A hybrid neural network‐first principles approach to process modeling

AICHE Journal - Tập 38 Số 10 - Trang 1499-1511 - 1992
Dimitris C. Psichogios1, Lyle Ungar1
1Department of Chemical Engineering , University of Pennsylvania , Philadelphia, PA, 19104

Tóm tắt

AbstractA hybrid neural network‐first principles modeling scheme is developed and used to model a fedbatch bioreactor. The hybrid model combines a partial first principles model, which incorporates the available prior knowledge about the process being modeled, with a neural network which serves as an estimator of unmeasured process parameters that are difficult to model from first principles. This hybrid model has better properties than standard “black‐box” neural network models in that it is able to interpolate and extrapolate much more accurately, is easier to analyze and interpret, and requires significantly fewer training examples. Two alternative state and parameter estimation strategies, extended Kalman filtering and NLP optimization, are also considered. When no a priori known model of the unobserved process parameters is available, the hybrid network model gives better estimates of the parameters, when compared to these methods. By providing a model of these unmeasured parameters, the hybrid network can also make predictions and hence can be used for process optimization. These results apply both when full and partial state measurements are available, but in the latter case a state reconstruction method must be used for the first principles component of the hybrid model.

Từ khóa


Tài liệu tham khảo

Anderson B. D. O., 1979, Optimal Filtering

10.1080/00207178108922522

10.1080/00207178308933084

10.1016/0098-1354(90)87028-N

Bhat N. andT. McAvoy “Information Retrieval From A Stripped Back‐propagation Network ”AIChE Meeting(1990).

10.1021/ie00096a013

10.1016/0098-1354(85)85014-6

10.1016/0098-1354(85)80004-1

10.1080/00986449008940684

Goodwin G. C., 1984, Adaptive Filtering Prediction and Control

Gostomski P. A. B. W. Bequette andH. R.Bungay “Multivariable Control of a Continuous Culture ”AIChE Meeting(1990).

Haesloop D. andB. R. Holt “A Neural Network Structure for System Identification ”Proc. of the American Control Conf. 2460(1990).

Hernandez E. andY. Arkun “Neural Network Modeling and an Extended DMC Algorithm to Control Nonlinear Systems ”Proc. of the American Control Conf. 2454(1990).

10.1021/i200034a037

10.1109/72.80236

10.1016/S0019-9958(75)90382-4

10.1002/aic.690330408

Lant P. A. M. J. Willis G. A.Montague M. T.Tham andA. J.Morris “A Comparison of Adaptive Estimation With Neural Based Techniques for Bioprocess Application ”Proc. of the American Control Conf. 2173(1990).

Leonard J. A. M. A. Kramer andL. H.Ungar “A Neural Network Architecture That Computes Its Own Reliability ”Comp. Chem. Eng. submitted (1991).

10.1016/0098-1354(92)80053-C

Mozer M. C., 1989, Advances in Neural Information Processing

10.1021/ie00060a009

Rivera S. L. andM. N. Karim “Application of Dynamic Programming for Fermentative Ethanol Production by Zymomonas mobilis ”Proc. of the American Control Conf. 2144(1990).

Rumelhart D. G. Hinton andR.Williams “Learning Internal Representations by Error Propagation ”Parallel Distributed Processing: Explorations in the Microstructures of Cognition: 1. Foundations MIT Press (1986).

Sistu P. andW. Bequette “Process Identification Using Nonlinear Programming Techniques ”Proc. of the American Control Conf. 1534(1990).

Stinchcombe M. andH. White “Universal Approximation Using Feedforward Networks With Non‐Sigmoid Hidden Layer Activation Functions ”Proc. of the Int. Joint Conf. on Neural Networks 613(1989).

10.1002/bit.260361009

10.1002/aic.690170433

Werbos P. “Beyond Regression: New Tools for Prediction and Analysis in Behavioral Sciences ” PhD Thesis Harvard University (1974).

10.1049/ip-d.1991.0036

Thông tin
Thông tin xuất bản

AICHE Journal

Tập 38 Số 10

1499-1511

Thông tin tác giả