Selection of Optimal Tunnel Route and Tunneling Method based on AHP Technique . Case Study – Broadlands Hydropower Project

Selection of optimal tunnel route and tunnelling methods for head raise tunnel of a hydropower project is one of the most critical and complicated steps in its design stage. Since factors involved are highly diversified, the decision making process is quite complicated. In this study, the Analytical Hierarchy Process (AHP), a multi criterion decision making tool was employed to establish the optimal tunnel route and the most appropriate tunnelling method for the established optimal tunnel route. The optimal tunnel route was established by evaluating two tunnel route options with ten factor criterion. Four tunnelling alternatives were evaluated against relevant ten factor criterion in selecting the most suitable tunnelling method for the selected optimal tunnel route. Criterion were limited to geotechnical, geological, geomorphological, safety, technical, economical and socioenvironmental aspects. Analysis reveals that, route B, which is having the overall composite weight of 0.513 (51.30%) is the optimal tunnel route and combination of Drill & Blast method (section III of route B) and New Austrian Tunnelling Method (sections I and II of route B) is the most appropriate tunnelling criteria along the selected tunnel route.


Introduction
Route selection is the critical initial step in the process of design and construction of a tunnel. In planning a suitable tunnel route, the geological, geotechnical, hydrological, hydro geological, socio-environmental, economic and technical factors need to be taken into consideration. However, these factors are heavily diversified and hence the planning and decision making process becomes complicated. Internationally, it has been practiced and proven that Analytical Hierarchy Process (AHP) can be used effectively in finding solutions to such complicated decision making problems and even the adoptability to tunneling industry has been proven [1].
Similarly, selection of tunneling method is also critical and is a complex decision making process in tunnel engineering. The selected tunneling method should ensure mine safety with least hazard levels and it should also be technically, economically and socioenvironmentally sound enough.

1.1
Scope and Objectives of the Study This study was aimed at selecting the most suitable tunnel route and tunneling method for BHP.
This study is focused on a part of the head raise tunnel of the Broadlands Hydropower Project (BHP), which is one of the major ongoing hydropower construction projects in Sri Lanka. It is located in the Central Highlands; approximately ninety kilometers east of Colombo city. This will be the last hydropower project of the Kelani River basin. The considered part of the tunnel has two proposed route alternatives, Route A and Route B and the lengths are 1.2 km and 1.5 km respectively. The specific objectives of the study are to,

Methodology
The following methodology was adopted in order to achieve the above mentioned objectives.

Data Collection
Data collection was performed in order to obtain maximum amount of relevant engineering geological information of the area such as rock mass characteristics and ground water conditions [2] as the accuracy of the decision totally depends on the reliability and adequacy of the data collected from both literature and field.
Number of field data collections were performed to collect necessary data using exposed rock outcrops.

2.1.1
Desk Study Initial study was performed on the case histories in applying AHP for projects of similar nature [1], [3], [4] and [5]. Geology and the Geomorphology of the area was studied using 1:50,000 scale topography and geology maps with contours at 10 m intervals prepared by the Survey Department of Sri Lanka and the geological maps of the study area developed by JICA at the feasibility study stage of the project [6].

2.1.2
Field Data Collection Rock texture, discontinuity orientations and their conditions were obtained from previously carried out borehole logs, rock outcrops at the tunnel trace as well as from the walls of partly excavated adit of tunnel route A.
Considerable amount of socio-environmental information were gathered through the discussions made with the experts engaged in the project and with the village community.

Engineering Geological Assessment Comprehensive
Engineering Geological Assessment for the area was carried out to study the engineering geological behaviour of the two tunnel routes. The assessment was initiated with the development of engineering geological maps for the particular area and subsequently the longitudinal cross sections along the tunnel routes. This was further extended by performing rock mass classification and stability analysis against possible rock slope failures along tunnel walls. The results are summarized in Annexure 01.

2.2.1
Rock Mass Classification. The rock mass classification was performed to examine the properties of the rock mass in a very systematic manner and classify rock mass in to different rock classes. This can effectively combine the findings from observations, experience and engineering judgment in providing a quantitative assessment of rock mass conditions. The rock mass classification was performed along two tunnel routes using Rock Mass Rating (RMR) [7] and Rock Quality Index (Q method) [8] to examine the quality of the surrounding rock masses in a systematic manner and to classify rock mass into different rock classes. Sample calculation has been included in Annexure 02 and the results are summarized in Tables 01 to 04.

Rock Mass Rating (RMR)
Rock mass along each tunnel route was sectioned, based on the engineering geological characteristics and then a rock mass condition assessment was performed for each section using rock mass rating system for parameters given below; I. Uniaxial compressive strength of rock material. II.
Spacing of discontinuities. IV.
Condition of discontinuities. V.
Orientation of discontinuities. Where; Li -Length of the specific tunnel section Qi -Q value of the specific section L -Total length of the tunnel route

Rock Wedge Stability Analysis
Failures along discontinuities, such as faults, bedding planes and joints are common during underground excavations in hard rock masses [9]. In this regard, stereonet analysis, which is a graphical technique [9], was performed considering only the major joint sets in the particular sections. Section 3.5 presents the results of the rock wedge stability analysis.

2.2.3
Tunnelling Cost Analysis The tunnel construction cost per meter run of the two route options were evaluated for the purpose of tunnel construction cost comparison. The construction cost includes excavation and supporting cost. Data for cost items were obtained from Kukuleganga and Upper Kothmale Hydropower Projects in Sri Lanka [10]. The results are summarized in Table 06 and 07 .

Applying Analytic Hierarchy Process
(AHP) There are many factors involved in the selection of tunnel route and tunneling method. Therefore, a multiple factor decision making approach was required for this analysis. In this regard, the AHP method, which was introduced by Prof. Thomas L. Satty in 1980 was adopted [11].
After identifying the problem, selection of criterion was carried out based on factors which are explicit in the problem stated.
Then the possible alternative solutions for the problem were identified and the solution process was proceeded to select the most appropriate alternative [11]. Figure 01 shows the structure of hierarchy for above three steps.

Figure 01 -Structure of Hierarchy for AHP
In selecting optimal tunnel route, the identified main criterion were: surrounding rock type, rock slope stability, tunnel overburden, tunnel geometry, geomorphology, tunnel construction safety, socio-environmental impacts, tunneling cost (drilling, blasting, supporting etc), construction duration and length of the tunnel route. The alternatives were Route A and Route B.
To identify the main criterion to be incorporated in the AHP model, the importance of above mentioned factors were assessed. For this purpose, a survey was conducted among 23 experts from different functional levels in local tunneling industry.
In selecting the suitable tunneling method, the identified factors were rock mass properties, tunnel geometry, tunnel overburden, ability of mechanization, groundwater level and expected water inflow, tunnel construction safety, socio-environmental impacts, surface disturbances, time/cost considerations and local experience [1]. These factors were evaluated over three alternative tunneling methods namely drill and blast (D & B), cut and ENGINEER 4 cover (C & C), and New Austrian Tunneling Method (NATM). The Tunnel Boring Machine (TBM) option was not considered since it is uneconomical for short length tunnels.
For the assessment, a questionnaire with a fivepoint scale for each of the criterion had been prepared. The respondents were asked to rate each factor according to a five-point scale, based on the priority that should be placed for each factor.
A weight had been assigned to each alternative through a pair-wise comparison based on the studied geological, geotechnical and other relevant conditions. The pair-wise comparison matrices were made with the aid of the scale of relative importance defined by Satty, 2008 [11]. The relative priorities of alternatives were obtained by this comparison and hence the relative suitability of the available alternatives to serve the objective was defined.

Specimen calculation
Step 1: Pair-wise comparison of route A and B against "Surrounding rock conditions" According to rock mass classification, the average importance of route A over route B in terms RMR and Q values is nearly 25%. It is almost 1/4 th of the scale. Therefore, surrounding rock type of route A is weakly important than route B. Referring to the table of relative importance defined by Satty, the intensity of importance is 3. This is marked on the Satty's scale of relative importance as follows.
Step 2: Build up pair wise comparison matrix According to the rules of AHP, the built up pair wise comparison matrix is as follows.
The normalized principal Eigen vector (also called priority vector) (w) is calculated and given by; The priority vector provides relative weights of the two route alternatives against "surrounding rock conditions". The weight of Route A is 75% and the weight of route B is 25%.
Step 3: Check for consistency When many pair wise comparisons are performed, some inconsistencies may typically arise. Therefore consistency of judgment is checked as follows.

…. (4)
Where CI is the consistency index, n is the size of the matrix and max is the principal Eigen value. According to AHP theory, if λMax = n, then the judgments have turned out to be consistent.
The composite weight of each alternative against all selected criterion can be determined using the decision matrix. Hence the relative suitability of the available alternatives is defined to serve for the objective.
The AHP method was applied twice separately to serve two initial objectives. First, to select the optimal tunnel route and then to select the most suitable tunneling method. The composite weights derived are summarized in Tables 08, 09 and 10.

Geographical Condition of the Area
The Broadlands Hydropower Project area is located in the central highlands; approximately ninety kilometres east of Colombo city. The area is widely underlain by Precambrian Gneiss. Two tributaries of the Kelani River, Maskeli Oya and Kehelgamu Oya join each other immediately downstream of the proposed dam site [6].
Kelani River forms a deep valley at the project area. The river bed elevation in this area is approximately 100 m from mean sea level. The slopes on both banks rise to ridges of more than 500 m elevation.
For this study, only the area marked in green colour dotted line in Figure 02, which enclosed the selected tunnel route alternatives up to common main tunnel, was considered.

Engineering Geological Mapping
Engineering geological map of the particular site under study was developed as depicted in Figure 03 and exhibits the information of existing rock types, the arrangement of different rock beds below the ground, the details of existing structural geological features of the area and the various topographical features.

Geological Subsurface Profiles along
Tunnel Routes Based on the available borehole logs and geological maps, sub surface profiles along the tunnel routes were developed. However, it should be emphasized that, most of the performed boreholes were not exactly aligned on the proposed tunnel routes and hence the sub surface conditions were interpreted using the nearest borehole data to come up with most reasonable geological sub surface profiles and presented in Figures 03 and 04.

Results of Rock Mass Classification
Relevant RMR and Q charts are provided in Annexure 03.

Stereonet Analysis Results of Rock Joints
The stereonet plots as shown in Figure 05 to 07 were developed based on the major joint sets in the area, listed in Table 05. Individual plane failures are possible along all three joint sets. Out of them, most vulnerable is joint set 1 and least is joint set 3, which has the least dip. Wedge failures are possible along the combination of joint sets 1, 2 and 1, 3. The most vulnerable of them is the wedging along the intersection of joint sets 1, 2 for both routes.

Cost Analysis for Tunnel Route Alternatives
Based on the engineering geological assessment, the applicable excavation and tunnelling methods were identified.
Considering the outcome, the excavation and supporting cost per meter run of the two alternatives were estimated and summarized in Tables 06 and 07.

Selection of Optimal Tunnel Route
As described in section 2.3, the ten criteria were prioritized using the results obtained from the questionnaire survey. Figure 08 depicts the priority chart of criterion. Then the pair-wise incorporated in the assessment of best tunnelling method. The priority chart for the criterion is presented in Figure 09.
The tunnel route was divided into following three sections based on the engineering geological characteristics;   Table 08.

Figure 08 -Criterion Priority Chart for Optimal Tunnel Route Selection
From the overall composite ranking of the two routes, it can be concluded that the tunnel route B, having a rating of 0.513 (51.3%) is preferred over route A. As depicted in Tables 09 to 11, the results of AHP reveals that, NATM is preferred for section A and B, and Drill and Blast method is preferred for section C.

Conclusions
The study was limited to the tunnel routes, A, chainage 0-1200m and for route B, chainage 0-1500m, from that point onwards, both the tunnel routes follow a common path.
According to the geological cross sectional profiles developed along tunnel routes, considered portion of the tunnel route A passes through charnokitic gneiss, marble and biotite gneiss and tunnel route B passes through charnokitic gneiss, silty sand and biotite gneiss. Top overburden part of the profile along both the routes are covered by rock blocks, boulders and soils of gravel silty clay and silty sand. The engineering geological assessment revealed that the surrounding rock masses of both tunnel routes belong to the poor rock class.
The overall RMR value of the tunnel route A is greater than that of the tunnel route B and overall Q value of tunnel route A is much greater than tunnel route B.
Both RMR and Q values of tunnel route A is much closer to the fair rock class. Therefore, the quality of the surrounding rock mass of the tunnel route A is considerably preferable than that of tunnel route B.
From the performed stereonet analysis for the major joint sets of the area, the rock wedge failures are possible along both tunnel routes.
The cost analysis revealed that tunnel route B is cost effective than tunnel route A. Per meter cost difference between two tunnel routes is approximately 21%.
For the considered conditions and criterion, the final result of the Analytical Hierarchy Process (AHP) for tunnel route selections revealed that the route B is the most optimal tunnel route.
According to the AHP analysis performed for tunnelling technique selection for tunnel route B, chainage 0 to 940 m, tunnel excavation has to be carried out using NATM and chainage 900 to 1500 m, the tunnel excavation has to be performed using Drill & Blast method. Therefore it can be concluded that a combination of Drill & Blast and New Austrian Tunnelling method should be applied for the tunnel route B.   As there is no borehole after 620 m chainage, it is assumed the geological condition is same as 508-620 section. Tunnel route pass through gray ~ dark gray biotite gneiss, attitude is N10°Wand SW∠ 40°. According to stereo net analysis based on three major joint sets in this region, surrounding rock is unstable and may generate unstable blocks and wedges.

850-900
As there is no borehole after 620 m chainage, it is assumed the geological condition is same as 508-620 section. But due to the fracture F4, the conditions may be slight differ. Tunnel route pass through gray ~ dark gray biotite gneiss, attitude is N10°Wand SW∠ 40°. From the stereo net analysis based on three major joint sets in this region, surrounding rock is said to be unstable and may generate unstable blocks and wedges. According to obtained RMR and Q values, the surrounding rock type is "Fair".

900-1200
As there is no borehole after 620 m chainage, it is assumed the geological condition is same as 508-620 section. Tunnel route pass through gray ~ dark gray biotite gneiss, attitude is N10°Wand SW∠ 40°. From the stereo net analysis based on three major joint sets in this region, surrounding rock is said to be unstable and may generate unstable blocks and wedges. According to obtained RMR and Q values, the surrounding rock type is "Fair". It should be noted that above table has not had a major revision since 1973. In many mining and civil engineering applications, steel fiber reinforced shotcrete may be considered in place of wire mesh and shotcrete.

Estimated support categories based on the tunneling quality index Q (After Grimstad and Barton, 1993)
ENGINEER -Vol. LI, No. 04, pp. [page range], 2018

Introduction
When planning a large hydropower project at a selected location, it is important to decide whether to construct a reservoir type or Run of the River (ROR) type based on an economic comparison including environmental and other considerations. There are environmental, social and economic impacts of hydropower projects which cannot be neglected, especially for large hydropower plants.
Owing to the severity of impacts predicted to be caused by climate change, GHG emissions need to be considered even for hydropower because of CO2 and CH4 emissions associated with large reservoirs. Disturbance to the ecosystem or loss of ecosystem is a negative impact to the environment including deforestation, impacts on fish owing to flow reduction and aesthetic effects on waterfalls. Reduced flow across waterfalls may cause reductions in tourism. Resettlement is a critical social impact caused by large hydropower projects. The economic advantage of reservoir type is the ability to store water and hence to be dispatched during both wet and dry seasons. Dry seasons affect the river flow of ROR type and hence the electricity generation.
The objectives of this research are to estimate the quantifiable advantages and disadvantages of reservoir type and ROR type hydropower generation, and to assess the economic impacts of converting reservoir type hydropower projects to ROR type hydropower projects at the design stage. In addition to costs related to power plants, costs related to environmental, social and economic impacts of hydropower are considered for project comparison.
Almost all the major hydropower potential has been developed in Sri Lanka by now. Out of the two existing large ROR type hydropower plants, Upper Kotmale Hydropower Project (UKHP) was taken as the case study for this research. However, as there are other countries with untapped potential of hydropower, this research outcomes will especially be important to them.

Literature Review
Worldwide research publications on comparison between reservoir type and ROR type hydropower plants were studied to identify the research gap.
A study [1] has been conducted focusing on the Amazonian regions of Bolivia, Brazil, Colombia, Ecuador and Peru, where a large untapped hydropower potential is available, qualitatively comparing the reservoir type and ROR type based on the climate change impacts. The study concludes that it will be necessary to invest in reservoirs to increase the margin of reserve and cope with climate change. The study also indicates the local, social and environmental impacts associated with the exploitation of hydropower. Another study [2] has also been conducted based on the climate change impacts of hydropower, focusing on Central and South American regions where 60% of the electricity demand is met through hydropower. Building new storage reservoirs is given in it as a potential adaptation measure.
A review [3] done for Yunnan in China, qualitatively compares small and large hydropower projects for their environmental implications and socio-economic consequences. A comparison [4] between large and small hydropower projects in Tibet, based on the CO2 equivalent states that small hydro performs better in terms of environmentally friendly development and low carbon energy than large hydro. In Tibet, large hydropower plants are essential to meet the large and growing electricity demand.
A study [5] from Western Himalayan region of India on environmental sustainability of ROR hydropower projects has been conducted. It presents a public perception cum data collection study on environmental impacts of small and large ROR hydropower projects. It concludes that every environmental impact of small hydropower is not "small" as compared to large hydropower and ignoring environmental impacts of small hydropower may not be a good practice in the Himalayan region. A case study [6] has been conducted for Uma Oya hydropower project in Sri Lanka incorporating economic and socio-environmental considerations into project assessment but GHG emissions from reservoir has not been considered in that.
Findings in the last two decades indicate that hydropower reservoirs produce GHGs as CO2 and CH4, raising the question whether hydro power based generation is a clean and green electricity source [7], [8]. Most of the past world studies [9], [10], [11], [12], [13] focused on GHG emissions from hydropower reservoirs due to flooded organic matter decaying under water and the quantifications were based on long term field measurements. The results were summarized for tropical and non-tropical regions separately.
Latest research findings on methods of GHG emissions from hydropower and quantification of them were studied because it is an emerging study area at present. The study done by A.B. Hidrovo et al., "Accounting for GHG net reservoir emissions of hydropower in Ecuador" [8], presents a rough and holistic estimate of net annual GHG emissions from reservoirs in the absence of long term field measurements. It is difficult to estimate the costs related to all ecological impacts of hydropower projects and their research covered only deforestation.

Case Study
There are many reservoir type large hydropower projects in Sri Lanka [14] but only two large ROR type projects, namely Upper Kotmale Hydropower Project (UKHP) (150MW) and Kukule Ganga Hydropower Project (70MW) are in operation at present. These two were constructed during the recent past, compared to the time period in which the other reservoir type projects were implemented in Sri Lanka. Out of the above two, UKHP was selected for the case study in this research.
For UKHP, an earlier planned (1985)(1986)(1987) Caledonia reservoir type project and the already implemented Talawakele ROR type project are taken for comparison. These two sites are located at Kotmale Oya in Nuwara Eliya District, Central Province, Sri Lanka. The distance between the two locations is about 10km. Table 1 gives the salient features of these two projects. The estimated cost of Caledonia reservoir project is given in 1986 prices. This price to be adjusted to 2014 prices for the economic comparison under this study and adjusted project cost is shown in Table 11. The original proposal of Talawakele ROR had suggested to take tributary diversions to the main stream (Kotmale Oya) in order to get a larger catchment area. Accordingly, the annual energy generation was originally estimated as 512.00 GWh in the Environmental Study in 1995, but five waterfalls were to be impacted due to that proposal. Therefore, all the tributary diversions were cancelled. That resulted in the reduction of the annual energy to 409.00 GWh in the implemented Talawakele ROR project.

GHG Emissions from Hydropower Projects
In reservoirs/ponds, there are decomposed organic matter (flooded and upstream organic matter). These organic matters produce CO2 and CH4 which reach the water surface layer and release to the atmosphere by diffusion. CH4 is also released by bubbling. These bubbles are produced in the methanogenesis process. Less CO2 bubbles are also produced because CO2 has a higher solubility than CH4. GHG are again released to the atmosphere by degasification when the water passes through the spillways of the dam and turbines. This is due to the change of temperature, pressure and turbulence. GHG are also released to the atmosphere by diffusion in the downstream river. The previously generated turbulence helps the gases to be easily diffused to the air. These emissions can be precisely determined by long term field measurements. In the absence of long term field measurements, a complete, rough and holistic estimate of net GHG emissions from hydropower projects per year (En) can be obtained from the Equation 1 [8].

Estimation of Emissions from Loss of Ecosystem
The formula of photosynthesis and respiration [8], [16] was applied to estimate Ee. This is given in Equation 2.

CO2(264g) + H2O(108g)
The particular element called amylase is related to growth of dry matter in a plant. Dry matter weight of a plant type can be obtained from its Net Primary Production (NPP) data [8]. Table 2 gives typical NPP values for tropical forest and cultivated land [17].  Table 3 gives the land use prior to inundation for the two selected projects. Based on the information in Table 2 and Table  3, the loss of dry matter production, and hence the CO2 emissions per year of the two sites are estimated as follows. NPP of tea plantations was assumed to be that of the available cultivated land data in literature.
Where, Ef = Mean reservoir emission factor Ae = Area of reservoir/pond Following long term field measurements data, past studies [8], [19], [20] have shown that reservoir GHG emissions decay exponentially against reservoir life.
According to Zhang et al [18], it is almost impossible to determine this reservoir GHG emissions precisely in the absence of long term field measurements. Therefore, Zhang et al has applied directly a constant mean reservoir emission factor for the total reservoir lifetime. It showed that, boreal regions have a significantly lower GHG emissions than tropical regions. The boreal region is defined as the zone having a definite winter with snow, and a short summer, generally hot [21].
Sri Lanka is a tropical country. Long term field measurements has not been done in Sri Lanka yet. Therefore, constant mean Ef of tropical regions was selected for this research for the total power plant lifetime as given in Equation 3.
Tropical Ef = 2,771.60 g CO2-eq m -2 yr -1 [18] . It should be noted that there may be differences between reservoir and ROR pond water behaviour. Water is always stored in the reservoir but in Talwakele pond, water is collected during the day time and used during the night peak hours. This may affect reservoir/pond GHG emission patterns, which have not been considered here.

Estimation of Emissions from Turbine, Spillway & Downstream River
According to Hidrovo et al [8], out of the total direct GHG emissions (Er and Etsd), 45% would come from the reservoir and 55% would come from turbine, spillway and downstream. Therefore, Etsd was determined by Equation 4 [8].

Estimation of Emissions from Construction, Operation & Maintenance
Life Cycle Impact Assessment (LCIA) is an accepted tool which allows to identify the potential environmental impacts associated with a product or service, throughout its entire lifespan [22]. A LCIA has been performed by Hidrovo et al [22] to determine Ecom.
For this research, LCIAs performed for reservoir type and ROR type hydropower projects were studied from literature. Ecom values of the two case study projects were estimated accordingly.

4.4.1
Ecom of Caledonia reservoir project Turconi et al [22] have critically reviewed case studies involving LCA of electricity generation from hydropower and found that Ecom was in the value range of 11-20 kg CO2-eq/MWh for dam reservoirs. Zhang et al [23] used LCA to compare two reservoir hydropower schemes, one with a concrete gravity dam (CGD) and the other with an earth-core rockfill dam (ECRD). They have found that the CGD scheme had a higher Ecom (11.11 kgCO2-eq/MWh) than ECRD.
Caledonia reservoir also has a CGD and hence the best available approximation of Ecom for Caledonia was made as 11.00 kgCO2-eq/MWh. Total Ecom= (11kgCO2-eq/MWh)×664×10 3 MWh = 7,304.00 ton CO2-eq/year [24] performed LCIA for ROR type hydropower projects and found that a ROR (plant life: 30 years, plant factor: 45%) with small reservoir had Ecom value as 11.00 kgCO2eq/MWh. The variation of that with plant life and plant factor is given in Table 5 and Table 6 [24]. Talawakele ROR has a plant life of 50 years. Therefore, Ecom value was approximated as 8.00 kgCO2-eq/MWh which was a reduction by 3.00 kgCO2-eq/MWh than the reference. Talawakele ROR has a plant factor of 30% and hence the values in Table 6 were plotted, as shown in Figure 1, to obtain Ecom at that plant factor.

Total Emissions
The net GHG emissions per year En of Caledonia reservoir project and Talawakele ROR project were calculated and tabulated as in Table 7. The specific GHG emissions have also been calculated and tabulated in Table 8. It can be seen that even the GHG emissions from reservoir type hydropower projects are negligible compared to that of thermal power projects where specific GHG emissions are several hundred grams per kWh.

Economic Comparison of Selected Hydropower Projects
When making a decision on selecting a project among several project alternatives, technical feasibility and economic viability are two important criteria to be considered in engineering projects. The technical feasibilities of the two selected projects for comparison were ensured from their feasibility study reports. The decision making has been based on their economic viability.

Levelized Cost of Electricity (LCOE)
The economic comparison between the two types of plants was carried out by determining LCOE for each case.

LCOE = Total Lifetime costs (LKR) …(5) Total Lifetime Energy Output (kWh)
The costs and benefits of the two types of plants were estimated as detailed below.

5.2
Loan Details Financing for the UKHP was provided by Japan International Cooperation Agency (JICA) and the details are shown in Table 9.

Project Cost and Loan Amount
The project cost and the loan amount of the implemented Talawakele ROR project were calculated from the annual cost distribution data during the construction period. The result summary is given in Table 10. Loan as a percentage from project cost (%) 89.00 Adjusted loan amount at the end of grace period (million LKR) 51,834 The project cost of Caledonia reservoir was taken from the Feasibility Study report [1]. The percentage of the original loan amount from the project cost was estimated as 89.00% to maintain the same conditions in both projects for comparison. The project cost summary for the Caledonia reservoir case is given in Table  11. For this study, loan repayment schedules of Talawakele ROR and Caledonia reservoir projects were considered during the 30 years of loan repayment period. The purpose was to obtain the cash outflow from capital repayments and interest payments (finance cost) in order to use for LCOE calculations.

Plant Operation and Maintenance Cost
Actual O&M costs of the implemented Talawakele ROR type hydropower plant were obtained for year 2014, 2015 and 2016. The plant was commissioned on 14 th July 2012 [25] and the O&M cost could be exactly obtained since 2014. Those annual O&M costs are given in Table 12. If UKHP was implemented under CDM, the GHG emission reduction compared to baseline emissions defined for CDM projects, could have been sold via carbon trading to earn an income. In this study, that was considered as benefits for both projects.
Grid Emission Factor (GEF) is a parameter to determine the baseline emissions for CDM projects in the renewable energy sector and waste heat/gas recovery sector. It refers to CO2 emission factor associated with each unit of electricity provided by an electricity system [27]

Levelized Cost of Electricity Calculation and Economic Comparison
With above costs and benefits, two cash flows were prepared and Net Present Value (NPV) of total costs and NPV of energy were calculated considering the lifetime of the plants. The results are given in Table 13. From Table 13, it can be seen that the LCOE from Talawakele ROR project is substantially lower compared to the Caledonia reservoir type project.

National Benefit
Reservoirs can store water and hence the expected operation during peak demand hours can be obtained in both dry seasons and wet seasons of a year. In ROR, the flowing water is collected and stored during the day time in dry seasons in order to operate during the night peak. The expected peak operation may not always be possible because the stored water amount may not be sufficient to cover the entire peak period which is from 18 According to the results shown in Case 1, even though the Talwakele ROR hydropower plant operates in its full capacity (assuming every year is a wet year), there is an annual loss of LKR 2,123 million to the country by not having the Caledonia reservoir plant for peak serving.
Case 2: based on the actual night peak operation data of UKHP obtained from System Control Centre (SCC) of Ceylon Electricity Board (CEB), annual financial loss, SCC daily records the maximum output (MW) of UKHP during the night peak operation and the time it occurred. In this study, it was assumed that the plant operated with that output for the total night peak period in the respective days. Accordingly, the generated night peak energy were calculated for each year from 2014 to 2016 as shown in Table 14.
The results of Case 1 and Case 2 are summarized and given in Table 15.

Conclusions and Recommendations
It is concluded that GHG emission considerations are not strong enough to discourage reservoir type new hydropower plant developments.
According to the LCOE results, it can be concluded that the ROR type has the overall economic benefit in the case of UKHP, but if the project objective is solely to capture the maximum hydropower potential or peak serving, it can be concluded that the reservoir type has a better overall economic benefit to the country.
It is recommended that for similar future large hydropower developments, a detailed similar study to be carried out before taking the decision on reservoir construction for hydropower generation. A case by case study is recommended to be conducted because the environmental factors such as the impact of flooding, melting snow and ice, etc. vary from location to location. The methodology presented in this study can be followed for such studies with suitable modifications where necessary.
Although hydropower projects do not have zero GHG emissions, as per this study, they are much less compared to thermal power plants. In view of grave need for limiting the average global temperature increase to 2 0 C by the end of this century and the GHG emission reduction commitments made at the COP21 at