No abstract available. / Thesis (Ph.D.)-University of Natal, Durban, 1983.

No description available.

The benefits of maintenance have been known for centuries, and throughout history these benefits have been the basis for much discussion. Recently such benefits have been formally proposed in terms of scheduled or cyclical maintenance programs. These programs have been developed for important historic buildings; yet the same benefits can be applied to any house. This concept is the basis of this thesis.The thesis begins with a discussion of the obvious need for some maintenance program and describes the recent interest of local governments in some form of maintenance planning, discussing the benefits of an organized approach to maintenance. Next it briefly reviews the beginnings and history of maintenance theories and is followed by a discussion of maintenance as a level of conservation of building. Basic statistics about homeowners and the development of the survey are addressed, which leads to the presentation of compiled results of the survey and a interpretation of the data. The conclusion indicates a direction for the development of an organized maintenance plan.

Pressures from environmental agencies contribute to the challenges associated with the disposal of waste tyres, particularly in South Africa. Recycling of waste tyres in South Africa is in its infancy resulting in the historically undocumented and uncontrolled existence of waste tyre stockpiles across the country. The remote and distant locations of such stockpiles typically complicate the logistics associated with the collection, transport and storage of waste tyres prior to entering the recycling process. In order to optimize the logistics associated with the collection of waste tyres from stockpiles, useful information about such stockpiles would include estimates of the types of tyres as well as the quantity of specific tyre types found in particular stockpiles. This research proposes the use of computer vision for categorizing individual tyres and estimating the number of visible tyres in tyre stockpile images to support the logistics in tyre recycling efforts. The study begins with a broad review of image processing and computer vision algorithms for categorization and counting objects in images. The bag of visual words (BoVW) model for categorization is tested on two small data sets of tread tyre images using a random sub-sampling holdout method. The categorization results are evaluated using performance metrics for multiclass classifiers, namely the average accuracy, precision, and recall. The results indicated that corner-based local feature detectors combined with speeded up robust features (SURF) descriptors in a BoVW model provide moderately accurate categorization of tyres based on tread images. Two feature extraction methods for extracting features for use in training neural networks (NNs) for tyre count estimations in tyre stockpiles are proposed. The two feature extraction methods are used to describe images in terms of feature vectors that can be used as input for NNs. The first feature extraction method uses the BoVW model with histograms of oriented gradients (HOG) features collected from overlapping sub-images to create a visual vocabulary and describe the images in terms of their visual word occurrence histogram. The second feature extraction method uses the image gradient magnitude, gradient orientation, and edge orientations of edges detected using the Canny edge detector. A concatenated histogram is constructed from individual histograms of gradient orientations and gradient magnitude. The histograms are then used to train NNs using backpropogation to approximate functions from the feature vectors describing the images to scalar count estimations. The accuracy of visible object count predictions are evaluated using NN evaluation techniques to determine the accuracy of predictions and the generalization ability of the fit model. The count estimation experiments using the two feature extraction methods for input to NNs showed that fairly accurate count estimations can be obtained and that the fit model could generalize fairly well to unseen images.

M.Ing. (Civil Engineering) / The: structural classification of granular base pavements by means of resilient deflection bowl parameters measured by the Impulse Deflection Meter and application of measured deflections at network level were investigated. The Investigation consisted of a mechanistic analysis and an analysis of observed field results. Curves for structural classification of granular base pavements obtained by regression are presented. The Curves obtained by the analysis of field observations compare well with local experience and theoretical failure criteria. It was found that the mechanistic model used in this study did not yield satisfactory results when simulating deflection behaviour through linear elastic programs. The parameters recommended for future use are Maximum Deflection (Y-rnax), Base Layer Index, or DLl (formerly SCI), Middle Layer Index, or MLl (formerly DOl) and Radius of Curvature. The use of deflection measurements at network level was also investigated. A method for calculating a Structural Stiffness Index to be incorporated into Pavement Management Systems as a Structural Indicator is proposed. Recommendations regarding the density of testing and identification of roads for testing at network level are also made. Several factors influencing deflections and structural bearing capacity are discussed and some recommendation." are made. The main factors influencing deflections and predicted pavement structural capacity were shown to be Temperature, Seasonal Influences and Condition of Surfacing. Further research into the specific influence of these factors on deflections and hearing capacity is needed.

M.Tech. (Mechanical Engineering) / The dissertation documents the design project of a low cost tractor prototype to be manufactured locally. The design and manufacturing of a South African tractor is part of a larger project namely “Roads for Africa” initiated by the local company Terragrader (Pty) Ltd. The design parameters of the tractor were established based on the power needs of a dirt roads refurbishing machine the “Terragrader” manufactured by the same company. Although the tractor was designed to power the Terragrader machine, it can successfully be used to power different agricultural machines or transport. The designed South African tractor has to be versatile, easy to adapt to different working conditions. The design aim is to produce a tractor able to provide adequate power at a cost of about R500 000, whereas the market offers similar machine at a much higher price starting from about R800 000. The concept of a South African tractor is not new but due to import options, there was never a known / well established South African brand. Due to the shrinkage of the South African mechanical manufacturing industry by nearly 40 % in the last years, a successful tractor design and manufacturing will contribute toward the much needed job creation. Due to its complexity the project is a combination of newly designed parts and outsourced components. The integration of design establishes the layout of the tractor with the aid of virtual 3D solid modeling.

M.Com. (Business Management) / The Rolling Stock Department (Rohnat) within Spoomet is responsible for the maintenance and repair of all classes of locomotives. One of the biggest problems at maintenance depots of Rolling Stock is to decide which faulty locomotives, withdrawn from service, must get preference for repairs. This problem is compounded at depots where different classes of locomotives are maintained and repaired. PURPOSE The purpose of this study is to determine critical factors and to create a decision model that can be used for project management at maintenance depots. This decision model must help maintenance depots of Rohnat to determine in which sequence different types of locomotives must be repaired to the advantage of Rohnat and Spoomet. METHOD OF STUDY The most suitable model that could be used to determine the repair sequence of locomotives was the MADM model. In this study, brain storming sessions and individual discussions were held with different Spoomet entities to determine critical factors that could be used in this model. The principles of the MADM model were explained to all parties concerned, whereby critical ABSTRACT factors affecting this model were discussed. Questionnaires were completed after the iscussions to confmn the critical factors that could be used in the MADMmodel. MADM - DECISION MODEL The MADM model is unique in the sense that financial and non-financial factors can be combined . The non fmancial factors which contribute to corporative success and which are left out in other models are however included in this model. The critical factors are further divided into two groups namely quantitative and qualitative. A conceptual model has been prepared in this study to give a better perspective of the workings and arrangements of weights, values and risks of this model. SUMMARY The most suitable model that could be used to determine the repair sequence of locomotives was the MADM It was possible to determine different critical factors that could be used in this model. Different Rolmat depots could use this model, but with different weights for each critical factor to suit their requirements. Critical factors could be added to or removed from this model to suit changing requirements.

Modern automobiles are painted with basecoat, technology which is either metallic, solid colour or pearlescent. This requires protection from chemicals, scratching, weathering and UV light by applying a protective top coat (clearcoat) over the basecoat. For the clearcoat to cure into a hard protective shell it undergoes an irreversible crosslinking process. This usually takes place over the first four to five hours, depending on the formulation and weather conditions. The speed of crosslinking can be enhanced by temperature. Pot life is important as it can affect the overall quality of the painted surface. If crosslinking occurs too quickly, before the clearcoat is applied onto the surface, the clearcoat cannot be used to produce a good quality finish. The “expired” mixture is thus discarded. If used, the quality of the finished product cannot be guaranteed to last, and the paintwork may have to be redone. This often means removing the underlying paint and primer as well with the clearcoat film. Besides the time lost, the discarded clearcoat mixture often lands in the landfill, polluting ground water and the environment. It is thus important from the point of view of both the environment preservation and waste management, that as much clearcoat as possible is used without being wasted. It was proven in an earlier study (BSc Hon Formulation Science Treatise, 2011) that adding eugenol to a clearcoat mixture after crosslinking had started could reduce its viscosity, which is an indicator of crosslinking progress. Crosslinking subsequently resumed at a lower rate than in traditional blends. If stored away from oxygen and high temperatures, this blend could maintain optimum viscosity indefinitely. In this follow up study an optimum formulation was developed using D - optimal experimental design. It sought to extend the pot life to avoid waste to spray painters while saving the environment from pollution. The formulation that gave the desired viscosity after five hours of pot life was adopted. It was hypothesised that the optimum formulated clearcoat mixture would have a longer pot life than its traditional counterparts, and that it would perform just as well as the traditional clearcoat mixtures. To study the rate of crosslinking (disappearance of functional groups and appearance of the urethane bond), FTIR spectrometry was performed on a mixture produced from the optimized formula in comparison to that of a traditional mixture (the control). The rate of disappearance of functional groups was found to be lower in the eugenol mixture than in the control mixture. After six hours, eugenol was added into the control mixture, and this seemed to reduce the viscosity with the re-emergence of functional groups in the mixture. After 24 hours of crosslinking, an FTIR scan was done on the solid sample and this revealed that the eugenol mixture had crosslinked fully, with no detectable functional groups in the sample.

Ideally, pavement management is a process of sequential decisions on a network of pavement sections. The network is subjected to uncertainties arising from material variability, random traffic, and fluctuating environmental inputs. The pavement manager optimizes the whole system subject to resource constraints, and avoids sub optimization of sections. The optimization process accounts for the dynamics of the pavement system. In addition to objective data the manager seeks information from a number of experts, and considers selected social-political factors and also potential implementation difficulties. Nine advanced schemes that have been developed for various pavement administrations are compared to the ideal. Although the schemes employ methods capable of handling the pavement system's complexities in isolation, not one can account for all complexities simultaneously. Bayesian decision analysis with recent extensions is useful for attacking the problem at hand. The method prescribes that when a decision maker is faced with a choice in an uncertain situation, he should pick the alternative with the maximum expected utility. To illustrate the potential of Bayesian decision analysis for pavement management, the author develops a Markov decision model for the operation of one pavement section. Consequences in each stage are evaluated by multi-attribute utility. The states are built of multiple pavement variables, such as strength, texture, roughness, etc. Group opinion and network optimization are recommended for future research, and decision analysis suggested as a promising way to attack these more complex problems. This thesis emphasizes the utility part of decision analysis, while it modifies an existing approach to handle the probability part. A procedure is developed for Bayesian updating of Markov transition matrices where the prior distributions are of the beta class, and are based on surveys of pavement condition and on engineering judgement. Preferences of six engineers are elicited and tested in a simulated decision situation. Multi-attribute utility theory is a reasonable approximation of the elicited value judgements and provides an expedient analytical tool. The model is programmed in PL1 and an example problem is analysed by a computer. Conclusions discuss the pavement maintenance problem from the decision analytical perspective. A revision is recommended of the widespread additive evaluation models from the standpoint of principles for rational choice. Those areas of decision theory which may be of interest to the pavement engineer, and to the civil engineer in general, are suggested for further study and monitoring. / Applied Science, Faculty of / Civil Engineering, Department of / Graduate

M.Ing. / Please refer to full text to view abstract