Neural network modeling hybrid

Neural Network based hybrid dynamic modelling and optimisation methods for conventional and unconventional column configurations... 

Chapter 12 describes three types of models production systems, neural networks, and hybrid models. The general structure of each is described as well as the most well known examples. 

Production systems, neural networks, and hybrid models... 

Application of artificial neural networks (ANN) for modelling of the kinetics of a catalytic hydrogenation reaction in a gas-liquid-solid system has been studied and discussed. The kinetics of the hydrogenation of 2,4-DNT over a palladium on alumina catalyst has been described with feedforward neural networks of dififerent architectures. A simple experimental procedure to supply learning data has been proposed. The accuracy and flexibility of the hybrid first principles-neural network model have been tested and compared with those of the classical model. 

In this study the application of AIW to modelling of the kinetics of a catalytic hydrogenation reaction in the gas-liquid-sohd system is studied and discussed. Also a hybrid first principle-neural network model, where conventional mass and energy balances are supported with a trained network, implemented to represent the reaction kinetics, is proposed and tested... 

The reactor model presented above has been used in two versions as a conventional model (CM) and as a hybrid model (HM), respectively. In the conventional model the reaction rates r j appearing in the balance equations of Eq. 7 have been calculated following the expressions for Langmuir-Hinshelwood kinetics [13] derived in our previous studies. In the hybrid first principles-neural network model (HM) the conventional kinetic subroutines in the CM algorithm have been replaced by the neural network, so in the HM the reaction rates r- have been supplied by the trained network. 

Trained nets have been implemented into the hybrid model. The net N4 - see Table 2 - represents the optimal topology. It assures a good accuracy of the predictions and simultaneously a relative simple architecture. A comparison of the results obtained with the hybrid first principles-neural network model and those predicted with the conventional one (CM) is shown in Figs. 4a-b. 

Nascimento, C.A.O. Giudici, R. NBeiler, I.C. Modeling of industrial nylon-6,6 polymerization process in a twin screw extruder reactor. II. Neural networks and hybrid models, J. Appl Polym. Sci., 1999, 72, 1, 905-912. 

Aguiar, H.C. and Filho, R.M., 2001, Neural network and hybrid model a discussion about different modeling techniques to predict pulping degree with industrial data, Chem. Eng. Sci., 56 565-570. 

Zorzetto, L.F.M., Filho, R.M. and Wolf-Maciel, M.R., 2000, Process modelling development through artificial neural networks and hybrid models. Comp. Chem. Eng. 24 1355-1360. 

Hayajneh, M.T., Hassan, A.M. and Mayyas, A.T. (2009) Artificial neural network modeling of the drilling process of self-lubricated aluminum/alumina/graphite hybrid composites synthesized by powder metallurgy technique, J Alloys Compd, 478 559-65. 

Y. Tian, J. Zhang, and J. Morris, Modeling and optimal control of a batch polymerization reactor using a hybrid stacked recurrent neural network model, Ind. Eng. Chem. Res., 40, 4525-4535, 2001. 

Phase-Switching Models Neural Network Models Pacemaker and Hybrid Models... 

Models of membrane reactors based on artificial neural networks and hybrid... 

Curdo, S., Calabro, V., lorio, G., Reduction and control of flux decline in cross-flow membrane processes modeled by artifidal neural networks and hybrid systems. Desalination, 2009,236(1-3), 234-243. 

Ou S., Achenie L.E.K., A hybrid neural network model for PEM fuel ceWs, Journal of Power Sources, 2005,140(2), 319-330. 

Psichogios, D. C., and Ungar, L. H., A hybrid neural network-first principles approach to process modeling. AIChEJ. 38, 1499 (1992). 

Greaves, M. A., Hybrid Modelling, Simulation and Optimisation of Batch Distillation Using Neural Network Techniques. Ph.D. Thesis, (University of Bradford, Bradford, UK, 2003). 

Dynamic sets of process-model mismatches data is generated for a wide range of the optimisation variables (z). These data are then used to train the neural network. The trained network predicts the process-model mismatches for any set of values of z at discrete-time intervals. During the solution of the dynamic optimisation problem, the model has to be integrated many times, each time using a different set of z. The estimated process-model mismatch profiles at discrete-time intervals are then added to the simple dynamic model during the optimisation process. To achieve this, the discrete process-model mismatches are converted to continuous function of time using linear interpolation technique so that they can easily be added to the model (to make the hybrid model) within the optimisation routine. One of the important features of the framework is that it allows the use of discrete process data in a continuous model to predict discrete and/or continuous mismatch profiles. 

In Greaves et al. (2001) and Greaves (2003), instead of using a rigorous model (as in the methodology described above), an actual pilot plant batch distillation column is used. The differences in predictions between the actual plant and the simple model (Type III and also in Mujtaba, 1997) are defined as the dynamic process-model mismatches. The mismatches are modelled using neural network techniques as described in earlier sections and are incorporated in the simple model to develop the hybrid model that represents the predictions of the actual column. 

So and Karplus [51] have developed a hybrid method that combines a GA for descriptor selection with an artificial neural network for model building. They found improved models for the Selwood data set when compared with the GFA and evolutionary programming methods, with the success being attributed to the ability of the neural network to select nonlinear descriptors. 

---