S. cerevisiae ile Remazol Sarı (RR) giderimine yapay sinir ağı (YSA) yaklaşımı
Loading...
Date
2019-05-24
Authors
Erdem, Fatma
Journal Title
Journal ISSN
Volume Title
Publisher
Bursa Uludağ Üniversitesi
Abstract
Bu çalışma kesikli sistemde S. cerevisiae ile Remazol Sarı (RR) gideriminin incelendiği bir önceki çalışmamızın devamı niteliğinde olup deneysel sistemden elde edilen veriler ile biyosorpsiyon işlemine Yapay Sinir Ağı (YSA) yaklaşımı 2 farklı aşamada uygulanmıştır. İlk kısımda biyosorpsiyon sistemi için YSA sistem girdi verisi olarak pH, başlangıç boya konsantrasyonu ve başlangıç biyosorbent konsantrasyonu tanımlanırken çıktı katmanında % boya giderim değeri tahmin edilmiştir. YSA eğitimi Levenberg–Marquardt ileri besleme algoritması ile yapılmış olup deneysel veriler %60 eğitim, %20 validasyon ve %20 test olarak bölünmüştür. Maksimum devir (epoch) değeri 12000 iterasyon olarak belirlenmiştir. Sisteme ait R2 değerleri eğitim için %98, validasyon için %96 ve tüm biyosorpsiyon sistemi için %98 olarak belirlenmiştir. Çalışmanın ikinci aşamasında biyosorpsiyon sistem parametrelerinin biyosorpsiyon işlemine ayrı ayrı etkilerinin YSA ile modellenmesi gerçekleştirilmiştir. Deneysel veriler bu kısımda %50 eğitim, %25 validasyon ve %25 test olarak bölünmüştür. Çalışma sonucunda deneysel ve model tahmini % giderim değerleri karşılaştırıldığında, YSA ile sistemin iyi bir şekilde modellendiği ve modelin iyi bir tahmin yeteneğine sahip olduğu görülmüştür.
This study was a continuation of the our previous study in which the Remazol Yellow (RR) removal with S. cerevisiae was examined in the batch system and Artificial Neural Network (ANN) approach has been applied in 2 different stages to the biosorption process with the data obtained from this experimental system. In the first part, pH, initial dye concentration and initial biosorbent concentration are defined as system input data and the % dye removal value in the output layer is estimated with the artificial neural network. ANN training was conducted with the Levenberg-Marquardt feed forward algorithm and the experimental data were divided into 60% training, 20% validation and 20% testing. The maximum epoch value is determined as 12000 iterations. The system R2 value was set at 98% for training, 96% for validation and 98% for the entire biosorption system. In the second stage of the study, the effects of biosorption system parameters (pH, initial dye and biosorbent concentrations and temperature ) on the biosorption process were modeled by ANN. Experimental data is divided into 50% training, 25% validation and 25% test in this section. At the end of this study experimental and model estimates % dye removal values were compared and it was seen that the system is well modelled with ANN and the model has a good estimation ability.
This study was a continuation of the our previous study in which the Remazol Yellow (RR) removal with S. cerevisiae was examined in the batch system and Artificial Neural Network (ANN) approach has been applied in 2 different stages to the biosorption process with the data obtained from this experimental system. In the first part, pH, initial dye concentration and initial biosorbent concentration are defined as system input data and the % dye removal value in the output layer is estimated with the artificial neural network. ANN training was conducted with the Levenberg-Marquardt feed forward algorithm and the experimental data were divided into 60% training, 20% validation and 20% testing. The maximum epoch value is determined as 12000 iterations. The system R2 value was set at 98% for training, 96% for validation and 98% for the entire biosorption system. In the second stage of the study, the effects of biosorption system parameters (pH, initial dye and biosorbent concentrations and temperature ) on the biosorption process were modeled by ANN. Experimental data is divided into 50% training, 25% validation and 25% test in this section. At the end of this study experimental and model estimates % dye removal values were compared and it was seen that the system is well modelled with ANN and the model has a good estimation ability.
Description
Keywords
Biyosorpsiyon, Yapay sinir ağı, Modelleme, Remazol Sarı (RR), S.cerevisiae, Biosorption, Neural network, Modelling, Remazol Yellow (RR)
Citation
Erdem, F. (2019). "S. cerevisiae ile Remazol Sarı (RR) giderimine yapay sinir ağı (YSA) yaklaşımı". Uludağ Üniversitesi Mühendislik Fakültesi Dergisi, 24(2), 289-298.