Category: Battery storage

Could the transition drive emissions up?

dchone July 12, 2022

One question that comes up quite regularly about the energy transition is the amount of energy, and therefore emissions, required for the transition itself. This is the energy required for making solar PV modules, wind turbines, batteries and so on. Further up the supply chain there is also the energy required for the additional minerals, such as the lithium, nickel, cobalt and copper found in an electric vehicle (EV). These not only have to be mined, but also go through extensive industrial transformation and refining processes to make the actual materials required for the end use. Today, most of these processes use oil, coal and gas for energy, giving rise to carbon dioxide emissions.

Perhaps the most energy intensive part of the energy transition is the manufacture of lithium-ion batteries, now being widely deployed in EVs. Some commentators have even questioned the effectiveness of the EV as a mitigation route, particularly when the battery is made in China (currently a heavy reliance on coal for energy) and the vehicle is driven in a country with a high electricity emissions intensity (e.g. a country like Poland still largely dependent on coal fired power stations). The problem with this argument is that transitioning in a series of steps (e.g. first decarbonise the electricity supply, then start deploying electric cars) would take decades longer than transitioning in parallel steps (i.e. decarbonising the electricity supply at the same time EVs are deployed). Nevertheless, the parallel approach could drive up emissions in the short term, the question is by how much?

The manufacture of batteries for EVs provides a good example of the problem. In a recent article, MIT report that the Tesla Model 3 holds an 80 kWh lithium-ion battery and the CO₂ emissions for manufacturing that battery would range between 3120 kg (about 3 tons) and 15,680 kg (about 16 tons), depending on the manufacturing location. The article notes that the vast majority of lithium-ion batteries—about 77% of the world’s supply—are manufactured in China, where coal is the primary energy source. That means most batteries are currently made with CO₂ emissions at the higher end of the range, although as battery factories spring up across the world and particularly in the EU and US, that picture will change.

Bringing together a few assumptions about battery manufacture, EV deployment and embedded CO₂ in both manufacture of EV batteries and driving EV cars, it is possible to get a back-of-the-envelope view of the scale of the issue. I will assume the following;

EV production rises from current levels (some 7 million vehicles per year) to all EV production globally by the mid-2030s (i.e. no more internal combustion engine cars are built after that time). This is an aggressive transition, but probably the minimum that is required for a 1.5°C goal.
Higher CO₂ emission battery manufacture is currently at 77%, but the share declines to 40% by 2060 and the higher CO₂ emissions also fall by 75% over the same timeframe as the manufacturing system decarbonises.
Lower CO₂ emissions manufacture is therefore 23% now, but rises to 60% by 2060 and the manufacturing CO₂ emissions fall to zero by 2050. Decarbonising industry to such an extent will require a variety of technologies, with carbon capture and storage playing a critical role.
The 80 kWh battery delivers 300 miles of range and the average vehicle travels 10,000 miles per year.
The electricity supply which EVs use is on average 0.4 tonnes CO₂ per MWh now, falling to zero by 2060. The actual global average grid intensity is higher than 0.4 today, but EVs tend to be driven in lower intensity regions at the moment, e.g. the EU, California etc.
An EV produced today has a 15 year life.
The EV mitigates emissions from internal combustion engine vehicles at a rate of 120 gms/km. As a simplification, this doesn’t change throughout the calculation. It assumes that smaller cars are replaced earlier and that the average fleet efficiency of internal combustion vehicles improves over time.
The battery represents a net increase in car manufacturing emissions with other emissions in the manufacturing process about the same for both EVs and internal combustion vehicles.

The calculation is for net-emissions, which is;

[Battery manufacturing emissions] + [Indirect EV emissions during driving] – [Gasoline / Diesel emissions backed out by EVs] = Net Emissions

What we see from the charts below is that global passenger car emissions rise before they start falling when net-emissions cross the zero line. This happens in 2035. Clearly the year in which this happens depends on the assumptions made, with the CO₂ from internal combustion vehicles not being used being a key determinant. For example, if this is raised to 180 gm CO₂/km, the crossover point is around 2030.

The outcome certainly points to the longer term benefit of the EV transition, with global cumulative emissions over 25 Gt lower in 2060 than they would otherwise be. This is a material reduction when thinking about a 500 Gt carbon budget for 1.5°C. However, it also highlights an issue with the current global goal to reduce emissions by 45% by 2030 relative to 2010, as set out in the Glasgow Climate Pact; the EV revolution that we are currently in the midst of is unlikely to contribute to that reduction. If anything, it could make the task even more difficult.

In a post some time back I noted that the only real opportunities for change which could make a material difference to global CO₂ emissions by 2030 are where replacement technologies are already being manufactured at scale or where governments are prepared to create social change. This quickly reduces the options to only three major opportunities:

Significantly curtailing coal-fired power generation through replacement with renewables;
Replacing internal combustion engine vehicles with electric vehicles; and
Ending deforestation.

Passenger vehicle emissions account for 4 Gt, or 10%, of global CO2 emissions today. If change in this sector can’t deliver any net reductions by 2030 and potentially adds to global emissions, then it calls into question any possibility of a 45% reduction in 8 years. Almost perversely, if EV production could be ramped up in the short term, the problem for 2030 gets worse while the longer term net global cumulative emissions picture gets better.

None of the above is to meant to argue against an EV transition, it is clearly the right way to go. But like many other aspects of the energy transition, it is more complex than it looks.

Battery trends

dchone August 2, 2021

Two key advantages of fossil fuels are the energy density of the fuel itself and the ease of storing energy that a molecule based fuel offers. Most homes have a huge energy store sitting in the car gasoline tank in the garage, or perhaps in an LPG / propane tank in another part of the house. The ease of storage makes transport relatively simple, with everything from passenger cars to A380 planes dependent on the need to carry fuel with them. But as the shift away from fossil fuels gathers pace and electricity grows in importance as the energy carrier of choice, one critical technology emerges that we all already use but will grow in size and scale – battery storage. We need batteries to store electricity for portable use and to store electricity at city level scale to manage the power grid, particularly as intermittent renewable sources become prevalent.

Battery technology dates back to around 1800, but domestic batteries were made popular over 100 years ago with the introduction of the AA battery in 1907 by the American Ever Ready Company, following on from their successful D cell flashlights. Today, we use batteries for a variety of household devices, but battery use across society is set to expand rapidly as the energy transition gathers pace. Further, as battery technology improves, these handy energy stores are making their way into more and more devices and applications.

In 2010, global battery production was less than 5 GWh, but with the arrival of the electric car and the growth in grid storage, production in 2020 was nearly 400 GWh (Source: Wood Mackenzie). There is also a significant and growing pipeline of Gigafactory projects, with manufacturing capacity around 1.3 TWh by 2030 based on known and expected projects. But what about the demand potential?

Numerous auto manufacturers have signalled their intent to bring internal combustion engine (ICE) passenger car manufacturing to an end, with dates between 2030 and 2040 often cited for the full switch to electric vehicles (EV). As a stretch, let’s assume that all passenger ICE production ends by 2035, which by then might mean 70 million EVs produced globally per year. If each car requires an 80 kWh battery, then that’s 5.6 TWh of new capacity required each year. Although recycling of batteries and battery components will eventually change the manufacturing landscape, that won’t be the case in the first half of the 2030s. At that time the availability of material for recycling will be the result of production today, which is a tiny fraction of our assumed production in 2035.

Grid storage requirements are a significant unknown. In a report published earlier this year by the research firm Frost & Sullivan, they predict additional global grid battery storage capacity additions will likely reach 135 GWh (0.14 TWh) in the next nine years from the 8.5GWh annual capacity additions that were recorded last year. But capacity additions are scaling rapidly, with the much talked about Tesla installed 100 MWh facility in South Australia in 2016 now easily eclipsed by multiple 300-400 MWh projects. In a 2020 study released by RethinkX, they estimated that for areas of the United States, a shift to 100% wind and solar would require some 40-90 average demand hours of battery storage. In 2020 US electricity demand was 4300 TWh, which would imply around 30 TWh of battery storage. However, it is possible that there is overlap between grid storage and EV storage, which by 2035 might have reached 12 TWh sitting in US garages and at charging points (assuming at least 50% EV penetration by then).

Assuming a rapid transition, the US alone might need 20-25 TWh of installed storage capacity by 2035, with global installed capacity perhaps reaching 100 TWh by that time. That would require a 35% year-on-year expansion of battery production capacity for the next 15 years as shown in the chart. That means in 2035 global battery production is close to 100 times current levels. It also requires manufacturing capacity in 2030 of 8 TWh, six times that of the current project pipeline for new facilities.

Batteries require particular minerals and chemistry, which today consist of lithium, nickel and cobalt in the current generation of Li-Ion batteries. The chemistry of batteries is the subject of extensive research, which points to much lower requirements for these minerals per kWh of storage. A recent analysis by the European Federation for Transport and Environment (Transport & Environment (2021), From dirty oil to clean batteries) states that over the period 2020 to 2030 the average amount of lithium required for a kWh of EV battery drops by half (from 0.10 kg/kWh to 0.05 kg/kWh), the amount of cobalt drops by more than three quarters, with battery chemistries moving towards a lower cobalt content (from 0.13 kg/kWh to 0.03 kg/kWh). For nickel the decrease is less pronounced – around a fifth – with new battery chemistry moving towards a higher nickel content as a fraction of the total, but still a decline per kWh (from 0.48 kg/kWh to 0.39 kg/kWh).

On the basis of the 2030 numbers above, production of 20 TWh (20,000 GWh in the chart above) battery storage per year in the early 2030s would need;

Metal	Early 2030s additional annual demand, million tonnes	Current global production, million tonnes	Required increase in global production
Lithium	1.0	0.09	10 times
Cobalt	0.6	0.15	4-5 times
Nickel	7.8	2.7	4 times

This is a significant step-up in metals production, with history pointing against achieving it.

Metals supply could well become a limiting factor in the energy transition given the potential demand for electricity storage. These levels of production increase are feasible over time, but in the space of 10-15 years they represent double to triple the historical trend, although that is true for almost everything in attempting to reduce emissions by 40+% in a decade. The above analysis also doesn’t account for other demands on batteries; from vans and trucks, small ships, barges, small planes, household and commercial devices and so on.

Equally, we shouldn’t discount innovation and different directions of travel. Apart from a slower energy transition, other factors that might influence the outcome are a more rapid evolution of battery chemistry towards more widely available minerals and/or a shift away from chemical batteries to other storage and balancing solutions in the electricity grid. However, don’t expect to see a truly novel solution scale sufficiently in just a decade. The first commercial Li-Ion battery was marketed in 1991 and has taken 30 years to scale to current levels. it was based on research and development over the previous 20 years.

In any case, the decade ahead could well be a period of rapid change and expansion for the global mining industry.

A ten year reduction challenge for the USA

dchone May 4, 2021

In April, at the opening of the White House Climate Summit, President Biden announced the United States Nationally Determined Contribution (NDC) for the period up to 2030, encompassing a target of greenhouse gas reductions 50-52% below 2005 levels. The US Greenhouse Gas inventory is available on the US EPA website and is summarised in the table below.

Sector and gases(Mt CO₂e)	2005	Latest data for 2019

CO₂ Fossil Fuel Combustion	5753.5	4856.7
CO₂ Non-Energy & Industrial	381	399.1
CH₄ from all sources	686.1	659.7
N₂O from all sources	455.8	457.1
HFCs, PFCs, SF₆, NF₃ from various industries	146.5	185.6
Total Emissions (Sources)	7423	6558.3

LULUCF Emissions	16.8	23.5
LULUCF Carbon Stock Change	(804.8)	(812.7)

Net Emissions (Sources & Sinks)	6635.0	5769.1

Given a baseline of 6635 Mt CO₂e, a 50% reduction in 2030 would require US emissions to be 3318 Mt or lower in that year. The 2019 data represents a reduction of 13% over 2005, so the reduction in the next 10 years must be nearly triple that of the last 15 years. The reduction seen to date from fossil fuel combustion is running slightly ahead of the overall improvement, with a reduction from 2005 to 2019 of nearly 16%.

The goal that President Biden put forward is aligned with a strategy that is designed to achieve a goal of net-zero emissions for the US in 2050. Recently, the Shell Scenario team released a US Sketch outlining one possible pathway to such an outcome. The Sketch is illustrated below (N.B. 2020 figures are a pre-pandemic view of US emissions which the US is approaching again as the recovery gains momentum).

The Sketch provides the opportunity to look at the nature of the transition required in the 2020s for the NDC. The Sketch energy system emissions in 2030 are 3440 Mt against a 2005 baseline of 6030 Mt (small differences in baselines result from different categorizations of emissions), or a reduction of 43%. While this isn’t exactly aligned with the NDC, it could be in that the NDC makes specific reference to a major expansion of sink capacity through enhanced soil carbon uptake in the agricultural sector.

Key elements of the journey to 2030 outlined for the US in the Sketch are as follows;

In 2020 (pre-COVID calculation) the US generated 3970 TWh of electricity, with 80% from coal, gas and nuclear. Solar and wind made up less than 10% of the total. By 2030 in the Sketch, the generation mix has shifted to natural gas and solar each making up about 25% of the total, with wind, nuclear and coal combined at around 40%. Notably, total electricity generation has increased to 5720 TWh, or 43% over 2020. This is in a pathway that still sees electricity system emissions at 600 Mt in 2035, the year President Biden has targeted for a zero emission system. His goal is about 5 years ahead of the Sketch.
The increase in electricity consumption comes from an increasing proportion of electricity in final energy, displacing gasoline in cars, natural gas in homes and coal in industry. Electrification is a major lever for decarbonisation. The proportion of electricity as final energy rises in a decade from 22% to 35%, breaking a near century long trend which if continued, would otherwise see it rise to 25% at most.
Over the 2020s, solar generation of electricity grows fastest, increasing by a factor of nearly 12 in a decade. The U.S. installed 19.2 gigawatts (GWdc) of solar PV capacity in 2020 to reach 97.7 GWdc of total installed capacity. With a utilisation factor of 15-20%, this could be expected to generate around 150 TWh of electricity, although the Energy Information Agency reported a total of 130 TWh (perhaps because the capacity at the start of the year was 79 GWdc). In any case, capacity by 2030 will need to grow to at least 1 TWdc, assuming some efficiency improvement in new installations compared to existing facilities. With a year-on-year installation growth of 30%, the US could be installing some 260 GWdc per year by 2030, bringing the total installed capacity to 1.2 TW by the end of that year. Very large scale energy storage will also have to deploy to support solar.
The Sketch also shows a three-fold increase in wind generation throughout the 2020s and importantly, no decline in nuclear generation. The latter becomes increasingly important as other non-intermittent generation sources decline in use.
Electric vehicle (EV) deployment is an important part of the trend towards electrification and it emerges rapidly in the Sketch. By 2030 the US will need to see 20% of all passenger vehicle use as electric, with hydrogen fuel cell vehicles rapidly appearing in the large SUV sector. At the moment, there are about 1.5 million EVs on the road, out of a total US passenger vehicle fleet of some 280 million. To reach a level of 20% EVs in terms of use, sales will need to climb rapidly from some 400,000 vehicles per year now, to over 13 million in 2030, or nearly 80% of all sales. This represents a year-on-year growth rate in EV sales of 42% through the decade.
Road freight makes less use of electricity, although it does impact the lighter end of the freight market. For larger, longer haul freight trucks, hydrogen emerges during the 2020s, with some 200,000 trucks on the road by end of the decade. This is still relatively small, but the sector is non-existent today. In the second half of the century in the Sketch, hydrogen becomes the dominant fuel for heavy road freight.
Natural gas, propane and oil are used widely in the US for cooking and/or heating in homes. Although the majority use electricity for cooking, heating is provided mostly by gaseous hydrocarbon fuels. In the US Sketch, heating and cooking in residences shifts from 13% electricity to 42% electricity in just a decade. Of all the transition tasks ahead, this may be the most challenging in that it involves convincing millions of households, individual consumers and landlords to change their behaviour, refit existing properties, invest money for the change and in some instances make use of a less preferred energy option.
While electricity already plays a significant role in light industry, this is not the case for heavy industry where very high temperatures and high thermal loads are often required for conversion processes. Heavy industry energy use in the US is currently three-quarters gas, coal and biomass and one quarter electricity. This begins to shift rapidly and by 2030 in the Sketch, 40% of energy demand for heavy industry is supplied by electricity. In addition, hydrogen is emerging, with the first industrial installations making commercial use of this fuel. While the amount is small in 2030 it could mean significant development, design and engineering of new processes in a limited amount of time. Alternatively, the amount could point to natural gas being topped up with hydrogen as a mixed fuel for industrial furnaces, prior to converting whole processes over to hydrogen based fuels.
The many changes across the energy landscape result in steep decline in the use of fossil fuels in the Sketch, with coal falling fastest. By 2030 there is a 40% reduction in coal use, a 14% reduction in natural gas and a 25% reduction in oil within the economy, compared to 2020 (pre-COVID) use.
But even by 2050 in the Sketch, natural gas and oil still play an important but significantly diminished role in the US economy. That points to the need for carbon capture and storage (CCS), both directly linked to facilities and indirectly as an atmosphere removal mechanism (e.g. BECCS). The US already has the most mature CCS industry in the world, underpinned by years of enhanced oil recovery using carbon dioxide. But an order of magnitude change is required here as well during the 2020s, with a shift from some 30 mtpa carbon dioxide stored to some 300 mtpa.

While the Sketch does not require nature based solutions (NBS) for the US energy system to reach net-zero emissions by 2050, natural carbon sinks do provide an additional lever to support the transition and the NDC also discusses this aspect of emissions mitigation. In the US, national forests cover 193 million acres, an area larger than Texas. Apart from their significant bioenergy potential, forests act as natural carbon sinks. The US could reduce CO₂ emissions by up to 300 million tons a year by 2050 through afforestation – the planting of new trees where there were none before. While new forests represent significant additional sinks, there is also potential to expand areas such as wetlands and mangrove swamps, and perhaps most importantly to encourage farmers to adopt practices that enhance soil carbon. The US NDC references this opportunity.

The rates of change outlined above are substantial, with year-on-year growth rates for several new energy technologies of 20-50%, depending on the sector. While growth at this level can be sustained when a sector is very small and this has been observed for renewable energy deployment, it becomes increasingly difficult as the sector grows and absolute levels of deployment become very large. Many of the sectors mentioned will face this issue during the 2020s. Nevertheless, the Sketch was designed to show what is possible given the right policy levers for innovation and deployment, sufficient and targeted financing and overall transition design.

Note: Scenarios don’t describe what will happen, or what should happen, rather they explore what could happen. Scenarios are not predictions, strategies or business plans. Please read the full Disclaimer here.

Exploring the energy future

dchone February 10, 2021

Three scenarios, Waves, Islands and Sky 1.5 explore the energy landscape out to 2100 and reveal very different possible pathways forward. All three are transformational in nature, with net-zero emissions now within sight, albeit on quite different timescales.

Powering remote activities

dchone March 28, 2019

Over the past three weeks I have been on a voyage from Cape Horn to the Cape of Good Hope, specifically Ushuaia to Cape Town. With good weather for most of the trip, we were fortunate to stop in the Falkland Islands, South Georgia and Tristan da Cunha (some pictures below). Each of these have communities ranging from a few people to two thousand in the case of Stanley and each has found its own solution to providing energy. Of course the other remote activity out here is the need of the ship itself which will have travelled for some 20 days without refuelling and carried 400 passengers and crew across the South Atlantic.

The Falkland Islands has the major settlement of Stanley, a couple of very small towns, a military garrison and numerous remote farms. The Islands have settled on wind power to displace diesel generators, achieving an average of 35-40% displacement and a peak of 54%. Rural wind power has also been a success for the remote farms. The next step is to look at the potential offered by modern energy storage technologies, although flywheels have been used since 2010 for some storage in association with the wind turbines. Given the geography and climate, neither solar or hydro have been a success, apart from some niche applications. But the Falkland Islanders also have a history of burning peat for heat, although this is in decline. Kerosine and diesel are the most common fuels used today.

South Georgia is completely different. Since 2008 South Georgia’s two settlements Grytviken and King Edward Point (KEP) have been powered by hydro electricity. On the slopes above Grytviken there is a dam, originally built by the whalers at the turn of the 20th century. The dam increased the capacity of Gull Lake to feed water to the first hydroelectric power plant in 1914. The electricity produced then was mainly used for lighting the whaling station. The plant was expanded in 1928 to reduce the station’s reliance on imported coal for steam to power the factory. The electricity produced was then used to power winches and other factory equipment. The new turbine house has been built just off the pathway from the settlement to Shackleton’s grave, with the only visible sign of it as a generating station being the small stream of water seemingly running from under the building and into the bay.

Tristan da Cunha has a single settlement of about 270 people and an export factory to process the fish and lobster that are caught around the island. Although much of the electricity system was replaced about a decade ago after a fire, an entirely diesel based system was rebuilt. In recent years some changes have taken place with the installation of a few home solar water heaters, saving on bottled LPG which is used to heat water in Tristan houses. A small solar farm was also constructed west of the fishing factory. It consists of 26 solar panels, aligned to face the northern midday sun and each capable of generating 250 watts, so a combined capacity of 6.5kW. The connection to Tristan’s electricity grid was made on 30th April 2015. There are further plans for change, but the logistics of getting equipment to Tristan is enormously challenging. The island can only be approached by sea and few ships stop there. As we discovered on a second stop there (to pick up a local government person so we could land on Inaccessible Island), weather can quickly close the port and conditions can persist for days. But the case for further change is strong, given the community dependency on the import of diesel fuel and LPG. The eventual solution for Tristan may be a combination of parts. Although there is excellent wind, it can be ferocious at times, bordering on hurricane conditions, which perhaps isn’t ideal for turbine operation. Solar can also be challenging, with thick cloud shrouding the island at times. And although there is ample rainfall, collecting this and channelling it through a hydro plant would also be very difficult given the geography.

This then brings the focus to the key dependency for all remote activities, the transport to get there and transport once there. As was the case for our vessel, all these locations are completely dependent on long distance, self-powered transport and that remains almost entirely powered by liquid fuels coming from petroleum. While renewables are starting to provide local energy solutions for remote activities, the energy for the transport associated with such activities has no immediate zero emission alternatives. Synthetic fuels, either from a biomass / biowaste starting point or formulated from hydrogen and carbon dioxide offer a simple drop-in possibility, but the bio-alternatives are still relatively small scale and the pure synthesis route is still at the pilot plant stage of development. It should be noted that large scale synthesis of fuels from hydrogen and carbon monoxide does exist, but the starting points are coal (SASOL in South Africa) or natural gas (Shell in Qatar). For a net-zero emission synthetic fuel, the hydrogen would need to be produced by electrolysis of water using renewable energy and the carbon extracted from the air as carbon dioxide.

Apart from synthetic fuels, the best prospect for change is perhaps hydrogen itself, in that there is good experience containing and carrying it and fuel cells can power even large motorised vessels such as ships. Nuclear exists on ships in the military, but after an attempt to demonstrate the feasibility of nuclear powered commercial ships in the 1960s, nothing more has come from this form of propulsion. The challenge will lie with the providers of heavy transport; shipping companies, airlines and aerospace companies and large road haulage entities. Perhaps like the remote activities themselves, different solutions will emerge over time for the various requirements faced. Some remote locations may even be well placed to provide hydrogen in that they could have an abundance of renewable electricity to put towards hydrogen production via electrolysis. Scotland’s Orkney Islands are starting to experiment with such a route forward, as recently reported by the BBC.

Albatross chick on West Falkland

King penguins on South Georgia

A king penguin colony on South Georgia

King penguins on South Georgia

Me on Tristan da Cunha

The worlds most remote inhabited island, Tristan da Cunha

Northern rockhopper penguins on Gough Island

Fur seal on Gough Island

Fur seals on Gough Island

Inaccessible Island

Birds returning in the evening to Nightingale Island

The potential for rapid energy transition in North America

dchone March 5, 2019

Over recent weeks and months there has been considerable discussion in the USA and Canada on clean energy transition pathways, carbon pricing and the Paris Agreement. This has been catalysed by the IPCC 1.5° report, the arrival of new political figures on the scene and the prospect of elections in 2019 and beyond. As I discussed in my recent post on a new report that outlines a pathway for global transition consistent with 1.5°C, very steep emission reductions are called for in North America over the coming decade, as shown in the chart below. For the 1.5° case, a reduction of about 80% is proposed from 2015 to 2030.

Source: Achieving the Paris Climate Agreement Goals – Sven Teske (Editor)

But is such a rapid transition possible for an energy system as large as that found in the USA? This is similar to a pathway that has been proposed by those calling for a Green New Deal in the USA. But it is very different to that proposed by the Sky Scenario, released by Shell in 2018.

In Sky, to reach the goals of the Paris Agreement, global emissions need to peak in the 2020s and be falling by 2030, which happens in large part through an initial mobilization of effort throughout the 2020s. But emissions in the USA don’t get even close to an 80+% reduction in that time. North American emissions peaked around 2005 and are slowly falling, but in Sky the pace of change triples in the 2020s with a significant boost in effort, led by government policies such as carbon pricing and electric vehicle mandates and incentives. But by 2030, emissions have fallen by only 23% compared with 2015.

This would constitute a major shift in direction for the US economy. Five significant changes in the energy system in the decade to 2030 in Sky for North America are;

Doubling of wind generation. This should be a relatively easy goal, but on the basis of current build rates it would not be met. Current US capacity is around 90 GW with a recent quarter seeing 0.6 GW added.
A six-fold increase in solar generation. Current build rates would deliver about a third of the required capacity by 2030.
Coal use down by a third.
Natural gas demand plateaus and is the same in 2030 as it is in 2020.
Electric vehicle (EV) sales boom and by 2030, 20% of all vehicle miles traveled are in electric vehicles.

To imagine even more than this is quite challenging. In the EV sector, sales in the USA in 2018 grew rapidly, but still only amounted to 2.1% of the total (Source: EVVolumes.com). A looming issue is the lack of EV models in the SUV / light truck category, although this should start to correct in 2020 with several expected. Nevertheless, this category size makes up a large portion of the US market and it may take some time for consumers to adopt electric versions of their favourite model. While an all EV sales line up might be possible by 2030, an all EV fleet isn’t. For a 100% EV fleet to occur in 2030, all new passenger vehicle sales need to be electric in the next year or two.

In the power generation sector, the multipliers given for wind and solar above, together with an unchanged role for nuclear, results in over half the power generation coming from non-fossil sources by 2030. In Sky, it isn’t until after 2040 before coal is phased out, but natural gas persists in the power mix until the 2060s. A faster transition would require not only a further acceleration in solar and wind deployment, but also the development of major grid storage capacity to replace natural gas. Although there is some grid battery storage today, it amounts to some 2 GWh (Source: PV Magazine), a tiny fraction of the amount that might be needed for a 100% renewable energy system. Further technical developments will be required, or perhaps storage will be combined with ultra-high voltage long distance transmission.

But the most static sector over a ten year period may well be heavy industry; sub-sectors such as cement, iron ore smelting and petrochemicals. While Sky sees some shift towards electricity, the change from 2020 to 2030 is modest. Carbon dioxide emissions remain unchanged at 400 million tonnes per annum over this period. Change is measured in decades rather than years. This is because of the time it takes to develop new industrial processes that don’t use fossil fuels for energy (e.g. hydrogen based iron ore smelting), retrofit exiting capacity and build the necessary infrastructure to support such change (e.g. hydrogen generation).

The next ten years is critical for a transition that meets the goals of the Paris Agreement and this was a feature of the House Resolution for a Green New Deal. Significant changes must be achieved to set the scene for net-zero emissions in the decades following. But to imagine wholesale energy system change in s decade is not realistic. There may well be narrow pockets of very deep change found in parts of the system, but complete change remains a fifty year project.

Note: Scenarios are not intended to be predictions of likely future events or outcomes and investors should not rely on them when making an investment decision with regard to Royal Dutch Shell plc securities. Please read the full cautionary note in http://www.shell.com/skyscenario.

Stock, flow and cobalt. Issues for the EV transition.

dchone July 11, 2018

In Shell’s recently released Sky Scenario, around 1 billion electric passenger vehicles (EVs) could be on the roads as early as the 2040s. These 1 billion vehicles will each need a battery and while there are different battery chemistries available, it is likely that these batteries will need to include the vital element cobalt, a material that confers upon the battery good cycle ability and power output.

The new Tesla Model S has a Nickel-Cobalt-Aluminium Lithium Ion battery which contains just under 5 kg of cobalt, the lowest cobalt content of the various batteries that Tesla use.

While cobalt use per battery has fallen, with one manufacturer releasing information on the possible development of a cobalt free battery, some industry insiders believe that the Tesla Model 3 use of 10% cobalt in the battery cathode (8:1:1 Nickel:Cobalt:Manganese) is the minimum required for safety and battery longevity reasons.

If 8:1:1 were to be the dominant battery chemistry in the coming decades, an interesting dilemma emerges for automakers as they pursue electric vehicle development and rapid deployment. A look at global cobalt reserves, published recently in the BP Statistical Review of World Energy but drawing on data from the US and British Geological Survey, shows a current estimate of some 7 million tonnes, or 7 billion kilograms of cobalt. This is against current global production of 140 thousand tonnes, which has doubled over the last decade, giving a reserve to production ratio of 52 years.

The cumulative amount of cobalt (the stock) needed to be mined for passenger vehicle batteries will depend on the eventual number of cars in service (or batteries produced) and the amount of cobalt per battery. This stock is unrecoverable until cobalt is no longer used in batteries (or until cars transition again to some other propulsion mechanism), but it is reused repeatedly as older cars are scrapped, the cobalt recovered and recycled as new cars are built (the flow).

As the EV transition depicted in Sky gets going, the draw on cobalt increases and the stock builds. For example, by 2040 in Sky, 12.7 trillion vehicle kilometres are supplied by EVs, which could mean about 800 million to a billion vehicles on the basis of current usage patterns. At 4 to 5 kg each, that would require nearly 4 billion kilogrammes of cobalt, or more than half the global reserves. Before 2050 in Sky, the current global reserve of cobalt would be exceeded by EV deployment, without considering all the other demands for cobalt.

The above model assumes 100% recycle of cobalt as well, although this has little early impact in that recycle in 2040 will only come from production of EVs in the 2020s, which is still low compared to the potential 2040 stock. In any case, recycle is largely about annual flow, not stock. But an alternative model could emerge, which would see a much lower stock requirement for cobalt and therefore lessen the need to find alternatives, even with the same demand for passenger transport.

Autonomous vehicles offer the potential for vehicle sharing. If autonomous technology emerges rapidly, then the total number of cars in society can sharply decline under a vehicle sharing model, in that each car is used for a high percentage of the day, rather than parked awaiting the single use by its owner. While an individual car would likely have a much shorter life, say two to three years instead of fifteen, the stock requirement for materials such as cobalt would be much lower.

If by 2040 the number of EVs serving the above 12.7 trillion km demand is 100 million instead of 1 billion, then each vehicle would be driving 127,000 km per annum, leading perhaps to a two-year life after which there is either major refurbishment or even replacement. Annual production, or flow, of vehicles might remain largely unchanged, depending on the ratio of vehicles shared to vehicles owned and the life of a single vehicle under a high usage scenario. Whether the trend is towards refurbishment or replacement, the important materials such as cobalt are recycled and the total stock required by society is a fraction of the traditional ownership model. In a 100 million vehicle system, the required stock of cobalt could be as low as 500 million kilograms, just a fraction of current estimates of global reserves.

The above discussion introduces a new dimension into EV scenario thinking for auto manufacturers. The future might be dictated more by total vehicle stock considerations, in that manufacturers could encourage an alternative ownership model as the more cost-effective route forward to balance the demands on certain materials like cobalt. The alternative requires a continual search for different battery designs or seeking new deposits of hard-to-find minerals.

Note: Scenarios are not intended to be predictions of likely future events or outcomes and investors should not rely on them when making an investment decision with regard to Royal Dutch Shell plc securities. Please read the full cautionary note in http://www.shell.com/skyscenario.

Real costs beginning to emerge?

dchone December 12, 2017

There was great excitement coming from South Australia recently as the Tesla 100 MW generation capacity, 129 MWh storage battery commenced operation within the electricity grid. This is part of a much-needed upgrade to the grid such that it can better support the high level of renewable energy now available across the state. South Australia has been hit by a number of high profile blackouts over the past two years. While these were the result of various sets of unusual events unrelated to renewable energy deployment, the presence of a very high level of renewable energy had left the grid less robust, leading in turn to the blackouts.

The State Government has been under pressure to rectify the situation and the installation of a battery facility is one of the measures implemented, along with additional natural gas capacity. The addition of the battery also offers the opportunity to better understand the underlying cost of the renewable energy transition.

From the outset, the cost to consumers of renewable electricity generation has been hard to pin down. In all cases, solar PV and wind have been added incrementally to an existing grid, typically powered by some combination of coal, hydroelectricity, nuclear and natural gas. More recently, that incremental cost has fallen to very low levels, in some cases to just a few cents per kilowatt hour. These very low costs have been highly visible in the media, but do not represent the true cost for the system as a whole.

While very low marginal installation costs are great news for renewable energy deployment, the rapid increase in available capacity in some grids has played havoc with the existing power generation business model. When the renewable energy is available it reduces the dispatch of existing generation, thus cutting into the profitability, and possibly viability in extreme cases, of these operations. However, intermittent renewable energy cannot exist alone on the grid, so it is effectively undermining its own existence by driving existing generators away. This is a well-known issue and is being tackled in some jurisdictions with new business models; for example, the use of capacity payments to existing generators to ensure the viability of their facilities. This increases the overall cost of renewable electricity but doesn’t change the low marginal generation cost offered by additional renewable energy added to the grid.

In additional to capacity payments, the grid operator can provide further backup generation and energy storage, such as in South Australia. In the case of the latter, Tesla have built the 129 MWh storage facility, located adjacent to an existing 99 turbine, 315 MW wind farm. The Tesla battery is reported to have cost A$50 million. If it operated as a back-up to the windfarm itself, it would provide less than 30 minutes of storage, so this isn’t true backup for extended windless periods. Rather, the battery is there to address shorter term grid fluctuations and natural gas is added as a reliable back up for periods of low renewable generation. The need to keep the battery charged also reduces the output of the wind system for periods of time, however, that energy isn’t lost, just time shifted (apart from storage and transfer losses).

Exactly how the total cost of supply changes depends entirely on the situation for each individual grid, but that cost is greater than the marginal cost of renewable energy supply, even when expressed as a levelized cost of electricity (LCOE). This higher system cost for renewable electricity is how we should view this resource within a grid.

A solar transportation system?

dchone October 19, 2017

While Elon Musk is waxing lyrical about the Tesla Model 3, a house with solar roof tiles and batteries both in the house and in the car, an innovative group from Eindhoven in the Netherlands have combined all of this into a single offering and produced a solar powered family car that can also be used to supply energy to the grid.

The Stella Vie, claimed to be the most efficient family car ever built, recently featured in the World Solar Challenge, a biennial event which sees solar vehicles travel the 3000 km route from Darwin to Adelaide in Australia. The car competed in the Cruiser category. The 2017 event saw the third running of this class, which was introduced in 2013 when a four-seater ‘family car’ travel over 3000 km with an external energy input of just 64 kWh! To put this into context, a modern Tesla S (P100D) family car has a 100 kWh battery with a range of around 500 km, although in a very slightly modified Tesla S, a distance of over 1000 km has been achieved.

A report by The Guardian on the event, claimed that these vehicles represented ‘the future’. All of this begs the question as to whether solar powered cars are in our future?In fact, it is possible to buy a part solar powered car today. The new Toyota Prius plug-in hybrid has the option of a solar roof panel, although according to Toyota this adds just 5 km a day to its range. But to open this discussion further requires a deeper look at the numbers.

Jaguar solar powered concept vehicle (Source: Jaguar)

We could start by assuming that a future solar powered car doesn’t just have a single roof panel, but a surface made largely of a solar material, akin to the Tesla roof tiles on a house. It is conceivable that this extends to the windscreen glass as well. That means for a car such as the Tesla Model S, there is up to 10 m² of solar panelling. In the USA in June, solar insolation (i.e. the energy arriving from the sun) would be around 7 kWh/m2/day. If the solar material on the vehicle had a conversion efficiency of up to 40% (a leading-edge research result today), then 28 kWh of energy might be collected, but by December this would fall to as little as 8 kWh. These amounts are far in excess of the typical daily needs of the Stella Vie for family use (say 50 km/day, or ~1 kWh), but not so for the Tesla S, with heating, air-conditioning, media, various IT systems and heavier paneling for safety. In winter (50 km in a Tesla S equates to 10 kWh) there is an energy shortfall, although potential generation is in excess of typical daily needs in summer. The balance could be exported to the grid in the summer.

However, in central London, the vehicle would generate very little of its needs in winter (<2 kWh), but should still cover its own requirements in mid-summer (~18 kWh). On an annual basis, it would likely fall short on its needs (~10 kWh/day maximum generation vs. a similar daily requirement).

The numbers indicate that in some geographies, assuming efficiency improvements in both solar PV and the vehicles themselves, a family car that requires no net electrical energy on an annual basis is plausible, although considerable advances in material engineering to integrate solar PV into various surfaces would be required. In other geographies, the balance wouldn’t be as favourable, but the contribution to the annual energy requirement of the vehicle could be substantial.

But this isn’t the end of the story. For the vehicle to collect such an amount of energy it must be parked in the sunshine all day, or be on the road. It would also have to be kept relatively clean. This might be possible in towns or some (treeless) suburbs, but in larger cities with outdoor parking limitations and taller buildings, the energy collection potential could drop substantially. Nevertheless, there would be some contribution when on the road.

Given that Toyota have already started down this track with the Prius, it can’t be too long before others follow. Innovation in solar PV, materials technology and vehicle efficiency could see solar augmentation becoming widespread by 2030, with a global vehicle fleet potentially requiring little to no net energy by the early years of the second half of this century. While this may seem a long way off, even a high deployment scenario for electric vehicles takes until the mid 2050s to completely dislodge internal combustion engine vehicles from the market.

Other scenarios for vehicle development could change this balance. For example, a much smaller fleet of autonomous ride sharing vehicles would still require the same amount of total fleet energy, but with far fewer cars deployed, energy generation would likely be much lower. Nevertheless, the World Solar Challenge, like the Shell eco-Marathon where vehicle efficiency is tested to the limits, points to an interesting future for mobility.

A new era for aviation?

dchone October 9, 2017

The pressure to reduce carbon dioxide emissions and the prospect of a world running largely on renewable electricity has sent research and development teams in every sector back to their respective drawing boards to look at options that might exist for electrification. Perhaps the most challenging sector is aviation, where liquid hydrocarbon fuels are the only form of energy carrier available (mainly of fossil origin, but with some bio-origin fuels now appearing). The dependency on hydrocarbons is due to their high energy density and the challenge with fuel to weight ratio that planes have. However, fuel costs can represent up to 70% of total costs for an airline, so the business model tends to focus on efficiency as a primary consideration. Efficiency isn’t just about the plane itself, but about maximising passenger load, minimising extraneous weight, limiting taxiing and air traffic delays, using electricity for power at departure gates and optimising routes.

In the light of all the above, the idea of electrification in aviation is tantalizing, but there is little sight of this happening. At one end of the spectrum there is the prospect of single person electric drone taxis for short hops in cities, but after that there is nothing, until last week. EasyJet, in conjunction with the startup, Wright Electric, announced that they hoped to be flying a short-haul battery electric passenger plane within a decade.

The concept illustration looks impressive, but is this really feasible? Battery energy density is a key determinant and it is moving rapidly in the right direction. The energy density difference between the traditional Lead-Acid battery, still the standard for starting most cars and the best lithium based batteries is nearing a factor of 10, but lithium based batteries are still a long way from Jet A1 fuel as shown in the table below. The difference in energy density on a weight basis is around twenty times, in favour of Jet A1.

	Jet A1	Current Li-Ion Battery	Emerging Li-Metal batteries
Energy MJ/kilogram	42.8	~0.7	1.8
Energy MJ/litre	37.4	~2.0	4.3

Like an electric car, the efficiency of a battery electric aeroplane would be significantly higher than the combustion engine equivalent, although the starting point for a modern jet engine already exceeds that of vehicles. The chart below (IPCC Report on Aviation 1999, chart from 1991) shows that overall efficiency of jet aircraft falls in the range 20-40%, but significant improvements have been made since then. A modern Boeing 787 would show an overall efficiency approaching 50% on the same chart.

Even with near 100% efficiency for the battery electric aeroplane, the energy density of Jet A1 still gives that fuel a factor of ten advantage. As such, it will be distance that suffers, given there is a weight restriction for aeroplanes.

I am a regular traveler out of London City Airport and often see the Embraer 190 plane, which is similar in size to the easyJet concept photograph distributed with their announcement. But the Embraer 190 has a range of over 4,500 km, so one tenth of this gets near to the 335 mile range goal mentioned by easyJet, ideal for the flight I often take from London City to Rotterdam. So on paper, this would appear to work, but a plane with a range limit of a few hundred miles might significantly restrict the operational flexibility that airlines enjoy; for example, it couldn’t be swapped at short notice for a London City to Rome flight, should that be necessary.

Battery electric planes also bring with them a particular design change – apart from the obvious. Currently, planes land some 20% lighter than they take off, as they burn the fuel. With battery electric planes, they will land heavier than they take off, because the discharge of the battery means oxidation, meaning it gains mass. This will require very different landing gear.

Another facet of the EasyJet announcement is the desire to see these planes carrying passengers within 10 years. Given that the plane is concept only and doesn’t come from an existing family of similar planes, this may be ambitious. The 787 Dreamliner was announced in concept by Boeing in early 2003, finally receiving certification in late 2011. That’s nearly nine years for what is essentially a new version of an existing product, albeit with some significant changes such as the use of carbon fibre in the fuselage. But the 787 then had problems with its battery system for on-board electronics, leading to a temporary grounding and eventual regular flights from April 2013.

The 787 was not Boeing’s first attempt at a new aircraft following the 777 series. In the late 1990s it started development of the Sonic Cruiser, releasing a concept proposal in March 2001. With aviation business models changing rapidly at that time, Boeing abandoned this concept and instead moved to the 787, but this evolutionary process still consumed valuable design years. At least from one perspective, it could be argued that the 787 took over 13 years to go from concept to regular use.

The Wright Electric concept represents a revolutionary change in aircraft design and propulsion, so there is every chance that this may take longer than anticipated to get going. It will require extensive certification and testing by airlines, airports and the aviation authorities and may go through more than one design iteration, depending in part on the evolution of battery technology and the resultant changes in energy to weight and volume ratios. The story of the Mitsubishi Regional Jet is outlined here, and is a saga of changes and delays spanning 17 years. It is a conventional jet aircraft, but represents a first for Mitsubishi for a very long time.

A further ambitious aspect of this project is the notion that a start-up can take on the likes of Boeing and Airbus and find the necessary investors to back what is typically a multi-billion dollar investment in design, engineering, prototype development, manufacturing, testing and certification of a new aeroplane.

Behind all this is the pressing need for electrification across society, so this type of thinking and risk taking is essential. The question that remains though, is whether I will be able to ride in such a plane from London City Airport before 2030.