I recently attended ecoConnect's session on the future of the UK Solar Industry, which featured a panel debate on the effects of proposed changes to the Feed-in-Tariff (FiT) on the the UK solar industry. To help get my head around the rather complex dynamics, I have been creating an online simulation model with a dashboard. The model makes some interesting and counter-intuitive suggestions about where the UK solar industry is headed, and how we can influence the path.

Using the Dashboard

This is work in progress: the text is still being written, to explain the model and its implications, and some of the data in the model is based on crude estimates. Also, this dashboard currently works best with recent versions of Firefox, Chrome, Safari, or Opera. It will work poorly, if at all, in Internet Explorer, which does not support the HTML5 features used by the dashboard.

The left side of the dashboard has 13 sliders, on which you can click or drag to change the underlying model parameter:

  • Initial p/kWh: the price of electricity at the beginning of the simulation, in pence per kilowatt-hour
  • Price trend: the annual increase in the price of electricity, e.g., .05 for 5% per year
  • Household electricity demand: the annual demand for electricity for the average household, in kilowatt-hours (kWh)
  • Generation feed-in-tariff: the tariff a homeowner receives for each kWh generated from solar
  • Export feed-in-tariff: the tariff a homeowner receives for each kWh generated from solar beyond the household's requirements, and thus exported to the grid
  • Starting installers: the number of solar installation companies operating at the beginning of the simulation
  • Installer maximum output: the maximum number of installations that can be performed by an average solar installation company in a month
  • Total households: the total number of households in the UK that could potentially install solar panels (estimate of 1 million obtained from the interesting analysis here)
  • Initial installed: the initial fraction (e.g., .05 for 5%) of eligible households that already have solar panels installed at the beginning of the simulation (i.e., at the beginning of 2010)
  • Initial cost: the average cost of installing solar panels, at the beginning of the simulation (i.e., in 2010)
  • Learning curve: the reduction in cost from each doubling of installed base; for example, 0.2 means that the cost will decrease by 20% each time the number of households with solar panels doubles
  • Capacity: the average nominal size of a new domestic solar installation, in kW
  • Efficiency: the output as a fraction of nominal capacity, e.g., 0.75 for 75%

The right side of the dashboard shows graphs for 6 different views, selectable by tabs:

  • Overview: shows total number of houses with solar installed, and the number of solar installation companies operating
  • Electricity: the annual electricity demand for a typical household (assumed constant, but can be changed by slider), and the price of electricity (based on initial level and annual trend, both settable via sliders on the left)
  • Technology: shows the cost of a new solar installation (declines because of learning curve, settable), and the payback period for a homeowner
  • Homes: shows the number of eligible homes with solar installed, and the number of homes installing over time
  • Industry: shows the number of companies, and the rate of entry/exit
  • FiT Scheme: the feed-in-tariff scheme, with graphs of typical benefits to a participating household, and overall cost of the scheme

You cannot yet save the value of your sliders (let us know in the comments if this is something you would like us to add). You can restore the starting values for all sliders at any time by clicking the refresh button on your browser.

How the Model Works

The model underneat the dashboard is a system dynamics simulation model that runs inside your browser to simulate the UK domestic solar installation market for 10 years. This model attempts to tie-together the economics of the system as a whole, i.e., from the points of view of consumers, energy companies, solar installers, and government.

Here is a high-level schematic of the model:

In this diagram, rectangles are stocks (resources which accumulate over time), circular valves on the straight double arrows are flows (which increase/decrease stocks over time), and thin arrows are influences. Words in green are variables that can be changed by moving the sliders.

Here is what is going on:

  • The price of electricity starts at a defined level, and increases over time at the rate you set (default 5%)
  • Prospective customers evaluate the payback period for a solar installation, based on the current cost, balanced against three sources of benefit:
    1. Cost saving from electricity generated instead of bought
    2. Generation tariff received for every kWh generated (even those exported to the grid)
    3. Export tariff received for every kWh exported to the grid
  • An attractive payback period creates demand for new installations; the model represents this by a table-lookup function, with low payback period translating to more demand
  • New installations is a flow which moves households from the stock "Houses without solar" to "Cumulative installations"; it is the lower of demand and industry capacity (see below)
  • The current cost of a new solar installation declines over time as installed base increases, due to the learning rate (e.g., if the learning rate is 20%, then after the first doubling of installations the cost becomes 80% of its former value); this of course reduces the payback period further, creating a postive feedback loop
  • The number of installers acts as a constraint to new installations, and also changes in response to supply and demand: if the workload is high (too few installers), new companies enter the market, and if demand falls so that there is excess supply, installers leave the market; this is a negative feedback loop

What the Model Tells Us

If you experiment with the sliders, you will see the following behaviour (described for each set of sliders):

Electricity Price

  • If electicity prices go up (either through the initial price, or the price trend) the payback period goes down, and this accelerates the adoption of solar
  • On the "Technology" tab, also notice how a higher electricty price causes the cost of a new installation to decline; this is because as more people install solar, the learning curve kicks in and lowers the cost of solar
  • On the "Industry" tab, notice how increasing the electricity price increases the number of solar installers (to keep up with the higher demand)
  • On the "FiT Scheme" tab, you will see that higher electricity prices naturally increase each household's benefit from "Cost saving" (because of the higher price per kWh saved)
  • Also on the "FiT Scheme" tab, the total cost of FiT also increases with higher electricity prices; this is due to the larger number of participants
  • The model assumes that changing the price of electricity does not affect the amount used (no price elasticity); this is clearly unrealistic, and will be changed in a future version of the model

Feed-in Tariffs

When you change either of the Feed-in Tariff (FiT) sliders, you will see a "kink" in some of the graphs at time around 2012.25, because the new FiT you set only comes into effect at the end of the first quarter 2012 in the model.

  • On the "Technology" tab, changing either FiT increases or decreases the payback period, because it changes the respective benefit received (directly visible on the "FiT Scheme" tab)
  • Also on the "Technology" tab, you will see that increasing the FiT levels decreases the cost of a new installation; this is because of the increased adoption, leading to reductions from the learning effect
  • Changing the FiT levels has a significant effect on the size of the industry; if you view the "Industry" tab and move the FiT sliders to zero, the number of solar installers drops precipitously (rate of leaving/entering shown on the graph on the right)
  • For all of the above points, changing the generation tariff has a much larger effect than changing the export tariff, because of the higher magnitude of the tariff

Solar Industry (Installers)

  • I had a hard time getting data on the actual number of installers, and went for a low estimate of 1,420 (from DECC), although there are estimates of up to 50,000 (see here); I suppose it depends on how you define an installer (comments welcome)
  • In general, increasing the number of installer will only cause them to drop out later in the game, since in the model the industry size adjusts to demand
  • In some scenarios (i.e., if you increase demand by increasing the price of electricity or increasing the FiT), adding more installers at the beginning will increase overall installations, because the industry does not need to "catch up" to demand (industry size is a constraint on installations, remember)
  • The "Max output" slider is there to let you set the productivity of an average installer; the starting value of 12 per month does not really mean anything, and was created to allow the initial number of installers in the model meet demand; comments on this welcome

Households

  • The number of households represents the number of eligible households in the UK, i.e., those for which installing PV is makes sense; the number 1 million (out of a total 26 million households in the UK) was obtained here
  • The initial installed fraction is set to 0.05 (5%), to give 50,000 as the approximate number of households with solar installed at the beginning of the simulation (i.e., at the start of 2010); this needs to be checked
  • These sliders are provided mainly as a way of checking different assumptions about the market size and initial installed base; their immediate effect is mainly to cause the installer industry to expand or contract to balance with the change in demand

Technology

  • Changing the initial cost or the learning curve has the effect of changing the payback period, and thus the adoption of PV
  • The 20% assumption for the learning curve is a guess, and needs to be checked
  • This section needs to be completed

Implications

This section is still being written.

Taking it Further

This first cut is a simple but illustrative, and invites several avenues for development:

  • Demand elasticity for electricity probably needs to be introduced (currently, electricity usage does not change with price)
  • Overall tax implications need to be included to reflect government trade-offs, i.e., tax effects of reduced employment
  • Other policy impacts, such as carbon reduction targets, could be included
  • Need to check some data, as noted above

I'm planning to refine sections of this model going forward, and to post entries as things happen. In the meantime, comments are welcome.

Add a comment