Modeling Methodology Based on Stacked Neural Networks for Inferential Prediction of Polypropylene Melt Index

A modeling methodology based on stacked neural networks by combining several individual networks in parallel is proposed. Stacked neural network as an effective method for modeling of inherently complex and nonlinear systems especially a system with a limited number of experimental data points is chosen for yield prediction. Proper determination of the stacking weights is essential for good stacked neural networks model performance, so determination of appropriate weights for combining individual networks using robust least squares estimation is proposed. Inferential prediction of melt index as the most important characteristic process of polypropylene polymerization has been carried out. The application of the proposed modeling method based on stacked neural networks to the development of melt index soft sensor in an industrial propylene polymerization plant demonstrates its effectiveness.

Download Full-text

Estimation of Polypropylene Melt Index through the Application of Stacked Neural Networks

Advanced Materials Research ◽

10.4028/www.scientific.net/amr.187.411 ◽

2011 ◽

Vol 187 ◽

pp. 411-415

Author(s):

Lu Yue Xia ◽

Hai Tian Pan ◽

Meng Fei Zhou ◽

Yi Jun Cai ◽

Xiao Fang Sun

Keyword(s):

Neural Network ◽

Neural Networks ◽

Network Model ◽

Neural Network Model ◽

Model Performance ◽

Measurement Interval ◽

Melt Index ◽

Polypropylene Melt ◽

Single Neural Network

Melt index is the most important parameter in determining the polypropylene grade. Since the lack of proper on-line instruments, its measurement interval and delay are both very long. This makes the quality control quite difficult. A modeling approach based on stacked neural networks is proposed to estimation the polypropylene melt index. Single neural network model generalization capability can be significantly improved by using stacked neural networks model. Proper determination of the stacking weights is essential for good stacked neural networks model performance, so determination of appropriate weights for combining individual networks using the criteria about minimization of sum of absolute prediction error is proposed. Application to real industrial data demonstrates that the polypropylene melt index can be successfully estimated using stacked neural networks. The results obtained demonstrate significant improvements in model accuracy, as a result of using stacked neural networks model, compared to using single neural network model.

Download Full-text

Inferential estimation of polypropylene melt index using stacked neural networks based on absolute error criteria

2010 International Conference on Computer, Mechatronics, Control and Electronic Engineering ◽

10.1109/cmce.2010.5610339 ◽

2010 ◽

Author(s):

Luyue Xia ◽

Haitian Pan

Keyword(s):

Neural Networks ◽

Absolute Error ◽

Melt Index ◽

Polypropylene Melt

Download Full-text

Predicting Viscosities of Heavy Oils and Solvent–Heavy Oil Mixtures Using Artificial Neural Networks

Journal of Energy Resources Technology ◽

10.1115/1.4049603 ◽

2021 ◽

Vol 143 (11) ◽

Author(s):

Zehua Chen ◽

Daoyong Yang

Keyword(s):

Neural Networks ◽

Artificial Neural Networks ◽

Model Performance ◽

Viscosity Data ◽

Heavy Oils ◽

Ann Model ◽

Wide Range ◽

Data Points ◽

Ann Models ◽

Artificial Neural

Abstract This study investigates the potential of artificial neural networks (ANNs) to accurately predict viscosities of heavy oils (HOs) as well as mixtures of solvents and heavy oils (S–HOs). The study uses experimental data collected from the public domain for HO viscosities (involving 20 HOs and 568 data points) and S–HO mixture viscosities (involving 12 solvents and 4057 data points) for a wide range of temperatures, pressures, and mass fractions. The natural logarithm of viscosity (instead of viscosity itself) is used as predictor and response variables for the ANNs to significantly improve model performance. Gaps in HO viscosity data (with respect to pressure or temperature) are filled using either the existing correlations or ANN models that innovatively use viscosity ratios from the available data. HO viscosities and mixture viscosities (weight-based, molar-based, and volume-based) from the trained ANN models are found to be more accurate than those from commonly used empirical correlations and mixing rules. The trained ANN model also fares well for another dataset of condensate-diluted HOs.

Download Full-text

Determination of S-N curves with the application of artificial neural networks

Fatigue & Fracture of Engineering Materials & Structures ◽

10.1046/j.1460-2695.1999.00198.x ◽

1999 ◽

Vol 22 (8) ◽

pp. 723-728 ◽

Cited By ~ 1

Author(s):

Artymiak ◽

Bukowski ◽

Feliks ◽

Narberhaus ◽

Zenner

Keyword(s):

Neural Networks ◽

Artificial Neural Networks ◽

Artificial Neural

Download Full-text

Determination of the International Sensitivity Index of a New Near-Patient Testing Device to Monitor Oral Anticoagulant Therapy

Thrombosis and Haemostasis ◽

10.1055/s-0038-1657641 ◽

1997 ◽

Vol 78 (02) ◽

pp. 855-858 ◽

Cited By ~ 25

Author(s):

Armando Tripodi ◽

Veena Chantarangkul ◽

Marigrazia Clerici ◽

Barbara Negri ◽

Pier Mannuccio Mannucci

Keyword(s):

Linear Relationship ◽

Whole Blood ◽

Oral Anticoagulant ◽

Oral Anticoagulant Therapy ◽

Oral Anticoagulants ◽

Calibration Model ◽

Testing Device ◽

Data Points ◽

Near Patient Testing

SummaryA key issue for the reliable use of new devices for the laboratory control of oral anticoagulant therapy with the INR is their conformity to the calibration model. In the past, their adequacy has mostly been assessed empirically without reference to the calibration model and the use of International Reference Preparations (IRP) for thromboplastin. In this study we reviewed the requirements to be fulfilled and applied them to the calibration of a new near-patient testing device (TAS, Cardiovascular Diagnostics) which uses thromboplastin-containing test cards for determination of the INR. On each of 10 working days citrat- ed whole blood and plasma samples were obtained from 2 healthy subjects and 6 patients on oral anticoagulants. PT testing on whole blood and plasma was done with the TAS and parallel testing for plasma by the manual technique with the IRP CRM 149S. Conformity to the calibration model was judged satisfactory if the following requirements were met: (i) there was a linear relationship between paired log-PTs (TAS vs CRM 149S); (ii) the regression line drawn through patients data points, passed through those of normals; (iii) the precision of the calibration expressed as the CV of the slope was <3%. A good linear relationship was observed for calibration plots for plasma and whole blood (r = 0.98). Regression lines drawn through patients data points, passed through those of normals. The CVs of the slope were in both cases 2.2% and the ISIs were 0.965 and 1.000 for whole blood and plasma. In conclusion, our study shows that near-patient testing devices can be considered reliable tools to measure INR in patients on oral anticoagulants and provides guidelines for their evaluation.

Download Full-text

Operational Determination of the Activated Sludge Process Using Neural Networks

Water Science & Technology ◽

10.2166/wst.1992.0762 ◽

1992 ◽

Vol 26 (9-11) ◽

pp. 2461-2464 ◽

Cited By ~ 2

Author(s):

R. D. Tyagi ◽

Y. G. Du

Keyword(s):

Neural Network ◽

Neural Networks ◽

Steady State ◽

Activated Sludge ◽

Feedforward Neural Network ◽

Training Data ◽

Activated Sludge Process

A steady-statemathematical model of an activated sludgeprocess with a secondary settler was developed. With a limited number of training data samples obtained from the simulation at steady state, a feedforward neural network was established which exhibits an excellent capability for the operational prediction and determination.

Download Full-text

Determination of the geophysical model function of the ERS-1 scatterometer by the use of neural networks

Journal of Geophysical Research Atmospheres ◽

10.1029/97jc02178 ◽

1998 ◽

Vol 103 (C6) ◽

pp. 12853-12868 ◽

Cited By ~ 10

Author(s):

Carlos Mejia ◽

Sylvie Thiria ◽

Ngan Tran ◽

Michel Crépon ◽

Fouad Badran

Keyword(s):

Neural Networks ◽

Model Function ◽

Geophysical Model ◽

Geophysical Model Function

Download Full-text

Determination of application volume for coffee plantations using artificial neural networks and remote sensing

Computers and Electronics in Agriculture ◽

10.1016/j.compag.2021.106096 ◽

2021 ◽

Vol 184 ◽

pp. 106096

Author(s):

Mailson Freire de Oliveira ◽

Adão Felipe dos Santos ◽

Elizabeth Haruna Kazama ◽

Glauco de Souza Rolim ◽

Rouverson Pereira da Silva

Keyword(s):

Remote Sensing ◽

Neural Networks ◽

Artificial Neural Networks ◽

Coffee Plantations ◽

Artificial Neural ◽

Application Volume

Download Full-text

Non-Destructive Micromagnetic Determination of Hardness and Case Hardening Depth Using Linear Regression Analysis and Artificial Neural Networks

Metals ◽

10.3390/met11010018 ◽

2020 ◽

Vol 11 (1) ◽

pp. 18

Author(s):

Rahel Jedamski ◽

Jérémy Epp

Keyword(s):

Neural Network ◽

Neural Networks ◽

Regression Analysis ◽

Artificial Neural Networks ◽

Linear Regression ◽

Linear Regression Analysis ◽

Case Hardening ◽

Artificial Neural ◽

Non Destructive

Non-destructive determination of workpiece properties after heat treatment is of great interest in the context of quality control in production but also for prevention of damage in subsequent grinding process. Micromagnetic methods offer good possibilities, but must first be calibrated with reference analyses on known states. This work compares the accuracy and reliability of different calibration methods for non-destructive evaluation of carburizing depth and surface hardness of carburized steel. Linear regression analysis is used in comparison with new methods based on artificial neural networks. The comparison shows a slight advantage of neural network method and potential for further optimization of both approaches. The quality of the results can be influenced, among others, by the number of teaching steps for the neural network, whereas more teaching steps does not always lead to an improvement of accuracy for conditions not included in the initial calibration.

Download Full-text