MODEL-BASED CONTROL WITH STOCHASTIC SIMULATORS: BUILDING PROCESS DESIGN AND CONTROL SOFTWARE FOR CATALYTICALLY ENHANCED MICROSYSTEMS

Loading...
Thumbnail Image
Penn collection
Senior Design Reports (CBE)
Degree type
Discipline
Subject
Funder
Grant number
License
Copyright date
Distributor
Related resources
Author
Curnan, Matthew T
Gopalratnam, Arjun
Slominski, Charles G
Wang, Eric
Contributor
Abstract

The production, characteristics, dynamics, and economics of microreactors were studied in this report. Overall it was found that the best microfabrication techniques for small scale processes were laser ablation, the LIGA process, soft lithography, and anisotropic wet chemical etching, roughly in ascending order of effectiveness. One of the few viable bonding techniques was found to be diffusion bonding followed by microlamination, whereas many coating methods -- such as solgel coating, modified anodic oxidation, and electrophoretic deposition -- were effective in μTAS integration. The high surface area to volume ratio of microreactors enables precise control of the temperature of the reactor along its axial dimension. Taking advantage of this feature in the design of microreactors leads to better control of complex reaction networks and generates more valuable effluent streams. A model predictive controller was implemented for the common, archetypical reaction network involving the hydrogenation and dehydrogenation of cyclohexene with various control objectives. It was found that the highest rate of production of benzene and cyclohexane occurred at 600 K while the most pure stream of benzene occurred at 200 K. Model predictive control was found to be highly resistant to the inherent stochasticity of small scale processes. The market for a software-based controller for microreactors was surveyed and found to still be in the early stages of development. A profitability analysis was conducted for a start-up company using microreactors to make cyclohexane. A price of $18,000 for the product was found to be a reasonable selling price yet allowed the start-up to remain profitable.

Advisor
Date Range for Data Collection (Start Date)
Date Range for Data Collection (End Date)
Digital Object Identifier
Series name and number
Publication date
2010-04-01
Volume number
Issue number
Publisher
Publisher DOI
Journal Issue
Comments
Recommended citation
Collection