Abstract

Reservoir management is an intricate task and the calculator proposed herein helps in assessing the storage in the reservoir using the elevation level. Based on the empirical values of storage capacity and levels the code developed here is able to generate bi-directional query by returning closest approximation corresponding to the given input. The closest approximation is heavily dependent on the resolution level of data for capacity versus elevation that is provided to the program. The efficacy of this code is tested on the data from Bhakra Dam. The code correlates the water level with the volume of water available in the reservoir. This precise assessment of available water will aid in the management of issues related to availability of water for agriculture, industry, hydropower, and domestic use among others. When the inflow and release volume is provided to the code as an input, it calculates the current water level and storage in dam. The output provided by the code can be utilized for effective operation of reservoir systems. The Bhakra Nangal and Beas project is a lifeline of the Northern states and its optimal operation by continuous monitoring of the available water will contribute to its better management.

1. Introduction

Efficient operation of reservoirs is a challenging task due to stochastic nature of inflow. The objective of the present research is to develop a code that can be used for the operation of reservoirs in an efficient manner. The practicality of the approach developed herein is demonstrated through application to Bhakra reservoir located on Satluj River Basin – a key basin in the Himalayan region. The Satluj river originates from Mansarowar lake situated in Tibet at an elevation of about 4572 m and extends from 30° N to 33° N latitudes and 76°E to 83°E longitudes. It is one of the major rivers of the Indus system. In India Satluj river enters from Shipkila situated in Himachal Pradesh and flows in the South Westerly direction. The length of the river is approximately 1448 km. From Himachal Pradesh the Satluj river enters into the plain region at Bhakra located in Punjab. India's highest gravity dam is located at Bhakra situated in Punjab. It is a straight concrete gravity dam having 56,980 km² catchment area upstream of Bhakra dam. The height of the dam above the deepest foundation level is 225.55 m (740 ft). The height of the dam above river bed is 167.64 m (550ft). Govindsagar, the reservoir formed by the Bhakra Dam, has an area of 168.35km². The length of the reservoir is 96.56 km with a gross storage capacity of 9621 million cu.m, and live storage capacity of 7191 million cu.m. The minimum dead storage level to which the reservoir may be depleted is fixed at a mean sea elevation of 1462 ft (445.62 m) from irrigation and power generation considerations. The dam has ample spare capacity for silting during the expected life of the reservoir. The reservoir has an available capacity of 6 million acre-feet between the maximum reservoir level at 1690 ft (515.11 m MSL) and dead storage level at 1462 (445.62 m MSL) that can safely store the entire surplus flow of an average year.

2. Literature Review

Accurate estimation of reservoir level, storage, inflow and release are important parameters for deciding optimal operation policy of reservoirs. Reservoirs provide balance between the inflow, which are highly variable in time, and volume of water required to meet specific demand. In the past several procedures have been devised to estimate storage requirements. Reservoir inflow is estimated using conventional simple water balance equation. An analytical method based on simultaneous minimization of error in estimating reservoir water level and inflow variation was developed by ¹. A recursive filter based on state and observation equations was developed by G.Evensen ². ³ Estimated local inflows using the water balance equation and there after this methodology has been successfully applied in Ontario Hydro's new Energy and Resources Information System (ERIS) for the determination of daily local inflows. ⁴ estimated reservoir capacity of Dibang Multipurpose Dam on Dibang River, Arunachal Pradesh using residual mass curve technique. ⁵ developed a model which can estimate current storage capacity of reservoir. The model developed by him is essentially an ASC curve which is based on Python script. This tool uses ArcPy site package and it works with triangulated irregular network (TIN) model.

⁶ Amnatsan et al. developed a variation analogue method (VAM) for reservoir inflow forecasting which is able to capture the peak flows as rare extreme flows are crucial for effective reservoir management. The monthly inflows were also forecasted for Sirikit Dam, Thailand using wavelet artificial neural network (WANN) model, and the weighted mean analogue method (WMAM). However, better forecasts were obtained using the VAM model. The model was able to adequately capture extreme inflows at Sirikit Dam. ⁷ researcher used mass curve technique at Gizab multipurpose dam in Afghanistan to estimate maximum potential head and reservoir storage capacity. ⁸ Furnans, et al. described the utility of hydrographic survey. Authors have analyzed more than 100 bathymetric surveys of Texas reservoirs and estimated reservoir volumes and surface areas corresponding to reservoir stages. These surveys are also useful in estimating loss in capacity due to sedimentation over time.

⁹ Alrayess, et al. developed a model to determine required reservoir capacity to meet demand for a given annual inflows using mass curve. Reservoir capacity-yield-reliability relationships were investigated for the Sami Soydam Sandalcik reservoir in Turkey. Data of monthly and annual mean flow for the duration 1962-2013 was used.

¹⁰ Salih et al. used GIS to analyze the hydromorphometry of river basin. Based on 2D and 3D models of three sites. The maximum level, volume, surface area, circumference, shape factor of reservoirs is calculated. This experimental procedure can be used for any reservoir.

¹¹ EL-Hattab used bathymetric data to estimate the depth of channel. This depth will help to decide desired dredging and maintenance of channel in real time. For digital terrain modelling applications, it is concluded that a triangulated irregular network technique could be used. It is the fastest interpolation technique requiring only 0.35s to create a model with 5m grid size. ¹² Rodrigues et al. used remote sensing data to estimate storage capacity of small reservoir in the Sao Francisco, Limpopo, Bandama and Volta river basins. ¹³ Khattab et al. developed bathymetric map of Mosul Lake by using a digital elevation model (DEM). In many developing countries, Bathymetric maps are still not available for lakes and reservoirs.

¹⁴ explored the application of a deep learning algorithm, a recurrent neural network (RNN), long short-term memory (LSTM), and gated recurrent unit (GRU) to predict inflows and to do optimal reservoir operation for the Xiluodu (XLD) reservoir.

3. Appurtenances Used for Release of Water

The following facilities are available for discharging water to the downstream of the Bhakra Dam and the release can be estimated easily depending upon the appurtenances used for release.

Ÿ Penstocks

Ÿ River outlets

Ÿ Spillway

Out of these the discharge of water through the operating penstocks of the power plant is limited to the extent of power load coming on the turbine, while in the case of spillway the release is possible only when the reservoir is above the spillway crest level of 1645.211 ft.

Five penstocks are located on each side of the abutment. These are numbered from 1 to 5 on the left side, and 6 to 10 on the right side. The penstock are controlled automatically by the turbine wicket gates and for use in emergency. For the periodical shut down, head gates have been provided at the upstream end. Power output depends upon the:

Ÿ Difference in reservoir water elevation and tail water elevation.

Ÿ Discharge through one unit of the power plant.

Curves are prepared based on theoretical considerations and model results. Curves are modified according to actual performance of the units.

There are sixteen 96" (2438.4 mm) diameter outlets in the spillway portion of the dam located in two horizontal tiers of eight river outlets each, with centre lines at El. 1320 (402.34 m MSL) and EL 1420 (432.82 m MSL). These river outlets have been allotted numbers in a sequence from left to right, numbers 1 to 8 for El. 1420 (432.82 m MSL) tier of river outlets and 9 to 16 for El. 1320 (402.34 m MSL) tier of river outlets.

The overflow spillway has crest level at El. 1645.21 ft (501.46 m MSL) and each of the four spillways is provided with radial gates for controlling the flow of water and disposal of floods. The spillway is located in the central block numbering 18, 19, 20, 21 and 22 of the dam. A central training wall divides the overflow spillway into two compartments for repair facilities. Tables are prepared which indicate discharge capacities for the spillway for various elevations of the reservoir water level with one and more gate open. Tables that indicate the quantity of water that will be released during part opening of the gates have been prepared.

There are sixteen river outlets main facility for releasing the water for irrigation and power indents downstream of Bhakra Dam. These river outlets in addition to the spillway are used for releasing the flood discharge. For operating the river outlet, jet flow gates are used. The spillway together with outlet can discharge 290000 cusecs of flood water. Inside the dam 5 Km long galleries are provided at various elevation for foundation grouting, drainage and inspection of the dam. A 9.14 m wide road is provided at the top of the dam.

4. Reservoir Storage and Area Information

Table 1 indicates the volume of water in million acre feet available and area in thousand acres at different elevations of the reservoir. Using contour maps and area covered at particular contour and using Trapezoidal or Simpson’s rule, the reservoir storage capacity can be estimated. As per the present practice the capacity is calculated using the water level from the charts available. Curve plotted between the elevation in feet and capacity in million acre feet is shown in Figure 2. Another curve plotted between elevation in feet and area in thousand acres is shown in Figure 3.

Figure 4 indicates daily reservoir level, reservoir capacity in acre feet, inflow and release and Beas contribution. Beas contribution is water released from Dehar power plant. Table 2 indicates average monthly inflow and release for the duration 1999-2018.

Figure 5 indicates the reservoir minimum and maximum level in feet for duration 1975 -2019.

The dead storage level of reservoir is at an elevation of 1462 ft, whereas the maximum reservoir elevation is 1680 ft. In order to mitigate floods the reservoir level can be raised up to elevation of 1690 ft in consultation with Chairman BBMB. It is observed from Figure 5 that the water level reaches maximum storage elevation of 1680 ft every year by 31 August. from 1975 to 2019 except in few years. Figure 5 also indicate no significant change in maximum or minimum levels from 1975 to 2019.

5. Bhakra Reservoir Capacity Calculator

The Python script uses binary search to predict various parameters of Bhakra reservoir. Binary search is half-interval search algorithm that predicts the position of a target value within a sorted array. Binary search compares the target value to the middle element of the array. In binary search, an array is sorted by repeatedly dividing the search interval in half. From excel sheet data, the Python script is able to locate the volume corresponding to a particular elevation. The code automatically updates the status of reservoir with time. Initial water level data is fed to code and the corresponding closest reservoir capacity is evaluated depending upon the resolution of the data.

Water balance equation

(1)

is storage at time (t+1) in thousand acre feet, is the total inflow volume ( thousand acre feet) in time t and is the total release volume (thousand acre feet) from the dam in time t.

(2)

Present storage S(t+1), previous storage St can be estimated from Python script from measured levels which are monitored continuously. Release depending upon demand is a known parameter. Inflow volume Qt can be determined using the following equation.

(3)

Figure 6 shows the results obtained after executing the python script inflow_calculator_bhakra.py. The results indicate the computation of reservoir capacity when reservoir level is provided to code. When inflow volume and release is fed in the code, it provides next state of reservoir storage and its level.

6. Conclusions

Python software based inflow calculator will enable the engineers and policy makers to determine the storage available in the dam based on water levels measured at gauging stations daily. The levels in the dam can be effectively monitored hourly during peak flows in filling period with the help of the script developed herein. Continuous monitoring of reservoir storage is necessary in order to ensure effective flood mitigation. Dam is required to be emptied to absorb excess flow. Release from the dam depends upon the demand, storage available and downstream capacity to absorb flow. Inflow can also be assessed from the water balance equation. The software described in the present paper has the potential to enable continuous monitoring of the reservoir state. Using the software developed here, human error in the operation of Bhakra reservoir can be completely eliminated.

Acknowledgements

The authors acknowledge the funding received by Science and Education Research Board (SERB) of the Department of Science and Technology (DST) for undertaking this research.

References