W. Harmon Ray’s research while affiliated with University of Wisconsin–Madison and other places

... So, there is a need to be more focused in this direction in order to improve the productivity and economic aspect of the process industries by minimizing the requirement of more chemicals, water IOP Publishing doi: 10.1088/1757-899X/1022/1/012010 2 and man power.This can be achieved by optimizing various design parameters by using various modern control system techniques. ...

Reference:

A Review on Control System Applications in Industrial Processes

... As observed in our previous SMAD model, a sharp drop in the level of functional TGF-β receptors leads to a marked decrease in the activity of receptor-activated SMAD complex. From a control theory perspective, the loss of functional TGF-β receptors corresponds to a reduction in the "process gain," as a result of which only a small portion of the input signal is "transduced" into the response [47,48]. If, in addition to the loss of receptors, receptor internalization is biased toward lipid rafts, then activity of the SMAD complex is reduced even further, to an almost undetectable level. ...

Reference:

Computational Modeling and Analysis of the TGF-β-induced ERK and SMAD Pathways

... Villa et al first established the dynamics of polymerization reactors with evaporative cooling and wall heat transfer [17] . They showed that the successful design and operation of such reactors, which relies on reflux condensers for heat removal , requires careful consideration of stability issues. ...

Reference:

Scale-up of polymerization processes

... Although, the industrial processes are generally nonlinear yet FLC technique has proven to be very effective for it. Scott, Ray (1993) has extensively studied on applications of knowledge-based artificial neural network for various practical solutions of many intricate problems of linear model structures. He cited a comparison between conventional PID and ANN-based control and demonstrated how refining of PID controller using neural networks has been possible. ...

Reference:

IGDTUW AIST-2020

... Meanwhile, the feed forward and back propagation is the most famous training algorithm to train the ANN [36]. The multilayer perceptron (MLP) feed forward neural network is the commonest prediction architecture [22,37]. The MLP architecture includes one input layer, one output layer and one or more hidden layers depending on the complexity of the problem being examined. ...

Reference:

Application of the Artificial Neural Network to Predict the Bending Strength of the Engineered Laminated Wood Produced Using the Hydrolyzed Soy Protein-Melamine Urea Formaldehyde Copolymer Adhesive

... The wave model has the advantage of providing a low-order approximation that describes the property of interest directly and can be used in model-based control schemes [11,12]. Profile control strategies have also been used for control of packed bed reactors to ensure that the reactor hotspot is below the safety limit, which in turn avoids catalyst deactivation [13,14]. A full profile control of temperature in an FBR has been experimentally demonstrated by Yoshida and Matsumoto [15]. ...

Reference:

Profile control in distributed parameter systems using lexicographic optimization based MPC

... Nevertheless, the theory for systems governed by differential algebraic equations (DAEs) is not so well developed and is modestly discussed in the open literature, and the main results available are usually obtained for simple cases, such as those described by index-1 DAE systems. Besides, open and public domain computational codes for nonlinear analysis of DAEs are almost nonexistent in the literature (see, for instance, Hyanek et al., 1995; Kienle, et al., 1995; Ochs et al., 1996; Mangold et al., 2000). ...

Reference:

Stability analysis of differential-algebraic equations in AUTO_DAE

... Hence, it is required to obtain using input-output data. An example fascinating this approach is singular value decomposition (SVD) method to estimate Green's function model [97]. The disadvantage concerning this approach is that it provides lower order model only for time-invariant systems. ...

Reference:

IETE Technical Review Volterra and Wiener Model Based Temporally and Spatio-Temporally Coupled Nonlinear System Identification: A Synthesized Review Volterra and Wiener Model Based Temporally and Spatio-Temporally Coupled Nonlinear System Identification: A Synthesized Review

... Several mathematical models have been proposed for the free-radical polymerization of styrene (Berkenwald et al. 2016;Kalfas and Ray 1993;Kovenklioglu 1980;Seyedi et al. 2020; Soriano-Corral and Morales 2011). However, the number of publications discussing silver nanoparticles' effect on the free-radical bulk polymerization of styrene is limited (Soriano-Corral and Morales 2011; Yanagihara et al. 1999;. ...

Reference:

Nanocomposites production of polystyrene/silver obtained by embedding silver nanoparticles in situ with styrene polymerization

... Vinyl acetate Semibatch FPE Galerkin Solubility data - [13] Styrene-methyl methacrylate CSTR PB OCM Partition coefficients - [14,15] Vinyl acetate-butyl acrylate Semibatch PB OCM Partition coefficients Yes [16] Sty-butadiene, acrylonitrile-butadiene CSTR 0-1-2 Method of moments Partition coefficients - [17] Vinyl chloride CSTR 0-1-2 Finite volume (not reported) - [18] Vinyl chloride Batch 0-1-2 Finite volume Solubility data - [19] Vinyl chloride Batch 0-1 Finite volume Partition coefficient Yes [20] Styrene, butyl acrylate Batch 0-1/PB Backward finite difference Solubility data Yes [12] Styrene Semibatch FPE Finite volume Morton equation Yes [11] Styrene-butyl acrylate Batch and semibatch PB Fixed pivots Partition coefficients Yes [21] The numerical solution of hyperbolic PDE's is challenging due to the possible presence of steep fronts as in the case of the inviscid Burger's equation representing shock-waves [22]. One way to reduce the set of PDEs to a set of ordinary differential equations (ODEs) in time is using the weighted residual method (MWR) on finite elements, in which the solution in each element is approximated with either first-, second-, or third-order polynomials of the internal coordinate (mass or diameter particle) with time-dependent coefficients [23]. ...

Reference:

Precise Modeling of the Particle Size Distribution in Emulsion Polymerization: Numerical and Experimental Studies for Model Validation under Ab Initio Conditions