Main Article Content

Abstract

Abstract
Pneumatic conveying recirculate dryer (PCRD) is an artificial drying machine which is suitable for flour drying. Previous research has designed PCRD machine to dry the sago flour. The change of sago flour color in PCRD machine is very difficult to be directly measured during the drying process. The aim of this research was to develop an artificial neural network (ANN) model to predict the color difference (ΔE) between wet sago flour before drying and dried sago flour after drying by PCRD machine. The value of ΔE observation was obtained based on the sago color data calculation. The color of sago flour was measured using a color meter (TES 135A). The observation ΔE data were trained and tested on the ANN model using Graphical User Interface (GUI) application, a neuralnetwork- toolbox-based ANN on Matlab R2014a. The training and testing results of the ANN model showed that the best network structure were 12 input neurons, 5 neurons of the first hidden layer, 5 neurons of the second hidden layer, 1 neuron of the third hidden layer, and 1 output neuron (12-5-5-1-1). The value of MSE obtained by the ANN model structure was 0.0005121 with 16 times epoch. The validity test result showed that the coefficient of determination value for the training process (R2 train) equal to 0.987 and for the testing process (R2 test) equal to 0.976. Meanwhile, the optimization analysis result showed that the value of MSE and MRE were quite small, as well as the MSE and MRE value on each parameter variation. It showed that the ANN model is valid to be used to predict the color difference of sago flour drying on PCRD machine.

Abstrak
Pneumatic conveying recirculate dryer (PCRD) adalah salah satu mesin pengering buatan yang cocok digunakan untuk mengeringkan bahan tepung. Pada penelitian terdahulu telah dirancang mesin PCRD untuk mengeringkan tepung sagu. Pengukuran perubahan warna tepung sagu pada mesin PCRD sangat sulit dilakukan secara langsung selama proses pengeringan. Tujuan penelitian ini adalah mengembangkan model jaringan syaraf tiruan (JST) untuk memprediksi perbedaan warna atau color difference (ΔE) antara tepung sagu basah sebelum dikeringkan dengan tepung sagu kering setelah dikeringkan dengan mesin PCRD. Nilai ΔE observasi diperoleh berdasarkan hasil perhitungan data warna tepung sagu. Warna tepung sagu diukur menggunakan color meter (TES 135A). Data ΔE observasi tersebut dilatih dan diuji pada model JST menggunakan aplikasi Graphical User Interface (GUI) JST berbasis neural network toolbox pada Matlab R2014a. Hasil pelatihan dan pengujian model JST menunjukkan bahwa struktur jaringan yang terbaik adalah 12 neuron input, 5 neuron lapisan hidden layer 1, 5 neuron lapisan hidden layer 2, 1 neuron lapisan hidden layer 3, dan 1 neuron output (12-5-5-1-1). Nilai MSE yang dicapai struktur model JST tersebut, sebesar 0,0005121 dengan epoch 16 kali. Hasil uji validitas menunjukkan bahwa nilai koefisien determinasi untuk proses pelatihan (R2 latih) sebesar 0.987, dan proses pengujian (R2
uji) sebesar 0.976. Sedangkan hasil analisis optimasi menunjukkan bahwa, nilai MSE dan MRE yang dihasilkan cukup rendah, begitupula nilai MSE dan MRE pada setiap parameter variasi. Hal ini menunjukkan bahwa model JST tersebut valid digunakan untuk memprediksi color difference pengeringan tepung sagu pada mesin PCRD.

Keywords

artificial neural network color difference modeling pneumatic conveying recirculated dryer sago flour

Article Details

Author Biographies

Abadi Jading, Universitas Papua

Jurusan Teknologi Pertanian, Fakultas Teknologi Pertanian, Universitas Papua

Nursigit Bintoro, Universitas Gadjah Mada

Departemen Teknik Pertanian dan Biosistem, Fakultas Teknologi Pertanian, Universitas Gadjah Mada

Lilik Sutiarso, Universitas Gadjah Mada

Departemen Teknik Pertanian dan Biosistem, Fakultas Teknologi Pertanian. Universitas Gadjah Mada

Joko Nugroho Wahyu Karyadi, Universitas Gadjah Mada.

Departemen Teknik Pertanian dan Biosistem,
Fakultas Teknologi Pertanian. Universitas Gadjah Mada.

References

  1. Aichayawanich, S., M. Nopharatana, A. Nopharatana, dan W. Songkasiri. 2012. Effects of pneumatic conveying drying conditions on damaged starch content and particle size of cassava starch. Advanced Materials Research 554–556:1433–1438.
  2. BSN. 2008. Tepung Sagu: SNI 3729. Badan Standar Nasional (BSN).
  3. Correa, J.L.G., D.R. Graminho, M.A. Silva, dan S.A. Nebra. 2004. The cyclonic dryer- A numerical and experimental analysis of the influence of
  4. geometry on average particle residence time. Brazilian Journal of Chemical Engineering 21(1):103–112.
  5. Jading, A., N. Bintoro, L. Sutiarso, dan J.N.W. Karyadi. 2016. Analisis efisiensi pneumatic conveying recirculated dryer untuk pengeringan
  6. bahan-bahan tepung. Di dalam Partoyo, Y. Ratih, D. Mulyanto, dan E. Al (Eds.), Seminar Nasional Reaktualisasi Pemberdayaan Masyarakat pada
  7. Era Persaingan Bebas. Yogyakarta: Fakultas Pertanian UPN Veteran Yogyakarta.
  8. Jading, A., N. Bintoro, L. Sutiarso, dan J.N.W. Karyadi. 2017. Model jaringan syaraf tiruan untuk memprediksi kadar air bahan pada pneumatic
  9. conveying recirculated dryer. Jurnal Teknologi Industri Pertanian 27(2):141–151. https://doi. org/10.24961/j.tek.ind.pert.2017.27.2.141
  10. Limbongan, J. 2007. Morfologi beberapa jenis sagu potensial di Papua. Jurnal Litbang Pertanian 26(1):16–24.
  11. Poonnoy, P., A. Tansakui, dan M. Chinnan. 2007. Artificial neural network modeling for temperature and moisture content prediction in tomato slices
  12. undergoing microwave-vacuum drying. Journal of Food Science 72(1):E42–E47.
  13. Rosas, A., O. Baez, G.R. Urrea, dan G. Luna. 2011. Experimental and neural network prediction of a tray drier for drying getal pear. In ICEF11
  14. International Congress on Engineering and Food. p. 1–6.
  15. Satish, S., dan P. Setty. 2005. Modeling of a continuous fluidized bed dryer using artificial neural networks. Heat and Mass Transfer
  16. 32:539–547.
  17. Scala, K.D., G. Meschino, A. Vega-gálvez, R. Lemus-mondaca, S. Roura, dan R. Mascheroni. 2013. An artificial neural network model for
  18. prediction of quality characteristics of apples during convective dehydration. Food Science and Technology 33(3):411-416
  19. Shen, Y., C. Lu, P. Xu, dan Y. Yang. (2010). Color difference detection based on BP neural network. In Second International Conference
  20. on Computational Intellegence and Natural Computing (CINC).
  21. Shrivastav, S., dan B.K. Kumbar. 2009. Modeling and optimization for prediction of moisture content, drying rates and moisture ratio. International Journal of Agricultural and Biological Engineering
  22. 2(1):58–64.
  23. Widaningrum, E.P. Purwani, dan S.J. Munarso. 2005. Kajian terhadap SNI mutu pati sagu. Jurnal Standardisasi 7(3):91–98.