Abstract

Experimental investigations have been carried out to measure axial effective thermal conductivities of

packed beds for a number of particles and catalyst pellets. Measurements were made for three gases (air,

nitrogen and carbon dioxide) in beds packed with ball bearings, copper chromite, chromia alumina, alumina

hollow cylinders and alumina spheres. A glass vacuum vessel was employed for most measurements,

but a thin wall stainless steel vessel was used in a few experiments.

Empirical correlations to predict the axial effective thermal conductivity of packed bed reactors have

been derived from the experimental results.

2002 Elsevier Science Ltd. All rights reserved.

Keywords: Axial thermal conductivities; Packed beds

This paper concerns the adequacy of existing detailed reaction mechanisms for use in computer simulations of combustion systems with injection of gaseous fuels such as hydrogen, and methane. Shock tube induction time data are compiled from the literature and compared to thermodynamic conditions of gas combustion systems to establish validation limits. Existing detailed reaction mechanisms are then used in constant-volume explosion simulations for validation against the shock tube data. A quantitative measure of mechanism accuracy is obtained from the validation study results, and deficiencies in the experimental data and reaction mechanisms are highlighted. 

This paper presents empirical methods for enhancing the accuracy of inductive learning systems. It

addresses the problems of: learning propositional production rules in multi-class classification tasks in

noisy domains, maintaining continuous learning when confronted with new situations after the learning

phase is completed, and classifying an object when no rule is satisfied for it.

It is shown that interleaving the learning and performance-evaluation processes, allows accurate

classifications to be made on real-world data sets. The paper presents the system ARIS which

implements this approach, and it is shown that the resulting classifications are often more accurate than

those made by the non-refined knowledge bases.

The core design decision that lies behind ARIS is that it employs an ordering of the rules according to

their weight. A rule’s weight is learned by using Bayes’ theorem to calculate weights for the rule’s

conditions and to combine them. This model focuses the analysis of the knowledge base and assists the

refinement process significantly.

The system is non-interactive, it relies on heuristics to focus the refinement on those experiments that

appear to be most consistent with the refinement data set. The design framework of ARIS consists of a

tabular model for expressing rule weights, and the relationship between refinement cases and the rules

satisfied for each case to focus the refinement process. The system has been used to refine knowledge

bases created by ARIS itself, as well as to refine knowledge bases created by the RIPPER and C4.5

systems [6,25] in ten selected domains. Two main advantages have been observed. First, the ability to

gradually improve the knowledge base as the refinement proceeds. Second, is the ability to learn strong

rules utilising condition’s weight using minimum covering algorithm. Thus, by taking account of this

information we improve the applicability and quality of refinement

Configurable Computer Design for Guidance and Control Algorithms

In this paper mappings of several different algorithms from the fields of operating systems and multimedia are discussed, together with the underlying system architecture. The system architecture is based on a board with four FPGA circuits. For the purpose of this research, a simulator of the system has been developed, and different interconnection networks, memory, and bus architectures have been tested. The results of the simulation, as well as analytical results, are presented.