Exawatt, Open Climate Fix, The University of Sheffield              OSE0040

Written evidence submitted by Exawatt, Open Climate Fix, and The University of Sheffield

The following response is authored by a group of experts in Global and UK solar PV from:

●        The University of Sheffield (Department of Physics - Sheffield Solar) provides benchmark solar PV generation monitoring for National Grid ESO [1] and works with the Analysis Directorate at BEIS on tracking solar capacity [2]. Further contributions come from the Biosciences department with research on agriculture / solar land use (agrivoltaics) [3].

●        Open Climate Fix Ltd - using data and AI to provide solar forecasts for National Grid ESO and have worked on the UK Modernising Energy Data, Open Energy Data initiative [8].

●        Exawatt Ltd - provides supply chain analysis and market intelligence to the largest global solar PV manufacturing companies [5].

Our analysis of historical installation data in the UK is both contrary to, and in support of, the premise of the call for evidence. The UK has, on average, between 2010 and 2020, installed 2.1 GW above that expected for nations with similar economies, level of technological development and land available for PV installation [4] & (figure 1).  However, since 2017, and the closure of the Renewable Obligation Certificate, the UK has been installing around 1 GW per year less than expected. More recently, in the last 12 months, small scale installations (see answer to question 5) show a strong rebound; revitalised by increasing prices of electricity. However, the installation of larger ground mount systems has not yet rebounded in the same way, despite a strong pipeline.

While the current Global market for solar PV is strong, and is not limited by solar cell technology performance or price, in the UK it is already showing signs of slowing and it is likely to stall prematurely (within a few years) due to lack of permission being granted for grid connection of large solar PV systems and the possibility of recently introduced windfall taxes making large scale solar less attractive as an investment. The grid connection barrier is partly due to a lack of physical infrastructure (ie. copper wires, transformers and control systems) but it is compounded by a lack of operational visibility of changing electricity usage profiles (due to solar and electric vehicles). Unless both the physical and digital energy infrastructure is developed quickly the UK will not achieve its target of 70 GW capacity and it is likely that small scale systems (that currently don’t need grid connection permission) may need to be regulated to prevent network instability and/or overload.  These barriers are summarised by type of solar system in table 1.

Suggested actions:

●        DNOs should be incentivised through targets set by Ofgem to provide defined levels of new solar generation and energy storage grid connections;

●        DNOs should work with local planning authorities to proactively plan for these new grid connections on the most appropriate local land and buildings. Digital decision support tools are needed to transform what is currently a reactive planning process to a proactive one;

●        the governance of electricity metering data should be simplified to allow wider sharing and use for electricity system planning and operation. Minimum rules for anonymisation of property level electricity metering (to comply with GDPR) should be clarified by Ofgem;

●        DNOs should continue to extend the data quality of the Embedded Capacity Register. The scope of this register should be confirmed for all systems > 50 kW.

●        the renewable electricity windfall tax level (set by the treasury) should be reviewed quarterly to ensure that investment in large scale solar PV continues and that the industry is not inadvertently squeezed between the cap on electricity sales and rising costs of installation;

●        property level Energy Performance Certificates should include accurate and comprehensive information about solar PV systems (already implemented in Scotland) along with other significant electricity equipment such as heat pumps, electric vehicle chargers & battery storage.

Table 1 - Summary solar PV deployment opportunities and barriers by system size and type. Future potential capacity estimates are based on a combination of historical maximum installation rates (see figure 9), total available roof space and the ease/efficiency of installation (see figure 8). For a summary of historical, current and future capacity see figures 10 and 11.

1.      What role can developments in solar panel technology play in the UK’s transition to net zero?

Developments in solar PV technology, while important for policy makers to understand, should not be a key concern of the government in terms of driving further deployment. Solar PV technology is now mature and has already undergone substantial cost reduction as the technology has transitioned to a mass produced commodity over the last 20 years. Solar PV is directly investable in the current electricity market (although the recently implemented windfall tax does create some risk) and no longer needs subsidy to incentivise deployment. Current and future development in solar panel technology will further reduce costs and improve efficiency and this will lead to further reductions in the cost of electricity from solar PV, making it even more investable in the future.

In terms of solar panel technology development specifically, over the past decades, the solar PV manufacturing industry has made enormous progress in reducing the cost and increasing the efficiency of mass produced solar panels. These advances have played an important role in reducing the levelized cost of electricity (LCOE: ie. cost per kWh) for solar PV installations around the world and thereby assisting countries in decarbonising electricity generation. In the coming years, improvements in solar cell and module (panel) technology will undoubtedly continue to drive higher efficiency and therefore enable lower system costs. Note that higher efficiency can assist not only in driving down module costs when considered as cost per kW of panels produced, (the key concern of the PV manufacturing industry) but also in lowering numerous balance of system (BOS) costs such as labour costs for installation, land costs, mounting costs, wiring costs. The industry is poised to transition over the next 2-4 years from the incumbent technology (PERC cells, based on p-type monocrystalline wafers) to higher-efficiency alternatives (TOPCon and HJT cells, predominantly based on n-type monocrystalline wafers). Modules based on these technologies currently offer average efficiency of ~21.5% - 22.0% vs ~20.8% for PERC, and this lead will grow significantly over the coming years ([5] Exawatt Q4 2022 Module Tracker; not publicly available). In addition to higher efficiencies, these technologies also offer other advantages such as lower degradation and improved temperature coefficients - both of which ultimately help to lower LCOE. Note that these higher-efficiency technologies are, broadly, evolutions on the incumbent technology rather than revolutions in the industry, and that this trend - of incremental improvements driven by the major manufacturers - is unlikely to change in the near future. In the longer term (2030 and beyond) the industry can be expected to continue its transition to higher efficiencies by adopting still more advanced technologies such as tandem cells (multiple cells stacked on top of each other). These cells are also likely to be built, initially at least, on the existing technology platforms used by major manufacturers - i.e. they will largely be treated as ways to enhance the offerings of the existing silicon-based industry rather than as a disruptor that will replace the existing industry.  Adoption of these newer technologies is dependent on them achieving cost-reductions such that they can enable a lower LCOE - ultimately utility-scale module buyers are almost entirely driven by achieving the lowest LCOE possible for their system (meanwhile residential and commercial/industrial customers will typically focus on a combination of LCOE, payback time and initial cost). It is worth noting that over the past decade the PV manufacturing industry has transitioned from a “cost-reduction era” to a “performance-improvement era” (figure 2). Until around 2017, avenues for absolute cost reductions – in a $/m2 sense – were plentiful, while module efficiency improved slowly but surely. More recently, such cost reductions have become harder to realise, in part due to the enormous successes already achieved by the industry in optimising processes, reducing materials usage and reducing input materials costs. While some opportunities for minor $/m2 cost reductions still exist with conventional manufacturing techniques, these are comparatively limited, and in the near term will simply mark the regaining of ground lost due to recent price rises such as those for polysilicon (the material feedstock for silicon wafer manufacturing) in 2021/2022.

In comparison, since 2017 module efficiencies have increased rapidly and module efficiency gains are now the key to driving lower system costs and lower LCOE. This is not to say that cost has become – or will become – unimportant, simply that the PV industry is now in an era where its evolution is principally driven by improvements in performance at acceptable cost, rather than reductions in absolute cost at acceptable performance (figure 2).

2.      To what extent is the contribution of onshore solar technologies to the UK’s renewable energy mix limited by storage and distribution capacity?

Solar power is clearly only generated during daylight hours with the national output from solar with peak output varying by an order of magnitude between the sunniest and cloudiest days. I.e electricity is generated at the same time of day and on mostly the same days across the grid. This means that in our current market structure, where the wholesale price of electricity is set by the most expensive generator needed to fulfil demand, as solar increases its share of the generation mix, the value of solar electricity to the grid decreases as the power is concentrated in a small number of hours (figure 3).  This effect is colloquially termed price cannibalisation, and is more acute in solar energy as the generation is focussed in the smaller number of daylight hours than other generation types. It can be is measured by a “Value Factor” [6]. Figure 3 shows that as the share of solar in an electricity market increases, the value of the solar energy reduces by about 30% per 10% solar market share. 

This price cannibalisation effect will ultimately lead to a natural saturation level for solar in the electricity market. However there are ways of mitigating the effect and increasing the value and contribution of solar.  One approach is to deploy more fast-start gas fired generation, but this would conflict with the aim of decarbonising the grid.  Two other approaches (beneficial to both solar and decarbonisation) are:

1. Increased use of batteries to “smooth” the electricity generated by solar installation - moving the electricity from the time of generation to the time of use - usually within-day, but also some day to day time shifting. 
2. Increased interconnections with solar or other generation sources produced in different electricity grids, where the factors driving the time of generation are different to those in the UK.  (See Question 10) For example, connections with Europe allow for solar power from Italy to be used on the UK grid, smoothing out the diurnal cycle of the UK; or using hydro electric power from Switzerland to match customer load shapes.

To address (1) with reference to the question: energy storage is already at a point of technological development where the business case for unsubsidised utility scale battery projects, often co-located with solar generation, is strong.  Utility scale batteries co-located with solar generation is beneficial to the development of solar sites for a number of reasons - the ability to redistribute the energy generated in the day to later in the evening or the morning after helps increase the Value Factor, and Lithium-ion batteries, the most common technology today, are suited to such short-duration storage.  These benefits are easily accrued by developers via the electricity market.  Further revenues can be accrued by developers through grid balancing services, which also enhance the ability of the grid to accept increased renewable penetration.

From a grid & market perspective, the current rules for gaining grid connections do not reflect the almost completely beneficial nature of co-located solar and storage.  One of the key constraints for connections is the line capacity - a measure of how much power can be passed through the line at the limit.  By co-locating the assets, the economically rational developer will store the solar electricity generated at peak generation times, so “shaving” the peak, and discharge the stored energy at times when solar is not generating.  This results in a reduction in the peak line capacity required when adding storage.  Unfortunately, current rules require that the peak capacities of each PV and battery asset are summed when making the connection, penalising such development and slowing their rollout.  Other network connection issues are addressed in the answer to question (4 & 5) and current research is evaluating the links between the wholesale electricity market and the locations of solar and storage across the UK [7, 8].

The strength of the business case for utility scale energy storage can be observed in the demand for utility-scale batteries, where battery supply is already committed for 12-18 months ahead, and the limitation for new projects is the availability of battery cells.  This is compounded by broader supply chain issues.

Energy storage technologies for commercial or residential have not yet become widespread and the business case, while being developed by a number of innovative start-up companies, is not yet strong.  The key issue for energy storage in the home or for smaller businesses is that individually they cannot take advantage of the wholesale market prices and grid service provision as utility scale storage can.  The price signals available for household or even business use cases are not strong enough to justify the investment.  This is an intentional construct of the electricity market as originally created in 1989, to protect households and businesses from the wholesale market fluctuations. 

Market structures are extremely complex in having to balance multiple societal, economic and environmental demands.  However, if the country is to take maximum advantage of the flexible tools such as distributed energy storage, having the ability for individuals to benefit financially is a challenge that should be looked at in much more detail and it should be possible to protect the most vulnerable while allowing incentives for the middle to better off.

The other barrier to residential small scale commercial energy storage is the communication with and control of thousands of batteries located in a wide variety of locations.  The technological solutions for this exist, but integrating them in real world situations is still a fast evolving space. 

3.      How significant are current technological developments in solar energy storage and distribution for the potential contribution of onshore solar to the UK’s renewable energy mix?

Storage capacity on the electricity system is currently negligible compared with electricity demand and generation capacity. While we believe that the installation of utility scale battery storage currently makes sense as an investment (ie. is commercially viable) we believe that there is approximately an 18 month lead time for battery cells and so the current installation rate of utility battery storage is very limited. Once the supply chain for battery cells matures deployment of utility storage will proceed quickly in locations where the value is highest. Battery storage will improve the value of solar PV, but it will not resolve the need for alternative sources of electricity generation in the winter months. Interseasonal electricity storage is still an emerging technology.

In terms of distribution capacity there are two levels to consider. The first is the capacity of the transmission network and the second is the capacity of the distribution network.

Constraints on the transmission network will ultimately limit solar PV deployment - and specific parts of the network will need to be reinforced (as is currently in progress to transfer North Sea offshore wind into the transmission grid). Reinforcement plans are the ongoing responsibility of National Grid through the Network Operations Assessment. Current issues on the transmission network that are a direct result of solar PV are reverse flows occurring at several grid supply points (such as at Grid Supply Points in Cornwall and Norfolk), however these flows are currently managed effectively by National Grid ESO. While National Grid ESO don’t directly control access of solar PV to grid connections the planning actions that they take impact on capacity limits that are controlled by the distribution network operators (DNOs).

Constraints on the distribution network are currently more significant, but much harder for organisations external to the distribution operators to assess. For large scale ground mount solar DNOs engage with the sector actively and try and recommend locations for future development. For building integrated (or roof mounted) solar the process is much more reactive and investors have less visibility of the likelihood of approval of an application for a grid connection.

DNO and ESO decision making is continually evolving and improving along with digital support tools for local and national energy planning and the network is being continually reinforced and developed to take greater capacity. To speed this process there are two recommendations: First is that while DNOs are regulated to respond to applications for grid connections of solar pv, they do not have targets for the new connection capacity they must accept.  It is worth the regulator considering such targets as a way to better support the sector.

The second recommendation relates to the digital decision support tools - that need to reach maturity to prevent conservative decision making. The issue here is to be able to make accurate local estimates of electricity demand as well as being able to monitor embedded generation, predominantly from solar.

The central issue for these digital tools is that GDPR and commercial sensitivities restrict access to raw electricity metering data on both the demand and generation side and this in turn limits the ability of DNOs and the ESO to model electricity demand at high enough spatial resolution to create the network power flow models needed for “proper” data driven decision making for network planning. In some cases workarounds to full data access simply take time to develop and validate, while in other cases low data quality and in some cases complete lack of data result in more substantial gaps in visibility. As these tools develop the network will be able to further evolve (through a combination of reinforcement, storage and new markets) and adapt to support increasing levels of solar generation to the those proposed in net zero models (35 - 70 GWp) However without the development of these planning and operating tools deployment will naturally stall as DNOs fail to be able to authorise new connections or a lack of knowledge of the time taken to authorise makes investors look elsewhere.

Previous Governments consultations (such as [9] the Public Interest Advisory Group on Smart Meter Data) on access to smart meter data have repeatedly recommended that data is made open wherever possible, and Innovate UK programmes have developed case studies of how this is possible (Modernising Energy Data Access Phase 2 [10]), but unfortunately the central issue remains that strong regulation by Ofgem and conservative decision making by DNOs, ESO and other actors means data is largely not open and work in this space extremely challenging. While it is clear that GDPR cannot be removed, there are opportunities for the development of legal precedence on less conservative approaches to smart meter data anonymisation, and this would mean that much more useful information is retained in electricity metering data. 

4.      What are the current barriers (regulatory, technological or otherwise) to expanding the number of small and large-scale solar installations in the UK?

and

5.      What needs to be done to facilitate solar farm access to grid connection, to enable wider distributed energy generation from solar installations?

In Jan 2022 Solar Media and Solar UK reported a pipeline of more than 37 GWp of planned large scale solar PV systems, compared with 13.8 GWp of currently installed capacity of which 8.7 GW are ground mounted.  By October 2022, less than 0.5 GW of this pipeline has been installed (Sources: MCS, Solar Media, REPD and FIT). This suggests that there are currently significant barriers to large scale system deployment.

Grid integration provides a major limiting factor to continued large scale solar PV deployment:

●        Belated network connection delays are becoming widespread in the industry.  Projects initially granted connections in 2023-24 are having their connection dates pushed out as far as 2028-32 by DNOs.  In the short term, these multi-year delays delay the expansion of solar energy.  In the medium term, they increase costs and create uncertainty for developers, damaging the appetite for future investment. 

●        Grid connections from DNOs are currently difficult to procure and are not available at many locations across the country.  Information from DNOs on connection locations has improved in recent years, but if the distribution network does not evolve to support more solar PV then distribution operators will reject applications for network connections and investors will look elsewhere.

While grid integration barriers are slowing the progress of solar development, they could become critical barriers in a short space of time. To take some examples from overseas - the worst case outcome is Vietnam, where from 2022 they have stopped all solar developments due to grid instability. The UK needs to ensure that energy security and safety constraints don’t mean that such barriers need to be put into place in the future.  Other regions are approaching these hard barriers, and are putting measures in place to ensure these issues do not prevent future solar developments: Germany currently limits export to the grid from building integrated systems; from 2022 the grid operator in South Australia has the ability to centrally control the export level from all solar systems to allow them to flatten the peak by turning off solar during hours of highest generation.

Recommendations for reducing these impacts:

●        Reinforcing the grid at both the high and low voltage levels, to ensure the grid can cope with the new solar flows. This includes both new hardware and control systems in transformers.  Regulation could assist this by requiring DNOs and the ESO to have minimum levels of solar connected over the coming licence period as part of the licence conditions

●        Deliver improved situational awareness of the power flows through the grid to enable proactive, forward looking modelling by network operators, commercial companies and researchers into the optimal network configuration to allow the maximum renewable penetration while minimising network reinforcement costs and ensuring grid stability.  This can be done by collecting more data: ensuring registration of all solar assets and requiring generation data for small and medium systems to be reported centrally and shared appropriately.

●        Requiring control systems to be installed in small and medium scale solar so it can be curtailed (turned off) in peak hours of generation.

The investor appetite and hence rate of deployment of solar PV is driven by the return on investment. Project costs include the price of the technology components, installation costs and grid connection fees. Revenues are dependent on the price of electricity, which for building large scale ground mount systems is in terms of the value of the electricity sold to the market and for building integrated commercial systems is in terms of the value of electricity purchases avoided. It has been a healthy market post subsidy and large scale solar has enjoyed a strong business case without any government intervention up to recently.

In recent months, there have been several external and internal factors which are damaging the business case for large-scale ground mount solar:

●        Interest rates have increased greatly over 2022, increasing financing costs

●        Costs of components have risen with post-Covid inflation

●        Energy prices have increased, but with the windfall tax of 45% to be applied to renewable generators from 2023, developers are not able to benefit from the full upside available

While the first two factors the government has no control over, the level of windfall tax should be closely monitored to ensure that it is not providing an unwanted barrier to further solar development.

The residential solar market was healthy prior to the Feed in Tariff ending in 2019, after which the industry contracted sharply.  Post 2019 the market was subdued before taking off in 2022 with the spike in energy prices.  It is being held back by two significant factors: a shortage of skills in the industry and a shortage of capital.  The first of these two points is well documented across many industries, but with the solar industry being medium skilled, it would seem to be a good candidate for government training incentives.

The capital barrier arises as while the financial case for an individual homeowner to invest in solar panels is strong, many homeowners do not have the available free capital to be able to make the investment.  In Germany, the state-owned development bank, KfW, makes low interest loans available to householders looking to invest in solar power, facilitating increased uptake.  If the UK government would support similar schemes, it has the potential to significantly increase residential solar. Across all solar markets there is a background risk that unexpected or frequent changes of government policy will create uncertainty in the investment community.  Where possible, cross party support for policy positions should be garnered, which will greatly increase personal and institutional investor appetite for UK solar. 

Electricity metering data

There are also some specific challenges around network planning largely concerning access to electricity metering data at different levels in the electricity network.

In the residential sector GDPR rightly limits data access. But the real problem is that residential metering data is not even routinely being recorded anywhere other than on the meter itself. We believe that while the smart meter data transfer operator has rights to read the meter they don’t have the rights to store the data. The DNO can store the data anonymously but they can’t routinely access it. These rules seem to be overly conservative and complex and need reviewing to ensure that residential metering data is available to support grid planning operations.  There are currently also currently no central government records of significant electricity technology (ie. solar PV, electric vehicle charger or ownership or heat pump) installed at a property.  The statistical likelihood of a Solar PV installation can be estimated from a combination of several different datasets (MCS and FIT largely). Electric vehicle ownership is known by DVLA but not available for use by electricity operators and heat pump installation is only known if installed and registered under the renewable heat incentive or via an MCS registered installer. The recording of significant residential electricity assets is a mess and needs to be consolidated. The Energy Performance Certificate (EPC) scheme is one such approach but would currently not be fit for purpose due to its requirement only on sale of property and the low quality of the data collected by surveyors during inspection.  By enhancing the EPC mandate and scope or using a new scheme to record installations by electricians, this challenge could be addressed.

Commercial properties are not subject to GDPR but electricity metering data is still confidential and is not always routinely accessible by DNOs for network planning purposes. This data is stored under the DCC scheme but cannot be shared easily. 

Network metering at substations is the alternative option in terms of network planning. However, without knowing exactly which properties are connected to which substations, and also knowing what significant electricity technologies are connected, this approach is limited and requires substantial machine learning to infer likely technology impacts on the electricity network. Work in this space is progressing under the Energy Entrepreneurs scheme funded by BEIS.

6.      Are government support schemes sufficient to encourage small-scale solar technology deployment by consumers? What role does the pricing of energy under these schemes play in the uptake of solar technology by domestic and commercial properties?

Current Government support for residential and small scale commercial solar PV is minimal and is limited to some grant aid. The Feed in Tariff was the main support scheme for these systems but closed to new entrants in April 2019. However, following several years (see figures 5 and 6) (from 2018 to 2021) of very low levels (<500 MWp / year) of solar PV deployment the residential and small scale commercial market (from Q4 2021 to now) has rebounded strongly due to the rising retail price of electricity. Figure 6 shows the growth of both residential (<4 kWp) and small scale commercial ( >4 kWp but < 50 kWp) markets alongside the yearly average price of electricity on fixed tariff (source: OFGEM). While the precise relationship between installation rate and electricity price will depend on distribution of tariffs amongst the potential market participants, the present analysis shows clear and strong correlation between the uplift in electricity price and the increase in the deployment of residential and small scale commercial PV systems.

Figure 5 shows conclusively that the increase in electricity price has stimulated the market for small scale PV. Given that the installation of residential systems does not require authorisation by the Distribution Network Operator there is actually a risk, in the event that the price of electricity does not fall, that, over the coming years, the market for small scale PV grows uncontrollably and the Government may need to intervene to control it.

An alternative scenario is that due to lack of access to capital the residential market declines quite quickly once households with enough cash to spend on a system dry up. Providing access to capital funding so that all households can benefit from solar PV is a major opportunity.

In conclusion, rather than the Government needing to incentivise small scale PV, they need to monitor deployment carefully and develop plans to either slow deployment and/or mitigate potential grid instability (such as regulating the co-deployment of storage, or limiting grid export as has been done in Germany), and/or provide access to capital for lower income households to benefit

The level of residential PV deployment that could cause an issue for the distribution network is currently unclear and is linked to local and regional electricity demand profiles that are in turn linked to the adoption of electric vehicles. While work is progressing to better understand these issues a lack of access to residential electricity metering data is hampering progress. Critically there is a need for a set of practical guidance when interpreting the application of GDPR to smart meter data. Our current work under the Smart Meter Data Repository programme (Innovate UK) seeks to address this, but due to the complexity of the overall governance of the electricity sector, changes in data governance practices and regulation by Ofgem are much slower than the adoption of solar PV and EV technology and more direct Government intervention on this point may be needed.

7.      Does Government policy and current planning guidance adequately address the issues raised by proposals to install solar farms on land with high agricultural or ecological value?

It is certain that to achieve an installed capacity of 70 GW (as in current UK policy and Future Energy Scenarios) land use conflicts will arise. High grade agricultural land is attractive to solar PV investors and installers. To achieve 70 GW (approximately 5 times the current capacity) we assume that between 5 and 7 times the current area occupied by solar PV will be required. Current land use by PV is 24,400 ha or 0.1% of the total land in the UK. Scaling up between 5 and 7 times leads to between 125,000 ha and 175,000 ha being needed (or 0.5 to 0.7 % of the total land). It is our opinion that while current planning guidance is in principle appropriate, much more attention should be paid to developing a mixed land use economy, where agriculture and energy from solar PV (and wind) not only co-exist but are mutually beneficial.

“Best and most versatile land” (BMVL) refers to agricultural grades 1 to 3a, and Utilised Agricultural Area (UAA) is 17.2 million ha, or 71% of the UK land. Ground-mounted solar parks will not be evenly distributed across the UK, so while 0.5 to 0.7% of the total UK land area may be considered small, this land use change will be concentrated in some areas e.g. south England. Current policy and guidance The National Planning Policy Framework (NPPF) sets out the national planning policy context for renewable energy in England. From BRE Planning Guidance for Large Scale Solar for large scale solar,

“Ground Mounted Solar PV projects, over 50kWp, should ideally utilise previously developed land, brownfield land, contaminated land, industrial land or agricultural land preferably of classification 3b, 4, and 5 (avoiding the use of “Best and Most Versatile” cropland where possible).”

While solar PV is not currently restricted on BMVL, it is advised against, and less likely to receive planning permission and more likely to face objection by local communities. There are also requirements to return land to previous use post decommission.

Our research is currently assessing the potential overlap between agriculture land grades and ease of solar PV installation (ie. gaining grid connection) in Great Britain. The premise of this work is that to alleviate land use tensions the coexistence of agriculture and photovoltaics (ie agrivoltaics) is necessary. This involves a certain amount of technological development in the way solar systems are implemented, and an understanding of energy yields in these kinds of installations, but also, more importantly, a proactive dialog between land owners, tenants and investors.

Initial data in the assessment of PV-agriculture land use conflicts

We assessed the location of land suitable for PV development that overlaps with agricultural land (figure 4), i.e. where there is the potential for PV-agriculture land use conflicts. Locations with high “PV potential” scores of 9-10 and overlap/conflict with 42.7% of agricultural land, with the following proportions of agricultural grades: Grade 1 land overlapping with high potential for solar PV: 58.9%; Grade 2: 50.1%; Grade 3: 44.2%; Grade 4: 32.2% (note: grades 3a and 3b in England and Wales were combined, as were 3.1 and 3.2, and 4.1 and 4.2 in Scotland). It is clear that high grade agricultural land is also the most suitable for PV development, despite only making up a minority of total agricultural land: e.g. the highest grade of agricultural land, grade 1, makes up only 2.5% of total agricultural land, yet more than half of this is also highly suitable for PV development.

A reasonable strategy might be to prevent the development of solar PV on grade 1 agricultural land since this would have little impact on the total land available for PV. This would alleviate some of the tensions without substantially affecting land available for solar PV and it may also provide greater certainty in the planning process. The downside to this approach is that it would complicate planning law and may be difficult to apply in practice and this might slow development rather than speed it up.

Table 2. The proportions of agricultural grades of land that overlap with high potential for PV development (high potential = suitability scores 9 & 10). Note: grades 3a and 3b in England and Wales were combined, as were 3.1 and 3.2, and 4.1 and 4.2 in Scotland.

Further results from this study will be available to share as we look to complete our manuscript and publish the findings in Q1 2023.

8.      How sustainable is the supply chain for solar panel manufacture? Do levels of sustainability differ between mature and emergent technologies?

The solar PV market is almost entirely dominated by crystalline silicon (c-Si) technologies, which are forecast to take ~96% market share in 2022 ([5] – Exawatt, PV Market Update Q2 2022). The remaining 4% market share is almost entirely attributable to thin-film modules based on cadmium telluride (CdTe) sold by US-headquartered First Solar. This situation is unlikely to change in the next 15 years. The idea of an emergent (ie. disruptive) solar PV technology is not consistent with the reality of the consolidated and globally dominant Chinese solar PV supply chain. Disruption by an emergent technology is highly unlikely given the maturity of c-Si PV technology, the high capital cost of solar PV manufacturing plant and the cumulative learning (and total investment) made in the last 20 years. 

Conventional c-Si manufacturing is divided into five key steps: the manufacturing of polysilicon, ingots, wafers, cells, and finally modules. Ingot manufacturers almost always also undertake their own wafer slicing, meaning that many in the industry refer simply to “wafer” manufacturing when discussing combined ingot and wafer manufacturing. At this stage, it is worth noting that China currently dominates all stages of the PV manufacturing value chain, with this being particularly acute at the ingot and wafer manufacturing steps (Figure 7). In addition, the polysilicon and ingot/wafer manufacturing stages of the value chain are highly consolidated among a small number of large players – Exawatt expects the top ten manufacturers to account for ~97% and ~93% of total global production respectively in 2022. Mature solar technologies have energy payback times in the region of 5 years (A comprehensive UK specific study was conducted in 2016 and 2017 [13, 14] using data from 2010 to 2015, but there are no more recent analyses. Since 2016 PV module costs[AB1] have substantially reduced and efficiency substantially increased (see Figure 3) and it is very likely that payback times are now < 5 years). CO2 emissions are principally driven by energy consumption in manufacturing of polysilicon and its feedstock MG-Si. Some manufacturers do claim meaningfully lower carbon footprints – for example Norwegian wafer manufacturer Norsun uses hydroelectricity for its ingot and wafer manufacturing, and modules based on its wafers have been reported to fare well in some PV tenders in France where carbon footprint is taken into account.

Emerging technologies that are likely to gain significant market share in the next decade are likely to build on the existing c-Si industry and so will have many of the same fundamentals when it comes to sustainability. Some manufacturers do hope to enable lower carbon footprints by replacing the conventional polysilicon, ingot and wafer manufacturing steps, but are likely to remain at relatively low market share (<10%) even if they hit their capacity expansion targets [21]. While energy payback time for PV modules is far lower than their lifetime, the rapid increase in production and deployment that will be required to achieve net zero targets means that reducing carbon footprint for PV modules is of increasing interest in order to avoid very significant emissions as production ramps up.  Materials requirements for PV will be significant as global demand grows (Exawatt expects global annual PV module production to exceed 1,000 GW by 2030). This will lead to significant demand on silver [15] , polysilicon [16] and even abundant materials such as copper, concrete and aluminium [17] and [18]. New technologies may experience some challenges with materials requirements – for example HJT cells employ indium tin oxide as a conductive and transparent layer, and indium will need to be replaced in these cells if HJT is to take significant market share and PV is to continue to grow as expected [15].

Looking at ESG concerns beyond carbon footprint and materials requirements, the most topical currently relates to concerns over possible human rights abuses in the Xinjiang region of China. The Xinjiang region is home to several large metallurgical-grade silicon (MG-Si) manufacturers (including the world’s largest, Hoshine) – MG-Si being the key feedstock for polysilicon manufacturing. In addition, Xinjiang is a key region for polysilicon manufacturing – accounting for over 50% of Chinese polysilicon capacity at the end of 2020, although this is set to fall to ~30% by the end of 2024 (Exawatt market intelligence). The USA has effectively blocked imports of goods containing material from Xinjiang (under the Uyghur Forced Labor Prevention Act (UFPLA) there is now a “rebuttable presumption” that such goods are made using forced labour); however, other regions have not yet followed suit.

Based on discussions with module buyers and module manufacturers, Exawatt believes that while most major manufacturers are now offering US-based buyers “Xinjiang-free” modules (as required to get these modules into the US), these manufacturers are pricing these modules out of the market in other regions by quoting very high premiums for them. Note that Exawatt believes that there is only a very small cost premium associated with manufacturing these modules – the price premium is not directly related to extra costs incurred by the manufacturers. Exawatt also expects that manufacturers are likelyto continue to price “Xinjiang-free” modules out of most markets where they are not legally required. As yet, it remains unclear whether the supply-chain traceability in place for “Xinjiang-free” modules is currently sufficient for US Customs to release detained shipments into the US market. Given the highly consolidated nature of the polysilicon and wafer industry and the high number of polysilicon manufacturers with operations in the Xinjiang regions, it is also worth noting that while many manufacturers are now offering “Xinjiang-free” modules, almost all manufacturers will have a supply chain in which one of their suppliers handles material from Xinjiang or has manufacturing facilities in Xinjiang, even if the materials for a certain batch of modules can be certified as being “Xinjiang-free”.

9.      Does the concentrated global distribution of solar panel supply chains (80% manufacture in China) pose a risk to solar technology expansion in the UK? If so, how could this be mitigated

China currently dominates all stages of the PV manufacturing value chain, with this being particularly acute at the ingot and wafer manufacturing steps (Figure 7). While there are some supply-chain risks associated with this level of concentration, the principal risk would come from worsening political and/or trade relations between the UK and China. Should the UK choose to implement major trade barriers to PV imports from China it would be reasonable to expect significant disruption to PV module imports and therefore a significant decline in PV installations. By way of an example we can look to the USA - here there have been a number of disruptions due to trade issues in recent years. Many of these involve import tariffs (both antidumping and countervailing duties on imports from China, as well as Section 201 tariffs on imports from essentially all countries), which have significantly pushed up the price of PV modules in the US when compared to other regions. Beyond this, there has been significant disruption due to the implementation of policies aimed at limiting imports of products tied to the Xinjiang region. First a Withhold Release Order banning the import of any products containing material from Hoshine (the world’s largest producer of metallurgical-grade silicon, the key feedstock for polysilicon manufacturing) and then the Uyghur Forced Labor Prevention Act (UFLPA) which is reported to have significantly disrupted imports from many of the largest module suppliers since its implementation in June 2022. Note that this has been a challenge despite the fact that the US market is predominantly served by cell and module manufacturers in SE Asia (a situation which arose due to a series of additional duties on cell and module imports from China). Mitigation of this risk would require significant expansion of non-China supply chains. However, this is not trivial since China is the dominant player not only in the areas of the supply chain shown in Figure 7, but also in supplying key input materials such as PV glass, aluminium frames, polymeric backsheets, encapsulants, manufacturing equipment and so forth.

As of late 2022, materials and consumables costs account for ~85-90% of PV module manufacturing costs for a Chinese producer (with labour, electricity and depreciation accounting for the remainder) [Exawatt; based on a manufacturer integrated from ingot to module as per leading manufacturers]. Given this, scaling cost-competitive PV supply chains outside of China is also likely to rely on scaling supply chains for these bulk materials. There are currently believed to be moves ongoing by several major existing manufacturers to establish non-China supply chains. However, these are currently very limited in nature and will take time to establish. The USA and India have recently introduced PV manufacturing incentives as part of the Inflation Reduction Act and a Production Linked Incentive scheme respectively. While these schemes appear likely to lead to the growth of significant PV manufacturing capacity in these two countries, it is highly uncertain whether this capacity will maintain cost competitiveness once the incentives come to an end. In addition, in many cases these factories will import significant volumes of input materials from China.  Note that while module manufacturing can be established in a relatively short time frame (typically 6-9 months in China or SE Asia), other manufacturing facilities such as polysilicon and PV glass can take 2-3 years to establish and ramp to full production, and will only be established if there is confidence in a stable and long term downstream demand for those materials. As such, any attempt to build supply chains outside of China will only succeed if it takes a long-term approach.

10.  Are emerging proposals to utilise solar energy overseas capable of supplying additional sustainable grid capacity via interconnectors to the UK?

The more interconnected the GB and Northern Ireland electricity grids are to mainland Europe the lower our carbon emissions will be and the cheaper our electricity. As a result, current energy policy already targets a tripling in interconnection from 6 GW to 18 GW by 2030 [17]. There are two linked factors that are important in this respect. The first is the effect of geographic smoothing on intermittent wind and solar PV electricity. The larger the geographic area that is connected within a single electricity grid the smoother electricity generation will be and the higher the utilisation of renewable energy as a result.  The second related effect is longitudinal shifting of solar energy and weather patterns. Electricity networks that have wide longitudinal spread have the lowest carbon intensity of all because while at times of darkness (ie. no solar PV generation) in certain places on the network, there are other places that still have sunlight. Research has shown that as the geographic area of an electricity network increases from 10 thousand km2 (the size of a UK county) to 10 million km2 (the size of the EU or US) the proportion of electricity that can easily and cheaply be generated from renewable sources (solar and wind) increases from and central case of 40% to 60% along with a reduction in the cost of electricity of 20% [18]. In cases of extremely high natural gas prices the same research shows that the lowest cost energy systems for the US have 90% renewable generation.

Clearly connections to nearby mainland Europe will be the cheapest to implement but other connections to Spain or even Northern Africa would lead to a similar benefit albeit at higher cost.

December 2022

Figures

Figure 1 - From [4]. Measurement of the actual PV deployment in the UK compared with expected levels based on a machine learning model of capacity as a function of GDP, tertiary education and land availability. Full details available on request.

Figure 2 - From [5]. The evolution of average module cost of goods sold (COGS, i.e. manufacturing cost including depreciation) and average module efficiency for major Chinese module manufacturers from 2010 to 2020, based on financial reports, company statements and published module datasheets (solid line) and in 2021 based on Exawatt’s bottom-up cost modelling and published module datasheets (dashed line).

Figure 3 - From Hirth et al. [6]. Measurements of the effect of different levels of solar PV market share in different national electricity systems. As the share of PV increases the value of the solar electricity reduces (by about 30% per 10% market share). The variation in the effect in different electricity systems is due to differing levels of interconnection and different generation mixes.

Figure 4. (From current research [3]) Theoretical potential for solar PV development to conflict with agricultural land use in Great Britain. The map has been constructed by overlaying the grade of agricultural land with proximity to grid connection. Dark blue (grade 10) is where the highest grade agricultural land is in close proximity to a grid connection.

Figure 5. (Data source: MCS [2] & Ofgem) Top pane shows the average fixed electricity tariff (£/year) alongside (bottom pane) the monthly installation of residential (<4 kW; crosses and dotted line) and small scale commercial (>4 kW and <50 kW; circles and full line) solar PV systems. The 282 excluded rows are data prior to 2020 that is available on request. The data source is the Microgeneration Certification Scheme (MCS) and while not fully comprehensive, is estimated to cover 90% of installations [1] for the range of system sizes it applies to.

Figure 6. Historical residential and small scale commercial system installation rate (Microgeneration Certification Scheme). The current installation rate of small and medium sized systems is 60 MW per month. While historical installation rates (within the MCS records) for commercial systems were higher than for residential systems this has currently flipped and residential installation rate is higher than the commercial rate.

Figure 7. From Exawatt Ltd. The solar PV value chain. The five key steps in solar PV manufacturing are illustrated in the top pane. Quantification of the location the supply chain is in the bottom pane. Polysilicon is produced by the electrolysis of raw silicon dioxide and this uses large amounts of electricity and is typically co-located with hydroelectric power. There are only a handful of Global polysilicon manufacturers.  Ingot and wafer manufacture is dominated by China while cell and module manufacture is slightly more diversified geographically.

Figure 8. Source Geomni Building Stock Database.  A summary of building stock data from Great Britain.  The total roof area of different classes of buildings; the total number of these buildings; the spatial efficiency of installation of solar PV as measured by area per building within each class.

Figure 9. Historical annual GB Solar PV deployment by system size. Ground mount systems (top) were largely installed under the Renewable Obligation Certificate incentive prior to 2017. Large roof mounted systems (second from top) were also incentivised by the Renewable Obligation certificate, and since the termination of that scheme there is no visibility of these systems in the capacity registers [2]. Residential systems (third from top) are currently rebounding as are  small scale commercial roof mounted systems (bottom).

Figure 10. Historical cumulative GB Solar PV deployment by system size. Note that the blue line showing large roof mounted systems stops reporting new installations in Q4 2019. There is a gap in the reporting mechanisms for these types of systems and the level of deployment is uncertain.

Figure 11. A summary of the historical, current and future projected capacity (GW) by size and type of solar system in stacked (top) and bar chart (bottom) presentation. Current data is measured using a combination of datasets as described in [2]. Future projections are estimates made using historical maximum deployment rates for the different types of system as shown in figure 8, available roof space (and an average roof utilisation factor of 10%) and an adoption efficiency (figure 8) that is the average area per installation divided by 10000 (that leads to commercial roof adopting solar PV  in around 30% of cases and residential roofs adopting in around 5% of cases).

References

[1] O. T. Huxley et al. The Uncertainties Involved In Measuring National Solar Photovoltaic Electricity Generation. Renewable and Sustainable Energy Reviews. Vol. 156 (2022) pp112000. https://doi.org/10.1016/j.rser.2021.112000

[2] W. Spry. Review of Solar PV Capacity Publications. Department for Business, Energy and Industrial Strategy. Energy Trends. March 2022.  Accessed 16/11/2022.

[3] T. Mctiernan, R. Randle-Boggis et al. Working manuscript. Mapping the Potential Benefits of Agrivoltaics in the UK. Available on request.

[4] H. Alghanem et al. Forecasting Global Installed Solar PV Capacity from Yearly Standard National Statistics.  Manuscript in preparation for submission to Renewable and Sustainable Energy Reviews. Authorised access available on request.

[5] Exawatt Ltd. Solar PV Market Tracker. Q4 2022. 

[6] L. Hirth.  The Market Value of Variable Renewables. (2012) FEEM working paper.  15.2012. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2030215

[7]  D. Travers. Working manuscript. Geographic Variation in the Market Value of Renewables. A UK case study.  Available on request.

[8] F. Biggins et al. Manuscript submitted to Elsevier. The economic impact of location on a solar farm co-located with energy storage. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4269782

[9] Sustainability First. Smart Meter Energy Data Public Information Advisory Group report. (2021). https://www.sustainabilityfirst.org.uk/projects-smart-meter-data

[10]  Modernising Energy Data programme.. https://www.gov.uk/government/groups/modernising-energy-data

[11] M. Raugei & Leccisi.  A comprehensive assessment of the energy performance of the full range of electricity generation technologies deployed in the United Kingdom. Solar Energy. Vol 90 (2016) pp. 46 - 59 https://doi.org/10.1016/j.enpol.2015.12.011

[12] M. Raugei et al. Energy Return on Energy Invested (ERoEI) for photovoltaic solar systems in regions of moderate insolation: A comprehensive response.  Solar Energy. Vol. 102  (2017) pp.377 - 384.  https://doi.org/10.1016/j.enpol.2016.12.042

[13] Exawatt Ltd. Report commissioned by Nexwafe. Enabling higher efficiencies, lower cost and lower carbon emissions in PV manufacturing.  https://www.nexwafe.com/whitepaper

[14] Yuchao Zhang, Moonyong Kim, Li Wang, Pierre Verlinden and Brett Hallam. Design considerations for multi-terawatt scale manufacturing of existing and future photovoltaic technologies: challenges and opportunities related to silver, indium and bismuth consumption.  Energy and Environmental Science. Vol.14 (2021) pp.5587. https://pubs.rsc.org/en/content/articlelanding/2021/EE/D1EE01814K

[15] Brett Hallam, Moonyong Kim, Robert Underwood, Storm Drury, Li Wang, Pablo Dias. A Polysilicon Learning Curve and the Material Requirements for Broad Electrification with Photovoltaics by 2050. Vol 6. (2022) pp. 2200458 https://onlinelibrary.wiley.com/doi/10.1002/solr.202200458

[16] Alison Lennon, Marina Lunardi, Brett Hallam & Pablo R. Dias. The aluminium demand risk of terawatt photovoltaics for net zero emissions by 2050. Nature Sustainability. Vol.5 (2022) pp.357-363   https://www.nature.com/articles/s41893-021-00838-9

[17] Current policy targets for interconnection are summarised  in National Grid ESO Future Energy Scenarios 2022.  https://www.nationalgrideso.com/future-energy/future-energy-scenarios

[18] A.E. McDonald et al. Future cost competitive electricity systems and their impact on US CO2 emissions. Nature Climate Change. (2016) http://dx.doi.org/10.1038/nclimate2921