Optimal reservoir operation for water supply in dry season: the case study of Cua Dat reservoir in the Ma – Chu river basin

each level depend on knowledge of expert. If number of membership function for inputs and output increases, the number of rules in fuzzy rules base will also increases as well as the error decreases. Regarding to analysis of observed and calculated data of inputs and output (appendix 3), the author determined the ranges of each variable and divided into distinct levels. With reservoir water level has ranges that varied from 70m to 115 m. It was classified into 7 levels (figure4-16) including ‘very low’, ‘low’, ‘low-medium’, ‘medium’, ‘medium-high’, ‘high’ and ‘very high’. The inflow into the reservoir as ‘very low’, ‘low’, ‘medium’, ‘high’ and ‘very high’ are in range of [0-145] Mi.m3. Figure 4-16: Membership function for reservoir level for Fuzzy Mamdani model Figure 4-17: Membership function for inflow for Fuzzy Mamdani model 59 Trinh Xuan Manh MSc Thesis In the same way, the demand is divided into class intervals as ‘very low’, ‘low’, ‘meium’, ‘high’ and ‘very high’ in range of [0 -145] M.m3. Finally, the release is divided into 8 class intervals (figure 4-18) as ‘very low’, ‘low’, ‘medium-low’, ‘medium’, ‘medium-high’, ‘high’ and ‘Very.very.high’. Figure 4-18: Membership function for water demand for Fuzzy Mamdani model Figure 4-19: Membership function for release for Fuzzy Mamdani model IV.2.3.2. Formulation of the Fuzzy rule base In a Fuzzy system, fuzzy rules base was built due to ‘IF-THEN” principle. The fuzzy rule set usually is formulated based on knowledge and experience of expert. 60 Trinh Xuan Manh MSc Thesis Besides this, the fuzzy rule base also based on the actual observed operation of the reservoir. According to the collected data and different categories of each input/output variable in dry season and using Fuzzy Mamdani method, the Fuzzy rules have been identified as below: IF (RESERVOIR WATER LEVEL IS LOW) AND (INFLOW IS LOW) AND (WATER DEMAND IS LOW) THEN (RELEASE IS VERY LOW) There are 157 fuzzy rules were defined in Fuzzy system of the Cua Dat reservoir regarding to 03 input variables as reservoir water level, inflow and water demand and 01 output variable as release (Appendix 5). The fuzzy output of release is also in a fuzzy set to achieve the crisp value which is needed a process of “defuzzification” method. The figure 4-20 presents the fuzzy rules base for operation of the Cua Dat reservoir. Figure 4-20: Fuzzy rules base for operation of Cua Dat reservoir IV.2.3.3. Application, Implication, Aggregation and Defuzzification Application, implication, aggregation and defuzzification are main parts to calculate fuzzy output and determine crisp value for each rule is known in a fuzzy system. Application is a process that it uses fuzzy operators such as AND or Or to 61 Trinh Xuan Manh MSc Thesis obtain value for presenting the result for one rule. Panigrahi and Mujumdar (2000) have discussed the fuzzy logic operators obey the classical two valued logic. The AND opertator can be conjunction of the classical logic or it can be the product of the two parameters involved in it. Implication is a process after application above. Application process uses fuzzy operators to calculate a value due to inputs of fuzzy system then Implication process will apply the value into membership function of outputs in order to obtain a fuzzy set for each rule. Especially, in implication process is input of it is a number value but the result is a fuzzy set. For each rule always implement an implication method. Aggregation is next step after implication step. Aggregation integrates the outputs which are results of implication method. The inputs of aggregation are fuzzy set and result of aggregation process is also a fuzzy set for each output variable. The figure 4-21 presents process of application, implication and aggregation. This figure was found in Panigrahi and Mujumdar (2000). Figure 4-21: Process of application, implication and aggregation The result is optimized from fuzzy system through the processes such as application, implication and aggregation is a fuzzy set. So that we need a defuzzification method in order to achieve a crisp value or crisp number. The defuzzification method is integrated ready in Fuzzy Tool Box of MATLAB software 62 Trinh Xuan Manh MSc Thesis including: Centroid; Bisection; Largest of Maximum; Smallest of Maximum, Middle of Maximum. Panigrahi and Mujumdar (2000) have shown that one of most common defuzzification method is “Centroid” evaluation. In this thesis, the Centroid method will be used for defuzzification. The equation of this method as following: 𝐺 = ∑ 𝑦𝑖 𝑚𝐵(𝑦𝑖)𝑛𝑖=1 ∑ 𝑚𝐵(𝑦𝑖)𝑛𝑖=1 (4-12) Where: G is the centroid of the truncated fuzzy output set B mB(yi) is the membership value of the element yi in the fuzzy out put set B, n is the number of elements IV.3. Hydraulic and hydrological model setup In this part the author is going to set up a hydraulic model and a hydrological model for the Ma – Chu river basin. The MIKE 11 HD and MIKE NAM models will be used as an efficient way as mentioned in the literature review of the thesis. These models were used to simulate the release from the reservoir to the river in downstream area as well as the flow at some of control points after routing that to be evaluated effective optimization. MIKE NAM is a hydrological model and MIKE 11HD is a hydraulic model, the combination between 02 models through calibration and validation process to determine optimal basin parameters. The detailed contents are presented in following steps. IV.3.1. Determination of the model inputs As presented above, the author used two models including the MIKE NAM and MIKE 11HD model. Each model will need different data as inputs. The inputs of the MIKE NAM hydrological model include rainfall, evaporation and discharge. The inputs of MIKE 11HD include discharge and water level at the stations (appendix 4). To build up hydrological model, the author collected rainfall data from the meteorological and point stations of two years of 2006 and 2008 of Cua Dat, Bai Thuong, Cam Thuy, Song Am, Thuong Xuan, Sam Son, Giang stations to simulate NAM model for Cua Dat catchment. Similarly, discharge data of the Cua Dat station was collected to calibrate and validate the basin parameters. 63 Trinh Xuan Manh MSc Thesis The hydraulic model has the boundary conditions such as upstream boundaries, downstream boundaries and topographic data as cross sections. The choice of boundary conditions depends on the physical situation and data available on the basin. In this study, the author selected the following stations and data itself as boundary conditions. The upstream boundary is time series of observed discharge at the hydrological stations as Cua Dat (Chu Rver), Cam Thuy (Ma River), Kim Tan (Buoi River) and Tu Thon (Bao Van River) stations. Downstream boundary is time series of observed water level at some of stations in the river mouth as Hoang Tan, Lach Truong and Lach Sung stations. The time period for calibration and validation of model is in 02 years (2006 and 2008). After data collection for those models the author implemented set up model network as below: IV.3.2. Model setup In this part, the author implemented set up hydraulic schematization with the boundary conditions for the Ma – Chu river basin. According to topography data and hydrological data and hydraulic regime of river network, the hydraulic network of the river basin was chosen as following: + Ma River: starting from Cam Thuy station to the estuary (Cua Hoi estuary) + Chu River: starting from Cua Dat station to conjunction point of the Ma River (Giang confluence) + Buoi River: starting from Thach Lam station to conjunction point of the Ma River (Vinh Khanh confluence) + Cau River: Starting from Cau Chay station to conjunction point of the Ma River at Cam Truong confluence. + Len River: from Bong confluence to Lach Sung Estuary + De Chanel: from separate branch of Len River to conjunction point of Lach Truong River. + Lach Truong River: from Tuan confluence to Lach Truong estuary. The boundary conditions of this model were distributed: - Upstream boundary: is time series of daily observed discharge from upstream stations: + Cam Thuy station is located on the Ma River 64 Trinh Xuan Manh MSc Thesis + Cua Dat station is located on the Chu River + Kim Tan station is located on the Buoi River + Thu Thon station is located on the Bao Van River - Downstream boundary: is time series of daily observed water level from downstream stations: + Hoang Tan station is located in the Estuary + Lach Truong station is located in estuary of Tao Khe River. + Thach Sung is located on Len River. - Tributary boundary: flow which occurs from tributary basin is identified through MIKE NAM model. List of tributary basins are in below table: Table 4-13: List of tributary basins on the Ma – Chu river basin Tributary basins Name Control areas Area (km2) Am River Am river basin area 817.5 Dat river Dat river basin area 304 Dang River Dang river basin area 344.7 Cau Chay River Cau Chay river basin area 118 - Control stations: is used to evaluate the efficiency of the model including the station as following. Those stations are water level stations: + Ly Nhan station is located on the Ma River + Giang station is located on the Ma river) + Xuan Khanh station is located on the Chu River The hydraulic network of Ma-Chu river basin is presented in the below figure: 65 Trinh Xuan Manh MSc Thesis Figure 4-22: Hydraulic network of the Ma – Chu river basin IV.3.3. Model calibration and validation This process aims are to obtain the optimal parameters for NAM model and MIKE 11HD model respectively. To evaluate the efficiency of model using observed data at some control stations are compared to simulated data from the model. To access performance of the models the author used Nash- Sutcliffe coefficient (NASH). NASH has been commonly used in hydrological models to evaluate flow hydrographs. The NASH is an improvement over the coefficient of determination for stream flow comparison, because it accounts for model errors in estimating the mean of the observed datasets. It enables the efficiency of the model to be compared with the initial variance that is defined by the observed datasets. The observed and simulated discharge values have been compared to assess the predictive accuracy of the model. NASH ranges from −∞ to 1. When the values are from 0.9 to 1 indicate that model performance is perfect. When the values are from 0.8 to 0.9 and 0.6-0.8, the model performances are good and acceptable, respectively. It is defined as following: 66 Trinh Xuan Manh MSc Thesis      2 2 )( )( 1)( QobsaverQobs QobsQcal EINASH (4-13) Where: calQ : Simulated discharge at the time ith obsQ : Corresponding observed discharge at the time ith Qobsaver : Average observed discharge Moreover, it is possible to use other indexes or coefficients to access the results of model such as: Correlation Coefficient; Peak Error; Volume Error and Peak time Error etc. The results for both models will be presented in following contents. IV.3.3.1. NAM model calibration and validation A- Calibration for NAM model of the Ma –Chu river basin: Calibration is one of main steps in modeling. Trial and error method was used to stimulate NAM Model for the Cua Dat Catchment. It was carried out by selecting a possible parameters, applies it to the model and, if it is not successful, selects another possibility that is subsequently tried. The process ends when a possibility yields a goodness of fit between simulated and observed data. The Cua Dat Reservoir is upstream part of Ma- Chu River Basin. Its area covers 5.708 square kilometers. To calibrate the hydrological model of the Cua Dat catchment on the Chu River, the author used daily rainfall of the Cua Dat station and surrounding stations in year of 2006. Average daily evaporation at Thanh Hoa station was also used for the same period. The observed discharge which was collected from the Cua Dat station was used to access the performances of the model. The simulated performance for calibration process is as the Figure 4-23. 67 Trinh Xuan Manh MSc Thesis Figure 4-23: Observed and simulated hydrograph at Cua Dat station in 2006 Based on two hydrographs on the above figure, it can be seen that the difference between them is not quite high. The simulated result is still higher than observed result. Those caused by number of rainfall stations on the basin is rather limited, it do not cover the whole basin. However, the shapes of them are quite similar with each other and NASH is 0.69 at the station indicates an acceptable value. With this result, it is possible to validate the model in next step as below. B- Validation for NAM model of Ma –Chu river basin: Model validation is one of the important steps in modeling to achieve confident model. The model used only the parameters which were obtained in previous step to simulate runoff from rainfall. Year of 2008 was used for model validation with the same data and stations in calibration step. The simulated performance for validation process is as figure 4-24. 68 Trinh Xuan Manh MSc Thesis Figure 4-24: Observed and simulated hydrograph at Cua Dat station in 2008 For the validation period, the above figure indicates that between observed and simulated hydrograph have a good fit and similar pattern. NASH = 0.73 was acceptable value and indicated well validated model. Therefore, developed NAM model could be confidently used for the purpose of this study. After calibration and validation of NAM model, a set of NAM parameters are determined and they would apply to the Ma-Chu river basin to reveal the flows of rivers in basin. The objective of application of NAM model was to simulate runoff from available rainfall data for Ma - Chu River and its branches. However, due to lack of sufficient discharge data, the application of direct calibration of a rainfall-runoff model for each sub-basin is impossible. So it were needed to borrow model parameters from the Cua Dat catchment is to apply for other sub-basins to recovery runoff. The tributary basins were simulated runoff presented in the Table 4-13. IV.3.3.2. MIKE 11HD model calibration and validation In this part, in order to determine the optimal parameter for hydraulic model, the author used manual method to simulate the flow from upstream station. To do that the hydrological data was collected for year of 2006 and 2008 respectively. Insides that, the period of model calibration was chosen in 2006 and year of 2008 was used for model validation. The performance evaluation of model could be measured by NASH coefficient as efficient way. 69 Trinh Xuan Manh MSc Thesis A- Calibration for MIKE 11HD model For calibration period, the calibration is implemented by changing roughness along the river and cross section, time step, etc. The calculated process will be stopped if the difference between calculated and observed data is less than certain value for every point. The results for calibration are presented in some figures below: Figure 4-25: Observed and simulated water level at Ly Nhan Station in 2006 Figure 4-26: Observed and simulated water level at Xuan Khanh Station in 2006 70 Trinh Xuan Manh MSc Thesis Figure 4-27: Observed and simulated water level at Giang Station in 2006 Through the results which were showed in the above figures, it is possible to see that having a good fit between observed and simulated data about pattern and peak value for the most of the stations. NASH for stations varied from 0.87 to 0.92, the difference of two peaks is rather low. Therefore, the author also used the Peak Error and Volume Error to evaluate the results (Table 4-14) Table 4-14: Results of MIKE 11HD model calibration at the Ma-Chu river basin in 2006 Name of station NASH Peak Error Volume Error Ly Nhan 0.97 1.92% 0.33% Giang 0.89 4.90% 9.50% Xuan Khanh 0.88 3.98% 6.07% So that with results of model calibration, the author basically determined the hydraulic parameter (roughness) for whole river basin, Manning coefficient vary from 0.025 to 0.045 from upstream to downstream area. To evaluate the confidence of the model the author implemented next step for Validation. B- Validation for MIKE 11HD model The results for model validation are presented in some figures below. For the validation period, it can be seen that there is a good fit between observed and simulated 71 Trinh Xuan Manh MSc Thesis data about pattern and peak value for most of the stations. NASH for stations varied from 0.83 to 0.94, the difference of two peaks is rather low. Figure 4-28: Observed and simulated water level at Ly Nhan Station in 2008 Figure 4-29: Observed and simulated water level at Xuan Khanh Station in 2006 72 Trinh Xuan Manh MSc Thesis Figure 4-30: Observed and simulated water level at Giang Station in 2006 Table 4-15: Results of MIKE 11HD model validation at the Ma-Chu river basin in 2008 Name of station NASH Peak Error Volume Error Ly Nhan 0.91 1.60% 3.52% Giang 0.91 8.25% 6.44% Xuan Khanh 0.84 4.07% 11.88% According to the Table 4-15 and Figure 4-28 to 4-30, it can be seen that observed water level is higher than simulated water level at the beginning of the hydrograph at Xuan Khanh station. However, this difference is still in the limited range. With the highest results, developed MIKE 11HD model could be confidently used for the purpose of this study. 73 Trinh Xuan Manh MSc Thesis CHAPTER V RESULTS AND DISCUSSIONS V.1. Optimizing the Cua Dat reservoir operation As presented above, the author sets up a Fuzzy system for operation of the Cua Dat Reservoir (Figure 4-31). The Fuzzy system includes 03 input variables as 03 input variables such as reservoir water level, inflow and water demand and 01 output variable as release. Beside that, the Fuzzy system also includes a fuzzy rules base with 157 rules. Figure 4-31: Structure of fuzzy system for the Cua Dat reservoir Through setting Fuzzy system on the Cua Dat Reservoir, the author defined the graph of optimal release in dry season in 2011 – 2012. The optimal release graph in the Figure 4-32 presents optimal release that can meet water demand more than actual release. Especially, in the driest period in year when water demand increased significantly but the reservoir operated effectively to ensure this demand, while the actual release was lower significantly and did not supply enough water. To evaluate the quantification of effect of fuzzy release in the condition of shortage of water in the dry season, the author used the Nash- Sutcliffe coefficient (NASH) in whole water supply period in dry season. The results are shown that it developed for Fuzzy operating systems for the Cua Dat Reservoir and determine the optimal discharge process in the condition of shortage 74 Trinh Xuan Manh MSc Thesis of water in the dry season, and reservoir can still meet appropriate 80% of water demand throughout the dry season from 2011 to 2012. Actual release was determined through observed data of the reservoir was only to supply about 73% of actual demand. Table 4-16: The NASH for calculation of alternatives Year Releases Nash 2011-2012 Actual release 73% Optimal release 80% Optimal operation of the Cua Dat Reservoir using Fuzzy Logic approach determined efficient release and satisfied the water demands in downstream area more than the actual release as well as optimal operation would decrease water stress and conflicts in dry season. Figure 4-32: Comparison of water demand and fuzzy and actual releases V.2. Routing the release to the downstream The flow in dry season of year of 2011-2012 after the optimal operation of Cua Dat Reservoir is to rout to the downstream by MIKE 11 hydraulic model. The results of flow characteristics are in some of control points will be showed in some of figures and tables below: 0.00 20.00 40.00 60.00 80.00 100.00 120.00 140.00 160.00 1 3 5 7 9 111315171921232527293133353739414345474951535557596163656769717375 R es er v o ir r e le a se /D e m a n d ( M i. m 3 ) Ten-day period Water demand Optimal Release Actual release 75 Trinh Xuan Manh MSc Thesis Figure 5-1: Hydrograph of optimal operation at the Bai Thuong weir Figure 5-2: Hydrograph of optimal operation at the Xuan Khanh station 76 Trinh Xuan Manh MSc Thesis Figure 5-3: Hydrograph of optimal operation at the Giang station Table 5-1: Flow characteristics at the Chu River downstream using optimal operation No. Flow characteristics Q Optimization (m3/s) Q Requirement (m3/s) Q Actual Release (m3/s) 1 Qday Baithuong min (m3/s) 35 30.42 29.3 2 Hday XuanKhanh min (m) 2.25 - 1.98 3 Qday XuanKhanh min (m3/s) 35.8 30.42 33.2 4 Hday Giang min (m) -0.091 - -1.06 According to the results above, it can be seen that the minimum daily discharge at the Bai Thuong weir is around 35 (m3/s) and the Xuan Khanh station is appropriate 36 (m3/s). The table 5-1 makes a comparison between the minimum daily discharges of the optimal operation and minimum daily discharge of requirement in the downstream area, Hence, the minimum daily discharges of the optimal operation is higher than minimum discharge of requirement and discharge of actual release. This showed that the optimal operation of the Cua Dat Reservoir using Fuzzy Logic approach determined efficient release and satisfied the minimum water demands in downstream area. 77 Trinh Xuan Manh MSc Thesis CHAPTER VI CONCLUSIONS AND RECOMMENDATIONS VI.1. Conclusions The study investigated optimal reservoir operation using Fuzzy Logic algorithm for the Cua Dat Reservoir, which is located in Thanh Hoa province, middle of Vietnam. To apply Fuzzy Logic in this study, the author used the Fuzzy Tool Box of MATLAB software to optimize the release from the reservoir and it also simulated to downstream river by using MIKE 11 hydraulic model. In order to have an accuracy and reliable model for the purposes of this study, Fuzzy Logic algorithm is used with the information as existing rule curve of this reservoir, inflow into the reservoir, water level of the reservoir, observation discharge of hydropower plant. The observed data was recorded during the operation of the Cua Dat reservoir having short period because this reservoir has just operated since 2010. Moreover, water demand for water users in the downstream were also one of the important input variables for Fuzzy system, this variable was determined according to data related to water users such as Agriculture, Industry, Domestic, Hydropower, and Environment. The author determined that total water demand is about 4547 Mi.m3. From the results which are optimized from the Fuzzy operating systems for the Cua Dat Reservoir, the optimal discharge process was determined in the condition of shortage of water in the dry season, and reservoir can still meet 80% of water demand throughout the dry season from 2011 to 2012. While that, actual release was determined through observed data of the reservoir was only to supply about 73% of actual demand. To evaluate the efficiency of the optimal operation in the downstream river, the author has used two models including MIKE NAM and MIKE 11HD as an efficient way as mentioned in the literature review of the thesis. For model calibration and validation process, the hydrological and meteorological data was collected such as rainfall, evaporation and discharge from the stations on Ma- Chu river system in 2006 and 2008. To access performance of the models, the Nash- Sutcliffe coefficient (NASH) was used. For calibration period of NAM model, the NASH is 0.69 at the station indicated an acceptable value. For the validation period of NAM model, the 78 Trinh Xuan Manh MSc Thesis result indicated that between observed and simulated hydrograph have a good fit and similar pattern. NASH = 0.73 was acceptable value and indicated well validated model. Similarly, within MIKE 11 HD model calibration and validation. The results of model calibration at Ly Nhan, Xuan Khanh and Giang stations showed that having a good fit between observed and simulated data about pattern and peak value for the most of the stations. The NASH for stations varied from 0.87 to 0.92, the difference of two peaks is rather low. The results for model validation of MIKE 11 HD, the NASH for the same stations varied from 0.83 to 0.94, so that within the highest results, developed MIKE 11HD model could be confidently used for the purpose of this study. As mention above, the author used MIKE 11 model to simulate the release after optimal operation from Cua Dat Reservoir and the results illustrated that the minimum daily discharge at the Bai Thuong weir is around 35 (m3/s) and the Xuan Khanh station is appropriate 36 (m3/s). The minimum daily discharge of the optimal operation is higher than minimum daily discharge of requirement and actual release. Finally, those showed that the optimal operation of the Cua Dat Reservoir using Fuzzy Logic approach determined efficient release and satisfied the water demands in downstream area. The optimal operation would decrease water stress and conflicts in dry season. VI.2. Recommendations During the process of this thesis, the author has some following recommendations and suggestions for the future possible studies: It is very necessary to collect observed data during the operation of the Cua Dat Reservoir. Because the inputs for Fuzzy system need a long time period to calculate and design a system exactly. To determine exactly the water demand for agriculture sector using CROPWAT model need to collect many information of different crop plants in cultivated area. Therefore, the information relevant to water use of domestic is also necessary to determine water demand of system. One of limitations of this study, the author only used triangular membership function for the inputs and outputs of Fuzzy system. Moreover, using other data as rainfall, reservoir storage should pay attention as inputs of the Fuzzy system of Cua Dat Reservoir. The fuzzy rules in this study were defined by opinion of the author, 79 Trinh Xuan Manh MSc Thesis however, it is strongly recommended that the fuzzy rules should derive from a program or software in order to increase the efficiency of the system. In future, it is necessary to set up more hydrological and meteorological stations on Ma-Chu river basin to measure discharge, water level and meteorological factors; Modeling future events should attend to the dynamic changing of factors driving the changing of hydrological and hydraulic scheme of the river basin. For next study, the author really wants to apply the Fuzzy Logic algorithm on operation of a reservoir system to optimize operation of them and address all issues relevant to the river basin which has many reservoirs located in. 80 Trinh Xuan Manh MSc Thesis REFERENCES Bahremand A, De Smedt F. (2007). Distributed Hydrological Modeling and Sensitivity Analysis in Torysa Watershed, Slovakia. Water Resources Management. 22:393– 408 Cheng C.T, Wang W.C, Xu D.M, Chau K. W. (2008). Optimizing Hydropower Reservoir Operation Using Hybrid Genetic Algorithm and Chaos. Water Resources Management 22:895–909 DHI software. (2011). MIKE 11 Reference Manual. Dansih Hydraulic Institute. Hirad Mousavi, A.S. Ramamurthy. (2000). Optimal design of multi-reservoir systems for water supply. Advances in Water Resources, 23: 613 - 624. Hydrology Engineering Center (1991). Optimization of Multiple-Purpose Reservoir System Operations: A Review of Modeling and Analysis Approaches. Research Document No.34. Retrieved from : www.dtic.mil/dtic/tr/fulltext/u2/a236080.pdf Habese M, Nagayama Y. (2002). Reservoir operation using Neural Network and fuzzy system for dam control and operation support. Advances in engineering software (33) 245 -26. Institute of Water Resources Planning. (2003). Master Plan for Use and Protection of Water Resources. 9-56 pp Jairaj P. G, Vedula S. (2001). Multi-reservoir System Optimization using Fuzzy Mathematical Programming. Water Resources Management 14: 457–472, 2000. Kmenl A.H. (2008). Application of Hydraulic MIKE 11 model for the Euphrates river in Iraq. Slovak journal of civil engineering 2008/2, 1-7 pp. Khai N.H and Others. (2011). Research on technology to operate reservoir systems which to prevent flood, regulate flood, operate reservoir safely and appropriately using water resources in dry season. National Technology – Science Project (in VietNamese) Long N. L, Henrik Madsen, Dan Rosbjerg. (2007). Simulation and optimization modeling approach for operation of the Hoa Binh reservoir, Viet Nam. Journal of Hydrology, 336, 269 – 281 81 Trinh Xuan Manh MSc Thesis Mukand S. B, Chien N. D, Md. Reaz Akter Mullick, Umamahesh V. Nanduri. (2011). Operation of a hydropower system considering environmental flow requirements: A case study in La Nga river basin, Vietnam. Journal of Hydro-environment Research, 6 63 – 73. Luhandjula M.K., Rangoaga M.J. (2013). An approach for solving a fuzzy multiple objective programming problem. European Journal of Operational Research, 232, 249 – 255. Moeini R, Afshar A, Afshar M.H. (2010). Fuzzy rule-based model for hydropower reservoirs operation. Electrical Power and Energy Systems, 33, 171-178. MARD (2013). Cua Dat reservoir operation policy. 5-8 pp (In Vietnamese) MARD (2014). Final engineering report of Cua Dat Reservoir in operation period. Project of Cua Dat Reservoir. Nagesh K. D, Falguni Baliarsingh, Srinivasa R.K. (2009). Optimal Reservoir Operation for Flood Control Using Folded Dynamic Programming. Water Resources Management (2010) 24:1045–1064. Omid B. H, Abbas Afshar, Miguel A.M (2008). Design-Operation of Multi- Hydropower Reservoirs: HBMO Approach. Water Resources Management (2008) 22:1709–1722. Piman T, Cochrane T. A, Arias M.E, Green A. and Dat N. D. (2012). Assessment of Flow Changes from Hydropower Development and Operations in Sekong, Sesan and Srepok Rivers of the Mekong Basin. Journal of Water Resources Planning and Management, 1061/(ASCE)WR.1943-5452.0000286 Panigrahi D.P, Mujumdar P. P. (2000). Reservoir Operation Modelling with Fuzzy Logic. Water Resources Management 14: 89–109, 2000. Rama Mehta, Sharad K. Jain. (2009). Optimal Operation of a Multi-Purpose Reservoir Using Neuro-Fuzzy Technique. Water Resources Management (2009) 23:509– 529. 82 Trinh Xuan Manh MSc Thesis Tuan L.Q. (2012) Impacts on flow regime caused by dam construction for hydropower generation in the upper Se San river basin, using SWAT and WAFLEX models, Msc thesis, UNESCO- IHE, Netherlands, 19 pp. Vasantharajan S., Optimal Decisions Inc.; R. Al-Hussainy, Amerada Hess Ltd.; and R.F. Heinemann, Berry Petroleum Co. (2006). Applying Optimization Technology in Reservoir Management. JPT Distinguished Author Series. Van Waveren R.H, Groot S, Scholten H, van Geer F.C, . Wösten J.H.M, Koeze R.D, Noort J.J. (nd). (1999). Good modelling practice handbook. Dutch Dept. of Public Works, Institute for Inland Water Management and Waste Water Treatment. Source: Tuyen, M.H. (2009). Research on optimal operation for reservoir system on Huong river basin in dry season. Journal of Meteorology and Hydrology. Hung, N.T, Hung L.N. (2010). Models for optimal operation of multiple-purpose reservoir. Journal of Meteorology and Hydrology. Nghia T.T. (2009). Building operation process for multi-reservoir including Hoa Binh, Thac Ba, Tuyen Quang supplying water in dry season for downstream of Hong- ThaiBinh river basin. Journal of Hydrology. i Trinh Xuan Manh MSc Thesis APPENDICES Appendix 1: Water demand of some of crops in downstream area Figure A-1: Seasonal period and chart of water requirement of Maize in 2011 Table A-1: Water requirement of Maize in 2011 Months Decades Stage Kc ETc ETc Eff rain Irr. Req. coeff mm/day mm/dec mm/dec mm/dec Sep 3 Init 0.3 1.08 1.1 6.1 1.1 Oct 1 Init 0.3 1 10 47.5 0 Oct 2 Deve 0.3 0.93 9.3 38.2 0 Oct 3 Deve 0.48 1.42 15.6 27 0 Nov 1 Deve 0.75 2.14 21.4 10.2 11.2 Nov 2 Deve 1.01 2.77 27.7 0 27.7 Nov 3 Mid 1.2 3.03 30.3 3.3 26.9 Dec 1 Mid 1.2 2.79 27.9 11.5 16.4 Dec 2 Mid 1.2 2.53 25.3 15 10.4 Dec 3 Mid 1.2 2.47 27.2 10.2 17 Jan 1 Late 1.1 2.2 22 1.5 20.5 Jan 2 Late 0.82 1.6 16 0 16 Jan 3 Late 0.52 1 11 0.4 10.6 Feb 1 Late 0.35 0.66 0.7 0.1 0.7 Total 245.5 171 158.5 ii Trinh Xuan Manh MSc Thesis Figure A-2: Seasonal period and chart of water requirement of Sweet potatoes in 2011 Table A-2: Water requirement of Sweet potatoes in 2011 Months Decades Stages Kc ETc ETc Eff rain Irr. Req. coeff mm/day mm/dec mm/dec mm/dec Oct 2 Init 0.5 1.53 1.5 3.8 1.5 Oct 3 Init 0.5 1.48 16.3 27 0 Nov 1 Init 0.5 1.43 14.3 10.2 4.1 Nov 2 Deve 0.56 1.54 15.4 0 15.4 Nov 3 Deve 0.77 1.96 19.6 3.3 16.3 Dec 1 Deve 0.99 2.31 23.1 11.5 11.6 Dec 2 Mid 1.15 2.43 24.3 15 9.3 Dec 3 Mid 1.16 2.38 26.2 10.2 16 Jan 1 Mid 1.16 2.32 23.2 1.5 21.7 Jan 2 Mid 1.16 2.26 22.6 0 22.6 Jan 3 Late 1.15 2.2 24.2 0.4 23.9 Feb 1 Late 1.03 1.95 19.5 1 18.4 Feb 2 Late 0.89 1.66 16.6 1.3 15.3 Feb 3 Late 0.78 1.49 8.9 5 5.6 Total 255.7 90.1 181.7 iii Trinh Xuan Manh MSc Thesis Figure A-3: Seasonal period and chart of water requirement of Spring paddy in 2012 Table A-3: Water requirement of Spring paddy in 2012 Months Decades Stages Kc ETc ETc Eff rain Irr. Req. coeff mm/day mm/dec mm/dec mm/dec Sep 2 Nurs 1.04 0.38 0.4 5.7 0.4 Sep 3 Nurs/LPr 1.04 0.7 7 53.8 20.4 Oct 1 Nurs/LPr 1.08 3.61 36.1 49.7 0 Oct 2 Init 1.09 3.52 35.2 46.9 70 Oct 3 Init 1.25 3.75 41.3 44.8 0 Nov 1 Deve 1.26 3.49 34.9 43.6 0 Nov 2 Deve 1.44 3.67 36.7 41.9 0 Nov 3 Deve 1.68 4.04 40.4 36.6 3.8 Dec 1 Mid 1.92 4.31 43.1 31.2 11.9 Dec 2 Mid 1.99 4.15 41.5 26.5 15 Dec 3 Mid 1.99 3.84 42.3 20.1 22.1 Jan 1 Mid 1.99 3.53 35.3 12.2 23.2 Jan 2 Late 1.97 3.19 31.9 5.3 26.6 Jan 3 Late 1.68 2.8 30.8 5 25.7 Feb 1 Late 1.32 2.25 22.5 4.6 17.9 Feb 2 Late 1.04 1.83 11 1.9 9.4 490.4 429.9 246.4 iv Trinh Xuan Manh MSc Thesis FigureA-4: Seasonal period and chart of water requirement of winter paddy in 2012 Table A-4: Water requirement of winter paddy in 2012 Months Decades Stages Kc ETc ETc Eff rain Irr. Req. coeff mm/day mm/dec mm/dec mm/dec Jul 1 Nurs 1.04 0.49 2.5 21.9 0 Jul 2 Nurs/LPr 1.06 2.76 27.6 44 21.2 Jul 3 Nurs/LPr 1.08 4.85 53.3 46.7 76.6 Aug 1 Init 1.18 5.12 51.2 50.1 55.3 Aug 2 Init 1.25 5.2 52 52.7 0 Aug 3 Deve 1.26 5.03 55.3 53.6 1.7 Sep 1 Deve 1.31 4.99 49.9 55.4 0 Sep 2 Mid 1.35 4.88 48.8 57.1 0 Sep 3 Mid 1.35 4.71 47.1 53.8 0 Oct 1 Mid 1.35 4.53 45.3 49.7 0 Oct 2 Late 1.29 4.16 41.6 46.9 0 Oct 3 Late 1.07 3.21 35.3 44.8 0 Nov 1 Late 0.93 2.58 5.2 8.7 5.2 Total 514.9 585.5 159.9 : v Trinh Xuan Manh MSc Thesis Figure A-5: Seasonal period and chart of water requirement of sugar cane in 2012 Table A-5: Water requirement of sugar cane in 2012 Months Decades Stages Kc ETc ETc Eff rain Irr. Req. coeff mm/day mm/dec mm/dec mm/dec May 1 Init 0.79 3.11 3.1 2.9 0 May 2 Init 0.4 1.73 17.3 39.7 0 May 3 Init 0.4 1.8 19.8 40.9 0 Jun 1 Deve 0.4 1.89 18.9 41.5 0 Jun 2 Deve 0.5 2.44 24.4 44 0 Jun 3 Deve 0.64 3.08 30.8 44.3 0 Jul 1 Deve 0.78 3.71 37.1 43.8 0 Jul 2 Deve 0.92 4.32 43.2 44 0 Jul 3 Deve 1.07 4.83 53.1 46.7 6.3 Aug 1 Mid 1.21 5.24 52.4 50.1 2.3 Aug 2 Mid 1.24 5.16 51.6 52.7 0 Aug 3 Mid 1.24 4.93 54.2 53.6 0.6 Sep 1 Mid 1.24 4.71 47.1 55.4 0 Sep 2 Mid 1.24 4.48 44.8 57.1 0 Sep 3 Mid 1.24 4.32 43.2 53.8 0 Oct 1 Mid 1.24 4.15 41.5 49.7 0 Oct 2 Mid 1.24 3.99 39.9 46.9 0 Oct 3 Mid 1.24 3.72 40.9 44.8 0 vi Trinh Xuan Manh MSc Thesis Months Decades Stages Kc ETc ETc Eff rain Irr. Req. coeff mm/day mm/dec mm/dec mm/dec Nov 1 Mid 1.24 3.44 34.4 43.6 0 Nov 2 Mid 1.24 3.17 31.7 41.9 0 Nov 3 Mid 1.24 2.97 29.7 36.6 0 Dec 1 Mid 1.24 2.78 27.8 31.2 0 Dec 2 Mid 1.24 2.58 25.8 26.5 0 Dec 3 Mid 1.24 2.39 26.3 20.1 6.1 Jan 1 Mid 1.24 2.2 22 12.2 9.8 Jan 2 Mid 1.24 2.01 20.1 5.3 14.8 Jan 3 Mid 1.24 2.06 22.7 5 17.6 Feb 1 Late 1.22 2.09 20.9 4.6 16.3 Feb 2 Late 1.18 2.06 20.6 3.2 17.4 Feb 3 Late 1.13 2.12 16.9 5.9 11 Mar 1 Late 1.09 2.16 21.6 10 11.6 Mar 2 Late 1.04 2.18 21.8 12.8 9.1 Mar 3 Late 0.98 2.44 26.9 11.1 15.8 Apr 1 Late 0.93 2.66 26.6 5.9 20.7 Apr 2 Late 0.88 2.85 28.5 3.2 25.3 Apr 3 Late 0.83 2.99 29.9 14.3 15.6 May 1 Late 0.79 3.11 28 26.1 0 Total 1145.4 1131.5 200.4 vii Trinh Xuan Manh MSc Thesis Figure A-6: Seasonal period and chart of water requirement of Maize in 2012 Table A-6: Water requirement of Maize in 2012 Months Decades Stages Kc ETc ETc Eff rain Irr. Req. coeff mm/day mm/dec mm/dec mm/dec May 2 Init 0.3 1.3 7.8 23.8 0 May 3 Init 0.3 1.35 14.8 40.9 0 Jun 1 Deve 0.37 1.73 17.3 41.5 0 Jun 2 Deve 0.62 2.98 29.8 44 0 Jun 3 Deve 0.87 4.16 41.6 44.3 0 Jul 1 Mid 1.12 5.28 52.8 43.8 9 Jul 2 Mid 1.19 5.55 55.5 44 11.5 Jul 3 Mid 1.19 5.35 58.8 46.7 12.1 Aug 1 Mid 1.19 5.14 51.4 50.1 1.3 Aug 2 Late 1.17 4.87 48.7 52.7 0 Aug 3 Late 0.94 3.72 41 53.6 0 Sep 1 Late 0.64 2.44 24.4 55.4 0 Sep 2 Late 0.42 1.52 9.1 34.2 0 Total 453 575.1 33.9 viii Trinh Xuan Manh MSc Thesis Figure A-7: Seasonal period and chart of water requirement of Sweet Potatoes in 2012 Table A-7: Water requirement of Sweet Potatoes in 2012 Months Decades Stages Kc ETc ETc Eff rain Irr. Req. coeff mm/day mm/dec mm/dec mm/dec Oct 2 Init 0.5 1.61 1.6 4.7 1.6 Oct 3 Init 0.5 1.5 16.5 44.8 0 Nov 1 Init 0.5 1.39 13.9 43.6 0 Nov 2 Deve 0.56 1.44 14.4 41.9 0 Nov 3 Deve 0.77 1.86 18.6 36.6 0 Dec 1 Deve 0.99 2.22 22.2 31.2 0 Dec 2 Mid 1.15 2.4 24 26.5 0 Dec 3 Mid 1.16 2.23 24.5 20.1 4.4 Jan 1 Mid 1.16 2.05 20.5 12.2 8.3 Jan 2 Mid 1.16 1.87 18.7 5.3 13.5 Jan 3 Late 1.14 1.9 20.9 5 15.9 Feb 1 Late 1.03 1.76 17.6 4.6 13 Feb 2 Late 0.9 1.58 15.8 3.2 12.5 ix Trinh Xuan Manh MSc Thesis Months Decades Stages Kc ETc ETc Eff rain Irr. Req. coeff mm/day mm/dec mm/dec mm/dec Feb 3 Late 0.79 1.49 8.9 4.4 6 Total 238.2 284.3 75.2 Table A-8: Monthly water demand of agriculture of whole area in the Cua Dat reservoir downstream in 2012 Water demand of Bac Song Chu cultivated area in 2012 Months 1 2 3 4 5 6 7 8 9 10 11 12 Year Q(m3/s) 12.7 15.1 16.0 17.1 12.7 26.6 18.9 8.03 4.01 5.19 8.03 33.5 14.8 W(106m3) 34.5 36.9 43.4 44.1 34.5 69.0 51.0 21.9 11.4 14.8 21.2 90.2 473.8 Water demand of Nam Song Chu cultivated area in 2012 Months 1 2 3 4 5 6 7 8 9 10 11 12 Year Q(m3/s) 42.4 42.7 61.3 57.1 22.3 0.226 16.3 0.117 0.389 0.489 0.231 14.5 21.5 W(106m3) 113.7 103.3 164.2 148.1 59.8 0.6 43.6 0.3 1.0 1.3 0.61 39.0 675.9 Water demand of agriculture of whole area in 2012 Months 1 2 3 4 5 6 7 8 9 10 11 12 Year Q(m3/s) 55.2 57.8 77.4 74.1 35.0 26.8 35.2 8.15 4.39 5.68 8.27 48.1 36.3 W(106m3) 148.3 140.2 207.7 192.6 94.3 69.5 94.7 22.2 12.4 16.1 21.8 129.2 1149.7 x Trinh Xuan Manh MSc Thesis Appendix 2: Water demand in ten-day period in year of 2012 Table A-9: Water demand in ten-day period in 2012 Months Period Demand in ten-day unit in M.m3 Jan 1-10 73.57 Jan 11-20 59.08 Jan 21-31 63.18 Feb 1-10 83.53 Feb 11-20 87.92 Feb 21-28 79.82 Mar 1-10 87.19 Mar 11-20 86.16 Mar 21-31 96.58 Apr 1-10 71.40 Apr 11-20 101.15 Apr 21-30 55.45 May 1-10 50.53 May 11-20 59.67 May 21-31 102.33 Jun 1-10 112.62 Jun 11-20 107.10 Jun 21-30 48.04 Jul 1-10 46.21 Jul 11-20 60.94 Jul 21-31 72.50 Aug 1-10 37.00 Aug 11-20 63.07 Aug 21-30 100.00 Sep 1-10 112.11 Sep 11-20 124.16 Sep 21-31 118.82 xi Trinh Xuan Manh MSc Thesis Months Period Demand in ten-day unit in M.m3 Oct 1-10 121.28 Oct 11-20 120.93 Oct 21-31 131.08 Nov 1-10 115.59 Nov 11-20 106.62 Nov 21-30 119.59 xii Trinh Xuan Manh MSc Thesis Appendix 3: Observed and calculated data of inputs and outputs in the Fuzzy system Figure A-8: Average daily discharge into the Cua Dat Reservoir in 2011 and 2012 Figure A-9: Average daily water level of the Cua Dat Reservoir in 2011 Figure A-10: Average daily water level of the Cua Dat Reservoir in 2012 0 500 1000 1500 2000 2500 3000 0 50 100 150 200 250 300 350 D is ch a rg e (m 3 /s ) Days2011 2012 75.00 80.00 85.00 90.00 95.00 100.00 105.00 1 51 101 151 201 251 301 351 W a te r l ev el ( m ) Days 70 75 80 85 90 95 100 105 1 51 101 151 201 251 301 351 W a te r l ev el ( m ) Days xiii Trinh Xuan Manh MSc Thesis Figure A-11: Average daily turbin discharge of the Cua Dat hydropower plant in 2012 Figure A-12: Average daily turbin discharge of the Cua Dat hydropower plant in 2011 0.00 20.00 40.00 60.00 80.00 100.00 120.00 140.00 160.00 1 21 41 61 81 101 121 141 161 181 201 221 241 261 281 301 321 341 361 T u rb in d is ch a rg e (m 3 /s ) Days 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 1 1 3 2 5 3 7 4 9 6 1 7 3 8 5 9 7 1 0 9 1 2 1 1 3 3 1 4 5 1 5 7 1 6 9 1 8 1 1 9 3 2 0 5 2 1 7 2 2 9 2 4 1 2 5 3 2 6 5 2 7 7 2 8 9 3 0 1 3 1 3 3 2 5 3 3 7 3 4 9 3 6 1 T u rb in d is c h a r g e (m 3 /s ) Days xiv Trinh Xuan Manh MSc Thesis Appendix 4: Inputs of MIKE 11 HD model Figure A-13: Average daily discharge at the Cam Thuy and Cua Dat station in 2006 Figure A-14: Average daily water level at the Hoang Tan and Kim Tan station in 2006 0 500 1000 1500 2000 2500 3000 1 51 101 151 201 251 301 351 D is ch a g re ( m 3 /s ) Days Cam Thuy Cua Dat -2 0 2 4 6 8 10 1 51 101 151 201 251 301 351 W a te r L ev el (m 3 /s ) Days Kim Tan Hoang Tan xv Trinh Xuan Manh MSc Thesis Figure A-15: Average daily discharge at the Cam Thuy and Cua Dat station in 2008 Figure A-16: Average daily water level at the Hoang Tan and Kim Tan station in 2008 0 500 1000 1500 2000 2500 3000 3500 4000 1 51 101 151 201 251 301 351 D is ch a rg e (m 3 /s ) Days Cam Thuy Cua Dat -2 0 2 4 6 8 10 12 14 1 51 101 151 201 251 301 351 W a te r l ev el (m ) Days Kim Tan Hoang Tan xvi Trinh Xuan Manh MSc Thesis Appendix 5: Fuzzy rules base for the Cua Dat Reservoir Rules No. Res.levels (and) Inflows (and) Demands (Then) Releases 1 If Med.High V.Low Low Low 2 If Low.med V.Low low Low 3 If Low.med V.Low low Low.Med 4 If Low.med V.Low Low Medium 5 If Low.med V.Low V.Low Low 6 If Low.med V.Low V.Low V.Low 7 If Low V.Low Very.High Low.med 8 If Medium V.Low V.High Med.High 9 If Low V.Low Low V.Low 10 If Low V.Low Medium Low.Med 11 If Low.Med V.low Medium Medium 12 If Low Low V.Low Low 13 If Medium Low Low Low 14 If Low.Med V.Low Low.Med Medium 15 If V.Low V.Low Medium Low 16 If V.Low V.Low High High 17 If V.Low V.Low Low Low 18 If V.Low V.Low V.Low V.Low 19 If V.Low Low Low Low 20 If V.Low Low Medium Low.Med 21 If V.Low Low Medium Medium 22 If Low Low Medium Medium 23 If Low Low Medium Low.Med 24 If Low Med.High High Med.High 25 If Low Med.High High High 26 If Low.med Med.High High High 27 If Medium Medium High High 28 If Medium Med.High High High 29 If Med.High medium High High 30 If Med.High Med.High High High 31 If Medium Very.High High High 32 If Med.High Very.High High High 33 If Med.High Medium Very.High V.V.High 34 If High Medium Very.High V.V.High 35 If High Med.High High V.V.High 36 If Medium Low Very.High V.V.High 37 If Med.High Low Very.High V.V.High 38 If Medium Very.High Very.High V.V.High 39 If Med.High Very.High Very.High V.V.High 40 If Medium Medium High V.V.High 41 If Medium Medium Very.High V.V.High xvii Trinh Xuan Manh MSc Thesis Rules No. Res.levels (and) Inflows (and) Demands (Then) Releases 42 If Medium High High High 43 If Medium High Very.High V.V.High 44 If Med.High Medium High V.V.High 45 If Med.High Medium Very.High V.V.High 46 If Med.High High High V.V.High 47 If Med.High High Very.High V.V.High 48 If Medium Low High High 49 If Medium Low Very.High High 50 If Med.High Low High High 51 If Med.High Low Very.High High 52 If High Low High V.V.High 53 If High Low Very.High V.V.High 54 If Medium Low Medium Medium 55 If Medium Low Medium Med.high 56 If Med.High Low Medium medium 57 If Med.High Low Medium Med.high 58 If Med.High Low High Med.High 59 If Medium Low high High 60 If Medium V.Low Low Low.Med 61 If Medium V.Low Low Med 62 If Medium V.Low Medium Med 63 If Medium V.Low Medium Low.Med 64 If Med.High Low Low Low 65 If Medium Low Low Low 66 If Med.High Low Low Low 67 If Med.High Low Low Medium 68 If High Low Low Medium 69 If Med.High Low Medium Medium 70 If high Low Medium Medium 71 If Med.High V.low Low Low.med 72 If High V.low Low Low.med 73 If Med.High V.low Low Low 74 If Med.High V.low Medium Low 75 If Med.High Low Low Low 76 If Med.High Low Medium Low 77 If High V.low Low Low 78 If High V.low Medium Low 79 If High Low Low Low 80 If High Low Medium Low 81 If Med.High V.low Medium Medium 82 If High V.low Medium Medium 83 If Medium V.low Medium Medium 84 If Medium V.low Medium Med.high xviii Trinh Xuan Manh MSc Thesis Rules No. Res.levels (and) Inflows (and) Demands (Then) Releases 85 If Medium V.low High Med.high 86 If Medium V.low High Medium 87 If Medium Low Medium Medium 88 If Medium Low Medium Med.high 89 If Med.High V.low Medium Medium 90 If Med.High V.low Medium Med.high 91 If Med.High V.low High Med.high 92 If Med.High V.low High Medium 93 If Med.High Low Medium Medium 94 If Med.High Low Medium Med.high 95 If Low.med v.low Medium Med.high 96 If Low V.low High Med.high 97 If Low.med V.low High Med.high 98 If Low V.low Low Low 99 If Low V.low Low Low.med 100 If Low.med v.low Low Low 101 If Low.med V.low Low Low.med 102 If V.Low V.low V.low Low.med 103 If V.Low V.low Low Low.med 104 If V.Low V.low Low Medium 105 If Low V.low Low Medium 106 If V.Low Low High Med.high 107 If V.Low Low High High 108 If Low Low High Med.high 109 If Low Low High High 110 If Low V.low High High 111 If Low.Med V.low High High 112 If Low.Med V.low V.high High 113 If Low V.low V.high High 114 If V.Low V.low Low V.low 115 If V.Low V.low Medium V.low 116 If Low V.low Low V.low 117 If Low M.low Medium V.low 118 If Low Med.high V.low Low.med 119 If Low Med.high Low Low.med 120 If Low.med Med.high High Med.high 121 If Low.med High High Med.high 122 If Low.med Med.high High High 123 If Low.med High High High 124 If Medium Med.high High Med.high 125 If Medium High High Med.high 126 If Medium Med.high High High 127 If Medium High High High xix Trinh Xuan Manh MSc Thesis Rules No. Res.levels (and) Inflows (and) Demands (Then) Releases 128 If Med.High High High high 129 If Med.High High V.high High 130 If Medium High V.high High 131 If Medium Medium High High 132 If Medium Medium V.high High 133 If Medium Med.high High High 134 If Medium Med.high V.high High 135 If High High V.high High 136 If High Medium High High 137 If High Medium V.high High 138 If High Med.high High High 139 If High Med.high V.high High 140 If Medium Medium High V.v.high 141 If Medium Medium V.high V.v.high 142 If High Medium High V.v.high 143 If High Medium V.high V.v.high 144 If Medium Low High V.v.high 145 If Medium Low V.high V.v.high 146 If High Low High V.v.high 147 If High Low V.high V.v.high 148 If Medium V.low High V.v.high 149 If Medium V.low V.high V.v.high 150 If Med.High V.low High V.v.high 151 If Med.High V.low V.high V.v.high 152 If medium V.low High Med.high 153 If Med.high V.low High Med.high 154 If Low.med V.low High High 155 If Low.med Low High High 156 If Medium V.low High High 157 If Medium Low High High xx Trinh Xuan Manh MSc Thesis Appendix 6: Existing operation rule curve of the Cua Dat Reservoir Table A-10: Operation curves of the Cua Dat Reservoir Periods Crumble curve Limited curve 30/VI 97 73 31/VII 100 82 31/VIII 104 85 30/IX 109 96 31/X 110 105 30/XI 112 106 31/XII 112 106 31/I 112 103 28/II 108 97 31/III 105 90 30/IV 103 83 31/V 99 77 30/VI 97 73 Figure A-17: The map of operation curves of the Cua Dat Reservoir 70 80 90 100 110 120 1/7 1/8 1/9 1/10 1/11 1/12 1/1 1/2 1/3 1/4 1/5 1/6 1/7 B A B C C D E Flood control increment level=110,00m Max Water level=119,05m Dam elevation 121.30 m 3 1 2 Dead storage level=73m Flood control level=112,00m