The South African Weather Service rainfall seasonal forecasts are verified for the period of January-February-March to October-November-December 1998-2004. These forecasts are compiled using different models from different institutions. Probability seasonal forecasts can be evaluated using different skill measures, but in this study the Ranked Probability Skill Score (RPSS), Reliability Diagram (RD) and Relative Operating Characteristics (ROC) are used. The RPSS is presented in the form of maps whereas the RD and ROC are analyses are presented in the form of graphs. The aim of the study is to present skill estimates of operational seasonal forecasts issued at South African Weather Service A limited number of forecasts show positive RPSS value throughout the validation period. From RD and ROC analysis, there is no skill in predicting the normal category as compared to below-normal and above-normal categories. Notwithstanding, the frequency diagrams show that the normal category was often given a large weight in the operational forecasts. The value of verifying seasonal forecast accuracy from the user’s perspective is important. The understanding of seasonal forecast performance helps decision makers to determine when and how to respond to expected climate anomalies. Therefore the frequent update of the seasonal forecast verification is important in order to help Users make better decisions. Copyright / Dissertation (MSc)--University of Pretoria, 2008. / Geography, Geoinformatics and Meteorology / Unrestricted

This thesis presents a reliability analysis of a technical system. It is divided into three main sections. The first section introduces some of the most significant problems of reliability. It deals with a definition and an expresion of reliability, a reliability diagram selection and a detailed description of the reliability analysis. This part also includes an overview of reliability analysis types. Some of the most widely used reliability analyses are briefly described; further advantages and disadvantages of using each method are listed. Failure Modes and Effects Analysis - FMEA is then described in a greater detail. The second section contains an analysis of aviation conditions as well as a design of a reliability analysis that concerns a selected digital system; the system under analysis is a pitch trim indicator. The main design issue lies in a choice of a most suitable method. This choice stems from the overview of reliability analyses presented in the first section of the thesis. In the last section, a FMEA reliability analysis of the pitch trim indicator is conducted. This part includes a discussion of the results as well as a design action that is to lead to an increase in reliability of the analyzed system.

This work deals with simulation of the urban distribution network reliability. For the evaluation of these networks are used indicators of reliability which are described in the first part. Next methods of analysis and posibilities for capabilities reliability simulation of distribution network are described. The third part deals with the real value of the reliability of electricity distribution in Italy. These values are presented in tables and graphs. The last part is a simulation of the urban network. Effect of different configurations on the reliability of electricity supply are studied. Sensitivity analysis is done for failure rate and proportional of the number of sampling points; the simulated values are fiven in tables and graphically.

Prior to radiation therapy planning, tumours and organs at risk need to be delineated. In recent years, deep learning models have opened the possibility of automating the contouring process, speeding up the procedures and helping clinicians. However, deep learning models, trained using ground truth labels from different clinicians, inevitably incorporate the human-based inter-observer variability as well as other machine-based uncertainties and biases. Consequently, this affects the accuracy of segmentation, representing the primary source of error in contouring tasks. Therefore, clinicians still need to check and manually correct the segmentation and still do not have a measure of reliability. To tackle these issues, researchers have shifted their focus to the topic of probabilistic neural networks and uncertainties in deep learning models. Hence, the main research question of the project is whether a 3D U-Net neural network trained on CT lung cancer images can enhance clinical contouring practice by implementing a probabilistic auto-contouring system. The Monte Carlo dropout technique was employed to generate probabilistic and uncertainty maps. The model calibration was assessed using reliability diagrams, and subsequently, a clinical experiment with a radiation oncologist was conducted. To assess the clinical validity of the uncertainty maps two novel metrics were identified, namely mean uncertainty (MU) and relative uncertainty volume (RUV). The results of this study demonstrated that probability and uncertainty mapping effectively identify cases of under or over-contouring. Although the reliability analysis indicated that the model tends to be overconfident, the outcomes from the clinical experiment showed a strong correlation between the model results and the clinician’s opinion. The two metrics exhibited promising potential as indicators for clinicians to determine whether correction of the predictions is necessary. Hence, probabilistic models revealed to be valuable in clinical practice, supporting clinicians in their contouring and potentially reducing clinical errors.