(For USM Staff/Student Only)

EngLib USM > Ω School of Chemical Engineering >

Adaptive control using feedforward neural network (fann) model for lactic acid production/Chew Sze Hui

Adaptive control using feedforward neural network (fann) model for lactic acid production_Chew Sze Hui_K4_2011_NI
Bakteria asid laktik (LAB) telah lama digunakan dalam industri makanan sebagai kultur permulaan dalam penapaian daging dan produk susu. Hasil sampingan utama penapaian, asid laktik merupakan suatu bahan kimia yang serbaguna kerana mempunyai berbagai jenis penggunaan dalam industry makanan, farmaseutikal, barangan buatan kulit, tekstil dan sebagai stok suapan kimia untuk bahan kimia yang lain. Ia juga berfungsi sebagai monomer plastic terbiodegrasi. Pengeluaran bioteknologi lebih menguntungkan berbanding sintesis kimia dengan penggunaan bahan mentah yang murah seperti produk sampingan industri agro. Selain itu, cara ini juga boleh menghasilkan isomer stereo yang lebih ekonomi. Pemodelan dan rekaan strategi kawalan adaptif untuk pengeluaran asid laktik dengan penapaian kelompok Enterococcus faecalis dikaji dalam kajian ini. Model ini mengambil kira kesan had substrat dan perencat, perancatan produk, pengeluaran lactic acid berkait pertumbuhan dan tidak berkait pertumbuhan serta kadar sel mati. Di dalam proses yang sebenarnya, proses penapaian akan menghadapi gangguan dari persekitaran. Oleh itu, gangguan pada kuasa 0.5 dan 1.0 dikenakan pada model yang telah dibina. Satu rangkaian neural yang mempunyai keupayaan menggambarkan model pengeluaran asid laktik telah dibina. Sistem kawalan rujukan modal dibina dengan menggunakan rangkaian neural yang dibina sebagai rujukan. Beberapa titk set diperkenalkan kepada sistem untuk pemerhatian tindakbalas sistem itu. Hubungan tak lelurus antara kepekatan gula, kepekatan biojisim dan kepekatan asid laktik berjaya dimodalkan oleh rangkaian neural suapan hadapan yang mempunyai struktur 3 masukan, 5 neuron dan 1 keluaran. Keputusan simulasi menunjukkan jumlah kuadrat ralat ialah kurang dari 0.1. Kemudian, objektif kawalan bergantung pada imbangan antara penggurangan substrat dan produktiviti yang optimum telah dijalankan. Penilaian dari simulasi sistem kawalan menunjukkan bahawa sistem ini tidak dapat mengekalkan kepekatan produk penapaian asid laktik dan akan dihentikan secara manual selepas mencapai titik set. ___________________________________________________________________________________ Lactic acid bacteria (LAB) have long been used in the food industry as starter cultures for the manufacture of fermented meat and dairy products. The major by-product of the fermentation, lactic acid is a versatile chemical having a wide range of applications in food, pharmaceutical, leather and textile industries and as chemical feedstock for so many other chemicals. It also functions as the monomer for the biodegradable plastic. Biotechnological production is advantageous over chemical synthesis in that we can utilize cheap raw materials such as agro industrial byproducts and can selectively produce the stereo isomers in an economic way. The modeling and design of an adaptive control strategy for lactic acid production by batch fermentation of Enterococcus faecalis at is studied in this thesis. The model includes the effect of substrate limitation and inhibition, product inhibition, growth and non-growth associated lactic acid production and cell death rate. In reality, the model are ought to have noise. Therefore, noise was introduced to the system at powers of 0.5 and 1.0. A neural network having the capability of modeling the lactic acid was developed and by using the neural network as the reference model, the model reference control was constructed. Several set point was introduce to study its response. The nonlinear relationship between sugar concentration, biomass concentration and lactic acid concentration was then described by feedforward neural network with 3 inputs, 5 neurons and 1 output. The simulations show that the sum of squared error of less than 0.1. Then, the control objective relies on a tradeoff between substrate depletion and optimal productivity. The evaluation of the control system shows that the system is not feasible and the fermentation will be stopped manually upon reaching the set point.
Contributor(s):
Chew Sze Hui_ - Author
Primary Item Type:
Final Year Project
Language:
English
Subject Keywords:
Lactic acid bacteria (LAB) ; fermented meat ; dairy products
First presented to the public:
4/1/2011
Original Publication Date:
10/6/2020
Previously Published By:
Universiti Sains Malaysia
Place Of Publication:
School of Chemical Engineering
Citation:
Extents:
Number of Pages - 57
License Grantor / Date Granted:
  / ( View License )
Date Deposited
2020-10-06 13:48:04.603
Submitter:
Nor Hayati Ismail

All Versions

Thumbnail Name Version Created Date
Adaptive control using feedforward neural network (fann) model for lactic acid production/Chew Sze Hui1 2020-10-06 13:48:04.603