Ontario electricity demands - an appendix.

Last week Ontario's Independent Electricity System Operator (IESO) released its first Annual Planning Outlook (APO), which is pertinent to the next article I wanted to write. This isn't that article - it's an explanation of graphics I created for the article I want to write.

While the APO the first of the new thing, it's just the latest in a line of forecasts intended to guide electricity planning in the province. This graphic shows 5 previous forecasts, but I felt I should explain their origins, and note they may not be strictly comparable. "Demand" is, in Ontario, a word that is, by itself, not very definitive.



From the APO:

This demand forecast is presented on a weather-normalized basis and at the net level, as opposed to actual weather-affected and grid or gross level. Gross-level demand is the total demand for electricity services in Ontario prior to the impact of conservation programs but including the effects of naturally occurring conservation. Net-level demand is gross-level demand minus the impact of conservation. Grid-level demand is net-level demand minus the demand met by embedded resources. It is equal to the energy supplied by the bulk power system to wholesale customers and local distribution companies.

So there's 3 types of demand, none of which are what the IESO has traditionally called, "Ontario Demand." From the last monthly Market Report the IESO produced:

Ontario Demand represents the total energy that was supplied from the IESO-Administered Market for the purpose of supplying load within Ontario.

This makes more sense with the breakdown provided in a tab of a spreadsheet that accompanies the IESO's - checks notes - Reliability Outlook. Table 3.3.3 in that spreadsheet shows "Energy Demand" composed of "LDC [local distribution companies] Consumption, Wholesale Consumption, Generator Consumption and Losses." Note the generator consumption and (line) losses do represent "energy that was supplied from the IESO-Administered Market for the purpose of supplying Load" without being supplied to "wholesale customers and [LDCs]". The portion of generation that makes it to a wholesale customer or LDC might be called an Allocated Quantity of Energy Withdrawn (AQEW) - which is only a number the IESO publishes for 10 hours each 12-month period (due to the industrial conservation initiative), although they have provided hourly AQEW data on request.

AQEW may be the energy supplied from the IESO-controlled grid (ICG), which they seem to refer to as the "bulk power system" in the new APO, but there is also supply generated within local distribution company networks. That generation is known as embedded, or distributed (Dx), and by other names. The reliability spreadsheet I noted previously has another tab, table 3.3.2, which includes a column with the monthly "Reported Embedded Generation (GWh)".

My guess is that what the IESO is charting as "Energy Demand" in the APO is AQEW plus embedded (Dx) generation. That figure coincides almost exactly to what is reported as "Consumption" in the IESO's global adjustment reporting - and I feel "consumption" would have been a better term than "Total Ontario Demand" to use in this graphic from the APO:

Within a week of the APO's release the IESO published year-end data for 2019, in which a section titled "Demand" begins: "Ontario’s energy demand in 2019 was 135.1 TWh." The casual observor might think the APO's forecast therefore calls for a big 9 TWh surge in "demand" for 2020, as the IESO doesn't report on embedded supply at all in their year-end, or most other, summaries - and has long used the term "Ontario Demand" to mean demand for supply from generators on their ICG.

So hopefully that explains what 'demand" means in the latest forecast. Now about the 5 earlier forecasts...

I won't attempt the same analysis of what is meant by "demand" in prior forecasts, but I will note the origin of the forecasts, and offer some comments on the choice, as well as the source, of figures in my graphic.

Forecasts are done by predicting consumption in a variety of sectors. What is removed from those forecast (ie. impacts of energy efficiency programming, and possibly embedded generation) isn't always clear - nor is it clear production forecasts account for line losses. In long-term plans which included multiple scenarios, I picked a single one - the one I felt was most intended to represent a business-as-usual (BAU) scenario.

The earliest scenario in my graphic is from the Integrated Power System Plan (IPSP) submitted by the Ontario Power Authority (OPA) to the Ontario Energy Board. I believe that in the past I've referred to this as the last forecast/plan done by an independent professional organization. The figures I've captured (transposed from a graphic) are for the "Net of Reference Conservation" scenario from the 2007 IPSP.

The 2011 IPSP was never released publicly. I've taken the figures from documents acquired by Shawn-Patrick Stencil, which were the motivation for a 2017 article on the history of the supply mix. The planning process pre-IPSP included a Supply Mix Directive and Long-Term Energy Plan (LTEP), which were to serve as the basis for an IPSP constructed by the OPA. I knew it existed, but after the government lost its majority in an election the IPSP was buried and the OPA folded into the IESO. Shown is the medium of scenarios shown in the 2011 IPSP.

By 2013 the LTEP was the end product. Shown is the "Ontario Net Demand".

In 2016 the IESO produced an Ontario Planning Outlook which would serve as the basis for the next LTEP. I've used "Outlook B".

In 2017 the Ministry of Energy produced the LTEP, and the "Net Electricity Demand Forecast" from the accompanying data files matches the OPO outlook until the mid 2020's.

Which brings us up to the Annual Planning Outlook introduced last week.

My "IESO Actual (With Dx)" is simply the IESO's reported "Ontario Demand" plus the annual total "Reported Embedded Generation (GWh)".

That's the one graph I wanted to show. Probably more important than the annual supply required is the peak supply required. Estimating actual annual peak consumption is difficult as the data is not available publicly, nor is it likely the IESO has it. The Industrial Conservation Initiative requires the IESO to know embedded generation for only the 5 hours a year constituting the 5 highest daily AQEW peaks - AQEW being the quantity withdrawn from the IESO-controlled grid for consumers. Embedded generation lowers AQEW, so the 5 peaks tracked by the IESO have moved to, in general, later on summer days. These AQEW peaks are the only hours, as far as I know, that the IESO reaches out to local distribution companies for information on hourly embedded generation.

To be clear, it's known the AQEW demand peaks aren't actual consumption peaks, precisely because we know the peaks moved to later in the day as more embedded solar generation was added. The best I can do is use my own hourly estimates of embedded supply to estimate an actual history of annual peak consumption values.

The alteration of the peak demand hours by solar has great implications for determining capacity values in modeling supply mixes to meet capacity requirements. Most notably, existing solar has a much greater capacity value than future solar is likely to have.


spreadsheet