MS Pegram, "Strain measurement via the inner surface of a rolling large lug tyre"
Supervisor: TR Botha
Co-supervisor: Prof. PS Els
The complex interface between tyre and terrain is a largely studied topic in terramechanics and vehicle dynamics research. This interface, known as the contact patch, is however hidden from view and cannot easily be measured. Several studies have focused on measuring tyre strain on the inside surface of the tyre to indirectly determine tyre parameters. The inner surface is separated from the contact patch by the tyre thickness however this difference can be considered small in comparison to the benefit gained by a safe environment for measurement systems. Static studies of tyre strain have been successful however lacks the important phenomena occurring in a rolling tyre. Tyre strain measurements in dynamic tyres have been limited to discrete points and/or once per revolution, which is an insufficient sampling rate for vehicle stability controllers such as ABS.
This study performs full-field and point strain measurements of the inner tyre surface of a rolling agricultural tyre at low speeds. Stereo cameras mounted on a mechanically stabilised rim will record full-field measurement of the contact patch kept in constant view. Digital Image Correlation techniques are used to determine full-field deformation and strain from successively captured images. Point measurements, such as strain gauges, are included in the study for a comparative measurement.
An agricultural tyre hosts large lugs which include large strain concentrations within the contact patch. The complex tyre structure significantly influences the strain measurements, other factors such as inflation pressure, vertical load and slip angle is also studied. Since most vehicle forces are transmitted through the tyre at the tyre-terrain interface, capabilities to measure this area will be a great benefit for tyre research and leading towards a smart tyre.
T.J. Bam, "A computer-based justification for using the simple bend test as the basis for predicting the performance of steel hooked-end fibres in reinforced concrete"
Supervisor: Prof. S. Kok
Co-supervisor: Prof. D.N. Wilke
The classical test to confirm the performance of a given fibre design for use in reinforced concrete is the pull-out test. While attempts have been made to simulate the performance of such pull-out tests, these simulations have been found to vary widely in terms of complexity and accuracy. It therefore seems attractive to calibrate the numerical model using some form of inverse analysis. This thesis examines whether the tensile test, simple bend test and pull-out test provide enough information to define a starting material model that may be used for further design and simulation of hooked-end fibres.
It is widely known that the properties of the fibre are directly influenced by the wire drawing process and all subsequent manufacturing evets. The information required to simulate the performance of reinforcing fibres includes knowing the wire-drawing schedule and the work hardening behaviour, information that is typically unknown in practice. Sensitivity studies investigating the relevance of various factors of the fibre manufacturing process in terms of predicting behaviour during pull-out concluded that variables relating to the work hardened state of the material after wire drawing were the most relevant. Furthermore, the behaviour during pull-out is insensitive to the chosen hardening law, i.e. kinematic versus isotropic. A bilinear stress-plastic strain curve was shown to be a suitable approximation for a material that has been subjected to the wire drawing process.
While inverse analysis demonstrated that all tests provide sufficient information to determine the required properties for this bilinear material model, the pull-out test was shown to provide more accurate approximations of the maximum pull-out force at the first and second peaks and the bend test was shown to produce more accurate approximations of the energy associated with pull-out.
Sensitivity analysis and characterisation of the material model using an experimental pull-out curve demonstrated the importance of the coefficient of friction. Full characterisation using the pull-out curve therefore requires the solution to a three-variable problem: yield strength, tangent modulus and coefficient of friction.
S. A. Dressler, "Simulation of Fibre Pull-out Using a Graphics Processing Unit Accelerated Discrete Element Model"
Supervisor: Prof. D.N. Wilke
The combination of brittle material with ductile fibres can produce competent composites. The fibres transmit tensile forces across cracks that form in the brittle matrix at relatively low tensile strains. The fibre reinforcing, therefore, acts to both increase the maximum stress a structural section can support and improve the post maximum stress behaviour from brittle to ductile failure. An essential aspect of defining the effectiveness of fibre reinforcing is resolving the behaviour of the interface between the fibre and the matrix as the load being transmitted between the matrix and fibre increases.
The interface behaviour for simple fibres is understood analytically, and several models exist that can predict the stresses in the interface. Numerical models using finite element methods (FEM) have been used to investigate this problem in a more general way. FEM, being inherently a description of a continuum, does not elegantly describe the debonding process that occurs during the debonding of fibres from the surrounding matrix. Discrete Element Methods (DEM) describe continuous and discontinuous materials as the interaction between multiple independent particles and are well suited for modelling fracture and evolving contacts.
For this study two different DEM contact models are compared to investigate the model complexity that is required to describe fibre/matrix interface stresses and debonding accurately. A simple model (a linear spring model that only transmits normal and tangential forces) and a more complex model (parallel bonds which transmit normal and tangential forces, moments, and torsion) were used. Two stages of fibre pull-out were modelled independently using a GPU accelerated DEM simulator developed by the author: a fully bonded stage and the de-bonding stage. It was found that both models were able to simulate all stages when compared to analytical solutions. No improvement to the model behaviour was evident from using a complex contact model; for this reason, a simpler, faster contact model should be used to analyse this problem.
The DEM code is written relying heavily on the Numba module which allows the compilation of Python syntax for execution on a GPU. Non-reversible bond damage is simulated, and each bond must, therefore, be stored and bond damage updated at each time step. The implementation of collision detection, particle force determination and equation of motion integration written for execution on GPU are discussed. The data structure and memory use are described. The method used to apply boundary conditions is described. The performance of the developed code is investigated by comparison with similar codes, using Numpy and Numba Python modules, written for serial execution on CPU only. It was found that the developed code was 1000 times faster than the Numpy+Python implementation and 4 times faster than the Numba+Python implementation for force determination and equation of motion integration. Collision detection was 900 times faster compared to Numpy+Python but performed slower compared to Numba+Python.
C. A. Bam, "Development of a three-dimensional mesh generator with analytical mesh sensitivities for use in gradient-based shape optimisation"
Supervisor: Prof. D.N. Wilke
Co-supervisor: Prof. S. Kok
Structural shape optimisation is a field that has been studied since early in the development of finite element methods. The sub-fields of shape and topology optimisation are continuously growing in industry and aim to leverage the benefits of technologies such as 3D printing and additive manufacturing. These fields are also being used to optimise designs to improve quality and reduce cost.
Gradient-based optimisation is well understood as an efficient method of obtaining solutions. To implement gradient-based optimisation methods in the context of structural shape optimisation, sensitivities describing the change of the domain stiffness are required. Mesh deformation sensitivities are required to obtain the stiffness sensitivities. In this study a mesh generating method is developed that provides these mesh sensitivities.
The meshing method developed in this study is analogous to a linear truss system. The system is solved for static equilibrium through a geometrically non-linear finite element analysis using Newton’s method. Sensitivities are made available by Newton’s method for use in generating mesh sensitivities for the system.
Multi-point constraints (MPCs) are used to define the relationship between the mesh and the geometry boundary allowing for an accurate description thereof. Two MPC methods are investigated in detail for boundary discretisation, namely, the Lagrangian and Master-Slave Elimination Methods (MSEM). These two methods are compared for robustness and efficiency. The MSEM presents several difficulties in obtaining convergence on non‑linear boundaries in general when compared to the Lagrangian method. The MSEM has reduced computational requirements for a single Newton step, especially when direct solvers are used. However, when indirect solvers are implemented the time difference between the two MPC methods reduces significantly.
The Lagrangian method was selected as the preferred MPC method for implementation in the mesher to avoid the equilibrium convergence problems associated with the MSEM. This is justified on the basis of reliability outweighing the 6% time penalty for what is intended to be a tool in the shape optimisation process.
Analytical mesh sensitivities were derived for the mesh boundaries and proven to be accurate through comparison with numerical sensitivities. The method was demonstrated to be able to accurately describe mesh deformation throughout the domain for both uniform and non-uniform meshes in the presence of convex and concave non-linear boundaries.
DC Human, "PREDICTING STAGE PERFORMANCE OF A MULTI-STAGE CENTRIFUGAL COMPRESSOR USING THE OVERALL COMPRESSOR PERFORMANCE CHARACTERISTIC"
Supervisor: Dr. G Mahmood
Co-supervisor: Prof. J van der Spuy
The reliable operation of Integrally Geared Centrifugal Compressors (IGCCs), used in the coal-fired power generation industry of South Africa, is essential for economic, environmental and safety considerations. However, due to the unavailability of individual stage performance curves, the ability of a compressor owner to identify underperforming stages to maintain these compressors proactively remains limited.
This study addresses the stage performance prediction of an IGCC when only the compressor’s overall performance characteristic, in conjunction with the impeller diameters and tip speeds, are known. The study is limited to IGCCs used in the coal-fired power generation industry of South Africa.
Based on the limited inputs, two performance modelling methods were considered for this application, namely stage stacking and 1-dimensional modelling. However, stage stacking requires known operating points on each stage performance curve from which the rest of the curve can be extrapolated while 1-dimensional models require detailed stage design information to model stage performance.
This study developed a revised stage stacking procedure which in contrast to the traditional stage stacking procedure, does not require a known operating point on each stage’s performance curve, for it assesses the relative stage performance at the compressor’s surge flow rate. The relative maximum pressure ratio of each stage is acquired through the application of similarity principles while a simplified 1-dimensional impeller analysis model is used to assess relative impeller head coefficients.
The modelling process was developed based on performance and design data for IGCCs obtained from a compressor manufacturer. Performance data of four IGCCs, consisting of 13 stages, were obtained, including the design data for ten impellers.
Hence, the IGCCs satisfy the requirements of geometric and aerodynamic similarity, unveiling a linear relationship between the stage impeller tip speed and maximum pressure ratio. A simplified 1-dimensional performance model was used to assess relative impeller head coefficients. A verification procedure ensured the integrity of the findings of the 1-dimensional model was maintained by comparing the model results to findings obtained using commercial compressor performance modelling software. A sensitivity analysis was conducted on the 1-dimensional performance model to ascertain which input parameters could be scaled as a function of the impeller tip diameter.
For the four IGCCs for which data were obtained, the stage-discharge pressure and isentropic efficiency curves were calculated using the developed model. The maximum variation between the measured and calculated pressure and isentropic efficiency curves equaled 8.20% and 10.84%, respectively. The prediction accuracy of the developed modelling procedure is similar to map-based models found in literature and is considered adequate for identifying an underperforming stage. Thus, the developed model could serve as a valuable conditioning monitoring tool for site-based compressor owners.
Keywords: IGCC, stage performance, stage stacking, Aungier, sensitivity analysis, similarity principles, parameterised, non-dimensional
T.M.G. Komana, "Tyre Inflation Pressure control to improve vehicle ride comfort on rough roads"
Supervisor: Prof. P.S. Els
Tyres behave similar to a spring and damper systems smoothing out road irregularities as the tyre rolls. Tyre stiffness and damping characteristics are largely influenced by the tyre pressure. As a result, tyre pressure has an influence on tyre enveloping and ride comfort.
To take advantage of the influence of tyre pressure on ride comfort, a pressure controlled tyre was developed to vary the tyre pressure, thus varying the tyre characteristics as the road conditions change to improve ride comfort. Literature exists on TIS (Tyre Inflation Systems) and smart suspension control strategies optimised for ride comfort, which indicate that a pressure controlled tyre can be developed by managing a TIS with a ride comfort controller.
A VDG (Vehicle Dynamics Group) test trailer was used to complete this study. The trailer was modelled on three platforms; MATLAB, Cosin and ADAMS View; with the co-simulation managed through Simulink. Three tests were conducted to parametrise and validate the model, namely; pneumatic system parametrisation tests, APG (Aberdeen Proving Ground) bump tests and Belgian paving tests. The pneumatic system parametrisation tests show that the discharge coefficient is approximately 0.07 for choked flow and tappers off to zero for unchoked flow. Also, tests show that the tyre can be inflated from to 1barto 3 bar in 8s and deflated from 3 bar to 1bar in 13s . The APG bump tests and Belgian paving tests were conducted to validate the model over discrete obstacles and rough roads, respectively. These tests indicate that the model correlates with the actual trailer.
The validated model was used to develop the TIPc (Tyre Inflation Pressure controller), which uses a running RMS control strategy to manage the TIS. The objective of the TIPc is to maintain a NOT uncomfortable ride comfort level. The TIPc achieves this by deflating the tyre to a suitable tyre pressure when the ride comfort level is above NOT uncomfortable. The TIPc is robust to evaluate ride comfort on smooth roads and rough roads, as well as, detecting discrete obstacles. The TIPc ignores discrete obstacles when evaluating ride comfort to determine a suitable tyre pressure to improve ride comfort. The TIPc was able to achieve a 4-7% RCI (Ride Comfort Improvement).
T. M. Nkosi, "ASSET OPERATIONAL READINESS ASSESSMENT OF NEW BUILD POWER PLANT EQUIPMENT"
Supervisor:
The delivery of medium and mega project has been a challenge for a number of decades, with attempts made to reduce the associated issues around engineering projects implementation. Project delivery before 1950 mainly relates to cost, time, and scope, with a lack of documentation pertaining to methods, as well as inadequate techniques to achieve a quality final product. The concept of Asset Operational Readiness (AOR) emanates in the 1950s from the military as means of providing the “developmental state of weapons systems”. The concept gained momentum as it was associated with “system safety” in the 1980s for decision-making.
AOR can be defined as an establishment of a state or configuration which, after completion of the project, “places the right people in the right places at a right time working with the right hardware according to the right procedures and management controls”.
The research work covered in this thesis, aims to propose a best-practice AOR framework for mega-projects in the power generation industry. A thorough Literature Review provides an overview of best practices on the AOR requirements for various industrial fields. The survey shows that AOR implementation follows the Project Life Cycle Management (PLCM) principles, from conceptual and pre-feasibility phase to commissioning and operation phase. In addition, the survey considers methodologies and techniques, which aids to enhance AOR framework development such as Root Cause Analysis (RCA) exercises.
The study has provided an opportunity to develop an AOR theoretical framework refinement methodology, inclusive of RCA, AOR assessment tools, qualitative survey tool, and scoring systems. The AOR best practice framework and refinement methodology application to a real mega project case study, with historical data, enables a stage wise assessment of each component for individualized performance rating. This provides an identification of the areas that require refinement to have an improved AOR framework as outcome.
The research outcome shows that there are implications for inadequate development and implementation of items in the proposed framework. The implications range from rework during manufacturing and construction, poor product quality delivery, poor performance post commissioning, and overall cost overruns. In addition, the study provides evidence that implementation of the AOR framework aids a project to realize its potential and yield positive results, which ultimately benefits an organization in terms of quality product delivery, cost reduction, and optimal Operations and Maintenance of the established asset.
S. Mkhabela, "INVESTIGATION INTO THE EFFICIENCY AND EFFECTIVENESS OF RISK BASED INSPECTION (RBI)"
Supervisor:
Risk Based Inspection (RBI) is a risk assessment and management tool that addresses an area not completely addressed in other organisational risk management efforts such as Process Hazards Analyses (PHA) or Reliability-Centred Maintenance (RCM) [11]. It complements these efforts to provide a more comprehensive risks assessment associated with equipment operations. RBI yields Inspection and Maintenance Plan for equipment which identifies the actions that should be implemented to provide reliable and safe operation. RBI is a preventive maintenance strategy that combines predicting the expected failure time and condition monitoring as an effort for preventing potential equipment failures.
A literature review was conducted to obtain a wide understanding of the RBI field and to capture some of the improvement methods which may have been highlighted by other researchers for baseline purposes. An RBI implementation case study was performed, focusing on the largest power producing utility in Africa, to investigate the efficiency and effectiveness of RBI implementation, thus the results obtained from this case study could have a significant impact in the RBI and Asset Management environment. Efficiency in this study refers to achieving financial benefits, while effectiveness means serving the technical and risk mitigation purpose of a RBI implementation.
To evaluate the efficiency of RBI implementation, costs for executing scopes of work prior to RBI implementation were compared to RBI scopes Pre-Outage and Post-Outage scopes execution costs on two power stations that were identified as the most advanced in RBI implementation roll-out plans. Effectiveness was evaluated by assessing the RBI implementation against typical organisation’s objectives through audit findings, interviews, and lessons learnt.
Data Analysis was performed as described below:
• Cost analysis was performed, comparing the RBI versus Prior RBI maintenance scopes execution costs on the two identified power stations through an excel Model.
• Most recurring audit findings were identified through reviewing the audit reports.
• A bow-tie risk assessment was performed for the identified eight most recurring audit findings, and probabilities and consequences mitigations were recommended.
• Lessons learnt were compiled from the audit findings, bow-tie risk analysis mitigations, interviews results, and RBI sharing sessions.
A proposed framework was developed for RBI implementation improvement methods. The results showed that RBI is generally a cost effective process when the prior RBI scope execution cost was compared to RBI scope execution cost. RBI could reduce the maintenance costs through scope optimisation and downtime reduction. The RBI implementation process was found deficient for the specific instances, based on the audit findings and bow-tie risk assessment conducted in the case study. The most significant improvement areas identified included, ensuring that RBI scopes have are uploaded into the Computerised Maintenance Management System and there is only one consolidated final inspection scope submitted to Outage Department and tracked for tasks completion during the outage.
This study revealed that the case studied organisation is currently not efficient in implementing RBI, and could benefit significantly if they improve through executing the RBI maintenance inspection scopes as planned. The conducted interviews, recurring audit findings, and lessons learnt analysis demonstrated that the organisation is not effective, as it was successful in meeting only one from a total of six RBI implementation objectives. Extending the inspection frequencies to 72 months and beyond for some low risk components through RBI implementation was the only RBI objective in which the organisation was met successfully.
W.J. Visagie, "Blade tip timing to determine turbine blade fatigue in high backpressure conditions"
Supervisor: Prof. P.S. Heyns
In service, blade failure in turbines is a known problem and with a significant safety risk, high repair cost, non-operational revenue loss and even retiring of the unit. Therefore, the reliability of these blades, especially the low pressure (LP) blades, is of utmost importance for the successful operation of a steam turbine. To aggravate the matter, low load and/or high backpressure conditions can lead to stall flutter, high cycle fatigue and failure of last stage blades (LSB). Although the mechanism is well known, the conditions under which this may occur in a specific unit are not well defined. Recommended minimum load levels and maximum backpressure (poor vacuum) are typically defined by the original equipment manufacturer (OEM). However, it is not always easy to consistently adhere to these requirements and conditions do arise where some units operate out of these specifications for varying timespans.
This dissertation focuses on an investigation of blade tip timing (BTT) data together with finite element analysis (FEA) as a possible procedure for calculating the fatigue life of a steam turbine LSB running under FSNL, or low load at high backpressure conditions in the condenser. The investigation concentrates on the non-synchronous vibration of the LSB while vibrating at a blade resonance frequency. The procedure makes use of BTT data and a method, which assumes that a dominant resonance frequency is present in the BTT signal, and that only the mode excited by the dominant frequency contributes to the fatigue damage. The procedure does not include the participation of other modes with the dominant mode, in terms of establishing life estimations.
The FEA was conducted on the complete LSB through a cyclic symmetric analysis of a single blade and rotor sector with two half damping pins connected to either side of the blade. Plasticity is applied where the blade root deforms plastically due to initial overspeed testing. Damping was added to the model by means friction and Raleigh damping coefficients. The modal analysis and correct extraction of natural frequencies and modes are important factors to consider. Two tests were conducted to verify the FEA. The first verification is a static impact test conducted by a modal hammer, to verify stationary blade natural frequencies. The second verification test is a balance pit test with strain gauges, to verify natural frequencies at operating speed.
The fatigue life calculated for the dominant frequency for mode two, nodal diameter seven was 4.24 years in operation, for which deflection was measured during low load conditions. A significant change in results can be expected if a higher stress is present or, the duration of exposure at peak stress is higher than what the test data has shown. At high load conditions, the blade life calculations presented not to be an area of concern and fatigue life was calculated to be twelve orders higher than during low load conditions. The locations of the lowest fatigue life, which are the top two serrations on the inlet side, corresponds to the location of failures were cracks initiated.
RO Harat, "Dynamic Investigation of Vibratory Screen Response in A FEM Environment"
Supervisor: Prof. P.S. Heyns
Vibratory screens are critical in the mining and material processing industry. The high loading conditions on a screen deck during operation can lead to failures of the screen. Despite their importance, too little has been done to investigate the faults and failures of screens and the development of numerical models which are able to simulate and help understand the nature of the failures. In this work finite element methods (FEM) are used in the development and validation of vibratory screen dynamics. Experiments are conducted on the physical screen where the stiffness coefficients of the rubber buffer supports are varied. The response of the FEM model is validated against the physical screen by comparing the dynamic responses. Intelligent methods are used to compare the dynamic responses as they allow for the identification of often indiscernible and minute differences in signals. It was found that it is possible to classify the faults on both the FEM model and the physical model accurately. Features are used to identify and measure the differences between the experimental results and the FEM model. Finally, the intelligent models trained using the FEM model are used to classify faults on physical screen data.
C.J. Qambela, "Investigation of damage detection methods for laminated composite materials using full field digital image correlation"
Supervisor: Prof. P.S. Heyns
Co-supervisor: Dr. H.M. Inglis
The aeronautical industry is among many which have focused on material with a high strength to weight ratios in order to accomplish as much efficiency as possible during flight. Among material of high strength to weight ratios are composite materials (Matthews and Rawlings, 1999).
A composite material is defined by Lee (1989) as the combination of two or more materials of different characteristics (composition or form) which remain bonded together. This yields a material which essentially has all the beneficial attributes of its parent materials and little of their shortcomings. In the aeronautical field composite materials are used to construct components such as airplane wings spoilers and panels, vertical and horizontal stabilizers etc. (Baker et al., 2004).
The wide use and range of composites has led to the development of various techniques for damage detection. One of the common types of damage which composites experience is barely visible impact damage, which can be caused by dropped objects during construction or maintenance. Laminated composites, unlike isotropic materials such as steel, show essentially no yielding prior to complete failure. This makes it crucial to assess them for damage before use.
This research presents a method for damage detection in laminated composites by making use of full field digital image correlation (DIC). The technique is often employed to assess the structural deformation characteristics of components under various loading conditions, which can then be correlated with finite element assessments. In this research three experimental methods are explored in terms of their performance in detecting barely visible impact damage: 1) modal analysis, 2) full field DIC under static loading, 3) full field DIC under dynamic loading. Modal analysis results showed no noticeable shifts in natural frequencies between an undamaged and damaged carbon/epoxy woven laminated composite, when damage was induced via static point load application at the center of the specimen.
Full field DIC under dynamic loading namely, drop impact tests designed to induce no addition damage, revealed no change in peak out-of-plane displacements between an undamaged and damaged specimen. The technique proved effective only when severe visible damage was induced, which falls out of the scope of the research.
Prior to full field DIC testing under static loading a number of experimental exercises were performed to check the accuracy of the method when measuring rigid body motion and out-of-plane displacements. The acquired results were compared to independent measurements obtained using a micrometer for rigid body motion and an eddy current probe for out-of-plane displacements.
The method of full field DIC under static loading condition showed significant reduction in stiffness between the undamaged and damaged composite, with increases in out-of-plane peak displacement and von Mises strain fields. The success of full field DIC under static loading paved the way forward for the investigation of damage detection in laminated composites under various impact energies (i.e. centered impact and off-center impact) resulting in barely visible impact damage.
The finding revealed an increase in compliance of the impacted laminated composite due to the damage. The barely visible damage can be detected using full field DIC only when the loading or excitation is located around the damaged region. There was no noticeable change in out-of-plane displacement fields between specimens with centered impact and off-center impact. Significant changes in von Mises strain fields exist between the undamaged and damaged laminated composites with centered impact.
B.C. van der Westhuizen, "Impact of coal quaility on equipment lifetime at coal-fired power stations "
Supervisor: Prof. J. Wannenburg
With the export of coal being more lucrative than selling coal to South African power producers, power station operators might consider accepting lower-quality coal. While the impact lower quality coal has on cycle efficiency is understood, the influence it has on equipment reliability and lifetime is often not understood. This study focusses on addressing the question of how different characteristics of coal influences different damage mechanisms of common power station equipment.
The influence that coal quality has on air-heater element erosion was evaluated. This was accomplished by creating an air heater erosion model. This model was used along with historical coal quality and air heater erosion history to develop and fit the model for full boiler load. The model was verified against data not used during the development of the model, and a seemingly good prediction was made when compared to the measured result.
The influence abrasiveness index has on mill liners was also investigated as part of this study. Historical liner ultrasonic thickness and coal abrasiveness index results were used to fit a mathematical formula. The relationship was established to be directly proportional to increased abrasiveness index resulting in an increased wear rate.
A case study of the thermal damage mechanisms caused by variation of calorific value was completed by creating a mathematical thermo-hydraulic model of a hypothetical boiler and calculating the effect calorific value would have on the boiler temperature distribution. The results were then compared to temperature-related damage mechanisms; the comparison indicated that a variation in calorific value, whether up or down from the designed value would be negative for overall boiler health.
The final case study on a drum reclaimer was not completed due to the unavailability of related equipment.
J.F. Potgieter, "Numerical investigation on the effect of gravitational orientation on bubble growth during flow boiling in a high aspect ratio microchannel"
Supervisor: Dr. M. Moghimi Ardekani
Co-supervisor: Prof. J.P. Meyer
Co-supervisor: Dr P. Valluri
This study focused on numerically modelling the growth of a single bubble during the flow boiling of FC-72 in a microchannel with a hydraulic diameter of 0.9 mm and an aspect ratio of 10. The numerical domain was limited to a 10 mm section of the microchannel where bubble nucleation and detachment were observed in an experimental study on a similar microchannel setup. The high cost of 3D simulations was offset by an interface-tracking mesh refinement method, which refined cells not only at the interface, but also a set distance on either side of the interface. A user-defined function was used to implement a mass transfer model that is suited to micro-scale applications. A simplified approach is used to focus on the effects of buoyancy and bubble detachment. Simulations are first performed in a 2D section through the centre of the microchannel, and then in the full 3D domain.
In both the 2D and 3D results, the bottom heated case had the lowest maximum temperature and the highest heat transfer characteristics, which were influenced by the detachment of the bubble from the heated surface. This observation indicates that the gravitational orientation of the channel can have a significant effect on the heat transfer characteristics of microchannel-based heat exchangers, and that more investigation is required to characterise the extent of this effect.
R.M. Steyn, "Local heat transfer coefficients in an annular passage with flow turbulation"
Supervisor: Prof. J Dirker
Co-supervisor: Prof. J.P. Meyer
In this experimental and numerical investigation, the use of flow turbulation was considered as a method to increase local heat transfer coefficients in annular heat transfer passages. Experimental data was obtained for cases with and without inserted ring turbulators within a horizontal annular test section using water for average Reynolds numbers ranging from 2000 to 7500 and average Prandtl numbers ranging from 6.73 to 6.79. The test section was heated uniformly on the inner annular wall and had a hydraulic diameter of 14.8mm, a diameter ratio (inner wall diameter to outer wall diameter) of 0.648, and a length to hydraulic diameter ratio of approximately 74. A set of circular cross sectioned ring-type turbulators were used which had a thickness of 1mm, a ring diameter of 15.1mm and a pitch of 50mm. It was found that the presence of the flow turbulators increased the average Nusselt number by between 33.9% and 45.8%. The experimental tests were followed by numerical simulations to identify the response in the heat transfer coefficient by changing the geometry of the turbulators. For this, the turbulator diameters were ranged from 0.5 mm to 2 mm, and the gap size (between the inner wall and a turbulator ring) ranged from 0.125 mm to 4 mm at a pitch of 50 mm. The results showed that the use of turbulators increased the Nusselt numbers by a maximum of 34.8% and that the maximum can be achieved for a turbulator diameter of 2 mm and a gap size of 0.25 mm, for all the Reynolds numbers tested. From the numeric determined pressure drop values it was found that the smaller gap size had the lowest pressure drop and the smallest turbulators also produced the lowest pressure drop.
O. Scholtz, "The effects of macro and micro road texture on tyre simulations"
Supervisor: Prof. P.S. Els
The tyre contact patch is important to vehicle dynamics analysis since this is where the tyre-road interaction forces originate. If the contact patch mechanism is understood in depth and can be simulated accurately, it could improve various aspects of vehicle and tyre design. This study investigates the effect of changes in macro and micro texture on the forces generated in the contact patch. This is done by inspecting the effect of the resolution with which the surface is measured, on simulations of a vehicle driving over a rough road.
The macro texture largely affects the normal force that is created at the contact patch and influences the ride comfort. This is evaluated with the use of an FTire tyre model on a quarter car and a full-scale model respectively through simulations on a rough road. The quarter car model is used to evaluate the effects of speed, tyre model resolution and road geometry resolution on the reaction forces and moments based simulation. This analysis concludes that the speed and tyre resolution had little to no effect on the results. The road profile resolution, however resulted in a 20% error when a 15x15mm resolution was compared to a 1x1mm road resolution. A 10x10mm resolution road is ideal for a quarter car simulation. The Belgian paving, representing the rough road, was created at various resolutions on a macro texture level for full-scale vehicle simulations. This simulation data is compared to the experimental data. This experiment is solely based on evaluating the road resolution. It is found that the percentage error between the simulations and experimental data differed, depending on the resolution of the road. However, the differences were less than 5%. The simulations were compared to the most accurate road resolution and the percentage error differed by less than 10%. This indicates that a road resolution smaller than 15x15mm is only significant to highly sensitive results or quarter-car studies over a rough road when ride comfort is being evaluated.
The micro texture largely has an effect on the friction coefficient at the contact patch in the form of lateral and longitudinal force. The friction coefficient in this case is estimated with the use of a physics based model, which was validated on a flat concrete block. By comparing experimental data with the estimation data, the estimation model was found to be sufficient to estimate the friction coefficient on a rough road. The comparison resulted in a ± 6% error for the estimations. The micro texture resolution, at which the surface was measured, was subsampled to evaluate the effect it had on the estimation. It was determined that the resolution at which the micro texture is measured, influences the friction coefficient estimation. The percentage error of the estimation model changed by up to 10% for an increase of increments of 0.12x0.12mm. The change in micro texture resolution has a larger effect on the rough road than on the flat concrete road. To ensure that the percentage error stays below 10% the road should be sampled at 0.144x0.144mm for a rough road and 0.216x0.216mm for a flat road. Since the measuring capability was limited, no conclusion can be drawn for results with a road resolution of smaller than 0.12x0.12mm.
W.N. Niehaus, "NIC methodology: A probabilistic methodology for improved informative frequency band identification by utilizing the available healthy historical data under time-varying operating conditions."
Supervisor: Prof P.S. Heyns
Co-supervisor: Dr. S. Schmidt
Effective incipient fault detection requires a method that can separate fault signatures under constant and time-varying operating conditions. Identification and optimal selection of the informative frequency band which contains fault information is the focus area of the research in this article. Many automatic band selection techniques exist and have proven effective under constant speed conditions. However, it has been shown that these techniques occasionally identify frequency bands that contain non-damage related information, especially under fluctuating speed conditions and at low damage levels. With this research, a new methodology is proposed which makes use of popular informative frequency band selection techniques, such as the Fast Kurtogram amongst others, to effectively identify damage under constant and fluctuating speed conditions. A key step in this methodology, the NICogram, requires healthy historical data, which is used to identify frequency bands that contain novel information in unclassified signals. The methodology uses multiple signals to identify whether a component is damaged or not through a probabilistic approach. It is shown that the method performs much better than the conventional informative frequency band identification methods on synthetic and experimental data.
B. Ellis, "A real-time hybrid method based on blade tip timing for diagnostics and prognostics of cracks in turbomachine rotor blades"
Supervisor: Prof P.S. Heyns
Co-supervisor: Dr. D.H. Diamond
This dissertation proposes hybrid models for (i) diagnosis and (ii) remaining useful life estimation of a single fatigue crack in a low-pressure turbine blade. The proposed hybrid methods consist of physics-based methods and data-driven methods.
In this dissertation, blade tip timing is used to measure the relative tip displacement of a rotor blade. The natural frequency of the blade is determined by detecting the critical speeds of the blade using a newly derived least squares spectral analysis method. The method shares its origin from the Lomb-Scargle periodogram and can detect resonance frequencies in the blade’s displacement while the rotor is in operation. A Campbell diagram is then used to convert the critical speed into a natural frequency. Two kinds of shaft transients are considered, a run-up run-down crossing the same critical speed, is used to test the new method. This dissertation shows that the relative displacement of the blade tip is comparable to those simulated from an analytical single degree of freedom model. It is also shown that the newly proposed resonance detection method estimates the natural frequency of the blade to a high degree of accuracy when compared to the measurements from a modal impact hammer test.
The natural frequency obtained from the real time measurement is then used in a pre-constructed hybrid diagnostics model. The diagnostics model provides a probability density function estimation of the surface crack length given the measured natural frequency. A Gaussian Process Regression model is trained on data collected during experiments and finite element simulations of a fatigue crack in the blade.
The final part of this dissertation is a sequential inference model for improving the estimation of the crack length and the prediction of the crack growth. The suggested model uses an unscented Kalman filter that improves estimations of the crack length and the rate of crack growth from Paris’ Law coefficients. The model is updated each time a diagnosis is performed on the blade. The RUL of the blade is then determined from an integration of Paris’s Law given the uncertainty estimates of the current damage in the blade. The result of the algorithm is an estimation of the remaining number of cycles to failure. The algorithm is shown to improve the overall estimation of the RUL; however, it is suggested that future work looks at the convergence rate of the method.
A.J. Peenze, "Model predictive suspension control on off-road vehicles"
Supervisor: Prof P.S. Els
Reducing the rollover propensity of off-road vehicles while maintaining good ride comfort and off-road capabilities is a well-known challenge. With controllable suspension systems, the dynamics of the vehicle can be altered to give better performance than passive suspension systems. The semi-active hydro-pneumatic suspension system under consideration can switch between a soft and stiff spring, as well as between low and high damping.
In this study, a model predictive controller, based on a linear quadratic regulator and receding horizon theories, was developed to control individual struts of the suspension system. A combined handing and ride comfort metric was developed to determine the input weights of the model predictive controller based on the driving conditions. The metric discerns between a handling or emergency manoeuvre and normal driving on rough roads. It changes the input weight accordingly to bias the controller towards a handling setting or a ride comfort setting.
A 16 degree of freedom simulation model was validated for both the lateral and vertical dynamics and found to be a close representation of the real Land Rover Defender 110 used for the experiments. The controller was implemented into the simulation model to test and ensure the controller worked as intended. The simulation model was validated at speeds varying from 50 km/h to 80 km/h for severe double lane change handling manoeuvres. The ride validation was performed over a rough Belgian paving road at speeds of 21 km/h and 47 km/h.
The controller was also experimentally validated for the double lane change, Belgian paving and various other handling and ride comfort tests. In the handling test, the controller performed well keeping the suspension in handling mode and reducing the roll angle as compared to the ride comfort mode. Over the rough tracks, the performance of the controller was not good and the suspension controller did occasionally switch to the handling mode. Although the controller did switch over to the handling mode the vehicle’s ride comfort wasn’t detrimentally influenced.
Overall key aspects of the controller were identified for improvement to overcome the problem experienced in ride comfort settings and also improve the handling of the vehicle.
M. Meyer, "Modelling and multi-objective optimisation of heat transfer characteristics and pressure drop of nanofluids in microtubes"
Supervisor: Dr. M. Mehrabi
Co-supervisor: Prof. J.P. Meyer
A literature study was performed on the inner mechanisms of nanofluids and flow in microchannels. With ever changing technology, the need for smaller and more efficient devices has come about in the last couple of years. With the shrinking in size of components in electronics, an increase in heat has become a notable problem. With conventional heat transfer fluids not being able to handle the required heat removal rates, research into fluid enhancing has been of great interest. A nanofluid is a fluid with enhanced heat transfer potential, which can solve the problem of extracting enough of the added heat of new-age components. This will allow electronics to work with increased power and accomplish tasks faster. Nanofluids have been a very controversial method of heat transfer as problems with stability were keeping the fluid from replacing traditional heat transfer fluids.
Some research has been done on the models used for simulating and defining the thermal properties of nanofluids. Added accuracy of the models has been seen in recent years. However, no optimal setup for nanofluids has been found in terms of combining parameters like the base fluid and nanoparticle, as well as the concentration and diameter of the nanoparticle. An optimal setup of this kind would produce the best heat transfer rates at the lowest pressure drop. The simulation of nanofluids was done in Ansys CFD. The validation was done with previous literature that had experimental and numerical results. The validation had a very good outcome as some of the temperature data inside the microchannel presented a good correlation to previous work. The setup of the model for simulation and duplication to create a design study was also described and shown. This was done to ensure that the model can be used again if further investigation is needed. This will enable one to determine the effect of a new nanoparticle on the field of study to continuously improve on the model.
The results indicated the best nanoparticle to use with the best base fluid to ensure the lowest pressure drop and highest heat transfer. This was done with a multi-objective optimisation general algorithm. The outcome of the optimisation was that silicon dioxide, as nanoparticle, and water, as base fluid, would give the optimal setup. The diameter also appeared to have a very small effect on the outcome.
M. van der Westhuizen, "Effect of vehicle vibration on healthy term infants: Method and infant car seat vibration quantification"
Supervisor: Dr C. Kat
Co-supervisor: Prof P.S. Els
Even though infants are frequent vehicle travellers, little is known about the effect of vibrations on their comfort and health. Some studies have characterised the vibrational response of infant car seats, while others have focussed on the physiological responses of infants secured in infant car seats under varying operating conditions (stationary, vertically simulated or varying infant postures). The void identified in the current literature is - the relationship between in-vehicle multi-axis vibration input to a new-born infant seated in an infant car seat, and the resulting changes in posture and cardiorespiratory response of the infant. This study aims to form a basis for future studies seeking to investigate the effect of vehicle vibration on healthy, term infants by establishing a comprehensive method and investigating the multi-axis transmissibility of the infant car seat and subsequent vibration input that infants may experience during testing. The relationship between the responses of infants to vehicle vibration and the vibration frequencies, -magnitudes or -directions can be used to design infant car seats that not only provide protection in the event of a crash, but also promote the health of the infant under normal driving conditions.
A comprehensive method was developed for an investigation into the effect of vehicle vibrations on new-born infants. Aspects addressed in this method include quantifying the vibration input to the infant, tracking the infant’s change in posture and measuring their cardiorespiratory response.
As the proposed method relies on in-field testing, the road inputs and speed of the vehicle have been identified as aspects that may influence the vibration input to infants. The mission profile was selected such that it includes road inputs typically found during suburban driving. The vehicle speed was found to have a statistically significant (p < 0.05) influence on the vibrational response of the vehicle body, and based on available difference thresholds, will be perceptible by adult occupants. Whether these differences will result in different responses between infants remains to be investigated. The transmissibility of the infant car seat was determined over the selected mission profile. Although frequencies could be identified where the infant car seat amplified the vibration input to infants, the coherencies determined for the in-vehicle measurements made it challenging to interpret the transfer functions obtained.
If the findings presented in this study are considered during investigations into the response of infants to vehicle vibration, a holistic approach will be followed which considers many of the aspects that may influence the observed responses. This should provide meaningful insight into the effect of vehicle vibration on the health of infants under normal driving conditions, as this effect has not been investigated before.
J.K. Swanepoel, "Helically coiled cavity receiver for a micro-scale direct generation steam Rankine cycle using a novel solar dish design."
Supervisor: Dr. W.G. le Roux
Supervisor: Dr. A.S. Lexmond
Co-supervisor: Prof. J.P. Meyer
Sub-Saharan Africa has an underutilised solar resource that is available to provide distributed-scale power to rural communities that are not reached by the current grid systems that are in place. Renewable power generation systems are typically infeasible for this purpose due to their high manufacturing and maintenance costs. This research proposes to address this by experimentally investigating the performance of a micro-scale, solar thermal Rankine cycle with direct steam generation using an affordable solar collector design. The intended power generation range is between 0.1 and 1 kW, which can be used for the direct mechanical pumping of rural irrigation systems or municipal water supplies. The study focuses on the performance of the solar collector (the solar reflector and the solar receiver) under realistic solar thermal conditions in sub-Saharan Africa, as it is considered to be one of the most important components in the solar thermal Rankine cycle. The performance of the solar collector was first characterised in an optical analysis of the new faceted dish reflector design, which comprised six Mylar membranes stretched over the elliptical rims of television satellite dishes. The optical analysis was conducted with the main goal of determining how much concentrated solar radiation would intercept the experimental receiver aperture during solar testing. This was determined by measuring the reflectivity of the reflector facets through ultraviolet and visible light (UV-VIS) spectroscopy and producing an intercept factor trend as a function of receiver aperture size through photographic lunar flux mapping. Based on the spectroscopy analysis, the spectral reflectivity of the clean Mylar facets was determined to be 97% for the visible light spectrum. The intercept factor was determined to be 87% for the designed receiver aperture diameter of 135 mm. The thermal testing took place on a clear, sunny day with low wind velocities. The solar collector followed the arc of the sun throughout the day with a 1° tracking error. During the testing, municipal water was passed through the receiver at 0.284 g/s and the temperature and pressure within the receiver were recorded. A 91-minute testing period occurred, where the inlet and outlet process flows obtained relatively constant operating temperatures with the inlet temperature at 54 °C and the outlet temperature at 343 °C. The total solar irradiance dropped from 801 to 705 W/m2 during this testing period. Using a reflector with a total incident area of 2.73 m2, the total radiation intercepted at the aperture dropped from 1 845 to 1 625 W during this period. The total power capture by the working fluid averaged at approximately 861 W and the total rate of heat loss was determined to be between 1 000 and 750 W. An average collector efficiency of 42% and an average receiver efficiency of 49% were determined for the testing period. An analysis of the heat loss showed that approximately 84% of the heat was lost through the aperture, of which, 31% was from reflected radiation. The second-law analysis showed that most of the irreversibility in the solar collector was caused by absorption of the concentrated radiation at the coil surface. This was because of the large temperature difference between the sun and the receiver coils. The experimental thermal analysis highlighted the design challenges of the micro-scale thermal Rankine cycle. However, the analysis showed that a solar collector can be constructed using locally sourced, affordable materials and can be used to produce power at a micro-scale. With appropriate attention given to optimising the collector design and determining the optimum operating conditions of the solar receiver, the power cycle would be able to compete with current technologies to provide decentralised power to communities in need.
J.W. Wannenburg, "High-Cycle Fatigue Condition Based Maintenance of Reinforced Concrete Beams"
Supervisor: Dr. H. M. Inglis
Co-supervisor: C. Roth
If a critical reinforced concrete beam fails due to fatigue, it can have severe safety and production consequences, and preventative repair/replacement of such a beam is expensive. It would therefore be beneficial if repair/replacement can be done based on an accurately and conservatively predicted remaining useful life.
Condition based maintenance is a maintenance approach that uses empirical/analytical models and a measurable condition to predict remaining useful life. The P-F curve (condition-life) is a useful tool that can aid in making these decisions.
Flexural concrete cracks are not uncommon occurrences in load-bearing reinforced concrete beams, and it does not necessarily compromise the load bearing integrity of the structure. In this study, a condition based maintenance strategy is developed that uses concrete crack depth as the measurable condition in reinforced concrete beams that are subjected to cyclical loading.
A model to create a P-F curve is developed using rebar fatigue test results (in the form of an S-N curve) and the Palmgren-Miner law of damage accumulation. A Monte Carlo simulation and Weibull distributions are also used in the model. The effects of different variables on the model are then demonstrated.
In a case study, the developed model is applied to a real scenario of a beam supporting a coal plant crusher. The case study successfully proves the model's usefulness.
I. Marsden, "Preventative maintenance optimisation in a capital-constrained environment"
Supervisor: Prof. J. Wannenburg
Co-supervisor: Prof. P. S. Heyns
Due to the recognition of the importance of maintenance from an organisational perspective, a number of different maintenance-related approaches have been developed. These approaches include reliability centred maintenance, business-centred maintenance, total productive maintenance and life cycle costing. They consider maintenance from specific different viewpoints and no single approach can be applied to all circumstances. Common to all these approaches are techniques to optimise the maintenance strategies using mathematical models. A variety of mathematical approaches are described in the literature, all of which involve the minimisation of the total costs incurred in relation to the required maintenance activities. This study focuses on data-driven optimisation models that consider costs and the reliability performance of equipment. The practical implementation of these optimising maintenance models presents two main challenges. First, the decision on when to use which model would depend on the type of system/equipment under consideration, as well as on available data. Different models based on analysing the historical failure data of the system or component are considered in order to optimise the maintenance strategies to be applied to these two types of individual systems. In the case of having a number of identical components or systems in series, where a shutdown of one of the systems results in the shutdown of the entire series, models are considered to allow for analysis with the correct maintenance technique of components or systems showing these trends. A major limitation of these maintenance optimisation models is that they all require failure data for their implementation, which is not always obtainable. Historical maintenance cost data, however, is mostly available, therefore forecasting techniques and life cycle cost modelling are also considered. Second, the successful implementation of optimised maintenance strategies will be dependent on informed budgetary decisions being made. Therefore, the challenge of integrating the outputs from the variety of optimisation models utilised into a cohesive compilation and sensible presentation of an overall maintenance budget for a complex plant needs to be addressed. This study presents an integrated maintenance optimisation model that uses the appropriate sub-models described individually in the literature to enable the integrated compilation and sound presentation of an overall maintenance budget for a complex plant for appropriate decision-making.
The use of the case study validates this methodology. It illustrates that a concise, integrated overall budgetary maintenance decision model is highly beneficial in communicating the budgetary requirements for an organisation. It was found that the outcome resulted in an effective decision-making tool with significant potential for implementation in a variety of organisations in search of optimal maintenance planning and budgetary requirements.
A. G. Vosloo, "Hydropneumatic semi-active suspension with continuously variable damping"
Supervisor: Prof. P.S. Els
A well-known challenge in vehicle dynamics is to design a vehicle that will not only keep the occupants comfortable, but will also ensure safe and stable operation during various manoeuvres over multiple driving surfaces. A soft and compliant suspension is generally required for good ride comfort, while a stiff suspension with a low centre of mass is required for improved handling. These contradicting factors in the design process is commonly referred to as the ride comfort versus handling compromise.
A newly developed semi-active hydropneumatic suspension system is proposed to reduce or negate this compromise by being able to change its characteristics according to the dynamic state of the vehicle. The unit is equipped with two proportional solenoid valves that can provide continuously variable damping. In addition, the valves are able to completely close off flow to compressible gas volumes to provide four discrete stiffness characteristics.
This suspension system is based on a previously developed suspension that had only two state (open or closed) valves, which provided discrete damping characteristics. A thorough investigation of the older system proved that the system was capable of addressing the ride comfort versus handling compromise. The purpose of this study was to investigate whether the updated design could deliver improved performance and to recommend focus areas for future research initiatives.
The suspension system’s characteristics were determined experimentally by actuating the unit on a test bench. Results indicated that the unit produced the desired stiffness, low damping and response time characteristics. A mathematical model of the suspension unit was developed and validated against experimental data. The model was used in single degree of freedom simulations to investigate both passive and semi-active controlled performance. Preliminary results suggested that the suspension could be semi-actively controlled for improve ride comfort. However, the magnitude of improvements with continuous semi-active control, which includes a suitable response time, proved to be rather insignificant compared to the optimum passive suspension.
The findings proved that the suspension system was capable of successfully reducing the ride comfort versus handling compromise. Compared to the previous model, it is expected to deliver improved ride comfort and handling performance when implemented on the vehicle.
N. Wilken, "Experimental investigation of free-surface jet-impingement cooling by means of TiO2-water nanofluid"
Supervisor: Prof M. Sharifpur
Co-supervisor: Prof. J. P. Meyer
The exponential advancements in the field of electronics and power generation have resulted in increased pressure on the thermal management of these systems where the desire for enhanced heat transfer is prevalent. A technique for enhancing heat transfer that has gained sufficient attention over the past two decades is to suspend nano-sized metallic particles in a base fluid in order to enhance its thermophysical properties. Fluids produced in such a manner are commonly termed nanofluids. Due to the promising heat transfer capabilities of nanofluids, many industrial applications are beginning to implement these fluids in their thermal practices. One of the potential applications where nanofluids may be used which has received a great deal of research attention is jet-impingement heat transfer. Concerning the existing publications on nanofluid jet impingement, most works within the steady-state regime are limited to the cooling of Al2O3-water nanofluids, while transient studies do not account for cooling without the effects of boiling phenomena and for surfaces other than steel.
In this study, six particle volume fractions of TiO2-water ranging between 0.025 and 1% were prepared and characterised for appropriate cooling tests. The study was conducted within both the steady and transient states with the main objective of evaluating the thermal performance of the selected nanofluid and to determine the optimum particle concentration for jet-impingement cooling applications. Therefore, an experimental rig was designed and manufactured where a copper target surface of 42 mm was impinged upon by a 1.65 mm orifice nozzle at a non-dimensional nozzle-to-target height of 4. The results indicated that the use of nanofluids in impingement applications produced adverse effects, depending on the particle fraction considered.
With respect to the steady-state cooling tests, the copper surface was subjected to a constant heat flux of 145 watt and cooled by the different fluids at Reynolds numbers ranging between approximately 10 000 and 30 000. A maximum enhancement of 14.75% was observed in the measured Nusselt numbers, which occurred at a particle volume concentration of 0.05%. When increasing the volume fraction above 0.1%, unfavourable effects were observed for the heat transfer of the system in comparison with the base case tests of DI-water. Such trends were characterised by the trade-off between the enhancement in thermal conductivity and viscosity, both of which were increased with an increase in particle concentration. As for the effect of Reynolds number on the resulting thermal performance, a directly proportional relation was shown and could be described by the forced convection effect. The transient impingement tests showed that particle concentrations less than 0.1% produced an enhancement in cooling efficiency, while those of higher volume fractions showed negative effects. According to these tests, the maximum enhancement was also obtained at a volume fraction of 0.05% and produced an average cooling efficiency enhancement of 16%.
The results of the investigation clearly showed that the use of TiO2-water nanofluids in jet-impingement cooling applications produced thermal enhancement depending on the selected particle concentration.
C. Versteeg, "Pareto optimal comparison of numerical optimisation formulations for minimum-time and minimum end-state error motions on rigid and flexible robotic manipulators"
Supervisor: Prof D. N. Wilke
Numerous strategies have been established for robots and robotic arms to navigate physical space. For each strategy, a vast number of algorithms and formulations have been proposed over the years. The aim of this study is to propose a strategy to compare the various trajectory optimisation formulations with each other, to discern their performance and applicability on a specific task or a number of tasks. As we demonstrate, the proposed strategy can map performance between formulations, over various operating conditions, for a given task. The formulations are based on multi-objective optimisation principles that may treat or enforce criteria vastly differently for example; a criterion can be applied through either strong or weak enforcement. To compare formulations, a set of problems are optimised with the formulations, and the Pareto-optimal solutions are mapped to identify which formulation is optimal for a part of the solution domain.
Also, we propose a scalarisation of the Pareto front to assist the comparison of formulations in general, i.e. when a specific part of the solution domain is not important but the general performance of a formulation to reliably produce superior solutions. The Pareto front refers to the solution domain of all the non-dominated solutions, and the proposed scalar metric reflects the ease and robustness of the formulation to provide these optimised solutions. The proposed metric consists of a count of the number of solutions that are non-dominated over the entire Pareto front, which is denominated by the efficiency of the formulation, that is calculated from the number of function evaluations performed to solve the set of solutions.
This study looked beyond the effect of the differences in the formulations and also evaluated the effect of variations in the problem statement on the performance of the formulations. One such variation is the robotic system that could consist of rigid or flexible bodies, and another variation is the parametrisation domain that allows the movement to be prescribed either as joint positions or joint torques. To study the metric’s feasibility, a range of formulations were identified and used to optimise three basic trajectory optimisation problems, where the problems differed in complexity and dimensionality.
K. E. Dellar, "Clamped plate-style recuperator for a small-scale solar thermal Brayton Cycle using high-temperature sealant"
Supervisor: Dr W. G. le Roux
Co-supervisor: Prof J. P. Meyer
South Africa is plagued by rolling blackouts, and many citizens do not have access to electricity or clean water. A personalised micro-turbine power generation system presents a solution to this issue and may become as commonplace as a personal computer. With South Africa’s excellent solar direct normal irradiation (DNI) levels, a small-scale recuperated solar thermal Brayton cycle (STBC) shows enormous potential. However, a recuperator comprises up to 30% of the capital cost associated with a micro-turbine package and requires complex and costly manufacturing methods within a South African context. Thus, the objective of this research is to investigate a clamped plate-style recuperator that can be cost-effectively manufactured locally. Literature was consulted and criteria were outlined that a recuperator in a Brayton cycle should adhere to. To uphold these requirements, a counterflow plate-style recuperator is mandatory, and to combat complex manufacturing methods, a gasketed stacked-plate design, which requires a gasket material, was proposed. A sodium silicate-based sealant called Soudal Calofer is available locally and can withstand the operating conditions of an STBC. Experimental testing was carried out successfully on two small-scale versions of the proposed recuperator design. Testing showed that the physical construction was simple and cost-effective and the clamped plate-style high-temperature sealant combination worked well to form the recuperator core, facilitating an easy assembly and disassembly process. The construction sustained an airtight seal (Mark I) for the entire testing period at various pressures and high temperatures. Despite the occurrence of heavy soot-based fouling deposits during Test 1 due to incomplete combustion of the LPG as a result of the very low air mass flow rates, a mathematical model was able to match the values gathered from the testing. The data showed a cold-side effectiveness of 58.6% and a total pressure loss of 17.78%. For Test 2, a cold-side effectiveness of 82.5% and a total pressure loss of 11.48% were found for the recuperator core, which also validated the mathematical model. A case study was performed for the small-scale STBC. The results showed that the combination of a cold-side effectiveness of 84% and a total pressure loss of less than 5% could be attained when implementing the recuperator within the STBC for a channel height of 1 mm and width of 50 mm. Alternatively, if pressure loss is of less concern, a cold-side effectiveness of 89% could be achieved by increasing the total pressure loss to 19 kPa, which equates to an 8.8% pressure loss. It is recommended that a large-scale recuperator be built and tested to confirm the performance characteristics of larger mass flow rates and that the insulation of the unit be varied to determine its effects. Gasket geometry and the assembly method also need to be further researched to develop a uniform and consistent assembly technique that results in an airtight seal for every unit assembled. This may be achieved by regulating the amount of water added to the Soudal Calofer for thinning purposes to achieve a consistency which facilitates uniform application and by extended drying time to allow for the assembly to be completed, while not thinning the sealant so much as to lead to a seal failure. In conjunction, the clamping force distribution is critical to sealing the inner channel division. It is also recommended that the usable lifespan of such a recuperator be determined. Most crucially, thermal and pressure cycling must be investigated, especially where seal integrity is concerned.
H. Jordaan, "Thermal Analysis of a Feedwater Heater Tubesheet through Coupling of a 1D Network solver and CFD"
Supervisor: Prof PS Heyns
Co-supervisor: Dr S. Hoseinzadeh
A feedwater heater is a typical component in power plants which increases the cycle efficiency. Over the last decade, renewable energies have significantly developed and been employed in the power grid. However, weather conditions are inconsistent and therefore produce variable power. Fossil fuel power stations are often required to supplement the variable renewable energies, which increased the rate of power cycling to an unforeseeable extent over the past decade. Power cycling results in changes in the flow rate, pressure, and temperature of a feedwater heater’s inlet flows. In a tubesheet-type feedwater heater, these transients induce cycling stress in the tubesheet and failures due to thermal fatigue occur. The header-type feedwater is currently employed in high pressure applications as it is more resistant to thermal fatigue compared to the tubesheet-type. However, the tubesheet-type is more cost effective to construct and maintain. It would be advantageous if the cyclic thermal stresses in the tubesheet can be better analysed and alleviated to support the use of the tubesheet-type.
A detailed transient temperature distribution of the tubesheet is required to understand the thermal fatigue. Normally, engineers opt towards a full CFD to obtain such results. However, the size and complexity of a feedwater heater is immense and cannot be simulated practically solely using CFD spatial elements. This study developed a multiscale approach that thermally couples 1D network elements, CFD spatial elements, and macroscopic heat transfer correlations to reduce the computational expense substantially. The combination of the various selected techniques and the specific application of this methodology is unique. This approach is capable of obtaining the detailed transient temperature distribution of the tubesheet in a reasonable time, as well as include the effects of the upstream and downstream components within the network model. The methodology was implemented using Flownex and Ansys Fluent for the 1D network and CFD solvers, respectively. The internal tube flow was modelled using 1D network elements, while the steam was modelled with CFD. Thermal discretisation, mapping, and convergence were considered to create a robust methodology not limited to feedwater heaters only. Additionally, a method was developed to analyse flow maldistribution in tube-bundles using the coupled 1D-3D approach.
The implementation of the methodology consists of two parts, of which one is for development purposes, and the other serves as a demonstration. The development was done on a simple TEMA-FU heat exchanger which is representative of a feedwater heater. The methodology was tested by varying the primary fluid’s flow rates, changing the fluid media, and conducting transient simulations. The temperature distributions obtained were compared against a full CFD model and corresponded very well with errors less than 4%. A reduction in computational time of more than 40% was achieved but is highly dependent on the specific problem. Improvements to be made in future studies include the accuracy of the laminar case method and the stability of the flow maldistribution algorithm.
The methodology was demonstrated by applying it to an existing industrial feedwater heater. No plant data was available to use for input conditions and therefore were assumed. The steam in the DSH was modelled using 3D CFD elements and the tube flow with 1D network elements. The condensing zone’s heat transfer was approximated using an empirical correlation. A steady state case was simulated and the outlet temperatures corresponded well with the manufacturer’s data. The temperature distribution of the tubesheet and surrounding solids were obtained. Finally, assumed sinusoidal transient perturbations to the inlet conditions were imposed. It was evident that the thermal gradients of both sides of the tubesheet were misaligned which highlights the thermal lag and inertia that cause differential temperatures.
The 1D-CFD methodology was developed successfully with results that proved to correspond well, for a wide range of conditions, to full CFD. The methodology was applied and can be, in future work, validated with experimental results or extended by modelling upstream and downstream components in the network solver.
S. Baggeröhr, "A Deep Learning Approach Towards Diagnostics of Bearings Operating under Non-stationary Conditions"
Supervisor: Prof. P. Stephan Heyns
Co-Supervisor: Prof Daniel N. Wilke
Faults in bearings usually manifest as marginal defects that intensify over time, allowing for well-informed preventative actions with early Fault Detection and Diagnosis (FDD) protocols. Detection of the fault begins with capturing, for example, acceleration signals from a machine. Traditionally, handpicked descriptive statistical features (mean, RMS, skewness, kurtosis, etc.) or spectral diagrams obtained from these signals are then used for FDD. However, machine signals are often generated under non-stationary operating conditions of varying loads and speeds, requiring further intervention. More advanced signal processing techniques (spectral kurtosis, or cyclostationary analysis) are hence used to account for the non-stationarity of the signal. This is usually done by separating acceleration signals into deterministic and random components. Fault detection in bearings is possible by observing the random components of the signal.
A wealth of research has been invested in machine learning based techniques to circumvent the problems associated with non-stationary signals. Many of these methods require vast amounts of historical data to train. Machines typically spend most of their life operating in a healthy condition, therefore, most historical data is occupied with data that comes from a healthy machine condition, training these methods are difficult, due to the shortage of data from a machine running in an unhealthy condition. Furthermore, well-performing machine learning algorithms still require a domain expert to extract features that are known to be fault sensitive. Deep learning is a recent approach in data analysis whereby feature extraction is incorporated within the training of the algorithm. The algorithm is given the ability to find and extract its own features. The architecture of the algorithm allows for the extraction of complex hierarchical non-linear features. To the author's knowledge, no attempt has been made to make full use of the power of deep learning together with the known structure of bearing acceleration signals to perform FDD.
In this work, a bearing FDD methodology is developed using deep learning approaches. A model based on Generative Adversarial Networks (GANs) and Variational Autoencoders (VAEs) is used to learn a lower dimensional representation of an acceleration signal. A regularization strategy based on information maximization is used, which allows deterministic and random components of the signals to be learned separately. This representation is subsequently used to perform bearing FDD. The algorithm is trained in a completely unsupervised manner on exclusively healthy data and requires no pre-processing of that data. Furthermore, no auxiliary signals such as a shaft encoder, which contains information about the machine operating condition, is required for the algorithm to work. The methodology was tested on well-known benchmark datasets, and it was shown to be robust against non-stationary operating conditions. The algorithm can learn its own fault metric and by observing the trajectory of the signal representation, it is also able to diagnose the type of fault.
S. Machingauta, "An investigation into the benefits obtained by introducing Reliability Centred Maintenance in industrial organisations"
Supervisor: Prof J. L. Coetzee
Reliability Centred Maintenance (RCM) is a well-known maintenance process developed in the aviation industry. It has yielded great success and hence was adapted to be used in more industrial environments, such as the process developed Coetzee (2015) called ProaktivTM and the process developed by Moubray (1997) called RCM2. The RCM process is considered by many to be a very effective and comprehensive maintenance process that can, if implemented correctly, improve reliability and plant availability substantially.
However, many maintenance practitioners and maintenance experts who have used RCM believe that it is an overcomplicated process that is difficult to implement. In many cases the process is abandoned and left incomplete due to the amount of resources required and the slow results it delivers initially.
This dissertation investigates the benefits of implementing RCM on a mineral sizer at Sierra Rutile in Sierra Leone. In that regard, key performance indicators (KPI) of the sizer were recorded before and after the implementation of RCM. The main KPIs taken into consideration were:
This research centres mainly on the maintenance strategy improvement plans derived from the RCM process and the improvements to the production process that resulted from that exercise.
G. Janse van Vuuren, "Extracting Blade Condition Information from the Pressure Field around a Turbine Blade"
Supervisor: Prof. P.S. Heyns
Turbine stages are exposed to a variety of excitation sources in the power industry. The resulting forced vibration excitation of the blades may occur near a blade’s natural frequency. Blade vibration is an inevitable, inherent characteristic of turbines as the rotor blades travel through the trailing wakes of the upstream stator blades. Blade vibration can be worsened by other mechanisms such as pitting, corrosion fatigue and stress corrosion cracking commonly experienced in the power industry. Measuring turbine blade vibration allows for condition monitoring of the blades for damage. This is often coupled with finite element models of the blades or with computational fluid dynamic models of the flow field around the blades. These numerical methods, although well-established, lack the complexity of the true multiphysics phenomena within a turbine. As the blade vibration measurement techniques essentially capture blade vibration that is the result of fluid-structure interaction (FSI), blade vibration should be modelled as a coupled problem, but this is usually computationally expensive.
A rudimentary yet fundamentally correct numerical model of a turbine stage is thus required to model the fluid-structure interaction while minimising computational costs and retaining accuracy. If this can be achieved and blade health information can be detected in the flow field within the model, further analyses can then be put forth to predict blade health over time. The main objective of this study is to investigate the extent to which blade condition information can be extracted from a transient three-dimensional two-way FSI model of a blade passage containing a single rotor and stator blade. An experimental single-stage test turbine with five stator and five rotor blades is used to gather experimental data. The experimental data is used to validate the FSI model in the time and frequency domains. Two rotor blade assemblies were tested with the first configuration consisting of five healthy blades, and the second configuration consisting of four healthy blades and one damaged blade. All simulations are performed at constant rotational speeds for one single revolution of the rotor. Structural damping of the rotor blades is not considered. All numerical simulations are carried out using the commercial multiphysics software package of Ansys R2 2019 and the explicit use of CFX for the CFD simulations.
The results of the FSI model compare well to the experimental results when considering the simplifying assumptions made for the development of the numerical model. The first natural frequency and blade passing frequencies of the healthy and damaged blades can be extracted from the pressure field of the FSI model at critical speeds. Similar findings were observed in the fluid mesh deformation time profiles around the blade tips. Blade excitation is strongly coupled to engine- ordered vibration frequencies, specifically the blade passing frequencies and its first harmonic. Challenges are realised when modelling a single damaged blade that is part of a larger, healthy assembly of rotor blades. The compromise of reducing computational effort is highlighted here.
However, very promising results pertaining to blade condition information extraction from the two- way FSI model pressure field are obtained. These results have established a foundation on which a more complex FSI model can be built and coupled with a fatigue or remaining useful life study. It is suggested that future work should include structural damping of the rotor blades, a larger computational domain, and investigation of longer simulation times.
R.D. Bennet, 2020, "Vehicle Stability Analysis of Transport Latencies and Dropped Packets on Wireless Communication of an Off-site Steering Controller"
Supervisor: DR TR Botha (supervisor)
With autonomous vehicles being introduced around the world the possibility of controlling these vehicles from off-site locations presents itself as an opportunity to increase occupant/user safety as well as system efficiency. This applies not only to passenger vehicles but to vehicles in the industrial sector as well. To safely implement off-site control requires an understanding of how vehicle control is affected by larger transport latencies and dropped packets, which are inherent properties of a wireless network. This study aimed to shed some light on these effects on lateral control of a passenger vehicle when the vehicle's controller was placed off-site. In this study a Linear Quadratic Self-Tuning Regulator (LQSTR) controller, making use of a vehicle model based on auto-regressive theory describing the relationship between the yaw rate and steer angle of a vehicle, was used. A simulation study showed that the controller was able to control the vehicle through a Double-Lane-Change (DLC) -manoeuvre at vehicle speeds of at least 120 km/h while maintaining very low path tracking errors. With only minor alterations made to the controller's parameters the system was then subjected to transport latencies and packet drops between the vehicle and its controller to simulate off-site vehicle control. It was found that control of the vehicle was lost when latencies between the vehicle obtaining sensor data and a subsequent control action being realised were in excess of 240 ms and packet drop percentages reached 40% to 50%. If no packets were dropped, i.e. 0%
packet drop, transport latencies could be as high as 460 ms before control of the vehicle was lost entirely.
S.L.Robbines, 2020, "The feasibility of rotor fault detection from a fluid dynamics perspective"
Supervisor Prof P.S. Heyns
Co-supervisor: Dr J.A. Heyns
The majority of condition monitoring techniques employed today consider the acquisitioning and analysis of structural responses as a means of profiling machine condition and performing fault detection. Modern research and newer technologies are driving towards non-contact and non-invasive methods for better machine characterisation. In particular, unshrouded rotors which are exposed to a full field of fluid interaction such as helicopter rotors and wind turbines, amongst others, could potentially benefit from such an approach. Current literature lacks investigations into the monitoring and detection of anomalous conditions using fluid dynamic behaviour. This is interesting when one considers that rotors of this nature are typically slender, implying that their structural behaviour is likely to be dependent on their aerodynamic behaviour and vice versa. This study sets out to investigate whether a seeded rotor fault can be inferred from the flow field. Studies of this nature have the potential to further a branch of condition monitoring techniques. It is envisaged that successful detection of rotor anomalies from the flow field may aid in better distinction between mass and aerodynamic imbalances experienced by rotor systems. Furthermore, the eventual goal is to better describe the adjustments made to helicopter rotor systems when performing rotor track and balance procedures. In this work, time-dependent fluid dynamic data is numerically simulated around a helicopter tail rotor blade using URANS CFD with the OpenFOAM software package. Pressures are probed at locations in the field of the rotor and compared to results attained in an experimental investigation where good correlation is seen between the results. A blade is modelled with a seeded fault in the form of a single blade out of plane by 4°. Comparisons are drawn between the blade in its ‘healthy’ and ‘faulty’ configurations. It is observed that the fault can be detected by deviations in the amplitudes of the pressure signals for a single revolution at the probed locations in the field. These deviations manifest as increases in the frequency spectrum at frequencies equivalent to the rotational rate (1 per revolution frequencies). The results described are assessed for their fidelity when the pressure is probed at different locations in the domain of the rotor. Deviations in the pressure profiles over the surface of the blades are also seen for the asymmetric rotor configuration, but may prove too sensitive for practical application.
K. Govinder, 2020. "THEORETICAL ANALYSES AND DESIGN, CONSTRUCTION AND TESTING OF A FLOW LOOP FOR THE STUDY OF GENERALISED FORCED AND NATURAL CONVECTION BOILING HEAT TRANSFER PHENOMENA ON TYPICAL LIGHT-WATER UCLEAR REACTOR FUEL PIN CONFIGURATIONS"
Supervisor: Prof J.F.M. Slabber
Co-Supervisor: Prof J.P. Meyer
In a worldwide quest to improve light-water-type nuclear power reactor safety, a more accident- tolerant fuel is required. In partial fulfilment of this initiative, an alternative fuel cladding to the currently used zirconium alloy-based materials had to be investigated and qualified. Silicon carbide was proposed as a probable replacement material. Accordingly, an experimental program was established at the University of Pretoria to conduct comparative boiling heat transfer studies of silicon carbide cladding tubes versus typical zirconium alloy-based nuclear fuel cladding (Zircaloy-4® was used for comparisons in this study). For the study, a purpose-built flow loop test facility, the University of Pretoria – Thermal-Hydraulic Flow loop Cell (UP- HFC), was subsequently designed, built and commissioned.
The test rig used water as the coolant and had capabilities for forced convection to be selected either vertically up or down, or horizontally through the test section. The test-clad tubes were internally heated with electrical resistance-type cartridge heaters, which were located and sealed within the clad tube bores to form heated pins. The arrangement of the pins in the test section allowed for external coolant flow over the clad tube surfaces. Operating test conditions in the test section were configured according to subcooled or saturated regimes, forced or natural convection, with the test section pressure being adjusted from ambient to around 200 kPa, and mass flux spanning from around 450 to 1 800 kg/m 2 .s. Observation windows in the heated test section provided for visual inspection and digital camera recording of the boiling processes. This dissertation reports on the evaluation and design criteria used for the building of the flow rig and the findings of the initial heat transfer tests on both tube materials. The set-up criteria used for the tests included the mean test section pressure maintained at around 200 kPa, four flow rate settings, three different bulk coolant inlet temperatures, and six heater power settings. The combination of these variables provided 192 test data points. Preliminary results indicated that although the silicon carbide had a higher thermal conductivity rating than Zircaloy-4®, it, however, produced marginally lower heat transfer capacity than Zircaloy-4®. The conducted heat transfer tests verified that the theoretical design analysis of the UP-THFC was satisfactory.
Copyright © University of Pretoria 2024. All rights reserved.
Virtual Campus
Get Social With Us
Download the UP Mobile App