Issues regarding artificial neural network modeling for reactors and fermenters

Abstract

In recent years researchers in many areas have used arti®cial neural networks (ANNs) to model a variety of physical relationships. While in many cases this selection appears sound and reasonable, one must remember than ANN modeling is an empirical modeling technique (based on data) and is subject to the limitations of such techniques. Poor prediction occurs when the training data set does not contain adequate``information'' to model a dynamic process. Using data from a simulated continuousstirred tank reactor, this paper illustrates four scenarios:(1) steady state, (2) large process time constant, (3) infrequent sampling, and (4) variable sampling rate. The ®rst scenario is typical of simulation studies while the other three incorporate attributes found in real plant data. For the cases in which ANNs predicted well, linear regression (LR), one of the oldest empirical modeling techniques, predicted equally well, and when LR failed to accurately model/predict the data, ANNs predicted poorly. Since real plant data would resemble a combination of situations (2), (3), and (4), it is important to understand that empirical models are not necessarily appropriate for predictively modeling dynamic processes in practice. IntroductionArti®cial neural network (ANN) models have recently been used to model a variety of complex nonlinear physical relationships. Empirical techniques that have been employed to predictively model a dynamic process include radial basis function models for a continuous-stirred tank reactor (CSTR) [1], autoregressive moving average with external input (AR-MAX) models for a distillation process [2], an ANN for a continuous-stirred tank fermenter [3], and the recently introduced semi-empirical technique (SET) demonstrated for a CSTR [4]. ANNs fall into the class of purely empirical methods since their structures are seldom, if ever, phenomenologically inspired, and their coef®cients are determined strictly by the data and seldom have physical meaning. As an empirical technique, ANN models are¯exible, relatively easy to ®t, and typically perform very well in applications suited for empirical modeling. However, due to the ease of obtaining accurate ®ts to data sets, they can be misapplied and misused just as easily.This paper demonstrates important limitations of empirical models in accurately predicting model outputs (i.e., responses) to changes in inputs (disturbances) for real data taken from dynamic chemical, or biochemical process. In addition to ANN models, linear regression (LR), the most common empirical modeling technique, is tested. Both ANNs and LR predict well under steady state conditions, with data sampled at every input change, and under dynamic conditions, with constant time intervals between input changes and data sampled at very input change. However, under more complex sampling conditions found in real dynamic plant data such as infrequent sampling and variable sampling rate, empirical methods will likely predict poorly. When process data are sampled infrequently, input changes (there c...

Hybrid filtering of feed stream noise from oscillating yeast cultures by combined Kalman and neural network configurations

“…However, commercial and proprietary interests restrict the disclosure and public use of industrial data. Therefore, a common practice is to add noise to an experimentally validated laboratory-scale model and solve the model to generate data simulating a real noise-affected fermentation [17,23,24]. This practice was followed in previous studies of oscillating S. cerevisiae fermentations [12,15,16,18], and the data generated there were used in the present analyses also, thus maintaining consistency.…”

Lyapunov comparison of noise‐filtering methods for oscillating yeast cultures

“…Moreover, industrial data are sometimes not sufficiently detailed or sampled in a manner suitable for modeling; interventions to sample more data or more variables may be costly or may upset production schedules. Therefore many authors (Chen and Rollins, 2000;Schubert et al, 1994;Simutis and Lubbert, 1997;Tian et al, 2002) have found it expedient to generate data mimicking industrial conditions by adding noise to a model established through laboratory experiments and solving the model under the conditions of interest. This method also allows exploration of wide ranges of parameters and operating conditions, so that a few promising choices thus obtained may be applied in plant operation.…”

An Unrestricted Algorithm for Accurate Prediction of Multiple-Input Multiple-Output (MIMO) Wiener Processes

“…Another limitation of discrete-time models is that they are adversely affected when sampling is inconstant or infrequent. 8 A significant advantage of DTM over CTM is that prediction requires only a few recent input changes at most, whereas CTM can be dependent on all previous input changes and requires a fading memory treatment. 3 We recently developed constrained DTM for both decomposed MISO Hammerstein and Wiener systems that can effectively build models from sequential step tests in a two-stage approach.…”