AS O

SI

SB A N A UM I I N IP D AS

PROCEEDING OF THE 5th INAGA ANNUAL SCIENTIFIC CONFERENCE & EXHIBITIONS Yogyakarta, March 7 10, 2001

Key words: Investment cost, geothermal power plants, expectation value, economy of size. ABSTRACT Stepwise development strategy is considered a suitable method for securing a cost effective way for the development of geothermal power plants. This strategy has been in use in Iceland for the last decade. Geothermal high-temperature fields are developed in steps of 20-30 MW. About 6 years are required for each step in the development. Parallel development of several fields in a country might be preferable, especially when a rapid increase of the generation capacity is required in that country. The capacity factor of geothermal power plants depends on the mix of power plants serving the electricity grid. Where geothermal power plants can be operated as base load, the capacity factor is usually in excess of 0.9. The investment cost of geothermal power plants is divided into the cost of surface equipment, including the cost of the plant and steam gathering system, and the cost of subsurface investment (drilling cost). The cost of surface equipment can be estimated with the same accuracy as other construction works at the surface (buildings, roads, and bridges), whereas higher uncertainty might be associated to the cost of drilling. Analyses of the surface cost of 5 power plants in Iceland shows that the investment cost is linear with the size in the range 20-60 MW. The surface cost is found to be about 1000 $/kW with a relative error of 10%. Stefanson (1992) has published a statistical study of the drilling results in 31 high-temperature fields in the world. By using these results of Stefansson, it is possible to estimate the expectation value and its limits of error for the subsurface investment in an arbritary geothermal field. The results obtained are summarized in the following Table-1

1. INTRODUCTION Analyzes of the investment cost for geothermal developments are not frequently published. In some cases, it is assumed that the development cost can hardly be predicted because of the uncertainty involved in geothermal drilling. Another common misconception is that the cost of geothermal drilling is the main obstacle for geothermal development. Given the facts that the drilling cost of a low-temperature geothermal development is typically some 10-20% of the total development cost, and that the drilling cost for a typical geothermal power plant in a high-temperature field is usually in the range 20-50% of the total cost of such a plant, it is quite obvious that geothermal drilling cannot be blamed exclusively for the uncertainty in geothermal development. The present paper describes an analysis of the investment cost of geothermal power plants and estimates the expected investment cost level in unknown geothermal fields. Experience and data from Iceland are used in the paper, but it is expected that similar conditions are applicable elsewhere in the world. Figures for cost are based on actual cases in Iceland. 2. DEVELOPMENT STRATEGY In the past, geothermal power plants have frequently been built as large as possible, meaning that a large effort has been made to determine the generating capacity of the geothermal field prior to development, and the size of the power plant has been selected close to this estimated level. There are several drawbacks to this development strategy and we will here mention four unfavorable conditions: a. It is of economic importance that there is a short time between the investment in drilling and the moment when the investment can begin generating income (start of production). Drilling cost is typically in the range 20-50% of the total investment cost of a geothermal power plant.

E SI ON
A

INVESTMENT COST FOR GEOTHERMAL POWER PLANTS

The feasibility of the power plant may therefore critically depend on the production wells being drilled a short time before the start of production. These circumstances can create conflicting goals for the development. Many wells should be drilled and tested at an early stage of the development in order to minimize the risk in the estimate of the generating capacity of the geothermal field, whereas most wells should be drilled as late as possible in order to minimize the investment cost. b. Reliable knowledge on the generating capacity and other production characteristics of a given geothermal reservoir can only be obtained by observing, for some years, the response of the reservoir to a given production. If it is adopted that the power plant should be developed in one step, it will be necessary to test the reservoir with some testproduction for some years before the size of the power plant is decided. This will result in a relatively high cost of investigation to be paid long before the generation can start. If a stepwise development is adopted, on the other hand, the production from the first unit can be used to monitor the reservoir response and estimate the generating capacity of the geothermal system. c. Historically, there is a limited number of geothermal fields where the capacity of installed surface equipment has been larger than the generating capacity of the field. Such a situation can be avoided by applying a stepwise development of the field. d. It is frequently observed that the enthalpy of the geothermal fluid changes with time during the exploitation. This can easily result in a mismatch between the turbines and the geothermal fluid available in the field. A stepwise development of the field will be of help to maintain a suitable match between the turbine design and the production characteristics of the geothermal fluid. This list shows that a stepwise development of a geothermal field has several benefits. A major issue here is that the production characteristics of a geothermal reservoir cannot be

Investment Cost For Geothermal Powerplants

Valgardur Stefansson

determined without a proper investigation of the production from the same reservoir. By selecting a relatively small power plant as the first step in the development of the field, the actual production can be used to monitor the response of the reservoir and to estimate the production characteristics of the field. This knowledge will allow for determining when the next step in the field development can be taken. This means that with this development strategy, the final value of the generating capacity of the field will not be known in detail until the field has been fully developed. Initially, when the stepwise development strategy was adopted in Iceland, it was assumed that 20 MW was a suitable step for developing a high temperature geothermal field for power generation. This size was selected because this was the largest skid mounted equipment available at that time (1990) and it was considered of interest that the equipment could be moved to another location, if the field under development turned out to be unsuitable for power production. Newer designs of turbines with overlying condensers are easier to install than the older types, so larger step units (30 MW) are frequently used today as step units in the geothermal development. This new strategy, to develop all geothermal fields in 20-30 MW units, can make it necessary to have more than one field under development simultaneously. A typical time unit for developing each step would be 6 years with the following components: Reconnaissance Surface exploration Exploration drilling Production drilling and power plant Total time 1year 1 1 3 years 6 years

grid are only hydro and geothermal plants (with some oil fired reserve units) and it is natural that the geothermal plants are used as base load and the hydro is taking up the variable load. For other composition of the production units serving the grid, a different operational time for the geothermal plants might be more suitable. In general, however, the capacity factor of geothermal plants is high resulting in a favorable production price of each kWh generated.

4. COST PARAMETERS
It is suitable to split the investment cost of geothermal plants into two categories. On the one hand comes the cost of the surface equipment (the plant itself and the steam gathering system) and on the other hand comes the cost of the subsurface investment (drilling the wells). One of the benefits of this cost break-down is that the limits of error in the cost estimate for the surface equipment are the same as for other construction work on the surface (buildings, roads, bridges), whereas higher limits of relative error might be expected due to the difficulties to predict the yield of drilling, and thus the number of steam wells required for the plant. Figure-1 shows how the total investment cost of a 20 MW power plant in Namafjall, Iceland, depends on the total number of wells required for this power plant. Figure 1 is based on a detailed project planning report carried out in 1994 (Orkustofnun and VGK, 1994). Previous drilling in the Namafjall field indicates the expected yield of future wells to be on average 4 MW per well. Some 56 production wells are therefore required for this 20 MW power plant. Assuming that additional 2 wells are needed for reinjection, the total number of wells required can be estimated as 8. From Figure-1, it can be seen that the total investment cost for this power plant is about 35 M$, or 1750 $ per installed kW. It should be noted that the subsurface cost for the 20 MW power plant in Namafjall is estimated 13 M$ out of the total cost of 35 M$, or 37%. The limits of relative error for the subsurface part of the cost estimate are higher than for the cost of the surface equipment. By assuming that the relative error for the surface part is 15% and that for the subsurface part is 40%, we see that the relative error for the estimate of the total investment is 24%. Figure-1 shows that if there was any economy of size present for geothermal power plants, this effect would show up in the surface part of the investment cost. In a given geothermal field, there is a direct proportionality between the size of the power plant and the number of wells required providing enough steam for the power plant. Recent cost figures for the cost of surface equipment are available in Iceland. The cases regarded here are given in Table-1. The new power plants at Svartsengi and Nesjavellir are cogeneration plants producing both heat and electricity, whereas the plants at Bjarnarflag and Krafla are designed only for electricity generation. The data on the surface cost in Bjarnarflag and Krafla are extracted from the detailed project planning reports for the plants at these locations. Figure-2 shows the relation between the surface cost and the size of the power plants listed in Table-1. The data points define a linear relationship and the best linear fit to the datapoints is:

Decision on the power plant is to be taken on basis of the results of drilling 2 exploration wells in the third year of the investigation for the first step. Production drilling is carried out simultaneously with the construction of the power plant. When the electricity generation starts, 6 years after the beginning of the investigation, the monitoring of the reservoir starts, and after some 3 years of reservoir analysis it should be possible to decide whether the next step (20-30 MW) is advisable or not. Allowing 3 years for this phase, the next unit should be ready 6 years after the start of the first unit, and so on. Financial analysis of the stepwise strategy shows that the long time investment cost for the development of a geothermal field is considerably lower than by developing the field in one step. Much shorter time elapses before some revenue is obtained from the investment and the production cost of electricity becomes lower. The geothermal industry in Iceland has adopted the stepwise development of geothermal fields.

3. CAPACITY FACTOR
Geothermal power plants are independent of climate and seasons. They can operate at constant load day out and day in all year round. Therefore, it is suitable to operate the geothermal power plants as base load when they are connected to a grid together with other types of power plants. In Iceland, the average operational time of geothermal plants is about 8000 hours per year. In this case, the production units serving the

Investment Cost For Geothermal Powerplants

Valgardur Stefansson

Surface cost (M$) = (- 0.9 4.6) + (1.0 0.1)*MW

(1)

67.7 and 42.5 M$ respectively. This gives the range 1062 1692 $ per installed kW with the most likely figure 1267 $/kW. For the first step of the field development, the learning cost has to be added to the cost estimate. This cost is associated with drilling a sufficient number of wells in order to know to site the wells for maximum yield of drilling. The average number of wells required for this is 9.3 6.1 wells. By assuming that the average yield in the learning period is 50% of the yield obtained as the asymptotic value of the field, these circumstances are equivalent to that an additional cost of drilling 4.63.0 wells is to be added for the first development step. Assuming that the cost of each well is 1.5 M$, the additional cost is 6.94.5 M$. For the first development step in developing a geothermal field, the expected investment cost can be estimated as: Investment cost (M$) = (6.0 9.1) + (1.29 +0.31/-0.19) * MW

and the correlation coefficient is R2 = 0.97. This equation is valid in the range 20 60 MW. The data presented in Figure-2 indicate clearly that there is no economy of size for geothermal power plants in the range 20 60 MW. In fact, an attempt to fit a second order function to the data in Figure-2 results in a small second order term for a concave function, meaning that the investment cost is increasing faster than linearly with size. 6. SUBSURFACE COST Equation (1) could be adopted as an average cost estimate for the surface cost of geothermal power plants. For the sub surface part of the cost estimate we use the results of Stefansson (1992). In his paper Stefansson analysed the result of drilling in 31 high temperature geothermal fields in the world. He showed that the average yield of wells in each geothermal field is fairly constant after a certain learning period has been passed and a sufficient knowledge on the reservoir has been achieved to site the wells in such a way that the maximum yield is obtained. Figure-3 shows examples of the results presented by Stefansson (1992). The average output per drilled km in four geothermal fields is shown as a function of well number in each field. For these fields in the Philippines and New Zealand the average yield is about 4-5 MW for each drilled km in the fields. The figure shows also that even though the asymptotic yield of all the four fields is similar, a different number of wells were required to reach the level of maximum yield of wells in the field. Table-2 shows the average values for the 31 geothermal fields studied by Stefansson (1992). Figure-4 shows the histogram of the yield per drilled km for the 31 fields presented by Stefansson (1992). These figures can be used to estimate the average subsurface cost for geothermal power plants. We assume that the average depth of the wells is 1500 m, and that the average cost of such wells is 1.5 M$ (drilling cost in Iceland as reported by Stefansson, 2000). The average yield of the 1500 m wells is 1.5*(3.41.4) = (5.12.1) MW, and the cost per MW is 1.5/(5.12.1) = 0.29 (+0.21/-0,09) M$. The cost range is thus from 0.20 to 0.50 M$/MW with the most likely value of 0.29 M$/MW. Combining this result with equation (1), the expected investment cost in a known geothermal field (second or later steps in the field development) can be estimated as: Cost (M$) = (-0.9 4.6) + (1.29 +0.31/-0.19) *MW (2)

(3)

The most likely investment cost for a 40 MW unit in this case is 57.6 M$ with the upper and lower limits 79.1 and 44.9 M$ respectively. This gives the cost range 1122 1992 $/kW with the most likely value 1440 $/kW. 7. CONCLUSIONS Statistical methods have been used to estimate the investment cost of geothermal power plants where a stepwise development strategy is applied. The cost figures obtained are summarised in Table-3 8. ACKNOWLEDGMENTS The author thanks Sveinbjorn Bjornsson and Ingvar B. Fridleifsson for reviewing the manuscript and suggesting several improvements in the presentation. 9. REFERENCES Hitaveita Sudurnesja (2000) Annual report for 1999, (In Icelandic) Hitaveita Sudurnesja. Orkustofnun and VGK (1994) Geothermal power plant in Bjarnarflag. Project planning report. (In Icelandic) Landsvirkjun and Orkustofnun Report. Stefansson, V (2000) Competitive status of geothermal energy in the 21st century. Paper presented at Plenary Session III at the WGC2000. Stefansson, V. (1992) Success in geothermal development. Geothermics, 21, 823-834. VGK, Rafteikning and Orkustofnun (2000) Extension of Krafla Power Plant. Project Planning Report, (In Icelandic) Landsvirkjun.

This means that the most likely investment cost for a 40 MW power plant in a known geothermal field is 50.7 M$ and with one standard deviation limits, the upper and lower levels are

Investment Cost For Geothermal Powerplants

Valgardur Stefansson

Table-1 The results obtained are summarized Expectation value $/kW Surface cost only Total cost in a known field Total cost in a unknown field 977 1267 1440 Range within one standard deviation $/kW 762 - 1192 1062 - 1692 1122 - 1992

Table-2 Cost of surface equipment for various geothermal power plants in Iceland Location Svartsengi Nesjavellir Bjarnarflag Bjarnarflag Krafla Size of power plant MW 30 60 20 40 40 Cost of surface equipment M$ 26.16 62.38 23.00 38.80 41.89 Type of cost estimate Actual cost Actual cost Project planning Project planning Project planning Year 1999 1998 1994 1994 1999

Table-3 Average values for 31 geothermal fields (Stefansson, 1992). Average MWe per well Average MWe per drilled km Average number of wells before max. yield achieved 4.2 2.2 3.4 1.4 9.3 6.1

Table-4 Investment cost of geothermal power plants Most likely value $/kW 977 1267 1440 Range within one standard deviation $/kW 762 - 1192 1062 - 1692 1122 - 1992

Surface cost only Total cost in a known field Total cost in a unknown field

20 M W po w e r plant in N m afjall
60 55 50 60 70

Surface cost

Investm ent cost [ M $ ]

45 40 35 30 25 20 15 10 5 0 -2 0 2 4 6 8 10 12 14 16 18 20 22 0 0 10 20 30 40 50 60 70 40 50

M$
S u rface eq uip m en t

30

20

10

N u m b e r of w e lls

Figure-1 Investment cost of a 20 MW power plant in Namafjall

Figure 2. MW Relation between surface cost and size of power plant

Investment Cost For Geothermal Powerplants

Valgardur Stefansson

Average output per km drilled as a function of well-num ber

Average output per km drilled [ MW ]

11 10 9 8 7 6 5 4 3 2 1 0 0 20 40 60 80 100

W airakei Tongonan Tiwi Ohaki

-1 120 140

N um ber of well

Figure-3 Average output per drilled km as function of well number in four high-temperature geothermal fields.

10

0 -2 -1 0 1 2 3 4 5 6 7 8 9 10

M W /k m

Figure-4 Histogram for the parameter MW/km from 31 high-temperature fields.