Exploring Hydrocarbon Depletion
Page added on March 13, 2017
Current US energy policies can be broadly summarized as strong support for renewable wind and solar, but no effective penalties for fossil fuels. The politics that have led to those policies are fairly obvious. We needn’t discuss them here. Less obvious, though, are the consequences. It’s only a slight exaggeration to say that we’ve saddled ourselves with the worst of both worlds. Renewables have increased the cost and reduced the economic efficiency of our electricity system, but haven’t bought a proportionate reduction in carbon emissions.
I know those statements will raise hackles among renewable energy advocates. Let me say upfront that I’m not anti-renewables. Next to avoiding nuclear war, I think rapid global warming due to high and rising greenhouse gas levels is the most critical problem that the world faces. Sustainable, zero-carbon energy resources are vital to addressing that problem. So I have no problem with measures that increase the cost of energy — as long as those measures are effective in fixing the carbon problem. What bothers me is that they’ve been, if not entirely ineffective, at least much less effective than they should have been.
The fundamental reason for that impaired effectiveness is the incentive structures we’ve set up. Specifically, it’s our reliance on subsidies that mask and distort the true cost issues, while failing to effectively “de-incentivise” fossil carbon emissions.
The path not taken
Ask any economist, liberal or conservative, what the most effective way to reduce carbon emissions would be. You’ll get pretty much the same answer: set a price on carbon emissions that will capture the external costs of using the atmosphere as a dump for CO₂ emissions. We haven’t done that. Instead, we’ve implemented various subsidies that strongly encourage the deployment of wind and solar resources, while not overtly penalizing carbon emissions.
The mental model to rationalize that policy is that there’s a set demand for energy, and that every unit delivered by a zero-carbon wind or solar resource will reduce carbon emissions by the amount that fossil fueled resources would emit to generate the same unit. So supporting renewables ought to be enough; no need to go head-to-head with fossil fuel interests to levy a price on carbon emissions. Sounds reasonable, right? One can find that model reflected in most calculations by proponents of the carbon benefits of renewable energy.
Unfortunately, the model is wrong. It ignores the cost of intermittency on carbon emissions. In a policy environment that does not penalize carbon emissions, intermittency has a perverse effect on the average efficiency of carbon-fueled generation used to supply whatever isn’t supplied by the intermittent resources. It comes down to how economics play out in combination with the characteristics of the various energy resources that could be used to supply the backup for intermittent assets.
Economics of backing supply
From an economic perspective, the defining attribute of wind and solar resources is that costs are overwhelmingly in up front capital. Once those costs have been sunk — the resources have been built — the marginal cost of using them is essentially zero. Sunlight and wind are free, and while solar and wind farms do have operational and maintenance costs, even those costs are almost entirely independent of the actual energy output. That means that, in the wholesale electricity market, renewables can always underbid any other electricity resource for whatever amount of energy they have to sell. Yet their supply is totally inelastic; price has no effect on output.
Being able to underbid any other resource in the wholesale market does not mean that wind and solar are necessarily cheaper overall. The wholesale price may be well below the average price the solar or wind farm owners need to recover their investment. But since electricity is not a storable commodity, any amount that could be sold, but isn’t, is wasted. That makes any price above the marginal production cost better than nothing. And since that marginal cost is zero …
All other power resources are left playing second fiddle. They get the pieces of unmet demand that wind and solar leave on the table. In the absence of an effective price on carbon emissions, that second fiddle role has strong negative consequences for overall system cost and carbon emissions.
After renewables, the next lowest marginal cost resource seems like it should be high efficiency fossil fueled generation: combined cycle gas turbine units (CCGT) or supercritical steam coal fired plants. Those units would be expected to have the lowest specific fuel costs and hence the lowest marginal production costs. But it’s not quite that simple.
CCGT’s and coal-fired plants alike cannot simply be switched on and off like electric lights or motors. They require lengthy warm-up periods while they ramp up to rated capacity, and equally lengthy cool-down periods while they ramp down. To minimize damage from thermal stresses, It can take an hour or more to ramp one of these plants up to rated power, and a comparable time to shut it down. During ramping periods, they’re burning fuel, but producing either no electricity or much less per unit of fuel than they do when operating at rated power.
The result is that their marginal cost of production depends strongly on the duration over which steady production will continue. If a unit will be running at rated power for at least a full day, then low efficiency during startup and shutdown won’t matter much. But if a unit is being activated just to cover a temporary lull in solar or wind farm output, its net efficiency over an hour of production might be only half of its rated efficiency for steady operation. Moreover, maintenance costs will soar. “Wear and tear” on thermal power units is mostly due to metal and weld fatigue from thermal cycling stresses. And the faster the ramp rate, the larger the thermal gradients and the greater the stresses.
Those physical considerations mean that high efficiency CCGTs are almost never dispatched as short-term fill-in for temporary lulls in wind farm output or for drops in PV output from passing clouds. CCGTs don’t compete in the market for short-term backing of renewables.
What does compete most commonly are peaking units: simple gas combustion turbines or reciprocating engines. Most of the time they sit idle, but they can be dispatched quickly when needed to meet peaks in demand. They’re cheap, tolerant of frequent cycling, and don’t suffer undue loss of efficiency from short periods of operation. Of course, their efficiency is pretty low to begin with: typically about half of what a CCGT can deliver, or two thirds of what supercritical steam coal-fired plants manage. But efficiency is unimportant, so long as fossil fuels are cheap and there’s no penalty for carbon emissions.
At this point, advocates for renewable energy should be protesting. What I’ve just described is not the way it has to be! There are ways to deal with the intermittency of wind and solar resources that don’t lead to reduced efficiency and higher carbon emissions from fossil-fueled generators remaining on the grid.
That’s true. Certainly, in regions well-supplied with hydroelectric power, there’s no problem. Within bounds, hydroelectric is fully dispatchable. It can fill in for non-producing wind and solar resources at very little cost. But even when adequate hydroelectric capacity is lacking, there are ways to provide backing power that avoid the poor efficiency and carbon emissions of peaking units. Ideally, backing power can be supplied from storage — batteries, pumped hydro, or other prospective options.
Even if available storage options are too expensive for long duration supply — as indeed they are — there are proven options for high efficiency fossil-fueled generation that play well with renewables. New “flexible generation” units don’t suffer excessive wear and tear from cycling. They also have broader throttling ranges and can be operated well below rated output without undue loss of efficiency.
Unfortunately, in the absence of a meaningful price on carbon emissions, none of these cleaner options for coping with intermittency is economically viable. Technical merits don’t matter when the cheapest solution for intermittency is simply to fire up a peaking unit and burn a bunch of cheap fossil fuel. Carbon emissions be damned!
Readers will notice that I haven’t answered to two obvious questions: (1) what’s the actual quantitative impact on carbon emissions from inefficient backing of intermittent renewables; and (2) what price on carbon emissions would be needed to make efficient backing options more economically attractive?
Unfortunately, neither question has a firm answer. In both cases, “it depends”.
The actual impact on net carbon emissions depends very much on regional particulars. Regions blessed with a lot of dispatchable hydroelectric capacity, as already mentioned, don’t have to resort to inefficient backing at all. When available, hydroelectric is more than competitive with power from any source that has to purchase fuel. But it’s limited.
In regions without adequate hydro, the answer still depends on the particular mix of assets, as well as the comparative growth rates of renewable energy vs. total demand. If demand is growing faster than renewable energy capacity, local utilities will need to add new conventional capacity. They may be able to justify the capital cost of acquiring flexible generation. Flexible generation units are nearly equal in efficiency to the newest baseload units. More expensive, but their ability to provide efficient load following capacity and to back intermittent renewables can make them good investments.
Among OECD nations, regions in which total demand is growing faster than renewables are being added are the exception. More commonly, growth in renewables exceeds growth in overall demand. Then local utilities are facing a progressive reduction in the capacity factors for their conventional generation assets. They can’t simply scrap them, however, because their generating capacity will still be needed at times.
In that situation, it’s hard to justify the cost of acquiring any new capacity, regardless of how well it performs as a backing source. Unless a stiff price on carbon emissions puts a premium on carbon-efficient generators, the tendency will be to rely on available peaking units. Net efficiency of a utility’s conventional generation fleet will take a hit. However, only detailed simulations including the fleet makeup, regional weather patterns, and demand profile can really quantify the impact.
What price emissions?
The question of what price on carbon emissions would be needed to make efficient backing options competitive also has no simple answer. It depends on the level of grid penetration by renewables that we’re talking about, and on what technologies for efficient backing are available or will become available in the near future. It also depends on how close to being competitive those technologies already are. Then there’s the added complication that costs depend heavily on time and scale of deployment. Those factors make the whole notion of what is “economically competitive” a bit fuzzy. For technology and business development with high capital investment, perceptions of risk versus expected return on investment (ROI) weigh heavily.
There are almost certainly options for clean backing of intermittent resources in particular cases that could be more economical than peaking units, even without a price on carbon emissions. But there’s no consensus on which, if any, are scalable and offer a sufficient cost advantage (again, absent a price on carbon emissions) to justify the financial risk in pursuing them. The power business is notoriously conservative.
An example of a backing solution that should be economically viable now would be a seawater reverse osmosis (SWRO) desalination plant built as an adjunct to an existing nuclear power plant (e.g., Diablo Canyon). NPPs are not, by themselves, dispatchable resources. However, the pairing of NPP and flexible SWRO plant creates a zero-carbon flexible generation source. Its economic viability is aided by the particular nature of the RO desalination process.
Capital-intensive facilities — such as an SWRO desalination plant — are not ordinarily attractive for use as flexible loads. That’s because so much of the cost of their output derives from the cost of capital. Reducing their output to let them serve as flexible loads increases their cost per unit of output. However, in an RO desalination plant, the relation between freshwater output and energy consumption is non-linear. Operating the plant at a lower power level will reduce its output, but it also reduces pumping pressure and improves net energy efficiency. The reduction in output is only about half as much as a linear input-output model would predict.
The effect of that non-linear behavior is that 20% of a desalination plant’s nominal power draw (for example) can be diverted to the grid, but its freshwater output will only be reduced by 10%. The plant’s capital productivity for fresh water will take a 10% hit while it’s operating at the lower power level, but in return, it will be releasing 20% of its nominal power draw back to the grid as high value dispatchable power. It’s a cheap zero-carbon virtual power plant (VPP) that’s called into existence when needed and vanishes when not. As a VPP, it has no fixed capital cost that contributes to its LCOE; when the VPP is not “producing” power, the portion of the RO plant’s capital cost temporarily allocated to the VPP function just reverts to desalination.
Alas, as with hydroelectric power, opportunities for paired production of nuclear electricity and desalinated water are limited. It’s not a general solution to the problem of backing intermittent renewables. But the strategy of pairing baseload generation with flexible loads does have broad applicability. It can be used with fossil fueled baseload plants paired with other types of flexible loads. The result won’t be zero-carbon backing power, but if the capital cost of the flexible load is low enough, the result will be better than reliance on peaking units. It effectively creates a high efficiency flexible generation source without the cost of new generating capacity. There will be costs in developing and managing the required flexible loads, but there are good options.
One of the best is ice-based domestic refrigerators. Refrigerators collectively account for roughly 10% of the nation’s household electricity demand. Since they make ice anyway, it’s an easy, low cost design change to have them use stored ice as the cooling source for the refrigeration compartment. That makes timing of their power draw for ice-making flexible. If the unit is linked by the internet of things (IOT) to an aggregator service for the local transmission system operator, it becomes part of an efficient VPP. The power that it releases to the grid will be both cheaper and cleaner than output from a peaking unit.
The storage conundrum
At high penetration levels for renewables, coupling baseload generation with flexible loads likely won’t be enough. To avoid resorting to inefficient peaking units when renewables are not producing sufficiently, it will be necessary to tap energy previously stored when production exceeded demand. Advocates for renewable energy have made much of the rapidly falling costs of battery storage resulting from the EV market. Low cost battery storage is widely seen as a critical factor that will enable penetration levels for wind and solar that the grid would not otherwise be able to handle.
The problem is that the cost of battery storage — not to mention battery lifetimes — will need to improve a lot more before battery storage is any threat to untaxed fossil carbon. Battery storage is viable today for applications where there’s no good fossil fuel based alternative. Those are applications that require fast response power delivery over intervals of minutes to maybe an hour. That includes such things as peak shaving (for businesses billed on the basis of their peak usage), output smoothing for wind and solar farms, and emergency supply while other backup resources are ramped up. But battery storage is still several times too costly for general load shifting or primary backup of renewables. Not if it has to compete with peaking generators.
Power from stored energy derived from electricity can never be truly cheap, because its “fuel” is electricity that has already been produced from another source. And while surplus power from wind and solar is often dumped at a small fraction of its average cost, it normally takes only a small amount of flexible power charging to consume the surplus and restore wholesale prices to close to their long term average. On top of that, conversion and storage losses further boost the cost of storage output, even before figuring in cost of capital and depreciation for the storage system itself. That’s true for any form of electrical energy storage, regardless of particular technology.
All that is not to say that storage isn’t a desirable solution; in fact it’s necessary for zero-carbon energy supply in a mostly renewable energy economy. But it means that large scale stored power output will always be substantially more expensive than the average cost of primary generation. It will be nearly impossible for any storage technology to compete with generation from fossil fuels if storage can derive no monetary credit for zero carbon emissions.
The bottom line is obvious. If we are truly serious about reducing carbon emissions and slowing the pace of global warming, then we really, really need to get away from the current non-market system of subsidies and mandates for wind and solar and move to an economically efficient market-based system that directly addresses the problem of carbon emissions. The key is well understood: a revenue-neutral carbon tax.