مقایسه روش ‌های شبکه عصبی و نروفازی برای بهبود چهار چوب عملی دراستیک (مطالعه موردی: آبخوان دشت شبستر)

نوع مقاله: مقاله پژوهشی

چکیده

رشد روزافزون جمعیت و افزایش نیازهای آبی،  استفاده از منابع آب شیرین بویژه منابع آب زیرزمینی را افزایش داده است. به همین جهت ارزیابی آسیب­پذیری آب­های زیرزمینی روشی مناسب برای شناخت مناطق آسیب­پذیر و محافظت از این منابع به شمار می­رود. دشت شبستر در استان آذربایجان­شرقی یک منطقه فعال از نظر کشاورزی است و استفاده از منابع آب زیرزمینی در آن به علت کمبود منابع سطحی از اهمیت فوق­العاده زیادی برخوردار است. در این مطالعه از چهار چوب عملی دراستیک برای ارزیابی آسیب­پذیری آبخوان دشت شبستر استفاده شده است. مقدار شاخص آسیب­پذیری دراستیک در منطقه مورد مطالعه برابر3/53 تا 3/118محاسبه شد. با توجه به اینکه ضرایب وزنی اختصاص یافته به هر پارامتردراستیک، تا حدودی از روی نظر کارشناسی است بنابراین هدف اصلی این مطالعه بهبود دراستیک با دو مدل شبکه عصبی و عصبی فازی بوده است. ورودی های دراستیک به عنوان ورودی هر دو مدل هوش مصنوعی قرار داده شدند. شاخص دراستیک تصحیح شده با غلظت نیترات به عنوان خروجی مدل­ها در نظر گرفته شد. مقادیر نیترات به دو دسته آموزش و آزمایش دسته­بندی شد. پس از آموزش هر دو مدل، نتایج مدل در مرحله آزمایش با غلظت نیترات مورد ارزیابی قرار گرفت. نتایج نشان داد هر دو مدل هوش مصنوعی توانایی بالایی جهت بهبود مدل دراستیک دارند. با این وجود مدل عصبی فازی با داشتن ضریب همبستگی بالاتری با نیترات روشی مناسب جهت ارزیابی آسیب­پذیری آبخوان دشت شبستر بوده است.

کلیدواژه‌ها


عنوان مقاله [English]

Comparison of Neural Network and Neuro-Fuzzy Techniques to Improve the DRASTIC Frame Work (Case Study: Shabestar plain Aquifer)

چکیده [English]

Increasing population and rising water requirements have raised the use of freshwater resources such as groundwater. Therefore, assessing the vulnerability of groundwater is a suitable method for identifying the vulnerable areas and protecting these resources. Shabestar plain in East Azarbaijan province is an active agricultural area and the use of groundwater resources in this plain is important due to the shortage of annual precipitation. In this study, the DRASTIC frame work was used to assess the vulnerability of the Shabestar plain aquifer. The amount of DRASTIC vulnerability index in the study area was calculated as 53.3to 118.3. Given that the weights of the DRASTIC frame work were somewhat expert, so the main purpose of this study was improvement of the DRASTIC by two methods of Neural Network and Neuro-Fuzzy. DRASTIC inputs were introduced as inputs of the both artificial intelligence models. The corrected DRASTIC index with nitrate concentration was considered as the outputs of the models. Nitrate values were categorized into two groups of train and test. After training the model the results of the model were evaluated at the test step with nitrate concentration. The results showed that the both artificial intelligence models had the high ability to improve the DRASTIC model. Nevertheless, the neuro-fuzzy model having a higher correlation coefficient with nitrate was a suitable method for assessing the vulnerability of Shabestar plain aquifer.

کلیدواژه‌ها [English]

  • Groundwater
  • Neural network
  • Neuro Fuzzy
  • optimization
  • Shabestar Plain
  • Vulnerability
Aller L, Bennet T, Leher JH, Petty RJ and Hackett G, 1987. DRASTIC: A Standardized System for Evaluating Groundwater Pollution Potential Using Hydro-Geological Settings. EPA/600/2-87/035. Ada, Oklahoma: U.S. Environmental Protection Agency.
Almasri MN, 2008. Assessment of intrinsic vulnerability to contamination for Gaza costal aquifer. Jornal of Environmental Management 88(4): 577–593.
Anonymous, 2017. Study of Shabestar Plain abstraction wells water quantity and quality. East Azarbaijan water and wastewater Company.Iran.
Asefi M, Radmanesh F and Zarei H, 2014. Optimization of DRASTIC model for vulnerability assessment of groundwater resources using analytical hierarchy process (Case study: Andimeshk plain). Journal of Irrigation science and Engineering 37(1):55-67(In Persian). 
Asghari Moghaddam A, Fijani E and Nadiri A, 2015. Optimization of DRASTIC model by artificial intelligence for groundwater vulnerability assessment in Maragheh- Bonab plain. Journal of Geoscience 94:169-176 (In Persian).
Bhatt A, Helle H.B, 2002. Committee neural networks for porosity and permeability prediction from well logs. Geophysical Prospecting 50: 645-660.
Civita M.V, De Maio M, 1998. Mapping groundwater vulnerability in areas impacted by flash food disasters. 13th ESRI European User Conference, France, Italy.
Dixon B, 2005. Applicability of neuro-fuzzy techniques in predicting groundwater vulnerability. a GIS- based sensitivity analysis. Journal of Hydrology 309 (1-4): 17-38.    
Fijani E, Nadiri AA, Asghari Moghaddam A, Tsai F, Dixon B, 2013. Optimization of DRASTIC method by supervised committee machine artificial intelligence to assess ground water vulnerability for Maraghe-Bonab plain aquifer, Iran. Journal of Hydrology 503: 89-100.                                                                                                                     
Ghanbari N, Rangzan K, kabolizade M and Moradi P, 2017. Improve the results of the DRASTIC model using artificial intelligence methods to assess groundwater vulnerability in Ramhormoz alluvial plain aquifer. Journal of Water and Soil Conservation 24(2): 45-65.
Hamamin DF, Nadiri AA, 2018. Supervised Committee Fuzzy logic model to assess groundwater intrinsic vulnerability in multiple aquifer systems. Arabian Journal of Geoscineces 11(8): 1-14.
Hoshangi N and Alesheikh A.A, 2015. Evaluation of ANN, ANFIS and Fuzzy system in estimation of solar radiation in Iran. Journal of Geomatics Science and Technology 4(3): 187-200 (In Persian).
Huan H, Wang J and Teng Y, 2012. Assessment and validation of groundwater vulnerability to nitrate based on a modified DRASTIC model: Acase study in Jilin City of northeast china, Science of the Total Environment 440: 14-23.  
Kazakis N, Voudouris K.S, 2015. Ground water vulnerability and pollution risk assessment of porous aquifers to nitrate modifying the DRASTIC method using quantitative parameters. Journal of Hydrology 525: 13-25.
Kadkhodaie Ilkhchi F, Asghari Moghaddam A, Barzegar R, Gharekhani M, 2019. Optimization of the DRASTIC and SINTACS models in assessing the vulnerability of the Shabestar plain aquifer. Iranian Journal of EcoHydrology 6(1): 77-88 (In Persian).
Khosravi H, 2005. Neural Network Classifier, The code project. http://www.codeproject.com /KB/cpp/MLP.aspx. 
Mahdavi A, Zare Abyaneh H, 2015. Determination of aquifer vulnerability potential based on DRASTIC and FUZZY Logic models (Case study: Hamedan- Bahar Plain). Water and Soil Science- University of Tabriz 26 (1-1): 1-17 (In Persian).
Nadiri AA, Gharekhani M, Khatibi R, 2018a. Mapping aquifer vulnerability indices using Artificial Intelligence-running Multiple Frame works (AIMF) with supervised and unsupervised learning. Water Resource Management 32(9): 3023-3040.
Nadiri AA, Sedghi Z, Khatibi R, Sadeghfam S, 2018b. Mapping specific vulnerability of multiple confined and un­confined aquifers by using artificial intelligence to learn from multiple DRASTIC frameworks. Journal of Environmental Management 227: 415-428.                                            
Neshat AR, Pradhan B, Pirasteh S and Shafri HZM, 2014. Estimating groundwater vulnerability to pollution using a modified DRASTIC model in the Kerman agricultural area, Iran. Environmental Earth Science 71 (7): 1-13.                                    
Panagopoulos G, Antonakos A and Lambrakis N, 2006. Optimization of DRASTIC model for groundwater vulnerability assessment, by the use of simple statistical methods and GIS. Hydrogeology Journal 14: 894-911.                                  
Piscopo G, 2001. Groundwater Vulnerability Map, Explanatory Notes, Castlereagh Catchment, NSW, Department of Land and Water Conservation, Australia.
Vrba J and Zoporotec A, 1994. Guidebook on Mapping Groundwater Vulnerability. International Contributions to Hydrogeology.Verlag Heinz Heise GmbH and Co. KG 
Zeidenberg M, 1990. Neural Network in Artificial Intelligence. Ellis Horwood, NewYork.