Yıl 2007, Cilt 3, Sayı 2, Sayfalar 39 - 47 2007-08-01

Determination of The Salt Rejection Efficiency by Nanofiltration Membranes Using Neural Networks
Yapay Sinir Aglarini Kullanarak Nanofiltrasyon Membranlari Ile Tuz Giderim Veriminin Belirlenmesi

Beytullah Eren [1] , Recep Ileri [2]

113 71

In this paper, a feedforward neural network (NN) model is used to determination the salt rejection efficiency of a Nanofiltration (NF) experimental setup, which uses a DS5 nanofiltration membrane. Experimental data were collected from literature. Two hundred thirty eight experimental data were used for training and testing the network. The NN was fed with five inputs: the feed pressure, pH, salt concentration, dye concentration and cross flow velocity to determination the salt rejection efficiency. The Scaled Conjuge Gradient Algorithm (SCG) was used as optimization algorithm for training the NN. The best network configuration was set as 5-9-1 with trying and testing. The model will determine the salt rejection efficiency of nanofiltration membranes based on input and output parameters. The network was trained with a hundred seventy eight experimental data and tested with sixty experimental data. The mean absolute percentage error method was used to evaluate performance of NN. The mean absolute percentage error of training and testing results were 4.22 and 3.84 respectively. It is shown that the agreement between NN predictions and experimental data was very good.
Bu çalismada DS 5 tipi nanofiltrasyon membran kullanan membran prosesinde tuz giderim verimini belirlemek için ileri beslemeli bir YSA modeli kullanilmistir. Yapay Sinir Agini (YSA) egitmek ve test etmek için kullanilacak olan 238 adet deneysel veri literatürden toplanmistir. Bu deneysel verilerin 178 tanesi YSA’nin egit im setinde ve 60 tanesi de test setinde kullanilmak üzere rasgele seçilerek iki kisma ayrilmistir. YSA modeli bes girdi ve bir çiktidan olusmaktadir. Girdiler basinç, tuz konsantrasyonu, boya konsantrasyonu, yatay akis hizi ve pH iken çikti yalnizca tuz giderme verimidir. YSA’nin egitimi için optimizasyon algoritmasi olarak ölçeklenmis konjuge gradyent algoritmasi kullanilmistir. Test setinde YSA’nin performansini degerlendirmek için ortalama mutlaka yüzde hata (OMYH) ve ortalama karesel hata (OKH) metotlari kullanilmistir. Egitim ve test islemleri sonucunda en uygun gizli katman ve islem elemani sayisi deneysel veriler ile YSA sonuçlari arasindaki en düsük OMYH ve OKH performans degerlerine sahip olan tek gizli katmanli ve 9 islem elemanli yapi olarak belirlenmistir. Egitim ve test sonuçlari için ortalama mutlak yüzde hata degerleri sirasiyla 4,22 ve 3,84’dür. Bu degerlere göre deneysel sonuçlar ile YSA sonuçlari arasinda çok iyi bir uyum oldugu görülmektedir.
Konular
Diğer ID JA57HK87TG
Dergi Bölümü Makaleler
Yazarlar

Yazar: Beytullah Eren

Yazar: Recep Ileri

Bibtex @ { else313771, journal = {Electronic Letters on Science\&Engineering}, issn = {1305-8614}, address = {Bozok Üniversitesi}, year = {2007}, volume = {3}, pages = {39 - 47}, doi = {}, title = {Determination of The Salt Rejection Efficiency by Nanofiltration Membranes Using Neural Networks}, key = {cite}, author = {Eren, Beytullah and Ileri, Recep} }
APA Eren, B , Ileri, R . (2007). Determination of The Salt Rejection Efficiency by Nanofiltration Membranes Using Neural Networks. Electronic Letters on Science&Engineering, 3 (2), 39-47. Retrieved from http://dergipark.gov.tr/else/issue/29322/313771
MLA Eren, B , Ileri, R . "Determination of The Salt Rejection Efficiency by Nanofiltration Membranes Using Neural Networks". Electronic Letters on Science&Engineering 3 (2007): 39-47 <http://dergipark.gov.tr/else/issue/29322/313771>
Chicago Eren, B , Ileri, R . "Determination of The Salt Rejection Efficiency by Nanofiltration Membranes Using Neural Networks". Electronic Letters on Science&Engineering 3 (2007): 39-47
RIS TY - JOUR T1 - Determination of The Salt Rejection Efficiency by Nanofiltration Membranes Using Neural Networks AU - Beytullah Eren , Recep Ileri Y1 - 2007 PY - 2007 N1 - DO - T2 - Electronic Letters on Science&Engineering JF - Journal JO - JOR SP - 39 EP - 47 VL - 3 IS - 2 SN - 1305-8614- M3 - UR - Y2 - 2019 ER -
EndNote %0 Electronic Letters on Science&Engineering Determination of The Salt Rejection Efficiency by Nanofiltration Membranes Using Neural Networks %A Beytullah Eren , Recep Ileri %T Determination of The Salt Rejection Efficiency by Nanofiltration Membranes Using Neural Networks %D 2007 %J Electronic Letters on Science&Engineering %P 1305-8614- %V 3 %N 2 %R %U
ISNAD Eren, Beytullah , Ileri, Recep . "Yapay Sinir Aglarini Kullanarak Nanofiltrasyon Membranlari Ile Tuz Giderim Veriminin Belirlenmesi". Electronic Letters on Science&Engineering 3 / 2 (Ağustos 2007): 39-47.