<p>Blood coagulation and protein adsorption were studied on a series of model surfaces. Silanes were used to covalently attach chemical groups (including sulfonate. amine, polyethylene oxide (PEO), lysine methyl, and octadecyl) to pure silica substrates. Surface chemical analysis by x-ray photoelectron spectroscopy (XPS) was used to quantify the elemental composition of these surfaces. All silane layers were estimated to be monolayers or less with thickness in the range 0.2 nm (for methyl groups) to 2.0 nm (for PEO). Contact angles were measured to help quantify the interaction between the model surfaces and pure solvents (water, glycerol, dimethyl sulfoxide (DMSO)). The data were fit to established interfacial tension (interfacial free energy) models but none was found to be adequate. A new empirical model, not related to surface energetics, was developed and demonstrated to be capable of fitting the data more precisely. Small deviations from the empirical model (± 2° to 8°) for DMSO and water contact angles were attributed to the possible influence of surface charge. Protein adsorption kinetics on the model surfaces was measured using ¹²⁵I-labeled fibrinogen and albumin as single proteins under static conditions. A simulation of transport and reaction in this experimental system was written and used to fit data to proposed models for protein adsorption. A simple bilayer mechanism was found to fit the kinetic data better than a number of multistate monolayer models for adsorption. The adsorption of fibrinogen from plasma was also measured in order to investigate a possible relationship between the Vroman effect and surface activated blood coagulation. A surface activated blood coagulation assay based on the partial thromboplastin time test was used to characterize the blood compatibility of model surfaces prepared on glass test tubes. The kinetics of cleavage of a fluorogenic substrate for thrombin was measured in real time on surfaces which were assumed to be closely comparable to those prepared on pure silica. Clotting times ranging between 2 and 11 minutes were measured. A correlation analysis was performed using surface chemistry, contact angle, protein adsorption, and blood coagulation data. Clotting times were found to be empirically related to surface chemistry, but not to fibrinogen adsorption or the Vroman effect. The Vroman effect for fibrinogen was found to be predictable in most cases given only single component fibrinogen adsorption data. Fibrinogen adsorption from buffer was found to increase with sulfur (sulfonate) content and decrease with nitrogen (amino) content, but contrary to suggestions in the literature was not a strong function of wettability. A multivariate approach to studying the interactions between blood, proteins and model surfaces proved to be a practical approach to a complex problem. It was demonstrated that univariate relationships are often inadequate in their ability to explain the interfacial behavior of blood and its elements, and in fact may lead to erroneous conclusions</p> / Doctor of Philosophy (PhD)

<p>This thesis presents an investigation into the effect of recycle on the dynamic controllability and operability of chemical processing systems. In particular, this investigation focuses on the effect that recycle has on the behaviour of individual chemical units when they are configured such that the product of the last unit is returned as a partial feed to the first unit. An experimental extractive distillation unit, which separates a mixture of methanol and acetone using a water solvent, was chosen for this study. Two aspects of this system were highlighted in this thesis. The first is that the system is subject to, and is to be controlled against, feed flow and compositional disturbances. The second aspect is that the system contains deadtime, and the analysis techniques must take this into account. There are currently several techniques for analyzing the dynamic behaviour of chemical systems (without resorting to exhaustive numerical simulations). It is demonstrated that these techniques contain severe limitations which preclude their use for a recycle system. Modifications presented in this thesis to these techniques allow for their successful application on a recycle system; furthermore, these modifications are not limited to recycle systems exclusively, but should be utilized on chemical engineering systems in general. In particular, these modifications are: (1) A methodology was formulated for determining interaction for regulatory control systems. To date, dynamic interaction techniques were constrained to servo systems; the technique presented in this thesis is the first to consider the regulatory interactions of the system. (2) Interaction techniques were modified to take into account deadtime, as all present techniques have the shortcoming of being invariant to deadtime. (3) A technique for simplifying recycle transfer functions was formulated. This was necessary because recycle transfer functions formed from their constituent subsystems cannot be inverted to the time domain or used to design controllers. These techniques were applied to the extraction unit and compared against simulations to demonstrate their effectiveness. The large effect that recycle has on the open and closed loop response of this system was also demonstrated, illustrating the importance of considering the response of the entire system, and not just its component parts.</p> / Doctor of Philosophy (PhD)

<p>Characterization of the Rushton turbine (RT) and pitched blade turbine (PBT) impellers was undertaken with two objectives: (1) development of impeller boundary conditions which can be generally applied to CFD simulations of stirred tank flows for predictive purposes, and (2) accurate characterization of the turbulence quantities k and ε for use in model development and verification. Since the flow field for the RT is well characterized in the literature, work on this impeller proceeded directly to a fundamental model of the discharge flow. Using this swirling radial jet (SRJ) model, it was possible to accurately predict the impeller boundary condition, including k and ε. The predicted boundary condition was used in a three dimensional CFD simulation of the entire flow field. The results of this simulation accurately reproduce mean velocity profiles, the decay of k and ε in the bulk of the tank, and details of the flow behind the baffles. Characterization of the PBT commenced with mean velocity and flow visualization studies of the impeller discharge zone for various geometries. It was found that there is a definite transition of the entire time averaged flow field from a single circulation loop, to two circulation loops, which occurs at a specific off bottom clearance. It was also shown that the boundary condition is affected by feedback from this transition in the flow field. The time varying characteristics of the flow field were examined in some detail. Various proposed methods of estimating ε were applied, and the results compared. It was found that autocorrelation function methods revealed useful physical information about the flow, particularly the existence of not only the dominant blade passage frequency (BPF), but also a second, low frequency component in the flow. The effects of these frequencies on the calculated value of ε were examined. It has been possible to characterize the boundary conditions for both the RT and the PBT impellers, using both theoretical and experimental methods; and to show that mathematical modelling and CFD simulations based on the k-ε model can yield accurate and useful information about the flow field.</p> / Doctor of Philosophy (PhD)

<p>The objectives of this work were to investigate the applicability of using Partial Least Squares (PLS) in three areas of chemical engineering: (1) for performance monitoring, (2) for inferential model development, and (3) in dynamic model identification. One further goal of this work was to gain understanding of the properties of the PLS method in the context of chemical engineering. Properties investigated included: (1) effects of scaling, (2) the objective function in the algorithm, and (3) the configuration of the X and Y data blocks. The first application of PLS was in the area of performance monitoring of a multivariate processes. A multivariate Statistical Process Control (SPC) procedure has been proposed for handling large numbers of process and quality variables. PLS is used to reduce the dimensionality of these large and highly correlated data sets down to a few latent variables which contain most of the information about the process behaviour under normal operating conditions. By compressing all the information on the process down to low dimensional spaces, and using simple plots of the data in these spaces, together with meaningful control limits, the essential idea and philosophy of Shewart's (1931) SPC methods have been preserved and extended to handle the large number of variables collected in most process industries today. PLS has also been shown to be a very powerful approach to inferential model building when large numbers of highly correlated measured variables are available. By retaining all the measurements without overfitting the data PLS is able to utilize all the information obtained from the process measurements. Further, PLS models are extremely robust to missing data and sensor failures, an important feature of any inferential control scheme. The importance of obtaining a representative reference data set, (i.e., one that represents the process relationships for the expected range of process conditions, and the structure of the process and control configuration under which the model is to be used), when using empirical models has also been demonstrated. The comparison of PLS, Ridge Regression (RR) and Least Squares identification of non-parsimonious (Finite Impulse Response (FIR) and Autoregressive exogeneous (ARX)) models for multiple input single output (MISO) systems shows that the biased estimates obtained from PLS and RR provide smoother estimates of the impulse and step weights. The difficulty in characterizing structured noise when using these non-parsimonious models is discussed. Finally, the use of PLS for identification of multiple input multiple output (MIMO) processes was investigated.</p> / Doctor of Philosophy (PhD)

<p>The modelling and control of a pilot plant packed bed catalytic reactor is investigated. The reactions considered are those of the hydrogenolysis of n-Butane over a nickel on silica gel catalyst. These involve several independent species and are highly exothermic. A multivariate transfer function model for process and a time series model for the disturbances are constructed through an iterative procedure of process identification, parameter estimation and diagnostic checking. Using exit concentrations inferred from reactor bed temperatures and inlet flow measurements, a multivariate stochastic controller, designed to minimize a quadratic function of the production rates of various intermediate products subject to constraints on the variances of the manipulated input flowrates is implemented on-line using a process minicomputer. The parameters of the inferential concentration equation are updated on-line using infrequent data from a process gas chromatograph. The performance of this feedback controller is found to be quite satisfactory.</p> / Master of Engineering (ME)

<p>The work described in this thesis can be broken down into two separate parts. The first deals with the mathematical and computer modelling of industrially important methods for the production of polyolefins, specifically: (a) the high pressure tubular reactor process, (b) the high pressure autoclave type reactor process, and (c) the low pressure Ziegler Natta reactor process using heterogeneous catalysts. The second part deals with the chemical modification of these polymers by free radical methods during the extrusion process. Free radical copolymerization of ethylene in high pressure reactors is considered. Kinetic mechanisms to describe the polymerization rate and polymer properties, including copolymer composition, molecular weight, branching frequencies, melt flow index and polymer density, have been proposed. The method of moments is used, in conjunction with pseudo kinetic rate constants to allow for copolymerization, to calculate the molecular weight averages. Based upon this kinetic scheme, a mathematical model has been derived and implemented as computer programs to simulate commercial tubular and autoclave type reactors. The model parameters were fit to steady state data from industrial reactors. The dynamic autoclave reactor model includes the two phase kinetics due to polymer-monomer solubilities and phase separation in the reaction mixture. Gel formation from crosslinking reactions is also analyzed. A mixing model is developed to represent the flow inside the reactor. In the simulation PID control equations are used to maintain operation at the unstable steady state. A sensitivity study is performed on the mixing model parameters and on some of the kinetic parameters. The model is compared to steady state temperature, initiator flow and conversion data from commercial reactors. Many polyolefins are produced using heterogeneous Ziegler Natta catalysts either in gas phase, bulk or slurry type reactors. These properties include broad molecular weight and copolymer composition distributions. A kinetic model has been derived, based upon multiple catalyst sites of differing reactivities, for the Ziegler-Natta copolymerization of olefins. The model predicts the rate of polymerization, the copolymer composition and the molecular weight distribution of the polymer produced as well as accounting for the observed broad copolymer composition and molecular weight distributions. This dissertation deals with the development of mathematical models to relate the molecular modifications, scission, branching, crosslinking and grafting. The models, based upon generally accepted kinetic mechanisms and certain assumptions about the nature of simultaneous scission and crosslinking, can predict the molecular weight averages, degrees of crosslinking, scission and grafting, and the sol/gel ratio. A mathematical formulation to describe simultaneous random scission and crosslinking has been presented in the literature but only solved by assuming that scission and crosslinking occur serially. For this work, an algorithm to numerically solve this equation was developed. The results of these calculations were compared to the classical solutions to this problem and to the experimental data gathered in this project. This model, although applied to polyolefin modification, has applications for other systems where simultaneous random scission and crosslinking of polymers is encountered. (Abstract shortened by UMI.)</p> / Doctor of Philosophy (PhD)

<p>This thesis investigates the design, analysis and implementation of feedback control systems. A new definition of robustness for continuous feedback control systems is advanced: the region of the joint allowable variation in the gain and the dead-time of the process. This region is constructed from the open loop system frequency response. The continuous time Proportional-Integral-Derivative (PID), Internal Model Control (IMC) and Linear Quadratic Optimal Control (LQOC) design procedures for a first order plus dead-time process are compared for Integral Square Error (ISE) performance when tuned for the same level of robustness. The order from highest to lowest performance is: PID, IMC, LQOC. The robustness analysis method is used to explain the adverse effects of dead-time compensation on robustness. The continuous ISE performance of the discrete LQOC, and State Deadbeat IMC controllers are compared as the size of the control interval is changed; which one is better is shown to depend on the choice of the control interval. A new modified LQOC controller and a new Extended Horizon controller are proposed. The robustness analysis developed for continuous systems is extended to discrete systems as the region of joint allowable variation in the process gain and quasi dead-time. This region is constructed from the discrete open loop frequency response. The state deadbeat IMC with an augmented filter, and modified LQOC controllers are shown to have better ISE performance than the LQOC controller when tuned for the same level of robustness. A procedure is presented for saturation correction in multivariable one-step optimal controllers. A simultaneous correction is included which adjusts the remaining control inputs if some inputs saturate. The procedure is applicable to all related controllers including IMC. Results on saturation correction for PID controllers are also presented.</p> / Doctor of Philosophy (PhD)

<p>Methods for molecular weight characterization of polyelectrolytes, poly(acrylamide-co-sodium acrylate) by viscometry and gel permeation chromatography have been developed based on the use of fractionated polyacrylamide with known average molecular weights. A set of Mark-Houwink equations for the copolymer with various acrylate content has been established. With the viscosity measurements, these equations can be used to calculate the molecular weights of the copolymer with known composition, and to estimate the acrylate content when the weight average molecular weight of the polymer is known. The universal calibration method has been proven valid for poly(acrylamide-co-sodium acrylate) using nonionic polyacrylamide standards. The correlation of elution volume shift with the content of ionic groups has been found, from which the molecular weight of the copolymer can be estimated by gel permeation chromatography. Heterogeneous free-radical polymerization of N-isopropylacrylamide (NIPAM) and NIPAM with N,N'-methylene bisacrylamide (BAM) has been conducted in aqueous media in the presence of low levels of sodium dodecyl sulfate (SDS). The influence of temperature, BAM level, SDS level, pH and ionic strength on the polymerization rate, particle size and size distribution, and particle concentration has been investigated. Studies on the formation mechanism of latex particles of poly(NIPAM) and poly(NIPAM/BAM) have been carried out by measuring the kinetics of polymerization and the micro-structure of the particles. A mechanism of homogeneous nucleation (dispersant-limited agglomeration) and dispersion polymerization has been proposed. The microstructure of the latex particles has been studied from the swelling extent of the particles at various conversion. It is found that the particles have lowest swelling at early stage and have higher swelling at later stage. This has been interpreted as that the cross-linking density changes along the radius of the particles. In addition to the heterogeneity within the particles, the density distribution among the particles has been observed from the comparison of the particle size distribution by dynamic light scattering and that by disk centrifuge. The interaction of NIPAM polymers with SDS has been investigated by conductometry and turbidimetry. It has been found that the critical micelle concentration (CMC) of SDS increases with the concentration of NIPAM polymers, and that the cloud point of aqueous solutions of NIPAM polymers is elevated in the presence of SDS.</p> / Doctor of Philosophy (PhD)

<p>The theory of self-tuning regulators and related topics in linear stochastic control theory have been examined. A unifying treatment of the theory of self-tuning regulators has been presented, and an attempt made to clarify the confusion surrounding certain aspects of these regulators. The notation is that of Box and Jenkins (1970).</p> <p>Self-tuning control of a steam jacketed stirred tank was successfully implemented. The experimental program was designed to illustrate points from the theory that have caused confusion in previous industrial and pilot plant implementations.</p> / Master of Engineering (ME)

<p>In this thesis, the copolymerization theory has been extended to account for the unique phenomenon of multiple initiation/propagation/termination cycles, experienced by copolymer chains in free radical copolymerization with bifunctional initiators. A comprehensive model for the most general free radical copolymerization scheme, involving simultaneous and competitive statistical and donor-acceptor propagation mechanisms, has been developed. The effects of multiple chain recombination during bifunctionally initiated batch free radical copolymerization, on copolymer microstructure and compositional drift, have been evaluated. As a result, the concepts of effective instantaneous copolymer composition, and number of segments per copolymer chain have been introduced as an addition to the copolymerization theory.</p> <p>The ability to control the segment composition of the multi-segment copolymers formed during bifunctionally initiated free radical copolymerization, has been the basis for the design and synthesis of four different high T₀ engineering copolymers. Styrene/α-methylstyrene (T₀-115°C), styrene/N-phenylmaleimide (T₀-215°C), and α-methylstyrene/N-phenylmaleimide copolymers (T₀-260°C), as well as styrene/α-methylstyrene/n-phenylmaleimide terpolymers (T₀-220ºC), have been synthesized in bulk and suspension, and fully characterized. The performance of these four copolymer systems in suspension polymerization with bifunctional initiators has been evaluated at pilot plant scale, to assess the technical feasibility for their commercial production.</p> <p>In addition, a novel route for the synthesis of suspension expandable polystyrene with bifunctional initiators, in a highly productive simultaneous polymerization/impregnation stage, has been developed and fully tested in pilot plant scale.</p> <p>From the comprehensive kinetic and micro-structure models proposed in this thesis, two computer simulation programs (BIPEN and BICOP), have been developed. Both programs, predicting the most important phenomena of the polymerization processes studied herein, have been used as a product design tool, throughout this thesis, with very good results.</p> / Doctor of Philosophy (PhD)