Malaysian Journal of Analytical Sciences Vol 17 No 3 (2013): 376 – 386

 

 

 

KAJIAN AWAL PENGGUNAAN JARINGAN NEURAL TIRUAN (ANN) BAGI MENENTUKAN NILAI PEROKSIDA TIGA MINYAK SAWIT KOMERSIL BERDASARKAN SPEKTRUM FTIR

 

(Preliminary Study on Application of Artificial Neural Networks (ANN) for Determining the Peroxide Value of Three Commercial Palm Oil Based FTIR Spectrum)

 

Azwan Mat Lazim¹*, Musa Ahmad¹, Zuriati Zakaria¹, M. Suzeren Jamil¹

Suria Ramli¹, Faiz Zainuddin¹,  Mohd Nasir Taib², M. Nasir Mat Arip³

 

¹Pusat Pengajian Sains Kimia & Teknologi Makanan,

Fakulti Sains dan Teknologi, Universiti Kebangsaan Malaysia, 43600 Bangi, Selangor

²Fakulti Kejuruteraan Elektrik,

Universiti Teknologi MARA, 40450 Shah Alam, Selangor

³Forest Products Division,

Forest Research Institute Malaysia, 52109 Kepong, Selangor, Malaysia

 

*Corresponding author: azwani79@ukm.my

 

 

Abstrak

Nilai peroksida sering digunakan bagi menentukan kualiti minyak masak di mana nilai ini mengukur sebatian peroksida dan kumpulan hidroperoksida yang terbentuk di peringkat primer pengoksidaan lipid. Dalam kajian ini sebanyak tiga minyak masak (sawit) komersil telah dipilih dan dilabelkan sebagai A, B dan C. Dua keadaan berbeza telah diaplikasikan ke atas sampel-sampel minyak tersebut. Pertama, sampel minyak telah didedahkan ke udara selama 3 bulan (minyak A) manakala yang kedua sampel minyak telah digunakan untuk menggoreng berulang kali (minyak B dan C). Bagi tujuan latihan ANN, sebanyak dua input telah dipilih daripada spektrum FTIR yang diperoleh, iaitu nilai serapan pada panjang gelombang 3444 cm^-1 dan 3450 cm^-1. Hasil kajian menunjukkan arkitektur yang sesuai bagi jaringan ANN bagi menentukan nilai peroksida bagi sampel minyak komersil adalah 2:20:1. Bagi ujian ramalan, nilai peroksida yang diramalkan oleh ANN adalah hampir menyamai nilai peroksida yang diperoleh daripada kaedah piawai. Keputusan telah menunjukkan bahawa ralat purata yang rendah telah dperoleh, iaitu 0.48 apabila neuron terlindung yang digunakan bagi arkitektur jaringan adalah 20 manakala bilangan ulangan latihan adalah 300.  

 

Kata kunci: nilai peroksida, FTIR, ANN, minyak sawit

 

Abstract

Peroxide value is one of the measurements that being used to determine the peroxide in oil samples produce from the peroxide compund and hydroperoxide group at the primary level of lipid oxidation. In this study, 3 commercial palm cooking oils were selected and labeled as A, B and C. Two different conditions were applied to the samples. First, the oil sample was exposed to the air for three months (labeled as A) while samples B and C were used for frying for many times. Two inputs from FTIR spectra (3444 cm^-1 and 3450 cm^-1) were chosen for the ANN training. The suitable architecture for this training is 2:20:1. The prediction made by ANN was very accurate and compatible to the result which obtained from the standard method. A low average error (0.48) was obtained when the hidden neuron (20) and the epochs (300) were used.

 

Keyword: peroxide value, FTIR, ANN, palm oil

 

References

1.       Cornelius, J.A. (1969). Some technical aspects influencing the quality of palm kernels. J. Sci. Agric. 17: 57-61.

2.       Choo, Y.M. (1999). Speciality products: carotenoids. Dlm. Yusof, B. Jalani, B.S. & Chan, K.W. (pnyt). Advance in oil palm research, hlm. 34-52. Kuala lumpur. PORIM

3.       Guillen, M.D & Cabo, N.(1997). Infrared spectroscopy in the study of edible oils and fats, J. Sci. Food Agric. 75: 1-11.

4.       Moh, M.H. Che Man, Y.B. Badlishah, B.S. Jinap, S. Saad. M.S. & Abdullah, W.J.W. (1999). Quantitative analysis of palm carotene using FTIR and NIR spectroscopy. JAOCS. 76 (2): 249-254.

5.       Van de Voort, F.R, Ismail A.A., Sedman, J., Dubois, J. & Nicodemo, T., (1994). The determination of peroxide value by FTIR, J. Am. Oil Chem. Soc. 71:921-926

6.       Fukuzumi, K. & Kobayashi, E. (1972). Quantitative determination of methyl octadecadienoate hydroperoxide by infrared spectroscopy. J. Am. Oil Chem. Soc. 49: 162-165.

7.       Moh, M.H & Tang, T.S. (1999). A review on the quantitative analysis of fats and oils using FTIR. PORIM Bulletin. 41:30-37.

8.       Hussain, S. Devi, K.S. Krishna,D. & Reddy, P.J. (1996). Characterization and identification of edible oil blend and prediction of the composition by ANN- A case study. Chemmometrics and Intelligent Lab. Syst. 35:117-126

9.       Nora Izma, A.A. (2003). Pembangunan kaedah analisis menggunakan FTIR dan NIR dalam membandingkan minyak tulen dan minyak campuran. Tesis Pusat pengajian Sains Kimia & Teknologi Makanan. UKM, Bangi

10.    Basiron, Y. (2001). Global oils and fats business: challenges in new millenium. Oil Palm Industry Economic Journal. 11: 1-9.

11.    Wesolowski, M. & Suchacz, B. (2001). Classification of rapeseed and soybean oils by use of unsupervised pattern recognition method and neural networks. J. Anal. Chem. 371: 323-330.

12.    Azwan Mat Lazim, Mohd Zuli Jaafar, Phang Wei Shong & Suzereen Jamil (2013). Penentuan Kualiti Minyak Masak dan Lemak Komersial Menggunakan Teknik Kimometik. The Malaysian Journal of Analytical Sciences, Vol 17 No 1 (2013): 146 – 152

13.    PORIM Test Method. (1995). Kuala Lumpur: PORIM

14.    Hendl, O., Howell, J., Lowery, J. & William Jones (2001). A rapid and simple method for the determination of iodine values using derivatives FTIR measurements, Anal.Chim. Acta. 427:75-81

15.    Kamaliah Mahmood & Norsaadah A.R. (1997). Kaedah spektroskopi dalam pengenalpastian sebatian organik. Kuala lumpur: Penerbit Universiti Malaya.

16.    Che Man, Y.B., Moh, M.H. & Van de Voort, F.R. 1999. Determination of free fatty acid in crude palm oil and refined bleached deodorized palm olein using FTIR. Journal of American Oil Chemist Society. 76(4): 485-490.

17.    Kuntom, A. (1990). Oxidation and palm oil. PORIM Bulletin. 20: 32-40.

18.    Sherwin, E.R. (1978). Oxidation and antioxidantin fat and oil processing. Journal of American Oil Chemist Society.. 55: 809-814.

19.    Fauziah, A. Razali, I. & Nor Aini, S. (2000). Frying performance of palm olein and high oleic sunflower oil during batch frying of potato crisps. Palm Oil Development. 33: 2-7.

20.    Berger, K.G. (1982). Refined palm oil quality as received. PORIM Bull. 4: 19-26.

21.    Svozil, D. Kvasnicka, V. & Prospichal, J. (1997). Introduction to multilayer feed-forward neural network. Chemmom. Intell. Lab. Sys. 39: 42-62.

22.    Taib, M.N & Narayanaswamy, R.  (1996). Multichannel calibration technique for optical fibre chemical sensor using ANN. Sens. Actuators. B. 38-39:365.

23.    Miller, J.N & Miller, J.C. (2000). Stastistics and chemometrics for analytical chemistry. Ed. Ke-4. England. Perason Education Ltd.

24.    March, J.G., Simonet, B.M. & Grases, F. (1999). Determination of phytic acid by catalytic fluorimetric. Analyst. 124:897-900.