Neural Networks in Process Control - A Survey

Thibault, Jules; Grandjean, Bernard P. A.

doi:10.1016/b978-0-08-041267-2.50046-2

Cited by 35 publications

(31 citation statements)

References 58 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The training set consists of normalized biomass and substrate data from steady state and transient regimes obtained by dynamic step response experiments (Figs. 6,7,8,9) [12]. In case of step-type changes of the dilution rate D the culture requires some time to adapt to the new environmental conditions.…”

Section: Resultsmentioning

confidence: 99%

“…6,7,8,9) that the neural network A cannot be trained to fit the specific dynamic responses of the cell population and substrate concentration the step changes of the dilution rate before reaching the steady state, and especially the undershoots (overshoots). In contrast, the neural network models with time delay connections AC predict much better the biomass evolution (Figs.…”

Section: Resultsmentioning

confidence: 99%

“…Comparison between the results according to the substrate concentration (Figs. 7,9) gives significant priority to the second case with larger number of time delay connections. The undershoots (overshoots) of the substrate concentration are successfully predicted.…”

Section: Resultsmentioning

confidence: 99%

“…Among the variety of neural network architectures that have been proposed, the feedforward neural networks are by far the most frequently applied [7,9]. They consist of a layer of input neurons, a layer of output neurons and one or several layers of hidden (intermediate) neurons.…”

Section: Neural Network Modellingmentioning

confidence: 99%

See 3 more Smart Citations

Neural network modelling of fermentation processes. Microorganisms cultivation model

Petrova

Koprinkova

Patarinska

1997

Bioprocess Engineering

View full text Add to dashboard Cite

In the present paper the problem of chemostat modelling using the neural networks techniques is considered. A feedforward neural network with time delay feedback connections from and to the output neurons, which take into account the culture memory is proposed. A model of the growth of a strain Saccharomyces cerevisiae on a glucose limited medium is developed. Simulation investigations are carried out. The results are discussed.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

Section: Neural Network Modellingmentioning

confidence: 99%

See 2 more Smart Citations

Neural network modelling of fermentation processes. Microorganisms cultivation model

Petrova

Koprinkova

Patarinska

1997

Bioprocess Engineering

View full text Add to dashboard Cite

show abstract

“…Typically, some sigmoidal function is used for the output function f . Among the variety of NN architecture, the feedforward neural networks are by far most frequently applied in control problems [16]. They consist of a layer of input neurons, a layer of output neurons and one or several hidden (intermediate) layers.…”

Section: 2mentioning

confidence: 99%

Neural network modelling of fermentation processes

Petrova

Koprinkova

Patarinska

et al. 1998

Bioprocess Engineering

View full text Add to dashboard Cite

This paper deals with an application of neural networks for chemostat modelling. A feedforward neural network, taking into account culture memory is proposed for the speci®c growth rate approximation within the framework of the classical unstructured model. The investigations are carried out for different network topologies on the example of the growth of a strain Saccharomyces cerevisiae on a glucose limited medium and a model suitable for control synthesis is proposed.

show abstract

Gas atomization of polymers. II. Computational neural network modeling

Otaigbe

Noid

Sumpter

1998

Adv. Polym. Technol.

View full text Add to dashboard Cite

A novel gas atomization process (GAP) was designed for mass producing high‐quality spherical powders from molten polymers. The dependence of the powder characteristics, such as particle size and size distribution on the GAP control variables, and physical properties of the polymers were determined experimentally. Statistical and computational neural network analyses were applied to the data in both the forward and reverse modes, leading to models for predicting the particle size distribution from the process variables and vice versa. The computational neural network (CNN) model predicts an increase in the weight fraction of the smallest particle size range (0–53 μm) by increasing the polymer melt temperature while decreasing the polymer melt stream size for the more crystalline polyethylene‐based material studied. Comparisons of the CNN model predictions with experimental data yields good agreement and demonstrates that the CNN can be used to qualitatively design new experiments for the production of spherical powders from molten polymers, thereby reducing the amount of experimental work required to optimize the process. © 1998 John Wiley & Sons, Inc. Adv Polym Techn 17: 161–173, 1998

show abstract

Neural Networks in Process Control - A Survey

Cited by 35 publications

References 58 publications

Neural network modelling of fermentation processes. Microorganisms cultivation model

Neural network modelling of fermentation processes. Microorganisms cultivation model

Neural network modelling of fermentation processes

Gas atomization of polymers. II. Computational neural network modeling

Contact Info

Product

Resources

About