Performance improvement of headworks: a case of Kalignadaki A Hydropweor Project through physical hydraulic modelling

Transcription

1 Performance improvement of headworks: a case of Kalignadaki A Hydropweor Project through physical hydraulic modelling Dr. Ing. Meg B. Bishwakarma General Manager, Hydro Lab Pvt. Ltd., Nepal ABSTRACT: The Kali Gandaki A Hydropower Plant (KGA) is a peaking run-of-river type of project with installed capacity of 144 MW. This is owned and being operated by Nepal Electricity Authority (NEA). Since its operation started in 2002, the headworks of this plant has been experiencing unfavourable hydraulics at the intake and has also got significant problems in trash handling. Due to the skewed flow distribution at the intake and very short transition between the intake and the settling basins, the flow pattern in the settling basins is unfavourable for settling of suspended sediment as it was designed for. This situation has caused excessive sediment induced turbine erosion problems since its first few years of operation far more than anticipated. Therefore, in order to improve the overall performance of the heawrosks including the flow patterns in the settling basins, a physical hydraulic model at 1:40 scale was built and tested at Hydro Lab in Nepal. At first the model study was carried out to observe the performance of the existing design of the headworks by reproducing it in the river model. After that a comprehensive study was conducted through several modifications and corresponding tests to achieve a satisfactory performance of the headworks in terms of improved approach flow condition to the intake, flow distribution among the intakes, flow patterns at the inlet to the settling basins, in the main basin and at the outlet. The most challenging task was to achieve acceptable performance by making minimum modification on the existing headworks, as this is an already constructed project. Moreover, the modified design should require minimum days for plant shutdown for its implementation, as this is the largest plant supplying power to the national grid. In this paper, the author will present how the satisfactory solution was achieved through physical hydraulic model tests keeping in view of the minimum work required for the design modification and thereby reducing the downtime of the plant for implementation. BACKGROUND: The function of headworks of any run of river hydropower project is to be able draw design discharge through its intake, safe passage of design flood, passage of sediment load, passage of floating debris and suspended sediment managed by the settling basins. The designers therefore put their utmost effort to take care of these parameters while planning and design of the headworks of hydropower/water resources projects. Normally, the design is verified through physical hydraulic model tests by simulating the likely scenarios in order to have the chance to make modifications if needed prior to its implementation. This helps to make the proposed design more mature before it is constructed at a particular setting. It is important to understand the fact that the headworks design is site specific as the river characteristics will not be similar for all projects. We cannot change the river to fit with our design however we can change our design attempting to fit with the river. In this paper, modification of the headworks of the 144 MW Kali Gandaki A Hydropower Plant (KGA) in Nepal is being presented. This plant is owned and being operated by Nepal Electricity Authority (NEA). This is a peaking run-of-river type of project located in Western Nepal. The project has been in operation since According to the operation experiences from site, the headworks has been experiencing problems in dealing with floating debris as well as suspended sediment management from time to time. This has resulted in increased sediment induced turbine erosion problems and associated effects since its first few years of operation far more than anticipated. The owner of this plant therefore intended to improve the performance of the power plant by doing some modifications to the headworks structures

2 among other issues. Accordingly, a physical hydraulic model study was carried out at Hydro Lab to improve the performance of the existing headworks through several modifications and corresponding tests. The model study project was financed by the World Bank through NEA. The existing headworks consists of river inlet (free over flow type), a retardation section called forebay, 6 intakes supplying water to two settling basins leading water to a collector channel, which is connected to the headrace tunnel. The intake and the river are separated by a forebay wall or sometimes referred to as flood wall. The headworks is operated at 518:00 masl during the monsoon season to minimise the sedimentation of the peaking reservoir whereas, it is operated at masl during the dry season to maximise the energy generation. The general arrangement of the heaworks is shown in Figure 1. Settling basins (only river side basin is seen) Kali Gandaki River Bascule gate Diversion dam and spillway gates Intake gates Forebay wall River inlet Access road to dam site Figure 1: Existing headworks of Kali Gandaki A HPP during wet season operation The salient features of the power plant are given below: Power plant type : Peaking Run-of-river Design head : 115 m Design flow 141 m 3 /s Installed capacity : 144 MW Average annual estimated energy : 842 GWh Turbine : 3 Francis units of 48 MW each, vertical axis Penstock : 243 m long, 5.25 m diameter, steel lined Intake : 6 Nos. 10m x 10.63m (WxH ) each Spillway : 3 radial gates of 15m x 19m (WxH) each Settling basins : 2 Nos. 187m x 80m x 14m (LxBxH) each; gravity flushing type Start of project construction : 1997 Commercial operation : 16 August 2002 Project financed by : ADB/JBIC/Nepal Government/NEA Model study conducted : Model study funded by : The World Bank through NEA

3 OBJECTIVE OF THE STUDY: The main objective of this study was to come up with a satisfactory performance of the headworks in terms of flow pattern at the intake, at the inlet and outlet of the settling basins in order to manage the floating debris and suspended sediment more effectively as intended in the design through physical hydraulic model study. The following are the sub-objectives of this study in order to achieve the main objective; to study the design, drawings and performance of the existing headworks and formulate possible remedial measures. to review the hydraulic design of the main components of headworks with respect to operation. to build a physical hydraulic model and conduct testing of the existing headworks, without and with modifications, and make recommendation for improvement of the headworks performance. to display the problems and shortcomings of the hydraulic design so that modified solutions can be explored and optimized with simple modification in the model with cost as low as possible. to provide insight and visual performance to facilitate realistic decisions on the design and recommended modification. REVIEW OF AVAILABLE DOCUMENTS: Documents and information of the power plant made available by the Client, collected during the site visits and from other sources have been reviewed by the Consultant. The Consultant identified the following three different stages in the history of the project which are relevant when it comes to an assessment of the performance of the plant: As model tested and reported in the Final Report: Model Study of Kali Gandaki A Hydroelectric Project by SINTEF-NHL in May As designed at the end of the detailed design phase and reported in the Final Project Formulation Report II, by KGAA in May 1994 and confirmed / modified in the Final Engineering Design Report by KGAA in January As built and reported in the Project Completion Report by KGAA in January The headworks structures arrangements for above three stages are quite similar when it comes to the dam and the peaking reservoir, but they are completely different when it comes to the intake and the resulting hydraulic performance of the settling basins. Based on the preliminary review, the following main areas for improving performance of the headworks were identified: Improvement in the intake hydraulics which in turn help improve the performance of the settling basins with respect to their ability to trap suspended sediment and thus reduce the amount of sediment which is passed on to the turbines. Improvement in the trash passage system will help in reducing the amount of trash entering in to the intake. Such improvement will help to improve the performance of the intake and settling basins and also to reduce the head-losses at the intake. Improvement in the inlet and outlet conditions of the settling basins to improve the flow pattern in the settling basins. As observed during the monsoon season operation, the flow pattern in the settling basins is not conducive for sediment settling primarily due to the flow pattern at the inlet caused by the turbulent and uneven flow distribution at the intake and the design of the inlet itself. A major task of the study therefore was to improve the upstream and downstream flow conditions so the actual trapping ability of the basins becomes as close as possible to the theoretical potential as envisaged in the original design.

4 SITE VISITS AND FIELD MEASUREMENTS During the study period, four site visits were carried out to observe the site condition as well as to conduct some field measurement. First site visit was made in September 2010 to observe the headworks performance during monsoon/high flow season and to collect different information needed for the model study. Second site visit was made in November 2010 to observe the operational regime of the headworks during dry/low flow season. Some personnel at the headworks site were also interviewed during this visit to collect information on operational aspects of the headworks and common problems being faced by the operators since the commencement of the plant. Third site visit was made on December The following field activities were performed during the site visit: Bathymetric survey of the reservoir and plunge pool areas. Installation of staff gauge downstream of the dam. Topographical survey of about 1.4 km Kali Gandaki River stretch downstream of the dam. Finally, the fourth site visit was made on August 2011 to measure different flow parameters inside the settling basins. Flow velocities, depth and discharge were measured at different cross sections inside the settling basins by using Acoustic Doppler Profiler (ADP). Besides this, dye tests were also performed in the forebay to observe the flow pattern during the monsoon operation. RIVER MODEL CONSTRUCTION The river model was constructed based on the geo-morphological data and other information made available during the initial stage of the model study. Froude s model law was used for the modelling and the model was built in 1:40 linear scale. The river model included the river stretch about 500 m downstream of the dam and the peaking reservoir area up to 800 m upstream of the dam including both Kali Gandaki and Andhi Khola rivers. Andhi Khola is a tributary to Kali Gandaki River having confluence about 500 m upstream of the dam site. HEADWORKS MODEL CONSTRUCTION The headworks structures model construction was carried out based on the drawings included in the Project Completion Report. Then the headworks structures model was placed in the river model in reference to the original coordinates and elevation. The model included all relevant components of the headworks for assessment of their hydraulic performance. The model over view of the existing headworks is given in Figure 2. 2 Nos. settling basins Forebay 6 Intakes Forebay wall River inlet weir 3 Nos. spillway gates Figure 2: General overview of model of the existing headworks of Kali Gandaki A HPP

5 PERFORMANCE TEST OF EXISTING HEADWORKS The model was initially tested for observing the performance of existing headworks structures for different flow situations (up to 5 years flood). The tested flow situations are given in Table 1. Table 1: Simulated flows for Kali Gandaki and Andhi Khola Rivers (m 3 /s) Description Flow at Dam site Kali Gandaki River Andhi Khola River Design flow Average monsoon flow (June October) year flood 1,350 1, in 2 years flood 2,260 1, in 5 years flood 2,880 2, The bed load estimates for the model study was based on the observed flow and suspended sediment concentration from In order to take account of the unmeasured sediment concentration, the values were increased by 20 % for total sediment estimates. The adopted bed load concentration was taken as 25 % of the suspended load as normally used in the study for Himalayan Rivers. In order to test the performance of the headworks for a given flow, corresponding bedload sediment and floating debris were introduced in the model at the upstream river. The flow patterns, water levels and velocities at different sections in the reservoir and the settling basins were studied and documented for different flow situations. Water levels in the model were directly measured with levelling instrument. Velocities in the river/ reservoir and settling basins were measured using a micro propeller and an ADV (Acoustic Doppler Velocimeter) respectively. Flow pattern and floating debris transport patterns were studied by using paper confetti and coloured dye in the model. DRABACKS OF THE EXISTING HEADWORKS The main drawbacks of the existing headworks arrangement is the adverse approach flow condition to the intake. The flow over the forebay inlet weir is concentrated at the right part only hitting directly to the trash rack of the leftmost intake. This concentrated and skewed flow created recirculating flows in both sides of the forebay. The observed drawbacks are illustrated in Figure 3 and Figure 4 respectively. Intake with trash rack Flow to the intake a) Flow pattern in front of the intake b) Flow condition at settling basin entrance Figure 3: Flow pattern at the intake (a) and at the inlet to the settling basin (b)

6 Trash handling of the existing headworks arrangement was also found to be unsatisfactory. The bascule gate provisioned for the trash passage was not found so effective for passing the trash even if it would have been operated as recommended in the original design. The floating debris and trash enter the forebay over the inlet weir and get stuck in the trash rack. There is no way out for the trash and must be raked out by the mechanical raker, which is also very difficult due to high velocity at the left most part of the intake. The clogging of trash rack results in loss of net head available for power generation. Highly turbulent flow a) Floating debris in front of the intake b) Turbulent flow in the settling basins Figure 4: Floating debris accumulated in front of the intake (a) and turbulent flow condition in the settling basin (b) The effect of the concentrated approach flow extends even inside the settling basins creating recirculation flow and formation of eddies in the basins. Submerged hydraulic jump type of phenomenon was observed at the inlet to the settling basins which creates additional turbulence primarily due to its design. The uneven distribution of flow into the settling basins and the recirculating flow and formation of eddies have reduced the sediment trapping efficiency of settling basins. MODIFICATIONS After testing the model for the existing headworks, several modifications were carried out to improve the hydraulic performance of the headworks focusing on the shorter downtime required for implementation in the prototype. Out of various modifications tested in the model, only some significant ones and the final recommended modifications are included in this paper. All of the modifications were tested mostly for the average monsoon flow situation, as this is the flow usually available during the monsoon (wet season), when the plant is in operation. Focus was also given to improve the flow distribution in the settling basins with less eddies formation and hence to improve the trapping efficiency of the basins. First modification labelled as Modification-A1 was proposed to provide six orifices through the existing forebay wall. Each orifice was 4 m high (EL to masl) and was oriented parallel to side walls of the intakes. The inlet weir was blocked to prevent the inflow to the forebay. The minimum reservoir operating level was maintained at masl. This modification is shown in Figure 5. After some trial runs with Modification-A1, the forebay wall was straightened and the orifice height was reduced to 3 m (EL to masl) which was labelled as Modification-A2. Furthermore, a ramp with 15 degrees slope was provided at the inlet of the settling basins. The hydraulic performance of the headworks after Modification-A2 was satisfactory and could have been fine-tuned by carrying out minor adjustments, but it would have required longer power plant shutdown time for its implementation. Therefore, another concept (Modification-B) was conceived and tested in the model with different minor adjustments.

7 Wall to block the flow from existing river inlet Intakes Settling basins Straightened forebay wall Orifices in the flood wall Existing forebay wall Orifices in the forebay wall a) Introducing orifices in the forebay wall b) Straightening the forebay wall with orifices Figure 5: Modification of headworks labelled as Modification A1 with series of orifices along the forebay wall for smooth withdrawal of water to the intake Modification-B1 comprised of a siphon type forebay inlet arrangement which draws water over the existing forebay wall at an elevation of masl. The minimum reservoir operating level was proposed at masl. The forebay was blocked by a wall at the left of the inlet arrangement to avoid the inflow from the inlet weir. In Modification-B2, some further modifications were done over the Modification-B1 and tested in the model. This included installation of flow deflectors at the inlet of the settling basins to substitute the 15 degrees ramp. The modified headworks arrangement showed satisfactory hydraulic performance and thus the Modification-B2 was recommended for further development. The proposed arrangement for the Modifications B2 is shown in Figure 6. Intakes Settling basins Flow to the intakes Flow deflectors at settling basin inlet a) Modification viewing from downstream b) Modification viewing from upstream Figure 6: Modification of headworks labelled as Modification B2 with inlet structure on top of the flood wall and introduction of flow deflectors at the inlet to the settling basins Later, it was proposed to test the recommended modifications for reservoir operating levels lower than masl. The Client would prefer to have the improved headworks performance when operating it at masl during wet season, as this is the originally designed level. It was therefore agreed to carry out further tests of the recommended modifications keeping the minimum reservoir operation level at masl during the operation in monsoon season, which was labelled as Moddification-C1. In modification-c1, the existing flood wall was trimmed down by 3 meters and a forebay inlet arrangement similar to modification-b2 was provided. The water level inside the settling basins was maintained at masl during wet season operation. This modification also included throttling of the

8 intake gates and provision for flow deflectors at the inlet of the settling basins. Hence the shape, size and location of the deflectors were optimised to achieve the desired hydraulic performance of the Modification-B2. The model tests demonstrated a satisfactory hydraulic performance of the modified headworks. COLLECTOR CHANNEL GATES OPENING Initially the model tests for Modification-B2 were carried out with collector channel gate openings as normally practiced at site. Then discharge through each gate was calculated with the help of head losses over the gates measured in the model. It was observed that most of the flow into the collector channel was drawn by gate no. 1 and gate no. 7 creating skewed flow distribution inside the basins. To minimize the skewness, the model tests were also carried out for different openings of the 12 collector channel gates at downstream of the settling basins to ensure even distribution of flow along the cross-section of the basins. Different gate openings were worked out in the model aided by theoretical calculations. While finalising the optimum opening for each of the 12 gates, a criterion was set to pass equal discharge through the individual gate as well as through both basins in sum. The recommended gate openings for 12 collector channel gates for normal operation of the power plant during wet season are shown in Table 2. Table 2: Recommended collector channel gate openings for normal operation of the power plant (for the design discharge of 141 m 3 /s) Settling basin Left Right CC Gate no Opening, m Note: The gates are numbered from upstream to downstream i.e. CC Gate no. 1 is the rightmost gate and CC Gate no. 12 is the leftmost gate. Several tests were carried out in order to define optimum opening of each of the gate for passing different discharges including the design discharge through the existing headworks (intake and the settling basins) without any modifications. The flows tested were 141 m 3 /s, 94 m 3 /s, 70.5 m 3 /s and 47 m 3 /s which represent design discharge, 2/3 of design discharge, 1/2 of design discharge and 1/3 of design discharge. The results are found to be satisfactory and it was recommended that the above mentioned collector channel gate openings should be adopted in practice starting from this monsoon as an immediate measure. It will help reducing the skewness in flow distribution inside the settling basins by some extent which will obviously improve the performance of the basins. RECOMMENDED HYDRAULIC DESIGN Two alternatives labelled as Modification-B2 and Modification-C1 were recommended for implementation based on the outcome from the study. Modification-B2 requires the minimum reservoir operation level to be raised by 3 meters but the advantage of this modification is that it will take shorter shutdown period of the power plant for its implementation, as the inlet proposed to be constructed could be fabricated outside and hooked up on top of the existing forebay wall at level. However, in Modification-C1, there is no need to raise the minimum reservoir operation level significantly but it will take longer shutdown period of the power plant for implementation. Accordingly, final documentation studies of both of the alternatives were performed to record the performance of the headworks. Brief descriptions of the modified headworks components for both alternatives are given below: Modification-B2 Forebay inlet arrangement A siphon type forebay inlet arrangement was proposed on the existing forebay wall. The top level of the forebay inlet arrangement will be at masl and the minimum reservoir operating level should be maintained at masl. The forebay is proposed to be blocked at the left of the forebay inlet

9 arrangement by a wall, which is parallel to the intake side wall, with its top elevation at masl. This prevents the flow over the inlet weir to enter the forebay and ensures smooth approach flow to the intake during monsoon season operation. Floating debris passage The screen wall of the forebay inlet arrangement extends up to the elevation of masl and prevents the passage of floating debris and trash into the forebay. To provide the free passage of the floating debris and trash to downstream of the dam, the bascule gate located at the leftmost spillway gate is proposed to be operated together with the spillway radial gate. Settling basins To improve the inlet flow pattern inside the settling basins, an arrangement of flow deflectors were made to be provided at the inlet to the basins. These structures help distribution of flow in the vertical plane of the settling basins and help to reduce the phenomena of submerged hydraulic jump at the inlet thus improve flow pattern significantly. Since the minimum reservoir operating level is proposed to be raised by 3 m in the recommended modification, the divide wall between two settling basins need to be raised by 2 m. This will prevent inter-basin flows over the divide wall and ensures independent operation of the basins required during flushing of the basins. In order to achieve more uniform flow across both settling basins, the openings of the collector channel gates was optimised for the operation level of masl. The recommended openings were determined based on the findings of several tests in the model combined with the numerical computation. Modification-C1 Forebay inlet arrangement The existing forebay wall is proposed to be trimmed down by 3 meters and a siphon type forebay inlet arrangement similar to modification-b2 is proposed to be installed on top of the wall. The top level of the forebay inlet arrangement will be at masl and the water level inside the settling basins is maintained at masl. Similar to modification-b2, the forebay is proposed to be blocked at the left of the forebay inlet arrangement by a wall, which is parallel to the intake side wall, with its top elevation at masl. This prevents the flow over the inlet weir to enter the forebay and ensures smooth approach flow to the intake. Floating debris passage The screen wall of the forebay inlet arrangement extends up to the elevation of masl and prevents the passage of floating debris and trash into the forebay. To provide the free passage of the floating debris and trash to downstream of the dam, the bascule gate located at the leftmost spillway gate is proposed to be operated together with the spillway radial gate. Settling basins Flow deflectors at inlet of the settling basins are proposed for this modification as well to improve the inlet flow pattern inside the settling basins. In order to have more uniform type flow across both settling basins, recommendation is made for collector channel gate openings to ensure uniform flow distribution through the gates.

10 CONCLUSIONS AND RECOMMENDATIONS On the basis of overall performances of the headworks with recommended modifications studied in the model, the following conclusions and recommendations are made. All the available relevant documents were collected and reviewed during the study. Field visits were made to observe the performance of the existing headworks and conducted required measurements. The physical hydraulic model of the KGA headworks was constructed based on the data and drawings provided by the Client and tested it successfully. After testing the performance of the existing headworks, several modifications were made to the existing headworks and model tested for performance improvement focusing more on the intake hydraulics, handling of floating debris and the flow patterns in the settling basins. Two alternatives, labelled as Modification-B2 and Modification-C1, were recommended for the modifications of the existing headworks to improve its hydraulic performance. Based on the observed performance of the headworks, any one alternative could be implemented to achieve the objective. Implementation of Modification-B2 will require shorter shutdown period of the power plant than it is required for Modification-C1. However, minimum reservoir operation level has to be raised by 3 m for Modification-B2 whereas there is no need to raise the minimum reservoir operation level in case of Modification-C1. The maximum reservoir operation level is at masl for both of the recommended alternatives. Both recommended alternatives were observed to be performing well in handling floating debris and trash in the model. The bascule gate should be operated along with the spillway gates to facilitate free passage of floating debris and trashes to downstream of the dam. The model study was mainly focused on the improvement of the hydraulic performance of the headworks structures and thus, it is recommended to carry out necessary structural and stability analysis for the different components of the headworks before implementation of the recommended modifications to the existing headworks. It is recommended to maintain the collector channel gates opening as calculated based on the model tests results starting from this monsoon season so that the unequal flow distribution in the basins could be minimised. In order to reduce the sediment induced effects to the turbines, plant operation measures must be implemented together with an optimized maintenance programme for the hydraulic machinery in the power plant in addition to the suggested design modifications of the headworks based on this study. ACKNOWLEDGEMENT The author is grateful to Nepal Electricity Authority and the World Bank for providing funds, required information and data for this study. The author is also equally grateful to the involved team from Hydro Lab and the Scientific Advisor, Dr. Haakon Stole from Sediment Systems AS Norway, for their valuable contribution in this research. The materials obtained from the following references have been extremely helpful. REFERENCES Morrison Knudsen International. Inc., USA; NORCONSULT International A. S.; Norway; IVO International Ltd., Finland,2002; Project Completion Report, Volume V B As Built Civil Construction Drawings, Lot C1 Headworks Facilities. Norwegian Hydrotechnical Laboratory (NHL), SINTEF, 1995; Final Report: Model Study of Kali Gandaki A Hydroelectric Project; Hydraulic Model Study, Volume I: General Studies. Norwegian Hydrotechnical Laboratory (NHL), SINTEF, 1994; Final Report: Sedimentology, Kali Gandaki A Hydroelectric Project. Hydro Lab Pvt. Ltd., Nepal, 2012: Final Report; Physical hydraulic model model study of Kali Gandaki A Hydroelectric Project.