On-line modelling

In gridpoint models, transport processes such as speed and direction of wind and ocean currents, and turbulent diffusivities (see Section 4.8.1) normally have to be prescribed. Information on these physical quantities may come from observations or from other (dynamic) models, which calculate the flow patterns from basic hydrodynamic equations. Tracer transport models, in which the transport processes are prescribed in this way, are often referred to as off-line models. An on-line model, on the other hand, is one where the tracers have been incorporated directly into a d3mamic model such that the tracer concentrations and the motions are calculated simultaneously. A major advantage of an on-line model is that feedbacks of the tracer on the energy balance can be described... 

Off-line analysis, controller design, and optimization are now performed in the area of dynamics. The largest dynamic simulation has been about 100,000 differential algebraic equations (DAEs) for analysis of control systems. Simulations formulated with process models having over 10,000 DAEs are considered frequently. Also, detailed training simulators have models with over 10,000 DAEs. On-line model predictive control (MPC) and nonlinear MPC using first-principle models are seeing a number of industrial applications, particularly in polymeric reactions and processes. At this point, systems with over 100 DAEs have been implemented for on-line dynamic optimization and control. 

Simulation Sciences, Inc. Documentation for ROMEO (Rigorous On-line Modeling with Equation-based Optimization. Brea, CA (1999). 

Ideally, an on-line analyzer will be calibrated before it is installed in the process. It may be possible to accomplish this by calibrating it off-line with process grab samples and/or synthetic samples. It may be possible to install the analyzer in a lab-scale reactor, or in a semi-works or pilot plant. It may be possible to transfer to the on-line analyzer a method developed on an off-line analyzer or on another on-line analyzer (e.g. at a different plant site). However, sometimes none of these are possible and the analyzer will need to be calibrated on-line. The challenges of on-line model development (calibration) and validation, as well as approaches to dealing with them, are discussed below. For information related to calibration transfer issues, please see Chapter 12 of this book. 

When available, fundamental process models are preferred. For many complex processes such as composite manufacturing in general and autoclave curing in particular, however, these models are often not available. This lack of availability is due to an inadequate understanding of the complex events that take place during the process. A fundamental process model is occasionally available, but it is still unsuitable for on-line model predictive control application due to the extensive computing time required to solve the model s equations. This lack of... 

Results from these experimental runs were used as x, q data records to fit the parameters of six ANNs. In the experimental effort, a different feedforward ANN was used after each intermediate secondary measurement was obtained in the simulation-based effort, only one ANN accommodates all secondary measurements, and averaged dummy inputs are used for those secondary measurements not yet obtained. In addition in the experimental effort, a different ANN was used for final thickness and final void content predictions in the simulation-based effort, one ANN was used to predict both final thickness and final void content. The advantage of using one ANN to predict all values of q is that the parameters of only one ANN need be fitted. Fitting the parameters of an ANN for each variable in q is much more time-consuming. The disadvantage, however, is that the parameters A and abias are the same for each variable in q when just one ANN is used as an on-line model. When a different ANN is used for each variable in q, the parameters in A and abias are unique for each of those output variables, which results in increased on-line prediction accuracy. Similar speed-versus-accuracy arguments apply to the choice of one ANN for all secondary measurements versus an ANN for each secondary measurement. 

An on-line model has been developed for conservators to help them estimate indoor atmospheric concentrations of the common museum pollutants (Blades et al., 2000). This is especially useful if used in conjunction with targeted analytical campaigns to assess the atmosphere in a museum or gallery, although the interpretation of the results of such measurements is usually not straightforward (Section 12.8). 

Table 2.1 Model systems and typical applications of on-line models... 

Atmospheric aerosol particles modify the radiative transfer in the atmosphere and they have an impact on the cloud formation. Therefore, they alter the weather and they have an impact on climate. The anthropogenic part of this modification of the state of the atmosphere is currently not well understood and it raises the largest uncertainties with respect to climate change (see the IPCC report 2007). We developed a new on-line model system to investigate the aerosol-radiation-interaction on the regional scale. 

Identification of the capabilities of on-line models for air quality applications... 

Towards on-line model-based design of experiments... 

Towards On-line Model-Based Design of Experiments... 

This paper presents an on-line model based level control of a batch reactor with reaction rate uncertainties. The analyzed chemical batch process is catalyzed by a catalyst which decomposes in the reactor therefore it is fed several times during the batch. The chemical reaction produces a vapour phase by-product which causes level change in the system. The on-line control method is based on the shrinking horizon optimal control methodology based on the detailed model of the process. The results demonstrate that the on-line optimization based control strategy provides good control performance despite the disturbances. 

A novel gradient-based optimisation framework for large-scale steady-state input/output simulators is presented. The method uses only low-dimensional Jacobian and reduced Hessian matrices calculated through on-line model-reduction techniques. The typically low-dimensional dominant system subspaces are adaptively computed using efficient subspace iterations. The corresponding low-dimensional Jacobians are constructed through a few numerical perturbations. Reduced Hessian matrices are computed numerically from a 2-step projection, firstly onto the dominant system subspace and secondly onto the subspace of the (few) degrees of freedom. The tubular reactor which is known to exhibit a rich parametric behaviour is used as an illustrative example. 

---