ﺑﻬﯿﻨﻪﺳﺎزي و ﻃﺮاﺣﯽ ﺳﯿﺴﺘﻢ اﺳﺘﻨﺘﺎج ﻓﺎزي ﻋﺼﺒﯽ ﺗﻄﺒﯿﻘﯽ ﺑﺎ اﺳﺘﻔﺎده از اﻟﮕﻮرﯾﺘﻢ ﺑﻬﯿﻨﻪﺳﺎزي ازدﺣﺎم ذرات و روش ﺧﻮﺷﻪﺑﻨﺪي ﻓﺎزي در ﭘﯿﺶﺑﯿﻨﯽ آﺑﺸﺴﺘﮕﯽ ﺑﻌﺪ از ﺳﺮرﯾﺰ ﺟﺎﻣﯽ

Optimization and design of Adaptive Neuro-Fuzzy Inference System using Particle Swarm Optimization and Fuzzy C-Means Clustering to predict the scour after bucket spillway

ﭼﻨﺎﻧﭽﻪ ﻣﺼﺎﻟﺢ ﭘﺎﺋﯿﻦدﺳﺖ ﺳﺮرﯾﺰ اوﺟﯽ داراي ﺑﺎﮐﺖ ﺟﺎﻣﯽ ﺷﮑﻞ ﻓﺮﺳﺎﯾﺶ ﭘﺬﯾﺮ ﺑﺎﺷﻨﺪ ﺑﺎ ﮔﺬﺷﺖ زﻣـﺎن اﺣﺘﻤـﺎل واژﮔﻮﻧﯽ ﺳﺎزه ﺳﺮرﯾﺰ اوﺟﯽ وﺟﻮد دارد. ﺑﻨﺎﺑﺮاﯾﻦ ﭘﯿﺶﺑﯿﻨﯽ ﻋﻤﻖ آﺑﺸﺴﺘﮕﯽ در ﺑﻌـﺪ از ﺳـﺮرﯾﺰ ﺟـﺎﻣﯽ از اﻫﻤﯿـﺖ ﺑﺴﺰاﯾﯽ ﺑﺮﺧﻮردار اﺳﺖ. در اﯾﻦ ﻣﻄﺎﻟﻌﻪ، ﻋﻤﻖ آﺑﺸﺴﺘﮕﯽ در ﭘﺎﺋﯿﻦدﺳﺖ ﺳﺮرﯾﺰﻫﺎي ﺟﺎﻣﯽ ﺷﮑﻞ ﺗﻮﺳﻂ ﯾﮏ ﻣـﺪل ﺗﺮﮐﯿﺒﯽ ﻓﺮا اﺑﺘﮑﺎري ﺟﺪﯾﺪ ﻣﺪلﺳﺎزي ﻣﯽﺷﻮد. اﯾﻦ ﻣﺪل ﺑـﺎ ﺗﺮﮐﯿـﺐ ﺳﯿﺴـﺘﻢ اﺳـﺘﻨﺘﺎج ﻓـﺎزي ﻋﺼـﺒﯽ ﺗﻄﺒﯿﻘـﯽ، اﻟﮕﻮرﯾﺘﻢ ﺑﻬﯿﻨﻪﺳﺎزي ازدﺣﺎم ذرات و روش ﺧﻮﺷﻪﺑﻨﺪي ﻓﺎزي ﺗﻮﺳﻌﻪ داده ﻣـﯽﺷـﻮد. ﻻزم ﺑـﻪ ذﮐـﺮ اﺳـﺖ ﮐـﻪ در ﻣﻄﺎﻟﻌﻪ ﺣﺎﺿﺮ از روش ﺧﻮﺷﻪﺑﻨﺪي ﻓﺎزي ﺑﺮاي اﯾﺠﺎد ﺳﯿﺴﺘﻢ اﺳﺘﻨﺘﺎج ﻓﺎزي اﺳـﺘﻔﺎده ﻣـﯽﺷـﻮد. ﻫﻤﭽﻨـﯿﻦ، ﺑـﺮاي ﺳﻨﺠﺶ ﺗﻮاﻧﺎﯾﯽ ﻣﺪلﻫﺎي ﻓﺮا اﺑﺘﮑﺎري از ﺷﺒﯿﻪﺳﺎزيﻫﺎي ﻣﻮﻧﺖﮐﺎرﻟﻮ اﺳﺘﻔﺎده ﻣﯽﺷـﻮد. ﻫﻤﭽﻨـﯿﻦ از روش اﻋﺘﺒـﺎر ﺳﻨﺠﯽ ﭼﻨﺪﻻﯾﻪاي ﺑﺮاي ﺑﺮرﺳﯽ ﺗﻮاﻧﺎﯾﯽ ﻣﺪلﻫﺎي ﻣﺬﮐﻮر ﺑﻬﺮه ﮔﺮﻓﺘﻪ ﻣﯽﺷﻮد. اﯾـﻦ ﻣـﺪل ﻣﻘـﺎدﯾﺮ آﺑﺸﺴـﺘﮕﯽ در ﭘﺎﺋﯿﻦدﺳﺖ ﺳﺮرﯾﺰ ﺟﺎﻣﯽ ﺷﮑﻞ را ﺑﺎ دﻗﺖ ﻣﻨﺎﺳـﺒﯽ ﺗﺨﻤـﯿﻦ زد. ﺑـﻪﻋﻨﻮانﻣﺜﺎل، ﻣﻘـﺎدﯾﺮ ﺷـﺎﺧﺺ آﻣـﺎري درﺻـﺪ ﻣﯿﺎﻧﮕﯿﻦ ﻣﻄﻠﻖ ﺧﻄﺎ و ﺧﻄﺎي ﺟﺬر ﻣﯿﺎﻧﮕﯿﻦ ﻣﺮﺑﻌﺎت ﺑﺮاي اﯾﻦ ﻣﺪل ﺑﻪﺗﺮﺗﯿﺐ ﻣﺴﺎوي 7/544 و 0/189 ﺑـﻪ دﺳـﺖ آﻣﺪ. ﻫﻤﭽﻨﯿﻦ ﻣﺪل ﺑﺮﺗﺮ ﺑﺎ ﻣﺪل ANFIS ﻧﯿﺰ ﻣﻘﺎﯾﺴﻪ ﺷﺪ ﮐﻪ ﺗﺠﺰﯾﻪوﺗﺤﻠﯿﻞ ﻧﺘﺎﯾﺞ اﯾﻦ دو ﻣﺪل ﻧﺸﺎندﻫﻨﺪه دﻗـﺖ ﺑﯿﺸﺘﺮ ﻣﺪل ﺗﺮﮐﯿﺒﯽ ﺑﻮد.

Additionally, if the materials at downstream of bucket spillway are erodible, the ogee spillway is likely to overturn by the time. Therefore, the prediction of the scour after bucket spillway is pretty important. In this study, the scour depths at downstream of bucket spillway are modeled using a new meta-heuristic model. This model is developed by combination of the Adaptive Neuro-Fuzzy Inference System (ANFIS), Particle Swarm Optimization (PSO) and Fuzzy C-Means Clustering (FCM). In addition, to assess the performance of meta-heuristic models, the Monte Carlo simulations (MCs) are used. Also, in this paper, the k-fold Cross Validation is used for examination of the models ability. Moreover, the superior model was introduced using analyzing the numerical results. The model predicted the scour depth in terms of all input parameters. The model estimated the scour at downstream of bucket spillway with reasonable accuracy. For example, the mean absolute percentage error and root mean square error for this model were obtained 7.544 and 0.189, respectively. In addition, the superior model was compared with ANFIS model that analyzing showed the compound model had more accuracy.