This thesis is concerned with the design, operation and control of pressure swing adsorption (PSA) systems, employing state of the art system engineering tools. A detailed mathematical model is developed which captures the hydrodynamic, mass transfer and equilibrium effects in detail to represent the real PSA operation. The first detailed case study presented in this work deals with the design of an explicit/multi-parametric model predictive controller for the operation of a PSA system comprising four adsorbent beds undergoing nine process steps, separating 70 % H2, 30 % CH4 mixture into high purity hydrogen. The key controller objective is to fast track H2 purity to a set point value of 99.99 %, manipulating time duration of the adsorption step, under the effect of process disturbances. To perform the task, a rigorous and systematic framework is employed comprising four main steps of model development, system identification, the mp-MPC formulation, and in-silico closed loop validation, respectively. Detailed comparison studies of the derived explicit MPC controller are also performed with the conventional PID controllers, for a multitude of disturbance scenarios. Following the controller design, a detailed design and control optimization study is presented which incorporates the design, operational and control aspects of PSA operation simultaneously, with the objective of improving real time operability. This is in complete contrast to the traditional approach for the design of process systems, which employs a two step sequential method of first design and then control. A systematic and rigorous methodology is employed towards this purpose and is applied to a two-bed, six-step PSA system represented by a rigorous mathematical model, where the key optimization objective is to maximize the expected H2 recovery while achieving a closed loop product H2 purity of 99.99 %, for separating 70 % H2, 30 % CH4 feed. Furthermore, two detailed comparative studies are also conducted. In the first study, the optimal design and control configuration obtained from the simultaneous and sequential approaches are compared in detail. In the second study, an mp-MPC controller is designed to investigate any further improvements in the closed loop response of the optimal PSA system. The final area of research work is related to the development of an industrial scale, integrated PSA-membrane separation system. Here, the key objective is to enhance the overall recovery of "fuel cell ready" 99.99 % pure hydrogen, produced from the steam methane reforming route, where PSA is usually employed as the purification system. In the first stage, the stand-alone PSA and membrane configurations are optimized performing dynamic simulations on the mathematical model. During this procedure, both upstream and downstream membrane configuration are investigated in detail. For the hybrid configuration, membrane area and PSA cycle time are chosen as the key design parameters. Furthermore, life cycle analysis studies are performed on the hybrid system to evaluate its environmental impact in comparison to the stand-alone PSA system.

Thermodynamic models that can accurately describe and predict phase equilibrium are essential for process design and simulation. For systems containing associating fluids, where there are strong intermolecular interactions, this is a very challenging task. The focus of this work is on modelling systems containing hydrogen fluoride (HF) and repJ,acement refrigerants, both of which can present strong- hydrogen bonding and polar interactions. In addition, in developing molecular models for refrigerant molecules, their non-spherical shape needs to be addressed. The statistical associating fluid theory with potentials of variable range (SAFT-VR) is especially well-suited for this task. In SAFTVR hydrogen bonding interactions are taken into account explicitly by introducing a number of short range attractive sites, which mediate this interaction, while the non-sphericity of the molecule is treated via a model of a chain of tangentially bonded segments. Here the non-sphericity of the molecule is obtained from ab initio quantum mechanical calculations; this is a major new contribution of this work. Novel molecular models for HF and a number of refrigerants are developed. These models are then used to treat mixtures where hydrogen bonding is examined in particular detail. The SAFT-VR approach is flexible enough to reflect the fact that different types of hydrogen bond can fonn between two species. Two approaches for the estimation of mixture parameters are presented. The first one relies on coexistence composition data. However, often these compositions are obtained from data reduction, which depends on the thennodynamic description assumed, as opposed to experimental sampling. Hence, an inherent bias is incorporated into the intennolecular potential model. To overcome this, a mathematical model of experimental setup with integrated parameter estimation is developed, allowing self-consistent intennolecular parameters to be obtained. An estimation of experimental uncertainty is incorporated into the fonnulation ofthe parameter estimation -problem, enabling an evaluation of the statistical significance of the parameters. The systematic approach to model development established in this work is general and can be applied to other systems and advanced equations of state.

No description available.

Over the last few decades, the availability of inorganic membranes which can withstand high temperatures and harsh chemical environments has resulted in a wide range of opportunities for the application of membranes in chemical reactions. In particular, the combination of membrane separation and catalytic reaction in a single operating unit is an attractive way to increase conversions, to achieve better yields and to make more efficient use of natural resources in many reactions. In this work, a highly compact multifunctional Pd and Pd-Ag/alumina hollow fibre membrane reactor (HFMR) have been developed and applied to catalytic chemical reactions. The developed HFMR consists of a thin and defect free Pd-based membrane coated onto the outer surface of an alumina hollow fibre substrate with a unique asymmetric pore structure, i.e. a sponge-like outer layer and a finger-like inner layer where catalyst is deposited. In one study, a Pd-Ag layer was coated onto the outer surface of the substrate followed by deposition of sub-micron sized Pt(0.5wt.%)/γ-alumina catalysts into the finger-like voids of the substrates. This design achieved propane conversion as high as 42 % at the initial stage of the reaction at 723 K and space-time yields (STY) of the HFMR were approximately 60 times higher than that of a fixed bed reactor (FBR). In order to further increase catalytic surface area in the reaction zone, a sol-gel method was used to deposit Pt(1 wt.%)/SBA-15 catalysts into the finger-like voids of a substrate to develop a Pd/alumina HFMR. Benefiting from this novel design, the functionalized alumina hollow fibre substrates with surface area/volume values of up to 1918.4 m2/m3 possess a specific surface area of about 31.8 m2/g for catalysts. It was observed that in comparison with a conventional FBR, greater propene selectivity and propene yield was achieved by using the HFMR for propane dehydrogenation. The generic advantages of the design of these compact HFMR systems can be applied to further applications such as the water-gas shift reaction, which was also carried out in this study.

No description available.

No description available.

Dc(XJsition of coke on FCC catalyst received considerable attention in Catalysis field~little '{Irk has been reported on non-oxidative regeneration especially hyd~ogen at 300°C and high pressure of 20 bar for 8hrs over coked catalysts followed by extraction. The deficient h)'drogen type cracked catalyst was treated for 8hrs at different temperatures of25, 110 and 300'C. Coked zeolite was left under hydrogen pressure overnight to increase the adsorption time of hydrogen on zeolite surface and pores. Cyclic regenerations were applied on heavily coked catalysts to enhance the. volatility of coke precursors at longer time. Dcconvoluted peaks of TGA-DTG were showing several volatile peaks at certain dissociation temperatures indicating the type of coke. There was correlation between peaks of adsorbed materials of coke precursors in TGA and TPD. Hydrogen adsorbed above 400'e in TPD spectra i.mplies that high temperature hydrogen diffused through catalyst matrix nnd caused significant reduction and morphological changes of coke species. FTIR ptnks showed the intensity of the coke profiles appeared in coked and diminished in regenerated ones. The acidic and metallic sites of the deactivated catalyst were restored lIccording to their functional groups in FTIR spectra. Continuous regeneration of coked 5nmples wns showing gradual loss of coke percentage weight. Toluene and more polar wh-cnt like pyridine leached the hydrocarbon cOffiFOunds adsorbed on the surface of the . zeOlite before or after hydrocracking. The resulted extract was sent for GC-Ms to identify the kind of hydrocarbons products restored. The activity of FCC catalyst was restored by more than 60% due to the removal of polycyclic type of coke depending on the feed type. The regenerated catalyst can be used in catalytic reactions with high activity and selectivity for light products.

No description available.

No description available.

In the recent years, the legislation related to the safety of industrial processes has become more rigorous and demanding than in the past. The design of inherently safe processes is currently' viewed as a necessary target of the engineering discipline (Gupta and Ed~ards, 2002).. Alkylpyridines and their N-oxides are being used widely as intermediates and final products in the Phannaceutical Industry. Under typical industrial conditions the N-oxidation reaction is performed catalytically, in the liquid phaSe in a semi-batch mode. Hydrogen peroxide is used as the oxidising agent. It is dosed over the alkylpyridine-catalyst mixture over a period of time. The desired N-oxidation is accompanied by one only additional reaction, the lpldesired decomposition ofhydrogen peroxide. The catalytic N-oxidation of selected picolines, lutidines and collidines were performed in an HEL SIMULAR® lL calorimeter at conditions very close to the industrial'ones, following a certain procedure. The catalytic decomposition of hydrogen peroxide alone, under the same conditions was also studied. A sensitive mass flow-meter MKS 1179A Mass-Flo® was used for the continuous measurement of the oxygen produced by the decomposition of hydrogen peroxide, thus .allowing the split of the power of the two reactions. The heat of reaction has been evaluated for all the selected N-oxidation and hydrogen peroxide decomposition reactions. The concentration of the a1kylpyridines and their N-oxides was measured using HPLC. These measurements confirmed that the hydrogen peroxide decompositio~ was the only parallel reaction. Integration of the thermal power evolution curves also provided a continuous measure of concentration.. These curves were found to be in good agreement with the HPLC measurements. The reactions were studied using heat-flow and power compensation calorimetry at subcooled conditions. Their study shows that their N-oxidation follows Langmuir type kinetics. The paraIlel decomposition ofhydrogen peroxide also follows the same type ofkinetics. Their kinetic study was based on the kinetic model of Sempere et al, 1998, whi~h was improved and refined. The model of Sempere et al., 1998 on a-picoline assumes paths of reaction, in which, the employed· catalyst, phosphotungstic acid, does not participate. However, as reported in our previous work (papadaki et al., 2002a), no reaction takes place in the absence of phosphotungstic acid. The kinetic model proposed by Sempere et al. 1998 has thus been modified to account for those paths. The modified model is, as before, based on the assumption of very fast always in equilibrium reactions or interactions for the formation of intermediates which subsequently react in a number of ways to give the final products. All paths of the modified model are catalytic. The new kinetic model has been tested and found to be,:in good agreement with experimental data. A number of runaway scenarios of the excess of hydrogen peroxide used during the N-oxidation of alkylpyridines, under closed and open conditions, were examined. It was found that, in most cases, if the volume of the liquid hydrogen peroxide solution occupies more than 10% ofthe total volume ofa closed system (e.g. reactor and vent line between reactor and blockage), the production of gases raises the pressure so quickly that evaporation is completely suppressed. Higher that 70% filling levels result in complete expansion of the liquid~The MTSR(t) (the inaximum temperature attained by the synthesis reaction) of the system falls rapidly if th~ normal process temperature is high, but if a runaway.occurs exactly at the end of dosing, MTST (the maximum temperature reached during the synthesis reaction) will be very high and secondary decompositions will rapidly develop. The results of this study are currently being used to critically assess the current approaches and to further the study ofinherently safer designs.