<p>This thesis has been written in three parts. Part I deals with the rheological response of dilute solutions of high molecular weight polyacrylamides at low shear rates. The non-Newtonian effects were found to be significant for polyacrylamides with number average molecular weights exceeding 10⁶. The molecular weight average-intrinsic viscosity relationship most widely used in literature was found to be valid when the intrinsic viscosity was measured at high shear rates where the polymer solutions approached Newtonian behaviour. A new relationship was developed relating the number average molecular weight to the intrinsic viscosity extrapolated to zero shear rate.</p> <p>Part II is an experimental investigation of the free-radical chain polymerization of acrylamide in water with potassium persulfate initiator. Conditions were such that the polymers produced had a number average molecular weight in excess of one million. Molecular weight averages were measured by viscometry, accounting for the non-Newtonian effects by the methods developed in Part I. Values for the transfer constants to the monomer and to the initiator were estimated at 25°C and 40°C and compared to the literature values.</p> <p>In Part III, a new method was developed to estimate the reactivity ratios from composition-conversion data, based on non-linear regression. Previously published experimental data for the copolymerization of acrylic acid and acrylamide were analysed by the new method, and the results compared to those reported by the original investigators. Composition-conversion data were collected for this copolymerization system at intermediate conversion levels and over a limited range of compositions. Values for the reactivity ratios at 40°C were obtained from these data by the new algorithm, and compared to the literature values.</p> / Doctor of Philosophy (PhD)

<p>A pilot scale non-adiabatic packed bed reactor has been designed and constructed to carry out the hydrogenolysis of n-butane over a nickel on silica gel catalyst. The complex reaction scheme is highly exothermic and results in steep radial gradients within the reactor. The apparatus was interfaced to a minicomputer. A sophisticated executive program was developed to assist with on-line studies of the apparatus. It provided measurements of axial and radial reactor temperatures and composition data for the reactor effluent. A third order, discrete time state space model of the reactor was postulated from mechanistic arguments and the parameters fitted to dynamic data. The model served as the basis of a derivation of a model reference adaptive controller designed to satisfy Lyapunov's second theorem of stability. The control algorithm included terms for proportional, integral and setpoint actions. The objective of the controller was to regulate the reactor's production rate of intermediate reaction species. The control algorithm was implemented experimentally and its performance discussed. Problems with the controller are examined and recommendations made.</p> / Doctor of Philosophy (PhD)

<p>From the analysis of a postulated loss-of-coolant and loss-of-emergency cooling accident in a typical horizontal pressure tube nuclear reactor, the need was established for a systematic investigation on the boiling heat transfer coefficient to be expected on a 12.7 cm diameter cylinder and a bundle of these cylinders under saturated and subcooled conditions. To accomplish this, a three-part experimental program was instigated.</p> <p>In the first part, a successful technique based on a quenching (transient) procedure and the use of a heat flux meter was developed to measure directly the local boiling heat flux density around a large horizontal cylinder (12.7 cm diameter) as a function of wall temperature under saturated and subcooled conditions. These experiments were complemented by pertinent theoretical analyses. Although the surface averaged critical heat flux densities are shown to be predicted very well by known correlations, these experiments and analyses do reveal that there is considerable variation of the local critical heat flux density (CHF) around the cylinder. It was shown by high speed motion pictures (1000 frames/s) and from quenching experiments on partially insulated cylinders that this variation results from the interaction of the upflowing liquid-vapour boundary layer with the boiling phenomena. These results for the partially insulated surface and for steam injection underneath a single copper cylinder are also presented.</p> <p>In the second portion of the experimental program, a five row by three column array of instrumented copper cylinders was quenched to determine the interaction effect on the boiling curve of the vapour-liquid mixture rising from tubes located below. This bundle was investigated under saturated and subcooled conditions. Results show that there is a significant critical heat flux increase over the lower part of the cylinder and no appreciable effect over the other parts. The critical heat flux increase is relatively independent of the amount of upflowing vapour.</p> <p>In the third part of the program, a steam-heated tube, 2.67 cm outer diameter, was used to study the steady-state boiling characteristics of heptane on single horizontal tubes. The average boiling heat flux for each of these steam-heated tubes in a four row by three column array was determined and compared with that observed when each was used alone. The results show a substantial increase in the critical heat flux value and essentially no increase in the nucleate boiling regime on any tube when the lower tubes were boiling. When all tubes were at their critical heat flux, the average CHF in the bundle was about 20% higher than the one obtained in a single tube experiment.</p> <p>It is recommended to initiate a study to calculate the pressure drop in a horizontal boiler. It is suggested that the apparent decrease in the overall average critical heat flux for a bundle of tubes, which has been reported in the literature for horizontal reboiler systems, may not be caused by vapour blanketing but may result from insufficient liquid being drawn into the bundle, by natural convection or the thermosyphon effect, to satisfy that required to generate the vapour if all tubes are operating at the critical heat flux. Some recommendations for future work relating to this problem are presented.</p> / Doctor of Philosophy (PhD)

<p>The hydrogenolysis of propane, n-hexane, 2,3-dime-thylbutane, and 2,2-dimethylbutane was studied over supported catalysts of ruthenium, nickel, cobalt, and iron, some containing structural promoters. The catalysts were prepared by impregnation of several supports, and were characterized using mainly atomic absorption and chemisorption techniques. A differential reactor system was used, consisting of a fixed-bed reactor and an external recycle pump. The data from propane hydrogenolysis were fitted to a power rate equation and a selectivity equation. The results are in general agreement with those of similar experiments with ethane. The products from the hydrogenolysis of the hexanes were fitted to selectivity equations that were based on reaction networks derived for reversible adsorption-desorption of the hydrocarbons and irreversible rupture of the carbon-carbon bonds of the surface species. The product distributions were measured over a wide range of conversion (10 to 80%), but in most cases only at one temperature. In the sequence -ruthenium, nickel, cobalt, iron - the distribution shifts toward smaller hydrocarbons. Ruthenium tends to split the carbon-carbon bonds in a straight chain with equal probability, while nickel preferably splits terminal bonds. Structural promoters increase the activity of the nickel and cobalt catalysts, and stabilize the iron catalysts. Because of this the catalysts on low area supports had activities very similar to those on high area supports; the amount of metal, however, was higher on the low area supports.</p> / Doctor of Philosophy (PhD)

<p>Linear Programming (LP) has a wide range of industrial applications, including closed-loop systems such as real-time optimization and the steady-state economic optimization at each execution of Model Predictive Controllers. This thesis presents new metrics for monitoring the performance of linear closed-loop real-time optimization systems, as well as new methods for improving their performance when necessary. A novel diagnostic method for ranking parameter importance with respect to the objective function is also presented.</p> <p>Many standard methods are available for estimating the effects of parameter uncertainty on the objective function without a basis change, and more powerful existing methods require enumeration or sampling. This work introduces new sensitivity methods in LP problems with uncertain coefficients that can be correlated, appear in equality and inequality constraints, and have uncertainties with large enough magnitudes to lead to basis changes.</p> <p>The new monitoring approach measures the uncertainty effect as the range between the maximum and minimum profit in the plant under closed-loop optimization, termed the Profit Gap, and both its maximum and expected values can be determined.</p> <p>If the monitoring indicates a substantial Profit Gap could exist, the improvement step designs experiments to reduce parametric uncertainty. The unique experimental design maximizes the total profit during and after the experiment to the end of a production run.</p> <p>Both the monitoring and improvement methods involve the solution of bilevel optimization problems, which include complementarity constraints. Results of application to a closed-loop gasoline-blending problem demonstrate the power of the methods. The studies include typical uncertainties and measurement noise and show the economic benefits possible through the application of real-time monitoring and improvement.</p> / Doctor of Philosophy (PhD)

<p>The main purpose of this research was to find, from a simulation of an existing reactor under rather heavy constraints, how upon removal of these limitation the same reactor could be operated more economically and, efficiently. An accurate, simplified model of the process was developed and used in the design of single and double bed reactors with very consistent results. A better understanding of the process variables, reactions and constraints was obtained by extrapolation of the model within reasonable bounds.</p> <p>The usefulness of a multivariate search technique applied to a relatively complicated process was proven. Such techniques as Pontryagin's Maximum Principle l and dynamic programming2 become unwieldy for processes involving many state variables. An engineer can quickly grasp the ideas of multi-variable search methods whereas it is difficult to understand the above more elaborate techniques without much study. This report may be particularly useful to plant-process engineers who seek a practical optimum-seeking method.</p> / Master of Engineering (MEngr)

<p>The dependence of the critical shear stress and the critical recoverable shear strain on molecular weight has been studied . Polystyrene of narrow and broad molecular weight distribution, low and high density polyethylene and polypropylene were the polymers used in this study. The behaviour of the melts at the onset of instabiIity was explained by using Graessley's correlation between "true" and Rouse shear compliance in Hooke's law. It was found that the critical stress increases linearly as the weight average molecular weight decreases, increases slightly with temperature and is nearly independent of the distribution of molecular weights . The critical recoverable shear, calculated using Tanner's dieswell theory was found proportional to the factor M₂M₂₊₁/M_w², the proportionality factor being about 2.5 for linear polymers.</p> / Master of Engineering (ME)

<p>The heat transfer problem of polymer melts flowing through narrow channels and tubes has been studied. Four types of flow with constant temperature boundary conditions were examined:</p> <p>(i) drag (or Couette) flow beD~een parallel plates,</p> <p>(ii) Poiseuille flow between parallel plates,</p> <p>(iii) Poiseuille flow through a tube with circular cross-section, and</p> <p>(iv) drag flow between converging plates.</p> <p>In each case, the equations of conservation of mass, momentum and energy were solved simultaneously by the implicit finite difference method. A power-law temperature-dependent viscosity model was used and viscous dissipation was taken into account. Velocity and temperature profiles, pressure distributions, bulk temperatures and local Nusselt numbers have been calculated and are presented as a function of the axial distance along the channel. Results obtained by using the power-law temperature dependent viscosity model were also compared with the power-law temperature-independent viscosity model and the Newtonian, constant viscosity model results.</p> / Master of Engineering (MEngr)

<p>[Missing pg 73]</p> / Master of Engineering (ME)

Master of Engineering (ME)