Rethinking Energy Conservation in Ontario – Results:Ontario Power Authority Demand Response Programs
| In November, 2010, the ECO released volume 2 of its Annual Report on the progress of activities in Ontario to reduce or make more efficient use of electricity, natural gas, propane, oil and transportation fuels. Click here for more information on this report, including videos and communications materials. | |||||
|
|||||
Contents |
Demand Management
Unlike most goods, electricity cannot be stored.71 The electricity system operator must continually ensure that electricity supply is available to meet demand at every moment. For this reason, the time that electricity conservation takes place is very important. Electricity conservation initiatives that place a priority on reducing energy use at specific times (times of high or peak demand) are known as demand management. Demand management may or may not reduce the total level of energy consumption – it depends whether the energy use that is curtailed during peak hours is undertaken at another time, or foregone altogether.
The opportunities for demand management can be seen in Figures 2 and 3.
Figure 2: Ontario Demand – Top 100 Hours in 2009
Source: Derived from Independent Electricity System Operator market data
Figure 2 shows the 100 highest hours of demand (about 1 per cent of total hours) in Ontario in all of 2009. Most of these hours would have been during hot summer weekday afternoons. The difference between the single highest hour of demand (24,380 MW) and the 40th highest hour of demand (22,110 MW) is 2, 270 MW – approximately the power generated by three large natural gas plants. In other words, to meet this extreme power demand, Ontario might need three large gas plants that would run for a mere forty hours or less each year.
Opportunities for demand reduction are not limited to a few extreme events during the year. There are even larger demand management opportunities on a daily basis.
Source: Derived from Independent Electricity System Operator market data
Figure 3 shows the daily demand cycle or load profile on a typical summer day (August 16, 2009). Demand is low in the early hours of the day, begins to ramp up around 6:00 a.m., peaks in the late afternoon, and falls off significantly after 9:00 p.m. The difference between the daily peak and the trough is enormous — some 8,000 MW.
If demand management can be used to reduce both the yearly and daily electricity peaks, the economic advantages are obvious. In the near term, the use of peaking resources (i.e., natural gas-fired generators) with high fuel costs is reduced. In the longer term, fewer power plants and transmission and distribution lines need to be built. This will reduce the high cost of meeting demand at peak times. Thrown in to make the case for demand management are environmental and social benefits. Off-peak generation can be supplied primarily by carbon-free nuclear and renewable energy resources, while many peaking plants are fossil-fuel based (natural gas, as well as coal until 2014). The reduction in air pollutants during peak periods is also a benefit, because summer peaks are usually hot days where smog may already be a health concern. Lastly, not having to situate new power plants close to populated areas is also a considerable social benefit, as the intense debate over recently proposed gas peaking plants has shown.
Demand Management Potential in Ontario
The draft Integrated Power System Plan proposed that approximately 20 per cent of the planned reduction in peak demand by 2025 would come from demand management programs72 (as opposed to conservation actions such as energy efficiency and fuel switching that would reduce electricity consumption at all times)..
Ontario has pursued demand management through several avenues. A key policy choice has been mandating time-of-use (TOU) pricing (for residential electricity consumers and small businesses) and real-time market pricing (for larger customers). TOU pricing induces demand management by changing customer behaviour. Customers react to TOU pricing by shifting some electricity consumption away from periods of high prices (and high system demand).
Another form of demand management is targeted demand response (DR) programs.
OPA Demand Response (DR) Programs
Directives from the Minister of Energy to the Ontario Power Authority (OPA) in 2005 and 2006 have authorized the OPA to procure up to 500 MW of demand response.73 The intent was to address reliability concerns in the Greater Toronto Area by acquiring additional supply and demand resources.
The OPA’s primary means of achieving this objective has been a suite of DR programs (DR1, DR2, DR3), targeted primarily at large industrial and commercial consumers of energy. DR1 launched in June 2006;74 DR3 launched in August 2008;75 and, DR2 started in July 2009.76 These programs replaced (and greatly expanded upon) pilot demand response programs that had been tested by the Independent Electricity System Operator (IESO).
The intent of the three DR programs is to work together to displace the need for new peaking energy supply resources,77 which in Ontario usually means single-cycle gas turbines,78 also known as gas peaker plants.
| Objectives of DR Programs |
|---|
| The specific intent of each DR program can be summarized as follows: DR1 is voluntary load shedding. |
Although most participants in the DR programs will provide demand response by simply reducing their electricity consumption, DR1 and DR3 also allow participants to provide demand response through on-site generation. On-site generation has the same effect as load reduction in reducing the imbalance between demand and supply on the electricity grid. On-site generation is typically provided by backup generators (e.g., diesel) that would not otherwise be costcompetitive with electricity from the grid. This raises the concern that air emissions from these generators may compromise air quality, defeating the government’s intention to replace coal with conservation and cleaner forms of generation. In response to this concern, the Ministry of the Environment has issued a policy that requires generators used for nonemergency purposes (such as demand response programs) to meet air quality emission standards comparable to modern natural gas combustion turbines.79
How the Programs Work
Large industrial and commercial energy consumers are the target participants. DR1 is open to participants that can reduce electricity consumption by more than 0.5 MW, while DR2 and DR3 are open only to participants that can provide 5 MW or more of demand response. All programs allow participation by aggregators,80 who can combine demand response from multiple smaller commercial and industrial participants in order to achieve the minimum desired level of demand response.
The programs are activated during periods of high value to the electricity system. DR1 is activated when the wholesale market price (known as the Hourly Ontario Energy Price) is high. During hours of projected high prices,81 participants may choose to curtail electricity consumption in return for compensation from the OPA.82
DR3 is activated when the difference between estimated market supply of electricity and demand (known as the supply cushion) is low. The OPA must give participants at least 2.5 hours advance notice of activation, making the IESO’s forecasting role in estimating the supply cushion critical.
DR2 does not have variable activation hours based on conditions in the electricity market, but instead makes use of the known difference in demand between night and day. Participants must shift load on a regular basis from peak hours (between 7 a.m. and 7 p.m.) to off-peak hours (between 7 p.m. and 7 a.m.).
Participants are compensated by the OPA for providing demand response during these periods. Payment from the OPA is in addition to any reduction in direct electricity costs that the participant realizes due to their changed pattern of electricity consumption. Payment for the contractual DR2 and DR3 programs is much higher than for the voluntary DR1 program.
Participation
The DR programs have been quite successful in attracting participants and achieving the goal of acquiring up to 500 MW of demand response as set out in the minister’s directive.
Table 4: Amount of Demand Response Capacity at Year-End (MW)
| Program | 2007 | 2008 | 2009 |
|---|---|---|---|
| DR1 | 317 | 440 | 175 |
| DR2 | - | - | 119 |
| DR3 | - | 84 | 170 |
| Total | 317 | 525 | 464 |
Source: Ontario Power Authority, letter to ECO, September 9, 2010
Until August 2008, DR1 was the only active program. Since then, many DR1 participants have migrated to the contractual DR2 and DR3 programs. Active participation in the DR programs has been dominated by several large firms. However, DR3 also obtains a significant amount of its load reduction capacity (58 per cent or 98 MW) from aggregators. The participation of aggregators, as well as multiple direct participants, is desirable because it improves the program’s reliability by reducing dependence on the actions of any individual firm.
Results
Table 5 shows the results for the OPA DR programs in 2007 - 2009.
Table 5: OPA Demand Response Program Results 2007 - 2009
| Program | Year | Number of Hours of Operation | Average Load Reduction (Settlement) (MW) | Average Load Reduction (Verified) (MW) |
|---|---|---|---|---|
| DR1 | 2007 | 1365 | 128.4 | Not Available |
| 2008 | 1201 | 139.3 | Not Available | |
| 2009 | 132 | 151.4 | 58.6 | |
| DR2 | 2009 | Weekdays, 7 a.m.-7 p.m. | 53 | 50.7 |
| DR3 | 2008 (100 & 200 groups*) | 36 | 35.5 | 30.6 |
| 2008 (200 group only) | 24 | 52.2 | 46.8 | |
| 2009 (100 & 200 groups) | 16 | 85.7 | 81.4 | |
| 2009 (200 groups only) | 8 | 12.8 | 11.1 |
| Note: *The DR3 program allows firms to commit to providing up to 100 hours of demand response (“100 group”) or 200 hours (“200 group”) per year. Activations during extreme peak events would typically include both groups, while only the 200 group might be activated for more moderate peak events |
| Source: Ontario Power Authority, letter to ECO, September 9, 2010 |
The difference between the average load reduction estimated for settlement purposes (i.e., the reduction amount for which
participants receive payment) and the verified load reduction (calculated through formal program evaluation) is primarily
due to difficulties in correctly determining each participant’s baseline electricity consumption. Baseline consumption is
an estimate of how much electricity the participating facility would have consumed, were it not participating in the demand
response program.
Amount of Demand Response (MW) = Baseline Electricity Consumption (MW) – Actual Electricity Consumption (MW)
Estimating the value for baseline electricity consumption is a difficult task, as electricity consumers may alter their patterns of consumption for reasons unrelated to the demand response program, such as plant shutdowns or process changes. In particular, because most of the participants in DR programs are paying the wholesale market price for electricity, their consumption patterns may rise and fall in the opposite direction to the market price.
DR1 proposed a standard method for calculating baselines, but also allowed participants to submit a customized baseline, which the four largest participants did. An evaluation report showed that, while the standard baseline methodology was quite accurate in estimating load, the customized baselines greatly over-estimated participant load (and thus load reduction). Due to the size of the four largest companies, this had a significant impact on program results. In 2008, the average amount of DR paid for by the OPA for a given hour of activation was 102 MW; however, the true amount of load reduction attributable solely to the DR program was just 35 MW – an enormous difference.
The OPA has recognized these problems. As a result, DR2 and DR3 now have more stringent procedures for baseline calculation, and the OPA does not plan to continue to offer customized baselines for DR1 in the future.
Reliability
If demand response is to be treated as a serious alternative to supply resources by system planners and system operators, it needs to deliver reliable results. Table 6 shows the load reduction capacity enrolled in the three OPA DR programs at the end of 2009, and the actual amount of load reduction that the OPA predicts these programs would be able to deliver on an ex ante (i.e., expected or going-forward) basis.
Table 6: OPA Demand Response Program Capacity and Load Reduction Potential
| Program | Load Reduction Capacity at end of 2009 (MW) | Ex ante Load Reduction Estimate (MW) |
|---|---|---|
| DR1 | 175 | 0.2 |
| DR2 | 119 | 94 |
| DR3 | 170 | 129 |
Source: Ontario Power Authority, letter to ECO, September 9, 2010
Table 6 shows that DR2 and DR3 deliver reliable demand response from a system planning perspective, as measured by the fact that the ex ante load reduction estimates are a substantial portion of the load reduction capacity. Several program features enable these programs to deliver reliable demand response. They require mandatory load reduction; have low baseline errors; and, impose financial penalties for non-compliance. DR1, in contrast, suffers from its voluntary nature and high baseline errors, and delivers an extremely low amount of reliable demand response. The total amount of demand response capacity currently in DR1 is not an accurate measure of the program’s importance since many of the firms remaining in DR1 do not actively participate. DR1 now functions primarily as a risk-free way for firms to experiment with participating in demand response programs, prior to migrating to DR2 or DR3.
On a short-term basis – the relevant time frame for the system operator – the reliability of the contractual DR programs is even greater. In 2009, DR3 delivered 82 per cent of the day-ahead contracted load reduction.
In comparison with natural gas plants, the DR2 and DR3 programs also offer a benefit that could be called “distributed reliability”. The availability of a gas plant is an all-or-nothing situation and an unanticipated outage reducing power by some 500 MW could place serious strain on the grid. In contrast, the load reduction provided by demand response is made up of the contributions of multiple firms, reducing variability.
Activation Timing
The above assessment measures how reliable participants are in reducing electricity consumption in response to program requirements. It does not, however, evaluate whether the DR programs are activated when they are needed most, that is, coinciding with times when demand is closest to exceeding available supply.
There is strong evidence that the activation mechanism for DR3 is not optimally hitting the hours of peak system need. Between August 2008 and October 2009, DR3 was activated 21 times by the OPA. If targeted perfectly, these 21 activations would have corresponded with a measure of system need, such as the 21 days with the highest hourly peak demand. Upon examination, only 5 of the DR3 activations occurred on these 21 days.
The OPA has attempted to address this issue by modifying the activation mechanism for DR3 to require both high prices and a tight supply-demand balance. However, additional improvements may be needed.
Cost-effectiveness
There is something that feels economically unsound or even profligate about paying firms not to consume electricity. And in the near term, it is usually true that demand response is more expensive than the marginal cost of obtaining more generation from existing plants.
The true value of DR is to offset future costs – to prevent new generators from being built, with their large capital costs that must then be recovered from ratepayers. How does the OPA’s DR3 program stack up on a cost basis with its likely alternative – new natural gas single cycle peaker plants?
The OPA has estimated that the total cost of a single cycle gas plant (including associated transmission and distribution costs) that would only run during the top 88 hours of system demand in a year would cost approximately $1,187-$1,642 per MWh ($1.19/kWH-$1.64/kWH).
By comparison, the cost of acquiring the same amount of demand response (88 hours) from a DR3 participant in the Toronto area is roughly comparable at approximately $1,000-$1,700 per MWh ($1.00/kWh-$1.70/kWh) depending on the contract length, at least at first approximation.
Other factors not captured in the market price comparison – avoided emissions of greenhouse gases and other pollutants, and social benefits from avoiding new generation – weigh in favour of demand response over natural gas peaking generators.
It should be noted, however, that the costs for either demand response or peaking generation are extremely high – some 20 times the average price of generation, and many times greater than on-peak rates.
Issues and ECO Comment
Within a few years, the OPA has brought a large amount of demand response under contract. With the shift in enrolment from the voluntary DR1 program to the contractual DR2 and DR3 programs, the reliability of demand response as a system resource has increased dramatically, and the problem of overpaying for demand response due to baseline error has declined. As noted in section 2.0, the OPA has also taken steps to accurately measure the true system impact of demand response, by developing a rigorous evaluation protocol. The ECO commends the OPA for these accomplishments.
At an operational level, the greatest remaining need for improvement is activating DR at the right time. The economic case for DR3 relies on its potential to replace gas peaker plants at times of extreme system need, so it is critical that the OPA is able to target activation to exactly these times. The ECO supports the recommendation of the OEB Market Surveillance Panel that the OPA work with the IESO to improve the advance forecast of supply and demand, which will enable more precise activation of DR3.
Questions also remain as to the appropriate scale of contractual demand response programs. The OPA is close to reaching its authorized capacity for demand response programs (500 MW), and would need to seek additional authority from the
Minister of Energy in order to expand participation in the DR programs. The OPA expects to seek this authority from the Ministry of Energy in the near future. Therefore, it is an appropriate time to consider the role that contractual demand response should play in Ontario’s electricity system.
Given the similar cost of DR to new natural gas peaker plants, and the additional advantages of reduced emissions and reduced social tension associated with building new generation, the ECO supports the principle that additional demand response should be chosen in preference to building new gas peaker plants, wherever possible. This disciplines the system and prevents unnecessary overbuilding of supply resources.
However, given the high cost of contractual demand response programs per unit of electricity, they should be treated as the “option of second last resort”. DR programs should only be expanded if the demonstrated need exists (based on nearto mid-term forecast load growth).
When the Integrated Power System Plan was drafted, plans were on the drawing board for at least three new single cycle peaker plants. However, since then, structural changes to Ontario’s economy have reduced demand, perhaps permanently, while a large amount of new gas-fired generation has come on line and the government is rapidly expanding its renewable capacity through the feed-in tariff program. These events have greatly improved the near-term reliability of Ontario’s electricity system, and may allow for some breathing room prior to procuring additional demand response (or building new peaking plants).
Since the government is enhancing the role of time-of-use pricing for both wholesale and electricity consumers, this breathing room is helpful. Time-of-use pricing may reduce peak demand at much lower cost than contractual demand response programs.
Given the above points, the ECO believes that the role of contractual demand response programs should be reviewed in the revision of the Integrated Power System Plan (now renamed the Long-Term Energy Plan). A role will certainly remain (particularly for peak shedding programs such as DR3 to provide insurance against extreme weather events or generator outages), but it may be smaller than was originally anticipated. On the other hand, a new role for some form of demand response will emerge – namely balancing fluctuations in supply from renewable energy sources. It is unclear whether the existing contractual demand response programs can respond quickly enough to variations in supply to be useful for this purpose. Updates to the Long Term Energy Plan will allow the OPA to re-assess the need for new contractual demand response on a regular basis, and expand demand response capacity as needed to avoid building additional gas peaker plants, while minimizing cost to ratepayers.
The ECO recommends that the Ontario Power Authority only expand contractual demand response programs when this will eliminate a demonstrated near-term need for new peaking generation.
