Time-of-use times and system-wide demand

With continued attention upon time-of-use pricing in the press (see, for instance, recent articles entitled ‘Business owners shocked by electricity bills’, from the 19 August 2011 edition of The Toronto Star and ‘PCs slam smart meters, Liberal MPP defends them’, from the 19 August 2011 edition of The Guelph Mercury), I was prompted to reflect further upon the different use of electricity – system-wide in Ontario – at differen times.  Before data about that, some additional context is appropriate.

The electricity system is one that moves a commodity – namely, electrons – that cannot easily (cheaply) be stored.  In other words, they have to be ‘delivered’ when generated; alternatively, they have to be ‘generated’ when demanded.  Thus, the logic behind time-of-use rates is that we try to lessen demand during those ‘peak demand periods’, so that we – over time – smooth out the system-wide load.  Why do we want to do this?  Well, remember that we have to make the electricity system work for that one time during the year in which system-wide demand is at its peak; at all other times, part of the capacity of the electricity system (the ‘grid’) sits idle.  All else being equal, then, we would like a system with flat load – using the same amount of electricity at every hour of the day, so that we are using the capital infrastructure as ‘fully’ as possible.

How are we doing in Ontario?  I have pulled out Ontario system-wide demand for the months of January and July 2011.    I have, for each month, constructed what is called a ‘load duration curve’ – for the 744 hours in each month, I have lined them up, highest to lowest (in terms of demand in MW) from left to right.  For each of these two months (which are, of course, the ‘heart’ of winter and the ‘heart’ of summer), I have prepared two graphs – one with the full 744 hours on it, and the other focusing upon the top 1% (or seven hours).  They now follow.

Load Duration Curve – all 744 hours in January 2011 (with 1% or 7.44 hour divisions along the x-axis)

Load Duration Curve – top seven hours in January 2011

Load Duration Curve – all 744 hours in July 2011 (with 1% or 7.44 hour divisions along the x-axis)

Load Duration Curve – top seven hours in July 2011

Interestingly, most – but not all – of those ‘top hours’ occurred during ‘peak periods’.  In the January case, four of the seven were during ‘on-peak periods’ – in particular, Monday evenings (17 January from 6-7pm, 24 January from 5-7pm and 31 January from 6-7pm).  One of them occurred during the mid-peak period (Monday, 24 January from 4-5pm) and two occurred during the off-peak period (Monday, 24 January from 7-8pm … note that this was formerly an ‘on-peak period’ and Sunday, 23 January from 6-7pm). In the July case, five of the seven were during ‘on-peak periods’ – from noon-5pm on Thursday, 21 July, with the other two (Thursday, 21 July from 5pm-7pm) during the ’mid-peak period’.

This led me to think about where the hours ‘fall’ vis-à-vis their assigned ‘time-period’ (on-peak, mid-peak or off-peak).  In response, I put together the tables below.

January demand hours	on-peak	mid-peak	off-peak
top quarter	72	43	71
2nd quarter	47	70	69
3rd quarter	1	7	178
bottom quarter	0	0	186
July demand hours	on-peak	mid-peak	off-peak
top quarter	72	45	69
2nd quarter	48	36	102
3rd quarter	0	39	147
bottom quarter	0	0	186

In these, I ranked the demand hours from ‘top’ (and ‘top quarter’) through down to bottom (and ‘bottom quarter’). If our province-wide demand periods followed ‘time-of-use times’ perfectly, we would expect the 120 hours for on-peak times (16% of the 744 hours in the month) to all fall in the ‘top quarter’; the 120 hours for the mid-peak times to fall largely in the top quarter, with some in the second quarter, and the other 504 hours (68% of the time) to fall in some of the second quarter, and fully in the bottom two quarters.  They seem pretty close, though the mid-peak hours in summer seem particularly problematic (at least relatively).  We can return to this in due course.

For now, however, the point remains that there is different system-wide demand at different times – thus, the motivation for time-of-use prices.