Three Levels of Modeling

In this chapter, we have discussed three levels of modeling for dense gas-solid flows, with the emphasis on the technical details of each of the models, which have not been published elsewhere. Up till now, the models have mainly been used to obtain qualitative information, that is, to acquire insight into the mechanisms underlying the gas-solid flow structures. However, the ultimate objective of the multiscale approach is to obtain quantitative... 

As shown in Figure 1.36, Catalysis addresses three levels of modeling the problem domain or business, the component or system specification (externally visible behavior), and the internal design of the component or system (internal structure and behavior). 

Three levels of models have been used to describe gas exchange ... 

Applying this training model results in what can be seen as three levels of training topics. 

As implied in the diagram representing the GEMS model (Figure 2.5) and discussed in Section 2.6.3, certain characteristic error forms occur at each of the three levels of performance. This information can be used by the human-reliability analyst for making predictions about the forms of error expected in the various scenarios that may be considered as part of a predictive safety analysis. Once a task or portion of a task is assigned to an appropriate classification, then predictions can be made. A comprehensive set of techniques for error prediction is described in Chapter 5. 

In this chapter, we present an attempt to put these ideas into practice that has three components. First, we use a Model of Modelling , an external representation of the mental processes that we postulate to be undergone as a person forms a model. Second, this model is exemplified by its apphcation to the design of a teaching sequence about a key topic in chemistry, the imderstanding of which requires fluency of mental movement between the three levels of representation. Lastly, this application is implemented in such a way that the processes of students thought taking place can be monitored. 

What capability did the students have both to visualize the three levels of representation and to move between them, during this modelling-based teaching ... 

From the point of view of the focus of this book, the most important outcome was the demonstration that the use of the Model of Modelling as a basis for producing and conducting teaching activities enabled the students to demonstrate their capability within and between the three levels of representation. Taking each of the five aspects of metavisual capability, given earlier, in turn ... 

This study suggests that much more research is needed into aspects of metacog-nitive capability and into the use of the Model of Modelling if we are to provide fuller explanations of the processes involved in understanding the three levels of representation in chemistry . For example ... 

What is the relatiorrship between the way stndents nse the three levels of representation and their capability to test their models As testing models (from thought experiments and/or empirical ways) is an essential stage of the modelling process, their expertise in dealing with the three levels of representation seems to be vital for their learning in modelling-based corrtexts. 

Justi, R., Gilbert, J. K., Ferreira, P. F. M. (2008). The application of a model of modelling to illustrate the importance of metavisualization in respect of the three levels of representation. In J. K. Gilbert D. F. Treagust (Eds.), Multiple representations in chemical education. Dordrecht Springer. 

In previous chapters we developed in several places three levels of complexity of a binding system first, the direct-interaction model second, the model with indirect correlations arising from conformational changes in the entire adsorbent molecule and third, indirect correlations mediated through subunits. 

What information could be obtained if all ten experiments were carried out at only three levels of pH Three levels of x, (f= 3) provides sufficient factor combinations for being able to fit a three-parameter model, but leaves no degrees of freedom for estimating lack of fit / - p = 3 - 3 = 0. Because one of our objectives was to determine if a parabolic relationship provides an adequate model for the observed rate, we must be able to estimate the variance due to lack of fit of the model the number of factor combinations (levels of x, in this example) must therefore be greater than three. 

The most complete investigation of model response would establish a network of experimental cases throughout the area of interest. The more complete the coverage, obviously the more cases that will be required. For a system involving many variables the number of cases increases rapidly for example, in a system of six controllable variables an investigation at only two levels of each variable would require 26 = 64 cases. Examining three levels of each of the variables would require 36 = 729 cases. 

---