The Problem: Current world energy policies are failing to reduce atmospheric CO2Most of the world’s energy comes from burning fossil fuels, which adds CO2 to the atmosphere. The consensus of the world scientific community is that rising CO2 levels are warming the planet and causing climate change. Warming of over one degree Celsius is already evident and climate variability is already exceeding expectations for that level of warming. Governments and a majority of most populations have generally accepted the truth of the situation for some time but the unwillingness to accept the economic costs and political consequences of the measures proposed to reduce CO2 emissions have resulted in little change in the status quo. World CO2 levels are projected to continue growing at an increasing rate for as many decades as projections cover. Alternative energy from wind and solar are expected to still be less than 20% of world energy supply by 2050.
There is no political will by governments to impose costly change or from citizenry to demand or accept costly change, so the default behavior is effectively to defer substantive action to a time when either climate change has a tangible cost (like very high food prices or increased flooding from rising sea levels) or resource constraints (like scarce, expensive oil) force the issue. This is likely to take a decade or two by which time we will have passed the generally agreed goal of not exceeding 450ppm of atmospheric CO2 and 2-degree Celsius average temperature rise. A more realistic approach is neededWithin current political and economic constraints, there is an emerging awareness that the only way to cause substantive change is by developing market-competitive carbon-free energy alternatives that are cheaper than fossil fuel energy. Market competition driven by profitable investment opportunities, combined with government prodding should then result in a much higher growth of carbon free energy alternatives than the current subsidy constrained market. This perspective was well articulated by Bill Gates in his TED 2010 energy talk. He is part of an emerging "technology led energy policy" consensus that thinks that the current rate of spending subsidizing the deployment of high cost wind and solar technologies is not and will not make a significant dent in CO2 reduction. At the current relatively low level of government funding, a better strategy is to focus investment on R&D for development of several high-risk “miracle” (a Bill Gates term) alternatives that have a real potential to provide sufficient cheaper energy. This is not basic research R&D, which is probably already adequately funded, but power plant R&D that currently has no real funding.
|
Get the StratoSolar-PV Book:
Download the .pdf version of the ebook: Related Sites: Energy: The Oil Drum EIA IEA BP 101 sites Climate: RealClimate NASA The Breakthrough Skepticalscience Solar Energy: PV RESOURCES PVeducation.org FuturePV CSP Today RenewableEnergyFocus Greentechmedia GreenWorldInvestor Airships: Airships.net Carnet de Vol Synthetic Fuel: Doty Htyp California: Renewable Energy |
StratoSolar-PV: A cost competitive green energy solution
Solar PV based on crystalline silicon (c-Si) is a well-established technology with a forty-plus-year history of cost reduction along a consistent learning curve. Currently the business has grown to a manufacturing capacity of around 60GWp/year and around 30GWp/year of PV installations. These consume over 100,000 tonnes/year of pure polysilicon material. However, even with the recent dramatic reduction in PV-panel prices, electricity generated by PV systems at most locations costs more than electricity generated using fossil fuels. To get costs down to competitive levels will take around 5 doublings of current cumulative installed generation capacity. This is a capacity of 1TWp to 2TWp compared to 2012's 80GWp installed capacity.
The recent growth in PV installations worldwide has been due to the support of massive government subsidies, mostly in Europe. In a world of growing fiscal austerity, subsidies for PV electricity are reducing or disappearing. As a result, PV manufacturing capacity is declining as demand has slowed dramatically and the industry restructures. With slow growth in PV capacity, costs can not reduce significantly. As a result, PV will remain too expensive to survive without subsidy and hence will remain a very small provider of electricity. Were governments worldwide to maintain current levels of subsidy (with say Japan and China making up for Europe) PV may reach an installed capacity of 1TWp to 2TWp and become nominally cost competitive in between ten and twenty years.
StratoSolar PV reduces the average cost of current PV electricity by a factor of three. In energy, a factor of three is of enormous consequence. Most energy technologies have no technological road-map that will reduce their costs. Almost all technologies are increasing in cost. With fossil fuels, the fuel costs are volatile and can temporarily drop, but the trend is inevitably to higher costs. StratoSolar's tripling of average PV effectiveness makes today’s PV technology immediately viable and cost effective without subsidy.
In addition, StratoSolar PV technology avoids other costs associated with the intermittent nature of both wind and solar power. It does not need spinning backup generation. It does not need massive re-engineering of the electricity grid to transport electricity thousands of kilometers from where it is abundant to where it is needed. At the current small scale of PV deployment these issues are ignored. However, as wind power demonstrates, when an intermittent energy source approaches 20% of the grid the additional costs are high, matching or exceeding the cost of generation. Even if PV electricity generation were cost competitive today, the financial costs and risks associated with these two constraints (indeterminacy and geography) would severely limit the deployment of ground-based PV. In addition, StratoSolar PV is an affordable alternative for northern cloudy locations like Germany and Japan where PV is unlikely to ever be viable without subsidy.
With government policy support out of the picture and limited private capital likely to be committed, viable solutions have to need little capital up front, scale from initially small to large while growing organically and making a profit. A tough set of requirements that StratoSolar satisfies.
The rest of this web page explains StratoSolar technology and the reasons why it is both practical and economically competitive.
The recent growth in PV installations worldwide has been due to the support of massive government subsidies, mostly in Europe. In a world of growing fiscal austerity, subsidies for PV electricity are reducing or disappearing. As a result, PV manufacturing capacity is declining as demand has slowed dramatically and the industry restructures. With slow growth in PV capacity, costs can not reduce significantly. As a result, PV will remain too expensive to survive without subsidy and hence will remain a very small provider of electricity. Were governments worldwide to maintain current levels of subsidy (with say Japan and China making up for Europe) PV may reach an installed capacity of 1TWp to 2TWp and become nominally cost competitive in between ten and twenty years.
StratoSolar PV reduces the average cost of current PV electricity by a factor of three. In energy, a factor of three is of enormous consequence. Most energy technologies have no technological road-map that will reduce their costs. Almost all technologies are increasing in cost. With fossil fuels, the fuel costs are volatile and can temporarily drop, but the trend is inevitably to higher costs. StratoSolar's tripling of average PV effectiveness makes today’s PV technology immediately viable and cost effective without subsidy.
In addition, StratoSolar PV technology avoids other costs associated with the intermittent nature of both wind and solar power. It does not need spinning backup generation. It does not need massive re-engineering of the electricity grid to transport electricity thousands of kilometers from where it is abundant to where it is needed. At the current small scale of PV deployment these issues are ignored. However, as wind power demonstrates, when an intermittent energy source approaches 20% of the grid the additional costs are high, matching or exceeding the cost of generation. Even if PV electricity generation were cost competitive today, the financial costs and risks associated with these two constraints (indeterminacy and geography) would severely limit the deployment of ground-based PV. In addition, StratoSolar PV is an affordable alternative for northern cloudy locations like Germany and Japan where PV is unlikely to ever be viable without subsidy.
With government policy support out of the picture and limited private capital likely to be committed, viable solutions have to need little capital up front, scale from initially small to large while growing organically and making a profit. A tough set of requirements that StratoSolar satisfies.
The rest of this web page explains StratoSolar technology and the reasons why it is both practical and economically competitive.
Contents:
What the StratoSolar system does:
· Weather independent, predictable photovoltaic solar power (PV)
· Locations up to latitude 60 produce market competitive electricity
· Electricity in utility scale systems from 10 MW to 1 GW in modular increments
· Cost competitive electricity without subsidy
Key insights:
The idea exploits two environmental facts. Firstly, the stratosphere is a permanent inversion layer in the earth’s atmosphere. Inversion layers effectively isolate gas bodies. The calm weather free stratosphere is isolated from the turbulent troposphere below. There is no rain, hail, snow, or moisture in the stratosphere and wind force is much reduced and stable. Even severe weather events like thunderstorms and hurricanes have no effect at 20km altitude. This means that buoyant platforms suspended in the stratosphere can be permanently stationed there without needing to be winched down in bad weather. It also means that PV panels in the stratosphere don’t suffer water based weather effects and can be simpler and cheaper to manufacture.
Secondly, light from the sun at 20km altitude is both strong and constant from dawn to dusk. At 20km a platform is above over 90% of the atmosphere, so sunlight is not significantly scattered or absorbed and there are no clouds to interrupt power generation. This means that on average PV panels produce multiples of the energy they can generate on the ground, and just as important, the energy is highly predictable and not subject to interruption by clouds or storms.
· Weather independent, predictable photovoltaic solar power (PV)
· Locations up to latitude 60 produce market competitive electricity
· Electricity in utility scale systems from 10 MW to 1 GW in modular increments
· Cost competitive electricity without subsidy
Key insights:
The idea exploits two environmental facts. Firstly, the stratosphere is a permanent inversion layer in the earth’s atmosphere. Inversion layers effectively isolate gas bodies. The calm weather free stratosphere is isolated from the turbulent troposphere below. There is no rain, hail, snow, or moisture in the stratosphere and wind force is much reduced and stable. Even severe weather events like thunderstorms and hurricanes have no effect at 20km altitude. This means that buoyant platforms suspended in the stratosphere can be permanently stationed there without needing to be winched down in bad weather. It also means that PV panels in the stratosphere don’t suffer water based weather effects and can be simpler and cheaper to manufacture.
Secondly, light from the sun at 20km altitude is both strong and constant from dawn to dusk. At 20km a platform is above over 90% of the atmosphere, so sunlight is not significantly scattered or absorbed and there are no clouds to interrupt power generation. This means that on average PV panels produce multiples of the energy they can generate on the ground, and just as important, the energy is highly predictable and not subject to interruption by clouds or storms.
Where a StratoSolar PV system sits in the atmosphere.
The figure above helps illustrate where a StratoSolar PV system sits in the atmosphere and its relationship to various phenomena at mid latitudes. The PV platform floats in the stratosphere at the bottom of the ozone layer where the air temperature is about -55 degrees Celsius. The cumulonimbus cloud illustrates the relative scale and altitude of a severe super-cell thunderstorm, the most violent weather event that affects the tethers in the troposphere. Some super-cell thunderstorms can punch through the tropopause and top out close to 20km, and a small percentage spawn tornadoes. The jet stream, another strong wind phenomenon that affects the tethers is not shown explicitly, but when present it would be positioned just below the tropopause and be a few km thick. Original artwork by Windows to the Universe staff (Randy Russell)
This image is derivative from content on Windows to the Universe® (http://windows2universe.org) © 2010, National Earth Science Teachers Association. This work is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.
This image is derivative from content on Windows to the Universe® (http://windows2universe.org) © 2010, National Earth Science Teachers Association. This work is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.
Why it generates electricity at a low cost:
For solar-power plants, almost the complete operating cost is the loan payment. The StratoSolar PV system generates electricity at a low market competitive cost mostly because the solar PV array (which dominates PV cost) has a similar capital cost to current ground based PV power plants but generates electricity with a higher utilization, resulting in a low cost of electricity that is market competitive without subsidy at current PV cell cost. The reasons for this are:
While there is currently no economically viable electricity storage technology, the uninterrupted highly predictable PV power generation does not need spinning backup generation, and the power plants can be where they are needed rather than in deserts thousands of kilometers away. These two benefits offer significant additional cost advantages to StratoSolar, but the benefits are far larger than economic. As PV is attempting to be accepted as a viable utility scale electricity option these two problems are the major barriers to PV's acceptance as a viable technology for large-scale electricity generation.
As part of an integrated grid, PV in the stratosphere could provide the bulk of electricity requirements at the lowest market cost, with nuclear, hydro and natural gas covering the rest. Natural gas would be the natural choice to cover the shortfall during darkness. Given that fuel costs dominate the cost of electricity generated from natural gas, the cost penalty from the reduced natural gas power plant utilization would be small. Such an integrated electricity grid could have a carbon footprint with a fraction of the current US electricity generation carbon footprint and produce electricity with the lowest cost and the highest price stability.
This is a commercially competitive alternative energy solution. By not covering huge land areas, it saves on an expensive, highly regulated, and uncertain resource that tends to delay construction and limit financing options. It also allows great flexibility in location. The competitive and highly profitable economics should lead to a business that is market financed and does not need government support or subsidy once demonstrated. It is a bonus that this energy is carbon-free, and solves energy security issues.
For solar-power plants, almost the complete operating cost is the loan payment. The StratoSolar PV system generates electricity at a low market competitive cost mostly because the solar PV array (which dominates PV cost) has a similar capital cost to current ground based PV power plants but generates electricity with a higher utilization, resulting in a low cost of electricity that is market competitive without subsidy at current PV cell cost. The reasons for this are:
- The PV panels are exposed to 1.5 to 4.0X the solar energy (in kWh) of ground-based PV panels (the amount depends mostly on Latitude)
- This means each square meter of PV panel gathers 1.5 to 4.0X the energy of ground-based PV panels
- The PV array uses no land. There is only a small land and site development cost.
- The PV array buoyant support structure uses very little material due to light structural loads.
- The PV array buoyant support structure is factory produced. Labor costs for construction and assembly are greatly reduced.
- All construction materials are standard aluminum and plastics, off the shelf, and low cost.
- The PV panels are 20% more efficient due to a -30C to -40C operating temperature at the -60C ambient at 20km altitude.
- The PV panels are lower cost than ground-based PV panels due to reduced materiel use; 4kg compared to 20kg.
While there is currently no economically viable electricity storage technology, the uninterrupted highly predictable PV power generation does not need spinning backup generation, and the power plants can be where they are needed rather than in deserts thousands of kilometers away. These two benefits offer significant additional cost advantages to StratoSolar, but the benefits are far larger than economic. As PV is attempting to be accepted as a viable utility scale electricity option these two problems are the major barriers to PV's acceptance as a viable technology for large-scale electricity generation.
As part of an integrated grid, PV in the stratosphere could provide the bulk of electricity requirements at the lowest market cost, with nuclear, hydro and natural gas covering the rest. Natural gas would be the natural choice to cover the shortfall during darkness. Given that fuel costs dominate the cost of electricity generated from natural gas, the cost penalty from the reduced natural gas power plant utilization would be small. Such an integrated electricity grid could have a carbon footprint with a fraction of the current US electricity generation carbon footprint and produce electricity with the lowest cost and the highest price stability.
This is a commercially competitive alternative energy solution. By not covering huge land areas, it saves on an expensive, highly regulated, and uncertain resource that tends to delay construction and limit financing options. It also allows great flexibility in location. The competitive and highly profitable economics should lead to a business that is market financed and does not need government support or subsidy once demonstrated. It is a bonus that this energy is carbon-free, and solves energy security issues.
A description of the system:
An array of lightweight PV panels is attached to the top surface of a rigid buoyant platform which is permanently tethered in the stratosphere at an altitude of 20 km. The PV array gathers sunlight, converts it to electricity and transmits the electricity down tether/high voltage (HV) cables to the ground where it connects to the electricity distribution grid.
An array of lightweight PV panels is attached to the top surface of a rigid buoyant platform which is permanently tethered in the stratosphere at an altitude of 20 km. The PV array gathers sunlight, converts it to electricity and transmits the electricity down tether/high voltage (HV) cables to the ground where it connects to the electricity distribution grid.
The figure above shows a range of systems from initial small scale 8MWp to large 10GWp utility scale. The novel element of a StratoSolar power plant is a buoyant tethered platform supporting an array of PV panels floating in the stratosphere. The strong and light tethers supports HV power cables that transfer electric power to the ground. Excess buoyancy in the floating platform pre-tensions the tethers and allows the platform to resist wind forces.
A rigid truss structure supports the PV array. Buoyancy is from gasbags within the truss framework. Models for the PV array power output are subject to simulation to a high degree of accuracy, with high confidence in the results. While the buoyant structure is novel, there is no new science or materials. The wind and buoyancy forces are well understood from an engineering perspective. There are detailed meteorological models and historical data to provide an accurate statistical profile of the wind and buoyancy forces on the structure and tether. The combination of accurate structural analysis and reliable meteorological data provide high confidence in structural integrity under worst case weather conditions. Accurate climate models for sunlight and how it varies with location and altitude, daily and seasonally, provide an equally high confidence level for the power output.
The large-scale systems are a collection of mechanically connected individual modular small-scale systems. The tethers will have less visual impact than shown, though lights and other visibility enhancements will be required by airspace regulators. The benefits of connecting multiple smaller systems to make a larger system are reduced aerodynamic drag on the platform and the reduced impact of fewer large systems on regulated airspace. The reduced aerodynamic drag ensures that the structure can withstand the highest wind forces with a large safety margin and is safe to deploy on a permanent basis. It also facilitates modular maintenance and repair, technology upgrades, and incremental overall system expansion. Individual small arrays can be winched down in a few hours when weather permits and can use adjacent tethers as guides to ensure safe control.
Operationally there should be no need for people at 20km. There is no need for large “hanger” structures on the ground, either for construction or maintenance. Maintenance on the ground only occurs during good weather and at night to avoid disruption in power output.
Another benefit of the modular approach is the system can grow and be financed incrementally, reducing the risk capital required to develop and demonstrate the system viability.
What are the benefits of 20km altitude?
· More sunlight results in lower cost per kWh. Under $0.08/kWh without subsidy with today's PV costs.
· Locations as far north as latitude 60 are practical and have the biggest relative cost advantage.
· Highly predictable power compared to ground based PV. Weather never affects power.
· No shading effects simplify the electrical design.
· The -50 degree Celsius cold environment at 20km altitude enhances c-Si PV efficiency
· At 20km altitude the PV panels don’t have to handle rain, moisture, hail, snow or strong wind forces. This reduces their packaging cost
· The worst-case wind force on the PV array and the tether is low enough to allow for permanent tethering.
A rigid truss structure supports the PV array. Buoyancy is from gasbags within the truss framework. Models for the PV array power output are subject to simulation to a high degree of accuracy, with high confidence in the results. While the buoyant structure is novel, there is no new science or materials. The wind and buoyancy forces are well understood from an engineering perspective. There are detailed meteorological models and historical data to provide an accurate statistical profile of the wind and buoyancy forces on the structure and tether. The combination of accurate structural analysis and reliable meteorological data provide high confidence in structural integrity under worst case weather conditions. Accurate climate models for sunlight and how it varies with location and altitude, daily and seasonally, provide an equally high confidence level for the power output.
The large-scale systems are a collection of mechanically connected individual modular small-scale systems. The tethers will have less visual impact than shown, though lights and other visibility enhancements will be required by airspace regulators. The benefits of connecting multiple smaller systems to make a larger system are reduced aerodynamic drag on the platform and the reduced impact of fewer large systems on regulated airspace. The reduced aerodynamic drag ensures that the structure can withstand the highest wind forces with a large safety margin and is safe to deploy on a permanent basis. It also facilitates modular maintenance and repair, technology upgrades, and incremental overall system expansion. Individual small arrays can be winched down in a few hours when weather permits and can use adjacent tethers as guides to ensure safe control.
Operationally there should be no need for people at 20km. There is no need for large “hanger” structures on the ground, either for construction or maintenance. Maintenance on the ground only occurs during good weather and at night to avoid disruption in power output.
Another benefit of the modular approach is the system can grow and be financed incrementally, reducing the risk capital required to develop and demonstrate the system viability.
What are the benefits of 20km altitude?
· More sunlight results in lower cost per kWh. Under $0.08/kWh without subsidy with today's PV costs.
· Locations as far north as latitude 60 are practical and have the biggest relative cost advantage.
· Highly predictable power compared to ground based PV. Weather never affects power.
· No shading effects simplify the electrical design.
· The -50 degree Celsius cold environment at 20km altitude enhances c-Si PV efficiency
· At 20km altitude the PV panels don’t have to handle rain, moisture, hail, snow or strong wind forces. This reduces their packaging cost
· The worst-case wind force on the PV array and the tether is low enough to allow for permanent tethering.
Ground PV electricity cost compared to StratoSolar PV electricity cost:
As the graph below shows, Crystalline Silicon (c-Si) PV panels have reduced in price considerably as cumulative production volume has increased. This shows a learning rate of approximately 20% which means for every doubling of cumulative volume, prices reduce by 20%. Current prices are below this projection, which would imply a faster learning rate, but current prices are depressed by overcapacity. The graph shows that reliable predictions can be made over decades, but prices are much less predictable over yearly time frames. The factors that drive the 20% learning rate are still operating and long term price predictions based on this rate have a high confidence level.
The graph that follows takes a while to absorb, but is worth taking the time to do so. It presents succinctly the true state of current and future PV electricity costs without the distortion of subsidies. It also clearly shows the benefit that StratoSolar brings to the table. For a more detailed analysis see this document. Link
The figure above shows a projection of the reduction in capital costs in $/Wp and resulting levelized cost of electricity(LCOE) in $/kWh with cumulative GWp installed capacity along the forty year learning curve shown in the previous graph. LCOE is shown for several sunlight intensities that correspond to different geographic locations. The oval on the graph represents the current 2013 approximate capital cost ($/Wp) and the range of the resulting current LCOE ($/kWh) for geographies from Northern Europe to deserts and StratoSolar for comparison.
A common way to refer to the variability in solar power input is to convert it to a utilization or capacity factor. This is useful when comparing different power plant technologies. Using this metric the lowest utilization (Northern Europe) is about 0.10, average for the US is about 0.15, desert is 0.20 to 0.25, and StratoSolar is about 0.35 to 0.40.
The horizontal band centered at $0.10/kWh represents electricity that is competitive in the marketplace without subsidy. The width of the band represents the variability in electricity costs for different markets. Europe and California tend to be at the high cost end of the band. This is an extremely significant barrier to cross. Above the band is the world of government subsidy politics and low volume. Below the band is the world of market economics and high volume.
The vertical distance from this $0.10/kWh band to the different utilization lines represents the approximate amount of subsidy needed. As can be seen, for the current cumulative GWp the subsidy is very large. For example the current subsidy for Germany is about $0.20/kWh. This subsidy shows up in different ways. In Europe the mechanism is mostly feed-in tariffs that result in higher costs of electricity to consumers. In the US, the mechanism is mostly tax credits and accelerated depreciation ultimately paid by taxpayers.
The synthetic liquid fuel and synthetic gas bands show the ranges where the reduction in LCOE enables the manufacture of synthetic fuels using conventional manufacturing technologies that have already been proven at scale.
This chart clearly illustrates several points:
StratoSolar represents a low cost way to leverage the historical investment in PV technology into commercial viability at locations where PV is unlikely ever to be commercially viable and with today’s PV capital cost in $/Wp, which will apply even if the historical rate of $/Wp improvement slows.
A common way to refer to the variability in solar power input is to convert it to a utilization or capacity factor. This is useful when comparing different power plant technologies. Using this metric the lowest utilization (Northern Europe) is about 0.10, average for the US is about 0.15, desert is 0.20 to 0.25, and StratoSolar is about 0.35 to 0.40.
The horizontal band centered at $0.10/kWh represents electricity that is competitive in the marketplace without subsidy. The width of the band represents the variability in electricity costs for different markets. Europe and California tend to be at the high cost end of the band. This is an extremely significant barrier to cross. Above the band is the world of government subsidy politics and low volume. Below the band is the world of market economics and high volume.
The vertical distance from this $0.10/kWh band to the different utilization lines represents the approximate amount of subsidy needed. As can be seen, for the current cumulative GWp the subsidy is very large. For example the current subsidy for Germany is about $0.20/kWh. This subsidy shows up in different ways. In Europe the mechanism is mostly feed-in tariffs that result in higher costs of electricity to consumers. In the US, the mechanism is mostly tax credits and accelerated depreciation ultimately paid by taxpayers.
The synthetic liquid fuel and synthetic gas bands show the ranges where the reduction in LCOE enables the manufacture of synthetic fuels using conventional manufacturing technologies that have already been proven at scale.
This chart clearly illustrates several points:
- The same plant with the same capital cost produces electricity with highly variable cost depending on location. E.g. at the 2012, $2.50/Wp capital cost, Northern Europe generates electricity for about $0.30/kWh, and StratoSolar generates electricity for $0.08/kWh. StratoSolar has the best location (which can be over northern Europe) and lowest cost.
- The $2.50/Wp capital cost is approximately the 2012 cost. At historical rates of improvement, the $1.50/W cost may occur by 2020 at best. Even in the best desert locations, the resulting ground based PV electricity will still cost $0.12/kWh which will not be competitive without subsidy in 2020 except in high priced markets.
- StratoSolar will produce electricity without subsidy with current $2.50/Wp PV technology and will benefit equally from the PV $/Wp cost improvement path, producing increasingly competitive lower cost electricity.
- StratoSolar can do this for northern climes for which PV is not an economically viable option.
StratoSolar represents a low cost way to leverage the historical investment in PV technology into commercial viability at locations where PV is unlikely ever to be commercially viable and with today’s PV capital cost in $/Wp, which will apply even if the historical rate of $/Wp improvement slows.
