The useful life of each of the plants in case W-Z was assumed to be 25 years based on a study on the performance of the Andasol-1 plant in [83]as well as the fact that the existing FiT rate applicable to CSP in Spain is a value of 27 € ¢/kWh for a PPA period of 25 years [83]
The plant capacities simulated are based on the two reference plants selected, that is the 20 MW Gemasolar SPT plant and the 50 MW Andasol-1 PT plant. In order to investigate whether there is a cost saving in terms of LCOE and capital cost/
Watt especially for the SPT plant, an arbitrary size of 100 MW was selected. The simulated output of the 100 MW case Z plant was compared to an existing SPT plant of similar capacity, that is, the 100 MW Crescent Dunes SPT plant.
The selection of the two reference plants was based primarily on availability of data on technical parameters as well as data on cost of principal components which are considered vital to enable the development of models that can estimate the performance of the plants in case W-Z as realistically as possible.
In terms of the economic analysis, the main parameter that was investigated is the LCOE. This cost in SAM accounts for expenses such as; procuring equipment, operation and maintenance costs, interest payments, tax remittances and benefits as well as the salvage value. There is also an alternative to include investment or capacity based cash incentives into the computation however these fields were not included in the simulation since there is currently no legislation in Kenya on any such incentives.
The computation of the nominal LCOE which is the main economic parameter that has been discussed for all the cases is indicated in Equation 2 [74],
𝑛𝑜𝑚𝑖𝑛𝑎𝑙 𝐿𝐶𝑂𝐸 = −𝑋𝑂−
where Qn is the total annual energy generated in kWh, Xo is the equity investment, N is the project useful life which in this case is 25 years, Xn is the annual project cost for a particular year (n), dr is the real discount rate which doesn’t account for inflation while dn is the discount rate which accounts for inflation.
46 4.3.1 Sizing the solar field
For the SPT plant, the sizing of major components of the solar field such as the tower height and receiver is carried out using the SAM optimization tool. The tool optimizes the length of the tower, length and diameter of the receiver as well as computes the heliostat positions. Reducing the cost of LCOE to the least possible value is the main objective and constraints include a limit of the maximum flux that can be incident on the receiver as well as instances where the power obtained from the receiver falls short of the design value after the piping losses and losses from the receiver surface are taken into account [74]. After obtaining an optimal heliostat field layout and optimal height of the tower as well as receiver size, a parametric analysis was carried out to determine optimal size of the solar field that yields the lowest LCOE as discussed in section 5.1.4.
For the PT plant, the layout in regard to the number of collectors in a loop was selected based on the Andasol-1 reference plant as indicated in Table 13. In as far as the actual size of the field is concerned; a parametric analysis is carried out to determine the solar multiple that results in the largest possible annual energy output but at a minimized installation cost. Checking for the lowest LCOE is therefore the most convenient way to narrow down an appropriate size of the solar field as has also been discussed in [74].
4.3.2 Model validation
Validation of the model developed in SAM was done by comparing simulated values to those reported of the actual performance of the respective reference plants. For the solar power tower plant, a 20 MW plant with technical and economic parameters indicated in Table 11 and Table 12 was simulated for Seville whose weather file is available in SAM. It is not possible to replicate all of the operational conditions of the Gemasolar SPT plant due to limitations in the software since for this case there is no option to integrate NG backup for a SPT plant. Nonetheless as discussed in
47 section 5.2.1), it is estimated that there is a 7 % error margin of the SAM model for the SPT plant. This is considered to be an acceptable error margin for the purpose of estimating the performance of the case Z plant.
For the PT plant cases W-Y; a model was first developed for the Andasol-1 plant based on parameters highlighted in Table 13 and Table 14. This plant is noted to have a 12 % NG fuel backup and this was included in the simulations as well [83].
The results of the comparison between the simulated and reported values are indicated in Table 15.
Table 15: Comparison of reported against simulated values for Andasol-1 plant
Parameter Reported Simulated Difference (%)
Annual energy 174 GWh 173 GWh 0.5
Capacity factor 40.20 % 39.90 % 0.7
Fossil fuel backup 12 % 10 % 2
land area 476.8 acres 460 acres 3.5
Since no weather file for the actual location in Aldeire, Spain exists on the SAM database, a location was selected that has a similar DNI value which is 2,136 kWh/m2/year and this value is equivalent to 5.85 kWh/m2/day. The location selected is Chula Vista brown field in California which has a DNI value of 5.75 kWh/m2/day.
Based on the results in Table 15, the PT plant model developed in SAM is assumed to be able to estimate the performance of the proposed plants in case W-X with an approximate error margin of 2 %.
Fossil fill fractions were varied for the Andasol SAM simulation in order to get as close as possible to the reported 12 % share of annual energy production from the NG backup boiler. In order to achieve the 10 % fossil fuel share in the simulation, fractions of 0.15, 0.2 and 0.35 were used for period 1, 2 and 3 respectively.
For both PT and SPT simulations for the Andasol and Gemasolar plants, a summer peak dispatch period is used as indicated in Figure 17 which matches the seasonal demand load curves for the respective locations in Spain.
48
Figure 17: Summer peak dispatch schedule [74]
49