The use of Human Reliability Analysis (HRA) for a particular safety assessment is still a difficult problem. In order to perform a comparison of the available methods, the best approach is simulation. In this regard, the pilot study proposed by the Organisation for Economic Co-operation and Development (OECD) Halden Reactor Project (HRP) is intended to provide a first guidance in HRA methods evaluation through experimental data on crew performance in simulated scenarios. The quantitative evaluation of the results of these simulations in terms of crew performance and Human Error Probabilities (HEPs) is quite a difficult task. In this paper, a fuzzy expert system for systematically assessing crew performance is presented. The feasibility of the method is proved on a case study concerning a scenario of an incomplete scram in a Boiling Water Reactor (BWR).

The assessment of dependence among human errors is an important aspect of human reliability analysis. When dependence exists between two tasks, the probability of the operators' failure on one task is higher if they have failed on the preceding task, compared to when they have succeeded. In current practice, the task of assessing dependence among successive operator actions relies largely to expert judgment. The conversion of the expert knowledge into a mathematical model can improve traceability and repeatability of the assessment. The goal of the work is to investigate the use of a Bayesian Network (BN) expert model on dependence assessment in Human Reliability Analysis (HRA). In particular, a BN is built from a dependence model designed for post-initiating event scenarios in Nuclear Power Plants (NPPs); then the paper defines how the analysts should interact with the BN in order to translate their assessments into a BN input and to convert the BN output into suitable information for Probabilistic Safety Assessment (PSA). In this respect, two interfaces for the input assessment process and a framework for the output conversion are proposed.

Problems characterized by qualitative uncertainty described by expert judgments can be addressed by the fuzzy logic modeling paradigm, structured within a so-called fuzzy expert system (FES) to handle and propagate the qualitative, linguistic assessments by the experts. Once constructed, the FES model should be verified to make sure that it represents correctly the experts’ knowledge. For FES verification, typically there is not enough data to support and compare directly the expert- and FES-inferred solutions. Thus, there is the necessity to develop indirect methods for determining whether the expert system model provides a proper representation of the expert knowledge. A possible way to proceed is to examine the importance of the different input factors in determining the output of the FES model and to verify whether it is in agreement with the expert conceptualization of the model. In this view, two sensitivity and uncertainty analysis techniques applicable to generic FES models are proposed in this paper with the objective of providing appropriate tools of verification in support of the experts in the FES design phase. To analyze the insights gained by using the proposed techniques, a case study concerning a FES developed in the field of human reliability analysis has been considered.

The assessment of dependence among human errors is an important aspect of human reliability analysis. When dependence between two tasks exists, the probability of the operators' failure on one task is higher if they have failed on the preceding task, compared to when they have succeeded. In current practice, the task of assessing dependence among successive operator actions relies to a great extent to expert judgment, often with lack of traceability and repeatability. To overcome these limitations, this work presents a systematic framework for the elicitation of expert knowledge on the factors influencing the dependence between two successive tasks. The framework is based on a fuzzy expert system in which a set of transparent fuzzy logic rules is used to represent the relationship between the input factors and the conditional human error probability. The proposed modeling approach is applied to two tasks required in response to an accident scenario at a nuclear power plant. Given the methodological scope of the work, the fuzzy expert system is tested directly on a working model of dependence, whereas no actual exercise of expert judgment elicitation is carried out.

In the present work, cellular automata are combined with Monte Carlo sampling to solve two critical problems related to the unreliability assessment of complex networks composed of nodes interconnected by binary arcs: the identification of the minimal cut sets of the network and the computation of the Fussell—Vesely importance measures assessing the criticality of each arc with respect to the network unreliability. The effectiveness of the method is tested on two literature case studies.

This paper analyzes the behaviour of a fuzzy expert system for evaluating the dependence among successive operator actions, through a sensitivity analysis on the fuzzy input partitioning and assessment. Preliminary results are presented with respect to a case study concerning two successive tasks of an emergency procedure in a nuclear reactor. Work is in progress to perform a thorough sensitivity analysis to generalize the results obtained.

We present a control system, which allows an automatic optimization of the pulse train stability in a mode-locked laser cavity. In order to obtain real-time corrections, we chose a closed loop approach. The control variable is the cavity length, mechanically adjusted by gear system acting on the rear cavity mirror, and the controlled variable is the envelope modulation of the mode-locked pulse train. Such automatic control system maintains the amplitude of the mode-locking pulse train stable within a few percent rms during the working time of the laser. Full implementation of the system on an Nd:yttrium lithium fluoride actively mode-locked laser is presented.

This paper presents a methodology for developing a Fuzzy Expert System to evaluate dependence between two human failure events in human reliability analysis. A working model of dependence is used to show the methodology. The main assets of the use of Fuzzy Expert System are the traceability, the transparency and repeatability given to the dependence evaluation process. Expert knowledge is elicited to identify the main influencing factors with respect to the dependence between two successive tasks. Then, the relationship of the input factors and the conditional human error probability is established through a set of transparent fuzzy logic rules. A case study is presented to demonstrate the proposed approach.

In the present work, Cellular Automata are combined with Monte Carlo sampling to solve two critical problems related to the unreliability assessment of complex networks composed of nodes interconnected by binary arcs: the identification of the minimal cut sets of the network and the computation of the FussellVesely importance measures assessing the criticality of each arc with respect to the network unreliability. The effectiveness of the method is tested on two literature case studies.

This paper considers the fuzzy version of the Cognitive Reliability and Error Analysis Method (CREAM) for computing the probabilities of human action failures. By means of a weight assignment via the Analytic Hierarchy Process (AHP), accountancy is given to the different effects that the characteristics of the context scenario (Common Performance Conditions, CPCs) may have on the human operator performance. A thorough sensitivity analysis on the AHP ranking procedure is carried out to investigate the robustness of the proposed approach. A scenario of emergency response to a Steam Generator Tube Rupture (SGTR) in a Nuclear Power Plant (NPP) is considered as case study.

The quantitative assessment of the reliability of network systems can be a quite difficult and computationally expensive problem in practice. In this respect, Monte Carlo simulation offers a valuable tool for capturing the complex stochastic behavior of distributed, interconnected systems.

A high power laser system called ‘ATTILA’ (Almost-Table Top Terawatt Intense Laser) (10 TW) is under development at the University of Milano, Bicocca. Once fully functional ATTILA will produce 10 J, 1 picosecond pulses, reaching the power of 10 TW. The system is based on the chirped pulsed amplification technique and uses Nd:glass single-pass amplification. The amplifier chain can be injected with two different oscillators: a Nd:YLF oscillator by Quantronix, stretched by an optical fiber, and a Nd:glass femtosecond oscillator by TimeBandwidth (Zurich), stretched by diffraction gratings, both delivering nanojoules pulses. A regenerative amplifier increasing the energy up to 1 mJ follows these. Recently, we have implemented an automatic control system for the stabilization of mode-locking in the Quantronix oscillator. This is based on the active control of cavity length and on accurate stabilization of mode-locker temperature. A PC using LabView software drives the system. We are now beginning the installation of an adaptive optics system with a deformable mirror in order to optimize the optical quality and focusability of the beam. Our final goal is to obtain an intensity of 1018 W/cm2 on target. This will allow the production of relativistic electrons and energetic protons, and the study of relativistic plasma physics and of matter in extreme conditions.