Employing HEC-ResSim 3.1 for Reservoir Operation and Decision Making

*Abhishish Chandel, Vijay Shankar and Sumit Jaswal*

#### **Abstract**

The land of Himachal Pradesh is full of small and big rivers which are perennial. This benefit pushes Himachal Pradesh to build more and more dams to generate electricity, provide better irrigation supply to downstream areas and provide flood protection. To better utilize the huge potential of water, management of such reservoirs is the key issue. For this purpose, HEC-ResSim 3.1 is practiced on Pong Dam situated in western Himachal Pradesh. HEC-ResSim is one of the simulation models that possess single or multi-reservoir simulators and can simulate water resources systems. In this study, reservoir elevation and reservoir storage volume management is the target objective. The presented study was subsidized by daily observed data from 1998 up to 2014 of pool elevation, inflow, and outflow discharge. In addition, geometry and hydraulic data from dam and reservoir were employed to develop the platform to create a simulation using HEC-ResSim. Using the available reservoir data the simulation was performed for the 4 months of 2012. Then simulation results were compared with the real-time recorded data at the site. To validate the results, coefficient of determination for operations like reservoir elevation and reservoir storage was generated through regression plot and found more than 95% accurate. Also, Root Mean Square Error (RMSE) was calculated for both reservoir elevation and reservoir storage simulation and found under an acceptable range. This paper shows the utility of HEC-ResSim 3.1 for reservoir operational management and also throws light on the further scope. Finally, there is a discussion of how useful is HEC-ResSim as a reservoir management tool and integration of HEC-ResSim 3.1 with other hydrologic monitoring systems.

**Keywords:** HEC-ResSim, simulation, reservoir management, irrigation

#### **1. Introduction**

For a better operation of reservoirs, it is extremely necessary to manage the reservoir storage and hence reservoir levels with the seasonal variations throughout the year. Highly variable inflows from the river and water demands from the reservoir make it more challenging to manage reservoir storage volume according to daily demand. In the 1950s, the Corps of Engineers (COE), USA, and the National Weather Service (NWS) jointly developed the Stream-flow Synthesis and Reservoir Regulation (SSARR) computer model. SSARR was used both as the stream-flow forecasting tool and as the real-time reservoir regulation tool. However, over the past few years, the stream-flow routing algorithms have been migrated to the HEC's ResSim model. Modini [1] studied and described all the challenges and strategies to completely migrate the AUTOREG/SSARR model to HEC-ResSim. An important objective was to ensure that all the provisions of the current Columbia River Treaty Flood Control Operating Plan (FCOP) must be migrated to HEC-ResSim. The main objective of the study was to develop a flexible model to accommodate FCOP strategy changes.

Eichert and Davis [2] generated the HEC-5 model for the study of flood control on the Susquehanna Reservoir System, USA. Model executed a decision support system to overcome the uncertainties of unevenly distributed water resources systems. Hickey et al. [3] used Hec-5 software in reservoir simulation for flood analysis in response to the destructive floods of 1983, 1986, 1995, and 1997. The main targets of study are model development, with a focus on headwater and major terminal reservoirs, and potential improvements to the flood damage reduction system. Emphasis is laid on model development and analyzing the influence of reservoirs in flood hydrology. Kim et al. [4] developed a deterministic optimization model named Coordinated Multiple Reservoir Operating Model (CoMOM) for real-time multi-reservoir operations in the Han River basin in Korea. Matondo and Msibi [5] created a DSS support with three major components, that is, the model input, modeling options, and outputs screens. The output of the DSS comprises the optimal rationing (%), monthly reservoir volume for the desired duration as well as a graphical representation of the reservoir response over the old and new scenario.

Bekele and Knapp [6] coupled storage routing and multi-objective evolutionary algorithms to simulate reservoir release rates of "Shelbyville and Carlyle Lakes" on Kaskaskia River, USA. The resulting coupled model can provide simulations of storage and reservoir pool elevations for the two lakes under varying water use conditions. All the long-term studies proved that the modeling techniques for the management of reservoirs are very helpful and efficient. Todd et al. [7] published studies on reservoir simulation using different techniques. Focuses on the four modeling systems: Reservoir System Simulation (HEC-ResSim), River and Reservoir Operations (RiverWare), River Basin Management Decision Support System (MODSIM), and Water Rights Analysis Package (WRAP). Though fundamentally similar, the four modeling systems differ significantly in their organizational structure, computational algorithms, user interfaces, and data management mechanisms. The Bureau and Tennessee Valley Authority jointly sponsored the development of RiverWare at the Center for Advanced Decision Support for Water and Environmental Systems of the University of Colorado [8, 9]. The Tennessee Valley Authority applied RiverWare in optimizing the daily and hourly operation of the system of multipurpose reservoirs and hydroelectric power plants. The Lower Colorado River Authority also applied RiverWare in daily time step modeling of water supply operations for reservoirs on the Colorado River of Texas [10]. MODSIM is a general-purpose reservoir/river system simulation model based on a network flow linear programming developed at Colorado State University [11, 12]. MODSIM has been used to study several reservoir/ river systems in the western United States and throughout the world. The objective function coefficients used in MODSIM are factors entered by the model used to specify relative priorities that govern operating decisions.

The development of WRAP at Texas A&M University began in the late 1980s. WRAP has been greatly expanded since 1997 in conjunction with implementing a statewide Water Availability Modeling (WAM) System [13]. WRAP simulates water resources

#### *Employing HEC-ResSim 3.1 for Reservoir Operation and Decision Making DOI: http://dx.doi.org/10.5772/intechopen.101673*

development, management, regulation, and use in a river basin or multi-basin region under a priority-based water allocation system. In WRAP terminology, a water right is a set of water use requirements, reservoir storage and conveyance facilities, operating rules, and institutional arrangements for managing water resources. Simulation results stored as DSS files accessed with HEC-DSSVue (a program used to manipulate data from HEC-DSS databases) for plotting and other analyses [14, 15]. According to the comparative studies of the basic modeling techniques for the reservoir operational studies, HEC-ResSim is recommended as the most productive and efficient modeling software. In 2004, for evaluation and reservoir management of the Tigris and Euphrates rivers system in Iraq, HEC-ResSim 2.0 was employed by Hanbali [16]. The study included six main reservoirs, three off-stream reservoirs, and seven small reservoirs, and many diversion dams for diverting water from Tigris and Euphrates rivers. HEC-ResSim 2.0 was used for simulation history events especially flood and drought periods.

Babazadeh [17] employed HEC-ResSim for reservoir modeling and stated that the application of simulation models is one of the most efficient ways of analyzing water resources systems. Model verification results indicate that this model can simulate the behavior of the system very well. Modeling resulted in increasing irrigation efficiency by 20% and reducing failures in the system by 12%. McKinney [18] developed a flow model of Lancang Cascade Dams, China, to maximize hydropower production and calibrate the model to match outflow at downstream gauge with the data of most recent year available data. HEC-ResSim came out to be capable to model dams in series. Piman [19] used HEC-ResSim and SWAT simultaneously for the assessment of flow changes from hydropower development and operations in Sekong, Susan, and Srepok Rivers of Mekong Basin, Vietnam. To access the magnitude of potential changes, daily flows were simulated over 20 years using the HEC-ResSim and SWAT models for a range of dam operations and development scenarios. Goodarzi [20] practiced a combination of LINGO and HEC-ResSim models to determine monthly operating rules for the Zayandehrud reservoir system in Iran. The results show that optimizing the operation of the Zayandehrud reservoir system could increase its storage by 88.9% as well as increase the reliability index of regulated water for all downstream demands by more than 10%. Lara [21] employed the HEC-ResSim model on Tucurui Dam, Brazil. It was subsidized by daily observed data from 2001 up to 2006 of pool elevation, inflow, and outflow discharge. HEC-ResSim was established as a powerful tool to support the decision-making of reservoir operations and an interesting alternative for risk management and flood control. Klipsch and Hurst [22] developed the HEC-ResSim 3.1 user's manual which provided great support and learning in employing HEC-ResSim 3.1 for the management of Pong Dam. Along with the software support, HEC-ResSim user's support by socio-networking website played a great role in the execution of the study (HEC-ResSim user's blog). In the present study, the reservoir storage volume and reservoir levels are modeled using Hec-ResSim 3.1 reservoir understudy is "Pong Dam" on Beas River in the state of Himachal Pradesh, India. Reservoir level and reservoir storage's target information are generated which enhance the decision-making capabilities related to reservoir management works. The input data is reservoir inflow and outflow time-series, reservoir physical data, reservoir area-volume-depth relationship, etc. In this paper, HEC-ResSim 3.1 is employed to the reservoir for the generation of a decision support system to regulate the reservoir elevation and reservoir volume which further enhances the reservoir operations (flood control, irrigation water supply, and hydropower generation). The study will help the reservoir management to tackle the future incoming water challenges and also to create simulations to practice the sudden situations like cloud

bursts, surprise snowfall, uneven rains, etc. which are very common events over the region. All these parameters demand an improved operational system for the reservoir. The main objectives of the presented study are:


## **2. Simulation software and data**

#### **2.1 HEC-ResSim 3.1 introducing**

The simulation software used in the present study is HEC-ResSim (version 3.1) created by the U.S. Army Corp of Engineers—Hydrologic Engineering Center. The software has three main modules: Watershed Setup, Reservoir Network, and Simulation. Res-Sim has a graphical user interface (GUI) and utilizes the HEC Data Storage System (HEC-DSS) for storage and retrieval of input and output time-series data. ResSim is used to simulate reservoir operations including all characteristics of a reservoir and channel routing downstream. The data requirements for HEC-ResSim include the physical and operational characteristics of the dam and reservoir. The physical reservoir data is described through the use of the volume-area and elevation curves (**Figures 1** and **2**).

**Figure 1.** *Elevation-volume curve of Pong reservoir.*

*Employing HEC-ResSim 3.1 for Reservoir Operation and Decision Making DOI: http://dx.doi.org/10.5772/intechopen.101673*

**Figure 2.** *Elevation-area curve of Pong reservoir.*

The physical data of the dam include the type and capacity of each outlet. The operational data includes the zone definitions along with the rules governing the operations in each zone. There are three main management zones or pools, that is, the inactive pool, the conservation pool, and the flood pool.

The model allows the user to define alternatives and run their simulations simultaneously to compare results. Network elements include reservoirs, routing reaches, diversions, and junctions. In ResSim, watersheds include streams, projects (i.e., reservoir, levees), gage locations, impact areas, time-series locations, hydrologic and hydraulic data for the specific area. In the present paper, the reservoir operational rule curves data shown in **Figure 3** are employed to perform the HEC-ResSim simulations and represent the operational patterns of the Pong Dam. "Top of the Dam" shows the physical top-most part of the dam and "flood control curve" shows the max level for emergency releases. Till now maximum pool level achieved by the reservoir is shown by the "maximum pool elevation curve." "Conservation curve" shows the saved usage and operational levels of the reservoir for the past years. Similarly, buffer storage level and inactive pool levels are also described in the **Figure 3**.

Pong Dam modeling is performed in a systematic pattern using HEC-ResSim 3.1. In the first step, the Pong watershed is set up to employ HEC-ResSim 3.1. To set up Pong watershed features are added, stream alignments are drawn, configurations created, and project elements placed into configurations. The next reservoir network is developed over the watershed. In this routing reaches and junctions are added and edited simultaneously. Reservoir data like physical data (pool, dam, and outlet properties) are added. Then operation sets are added by applying zones and rules. Also, reference to the observed data created here. Alternatives by selecting network are

**Figure 3.** *Observed operational rule curves for Pong Dam.*

defined. Run control settings are determined by the operation set for the simulation. The look-back date is defined. Simulation identifies time-series records and observed records from alternatives. To perform simulation a predefined alternative to the current simulation is selected. In the simulation module, the output result is analyzed, simulated results for the reservoir elevation and reservoir volume are extracted and the efficiency of simulated output is computed. The efficiency is computed using two methods: that is, graphical method, that is, to find the coefficient of determination, and statistical method, that is, to find Root Mean Square Error (RMSE).

$$\text{RMSE} = \left(\frac{\sum\_{i=0}^{n} \text{(observed output - simulated output)}^{2}}{n}\right)^{0.5} \tag{1}$$

where *n* is the total number of data points.

### **3. Study location and characteristics**

Pong Dam on Beas River is located in the wetland zone of the Shivalik Hills of western Himachal Pradesh, India at 32.0167°N, 76.0833°E. It is the highest earth-fill dam in India. The reservoir is a well-known wildlife sanctuary and one of the 25 international wetland sites declared in India. India. **Figure 4** shows the geographical presence of the Pong Dam in India.

Beas River is one of the biggest rivers of Himachal Pradesh. The starting point of this permanent river is the snow-covered mountain at Rohtang Pass in Himachal Pradesh, India. Annual volume of water carried by Beas River as measured by monitoring stations in 40 years' time series is 9701.82 MCM. Pong Dam collects water drained from the catchment area of 12,613.998 Sq. km. First of all, operation reservoir volume is 7290 MCM at elevation 426.72 m from sea level and inactive level is 384.5 m above sea level. The length of the dam crest is 1951 m and its height from the river bed is 105.86 m and its crest width is 13.72 m. This dam has six Francis turbines with a capacity of 66 MW each. This is a single-purpose dam designed and constructed to serve irrigation to downstream areas. However, flood control, hydropower, and fisheries are the complimentary

#### **Figure 5.**

*Irrigation water release from Pong Dam (1998–2012).*

**Figure 6.** *Pong reservoir inflow (1998–2012).*

**Figure 7.** *Model layout of Pong reservoir and component setup in river Beas.*

*Employing HEC-ResSim 3.1 for Reservoir Operation and Decision Making DOI: http://dx.doi.org/10.5772/intechopen.101673*

benefits. Based on field survey and published data from Pong Dam Operation Company, that is, Bhakhra Beas Management Board, agricultural sector use of water from dam reservoir is shown in **Figure 5**. Irrigation water release is the major operation of the Pong Dam. Release decisions depend upon many seasonal factors and crop water demands.

Also, there is a daily record for the reservoir inflow from 1998 to 2012 which acts as a very useful parameter for simulation (**Figure 6**). As presented earlier (**Figure 5**) the irrigation water supply is nearly constant but the incoming water to the reservoir is very flashy. Touching extreme peaks during monsoons and consistently very low during the rest of the year. Such flow patterns demand proper management and regulation of water releases from the reservoir.

HEC-ResSim 3.1 is employed to model the pong reservoir's operational setup and comes out as shown in **Figure 7**. It includes a stream, pong reservoir, computational points, penstock tunnel outlet, and irrigation water supply outflow. Penstock outlet is before the spillway and meets to the stream again downstream from where whole water diverted to the irrigation water supply. As hydropower generation is a complimentary operation over the irrigation water supply, the water indent for the powerhouse completely depends upon the irrigation water demand in the downstream fields.

#### **4. Results and discussion**

Secondary data of the reservoir (inflow, outflow, reservoir level, storage, and power generation as daily data) are collected from Bhakhra Beas Management Board (BBMB). **Figure 8** shows the behavior of the reservoir as generated by the real-time

**Figure 8.** *Observed parameters of Pong storage dam operation.*

data records. Reservoir inflow is touching extreme peaks during monsoons, that is, July–September whereas inflow is very low during the rest of the seasons. This makes the River understudy very flashy, hence requiring proper management.

For the study of accuracy and validation of the model, observed data from different years are compared with the respective simulated data. Therefore, input data, output data, and reservoir and dam properties were supplied to the model for variable durations of simulation for different years. Here results are shown for the 4 months of 2012 tagged with the time-series data of the whole year 2012.

#### **4.1 Reservoir elevation operation**

When we talk about the head formation for hydropower generation or preparing the reservoir for flood control, the first thing that comes to the manager's mind is the reservoir elevation. If one achieves the target elevation, the further operations of the reservoir become more efficient and safe. Hence if we can predict the future hydrologic condition of the reservoir we can simulate the target reservoir elevation in that situation. Also, we can practice some random expected flood values to check the elevation operation in that situation. In this way, we can prepare our-self for the worst condition even if that situation had never hit before. **Figure 9** shows the comparative plot for the actual reservoir elevations and the simulated reservoir elevations. It is showing results for 4 months of simulation and is utilized for the calculation of RMSE for reservoir elevation. RMSE for the operation of reservoir elevation is 0.78 m. This RMSE value is acceptable and recommends HEC-ResSim 3.1 for such reservoir modeling efforts.

To check the efficiency of the model both the simulated and actual values of the simulation were analyzed in the regression curve (**Figure 10**). Daily observed and simulated data related to reservoir elevation for the 123 days of the simulation period

#### **Figure 9.**

*Comparison of observed reservoir elevation and simulated reservoir elevation for the simulation period of August to November 2012.*

*Employing HEC-ResSim 3.1 for Reservoir Operation and Decision Making DOI: http://dx.doi.org/10.5772/intechopen.101673*

**Figure 10.**

*Regression curve for the observed and simulated reservoir elevations for the simulation period of August to November 2012.*

(August–November 2012) plotted in the regression chart to analyze the second efficiency parameter, that is, Coefficient of Determination.

The "Coefficient of Determination" for the plot was evaluated as 0.98, hence we can say that the simulated reservoir elevation values are 98% correct. If we want to employ HEC-ResSim for Pong Dam elevation simulation then it can be a trustful and powerful tool.

#### **4.2 Reservoir storage operation**

Reservoir storage is the main point of concern when we have to manage the outflow with the downstream demands like agricultural irrigation, etc. Also for hydropower optimization, the elevation-storage curve is the backbone of the process. If the reservoir is very prone to sedimentation then also the storage volume of the reservoir plays a key role in changing reservoir geometry. So there is a great deal if there is a technique to simulate reservoir storage volume. If one achieves the target storage volume, the further operations of the reservoir become more efficient and safe. Hence if we can predict the future hydrologic condition of the reservoir we can simulate the target reservoir storage in that situation. Also, we can practice some random expected flood values to check the storage operation in that situation. In this way, we can prepare our-self for the worst condition even if that situation had never hit before. **Figure 11** shows the comparative plot for the actual reservoir storage and the simulated reservoir storage. It is showing results for 4 months of simulation and is utilized for the calculation of RMSE for reservoir storage. RMSE was calculated for the reservoir storage values obtained from the simulation duration. RMSE was evaluated as 151.81 MCM for the simulation of reservoir storage. This RMSE value is acceptable and recommends HEC-ResSim for such reservoir modeling efforts.

**Figure 11.**

*Comparison of observed reservoir volume and stimulated reservoir volume for the simulation period of August to November 2012.*

**Figure 12.**

*Regression curve for the observed and simulated reservoir volumes for the simulation period of August to November 2012.*

To check the efficiency of the model both the simulated and actual values of simulation are analyzed in the regression curve (**Figure 12**). Daily observed and simulated data related to reservoir storage for the 123 days of the simulation period (August– November 2012) plotted in the regression chart to analyze efficiency parameter, that is, Coefficient of Determination.

*Employing HEC-ResSim 3.1 for Reservoir Operation and Decision Making DOI: http://dx.doi.org/10.5772/intechopen.101673*

The coefficient of regression for the plot was evaluated as 0.92, hence we can say that the simulated reservoir elevation values are 92% correct. If we want to employ HEC-ResSim for Pong Dam storage simulation then it can be a trustful and powerful tool.

**Figure 13.** *Simulated and predicted outflow for August 2012.*

**Figure 14.** *Simulated and predicted outflow for September 2012.*

**Figure 15.** *Simulated and predicted outflow for October 2012.*

**Figure 16.** *Simulated and predicted outflow for November 2012.*

### **4.3 Additional simulation results**

HEC-ResSim is also capable of predicting unknown flows from the reservoir. In the initial conditions, we have entered irrigation release as an outflow only and kept spill zero. But after simulation, HEC-ResSim found some extra unregulated water in the reservoir which should be thrown out except irrigation release. This aids the management of release decisions even if we do not have any previous record of any such releases. Calculating all the inflow, outflow, elevation, and storage patterns in a synchronized manner and considering rule curves, HEC-ResSim predicted the target water spillage for all 4 months of simulation. **Figures 13**-**16** show the predicted water spillage from the reservoir for the simulation months of August, September, October, and November, respectively for the year 2012.

## **5. Conclusion**

According to the results presented in this paper, HEC-ResSim is an interesting alternative to reduce the uncertainties of outflow forecasts and support the improvement of the flooding warning program of the Pong Dam. Putting the Pong Dam's hydrological database together with HEC-ResSim, one could reproduce operational aspects of the dam and test different operational scenarios, even in real-time. Since for this model, the availability of data and information falls short but concerning the available data, it was able to prove that Hec-ResSim is a very efficient way to analyze any reservoir for different climatic conditions.

This article attempts to study the performance of the Pong storage dam in actual conditions and simulated conditions using HEC-ResSim and evaluation indices. Results of model validation showed that the model was capable of simulation with suitable accuracy.

HEC-ResSim is a powerful tool, which can support the decision-making of the managers and operators at the Pong Dam. The model presents capabilities to improve the precision of the flooding warnings, reduce dam safety costs, and increase hydropower production. In addition, Hec-ResSim supports a minimum computation time of 15 minutes, which can generate decision support for every 15 minutes of the simulation period. HEC-ResSim is also an interesting alternative for risk management and water control. Moreover, it can further come out to be more precise with the attachment with GIS. It can work a long way with power plants up-to-the turbines level if there is the availability of detailed data regarding it. HEC-ResSim can support the real-time decision, using a real-time integrated database along with a userfriendly interface integrator like HEC-RTS or DELFT-FEWS. HEC-RTS—Real-Time Simulation is another U.S. Army Corps of Engineers tool, which provides support for operational decision-making. HEC-ResSim integrated with HEC-RTS or HEC-HMS allows the water control manager to make short-term (typically a few days or weeks) forecasts of hydrologic conditions at the catchment scale.
