Department of Food Science and Technology, Tokyo University of Marine Science and Technology, Tokyo , Japan
|
|
- Kenneth Jones
- 5 years ago
- Views:
Transcription
1 Artificial neural network model flexibly applicable to retort processes under various operating conditions Yvan Llave, Tomoaki Hagiwara, and Takaharu Sakiyama Department of Food Science and Technology, Tokyo University of Marine Science and Technology, Tokyo , Japan ABSTRACT An artificial neural network (ANN) model was developed for prediction of the cold spot temperature profile during retort processing using pre-gelatinized starch dispersion (STD) as a model food. Samples of 3%, %, and 2% of STDs were prepared as follows: Corn starch powder was mixed with distilled water at 9 o C for 3 min. Each of them was filled in a retort pouch (3 23 mm), vacuum sealed, and finally processed in a retort. Sterilization was conducted under the following conditions: retort temperature (8-22 o C), holding time (2-22 min), and rotational speed (-25 rpm). Back-propagation network was chosen as the network model. The input variables for the model were retort temperature, and current and past temperatures of the cold spot (T i, T i-, and T i-2 ) at every time step. The output variable was the temperature of the cold spot at the next time step T i+. A model with 2 hidden layers, which contained and 5 neurons, respectively, was the best model among tested. It gave relative errors lower than 4% in the prediction of F value. Using the model developed, prediction of a whole profiles of the cold spot temperature were tested, starting from temperature data of the first 3 time steps with a whole course of retort temperature. The results showed very good performance of the model not only for STDs but also for several foods. Keywords: Artificial neural network; retort process; starch dispersion. I TRODUCTIO In retort processing it is essentially important to know cold spot temperature of products as accurately as possible for the evaluation of lethality attained. In recent years food engineers have developed computer models capable of simulating thermal processing based on engineering mathematics and scientific principles of heat transfer [-3]. Use of numerical solutions to mathematical heat transfer equations makes it possible to predict cold spot temperature of products even if retort temperature varies during processing. However, many parameters are required in the model: thermo-physical properties (heat capacity, density, thermal conductivity) of products and heat transfer parameters. For precise simulation, we need precise values of the parameters. Moreover, due to dynamic nature of retort process, the properties and the parameters cannot be considered uniform and constant. It is often hard to obtain the precise change in all the parameters. In some cases, such as starch-based food products processed in thermal system under intermittent agitation, brokenheating (BH) behavior is known to ocurr. BH behavior is characterized by a heat penetration curve with more than two straight portions, indicating transition of heat transfer mechanism due to gelatinization of starch. Most of the traditional models have a weak point to cope with such change in heat transfer mechanism. A more convenient technique with flexible applicability would be desired for retort process simulation. The concept of artificial neural network (ANN) has recently gained widespread popularity in many disciplines of engineering and science. Attractive features of ANN are their ability to learn from data sets to find complicated relationships in them. In this study, an ANN model was developed as an alternative tool for prediction of the cold spot temperature profile of retort products. Starch dispersion (STD) was used as a model food because they could show transition of heat transfer mechanism, so-called BH behavior. ANN model was trained with cold spot temperature profiles, including those of BH behaviors, experimentally obtained. MATERIALS & METHODS Samples of 3%, %, and 2% of STDs were prepared as follows. Corn starch powder (SIGMA-ALDRICH, USA) was mixed with distilled water thoroughly and brought to heating in a jacketed kettle mixer (Shinagawa Twin Mixer STM-5) at 9 o C for 3 min for pre-gelatinization. During the pre-gelatinization, water was added to compensate for evaporation losses. After being cooled to room temperature, the pre-
2 gelatinized STD was filled in a retort pouch (3 x 23 mm) PET2/AL9/NY5/CP6, and vacuum sealed (Yoshikawa FVCII). For validation of accuracy of ANN model, several foods were prepared and subjected to retort processing: corn soup, rice, whole corn, curry paste, and fish fillet according to traditional way of preparation for each retort foods. After pre-treatment, each sample was filled in a retort pouch of variable size according to its size. A set of 4 thermocouples (Shibaura Thermistor type 3 THR) were positioned at the geometric center of each pouch placed at the cold spot of the retort to obtain heat penetration data. Sterilization was carried out using stationary or rotary mode in a pilot scale retort (Hisaka RCS-4RTGN). After a pre-heating step (9 o C x 3 min), sterilization was conducted under the following conditions: retort temperature (8, 2, and 22 o C), holding time (2, 5, 2, and 22 min), and rotational speed (, 5,, 5, 2, and 25 rpm). Heat penetration data were recorded every 2s intervals via a data logger. A selected number of heat penetration data were processed using an ANN software (NeuralWorks Professional II/Plus). Back-propagation (BP) network was chosen as the network model. The input variables for the model were retort temperature, and current and past temperatures of the cold spot (T i, T i-, and T i-2 ) at every time step. The output variable was the temperature of the cold spot at the next time step T i+. In order to obtain the best topology of the network, parameter functions, the number of hidden layers, and the number of neurons contained in each layer were determined on a trial-and-error basis to give the smallest difference between the predicted temperature and the experimental data. In addition, the predicted time-temperature profile was compared with the experimental data in terms of accumulated lethality (F value) calculated by Eq. (). F = t ( T T) / z dt () where T = 2. o C and z = o C. Finally, the ANN model was tested for prediction of the temperature profiles of not only STDs under other retort conditions, but also of real foods. Pouches containing those foods were sterilized at 2 o C for different times to assure a final F value of 6. at least. RESULTS & DISCUSSIO Heat penetration curves for STDs Major factors affecting retort sterilization process of STDs considered in this study are the starch concentration and the rotational speed of retort. Figure shows typical heat penetration curves for 3% and 2% STDs. For 2% STD, the curve gave a single straight line after several minutes of initial lag period. On the other hand, the curve for 3% STD was approximated by two straight lines, indicating two stages of heat penetration recognized as BH. The significant change in heating rate suggested increase in heat transfer resistance at the break point (intersection of two lines), which is probably attribute to sol-gel transition caused by starch gelatinization, as known for many starch-based foods [3-5]. 3% L T RT-T ( ) 2% L 2% S TD, 5rpm 3% L2 3% STD, 25rpm T im e (m in ) Figure. Typical heating curves for 3% and 2% STDs under retort treatment
3 Construction of A model Sets of 58,997 and 5,268 data obtained through thermal processing of STDs were used for the network training and testing respectively. The steps taken for construction of ANN model are summarized as follows. The optimal configuration was selected from 24 ANN configurations (from to two hidden layers, up to fifteen neurons in each hidden layer), using training data sets and taking mean relative error as a measure of predictive performance. The coefficient of determination, R 2, between the predicted values by ANN model and the observed ones (desired output) was also used as another measure. As the result, a model BP with 2 hidden layers that contained and 5 neurons, respectively, was found the best model among tested. The final BP model, obtained by using Extended Delta-Bar-Delta learning rule and hyperbolic tangent transfer function, gave a good prediction at each time step. For example, in the case shown in Figure 2, the relative errors were within.5% for the cold spot temperature of 3% STD. This resulted in a relative error of -3.5% for F value in this case (see Table ). Temperature ( o C) Experimental temp Predicted temp Retort temp. Fo value observed Fo value predicted Time (min) F (min) Figure 2. Experimental and Predicted time-temperature profile and calculated F value of 3% STD, rotated at 25 rpm and heated at 2 o C for 5 min. Figure 3 compares the predicted and observed values of the cold spot temperature for 3% STD (during the retort processing shown in Fig. 2). It can be observed that there is a very high correlation between them (R 2 >.999), indicating a nice adjustment of ANN model to the process behavior. 4 T predicted ( ) R 2 =.9998 BP T observed ( ) Figure 3. Predicted vs. observed temperature at the cold spot using BP for 3% STD. The ANN model developed predicts the next step temperature from the last three temperatures already known. The prediction performance mentioned above was based on such step-by-step mode of prediction. However, we can put an output temperature into the input at the next step. Thus prediction of whole cold spot
4 temperature profile was tested, starting from temperatures at the first 3 time steps together with a whole course of retort temperature. This prediction mode is referred to as continuous prediction, while the original one is referred to as simple step prediction, hereafter. Table. Comparison of error estimation for validation processes of STDs. Simple step prediction Continuous prediction Validation Processes Time-temperature profile F Time-temperature profile F R 2 MRE RE (%) R 2 MRE RE (%) 3%STD/25 rpm/2 o Cx5min ± ± %STD/ rpm/2 o Cx2min ± ± %STD/5 rpm/2 o Cx2min ± ± %STD/5 rpm/2 o Cx22min ± ± Table shows the comparison of prediction errors of the both modes for validation data sets obtained under several retort conditions. The mean relative errors (MREs) of continuous prediction were found slightly higher than those of simple step prediction for the time-temperature profile. However the relative errors (REs) for F value (calculated with each time-temperature profile) were found opposite. This can be because continuous prediction gave large REs of temperature at beginning, where temperature was too low to affect F value. At higher temperatures, continuous prediction gave rather good prediction, which lowered RE of F value. A similar ANN model architecture, taking known temperatures at three time steps as input, was employed by Gonçalves et al. [6]. They reported that relative prediction errors of F value were lower than 2.6 % for food cans showing purely conductive heat transfer behavior. Although the experimental situations and ANN model construction were not the same, we thus obtained similar, yet better, accuracy of prediction. Moreover, the ANN model developed in this study expanded the scope of the prediction to foods with BH behaviors. Application to various types of food Several foods were selected to evaluate the prediction by BP model obtained above. Results of continuous prediction of time-temperature profile for corn soup are presented in Fig. 4. As shown in Table 2 the good accuracy of the prediction was confirmed also for other foods. Temperature ( o C) Experimental temp Predicted temp Retort temp. Fo value observed Fo value predicted Time (min) F (min) Figure 4. Experimental and Predicted time-temperature profile and calculated F value of corn soup sterilized at 2 o C at rpm.
5 Table 2. Performance of prediction of temperature profiles and Fo value for real foods. Targed food Time-temperature profile F R 2 MRE (%) RE (%) Corn soup ±.5.3% Rice ±.8.8% Fish (T. murphyi) ±.8 3.6% Whole corn ±.6 2.3% Curry paste ±.9 2.9% CO CLUSIO ANN model developed by training with heat penetration data for STDs was found successful to predict time-temperature profile and thermal lethality with high accuracy under variety of processing conditions, irrespective of different heat transfer modes. Its excellent performance for retort foods other than STDs indicated that the obtained model will be useful for prediction and control of retort systems. REFERE CES [] Tucker G.S. 99. Development and use of numerical techniques for improved thermal process calculations and control. Food Control. January, 5-9. [2] Tucker G.S. & Holdsworth S.D. 99. Mathematical modeling of sterilization and cooking processes for heat preserved foods, applications of a new heat transfer model. Trans. IchemE, 69, 5-2. [3] Noronha J., Hendrickx A., Van Loey A. & Tobback P New semi-empirical approach to handle time-variable boundary conditions during sterilization of non-conductive heating foods. Journal of Food Engineering, 24, [4] Yang W.H. & Rao M.A Numerical study of parameters affecting broken heating curve. Journal of Food Engineering, 37, [5] Tattiyakul J., Rao M.A. & Data A.K. 22. Heat transfer to a canned corn starch dispersion under intermittent agitation. Journal of Food Engineering, 54(4), [6] Gonçalves E.C., Minim L.A., Coimbra J.S.R. & Minim V.P.R. 25. Modeling sterilization process of canned foods using artificial neural networks. Chemical Engineering and Processing, 44(2),
MODELING OF A HOT AIR DRYING PROCESS BY USING ARTIFICIAL NEURAL NETWORK METHOD
MODELING OF A HOT AIR DRYING PROCESS BY USING ARTIFICIAL NEURAL NETWORK METHOD Ahmet DURAK +, Ugur AKYOL ++ + NAMIK KEMAL UNIVERSITY, Hayrabolu, Tekirdag, Turkey. + NAMIK KEMAL UNIVERSITY, Çorlu, Tekirdag,
More informationOptimization of pulsed microwave heating
Journal of Food Engineering 78 (7) 1457 1462 www.elsevier.com/locate/jfoodeng Optimization of pulsed microwave heating Sundaram Gunasekaran *, Huai-Wen Yang Food & Bioprocess Engineering Laboratory, Department
More informationARTIFICIAL NEURAL NETWORK PART I HANIEH BORHANAZAD
ARTIFICIAL NEURAL NETWORK PART I HANIEH BORHANAZAD WHAT IS A NEURAL NETWORK? The simplest definition of a neural network, more properly referred to as an 'artificial' neural network (ANN), is provided
More informationThe Study of Heat Penetration of Kimchi Soup on Stationary and Rotary Retorts
Prev. Nutr. Food Sci. 2015;20(1):60-66 http://dx.doi.org/10.3746/pnf.2015.20.1.60 pissn 2287-1098 ㆍ eissn 2287-8602 The Study of Heat Penetration of Kimchi Soup on Stationary and Rotary Retorts Won-Il
More informationRA P.1 (1-12) APT:m v 1.73 Prn:28/01/2008; 15:51 apt2439 by:laima p. 1. J. Manickaraj and N. Balasubramanian
RA P.1 (1-12) APT:m v 1.73 Prn:28/01/2008; 15:51 apt2439 by:laima p. 1 Advanced Powder Technology 0 (2008) 1 12 www.brill.nl/apt Original paper Estimation of the Heat Transfer Coefficient in a Liquid Solid
More informationEffect of Baking Powder in Wheat Flour Dough on Its Thermal Conduction during
Food Sci. Technol. Res., 5 (), 7, 009 Effect of Baking Powder in Wheat Flour Dough on Its Thermal Conduction during Heating Tamako mizu and Keiko nagao * Tokyo Kasei University, Faculty of Home Economics,
More informationNEURAL NETWORK TECHNIQUES FOR BURST PRESSURE PREDICTION IN KEVLAR/EPOXY PRESSURE VESSELS USING ACOUSTIC EMISSION DATA
Abstract NEURAL NETWORK TECHNIQUES FOR BURST PRESSURE PREDICTION IN KEVLAR/EPOXY PRESSURE VESSELS USING ACOUSTIC EMISSION DATA ERIC v. K. HILL, MICHAEL D. SCHEPPA, ZACHARY D. SAGER and ISADORA P. THISTED
More informationThermal Death Time Module- 16 Lec- 16 Dr. Shishir Sinha Dept. of Chemical Engineering IIT Roorkee
Thermal Death Time Module- 16 Lec- 16 Dr. Shishir Sinha Dept. of Chemical Engineering IIT Roorkee Thermal death time Thermal death time is a concept used to determine how long it takes to kill a specific
More informationPortugaliae Electrochimica Acta 26/4 (2008)
Portugaliae Electrochimica Acta 6/4 (008) 6-68 PORTUGALIAE ELECTROCHIMICA ACTA Comparison of Regression Model and Artificial Neural Network Model for the Prediction of Volume Percent of Diamond Deposition
More informationModeling and Compensation for Capacitive Pressure Sensor by RBF Neural Networks
21 8th IEEE International Conference on Control and Automation Xiamen, China, June 9-11, 21 ThCP1.8 Modeling and Compensation for Capacitive Pressure Sensor by RBF Neural Networks Mahnaz Hashemi, Jafar
More informationMilena Stanga Technical Marketing Engineer SOLVAY GREEN PVDF FOR GREEN BATTERIES
Milena Stanga Technical Marketing Engineer SOLVAY GREEN PVDF FOR GREEN BATTERIES STATE OF THE ART PVDF powder in NMP solvent PVDF is a partially fluorinated semi-crystalline polymer with excellent thermo-mechanical
More informationArtificial Neural Network Method of Rock Mass Blastability Classification
Artificial Neural Network Method of Rock Mass Blastability Classification Jiang Han, Xu Weiya, Xie Shouyi Research Institute of Geotechnical Engineering, Hohai University, Nanjing, Jiangshu, P.R.China
More informationNeural Network Based Density Measurement
Bulg. J. Phys. 31 (2004) 163 169 P. Neelamegam 1, A. Rajendran 2 1 PG and Research Department of Physics, AVVM Sri Pushpam College (Autonomous), Poondi, Thanjavur, Tamil Nadu-613 503, India 2 PG and Research
More informationMaterials Science Forum Online: ISSN: , Vols , pp doi: /
Materials Science Forum Online: 2004-12-15 ISSN: 1662-9752, Vols. 471-472, pp 687-691 doi:10.4028/www.scientific.net/msf.471-472.687 Materials Science Forum Vols. *** (2004) pp.687-691 2004 Trans Tech
More informationEstimation of the Pre-Consolidation Pressure in Soils Using ANN method
Current World Environment Vol. 11(Special Issue 1), 83-88 (2016) Estimation of the Pre-Consolidation Pressure in Soils Using ANN method M. R. Motahari Department of Civil Engineering, Faculty of Engineering,
More informationLecture 7 Artificial neural networks: Supervised learning
Lecture 7 Artificial neural networks: Supervised learning Introduction, or how the brain works The neuron as a simple computing element The perceptron Multilayer neural networks Accelerated learning in
More informationEXPERIMENT #1 SEPARATION AND RECOVERY OF ORGANIC COMPOUNDS, THIN LAYER CHROMATOGRAPHY, COLUMN CHROMATOGRAPHY, CRYSTALLIZATION AND MELTING POINTS
EXPERIMENT #1 SEPARATION AND RECOVERY OF ORGANIC COMPOUNDS, THIN LAYER CHROMATOGRAPHY, COLUMN CHROMATOGRAPHY, CRYSTALLIZATION AND MELTING POINTS Overview In the first few weeks of this semester you will
More informationApplication of Artificial Neural Networks in Evaluation and Identification of Electrical Loss in Transformers According to the Energy Consumption
Application of Artificial Neural Networks in Evaluation and Identification of Electrical Loss in Transformers According to the Energy Consumption ANDRÉ NUNES DE SOUZA, JOSÉ ALFREDO C. ULSON, IVAN NUNES
More informationINVERSE RELIABILITY ANALYSIS IN STRUCTURAL DESIGN
INVERSE RELIABILITY ANALYSIS IN STRUCTURAL DESIGN David Lehký, Drahomír Novák Institute of Structural Mechanics, Faculty of Civil Engineering, Brno University of Technology, Brno, Czech Republic.6.0 ISUME
More informationSearch for Inspiral GW Signals Associated with Short GRBs using Artificial Neural Networks
Search for Inspiral GW Signals Associated with Short GRBs using Artificial Neural Networks Kyungmin Kim and Hyun Kyu Lee Hanyang Univ. In collaboration with Y.-M. Kim, C.-H. Lee (PNU), J.J. Oh, S.H. Oh,
More informationA FUZZY NEURAL NETWORK MODEL FOR FORECASTING STOCK PRICE
A FUZZY NEURAL NETWORK MODEL FOR FORECASTING STOCK PRICE Li Sheng Institute of intelligent information engineering Zheiang University Hangzhou, 3007, P. R. China ABSTRACT In this paper, a neural network-driven
More informationSolubility Modeling of Diamines in Supercritical Carbon Dioxide Using Artificial Neural Network
Australian Journal of Basic and Applied Sciences, 5(8): 166-170, 2011 ISSN 1991-8178 Solubility Modeling of Diamines in Supercritical Carbon Dioxide Using Artificial Neural Network 1 Mehri Esfahanian,
More informationShort Term Solar Radiation Forecast from Meteorological Data using Artificial Neural Network for Yola, Nigeria
American Journal of Engineering Research (AJER) 017 American Journal of Engineering Research (AJER) eiss: 300847 piss : 300936 Volume6, Issue8, pp8389 www.ajer.org Research Paper Open Access Short Term
More informationAPPLICATIONS OF ARTIFICIAL NEURAL NETWORKS IN SUPERCRITICAL FLUID EXTRACTION MODELLING AND SIMULATION
1 APPLICATIONS OF ARTIFICIAL NEURAL NETWORKS IN SUPERCRITICAL FLUID EXTRACTION MODELLING AND SIMULATION Jianzhong Yin* 1, Xinwei Ding 1, Chung Sung Tan 2 1 School of Chemical Engineering, Dalian University
More informationDeep Feedforward Networks
Deep Feedforward Networks Liu Yang March 30, 2017 Liu Yang Short title March 30, 2017 1 / 24 Overview 1 Background A general introduction Example 2 Gradient based learning Cost functions Output Units 3
More informationAP Chemistry: Designing an Effective Hand Warmer Student Guide INTRODUCTION
AP Chemistry: Designing an Effective Hand Warmer Student Guide INTRODUCTION AP and the Advanced Placement Program are registered trademarks of the College Entrance Examination Board. The activity and materials
More informationData and prognosis for renewable energy
The Hong Kong Polytechnic University Department of Electrical Engineering Project code: FYP_27 Data and prognosis for renewable energy by Choi Man Hin 14072258D Final Report Bachelor of Engineering (Honours)
More informationMODELING COMBINED VLE OF FOUR QUATERNARY MIXTURES USING ARTIFICIAL NEURAL NETWORK
MODELING COMBINED VLE OF FOUR QUATERNARY MIXTURES USING ARTIFICIAL NEURAL NETWORK SHEKHAR PANDHARIPANDE* Associate Professor, Department of Chemical Engineering, LIT, RTMNU, Nagpur, India, slpandharipande@gmail.com
More informationArtificial Neural Networks (ANN) Xiaogang Su, Ph.D. Department of Mathematical Science University of Texas at El Paso
Artificial Neural Networks (ANN) Xiaogang Su, Ph.D. Department of Mathematical Science University of Texas at El Paso xsu@utep.edu Fall, 2018 Outline Introduction A Brief History ANN Architecture Terminology
More informationEffect of number of hidden neurons on learning in large-scale layered neural networks
ICROS-SICE International Joint Conference 009 August 18-1, 009, Fukuoka International Congress Center, Japan Effect of on learning in large-scale layered neural networks Katsunari Shibata (Oita Univ.;
More informationAddress for Correspondence
Research Article APPLICATION OF ARTIFICIAL NEURAL NETWORK FOR INTERFERENCE STUDIES OF LOW-RISE BUILDINGS 1 Narayan K*, 2 Gairola A Address for Correspondence 1 Associate Professor, Department of Civil
More informationECE521 Lectures 9 Fully Connected Neural Networks
ECE521 Lectures 9 Fully Connected Neural Networks Outline Multi-class classification Learning multi-layer neural networks 2 Measuring distance in probability space We learnt that the squared L2 distance
More informationANOTHER FIVE QUESTIONS:
No peaking!!!!! See if you can do the following: f 5 tan 6 sin 7 cos 8 sin 9 cos 5 e e ln ln @ @ Epress sin Power Series Epansion: d as a Power Series: Estimate sin Estimate MACLAURIN SERIES ANOTHER FIVE
More informationECE Introduction to Artificial Neural Network and Fuzzy Systems
ECE 39 - Introduction to Artificial Neural Network and Fuzzy Systems Wavelet Neural Network control of two Continuous Stirred Tank Reactors in Series using MATLAB Tariq Ahamed Abstract. With the rapid
More informationCool Off, Will Ya! Investigating Effect of Temperature Differences between Water and Environment on Cooling Rate of Water
Ding 1 Cool Off, Will Ya! Investigating Effect of Temperature Differences between Water and Environment on Cooling Rate of Water Chunyang Ding 000844-0029 Physics HL Ms. Dossett 10 February 2014 Ding 2
More informationCombination of M-Estimators and Neural Network Model to Analyze Inside/Outside Bark Tree Diameters
Combination of M-Estimators and Neural Network Model to Analyze Inside/Outside Bark Tree Diameters Kyriaki Kitikidou, Elias Milios, Lazaros Iliadis, and Minas Kaymakis Democritus University of Thrace,
More informationECE662: Pattern Recognition and Decision Making Processes: HW TWO
ECE662: Pattern Recognition and Decision Making Processes: HW TWO Purdue University Department of Electrical and Computer Engineering West Lafayette, INDIANA, USA Abstract. In this report experiments are
More informationESTIMATING THE ACTIVATION FUNCTIONS OF AN MLP-NETWORK
ESTIMATING THE ACTIVATION FUNCTIONS OF AN MLP-NETWORK P.V. Vehviläinen, H.A.T. Ihalainen Laboratory of Measurement and Information Technology Automation Department Tampere University of Technology, FIN-,
More informationArtificial Neural Network Based Approach for Design of RCC Columns
Artificial Neural Network Based Approach for Design of RCC Columns Dr T illai, ember I Karthekeyan, Non-member Recent developments in artificial neural network have opened up new possibilities in the field
More informationBackground & Purpose Artificial Neural Network Spatial Interpolation of soil properties Numerical analysis of Kobe Airport Conclusions.
IACMAG 214 September 22, 214 2 Applicability of artificial neural network to estimating soil of Holocene clays in Osaka Bay Kazuhiro ODA Graduate school of Engineering, Osaka University 3 BACKGROUND 4
More informationSolid Food Pasteurization by Ohmic Heating: Influence of Process Parameters
Excerpt from the Proceedings of the COMSOL Conference 2008 Boston Solid Food Pasteurization by Ohmic Heating: Influence of Process Parameters Markus Zell 1, Denis A. Cronin 1, Desmond J. Morgan 1, Francesco
More informationA Robot that Learns an Evaluation Function for Acquiring of Appropriate Motions
A Robot that Learns an Evaluation Function for Acquiring of Appropriate Motions Katsunari Shibata and Yoich Okabe Research Center for Advanced Science and Technology, Univ. of Tokyo -6-1 Komaba, Meguro-ku,
More informationEEE 241: Linear Systems
EEE 4: Linear Systems Summary # 3: Introduction to artificial neural networks DISTRIBUTED REPRESENTATION An ANN consists of simple processing units communicating with each other. The basic elements of
More informationDesign Collocation Neural Network to Solve Singular Perturbed Problems with Initial Conditions
Article International Journal of Modern Engineering Sciences, 204, 3(): 29-38 International Journal of Modern Engineering Sciences Journal homepage:www.modernscientificpress.com/journals/ijmes.aspx ISSN:
More informationWhat Do Neural Networks Do? MLP Lecture 3 Multi-layer networks 1
What Do Neural Networks Do? MLP Lecture 3 Multi-layer networks 1 Multi-layer networks Steve Renals Machine Learning Practical MLP Lecture 3 7 October 2015 MLP Lecture 3 Multi-layer networks 2 What Do Single
More informationCivil and Environmental Research ISSN (Paper) ISSN (Online) Vol.8, No.1, 2016
Developing Artificial Neural Network and Multiple Linear Regression Models to Predict the Ultimate Load Carrying Capacity of Reactive Powder Concrete Columns Prof. Dr. Mohammed Mansour Kadhum Eng.Ahmed
More informationClocking the Effect of Molarity on Speed of Reaction. reaction. While most people do assume that the temperature of the solution is often the most
Ding 1 Chunyang Ding Mr. Rierson AP/IB Chemistry SL 28 January 2013 Clocking the Effect of Molarity on Speed of Reaction In basic levels of chemistry, most of the experimenter s attention is on the reaction
More informationForecasting Crude Oil Price Using Neural Networks
CMU. Journal (2006) Vol. 5(3) 377 Forecasting Crude Oil Price Using Neural Networks Komsan Suriya * Faculty of Economics, Chiang Mai University, Chiang Mai 50200, Thailand *Corresponding author. E-mail:
More informationInternational Journal of Scientific Research and Reviews
Research article Available online www.ijsrr.org ISSN: 2279 0543 International Journal of Scientific Research and Reviews Prediction of Compressive Strength of Concrete using Artificial Neural Network ABSTRACT
More informationModeling of the Bread Baking Process Using Moving Boundary and Arbitrary-Lagrangian-Eulerian (ALE) C. Anandharamakrishnan, N. Chhanwal, P.
Presented at the COMSOL Conference 2010 India Modeling of the Bread Baking Process Using Moving Boundary and Arbitrary-Lagrangian-Eulerian (ALE) Approaches C. Anandharamakrishnan, N. Chhanwal, P. Karthik,
More informationSimple neuron model Components of simple neuron
Outline 1. Simple neuron model 2. Components of artificial neural networks 3. Common activation functions 4. MATLAB representation of neural network. Single neuron model Simple neuron model Components
More informationArtificial Neural Network : Training
Artificial Neural Networ : Training Debasis Samanta IIT Kharagpur debasis.samanta.iitgp@gmail.com 06.04.2018 Debasis Samanta (IIT Kharagpur) Soft Computing Applications 06.04.2018 1 / 49 Learning of neural
More informationNeural Network Model for Predicting the Electrical Properties of Nano-Structure Materials
Neural Network Model for Predicting the Electrical Properties of Nano-Structure Materials S. Kumaravel Department of Computer Science, AVVM Sri Pushpam College (Autonomous), Poondi, Thanjavur S. Sriram
More informationCE213 Artificial Intelligence Lecture 13
CE213 Artificial Intelligence Lecture 13 Neural Networks What is a Neural Network? Why Neural Networks? (New Models and Algorithms for Problem Solving) McCulloch-Pitts Neural Nets Learning Using The Delta
More informationFORECASTING OF ECONOMIC QUANTITIES USING FUZZY AUTOREGRESSIVE MODEL AND FUZZY NEURAL NETWORK
FORECASTING OF ECONOMIC QUANTITIES USING FUZZY AUTOREGRESSIVE MODEL AND FUZZY NEURAL NETWORK Dusan Marcek Silesian University, Institute of Computer Science Opava Research Institute of the IT4Innovations
More informationInvestigation of complex modulus of asphalt mastic by artificial neural networks
Indian Journal of Engineering & Materials Sciences Vol. 1, August 014, pp. 445-450 Investigation of complex modulus of asphalt mastic by artificial neural networks Kezhen Yan* & Lingyun You College of
More informationForecasting of Rain Fall in Mirzapur District, Uttar Pradesh, India Using Feed-Forward Artificial Neural Network
International Journal of Engineering Science Invention ISSN (Online): 2319 6734, ISSN (Print): 2319 6726 Volume 2 Issue 8ǁ August. 2013 ǁ PP.87-93 Forecasting of Rain Fall in Mirzapur District, Uttar Pradesh,
More informationMultilayer Neural Networks
Pattern Recognition Lecture 4 Multilayer Neural Netors Prof. Daniel Yeung School of Computer Science and Engineering South China University of Technology Lec4: Multilayer Neural Netors Outline Introduction
More informationCPT-BASED SIMPLIFIED LIQUEFACTION ASSESSMENT BY USING FUZZY-NEURAL NETWORK
326 Journal of Marine Science and Technology, Vol. 17, No. 4, pp. 326-331 (2009) CPT-BASED SIMPLIFIED LIQUEFACTION ASSESSMENT BY USING FUZZY-NEURAL NETWORK Shuh-Gi Chern* and Ching-Yinn Lee* Key words:
More informationNeural Network to Control Output of Hidden Node According to Input Patterns
American Journal of Intelligent Systems 24, 4(5): 96-23 DOI:.5923/j.ajis.2445.2 Neural Network to Control Output of Hidden Node According to Input Patterns Takafumi Sasakawa, Jun Sawamoto 2,*, Hidekazu
More informationNeural network modelling of reinforced concrete beam shear capacity
icccbe 2010 Nottingham University Press Proceedings of the International Conference on Computing in Civil and Building Engineering W Tizani (Editor) Neural network modelling of reinforced concrete beam
More informationPATTERN RECOGNITION FOR PARTIAL DISCHARGE DIAGNOSIS OF POWER TRANSFORMER
PATTERN RECOGNITION FOR PARTIAL DISCHARGE DIAGNOSIS OF POWER TRANSFORMER PO-HUNG CHEN 1, HUNG-CHENG CHEN 2, AN LIU 3, LI-MING CHEN 1 1 Department of Electrical Engineering, St. John s University, Taipei,
More informationAdsorption behavior of methylene blue onto gellan gum-bentonite composite beads for bioremediation application
World Journal of Pharmaceutical Sciences ISSN (Print): 2321-3310; ISSN (Online): 2321-3086 Published by Atom and Cell Publishers All Rights Reserved Available online at: http://www.wjpsonline.org/ Original
More informationApplication of Artificial Neural Network Model in Calculation of Pressure Drop Values of Nanofluid
International Journal of Engineering and Technology Volume 3 No. 5, May, 2013 Application of Artificial Neural Network Model in Calculation of Pressure Drop Values of Nanofluid Mahmoud S. Youssef 1,2,
More informationUnit 11: Temperature and heat
Unit 11: Temperature and heat 1. Thermal energy 2. Temperature 3. Heat and thermal equlibrium 4. Effects of heat 5. Transference of heat 6. Conductors and insulators Think and answer a. Is it the same
More informationPrediction of Channel Diameter to Reduce Flow Mal Distribution in Radiators using ANN
Indian Journal of Science and Technology, Vol 8(S9), 341-346, May 2015 ISSN (Print) : 0974-6846 ISSN (Online) : 0974-5645 DOI: 10.17485/ijst/2015/v8iS9/65585 Prediction of Channel Diameter to Reduce Flow
More informationV a l i d a t i o n S u m m a r y R e p o r t D o m e t i c S e r i e s M T. Translation of. page 1
Translation of Validation Summary Report DOMETIC MT4B Original report in German established by page 1 Table of Contents Description of the Validation Method...3 Test Readings MT4B at +10ºC Outside Temperature
More informationUnderstanding uncertainties associated with the 5128A RHapid-Cal Humidity Generator
Understanding uncertainties associated with the 5128A RHapid-Cal Humidity Generator Technical Note The Fluke Calibration 5128A RHapid-Cal Humidity Generator provides a portable, stable test environment
More informationPattern Recognition Prof. P. S. Sastry Department of Electronics and Communication Engineering Indian Institute of Science, Bangalore
Pattern Recognition Prof. P. S. Sastry Department of Electronics and Communication Engineering Indian Institute of Science, Bangalore Lecture - 27 Multilayer Feedforward Neural networks with Sigmoidal
More informationStudy on Thermal Conductivities Prediction for Apple Fruit Juice by Using Neural Network
Study on Thermal Conductivities Prediction for Apple Fruit Juice by Using eural etwork Min Zhang *, Zhenhua Che, Jiahua Lu, Huizhong Zhao, Jianhua Chen, Zhiyou Zhong and Le Yang, College of Food Sciences,
More informationProceedings of 12th International Heat Pipe Conference, pp , Moscow, Russia, 2002.
7KHUPDO3HUIRUPDQFH0RGHOLQJRI3XOVDWLQJ+HDW3LSHVE\$UWLILFLDO1HXUDO1HWZRUN Sameer Khandekar (a), Xiaoyu Cui (b), Manfred Groll (a) (a) IKE, University of Stuttgart, Pfaffenwaldring 31, 70569, Stuttgart, Germany.
More informationMath 315: Differential Equations Lecture Notes Patrick Torres
Introduction What is a Differential Equation? A differential equation (DE) is an equation that relates a function (usually unknown) to its own derivatives. Example 1: The equation + y3 unknown function,
More informationNeural Networks for Protein Structure Prediction Brown, JMB CS 466 Saurabh Sinha
Neural Networks for Protein Structure Prediction Brown, JMB 1999 CS 466 Saurabh Sinha Outline Goal is to predict secondary structure of a protein from its sequence Artificial Neural Network used for this
More informationVacuum techniques (down to 1 K)
Vacuum techniques (down to 1 K) For isolation (deep Knudsen regime) liquid helium dewar / inner vacuum jacket Leak testing at level 10-11 Pa m3/s (10-10 mbar l/s) liquid helium dewar & transfer syphon
More informationDevelopment of an computational algorithm for canned food quality control used with batch sterilization processes
Development of an computational algorithm for canned food quality control used with batch sterilization processes Rubens Gedraite (UNMEP) rgedrait@unimep.br Newton Libanio Ferreira (UNMEP) nelferrei@unimep.br
More informationANN TECHNIQUE FOR ELECTRONIC NOSE BASED ON SMART SENSORS ARRAY
U.P.B. Sci. Bull., Series C, Vol. 79, Iss. 4, 2017 ISSN 2286-3540 ANN TECHNIQUE FOR ELECTRONIC NOSE BASED ON SMART SENSORS ARRAY Samia KHALDI 1, Zohir DIBI 2 Electronic Nose is widely used in environmental
More informationArtificial Neural Network
Artificial Neural Network Contents 2 What is ANN? Biological Neuron Structure of Neuron Types of Neuron Models of Neuron Analogy with human NN Perceptron OCR Multilayer Neural Network Back propagation
More informationIn Situ Adaptive Tabulation for Real-Time Control
In Situ Adaptive Tabulation for Real-Time Control J. D. Hedengren T. F. Edgar The University of Teas at Austin 2004 American Control Conference Boston, MA Outline Model reduction and computational reduction
More informationEXPERIMENT ET: ENERGY TRANSFORMATION & SPECIFIC HEAT
MASSACHUSETTS INSTITUTE OF TECHNOLOGY Physics Department Physics 8.01X Fall 2000 EXPERIMENT ET: ENERGY TRANSFORMATION & SPECIFIC HEAT We have introduced different types of energy which help us describe
More informationSurface Temperatures in Dry Friction and Boundary Lubrication*
Surface Temperatures in Dry Friction and Boundary Lubrication* by Toshio Sakurai**, Heihachiro Okabe** and A. Sethuramiah** Summary: Surface temperatures were measured by dynamic thermocouple technique.
More informationKeywords- Source coding, Huffman encoding, Artificial neural network, Multilayer perceptron, Backpropagation algorithm
Volume 4, Issue 5, May 2014 ISSN: 2277 128X International Journal of Advanced Research in Computer Science and Software Engineering Research Paper Available online at: www.ijarcsse.com Huffman Encoding
More informationDeep Neural Networks (1) Hidden layers; Back-propagation
Deep Neural Networs (1) Hidden layers; Bac-propagation Steve Renals Machine Learning Practical MLP Lecture 3 4 October 2017 / 9 October 2017 MLP Lecture 3 Deep Neural Networs (1) 1 Recap: Softmax single
More informationBackpropagation Neural Net
Backpropagation Neural Net As is the case with most neural networks, the aim of Backpropagation is to train the net to achieve a balance between the ability to respond correctly to the input patterns that
More informationDEVELOPMENT OF ARTIFICIAL NEURAL NETWORK MODELLING FOR ADSORPTION OF INDIUM USING CARBON MATERIALS
DEVELOPMET OF ARTIFICIAL EURAL ETWORK MODELLIG FOR ADSORPTIO OF IDIUM USIG CARBO MATERIALS D.S.P. Franco¹; Fábio A. Duarte ;.P.G. Salau¹, G.L.Dotto 1 1 Chemical Engineer Department Federal University of
More information(Refer Slide Time: 00:58)
Nature and Properties of Materials Professor Bishak Bhattacharya Department of Mechanical Engineering Indian Institute of Technology Kanpur Lecture 18 Effect and Glass Transition Temperature In the last
More informationData Mining Part 5. Prediction
Data Mining Part 5. Prediction 5.5. Spring 2010 Instructor: Dr. Masoud Yaghini Outline How the Brain Works Artificial Neural Networks Simple Computing Elements Feed-Forward Networks Perceptrons (Single-layer,
More informationSTUDENT LABORATORY WORKSHEET EXPERIMENT A: NATURAL NANOMATERIALS
STUDENT LABORATORY WORKSHEET EXPERIMENT A: NATURAL NANOMATERIALS Student name: Date:.. AIM: Learn about the existence of natural nanomaterials Light interaction with colloids Gelatine and milk as examples
More informationNeural Networks: Basics. Darrell Whitley Colorado State University
Neural Networks: Basics Darrell Whitley Colorado State University In the Beginning: The Perceptron X1 W W 1,1 1,2 X2 W W 2,1 2,2 W source, destination In the Beginning: The Perceptron The Perceptron Learning
More informationA simple empirical model for calculating gain and excess noise in GaAs/Al ξ Ga 1 ξ As APDs (0.3 ξ 0.6)
A simple empirical model for calculating gain and excess noise in GaAs/Al ξ Ga 1 ξ As APDs (0.3 ξ 0.6) Mohammad Soroosh 1, Mohammad Kazem Moravvej-Farshi 1a), and Kamyar Saghafi 2 1 Advanced Device Simulation
More informationRao, M.A. Department of Food Science and Technology, Cornell University, Geneva, NY, USA
FOOD RHEOLOGY AND TEXTURE Rao, M.A. Department of Food Science and Technology, Cornell University, Geneva, NY, USA Keywords: Shear viscosity, shear-thinning behavior, extensional viscosity, sensory stimuli,
More informationArtificial Intelligence (AI) Common AI Methods. Training. Signals to Perceptrons. Artificial Neural Networks (ANN) Artificial Intelligence
Artificial Intelligence (AI) Artificial Intelligence AI is an attempt to reproduce intelligent reasoning using machines * * H. M. Cartwright, Applications of Artificial Intelligence in Chemistry, 1993,
More informationThermochemistry/Calorimetry. Determination of the enthalpy of vaporization of liquids LEC 02. What you need: What you can learn about
LEC 02 Thermochemistry/Calorimetry Determination of the enthalpy of vaporization of liquids What you can learn about Enthalpy of vaporisation Entropy of vaporisation Trouton s rule Calorimetry Heat capacity
More informationApplication of Neural Network Analysis to Correlate the Properties of Plasma Spray Coating
Indian Institute of Technology Kharagpur From the SelectedWorks of Ajit Behera Winter December, 2012 Application of Neural Network Analysis to Correlate the Properties of Plasma Spray Coating Ajit Behera,
More informationWeight Initialization Methods for Multilayer Feedforward. 1
Weight Initialization Methods for Multilayer Feedforward. 1 Mercedes Fernández-Redondo - Carlos Hernández-Espinosa. Universidad Jaume I, Campus de Riu Sec, Edificio TI, Departamento de Informática, 12080
More informationThe particle swarm optimization algorithm: convergence analysis and parameter selection
Information Processing Letters 85 (2003) 317 325 www.elsevier.com/locate/ipl The particle swarm optimization algorithm: convergence analysis and parameter selection Ioan Cristian Trelea INA P-G, UMR Génie
More informationWeek 5: Logistic Regression & Neural Networks
Week 5: Logistic Regression & Neural Networks Instructor: Sergey Levine 1 Summary: Logistic Regression In the previous lecture, we covered logistic regression. To recap, logistic regression models and
More informationDr. L. I. N. de Silva. Student Name Registration Number: Assessed By: Lecturers Remarks
Module - CE 2042 Soil Mechanics and Geology-1 Assignment Tests for Particle Size Distribution Analysis Marks 10% Learning Ability to conduct particle size distribution analysis of soils Outcome Ability
More informationMolecular Diffusion Through a Porous Medium. Steve Cavnar, Marc Sehgal, Anthony Martus, and Abdullah Awamleh
Molecular Diffusion Through a Porous Medium Steve Cavnar, Marc Sehgal, Anthony Martus, and Abdullah Awamleh University of Michigan ChemE 342: Heat and Mass Transfer Introduction The demonstration and instruction
More informationSupplementary Information
Supplementary Information 1. Thermodynamic data The isomerization of glucose into fructose using can be represented as: The equilibrium constant K eq and equilibrium conversion were calculated as follows:
More informationST. STEPHEN S GIRLS COLLEGE Mid Year Examination PHYSICS Time Allowed: 1 hour 30 minutes NAME: F.3 ( ) MARKS:
F.3 Physics Mid Year Examination 2005-2006 page 1 Form 3 193 students ST. STEPHEN S GIRLS COLLEGE Mid Year Examination 2005-2006 PHYSICS Time Allowed: 1 hour 30 minutes YRKwong, WYYau NAME: F.3 ( ) MARKS:
More information