Saturday, May 28, 2016

Wind driving extreme pricing in Ontario's electricity market

"There were wind shortfalls in all but one of the 28 High HOEPs during the Current Reporting Period"

Industrial wind turbines are driving extreme prices in the Ontario market according to the latest report from The Ontario Energy Board (OEB) Market Surveillance Panel (MSP).

And the past week's market performance.

  • On Sunday May 22, during hour 8, the system operator's (IESO) 5 minute Market Control Price (MCP) bottomed out at it's minimum of -$2,000 per megawatt-hour (MWh)
  • On Tuesday May 24, during hour 20, the 5 minute MCP hit the maximum $2000/MWh

Both spikes appeared to be due to renewable and demand forecasts being poor predictors of actual market conditions. The OEB's new report is therefore very relevant today, despite being for the period from November 1, 2014 to April 30, 2015.
In the Current Reporting Period there were 28 hours in which the HOEP exceeded $200/MWh (High HOEPs). This Period also had the highest HOEP since market opening, reaching $1402/MWh in hour ending 8 on February 20, 2015. The High HOEPs were primarily caused by under-forecasts of demand and short-notice losses of supply (curtailing of imports and under-generation of wind facilities relative to their forecast production).
...
Relevance: 
Identifying the factors that lead to deviations between the PD-1 MCP and the HOEP provides insight into the root causes of price risks that participants, particularly importers and exporters, face as they enter offers and bids into the market.
This is the complicated data explanation of the simple mechanism through which forecasts can cause price spikes (for the reported period almost the exclusive cause of high priced hours). "Demand" to the MSP is demand from  IESO grid-connected generators. If the forecast for wind, and solar, is light the demand for grid-connected supply would be greater as embedded wind, and/or solar, would be less productive than expected too.
2.1.2  Wind Shortfalls, Demand Under-forecasting and High HOEPs 
A ‘wind shortfall’ occurs when real-time wind output is less than the hour-ahead (PD-1) forecast. Conversely, under-forecasting of demand occurs when real-time demand is greater than the PD- 1 forecast. Both of these conditions result in a greater need for supply in real-time than was contemplated in PD-1. There were wind shortfalls in all but one of the 28 High HOEPs during the Current Reporting Period, and an under-forecasting of demand in 22 of the 28 High HOEPs. Figure 2-3 maps the HOEP against wind and demand forecasts, and shows a data point for each hour during the Current Reporting Period. The coordinates represent the degree of demand forecast error (on the y-axis) and wind forecast error (on the x-axis). If a data point lies above the x-axis, then real-time demand was higher than forecast (the forecast underestimated real-time demand). If a data point lies to the right of the y-axis, then real-time wind production was less than was expected in the PD-1 timeframe (real-time wind production fell short of expectations). 
Figure 2-3: HOEP Map Against Ontario Demand Under-Forecasting and Wind Shortfall November 2014 – April 2015 (MW)



Most High HOEPs occur in the upper right quadrant because wind shortfalls and demand under-forecasts result in both tighter than expected supply conditions and higher than expected demand, creating upward pressure on the real-time price. The forecasting errors in the top left and bottom right quadrants have off-setting price impacts (with one error putting upward pressure on price and the other putting downward pressure on price). In the bottom left quadrant, both forecasting errors put downward pressure on price; there is more wind generation relative to forecast and less demand than forecast. Wind shortfalls and demand under-forecasting are not the only causes of High HOEPs. Other events which force the market to turn to flexible (and typically more expensive) resources to supply demand in real-time also put upward pressure on prices. However, as shown in Figure 2- 3, 75% of the High HOEPs in the Current Reporting Period occurred when there was both an under-forecast of demand and a wind shortfall.
There are some important lessons for other jurisdictions here - they'll require this background:

  • the IESO accepted responsibility for forecasting wind
  • Ontario has no day ahead market (which would enforce a market's wind generators to forecast their output in advance, and cover shortfalls on a spot market)
  • wind generators therefore have no reason to try and meet forecast supply - which a cynical person would phrase as wind (and solar) generators have no incentive to provide a product of any value to end users.
  • The OEB MSP graphic above indicates the IESO usually overestimates demand
  • the IESO altered market rules for 2016, making wind dispatchable before most nuclear
  • Ontario phased out coal at the end of 2013
There's a lot here. Starting with coal...
Coal has a dispatchable range of 20 to 100 percent full power compared to around 70 to 100 percent for combined cycle gas...For gas to provide the same dispatchable power as coal, with both operating at their respective minimum loading points, several times as much gas generation would have to be on line meaning very much more GHG emissions. Under these circumstances coal, rather than gas, would make a better partner for wind.
A long-term view of the incidence of high price hours shows the spike occurring as coal was eliminated.


It does coincide with colder winters too, but the OEB data indicates that was not the cause during winter 2015, which is the second highest incidence of high priced hours in the past 10 years. The most recent winter had far fewer high priced hours, but it also had huge curtailment of wind. Recalling the IESO averaged a significant "Demand Overcast" in the MSP's report, it's credible to conclude the over demand overestimate is deliberate, acting to protect against unreliable wind forecasts - with the operation of the grid therefore requiring the increased curtailments seen this past winter.

While the high priced hours dropped this past winter, negatively priced hour continues what has been a relatively consistent upward climb in recent years - a climb coinciding with the growth of wind and solar generators. The Hourly Ontario Energy Price (HOEP) was negative 1 in 5 hours from November 1, 2015 to April 30, 2016.



The Washington Post has a report this week based on newly published research. One issue reported on:
On a national scale, approximately 17.2% and 10.7% of the total wind power in China was curtailed in 2012 and 2013, respectively, resulting in annual financial losses of more than one billion US dollars
I noted on Twitter my estimates show Ontario now curtailing a higher percentage of wind than the numbers shown for China - and perhaps more than any other jurisdiction.






spreadsheet including data for "count of hours" charts
spreadsheet with Ontario Wind curtailment graphic.

No comments:

Post a Comment