With that said, there are still limits to renewable technology and Saskatchewan will require the baseload power produced from coal, natural gas and hydro.
– Brad Wall
Here, Mr. Wall is repeating a common but fundamental misunderstanding of the ways in which electrical grids can be configured. In discussion of electricity futures, it is never long before someone mentions baseload. This is a much misunderstood term, and it is important to the present discussion that it be understood correctly. I hope that you will read on, so that you can understand just why “baseload” supply is neither necessary nor helpful in integrating renewables into the grid, but for those with limited time here’s the simplified version:
- What is needed when the wind isn’t blowing and the sun isn’t shining is power sources which can be quickly ramped up and down – the technical term is dispatchable sources.
- Options traditionally used for baseload – such as coal and nuclear – are not dispatchable. Indeed their inflexibility means they can only be used for baseload.
- Research has shown that, in fact, putting more of these inflexible sources on the grid makes it more difficult to integrate variable renewables.
- Increasingly, grid managers across the world are choosing to move away from baseload-dependence towards grids which integrate variable renewables and dispatchables – often a more networked grid reliant on multiple small generators rather than a few big ones.
OK, here’s the fuller explanation:
Power demand varies through the day; it also varies from day to day through the year. But there is always some demand – it never reaches zero. Baseload refers to that portion of electricity demand which is consistently present 24 hours a day, 365 days a year. So “baseload” actually refers to electricity demand: the term is being used carelessly when referring to any particular supply option. Traditionally, however, it has been simplest to meet baseload requirements by use of one or two sources running continuously. Certain sources – notably nuclear power and large coal-fired facilities – can only sensibly be used in this way because it is difficult (and sometimes dangerous) to make much change in their output. They operate best at full capacity, continuously, because they can’t be ramped up or down quickly. They are inflexible – or, to use the technical term, non-dispatchable. Hence utilities are very familiar with graphs something like these, showing the variation in demand through the day and the way in which different power sources are combined to meed that demand:
In these graphs, nuclear and coal provide the baseload, and hydro and gas the “intermediate” (or “load-following”) and peaking power. Normal SaskPower practice to date has been to use coal for baseload, hydro for intermediate and gas for peaking – though some newer gas-fired capacity is now used for baseload.
Such an arrangement would still work well if carbon emissions and radioactive waste did not pose problems (and if nuclear power could be made affordable without subsidy). However, it is not the only way to run a grid, and indeed becomes difficult with large proportions from variable sources (such as wind and photovoltaics). The point is to supply total demand, and there is more than one way to achieve this. Wherever significant proportions of variable renewables (such as wind and photovoltaics) are used, a shift away from the old approach becomes necessary. The result is daily supply graphs more like these:
As the proportions of inflexible “baseload supply” sources decrease, the potential for greater flexibility increases. This is shown by the experience in Denmark, where closure of old large coal-fired plant has enabled increased targets for the proportion of wind power on the grid. It has also been shown in research at Germany’s world-renowned Fraunhofer-Institut. Even so, moving away from the old inflexible-baseload-focussed approach requires careful planning. It is important to maintain security of supply, to the same degree as is achieved currently under the old system. (Preferably to a greater degree – it should be noted that Saskatchewan’s grid is substantially more vulnerable to outages than the much more renewables-exposed grids in Germany and Denmark.)
The reality is that dispatchables are required in any grid, but “baseload” is not – and inflexible sources are essentially in competition with variable renewables for grid slots.
This has been recognised not only in Denmark’s two grids, but also by grid managers in the English-speaking world. In 2009, (US) Federal Electricity Regulatory Commission chair Jon Wellinghoff said:
“I think baseload capacity is going to become an anachronism. Baseload capacity really used to only mean in an economic dispatch, which you dispatch first, what would be the cheapest thing to do. Well, ultimately wind’s going to be the cheapest thing to do, so you’ll dispatch that first…. People talk about, ‘Oh, we need baseload.’ It’s like people saying we need more computing power, we need mainframes. We don’t need mainframes, we have distributed computing.“
David Olsen, a member of the board of governors of the California Independent System Operator (CAISO), serving 30 million people in most of California and a small part of Nevada, speaks for several of his colleagues when he says that:
“Having a 24/7 nuclear plant, from a grid operator’s standpoint – that is a real problem. Dealing with 2,200 MW coming in at every minute – we have to design our grid around that inflexibility. ‘Baseload’ refers to an old paradigm that has to go away.”
And Steve Holliday, CEO of National Grid, which operates the electricity transmission networks throughout Britain and in the north-eastern United States, has stated that:
“The idea of baseload power is already outdated.”
It makes much more sense for SaskPower to develop a grid configuration which is compatible with a shift to 100% renewables – some variable like wind and solar, some dispatchable like biosyngas and hydro, and incorporating appropriate amounts of storage backup, power trading with Manitoba and smart grid flexibility.
There is also an excellent analysis of the baseload fallacy here.
For answers to more talking points click here.