In this age of safety awareness, technological emergencies still happen, occasionally with catastrophic results. Often human intervention is the only way of averting disaster. Ensuring that the chosen emergency managers are competent requires a combination of training and assessment. However, assessment currently relies on expert judgement of behaviour as opposed to its impact on outcome, therefore it would be difficult to incorporate such data into formal Quantitative Risk Assessments (QRA). Although there is, as yet, no suitable alternative to expert judgement, there is a need for methods of quantifying the impact of emergency management on risk reduction in accident and incidents. The Task Performance Resource Constraint (TPRC) model is capable of representing the critical factors. It calculates probability of task success with respect to time based on uncertainties associated with the task and resource variables. The results can then be used to assess the management performance based on the physical outcome in the emergency, thereby providing a measure of the impact of emergency management on risk with a high degree of objectivity. Data obtained from training exercises for offshore and onshore emergency management were measured and successfully used with the TPRC model. The resulting probability of success functions also demonstrated a high level of external validity when used with improvements in emergency management or design changes or real data from the Piper Alpha disaster. It also appeared to have more external validity than other HRQ/QRA techniques as it uses physical data that are a greater influence on outcome than psychological changes - though this could be because the current HRA/QRA techniques view human unreliability as probability of error rather than probability of failure. The simulation data were also used to build up distributions of timings for simple emergency management tasks. Using additional theoretical data, this demonstrated the model's potential for assessing the probability of successf or novel situations and future designs.

No description available.

abstract: Bayesian networks are powerful tools in system reliability assessment due to their flexibility in modeling the reliability structure of complex systems. This dissertation develops Bayesian network models for system reliability analysis through the use of Bayesian inference techniques. Bayesian networks generalize fault trees by allowing components and subsystems to be related by conditional probabilities instead of deterministic relationships; thus, they provide analytical advantages to the situation when the failure structure is not well understood, especially during the product design stage. In order to tackle this problem, one needs to utilize auxiliary information such as the reliability information from similar products and domain expertise. For this purpose, a Bayesian network approach is proposed to incorporate data from functional analysis and parent products. The functions with low reliability and their impact on other functions in the network are identified, so that design changes can be suggested for system reliability improvement. A complex system does not necessarily have all components being monitored at the same time, causing another challenge in the reliability assessment problem. Sometimes there are a limited number of sensors deployed in the system to monitor the states of some components or subsystems, but not all of them. Data simultaneously collected from multiple sensors on the same system are analyzed using a Bayesian network approach, and the conditional probabilities of the network are estimated by combining failure information and expert opinions at both system and component levels. Several data scenarios with discrete, continuous and hybrid data (both discrete and continuous data) are analyzed. Posterior distributions of the reliability parameters of the system and components are assessed using simultaneous data. Finally, a Bayesian framework is proposed to incorporate different sources of prior information and reconcile these different sources, including expert opinions and component information, in order to form a prior distribution for the system. Incorporating expert opinion in the form of pseudo-observations substantially simplifies statistical modeling, as opposed to the pooling techniques and supra Bayesian methods used for combining prior distributions in the literature. The methods proposed are demonstrated with several case studies. / Dissertation/Thesis / Doctoral Dissertation Industrial Engineering 2016

Thesis (PhD)--Stellenbosch University, 2015. / ENGLISH ABSTRACT: Structural design standards based on the principles of structural reliability are gaining worldwide acceptance and are fast becoming the new basis for structural safety verification. The application of these principles to establish a standardised basis for structural design using partial factor limit states design procedures is done in the European Standard for the Basis of Structural Design EN 1990 from which it is adapted to the South African Standard Basis of Design for Building and Industrial Structures SANS 10160-1. South Africa (SA) is on the advent of adopting the European Concrete Design Standard EN 1992-1-1 (EC2) as the equivalent standard for local use. This investigation seeks to provide a transparent quantitative reliability basis for the SA’s adoption of EC2, as well as provide for its subsequent implementation under local conditions and practice. The investigation kicks-off with a critical review of the reliability framework for structural resistance. The review establishes the relationships between the key elements of the framework, shedding light on issues SA needs to consider as it adopts EC2. Important issues for SA to consider include (1) target levels of structural performance ( -values), (2) partial factors, (3) model uncertainties, and (4) quality control. Design for shear resistance was investigated in greater detail by comparison of EC2’s Variable Strut Inclination Method (VSIM) for stirrup design against alternative approaches, namely, (1) South Africa’s currently operational SANS 10100-1 procedure, and (2) the fib Model Code 2010 first Level of Approximation (LoA I) and fib LoA III, which are based on the Modified Compression Field Theory (MCFT). Unbiased capacity predictions from the MCFT-based sectional analysis Program Response-2000 (R2k) served as LoA IV best-estimate results during this assessment. Results of this investigation showed that EC2 offers higher capacity predictions in excess of 1 MPa of stirrup reinforcement, with significantly higher predictions in the range of 1 to 2 MPa. A reliability performance assessment was therefore commissioned to assess safety regimes in terms of achieved reliability across a parametric range of the amount of stirrup reinforcement (from 0.45 to 2.0 MPa). The First Order Reliability Method (FORM) was implemented as part of the reliability performance assessment of the EC2’s VSIM design procedure. The model uncertainty for shear resistance (stirrup failures) was characterised according to a database of published stirrupreinforced concrete beam shear tests. Three cases of the Model Factor for shear resistance were derived from the experimental database for alternative shear resistance prediction models; two of which formed part of basic investigations conducted using the conventionally formulated performance function, and the other was integrated as part of an independent validation procedure using R2k predictions to obtain the reliability model. Results obtained from the basic reliability model ( -values) generally indicated lower levels of reliability with an increase in stirrup reinforcement and concrete strength, compared to those estimated from the R2k-based reliability model ( -values). The disparity between and -values revealed that systematic effects affect each model’s ability to predict the expected value of true shear resistance . There is reasonable evidence to suggest that the predictions of can be improved by accounting for each model’s peculiar sensitivity to concrete strength, consequently providing more representative estimates of . However, in the interim, and -values, respectively, represent reasonable lower and upper bound estimates of the performance of EC2’s VSIM design procedure. / AFRIKAANSE OPSOMMING: Die beginsels van struktuur betroubaarheid word wêreldwyd aanvaar as basis vir struktuur ontwerp standaarde en die versekering van voldoende struktuur veiligheid. Hierdie beginsels word in die Europese Standaard Basis of Structural Design EN 1990 toegepas om gestandaardiseerde partiële faktor gebaseerde limietstaat ontwerp prosedures daar te stel, van waar dit aangepas is vir die Suid-Afrikaanse Standaard Basis of Design for Building and Industrial Structures SANS 10160-1. Suid-Afrika (SA) staan op die punt om die Europese beton ontwerp standaard EN 1992-1-1 (EC2) aan te neem as die ekwivalente standaard vir plaaslike gebruik. Hierdie ondersoek het as doel om ‘n deursigtige kwantitatiewe betroubaarheidsbasis daar te stel vir die aanneming van EC2 as SA standaard en om voorsiening te maak vir die implementering daarvan onder plaaslike toestande en –praktyk. Die ondersoek begin met ‘n kritiese oorsig van die betroubaarheidsraamwerk vir strukturele weerstand. Die oorsig stel die verhouding vas tussen sleutel elemente van die raamwerk en werp lig op aspekte wat SA moet oorweeg in die aanneming van EC2. Belangrike aspekte vir oorweging sluit (1) teiken betroubaarheidsvlakke vir struktuur gedrag ( -waardes), (2) partiele faktore, (3) model onsekerhede en (4) kwaliteitsbeheer in. Skuif weerstandsontwerp is in groter detail ondersoek deur die EC2 se Veranderbare Stut Hoek Metode (VSHM) vir skuifbeuel ontwerp te vergelyk met alternatiewe benaderings, naamlik, (1) Suid Afrika se huidig operasionele SANS 10100-1 prosedure, (2) fib Model Code 2010 se sogenaamde eerste Vlak van Benadering (VvB I) en fib VvB III, gebaseer op die Aangepaste Drukveld Teorie (ADT). Onbevooroordeelde kapasiteit voorspellings van die ADT-gebaseerde snit analise program “Response-2000 (R2k)” is in die evaluering gebruik as VvB IV bes benaderde resultate. Die ondersoek toon dat EC2 hoër kapasiteit voorspel vir skuifbeuel bewapening tot 1 MPa en beduidend hoër kapasiteite voorspel vir skuifbeuel bewapening tussen 1 en 2 MPa. ‘n Betroubaarheidsprestasie studie is vervolgens geloots om die veiligheid in terme van behaalde betroubaarheid te bepaal oor ‘n parametriese bereik van 0.45 tot 2.0 MPa skuifbeuel bewapening. Die Eerste Orde Betroubaarheids Metode (EOBM) is implementeer as deel van die betroubaarheidsprestasie beoordeling van die EC2 VSHM ontwerp prosedure. ‘n Databasis van gepubliseerde skuifbeuel-bewapende betonbalk skuiftoetse is gebruik om die model onsekerheid vir die verskillende skuifweerstandsmodelle statisties te beskryf. Drie Model Faktore is uit die eksperimentele databasis afgelei, twee waarvan gebruik is in basiese ondersoeke met die konvensioneel geformuleerde prestasie funksie en die derde as deel van ‘n onafhanklike bevestigingsprosedure gebaseer op R2k voorspellings. Resultate wat verkry is uit die basiese betroubaarheidsmodel ( -waardes) was laer (meer konserwatief), en het ook vinniger afgeneem met ‘n toename in skuifbeuel bewapening as die waardes wat uit die R2k-gebaseerde betroubaarheidsmodel ( -waardes) verkry is. Die verskil tussen β en β -waardes toon dat sistematiese effekte die vermoë van elk van die modelle beïnvloed om die verwagte waarde van die werklike skuifweerstand V te voorspel. Daar is redelike bewyse om aan te voer dat die voorspellings van V verbeter sal kan word deur elke model se unieke sensitiwiteit teenoor betonsterkte in ag te neem, om sodoende meer verteenwoordigende β waardes te verkry. Intussen verteenwoodig die β en β -waardes onderskeidelik redelike onder- en bogrens skattings vir die prestasie van EC2 se VSHM ontwerp prosedure.

In this thesis, a retail market mechanism that provides differentiated reliability services is proposed. The differentiated reliability services beyond the standard level utilize advanced metering infrastructure, automated distribution reconfiguration and distributed generation (DG). The service quality at the standard level is regulated, while high reliability services are offered through a market mechanism. This proposed market mechanism is designed in two different models of managing the distribution networks. The first model assumes that an independent distribution system operator (DSO) as an administrative firm provides operational support for delivery and reliability services in a retail market, while the second model does not have a DSO. Main reliability market participants are distribution utilities, retail electricity providers (REPs), non-utility-owned DG units, and end users. The REPs, as end users’ representatives and aggregators, purchase delivery service with high reliability level and backup power from the utilities and DG units, respectively. The prices for these services are based on bidding by all market participants. Bids are created by each market participant optimizing its objective with respect to its own interests; therefore, the market participant can assess the investment costs and manage its own risk in setting the service charge. Notably, the proposed market mechanism, which is based on knowing customers’ willingness to pay, and preferences for reliability, aims to give long-term investment signals to service providers for planning investments in new technologies at value. In addition, the provision of high reliability services can be considered a means that enables the service providers to improve system resilience. The modified IEEE Roy Billinton Test System Bus 2 is simulated to demonstrate proof-of-concept for the proposed retail market by showing the iii process of settling the service prices and utilities’ expected compensation design. By comparing the settled service prices between the two market models, we show that the service prices are quite similar, but the number of end users obtaining backup power is different.

Offshore oil and gas exploitation is principally conducted using dry or wet tree systems, otherwise called the subsea Xmas tree system. Due to the shift to deeper waters, subsea production system (SPS) has come to be a preferred technology with attendant economic benefits. At the centre of the SPS is the subsea control module (SCM), responsible for the proper functioning and monitoring of the entire system. With increasing search for hydrocarbons in deep and ultra-deepwaters, the SCM system faces important environmental, safety and reliability challenges and little research has been done in this area. Analysis of the SCM reliability then becomes very fundamental due to the huge cost associated with failure. Several tools are available for this analysis, but the FMECA stands out due to its ability to not only provide failure data, but also showcase the system’s failure modes and mechanisms associated with the subsystems and components being evaluated. However, the technique has been heavily challenged in various literatures for several reasons. To close this gap, a novel multi-criteria approach is developed for the analysis and ranking of the SCM failures modes. This research specifically focusses on subsea tree-mounted electro-hydraulic (E-H) SCM responsible for the underwater control of oil and gas production. A risk identification of the subsea control module is conducted using industry experts. This is followed by a comprehensive component based FMECA analysis of the SCM conducted with the conventional RPN technique, which reveals the most critical failure modes for the SCM. A novel framework is developed using multi-criteria fuzzy TOPSIS methodology and applied to the most critical failure modes obtained from the FMECA evaluation using unconventional parameters. Finally, a validation of these results is performed using a stochastic input evaluation and SCM failure data obtained from the offshore industry standard reliability database, OREDA.

Confront with the serious aging problem in aircraft structure field, the profession was tasked to unveil the mysterious in the mechanism of aging. In decades, many endeavours were put into different subjects such as, fatigue and crack calculation, corrosion analysis, reliability evaluation, life prediction, structure monitor and protection, structure repair, etc. In an effort of developing a reasonable model for life prediction and reliability evaluation, a wide range of topics in the field of aging structure reliability are reviewed. Many existing methods and tools are carefully studied to distinguish the advantages, disadvantages and the special application. With consideration of corrosion fatigue life, and based on the data obtained through investigating service status of the aging aircraft, a fuzzy reliability approach is proposed and presented. Initially, the thesis presents the literature review in the field, introducing the well-established theories and analysis tools of reliability and points out how such these methods can be used to assess the life and reliability of aging structure. Meanwhile, some characteristic parameters and distributions, as well as some crucial calculation formulations, procedures for aging aircraft reliability/risk analysis are given. Secondly, mathematical models are established to evaluate the initial crack size and to assess both randomness and fuzziness of the variables, which also successfully work out the probability of survival of existing structures over a time period and predict the operation time under specific reliability requirement. As a practical approach to the reliability of aging aircraft structure, example is presented and evaluated. While conduct the calculation, a few programs based on FORTRAN code are developed to solve the none-linear equation, to work out the multi dimension integration and to simulate the survival probability. The crack life prediction software AFGROW is selected for comparison of the calculation results, which also shows the appropriate accuracy of the established model. As conclusion, the effects of some variables including fuzzy factors on reliability and life of aging aircraft structure are finally discussed. It is apparent that the confines of the model are existing as fact because of the huge assumption of the parameters input and model uncertainties. Suggestions on further prospective research are proposed respectively.

Computational design has become a predominant element of various engineering tasks. However, the ever increasing complexity of numerical models creates the need for efficient methodologies. Specifically, computational design under uncertainties remains sparsely used in engineering settings due to its computational cost. This dissertation proposes a coherent framework for various branches of computational design under uncertainties, including model update, reliability assessment and reliability-based design optimization. Through the use of machine learning techniques, computationally inexpensive approximations of the constraints, limit states, and objective functions are constructed. Specifically, a novel adaptive sampling strategy allowing for the refinement of any approximation only in relevant regions has been developed, referred to as generalized max-min. This technique presents various computational advantages such as ease of parallelization and applicability to any metamodel. Three approaches tailored for computational design under uncertainties are derived from the previous approximation technique. An algorithm for reliability assessment is proposed and its efficiency is demonstrated for different probabilistic settings including dependent variables using copulas. Additionally, the notion of fidelity map is introduced for model update settings with large number of dependent responses to be matched. Finally, a new reliability-based design optimization method with local refinement has been developed. A derivation of sampling-based probability of failure derivatives is also provided along with a discussion on numerical estimates. This derivation brings additional flexibility to the field of computational design. The knowledge acquired and techniques developed during this Ph.D. have been synthesized in an object-oriented MATLAB toolbox. The help and ergonomics of the toolbox have been designed so as to be accessible by a large audience.

Master of Science / Department of Industrial & Manufacturing Systems Engineering / Malgorzata J. Rys / According to the Centers for Disease Control and Prevention (2015), urinary tract infections (UTIs) are the most commonly reported healthcare-associated infection (HAI), of which approximately 75% of infections are attributed to the presence of a urinary catheter. Urinary catheters are commonplace within hospitals as approximately 15-25% of patients receive a urinary catheter during their hospitalization, introducing the risk of a catheter associated urinary tract infection (CAUTI) during their stay (CDC, 2015). In recent years there have been efforts to reduce CAUTI in U.S. hospitals; however, despite these efforts, CAUTI rates indicate the need to continue prevention efforts. Researchers have investigated the use of human reliability analysis (HRA) techniques to predict and prevent CAUTI (Griebel, 2016), and this research builds on that topic by applying the Human Error Assessment and Reduction Technique (HEART) to develop a model for a patient’s probability of CAUTI. HEART considers 40 different error-producing conditions (EPCs) present while performing a task, and evaluates the extent to which each EPC affects the probability of an error. This research considers the task of inserting a Foley catheter, where an error in the process could potentially lead to a CAUTI. Significant patient factors that increase a patient’s probability of CAUTI (diabetes, female gender, and catheter days) are also considered, along with obesity which is examined from a process reliability perspective. Under the HEART process, human reliability knowledge and the knowledge of eight expert healthcare professionals are combined to evaluate the probability that a patient will acquire a CAUTI. In addition to predicting the probability of CAUTI, HEART also provides a systematic way to prioritize patient safety improvement efforts by examining the most significant EPCs or process steps. The proposed CAUTI model suggests that 7 of the 26 steps in the catheter insertion process contribute to 95% of the unreliability of the process. Three of the steps are related to cleaning the patient prior to inserting the catheter, two of the steps are directly related to actually inserting the catheter, and two steps are related to maintaining the collection bag below the patient’s bladder. An analysis of the EPCs evaluated also revealed that the most significant factors affecting the process are unfamiliarity, or the possibility of novel events, personal psychological factors, shortage of time, and inexperience. By targeting reliability improvements in these steps and factors, healthcare organizations can have the greatest impact on preventing CAUTI.

No / Quality assessment (QA) requires high levels of domain-specific experience and knowledge. QA tasks for toxicological data are usually performed by human experts manually, although a number of quality evaluation schemes have been proposed in the literature. For instance, the most widely utilised Klimisch scheme1 defines four data quality categories in order to tag data instances with respect to their qualities; ToxRTool2 is an extension of the Klimisch approach aiming to increase the transparency and harmonisation of the approach. Note that the processes of QA in many other areas have been automatised by employing expert systems. Briefly, an expert system is a computer program that uses a knowledge base built upon human expertise, and an inference engine that mimics the reasoning processes of human experts to infer new statements from incoming data. In particular, expert systems have been extended to deal with the uncertainty of information by representing uncertain information (such as linguistic terms) as fuzzy sets under the framework of fuzzy set theory and performing inferences upon fuzzy sets according to fuzzy arithmetic. This paper presents an experimental fuzzy expert system for toxicological data QA which is developed on the basis of the Klimisch approach and the ToxRTool in an effort to illustrate the power of expert systems to toxicologists, and to examine if fuzzy expert systems are a viable solution for QA of toxicological data. Such direction still faces great difficulties due to the well-known common challenge of toxicological data QA that "five toxicologists may have six opinions". In the meantime, this challenge may offer an opportunity for expert systems because the construction and refinement of the knowledge base could be a converging process of different opinions which is of significant importance for regulatory policy making under the regulation of REACH, though a consensus may never be reached. Also, in order to facilitate the implementation of Weight of Evidence approaches and in silico modelling proposed by REACH, there is a higher appeal of numerical quality values than nominal (categorical) ones, where the proposed fuzzy expert system could help. Most importantly, the deriving processes of quality values generated in this way are fully transparent, and thus comprehensible, for final users, which is another vital point for policy making specified in REACH. Case studies have been conducted and this report not only shows the promise of the approach, but also demonstrates the difficulties of the approach and thus indicates areas for future development. / U 7th Framework Programme Integrated Project “Integrated In Silico Models for Prediction of Human Repeated Dose Toxicity of Cosmetics to Optimise Safety” (COSMOS). Grant Number: 266835. Cosmetics Europe.