Uncontrolled segregation in particulate mixtures has long been considered as an annoying, and costly, feature encountered in many materials handling operations and although the onset is not clear, many believe it to be driven by the differences in particulate physical properties. An increasing number of usefully scaled laboratory and computer simulation investigations are being carried, particularly by the physics community, to help our understanding of this phenomenon. Physicists at the University of Nottingham have identified that through careful control of frequency and acceleration during vertical vibration, different types of particles can be positioned and/or segregated in a small rectangular cell. An extension of this work resulted in the design of a new small scale batch separator capable of recovering at least one separated particle layer in a different chamber. This work has explored the scale up of the small particle separator to operate in a semi-continuous mode. Since complete experimental know how of particle segregation phenomena is still deficient an empirical design strategy was used. This scaled up particle separator was driven by a pneumatically powered vertical vibration bench in which dry, non-cohesive particulate mixtures of varying densities and sizes (<1000µm) were vertically vibrated under different conditions to assess their separation behaviours. Experiments with regular (e.g. glass and bronze) and irregular shaped particle mixtures (e.g. comminuted glass and bronze) showed that lower magnitudes of vertical vibration frequency (30±10%), dimensionless acceleration (3±10%), particle bed heights (20 and 40mm in majority of the investigated cases) and partition gap sizes (5 and 10mm) were important for separation. Finally, the technique was employed to separate various industrially relevant particle mixtures (shredded printed circuit boards, iridium and aluminium oxide and shredded personal computer wires). Two-dimensional Discrete Element Modelling (DEM) with interstitial fluid interactions simulated with a maximum of 1000 virtual glass and bronze particles showed some important aspects of particle segregation such as; layered particle separation, high density particles ending on top and bottom of the particle bed, convection currents, particle bed tilting and partitioned particle separation. The application of Positron Emission Particle Tracking (PEPT) to glass, bronze, ilmenite and sand particles showed distinct trajectory maps in three dimension (X,Y and Z) with varying particle speeds in the vertically vibrated particle mixtures. The low density particles were mostly observed to move in the middle while the high density particles patrolled in the outer periphery of the separation cell. These distinct particle motions suggested that convection currents played an important role in controlling segregation. Furthermore, the application of a smoke blanket visualization technique showed the existence of air convection currents on top of the vertically vibrating particle mixtures. The experiments on the scaled up semi-continuous particle separator confirmed what was identified previously in that good particle separation could be achieved through careful control of the frequency and acceleration during vertical vibration. This information lays the foundations for a new breed of low cost, dry separator for fine particulate mixtures. Key Words: Vertical vibration, particle separator, fine particle mixtures, dry separation, PEPT, DEM, smoke visualizations.

Premature yeast flocculation (PYF) is a sporadic problem encountered during industrial brewing fermentations. Current hypothesis states that factors, thought to arise from fungal infection of the barley in the field and/or the malt in the maltings cause yeast to flocculate prematurely and/or heavily before the depletion of the sugars in the wort. This results in poorly attenuated worts, with higher residual extract and lower ABV, flavor abnormalities (i.e. diacetyl, SO2), lower carbonation levels, disruption of process cycle times and potential issues with the re-use of the yeast in subsequent fermentations. Consequently, PYF generates significant financial and logistical problems both to the brewer and the maltster. In the current study a small-scale fermentation assay was developed and optimized to predict the PYF potential of malts, as well as to investigate the importance of yeast strain in the incidence and severity of the phenomenon. Furthermore, the impacts of the PYF factor(s) (i.e. arabinoxylans, antimicrobial peptides) on yeast fermentation performance and metabolite uptake were also studied, whilst the Biolog detection system was investigated as a potential rapid tool which to detect PYF. The results obtained suggested that our in-house assay can be successfully used to predict the PYF potential of malts 69 or 40 h post-pitching depending upon the yeast strain used. Whilst ale yeasts were not found susceptible to PYF, lager yeasts exhibited different degrees of susceptibility even to the same PYF factor(s). More specifically, the more flocculent lager yeast SMA was found to be more susceptible than the medium flocculent lager yeast W34/70. However, interestingly, the fermentation performance of a PYF+ wort could be significantly improved by using a non-flocculent and relatively insensitive to PYF lager yeast. It was also shown that worts with lower amount of glucose and maltose could be responsible for poor fermentation profiles and/or heavy PYF as well as elevated residual sugars and lower fermentability. The observation that linoleic acid (6 mg.l-1) exacerbated PYF (P = 0.047) and made its detection more rapid was found to be contrary to the “titration hypothesis” (Axcell et al., 2000) which hypothesized that the addition of fatty acids might “titrate” out antimicrobial peptides so that they can no longer bind to the yeast cells. High gravity fermentations with worts inducing PYF did not have a significant effect (P > 0.05) on yeast physiological characteristics or fermentation performance suggesting that the PYF+ sample used in this study was inducing PYF though the ‘bridging’ polysaccharide mechanism rather than through the antimicrobial peptides. The Biolog system can be used for the metabolic characterization of different flocculence lager yeasts incubated in different fermentation media, whilst wort composition had a significant effect in redox reduction reactions.

Office Space Allocation (OSA) is the task of efficient usage of spatial resources of an organisation. A common goal in a typical OSA problem is to minimise the wastage of space either by limiting the overuse or underuse of the facilities. The problem also contains a myriad of hard and soft constraints based on the preferences of respective organisations. In this thesis, the OSA variant usually encountered in academic institutions is investigated. Previous research in this area is rather sparse. This thesis provides a definition, extension, and literature review for the problem as well as a new parametrised data instance generator. In this thesis, two main algorithmic approaches for tackling the OSA are proposed: The first one is integer linear programming. Based on the definition of several constraints and some additional variables, two different mathematical models are proposed. These two models are not strictly alternatives to each other. While one of them provides more performance for the types of instances it is applicable, it lacks generality. The other approach provides less performance; however, it is easier to apply this model to different OSA problems. The second algorithmic approach is based on metaheuristics. A three step process in heuristic development is followed. In the first step, general local search techniques (descent methods, threshold acceptance, simulated annealing, great deluge) traverse within the neighbourhood via random relocation and swap moves. The second step of heuristic development aims to investigate large sections of the whole neighbourhood greedily via very fast cost calculation, cost update, and search for best move procedures within an evolutionary local search framework. The final step involves refinements and hybridisation of best performing (in terms of solution quality) mathematical programming and meta-heuristic techniques developed in prior steps. This thesis aims to be one of the pioneering works in the research area of OSA. The major contributions are: the analysis of the problem, a new parametrised data instance generator, mathematical programming models, and meta-heuristic approaches in order to extend the state-of-the art in this area.

Process improvement has been identified as a central topic of operations management, being relevant to the different functional areas and assisting in providing the benefits operations management aims to realise. While extensive research has been conducted on specific process improvement methodologies, high resource requirement of specific process improvement methodologies make them inappropriate for many Small and Medium Sized Enterprises (SMEs). Compared to specific improvement methodologies, organisational learning was identified as an appropriate theoretical perspective from which to analyse process improvement activities within SMEs, leading to the presentation of the following three research questions: How do engineering-oriented SMEs undertake process improvement? What is the applicability of the three models of organisational learning within engineering-oriented SMEs? How does organisational learning contribute to understanding of process improvement within engineering-oriented SMEs? The research questions were addressed through in-depth, interpretive, interview based case studies with 14 Engineering Oriented SMEs. The six exploratory cases studies enabled the identification of specific process improvement practices that related isolated problems or opportunities with organisational level changes. These activities appeared to require management to implement formalised operational processes to ensure changes were captured within operational procedures and subsequently used by operational staff. Management support and culture then appeared to affect the ability of process improvement practices to provide firm level benefits to the case companies. Without directions by management or acceptance by operational staff, efforts directed towards process improvement tended to have limited impact on the benefits companies were able to realise from process improvement. Findings were then analysed from three conceptualisations of organisational learning identified within operations management literature. This provided theoretically underpinned insight to the exploration of process improvement, emphasising the importance of experience, involvement with external parties and the multi-level nature of organisational culture. Following the analysis of the exploratory phase, the findings were confirmed within 6 additional engineering-oriented SMEs (2 were excluded). The confirmatory case companies allowed the further exploration of the relationships between the emergent themes in order for the third research question to be addressed. Organisational learning provided justification for the interaction and bidirectional relationship between process improvement and culture. Organisational learning also provided justification for the important role of management, in relation to interpreting the operating environment and adapting how they provided resources to process improvement. The research thus contributes to operations management theory, by building upon organisational learning theory, in terms of how process improvement is conceptualised, factors affecting the benefits realised from process improvement and the importance of management to provide resources and direction to process improvement activities. Within all the case companies, this involved both providing sufficient resources in terms of training and time to engage in process improvement, but also selecting work that provided firms with sufficient process improvement opportunities. By effectively engaging in process improvement, firms appeared better equipped to compete against larger firms and low cost economies.

The transport and separation of oil and water is a vital process to the oil and chemical industries. Fluids exiting from oil wells usually consist of gas, oil and water and these three phases need to be transported and separated before they can be processed further. Operation of the primary separators has often proved to be problematic due to the change in composition of the fluids as the well matures, often accompanied by the build up of sand or asphaltenes. These vessels are very expensive to install so there is motivation to improve their design and performance. One major factor affecting separator performance is the phase distribution of the inlet flow, as reflected in the flow pattern and droplet size. In this work, flow pattern boundaries and drop sizes of liquid-liquid dispersions were measured for vertical and horizontal flow of a kerosene and water mixture in a 0.063m tube. Drop size was investigated by using two different laser optical techniques. A laser backscatter technique was employed for concentrated dispersions and a diffraction technique was used at low concentrations. In order to develop a greater understanding of separator performance, a 1/5th-scale model was constructed of diameter 0.6m and length 205m. Residence Time Distributions were obtained for a range of different internal configurations and flow rates using a colorimetric tracer technique. Flow rates of 1.5-4 kg/s oil and 1-4 kg/s water were used and the vessel was equipped with a perforated flow-spreading baffle at the inlet and an overflow weir. Experiments were performed with no internals and with dip or side baffles. The side baffles acted to create quiescent zones within the vessel while the dip baffle caused a local acceleration of both phases. These situations are similar to those that can be caused by blocked internals or existing baffling or structured packing within field separators. A Residence Time Distribution model of a primary separator, the Alternative Path Model, was developed using transfer functions. This model has the ability to reproduce features of the experimental data by representing the flow as a series of continuous stirred tanks in series or in parallel. The model was used to develop parameters that could be used to obtain information about the performance of the separator. This model was also applied to Residence Time Distribution data obtained from field separators by BP Exploration, to relate features of the pilot scale separator to the field vessels.

The overall aim of this research project was to expand the understanding of the deposition of titanium and the nickel-based superalloy Inconel 718 by spray deposition methods. The spray processes employed were cold spraying and high velocity oxy-fuel (HVOF) thermal spraying. The first part of the work was undertaken to expand the understanding of the deposition of titanium by cold spraying; the HVOF process is unsuitable for Ti because of the metal's high reactivity. The deposits were produced from commercially pure titanium using cold spray equipment designed in the University. Using helium gas, the effects of different powder particle size ranges, types of substrate, substrate preparation methods, and spray parameter conditions on deposit formation were investigated. Using a simple one-dimensional model of compressible gas flow and particle acceleration, particle velocity distributions were calculated to aid interpretation of experimental data. Results show that titanium can be successfully cold sprayed onto substrates of Ti6AI4V and mild steel, with the critical velocity for deposition of this powder type of approximately 690 m s-1. The level of porosity was generally in the range of 13-23% and the adhesive bond strength was dependent on surface preparation but independent of gas pressure with values ranging from 22 MPa to 10 MPa for ground and grit blasted substrates respectively. This compares with a value of around 80 MPa which is typical for well adhered HVOF sprayed coatings. The second part of the study was concerned with comparing the deposition of Inconel 718 by cold spraying and HVOF thermal spraying; the latter employed a JP5000 liquid fuel gun. A Tecnar DPV-2000 instrument was used to systematically investigate the effect of changes in spray parameters (spraying stand-off distance, oxygen/fuel ratio, total mass flow rate, combustion pressure), on particle velocity and temperature during HVOF spraying. It was found that generally the particle velocity was more strongly affected by the stand-off distance and combustion pressure of the spraying gun whereas the particle temperature was mostly influenced by the particle size and combustion pressure. The microstructures of coatings sprayed under 4 different well controlled conditions were investigated and changes in the morphology of splats and partially melted particles in the coating were related to the particle temperature and velocity at impact. The HVOF had high bond strength and low oxygen level of typically 0.45 wt% (corresponding to an oxide content of less than 1.6 wt.%). By contrast, in the cold sprayed coatings, the bonding was considerably low (-14 MPa), independently from the process conditions. It was found that the process parameter that mainly affected the properties of the cold sprayed deposits was the gas pressure. More specifically, the microhardness of the coatings increased with the pressure whereas the relative porosity decreased.

The aim of the project was to understand the science behind rapid cooling of chocolate used in the Frozen Cone® process. Differential scanning calorimetry was used to study the effect of slow and rapid cooling on tempered chocolate. On rapid cooling, lower melting polymorphs of cocoa butter were generated. Upon heating these recrystallised into the more stable Form V. Results were confirmed by similar observations with tempered chocolate fats. A hypothesis was formed whereby upon rapid cooling, lower melting polymorphs nucleate and grow at the expense of Form V nuclei produced during tempering. Upon subsequent warming, these polymorphs melt and recrystallise into Form V. Rapid cooling on untempered chocolate did not show any recrystallisation during warming; proving that tempering is required for the formation of Form V crystals in the final matrix. These results were confirmed by temperature-controlled X-ray diffraction on cocoa butter and chocolate fats. The polymorph generated upon rapid cooling was identified as Form I. This co-existed and eventually transformed to Form II and Form V upon warming. X-ray results showed that following rapid cooling, Form V crystals created during tempering did not grow until above 5 °C. Direct contact cooling at different temperatures was carried out to mimic the Frozen Cone® process. It was found that above -15 °C, the adhesion of the sample to the holder increases and seems to be correlated to the presence of Form II. These results suggest that the molecular structure and adhesive property of the polymorphs formed at specific temperatures are important for the release of chocolate. Stepscan differential scanning calorimetry was used to separate the simultaneous melting and recrystallisation events occurring in chocolate following slow and rapid cooling, by deconvoluting the total heat flow into reversing and non-reversing components. The general applicability and limitations of Stepscan DSC are also discussed.

Recognising the difficulties encountered in measuring the adsorption of hydrogen at high pressure, a reliable volumetric differential pressure method of high accuracy and good repeatability has been developed for measurement up to ca 100 bar. The apparatus used has two identical limbs, a sample and a blank limb, between which a high accuracy differential pressure cell measures changes in pressure. By simultaneously expanding the two limbs and closely controlling the temperature of the entire system, many of the errors due to expansion of the gas can be avoided. In addition, helium blank measurements are used as a base line correction, which substantially reduces the effects caused by the rapid expansion of gas through a small port. Using this method, the hydrogen storage capacities of relatively small samples (1.0-2.5 g) of a selection of carbon materials have been accurately measured to a conservative limit of detection of 0.05 wt% and an accuracy of +/-0.02 wt%. The accuracy of the apparatus has been proven using lanthanide nickel (LaNi5), which has a known hydrogen storage capacity of 1.5 wt%, as a standard. The method has also been developed in order to analyse samples at elevated temperatures of up to 270 C. This has been demonstrated using lithium nitride (Li3N) compounds. The carbon materials studied include a series of activated carbons, carbon nanofibres (CNF) and carbon nanotubes (CNT). The activated carbons have displayed almost instantaneous hydrogen uptake independent of the degas method used, which indicates that sorption occurs via a physisorption mechanism. The series of powdered activated carbons have displayed direct correlation between the BET surface area and the hydrogen sorption capacity. The largest hydrogen sorption capacity observed for activated carbons was for a chemically activated carbon with a surface area of 3100 m2 g-1, achieving an uptake of 0.6 wt%. The preparation of CNF, grown from ethylene over mixed copper, iron and nickel alloy catalysts, has been extensively investigated. Control of the parameters of preparation has allowed the formation of CNF with surface areas of 10 - 500 m2 g-1, diameters of 100 - 1000 nm, lengths of 1-10s nm, gas conversions of 0-90 % and the formation of herringbone and platelet CNF structures. The CNF studied have been observed to be capable of adsorbing a maximum of 0.5 wt% hydrogen at 100 bar and ambient temperature. Only one of the materials studied was observed to break by a significant amount the trend of surface area vs hydrogen sorption capacity, observed for the activated carbons. This was a single-walled nanotube (SWNT) sample which achieved ca 1.6 wt% after slow carbon dioxide activation at low temperature. This larger sorption is hypothesised to result from the hydrogen slowly diffusing into the SWNT through defects in the structure and between the graphite planes in the CNF.

This research studies Mass Customization as an operations strategy and model. Opinions differ over whether MC should be a label for a specific business model in which customers select from pre-engineered product options, or whether it should be interpreted as a performance goal that has wider relevance. In this research it is viewed as the latter and a manufacturing enterprise is considered to be a mass customizer if it gives its customers the opportunity to have a product any time they want it, anywhere they want it, any way they want it and in any volume they want it, and at the same time brings the benefits that are associated with mass operations, in particular those of price and quality. In the literature MC is not one operations strategy but a family of sub-strategies and there are several classification schemes, most of which delineate the sub-strategies by the point along the value chain that customization takes place. Other than for one scheme for which correlations between technologies and MC types has been sought by means of a survey, no progress has been made in developing operations configurations models. Through the study of primary and secondary case studies several classification schemes are appraised and a new framework of five fundamental operations Modes is developed. The Modes are the kernel of a theory of MC, with the other elements being: - A model for Mode selection that uses four factors to determine when a Mode is suitable; - Indicative models of the information infrastructures of two Modes that demonstrate the Modes to be different and that they can be a foundation for configurations models; - A set of product customizable attributes that reveals the multifaceted nature of customization and extends the terminology of customization; - The delta Value concept that links the motivation for customizing attributes to differences between customers. A theory of MC is proposed, which postulates: - An MC strategy is relevant when there are differences across customers in how they value the configurations of customizable attributes; - There are five operational sub-strategies of MC; - The choice of sub-strategy for an enterprise is contingent on its organisation and its business environment. One of the five modes, Catalogue MC, is the Mode that is commonly associated with MC. It is the Mode in which all product variants are fully engineered before being ordered. A diverse set of order fulfilment models of relevance to this Mode are reviewed and organised into four types: fulfilment from stock; fulfilment from a single decoupling point; fulfilment from several decoupling points; and fulfilment from a floating decoupling point. The term floating decoupling point is coined to describe systems that can allocate a product to a customer wherever the product lies, whether it be a finished product in stock, a part processed product or a product that does not yet exist but is in the production plan. In the automotive sector this system has been called Virtual-Build-to-Order (VBTO) and in this research the generic characteristics of VBTO systems are described and key concepts developed, in particular the concept of reconfiguration flexibility. Discrete event simulation and Markov models are developed to study the behaviour of the VBTO fulfilment model. The non-dimensional ratio of product variety / pipeline length is identified to be a fundamental indicator of performance. By comparing the VBTO system to a conventional system that can fulfil a customer from stock or by BTO only, the role of pipeline fulfilment is identified and a surprising observation is that it can cause stock levels and average customer waiting time to be higher than in a conventional system. The study examines also how customer differences, in particular their willingness to compromise and their aversion to waiting, affect fulfilment and how fulfilment is dependent on reconfiguration flexibility.

The present thesis designs, implements and evaluates a channel for interaction between office and field users through a collaborative mixed reality system. This channel is aimed to be used for civil engineering purposes and is thus oriented toward the design and construction phases. Its application should contribute to the reduction of the challenges faced by those involved in a civil engineering project dealing with communication, collaboration and mutual understanding. Such challenges can become real problems for multidisciplinary teams of architects, engineers and constructors when working on the same project. In the context of this thesis, outdoor users are equipped with a real-time kinematic global positioning system receiver, a notebook, a head-mounted display, a tilt sensor and a compass. A virtual environment representing components of a civil engineering project is displayed before their eyes. Outdoor users share this collaborative virtual environment with indoor ones. They can talk to and see each other through an avatar. Indoor users can take part from any location where Internet is available. The goal of this thesis is to show that a networked solution of at least two users (In this case, indoor and outdoor users) is an opportunity for outdoor users to perform complex tasks whilst experiencing an immersive augmented reality application. Indoor users interact with outdoor ones when handling and navigating the virtual environment, guiding their counterpart through the scene and making clear common points of understanding. The thesis evaluates how users interact within a prototype system using a formative approach. Users are introduced to the system and motivated to “talk loudly”, thus verbalising what they are experiencing during the tests. All users are video-recorded while performing the exercises and interviewed immediately after. The evaluation reveals that users end up experiencing a system that is too immersive, which ends up narrowing their “attentional spotlight” to the virtual environment and not, as desired, experiencing an augmented reality system. The evaluation also makes clear that the design of the virtual environment is eventually more important for users than the system itself, and it is completely the kind of application that it is being used to and who the users are.