No description available.

No description available.

Thesis (M.Sc.Eng.)-University of KwaZulu-Natal, Durban, 2008.

To provide an adequate level of safety at grade crossings, Transport Canada has allocated several millions annually to prevent collisions at grade crossings through the implementation of countermeasures, such as train-actuated warning devices and track devices. Railway companies and provincial agencies have also provided additional support to improve safety at highway-railway grade crossings. One of technical challenges in estimating safety effect of countermeasures at highway-railway grade crossing is an extremely rare occurrence of collisions. Given that the collision process is random with significant variation over time and space, it is hard to judge whether a specific crossing is safe or safer than other crossings solely based on the number of collisions in a given year. Decision makers are also required to make difficult decisions on safety investment accounting for uncertainty in effectiveness of countermeasures. The level of uncertainty is even higher when there is lack of observed collision data before and after the implementation of specific countermeasures. This study proposes a Bayesian data fusion method which overcomes these limitations. In this method, we used previous research findings on the effect of a given countermeasure, which could vary by jurisdictions and operating conditions, to obtain a priori inference on its expected effects. We then used locally calibrated models, which are valid for a specific jurisdiction, to provide better estimates of the countermeasure effects. Within a Bayesian framework, these two sources were integrated to obtain the posterior distribution of the countermeasure effect. The outputs provided not only the expected collision response to a specific countermeasure, but also its variance and corresponding probability distribution for a range of likely values. Some numerical examples using Canadian highway-railway grade crossing data illustrate how the proposed method can be used to predict the effects of prior knowledge and data likelihood on the estimates of countermeasure effects.

No description available.

Problems can be solved using existing knowledge and methods derived from past experiences; and in building design, where buildings are sufficiently similar to those already built, this process can be optimised by creating standardised solutions to common problems. There is significant demand for specialist engineers who can apply these standardised solutions to established problems quickly and accurately; but novel designs generate entirely new problems for which established solutions are not always applicable. Generalist engineers working on novel designs must first define the problems before they can develop options and if necessary, create optimised solutions. Fire safety engineering (FSE) is the process of achieving fire safety in our built environment. The field requires both specialists trained in current practice and generalists skilled in creative and critical thinking. Current fire safety engineering education is mostly aimed at producing specialists, yet there is growing demand for generalists in high-end architecture, hindered by a lack of generalist education. Current education literature in FSE explains in detail what to teach, however they do not explain how to motivate students to learn what is taught; how to create the ‘need to know’ - the purpose that drives learning. The purpose can either be intrinsically motivating (i.e. the subject is interesting) or extrinsically motivating (i.e. if you don’t learn it then you will fail the exam). The former is sustained by autonomy and choice; the latter is sustained by control. Control increases the likelihood that the predicted outcome will be realised, but by definition reduces the likelihood of realising any other outcome, including potential innovation.Initially a study was created to test the effects of creating an autonomous learning environment within a traditional lecture-based ‘fundamentals’ course at the University of Edinburgh. This study, along with observations at a range of US universities led to the formation of an overarching theory of education. Ultimately, purpose is the goal students strive to achieve; autonomy creates the opportunity to think and learn independently; and structure provides the constraints that converge students towards an optimised result, supported by sound evidence and reasoning. Thus the key to generalist education was to provide purpose, autonomy and structure (PAS) in that order. The PAS concept was trialled at EPFL (Switzerland) and the participating students, with no prior knowledge of fire engineering, produced work of exceptional quality. In summary, the present study offers an observational validation that Purpose, Autonomy & Structure (PAS) can be used to effectively support the generalist way of thinking and although the examples given in this paper are related to fire safety engineering (due to the need for generalists in that field), the qualitative evidence on which the conclusions are based is not subject-specific, implying that the PAS methodology could be applied to other disciplines.

This thesis aimed to evaluate the nature of safety climate; the component parts, the antecedents and the consequences. This work is based in the maritime domain, a sector with rising accident rates and many potential risk factors (such as managing a remote multicultural workforce on a high risk moving site), as identified through a literature review. An overview of the factors involved in human error within the maritime industry at an organisational and individual level showed that one contributory factor, which had been comparatively under-researched is safety climate. The review of safety climate literature post 2001 demonstrated wide variation in measurement across studies, although there was an element of consistency in terms of relationships with outcome variables and also in the topics measured as safety climate. Aligned with previous reviews, the role of management commitment to safety was highlighted, as expressed through its prevalence in studies of safety climate. A safety climate measure for the shipping sector, which loosely mapped the most common themes identified through the review, was designed. Study 1, a study of safety climate, safety behaviour and national culture, sampled 1067 seafarers from one company's 46 oil and gas shipping vessels and used this questionnaire to measure the relative contribution of management commitment benchmarked with other safety predictors when predicting behaviour. The results demonstrated that the sole predictor of safety behaviour was workers' perceptions of management commitment to safety. It is argued that the practice of including numerous other safety-related indicators (such as communication and satisfaction) in the prediction of behaviour is empirically unjustified and also theoretically misaligned. The role of grouped worker safety behaviour was also evaluated and group safety behaviour was posited as a mediator of the relationship between group safety climate and objective safety performance (as measured through company accident rates). Although there was a significant relationship between group safety behaviour and accident rates, behaviour did not act as the mediator in the relationship between safety climate and safety performance. Having illustrated that management commitment has a focal role in safety climate, it would appear logical to postulate a role for leadership in shaping the perceptions of managementc ommitment. This is also consistent with with previous research, which has demonstrated relationships between safety leadership and safety climate, although leadership for safety is still an area of research where little is known. A review of the leadership literature illustrated conceptual and levels of analysis issues in leadership. The second study, also with a seafaring sample from one company (n=447, across 27 ships), used the Multifactor Leadership Questionnaire (Avolio & Bass, 2004) to examine the relationship between leadership, safety climate (using Zohar & Luria, 2005 and Zohar & Luria, 2000 safety climate scales of group and organisational safety climate) and performance-both in safety and non-safety outcomes. The results demonstrated support for the role of transformational and passive leadership in shaping safety climate. Transactional leadership did not contribute significantly to the prediction of safety climate. Levels of analysis issues were considered in the research and leadership is proposed as a group-level construct with individual differences in outcomes, which aligns the findings of the second study with theoretical considerations in the literature.

Abstract This thesis measures safety climate in a sample of Scottish acute hospitals. It demonstrates how staff perceptions related to safety issues are linked to their safety behaviours and also to the consequences, both for the workers and the patients. Following a review of the industrial and healthcare safety climate literatures, a theoretical model was proposed to investigate the underlying mechanisms between safety climate and safety outcomes. Based on this review, the Hospital Survey on Patient Safety Culture (HSOPSC) was selected as part of a questionnaire to measure safety climate, safety behaviours and safety outcomes. A total of 1969 clinical staff from seven Scottish acute NHS hospitals were surveyed. The psychometric analysis, using EFA and CFA, showed that the original 12 factor structure of the HSOPSC scale was replicated. A focus group study (n = 25) was conducted in two of the hospitals to extend the survey findings. The qualitative data supported the theoretical model proposed based on the literature review by demonstrating the role of managerial practices on safety-related issues. The group discussions further contributed to a wider conceptualization of safety culture by illustrating the multi-level perspective of staff on safety-related issues, including both the external influences and the individual factors. Using structural equation modelling on the same quantitative data set, managerial aspects of safety climate were examined in relation to safety outcomes (safety behaviours, worker and patient outcomes). Results demonstrated the effects of managerial commitment to safety at hospital and unit level on safety outcomes. It also showed the intervening role of safety compliance and safety participation between supervisory practices and self-reported injuries, both for workers and patients. Overall, this thesis provided a psychometrically robust safety climate measurement tool tested in Scottish acute hospitals, and showed the influence of safety-related managerial activities at different levels of the organization on safety outcomes for workers and patients separately.

Highly flammable hydrocarbon liquids are involved in a high percentage of building fires, whether those fires are accidental or incendiary in origin. Their mere presence is often taken as proof of a particular fire cause by some investigators despite their limited knowledge of the behaviour of the vapours from these fuels as they spread and diffuse. They are sometimes assumed to vaporize completely and instantly upon exposure and to diffuse uniformly through any compartment. The available models address large scale spills in ambient conditions of sun and wind, which do not apply to typical building fires. This study addressed the problem of modelling the spread of vapours from small-scale (less than four litre) spills of highly flammable liquids by means of a series of overlapping and complementary experiments, all of which dealt with the conditions found in most interior building fires (moderate temperatures, still air, and no sun). It was determined that the surface area produced by a given quantity of liquid could be predicted for smooth, flat floors whose surfaces could be classified as non-porous (vinyl or painted wood), semi-porous (unfinished concrete or wood), or porous (carpet or sand). The type of surface also controlled the evaporation rate (per unit area of the pool). Evaporation rates from surfaces such as carpet saturated with pentane were 1.5 times the rate for a free-liquid pool at the same temperature. A granular substrate such as sand produced a pentane evaporation rate twice that of a pentane liquid pool. This effect is not related to the roughness of the surface itself, but rather to the capillary drive within the matrix. Such a drive is stronger for granular matrices with a small void space (high packing density) and lower for those with larger void space. The size of the pool also controls the evaporation rate (the mass loss rate per unit surface area). Smaller pools (0.05 - 0.1m diameter) exhibit much higher rates than do the larger ones (0.3m) in this study. This is due to the enhanced evaporation due to lateral flow of vapours from the edges of the pools. Larger pools have a large central quiescent area that does not contribute to the overall evaporation. Smaller pools have no such quiescent area and a higher initial rate. There are also predictable losses due to pouring and splashing of volatile fuels that are closely related to the vapour pressure of the liquid involved. Vertical diffusion of n-pentane and hexane vapours is very slow when the vapours are being generated by evaporation from a pool. The heat lost to evaporative cooling results in a pronounced thermal gradient in the atmosphere above a pool that suppresses the vertical diffusion. The diffusion rates of pentane, hexane, and octane vapours can be predicted and the height at which an ignitable vapour/air mixture is present can be calculated. The vapours also exhibit a pronounced advective flow which spreads the vapours in a viscous, laminar fashion. The spread rate of this advective flow can be calculated and agrees well with experimental data. The evaporation of n-pentane, hexane, and n-octane were found to be predictive of the evaporative behaviour of petrol and camping fuels, two of the consumer products more commonly encountered in fires. Petrol, with its high concentration of pentane-like hydrocarbons, evaporates at the same rate as does n-pentane, at least for the first 10 -15min. Camping fuels are dominated by hexanes and their evaporative behaviour is very similar to that of the hexane studied in detail here. Octane contributes very little combustible vapour at typical room temperatures due to its very low evaporation rates at these temperatures. The behaviour of the flame propagation in vapour/air mixture layers is predictable. Layer ignition is found to produce some characteristic features that may be observed by a witness to the fire or that may produce burn patterns that survive the fire to be found by a diligent investigator. Unfortunately, estimates of the quantity of flammable liquid present and its distribution prior to the fire cannot be reliably made by examination of the burn patterns on carpet or floors after the fire, particularly if the fire was not suppressed for some time after ignition. Finally, an operational model based on these findings is offered for the use of fire investigators. This model, while limited to incidents in closed compartments with no mechanical ventilation and limited activity, offers a means by which the physical distribution of ignitable vapours can be predicted as it varies with time. This enables the investigator to explore the viability of various hypotheses about the quantity and distribution of flammable liquids prior to a fire, the relative location (both vertical and horizontal) of a potential ignition source, and, most importantly, the time factors involved in the evaporation of a flammable liquid and distribution of its vapours.

Situé dans le contexte général de la sécurité routière, et plus particulièrement axé sur la sécurité des véhicules légers (VL) en virages, ce travail de thèse a pour objet de proposer une méthodologie fiabiliste de prédiction de trajectoires à risque, basée sur le traitement statistique et la modélisation probabiliste de trajectoires réelles de VL en virages. La première partie du travail concerne la construction de modèles probabilistes simples et robustes représentatifs des trajectoires réelles observées. Ces modèles sont des transformées de processus aléatoires scalaires normalisés du second ordre, faiblement stationnaires, ergodiques et non gaussiens, et permettent de décrire de façon réaliste la variabilité aléatoire observée du triptyque Véhicule-Infrastructure-Conducteur. Ils permettent aussi, par construction, de s'affranchir d'éventuelles difficultés dans l'alimentation des paramètres dominants qui les gouvernent. La seconde partie est consacrée au développement et à la mise en oeuvre d'une stratégie fiabiliste destinée à associer un niveau de risque à chaque trajectoire en entrée de virage. Basée sur l'emploi conjoint de méthodes probabilistes pour la modélisation des incertitudes, fiabilistes pour l'évaluation des niveaux de risque et statistiques pour la classification et le traitement des trajectoires, cette approche est une réponse réaliste au problème posé. De par sa conception et ses possibilités, la méthodologie fiabiliste proposée est une contribution significative au développement de procédures d'alerte destinées à réduire notablement le nombre d'accidents en virages. / This PhD thesis tackles the general context of road safety, focussing on the safety of light vehicles (LV) in bends. A reliability engineering methodology is proposed to predict dangerous trajectories, based on the statistical processing and probabilistic modelling of actual trajectories in a bend. In the first part of this work, simple and robust probabilistic models are built to describe trajectories measured in an instrumented bend. The models are transforms of scalar normalized second order stochastic processes which are slightly stationary, ergodic and non-Gaussian. They offer a realistic description for the observed random variability of the Vehicle-Infrastructure-Driver system. They also inherently circumvent possible difficulties in the identification of the dominant parameters which control the system. The second part of this work is devoted to the development and implementation of a reliability engineering strategy intended to associate a risk level to each trajectory at a bend entry. Based on the joint use of probabilistic methods for modelling uncertainties, reliability engineering for assessing risk levels and statistics for classifying and processing the trajectories, this approach provides a realistic answer to the tackled problem. From its design and its possibilities, the proposed reliability engineering methodology constitutes a significant contribution to the development of warning procedures the deployment of which are expected to notably reduce the number of accidents in bends.