Case-based selection of a model of a reverse flow reactor

Abstract

Key steps in configuration of a reverse flow reactor for industrial application include finding the most suitable operation model and identifying reactor behaviour. The design of a configured reactor requires knowledge about the influence of operating conditions on overall process performance and precise evaluation of operating parameters range for which a sustained dynamic behaviour can be obtained. In case of a reverse flow reactor and in general in case of a complex reactor the design requires knowledge about reactor behaviour, process operation, important system parameters and their optimum values, the way of mathematical formulation and phenomena description, solving methods, control specific aspects, etc. One possible way for finding such information is studying one by one all the articles related to the reverse flow reactors which is often a time consuming and moreover a tedious task. Despite of the relatively large body of the literature no detail procedure exists for the design of the reverse flow reactors in order to cover all possible applications. Moreover, the design is carried out by a trial and error approach coupled with extensive and detailed numerical simulations. Therefore, in this paper a much faster and simpler procedure to optimize the way of knowledge gathering and information gaining for design purposes is discussed. This procedure uses the capabilities of a computational tool which provides specific information that may be used inside a reasoning process to find new solutions for new problems based on past experience. The paper is dealing with the implementation and testing of a computer tool capable to provide design information for reverse flow reactors. The computer tool relies on a decision supporting method which involves a case-based reasoning (CBR) algorithm. The tool querying procedure was validated and once the reliability of the tool was proofed finding solutions for new problems based on past experience becomes less time consuming and much more efficient than a classical literature survey. In this paper, the reverse flow reactor model selection for design purposes has been identified using a case-based reasoning tool capable of interactive querying and abstraction procedure. The results of querying procedure have been adapted in order to fit to new needs. The case study in this paper has been the selective catalytic reduction of NO x with ammonia in a reverse flow auto-thermal reactor. The solution provided by the computer tool has been adapted to the new problem and validated through external simulation.