By: Donald Jones, P.Eng., retired nuclear industry engineer – 2014 August
Ontario’s Independent Electricity System Operator (IESO) has a pilot project that uses motor/generator flywheels, batteries, and aggregate loads as short term energy storage to, it says, provide regulation services (reference 1). These short term energy storage systems should not be confused with longer term storage systems like pumped water storage, compressed air, thermal etc. Regulation is secondary frequency control (reference 2) and can be automatic (AGC – automatic generation control) or manual and brings grid frequency back into its narrow control band after an under frequency or over frequency event has been arrested and the grid stabilized by fast acting primary frequency control/response. This regulation service is normally supplied by selected hydroelectric units at Niagara Falls and even by Ontario’s coal fired units before they were shutdown. Combined cycle gas turbine (CCGT) units can also be used to provide regulation service.
The addition of large amounts of unreliable wind generated electricity to the Ontario grid will/has caused a deterioration in frequency control (reference 3). Wind does not provide any passive inertial response/energy storage capability or active primary frequency control. The wind generation displaces conventional generation on the grid, initially the CCGTs and then some hydroelectric, with the consequent loss of the passive inertial response (energy storage capability) of the rotating masses of the conventional units to help limit frequency perturbations from such things as wind gusts on the wind generators. As well this results in the loss of the active primary frequency control capability of the conventional units. Primary frequency control is automatic and is provided by the speed governors of individual generating units to very rapidly arrest any drift in frequency due to mismatches in supply and demand on the grid. Primary frequency control is essential for grid stability. Reduced amounts of primary frequency response on the system can result in under-frequency load shedding and cascading outages. Since the nuclear units are presently operated in their turbine-following-reactor mode of operation they cannot provide primary frequency control and only provide passive inertial response (reference 2). This means that during periods of surplus baseload generation (SBG) that usually occur when demand is low and wind conditions are favourable all the large CCGT units and some hydro units are shutdown so frequency response on the Ontario grid will solely depend on the primary frequency response of the operating hydroelectric units with enabled speed governors and the energy storage (inertial response) capability of the nuclear and hydroelectric units. This results in a jittery grid caused by wind gusts and larger swings in frequency after an upset resulting in the need for more megawatts of secondary frequency control/regulation to return grid frequency back into its narrow control band. During periods of SBG without wind generation the grid would still not have the inertial response of most of the gas-fired generators but would be more stable because it would not be subject to the rapid frequency upsets from wind. The shutdown of the CCGTs and some hydro because of SBG, with or without wind generation, also means that the reactive power support and voltage control provided by these units is no longer available with, apparently, little if any affect on grid voltage? If voltage support were a concern and the CCGTs and hydro units did not have a synchronous condenser mode of operation, which they likely don’t, then other equipment that provide local voltage support would have to be used.
CANDU nuclear units supply around 60 percent of Ontario’s electricity so if these units provided primary frequency control it would reduce the magnitude of frequency control action from other generators and help stabilize the grid, as well as reducing the amount and use of secondary frequency control/regulation. The Darlington and Bruce units were designed to operate in the reactor-following-turbine mode as well as in the turbine-following-reactor mode, the present mode of operation in Ontario (reference 2). In the reactor-following-turbine mode primary frequency control would be active so each unit would be able to provide around +/-2.5 percent of full power to arrest frequency changes. 10,000 MW, say, of nuclear on the grid in this operating mode would provide up to +/-250 MW for primary frequency control. The units were not designed to provide regulation service (although they could – see later). To provide the +/-2.5 percent of power for primary frequency control the units would have to operate at less than full power, 97.5 percent of full power. However the nuclear operators, Ontario Power Generation and Bruce Power, have configured the adjuster rods in the reactor to give better fuel burn-up and operate their reactors at constant load at their full licensed power output in the turbine-following-reactor mode that does not provide primary frequency control. This was done long before there was serious wind on the grid. The operators say this mode provides more stable operation as well, incidentally, as providing more revenue. To enable reactor power to vary in the reactor-following-turbine mode would require approval from the nuclear regulator, the Canadian Nuclear Safety Commission (CNSC), but the change would likely not need the reconfiguration of the adjusters to provide this amount of primary frequency control.
There is a better way that the nuclear units can provide primary frequency control (more than +/-2.5 percent if necessary) as well as regulation service, operating reserve, and load following. This is to provide a steam bypass-following-turbine mode of control as on the Enhanced CANDU 6 proposed for Darlington B (reference 3). This approach does not require the reactor to vary power, a concern of the CNSC. Varying the amount of steam bypass provides a very responsive regulation service as well as primary frequency control. One or more units could have speed governors controlled by IESO generated signals to provide regulation service/automatic generation control if required. The upcoming refurbishment of the nuclear units gives the opportunity to do this to provide more operational flexibility to the grid.
Practical GW size long term energy storage could be extremely useful for nuclear, but not for unreliable wind. If energy stored overnight is to be relied upon to meet the day’s peak demand without the need for open cycle gas turbine units then it has to come from a reliable source like surplus nuclear generation from new build. This new nuclear could also supply primary frequency response, regulation, reactive power, operating reserve and load following as well as contributing its rotating masses to the grid, all greenhouse gas-free. However, the eventual conversion of Ontario from gas space heating to electric space heating (reference 4) will result in more electricity being used during those long cold winter nights and daytime heating load will tend to bring the winter heating load into better alignment with the summer cooling load. All this will affect the grid’s capacity factor and impact the rationale for long term energy storage and should be discussed in Ontario’s next Long-Term Energy Plan. Ontario should prepare for this conversion now by building more nuclear units that will have very flexible load manoeuvring capability (Enhanced CANDU 6) that will adapt to the changing demand on the grid. Either that or go back to coal when gas becomes unaffordable, or unavailable if it comes from south of the border. It would be dangerous to rely on Quebec to supply Ontario’s future electricity needs since last winter it was hard pressed to meet its own heating needs.
So why does the IESO want to add motor/generator flywheels and batteries to the Ontario grid? It can’t be for reactive support and voltage control, as the IESO states, because this service can be supplied by the conventional generators on the grid. And it’s not for regulation in the traditional sense of the term, as the IESO states, since regulation refers to the relatively slower secondary frequency control follow-up to the very fast primary frequency control and the IESO is asking for very fast second-to-second response. No, they must be needed because wind/solar generation does not provide the inertial frequency response necessary to help stabilize the grid after short term upsets in frequency from things like wind gusts. Flywheels and batteries cannot replace the primary frequency control function of conventional generators that need a continuous source of energy to stabilize the grid after a major grid upset, like loss of a generating unit, or to supply regulation service, secondary frequency control. They can only try to replace the lack of inertial response of the wind generators and the inertial response of the conventional generators that have been displaced by the wind. A reduction in grid inertial response means an increase in the number of times that the grid frequency will exceed its deadband and bring in primary frequency control from the generators with enabled speed governors resulting in wear and tear to the units. If the nuclear units operated with enabled governors, as discussed earlier, this would reduce the wear and tear to individual units and increase the amount of primary frequency response MW available to stabilize the grid after a major upset. Of course sufficient grid inertia could be maintained by not taking all the CCGTs off line but this would result in greenhouse gas emissions and less wind generation. It would be nice to know what frequency control on the grid was like during past episodes of SBG when most of the gas-fired units were off line and hydro at a minimum, with or without wind. It must have been bad enough to warrant these flywheels and batteries.
If the Ontario grid did not have wind/solar any frequency control problem would disappear with increased demand and more use of CCGTs that would provide inertial response to the grid. However, with the tremendous amount of wind/solar that will be on the grid within the next decade (around 10,000 nameplate MW) this problem could be with us for a very long time.
Since these short term energy storage systems only seem to be required because of the lack of inertial response of the wind generators the cost of these systems, that the IESO thinks are needed, should be included in the capital cost of wind generation and not be part of the IESO costs for regulation service as the IESO seems to suggest. Without wind the IESO would not have a rationale for these systems. Wind and solar are not needed on the grid anyway for environmental reasons since the grid already has reliable greenhouse gas-free nuclear and hydroelectric generation. In fact over 80 percent of Ontario demand is met by greenhouse gas-free generation. More nuclear should be built to replace gas-fired generation the aim being a greenhouse gas-free nuclear-hydro grid and, incidentally, all the frequency control one could ask for.
1. RFP for Energy Storage Services – Backgrounder, 2014 July, IESO, http://www.ieso.ca/documents/media/Backgrounder-Energy_Storage.pdf
2. A quick primer on how CANDUs fit into Ontario’s windy power grid, 2013 July, Don Jones, https://thedonjonesarticles.wordpress.com/2013/07/06/a-quick-primer-on-how-candus-fit-into-ontarios-windy-power-grid-2013-july/
3. How wind affects Ontario’s power system dynamics and effect on CANDU refurbishment, 2014 April, Don Jones, https://thedonjonesarticles.wordpress.com/2014/04/01/how-wind-affects-ontarios-power-system-dynamics-and-effect-on-candu-refurbishment/
4. Ontario’s electricity – greenhouse gases up, cost up, security down, 2013 December, Don Jones, https://thedonjonesarticles.wordpress.com/2013/12/21/ontarios-electricity-greenhouse-gases-up-cost-up-security-down-2013-december/