IESO is incorporating a wider range of risks in its Reliability Outlook (RO), an 18-month forecast used to manage generator and transmission outages, a move it says will make it slightly easier to schedule summer outages.

The new probabilistic approach departs from the outlook’s deterministic focus on “normal” and “extreme” weather scenarios. The new methodology also changes how the ISO projects hydro and wind output and how it accounts for imports and planned loads.

The changes will bring the RO in line with best-practice forecasting methodologies and is a better fit for the evolving supply mix, with increased solar, storage and other distributed energy resources, IESO officials said during a recent briefing.

The ISO uses the RO to comply with NERC and Northeast Power Coordinating Council reliability standards. Bonnie Chan, manager of planning assessments, said the ISO completed its shift with the publication of its RO methodology in late March.

The changes “will make sure that resource planning decisions are based on the most accurate, data-driven insights possible, helping us better align planning efforts with what’s happening in the market,” said Fatema Khatun, a stakeholder engagement adviser.

The new approach is a response to stakeholder feedback and a recommendation from the Market Surveillance Panel that the ISO improve the RO’s alignment with two other forecasts: the Adequacy Report, which focuses on Ontario’s electricity requirements for the next 34 days, and its Annual Planning Outlook, a 20-plus-year look used for evaluating long-term investment decisions and resource acquisitions.

IESO’s metric for determining resource adequacy is called the Reserve Above Requirement (RAR). The ISO begins its evaluation by derating its installed resources and imports based on factors such as effective forced outage rates. The resulting “available” resources are measured against the projected demand forecast and required reserves. If there is excess capacity, the ISO has a positive RAR.

Weather Scenarios

The RO previously used a deterministic approach to calculating resource adequacy based on an “extreme” weather scenario and the assumption it would be able to rely on up to 2,000 MW of imports year-round.

The outlook used 31 years of weather history — dry bulb temperature, dew point, wind speed and cloud cover — from six weather stations between Windsor, Ottawa, Toronto and Thunder Bay to generate normal (50/50 probability) and extreme (maximum) demand forecasts. The approach was limited in its simplified output projections for embedded wind and solar.

IESO will continue to use 31 years of weather history but add a new data source for predicting production from its 2,000 MW of embedded solar capacity: Global Horizontal Irradiance (GHI), which measures the total solar radiation received on horizontal surfaces.

“Cloud cover is not necessarily the most accurate input for our solar models,” said Andrew Trachsell, senior demand forecaster. “Unlike the previous methodology, where you’re looking at cloud cover at [a weather station] that could be 50 km away [from a solar farm], this [GHI] data is at a very granular level — at 2 by 2 km — so it is a much more detailed or accurate input.”

Demand Scenarios

IESO also is changing how it accounts for new demand in the province, which increasingly has been targeted by large loads seeking low-carbon power.

It will use a “planned” scenario to account for loads that are less certain to reach commercial operation during the forecast period but large enough to warrant consideration because of their potential impact on grid operations. Its “firm” demand scenario will be limited to loads with a high probability of going into operation during the forecast horizon.

The former approach, which used a monthly normalization, resulted in higher demand peaks, particularly in the summer.

“The previous methodology attempted to put a monthly peak — which is a peak at a higher level with certainty, and less uncertainty above it — into the forecast, and then apportion that monthly peak across the week,” Trachsell said. “Now we’ve moved to just strictly the weeks, and that means that you’re going to get a lower peak with certainty, but much more … variability above it.”

Wind and Hydro

IESO also has begun using probabilistic distributions to model wind and hydro production instead of using a median value for each week. That allows it to capture a broader range of risks, including the tail ends of low and high hydro and wind conditions.

In the past, modelers used a monthly median of production plus scheduled operating reserves; 2012 was the proxy for the driest year, with 800 MW of hydro subtracted for the summer months.

“Before we looked at it more deterministically, we had the normal and extreme, and that was kind of like the bookends and we used the extreme weather scenario to do outage management and decision making,” said Emanuel Moldovan, senior planner. “There’s no more bookends. It’s just one scenario.”

Trachsell said the ISO’s previous method also “forced everything into a normal distribution, with an assessment of what the uncertainty around that normal was that had to be calculated. And once again, this was a fixed approach that was not necessarily as robust as it could be.”

The new methodology models wind on a Weibull distribution in all months, which the ISO says is more accurate in accounting for low wind periods. A Weibull distribution also is more flexible than the normal distribution and can be used to model a wider range of data shapes, including skewed and non-normal data.

LOLE Allocation

As the region’s Planning Coordinator, IESO is charged with ensuring the loss-of-load expectation (LOLE) averages no more than one day every 10 years, the limit set by NPCC.

Going forward — “pending ongoing monitoring of the situation,” Moldovan said — the ISO will allocate 30% of the LOLE risk to the summer (May to October) and 70% in the winter (November to April).

The new approach will reduce the RAR in the winter and increase it in the summer, making it slightly easier to approve outages in the summer. “Although the winter RAR is decreasing, current system conditions do not indicate a need to reschedule outages due to resource adequacy concerns in the winter months,” IESO said.

Imports

Assumed imports will be reduced from 2,000 MW to 1,000 MW in winter in both the RO and the Adequacy Report to account for IESO’s capacity agreement with Hydro-Quebec, which needs firm capacity from Ontario during those months.

Overall, while the new methodology considers a wider range of risks, “the results are similar to the previous approach and should result in minimal changes for outage management,” IESO said.