استفاده از مدل درختی M5 برای تعیین ضریب دبی سرریز لبه پهن مستطیلی

نوع مقاله : مقاله پژوهشی

نویسندگان

دانشکده کشاورزی، دانشگاه تبریز، تبریز، ایران

چکیده

سرریز لبه پهن سازه‌ای ساده برای اندازه‌گیری دبی جریان در کانال‌های انتقال آب است. ویژگی شکل آن و وزن زیاد آن باعث می‌شود تا اغلب به عنوان سرریز سدها نیز بکار رود. گاهی نیز این نوع سازه برای بدنه سد مد نظر قرار می‌گیرد. در این تحقیق توانایی شبکه عصبی مصنوعی و مدل درختی M5 در تخمین ضریب دبی (Cd) سرریز لبه پهن بررسی شده و نتایج این دو مدل با روش رگرسیون غیرخطی چند متغیره لجستیک قابل اعمال روی داده‌های گسسته مقایسه شده است. برای این کار چهار سری داده حاصل از تحقیقات متفاوت روی سرریزهای لبه پهن مستطیلی استفاده شده و پارامترهای بی‌بعد H1/L و H1/P به عنوان ورودی مدل‌ها در نظر گرفته شده و پارامتر هدف Cd به عنوان خروجی از مدل‌ها استخراج شده است. نتایج حاصله نشان داد که هر سه روش مذکور نتایج نسبتا دقیقی را جهت تخمین ضریب دبی سرریز لبه پهن ارائه می‌دهند (ANN: R=0.966 ، M5Rule: R=0.935 و Regression: R=0.84) ولی به دلیل ارائه روابط خطی ساده و قابل فهم توسط مدل درختی M5، این روش می-تواند به عنوان روشی کاربردی و جایگزین برای محاسبه ضریب دبی مد نظر قرار گیرد. همچنین H1/L مهمترین پارامتر دخیل در محاسبه ضریب دبی سرریز لبه پهن مستطیلی می‌باشد. تحلیل مدل درختی M5 نشان داد که 4 قانون با معادلات خطی متفاوت، الگوی تغییرات Cd را مدل می‌کند. تحلیل رگرسیون غیرخطی نشان داد که نقطه H1/L=0.22 محل تلاقی کلیه منحنی‌های تغییرات Cd به شمار می‌آید.

کلیدواژه‌ها

موضوعات


عنوان مقاله [English]

Prediction of discharge coefficients for broad-crested weirs using expert systems

نویسندگان [English]

  • Farzin Salmasi
  • Farnaz Nahrain
  • Ali Taheri aghdam
Department of Water Engineering, Faculty of Agriculture, University of Tabriz, Tabriz, Iran
چکیده [English]

Broad-crested weirs can be used to make discharge measurements in irrigation canals; the entrance of stepped weirs or chutes is sometimes designed as a broad-crested weir structure. These structures are also sometimes used for the dam body. In this study, the Artificial Neural Network (ANN) and M5 model tree methods are used to predict discharge coefficients (Cd) for broad-crested weirs. The results from these two models are compared with nonlinear regression equations. Four series of data obtained from different rectangular broad-crested weirs have been used and important dimensionless parameters have been defined. Results show that the ANN procedure is superior to the M5 model and regression approaches. The accuracy for ANN is quantified by R=0.966 and RMSE=0.038. All three methods are able to provide a reasonable prediction for Cd; the M5 model tree provides four linear equations that can be used to estimate Cd. The shape of the Cd contours shows that the effect of weir height (P) exceeds that of the weir length (L).

کلیدواژه‌ها [English]

  • Discharge coefficient
  • broad-crested weir
  • artificial neural network
  • nonlinear regression
  • M5 model tree
[1] K. Badr,D.Mowla, Development of Rectangular Broad-crested Weirs for Flow Characteristics and Discharge Measurement, KSCE Journal of Civil Engineering 19(1)  (2015):136-141
[2] A. Masoodi,  P. Parsamehr,  F. Salmasi,  S. Pureskandar, Regression Analysis, Genetic Programming and ANN to Predict Discharge Coefficient of Compound Broad Crested Weir, Journal of Water and Soil, 26(4) (2012): 933-942.(in Persian)
[3] E. Goodarzi,  J. Farhoudi, N.Shokr, Flow characteristics of rectangular broad- crested weirs with sloped upstream face, Journal of Hydrology and Hydromechanics, 60(2) (2012): 87–100.
[4] W.H. Hager, M. Schwalt, Broad-crested weir. J. Irrig. Drain. Eng, ASCE, 120(1) (1994): 13–26. Discussion: 12(2) (1994): 222–226.
[5] R. E. Horton, Weir experiments, coefficients, and formulas. Dept. of the Interior, U.S. Geological Survey, Water-Supply and Irrigation Paper 200. (1907). Government Printing Office, Washington, D.C.
[6] C. Keutner, Strrmungsvorg inge an breitkronigen Wehrkrpern und an Einlaufbauwerken. Bauingenieur, 15(37/38) (1934): 366-371.
[7] W. D. Moss, Flow separation at the upstream edge of a square-edged broad crested weir. J. Fluid Mech., 52(2) (1972): 307- 320.
[8] J. Singer, Square-edged broad-crested weir as a flow measurement device. Water and Water Engrg. 68(6) (1964): 229-235.
[9] A. J. M. Harrison, Some comments on the square-edged broad-crested weir. Water and Water Engrg. 68(11) (1964): 445-448.
[10] K. G. Ranga Raju, I. Ahmad, Discharge characteristics of suppressed and contacted broad-crested weirs. J. Irrigation and Power (India), 30(4) (1973): 157- 166.
[11] A. D. Crabbe, Some hydraulic features of the square-edged broad-crested weir. Water and Water Engrg. 78(10) (1974): 354-358.
[12] A. S. Ramamurthy, U. S. Tim, M. V. J. Rao, Characteristics of square-edged round-nosed broad-crested weirs. J. Irrig. and Drain. Engrg. ASCE, 114(1) (1988): 61-73; 115(4) (1988): 766.
[13] H. Bazin, Expériences Nouvelles sur l’Ecoulement par Déversoir. Recent experiments on the flow of water over weirs.  Mémoires et Documents, Annales des Ponts et Chaussées, Séries 7, 12(2) (1896): 645–731 (in French).
[14] C. A. Gonzalez, H. Chanson, Experimental measurements of velocity and pressure distribution on a large broad-crested weir. Flow Meas. Instrum., 18(3–4) (2007): 107–113.
[15] A. Khosro jerdi, M.R. Kavianpour, A. Shamsae, A. Daemi, Hydraulic investigation of broad creast weir and curved arch effect, 3rd Iranian Hydraulic Conference, (2002), University of Tehran.( in persian).
[16] N. Shokri, J. Farhodi, A. Ghodarzi, Investigation of the Effect of Downstream Slope of Rectangular Width Overflows on Flow Conditions (Critical Depth ​​and Speed Profile), (2009)Tehran, 7th Iranian Hydraulic Conference. (In Persian).
[17] N. Ajlo, J. Farhodi, Investigating the Effect of Changing Overflow Length on Two Floating Rectangular Widths on Flow Characteristics. Master thesis. (2009), Univercity of Tehran.
[18] R. Daneshfaraz, O. Minaei, J. Abraham, S. Dadashi, A. Ghaderi, 3-D Numerical simulation of water flow over a broad- crested weir whit openings, ISH Journal of Hydraulic Engineering, (2019): 1-9.
[19] M.T. Sattari, A.S. Anli, H. Apaydin, S. Kodal. Decision trees to determine the possible drought periods in Ankara. Atmosfera 25(2012): 65-83.
[20] ZhA. Dayev, Application of artificial neural networks instead of the orifice plate discharge coefficient. (2019), Flow measurement and instrumentation.
[21] D.P. Solomatine, Y. Xue. M5 model trees and neural networks: Application to flood forecasting in the upper reach of the Huai River in China. J. Hydrol. Eng. 9(2004): 491–501.
[22] B. Bhattacharya, D.P. Solomatin. Neural networks and M5 model trees in modeling water level-discharge relationship. Department of Hydroinformatics and Knowledge Management, NESCO-IHE Institute for Water Education, P.O. Box 3015,2601 DA Delft, (2004), the Netherlands.
[23] M. Pal, S. Deswal. M5 model tree based modeling of reference evapotranspiration. Hydrol. Process 23(2009):1437-1443.
[24] A. Komakpanah, S. Bakhtiari, Use of neural network in the design of python injections, Special Issue of Civil Engineering 35 (2009) 91-98.
[25] H.R. Saba, M. Mohsen Kamalian, I. Raeisizadeh, Determining Impending Slip of Slop and Optimized Embankment Operation Volume of Earth Dams Using a Combination of Neural Networks and Genetic Algorithms (GA), Amirkabir Journal of Civil Engineering, 50(4) (2018) 747-754.
[26] T. Honar, S. Pourhamzeh, A neural network model to predict characteristics of Hydraulic Jump in Stilling Basins with Convergent Wall, Water and  soil science, 23(2) (2012) 99-109.
[27] F. Salmasi, G. Yıldırım, A. Masoodi, P. Parsamehr, Predicting discharge coefficient of compound broad-crested weir by using genetic programming (GP) and artificial neural network (ANN) techniques. Arabian Journal of Geosciences. (6), (2013) 2709–2717.
[28] F. Salmasi, J. Abraham, Discussion of “Hydrodynamics of Rectangular Broad-Crested Porous Weirs” by Safarzadeh and Mohajeri, J. Irrig. Drain Eng., 146(4): (2020) 07020003.
[29] B. Nourani, H. Arvanaghi, F. Salmasi, A novel approach for estimation of discharge coeffcient in broad-crested weirs on Harris Hawks Optimization algorithm, Flow Measurement and Instrumentation 79 (2021) 101916.
[30] F. Salmasi, M.T. Sattari, Predicting discharge coefficient of rectangular broad- rested gabion weir using M5 tree model. Iranian Journal of Science and Technology, Transactions of Civil Engineering (Shiraz University). 41 (2): (2017) 205–212.
[31] M. Sharif Zak, Gh.A. Barani, Flow discharge forecasting in compound channels using Artificial Neural Networks. 1th International Conference of Civil and Development, Zibakenar, Iran (In Persian).
[32] J.R. Quinlan, Learning with continuous classes. In proceedings AI, 92 (Adams & Sterling, Eds), (1992):343-348, Singapore: World Scientific.
[33] A. Alikhanzadeh, Data Mining. First Edition, Olom rayaneh press, 14p. (2006).
[34] M. Pal, M5 model tree for land cover classification. International Journal of Remote Sensing. 27(4) (2006): 825-831.
[35] M.R. Fallahi, H. Varvani, S. Goliyan, Precipitation forecasting using regression tree model to flood control. 5th international watershed and water and soil resources management, (2012), Kerman, Iran (In Persian).
[36] D. Alberg, M. Last, A. Kandel, Knowledge discovery in data streams with regression tree methods. WIREs Data Mining Knowl Discov 2(2012), 69-78.
[37] B. Bhattacharya, D.P. Solomatine, Neural networks and M5 model trees in modeling water level-discharge relationship. Department of Hydroinformatics and Knowledge Management, NESCO-IHE Institute for Water Education, P.O. Box 3015, 2601 DA Delft, (2004), The Netherlands.
[38] A.H. Azimi, N. Rajaratnam, Z. David, M. Zhu, Discussion of new theoretical solution of the stage-discharge relationship for sharp-crested and broad weirs by V. Ferro, J. Irrig. Drain Eng., in press (2013): 516-517.