This thesis is concerned with the safety of industrial controllers which incorporate software. Software safety is compared with software reliability as a means of discussing the special concerns of safety. Definitions are given for the terms hazard, risk, danger and safe. A relationship between these terms has been attempted and the philosophy of safety is discussed. A formal definition of software safety is given. The factors influencing the development of software are examined. The subjectivity of safety is discussed in the context of safety measurement being a conjoint measurement. Methods of assessing the risk resulting from the use of software are described along with a discussion on the impracticability of using state transition diagrams to isolate catastrophic failure conditions. Categories of danger are discussed and three categories are advanced. The structuring of the software for safety is discussed and the principle of using safety modules and integrity locks is proposed. In discussing the reasons for errors remaining present in the software after testing two methods of measurement are suggested; Plexus and Fallibility Index. The need to declare variables is discussed. An experiment involving 119 volunteers was conducted to examine the influence of the length of variable names'on the correct usage. It was found that variables with a character length of 7 have a better probability of correct interpretation than others. The methods of assessing safety are discussed and the measurements proposed were applied to a commercially available product in the form of a Software Safety Audit. It is concluded that some aspects of the safety of controllers incorporating software can be quantified and that further research is needed.

Les sorties de route en virage constituent une part importante des accidents de la route en France, causes d'environ 30 % des accidents mortels pour 2008. Ces accidents se produisent principalement sur route secondaire ce qui montre le besoin du conducteur d'être assisté dans sa tâche de lecture et de négociation du virage. L'objectif de ce travail est le développement d'une méthodologie d'évaluation du risque encouru par le conducteur lors du franchissement d'un virage afin de mettre en place sur les itinéraires routiers un système d'alerte déclenchée en cas de situation potentiellement dangereuse. La méthodologie menant à l'évaluation du risque est basée sur l'emploi des méthodes probabilistes, permettant de prendre en compte fidèlement les incertitudes inhérentes au conducteur, au véhicule et à l'infrastructure. A l'entrée du virage, un dispositif de mesure renseigne sur la position latérale et la vitesse du véhicule arrivant. Une famille de trajectoires de passage en virage est alors stimulée à partir d'un modèle dynamique de véhicule dont certains paramètres d'entrée dépendent des trajectoires mesurées sur le trafic réel. Des indices de risque associés à des critères de sécurité sont ensuite évalués par des méthodes fiabilistes. Les résultats obtenus sont la détermination et la hiérarchisation des paramètres influents sur les critères de sécurité, ainsi que l'évolution des indices de risque en fonction des conditions initiales en entrée de virage. Les applications réalisées dans le dernier chapitre démontrent le potentiel de la méthodologie fiabiliste proposée et son intérêt dans le domaine de la sécurité routière. / Roadway departure while cornering constitutes a major part of car accidents in France, accounting for nearly 30 % of casualties in 2008. Most of them occur on secondary roads and reveal that drivers need assistance for a safe curve negociation. The subject of this work is the development of a methodology to access the roadway departure risk in order to implement in the road infrastructure warning devices alarming the driver in case of potentially dangerous situations. The application of probabilistic methods for the risk assessment allows an accurate representation of the uncertainties arising from the driver, the vehicle and the road infrastructures. At the entrance of the curve, speed and position of the approaching vehicle are measured. Then, a family of trajectories is simulated from a dynamic model. Some of the model input depend on experimental trajectories measured on the real traffic. Finally, savety criteria are chosen and associated risk indices are calculated using reliability methods.The influence of various parameters of the model are studied, and the obtained results are the evolutions of the risk indices as a function of the measured initial conditions. The probabilistic methodology elaborated in the road safety domain constitute a valuable decision support tool for the alarm triggering.

Fire development in a vented enclosure can proceed in an explosive and disastrous manner called flashover. This thesis examines when, why and how flashover occurs and gives the answers in terms of a few determining dimensionless parameters. The mechanism of flashover considered in this thesis is an enhancement of the burning rate because of thermal radiation from a layer of hot smoke, produced in the course of the fire, to the fire bed. A model, which is proposed for the problem, describes the development fro~ the moment of ignition incorporating the traditional two-zone approach. During early fIre development the density and temperature of the lower zone are reasonably assumed to be close to their initial value. Flashover itself is assumed to occur early in the fIre development, within the fuel controlled combustion regime. The model is analysed using the techniques of classical thermal explosion theory. Explicit criteria are found analytically and graphically to determine if the fIre will achieve flashover or not. The temperature-time characteristics of the fIre development are obtained explicitly for the fIrst time. It is shown that the thermal inertia of the compartment walls can have a significant effect upon the development. The effect of geometrically scaling the compartment is considered. Nondimensional analysis makes such study effective and leads to a square root relationship for the temperature/time characteristics of the fire development. The correlation between the model, four prevIOUS models and small scale experiments is examined. Under reasonable assumptions all models are shown to be described by the same mathematical problem. This means that the criterion for flashover and the development characteristics can be used for any of the modified models observed. Results are illustrated for an experimental fire box used in many experiments.

The use of small scale facilities in experimental fire research studies is well estabhshed. This thesis concerns the use of small scale facilities to examine principally the influence of thermal inertia of the lining material on ventilated enclosure fire The radiation error of the thermocouple reading was studied using radiation network models. Previous theoretical studies were implemented to provide improved models appropriate to the more complex arrangements considered here. Modeling was used to assess the radiation error of different sizes of thermocouple in the hot layer measuring position for post and pre flashover fires, and the models were compared with experiment. The current range of thermal inertia values for building products is much wider than those used in the regression by which the classical theory of pre-flashover temperature was derived (McCaffery et al. 1981). The range considered here is greater then has previously been considered by systematic experimental testing whilst maintaining all other independent key variables constant. Using of reasonable assumptions, the mass loss rates of non-flashover and flashover conditions were predicted by numerical calculation integrated with a zone model. Successful prediction was also made for published tests where sufficient information was available, and good agreement was found irrespective of flashover, scale or geometry. Two important and necessary assumptions used in the zone model, concerning specifically radiation heat transfer in the flashover condition are: that an average temperature of hot gas and flashover flame may be represented by a single thermocouple measurement; and that the massive increase in production of flanu-nable vapours from the fuel surface during flashover leads to a "cool core" partial scattering or blocking of the incident radiant heat from the flashover flame and hot gas. A computer programme was developed to implement and test recent flashover theory (Graharn et al. 1995). A logarithinic relationship has here been suggested between the thermal inertia parameter ', 8' and thermal inertia value of lining material. That relationship matches the current experimental results and other published data. The occurrence of flashover and the value of hot gas peak or steady temperature can be predicted using the computer programme, based on the published theory.

Complex computer models, built on basic physical principles, have the potential to aid in the understanding and prediction of wildland fire behavior. However, there remain significant uncertainties and assumptions in the way such models describe the fire, the vegetation, and the interaction of the two. To understand a model’s capabilities, limitations, and the improvements which are still necessary, comparison of model predictions to experimental measurement is critical. Unfortunately, collecting such measurements is particularly difficult at the large scale over which real wildland fires occur and, as a result, this happens infrequently. To address this, an opportunity was seized to collect a detailed set of measurements of fire behavior in a real forest environment. These measurements are thoroughly analyzed for the description they provide of the fire behavior. They are then used as a benchmark to test the capabilities of a particular complex model to describe such a fire and to highlight the limitations and uncertainties. As a result of this evaluation, a set of recommendations for future research, both in experiments and modeling, are offered, in order provide a coherent strategy for the future which will significantly advance these models.

Made available in DSpace on 2014-10-09T12:31:51Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T14:01:09Z (GMT). No. of bitstreams: 1 02240.pdf: 4149371 bytes, checksum: 240f783a82612847f69b355550847fb0 (MD5) / Dissertacao (Mestrado) / IPEN/D / Instituto de Pesquisas Energeticas e Nucleares - IPEN/CNEN-SP

Made available in DSpace on 2014-10-09T12:38:02Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T14:04:52Z (GMT). No. of bitstreams: 1 05582.pdf: 2197823 bytes, checksum: 93021a704f9d753e4eb2ce62f8d73163 (MD5) / Dissertacao (Mestrado) / IPEN/D / Instituto de Pesquisas Energeticas e Nucleares - IPEN/CNEN-SP

Made available in DSpace on 2014-10-09T12:31:51Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T14:01:09Z (GMT). No. of bitstreams: 1 02240.pdf: 4149371 bytes, checksum: 240f783a82612847f69b355550847fb0 (MD5) / Dissertacao (Mestrado) / IPEN/D / Instituto de Pesquisas Energeticas e Nucleares - IPEN/CNEN-SP

Made available in DSpace on 2014-10-09T12:38:02Z (GMT). No. of bitstreams: 0 / Made available in DSpace on 2014-10-09T14:04:52Z (GMT). No. of bitstreams: 1 05582.pdf: 2197823 bytes, checksum: 93021a704f9d753e4eb2ce62f8d73163 (MD5) / Dissertacao (Mestrado) / IPEN/D / Instituto de Pesquisas Energeticas e Nucleares - IPEN/CNEN-SP

This work has investigated the possibility of developing a model, capable of being used to harmonise standards of fire cover within the EU. The model had to take into account social, humanitarian, economic and environmental factors and the built environment in determining an appropriate emergency response by fire brigades to rescues, fires and contamination of land, air and watercourses. The resulting standards had to be socially acceptable and economically defensible. After examining existing standards of fire cover and the means of determining those standards, all current research into standards of fire cover and related issues was reviewed. A study was also undertaken into fire science, the means of predicting frequency of fire and the means of limiting the incidence and size of fires. Economics, as applied to local and national government and as applied to fire safety systems and fire services was explored. From that initial research it was concluded that the fundamental concepts of fire cover had remained unchanged since the restructuring of fire brigades throughout Europe during the late 1940's and had no relevance to the modern built environment. While some aspects of existing policies and some elements of current research were of value, this work has developed new concepts, including. • Functional requirements for common fire and non-fire emergencies. • Entry preparation time based on the criteria, "The time at which fire fighting shall commence." • Limiting fire size in buildings to "As Small As Reasonably Practicable for a Set Duration" for property protection. • Quantifying the term "As Small As Reasonably Practicable". • Defining a Standard Predetermined Attendance (SPDA) for all incidents. • Determining optimum attendance time limits dependent on the frequency of calls and the installed fire protection features in buildings. • A method of calculating an economic base for determining critical call numbers. • A means of establishing the optimum location of fire stations within a given area. • The number of SPDA's required at each fire station related to total workload. • The provision of assistance to areas that are outside maximum attendance times. The economic model that has resulted, accurately includes all of the potential economic, social, and political variations as the basis of calculation of Standards of Fire Cover in individual Member States of the European Union. It is the contention of this thesis that it is possible to construct formulae, based on sound economic principals, which are capable of being applied to the different fiscal situation in different countries, thereby ensuring similar relative standards.