthermore, the dispersed nature of DR opportunities in these sectors may make implementing DR programs costprohibitive. In order to address this issue, however, SVP could consider partnering with a DR aggregator to simplify the implementation of DR programs. Despite some of the problems related to implementing programs for many small customers, small commercial and residential energy users may provide some relatively cheap opportunities for DR. Capturing DR capacity from these sectors could require fairly simple, cost-effective technology that could shed a small amount of non-essential load from the accounts. In any event, given that small commercial and residential customers comprise only a small share of SVP's load, these sectors would likely not be a high priority target for any DR programs that SVP chose to implement. Figure 5. SVP's Energy Consumption (all sectors), 2006 11 Small commercial customers, in this case, refers to commercial customers without demand data, i.e., commercial customers generally consuming less than 8000 kWh per month. Subsequent sections of this document look more closely at DR potential for large commercial customers (large commercial customers being those that do consume more than 8000 kWh and therefore do have demand data). 13 Industrial Customers
Estimating DR potential from industrial customers requires additional information about the particular customer's energy profile. Many traditional industrial customers, however, can provide significant amount of DR potential. Large industry can often shed a lot of load by rescheduling production shifts or running on-site generation when it is available. One way to estimate a share of DR potential from industrial customers could be to determine the total quantity of back-up generation used by these facilities. The fact that so many of SVP's industrial customers are high-tech companies also plays a role in how much DR they might be able to provide. High-tech industries place a high value on reliability and some of these may therefore have their own generators (that said, SVP has attracted many customers precisely because of their emphasis on providing reliable power, therefore their customers may not have felt the need to install redundant systems). Those that do have back-up power, however, probably place a high value on having on-site power at all times and would be unlikely to relinquish control of their generators. Convincing high-tech customers to participate in DR programs may also be difficult because it would be unlikely that utility programs could provide incentives significant enough to affect the economics of their facilities. Generally speaking, within the high-tech sector, data centers are probably the least likely to provide much DR because they have a relatively constant, flat load. Chip manufacturers and labs, however, may be more able to shift some of their loads when needed. Figure 6. Sum of Monthly Peak Demand for Industrial Sector 14 Large Commercial Customers
As mentioned above, SVP only records demand data for large customers. For commercial customers, demand data is available for customers whose monthly energy use is more than 8000 kWh for three consecutive months. In these cases a monthly Maximum Demand is recorded (and a $5.69 per kW-month charge applied). 12 In 2006, although the number of commercial accounts qualifying for the Maximum Demand charge was only 26 percent of commercial accounts, these accounts cumulatively consumed 90 percent of energy sold to the commercial sector. As Figure 7 shows, in 2006 these large commercial customers demonstrated a clear summer peak for the month of July. This is consistent with SVP's overall system peak, which also occurred during the month of July and was likely driven in large part by the commercial sector. Figure 7. Sum of Monthly Peak Demand for Large Commercial Customers 12 From SVP Rate Schedule CB-1. Note that that if monthly energy use drops below 6,000 kWh for 12 consecutive months, the Maximum Demand meter will be discontinued. 15 Figure 8 shows the sum of maximum demand by building type13 and Figure 9 shows the average maximum demand per account in each building type; both are for SVP's large commercial customers in 2006. As these charts demonstrate, lodging has the highest sum of maximum demand and, because there are only 30 accounts with demand data in this category, lodging also has high average maximum demand per account. The average peak demand for colleges is also high since, although the sum of total peak demand is relatively low, there are only six accounts in this category.14 Figure 8. Peak Demand for Large Commercial Customers, 2006 13 Sum of maximum demand for all accounts of each given building type. Note that maximum demand is each account's highest recorded demand. The data is not able to indicate when that demand occurs and it is therefore likely that not all accounts have coincident maximum demands. Therefore, the instantaneous sum of demand for large commercial customers is very unlikely to ever have been as high as the sum of their recorded maximum demand. 14 Note that in some cases there are multiple accounts for one customer. This is the case for colleges where Santa Clara University is the customer associated with all but one of the accounts in this category of large commercial (the one other college account belonging to West Valley College). In the case of lodging there is, for the most part, a one-to-one correlation between accounts and customer name. 16 Figure 9. Average Peak Demand Per Account for Large Commercial Customers Figure 10. Number of Accounts with Peak Demand Data for 2006
Health Large College Grocery Care Lodging Office Misc. 6 66 38 30 130 171 Ref. WareRestSmall Warehouse Retail aurant School Office house 1 132 144 31 320 91 17 RMI used information from SVP's energy audits to generate average end-use breakdowns by building type for the commercial sector (this information was also used for the efficiency modeling). These end-use breakdowns are detailed in Volume 2 of this report. Using time of use load shapes developed in the California Energy Efficiency Potential Study for PG&E, RMI generated estimates of peak demand breakdowns by end-use for each building type. RMI then used estimated percent electricity reduction potentials for the different end-uses to generate peak demand reduction potentials for each end-use. The process is summarized in Figure 11. The final SVP demand potential by building type is compiled in Table 3. Figure 11. Process for calculating SVP's DR potential for large commercial customers Table 3. SVP DR potential by end-use building type15
Total DR potential 1.52 MW 0.83 MW 1.85 MW 7.25 MW 1.13 MW 0.10 MW 1.39 MW 2.23 MW 0.62 MW 2.42 MW 2.11 MW 21 MW Building Type College Grocery Health Lodging Large Office Ref. Warehouse Retail Restaurant School Small Office Warehouse TOTAL: Cooking 0.01 MW 0.00 MW 0.02 MW 0.06 MW 0.001 MW 0.09 MW 0.002 MW 0.002 MW 0.01 MW HVAC 0.91 MW 0.11 MW 1.20 MW 4.78 MW 0.76 MW 0.001 MW 0.76 MW 0.98 MW 0.31 MW 1.44 MW 0.57 MW Lights 0.48 MW 0.10 MW 0.40 MW 1.46 MW 0.19 MW 0.04 MW 0.54 MW 0.43 MW 0.28 MW 0.78 MW 1.17 MW Misc. 0.06 MW 0.01 MW 0.15 MW 0.42 MW 0.18 MW 0.02 MW 0.07 MW 0.09 MW 0.17 MW 0.21 MW Refg. 0.07 MW 0.60 MW 0.08 MW 0.53 MW 0.002 MW 0.04 MW 0.01 MW 0.65 MW 0.03 MW 0.04 MW 0.15 MW 15 Note that water heating is not included because if most water heating is assumed to be natural gas, there is no DR potential for this end-use. 18 As the table above shows, there is total estimated DR potential is about 21 MW with lodging being the building type with the greatest DR potential of about 7 MW. In order to evaluate the most desirable candidates for implementing DR programs, RMI considered that it is easiest and cheapest to implement DR programs with a few, large customers that are similar to one another, as opposed to trying to implement programs with many smaller and more diverse types of customers. Following this reasoning, RMI looked at the number of accounts in each of the building types compared to the DR potential by building type. Figure 12 compares the number of accounts by building type with the cumulative DR potential by building type. The most attractive targets by building type are located in the upper left-hand corner of the chart. The building-type categories in the upper left of the chart have both large DR potential and few number of accounts (and therefore are better targets for DR because they have higher DR capacity per account). As the chart demonstrates, the lodging sector not only had the highest DR potential but it also had relatively few accounts. Based on these facts SVP's lodging customers might be an attractive candidate for DR programs. SVP could also consider targeting Santa Clara University (especially since this one customer encompasses the bulk of DR potential in the college sector) as well as the health facilities in their territory and large restaurants, warehouses, retail or large office customers. One potential limitation to DR from lodging customers could be the fact that many hotels could be hesitant to take any action that might diminish guests' comfort. A corollary to this hypothesis is that the willingness of hotels to participate in DR events may be related to their occupancy level. All this said, RMI has observed that DR programs can be effectively implemented in the lodging sector. 19 Figure 12. DR Potential vs. Number of Accounts by Building Type for Large Commercial Customers Increasing attractiveness of building types for DR programs implementation Potential Savings for SVP from DR
DR programs could potentially save SVP from consuming expensive energy during market peaks. Savings from DR programs are dependant on the amount and price of energy purchased, or potentially sold, by SVP on the market. Because RMI did not have specific data pertaining to the amount and price for energy purchased by SVP, we instead analyzed data from the CA ISO on market energy prices. As the following section will demonstrate, RMI found that, even using the spot market energy prices during price peaks (which is probably one of the highest metrics for the value of reduced power demand), DR does not appear to provide very significant financial savings. Since real time prices for energy on the California ISO are currently capped at $400/MWh, this value would be the maximum value of reducing energy demand. Although $400/MWh is the upper bound of the value of saving on-peak energy, market data from 2006 indicates that it is a realistic value nonetheless. Specifically, during the summer days with highest demand, the maximum real-time price of energy in NP15 was frequently $400/MWh and on occasions the real-time hourly average market-clearing price (MCP) also reached this cap. The charts below are taken from the California ISO's Market Watch for Market Participants for July 26, 2006 and illustrate the spot price of energy and the day-ahead prices for Ancillary Services respectively. 20 Figure 13: July 26, 2006 Real Time Energy Prices16 Figure 14: July 26, 2006, Day Ahead Ancillary Services Market Clearing Prices17 Note that the charts above are for July 26, 2006, one day after SVP's system peak that year. Real-time hourly average MCP for July 25th were lower than those shown above for the 26th
16 17 From California ISO's Market Watch For Market Participants For Operating Day 7/25/2006. Ibid. 21 (though the maximum price did reach $400/MWh, the highest hourly average was about $250). Day-ahead A/S MCP did reach the $400/MW level on July 25th and on the 26th. 18 For 2006, July appeared to have some of the highest market prices for on-peak energy. California ISO shows that daily peak­hour bilateral contract prices for 2006 were highest in July. As the figure below shows, the month of July hit the $400/MWh price cap for settlement prices of energy at least a dozen times. This is also consistent with SVP's own elevated demand during this month; in fact, as shown in Appendix A, out of SVP's top 50 hours of highest demand in 2006, 41 of them occurred during the month of July. Figure 15: NP15 Real Time Dispatch and Settlement Prices for July, 200619 In order to calculate an estimate of potential savings from DR for SVP, we looked at how much the utility would have saved by calling seven events each lasting of four or five hours during the month of July of 2006. We selected the days to model these DR events by looking at the figure above and selecting the days during which NP15 settlement price reached $400/MWh. Using the hourly averages during the major peak in these days in July, we then estimated how much money could be saved by calling a DR event lasting four, or in some cases, five hours (depending on the duration of the particular peak). In order to estimate different ranges of
18 "Daily Market Watch Reports" are available starting January of 2006 at http://www.caiso.com/186a/186acc7538710.html. 19 From California ISO, Market Performance Report for July 2006. 22 savings, we estimated DR potential at 7 MW (about equivalent to all potential in lodgings), 14 MW and the full estimated potential of 21 MW. The total potential savings in July of 2006 alone were approximately $34,000, $66,000 and $102,000, respectively, for the three DR potentials calculated (for the full calculations, see Appendix C). Table 4: Potential Energy Savings from calling seven DR events in July 2006. Assuming 7 MW of DR capacity $34,000 Assuming 14 MW of DR capacity $66,000 Assuming 21 MW of DR capacity $102,000 Cost-effectiveness of DR Programs
In order to successfully implement DR programs of the scale and scope described above, SVP customers would have to install monitoring and control technologies and SVP would probably have to help finance those installations. On the utility end, SVP would probably also have to acquire technology that would allow them to control and facilitate load reductions. The size of the initial investments will, in a large part, depend on the kind of DR program to be implemented. If SVP decides to implement any DR programs, they should conduct further analysis to establish what the best type of program and fitting technology might be. RMI has provided some general information on DR program costs in Appendix B. Summary of Findings
RMI has determined that there does not to appear to be a significant advantage for SVP to implement large-scale DR programs. However, further analysis is necessary to determine the full DR potential in industrial, residential and small commercial sectors. SVP should also consider that there may be specific applications where DR could still prove helpful to SVP. DR may also be more attractive in the future and SVP should continue to consider it an option as the utility's load and California's energy prices change with time. 23 Section 4: Conclusion
From RMI's current analysis, demand response and distributed generation (DR and DG) seem to offer little benefit to SVP. Potential savings from DR do not currently look convincing enough to justify large scale programs. However, it may be worth noting that developments within SVP or in the larger California energy context (e.g. predictably high prices for peak capacity or energy sales) could make DR dramatically more appealing for SVP in the future. The case for DG within SVP's territory is contingent on the heat from distributed resources also being captured and used in a combined heat and power system. Therefore, despite the requirement that SVP purchase the gas from another utility, CHP could make sense for serving large commercial customers with constant thermal loads. Such DG resources can also be particularly valuable in areas of rapid load growth and constrained capacity, or areas soon to be in need of distribution upgrades. Customer-hosted CHP can provide a variety of benefits that would accrue to SVP: · Strong business case, lots of attractive potential hosts Examining SVP's current electricity rates for large commercial customers and commodity prices for natural gas (as reported by PG&E for large commercial and industrial customers 20,21), it appears that the spark spread is large enough to potentially support CHP/CCHP projects for large commercial customers with constant thermal loads. This would include customers such as research laboratories, manufacturing, hospitals, data centers, 22 and traditional industrial. These types of customers comprise approximately 60 percent of SVP's load, and most of these accounts represent an electricity demand larger than 1000 kW. The ability of SVP to procure financing at rates lower than the average commercial customer is also a great help in these finances. Increased reliability Increased dispersion of SVP's generation sources intrinsically leads to higher reliability in several manners. First, diversity of the generation fleet reduces the potential for all of the generation assets to be down at the same time. Second, geographic dispersion of the generation fleet can provide a stronger basis for the grid, depending on where the host customers are located. Decreased distribution congestion As mentioned above, by targeting prospective host customers in particular sectors of SVP territory, CHP can reduce congestion problems and at least defer substation upgrades. In the same way, CHP can also be a relatively rapid method of responding to load growth in concentrated sectors of the grid. · · 20 PG&E current and historical natural gas rates for large commercial customers G-NR2 (Jan 2006 ­ Present). Retrieved April 7, 2007 from http://www.pge.com/rates/tariffs/GRF.SHTML#GNR2. 21 PG&E tariff Gas Transportation to Electric Generation (Apr 2004 ­ Present). Retrieved April 7, 2007 from http://www.pge.com/rates/tariffs/GRF.SHTML#GEG 22 Where the thermal energy is used to drive a chiller and help manage the air conditioning load. 24 · Reduced GHG emissions By improving the overall efficiency with which the gas is used, SVP can potentially capture the GHG offsets for the reduced emissions associated with CHP vs. independent generation and industrial heat. Of course, SVP is a municipal utility, and as such needs to have the best interests of its owner/customers at heart. CHP offers significant advantages to both the host customer and to the customer base generally. · Reduced cost for process heat loads Typically it is difficult to offer a host customer a significant rate break on the electric power without incurring charges of favoritism. However, the pricing of the process heat (typically steam) can usually be set in such a way as to incentivize the host company. This advantage can be used strategically to encourage the ongoing presence of a valued industry that provides jobs and economic benefits to the community. Increased reliability and improved overall economics High reliability is a key feature for SVP with the predominance of high-tech and data center customers. Anything SVP can do to cost-effectively increase reliability while simultaneously improving the overall economics of power generation will, ultimately, benefit the community as a whole. · Overall, CHP (or CCHP) can help SVP improve reliability, reduce the cost of providing heat and power, defer or eliminate congestion problems and provide a means of meeting load growth more effectively. In addition, SVP's customers will gain increased reliability while also reducing their overall energy expenditures. For these reasons, RMI recommends that SVP consider using CHP/CCHP as a strategic option with its larger industrial customers. 25 Appendix A: Data and Analysis
Highest Hours of Demand for SVP in 2006 (listed in chronological order)
Date 6/21/06 6/21/06 6/22/06 6/22/06 6/22/06 6/22/06 7/17/06 7/17/06 7/17/06 7/17/06 7/17/06 7/18/06 7/18/06 7/18/06 7/18/06 7/19/06 7/19/06 7/19/06 7/19/06 7/19/06 7/20/06 7/20/06 7/20/06 7/20/06 7/21/06 7/21/06 7/21/06 7/21/06 7/21/06 7/21/06 7/21/06 7/24/06 7/24/06 7/24/06 7/24/06 7/24/06 7/24/06 7/24/06 7/24/06 7/25/06 7/25/06 7/25/06 7/25/06 7/25/06 7/25/06 7/25/06 7/25/06 8/9/06 8/9/06 8/9/06 Hour of Day (1 through 24) 15 16 14 15 16 17 14 15 16 17 18 14 15 16 17 13 14 15 16 17 14 15 16 17 12 13 14 15 16 17 18 12 13 14 15 16 17 18 19 12 13 14 15 16 17 18 19 15 16 17 Load (MW) 450.19 450.93 456.01 460.81 461.2 455.21 456.6 466.41 462.19 459.42 452.72 452.16 458.31 461.72 458.3 451.41 449.89 450.59 454.02 454.36 455.89 459.84 459.87 456.5 450.58 456.86 465.44 472.9 471.06 466.33 456.3 465.1 477.98 481.26 480.72 479.59 477.22 468.46 449.86 462.41 472.78 481.39 486.46 484.74 480 472.14 450.71 454.61 457.86 455.05 Peak hourly load in 2006 26 Calculations for DR Potential Savings
For 7/26/06: Peak Max Peak Average $400.00 $113.00 Spot Market Price of Energy ($/MWh) $400 $350 $150 $75 Hour of the day 16 17 15 18 Amount of Power Reduced from DR event (MW) 7 7 7 7 Dollars Saved ($) $2,800 $2,450 $1,050 $525 Amount of Power Reduced from DR event (MW) 14 14 14 14 Dollars Saved ($) $5,600 $4,900 $2,100 $1,050 Amount of Power Reduced from DR event (MW) 21 21 21 21 Dollars Saved ($) $8,400 $7,350 $3,150 $1,575 Total
For 7/25/06: Peak Max Peak Average $400.00 $77.12 $6,825 $13,650 $20,475 Spot Market Price of Energy ($/MWh) $250 $90 $60 $60 Hour of the day 14 15 13 16 Amount of Power Reduced from DR event (MW) 7 7 7 7 Dollars Saved ($) $1,750 $630 $420 $420 Amount of Power Reduced from DR event (MW) 14 14 14 14 Dollars Saved ($) $3,500 $1,260 $840 $840 Amount of Power Reduced from DR event (MW) 21 21 21 21 Dollars Saved ($) $5,250 $1,890 $1,260 $1,260 Total
For 7/24/06: Peak Max Peak Average $399.65 $81.57 $3,220 $6,440 $9,660 Spot Market Price of Energy ($/MWh) $14