Efficiency differs across existing plants within each fuel category due to factors like age, variations within the technologies and plant specifications. We assume that for all existing plants in a country with the same type of technology, for example gas power, the thermal efficiency is a linear function of capacity utilization.
To determine a linear function, one requires two exogenous values. We let one point be the thermal efficiency of the most efficient plant in each country, which we define as the efficiency reported for new plants in the Projected Costs of Generating Electricity, Update 2005 (OECD 2005). The publication has efficiency estimates for plants coming online in 2010 for some of the model countries for the different technologies. For countries where there are no estimates (and for pumped storage production which is not mentioned in the publication) the average efficiencies reported for 2009 (in IEA Electricity Information 2011) have been used. Because the plants that were in operation in 2009 differ in age, and hence have varying efficiencies, this estimate is adjusted by multiplying with a factor of 1.05.
A candidate for the second point of the linear function could potentially be the observed efficiency, calculated as the net electricity production to fuel use. However, it is not straightforward to use the observed average efficiencies to determine the other fixed point of the linear efficiency function. First, the unused parts of all electricity capacities have unobserved efficiency. Assuming that these are mainly vintage plants with lower efficiency, the (true) average efficiency of total capacity will be lower than the observed average efficiency. Second, the different electricity-producing technologies do not have a constant rate of capacity utilisation throughout the year. These rates are not known from primary data. The data only provide information on annual rate of capacity utilisation for each technology and the distribution of total production over the four time periods. Instead of using average efficiency directly to determine the second point, we calibrate the capacity utilisation for each technology and period by imposing the requirement that, for each country, the outcome should be consistent with cost minimization in electricity production, given our data. The problem is solved by running the electricity production block of the model. The solution of the problem provides the efficiency of the least efficient plant actually used in any period (the marginal efficiency for each technology and country), which is used as the second point in the linear efficiency function.
For new power plants the efficiencies are mainly taken from Projected Costs of Generating Electricity, Update 2010 (OECD 2010). All estimates from this publication are for plants coming online in 2015. Because of the assumption that the cost of a new plant (of a given technology) is the same for all model countries, the same applies for the efficiencies. For new pumped storage there is constant efficiency within each country, but these efficiencies differ across countries because of, e.g. topological differences.
Actual thermal efficiencies for the fossil-fuel-based technologies are based on observed fuel use and production of heat and electricity in 2009 as reported by the IEA in their Electricity Information Statistics Database. The estimated heat-electricity trade-off coefficients were used to convert heat produced in 2009 to its electricity equivalent, ensuring that all electricity quantities (production and consumption) in the base year also include transformed heat.
Through using observations from the 1970s and up until today a trade-off between heat and electricity has been estimated across fuels for all model countries. There is a wide dispersion between countries and fuels. For gas, oil and coal, the results were highly significant; whereas for waste and biomass the results were only significant when these were estimated together (there are fewer observations for these).
To calculate the net electricity production for the combustible fuels the estimates for gross electricity production, along with the ratio between net and gross electricity production for combustible fuels (found through using data from IEA Electricity Information) has been used. This source also contains data on net electricity production from other types of technologies.