Takehisa Kohda

Finding modules in fault trees

A new method for identifying all possible modules is presented. There are two kinds of modules: (1) those whose output events are expressed by gate events, and (2) those whose output events are not expressed by gate events. The latter are logical OR or AND combinations of basic events and modules. The method requires as input only fault-tree structure data representing gate event output-input relations. The output is a hierarchical decomposition of the fault tree into modules. The order in which the modules are identified corresponds to both the hierarchy established when the gates are numbered and the order in which modules are analyzed. The method can be applied to noncoherent fault trees. Illustrative examples are given. It is demonstrated that, by modular decomposition, 100-fold reductions in computation time are obtained for a 70-gate, 67-basic-event problem. >

A Reliability Optimization Method for Complex Systems with the Criterion of Local Optimality

A method with a new criterion of local optimality is proposed to solve reliability optimization problems of complex systems, where the structure need not be series. Unlike previous methods, the proposed algorithm yields a solution which is optimal in 2-neighborhood. The effectiveness of the method is shown through examples.

Risk-based reconfiguration of safety monitoring system using dynamic Bayesian network

To prevent an abnormal event from leading to an accident, the role of its safety monitoring system is very important. The safety monitoring system detects symptoms of an abnormal event to mitigate its effect at its early stage. As the operation time passes by, the sensor reliability decreases, which implies that the decision criteria of the safety monitoring system should be modified depending on the sensor reliability as well as the system reliability. This paper presents a framework for the decision criteria (or diagnosis logic) of the safety monitoring system. The logic can be dynamically modified based on sensor output data monitored at regular intervals to minimize the expected loss caused by two types of safety monitoring system failure events: failed-dangerous (FD) and failed-safe (FS). The former corresponds to no response under an abnormal system condition, while the latter implies a spurious activation under a normal system condition. Dynamic Bayesian network theory can be applied to modeling the entire system behavior composed of the system and its safety monitoring system. Using the estimated state probabilities, the optimal decision criterion is given to obtain the optimal diagnosis logic. An illustrative example of a three-sensor system shows the merits and characteristics of the proposed method, where the reasonable interpretation of sensor data can be obtained.

Optimal Structure of Sensor Systems with Two Failure Modes

An optimal sensor structure is developed for a sensor system which consists of several channels. Each channel monitors a particular plant state, e.g., temperature or pressure. When some states become abnormal, an event occurs. The channels which monitor these abnormal states then initiate appropriate safety systems. Sensors are either good, or failed-dangerous, or failed-safe. More than one sensor is available for each channel. The problem is to obtain the optimal s-coherent sensor structures for the channels. A theorem is proven and a nonlinear integer programming (NLIP) problem is derived to minimize s-expected total damage. The NLIP problem can be solved by the extended Lawler & Bell algorithm. For a 1-channel sensor system, the optimal structure can be obtained analytically by a simple formula.

Identification of system failure causes using bond graph models

This paper proposes the use of bond graphs in system reliability analysis to obtain combinations of component failures and disturbances leading to the system failure. Since bond graphs show a direct correspondence between their component and physical phenomena to be modeled, it is easy to obtain not only a system model, but also component failure representations. The system bond graph can be easily modified by its stepwise refinement. Using the formal causality assignment procedure, propagation paths of disturbances can be identified. Further, the detailed analysis can be also performed using quantitative information. Illustrative examples show the details and merits of the proposed method.<<ETX>>

A simple method for phased mission analysis

Abstract In a large-scale system such as chemical and nuclear plants, the system function requirement changes during its operation and its component must also play various roles depending on the system state. Since the logical relation between the system and its components changes during the system operation, the Phased Mission Analysis (PMA) must be introduced into their reliability analysis. The basic event transformation method in PMA can evaluate their reliability exactly, but it makes the analysis more complicated and time-cosuming. This paper develops a new, simple, and practical method for PMA. A novel component condition expression is proposed, which can express time requirements explicity. The logical operations of conditions for a component are transformed into the modification of time parameters in its expression. Event sequence conditions are obtained as logical OR of logical AND combinations of the proposed component condition expressions and their probabilities can be easily evaluated based on them. Illustrative examples show the effectiveness and merits of the proposed method compared with the dual tree method in Event Tree Analysis and the basic event transformation method in PMA. Especially, in an example of an emergency core cooling system (ECCS), the proposed method (1) can obtain mission failure conditions in a compact from of 7 terms while 70 minimal cut sets are obtained by the basic event transformations method, and (2) can obtain the mission unreliability in an analytical form.

Accident cause analysis of complex systems based on safety control functions

State of the art computers have made many technology-based systems so complex that new types of accidents now result from dysfunctional interactions between system components, further adding to the number of accidents resulting from component failure. Other factors, such as management effectiveness and organizational constraints, also must be considered as part of a failure prevention strategy. Conventional event-based analysis methods such as fault trees and event trees cannot be applied to such types of accidents. This paper applies a concept of safety constraints in systems control to the accident cause analysis. The analysis shows the effectiveness of the proposed method for both the retrospective and prospective events

Life Cycle Simulation Applied to a Robot Manipulator - An Example of Aging Simulation of Manufacturing Facilities -

Abstract Evaluation of the aging process of manufacturing facilities is essential for life cycle facility management, which includes such activities as design for reliability and maintainability and maintenance planning. For this purpose, we have developed a life cycle simulation system for robot manipulators. The developed system can simulate the wear of gears and bearings of joints as component deterioration and evaluate the resultant positioning error of an end-effector as functional degradation. The system is applied to assembly robots of a car parts manufacturing plant. The simulation results correspond to the failure history of the robots fairly well.

Fault-tree analysis considering latency of basic events

Severe accidents can occur even in advanced systems such as airplanes and nuclear power plants. In the accident analysis, most of the attention focuses on the active failure which initiates the occurrence of system failure, but the latent condition which allows the active failure to cause system failure must be considered. This paper proposes a simple method to deal with latent conditions in FTA (fault tree analysis) based on the detectability of basic events. In FTA analysis, the system failure occurs if all basic events in a minimal cut set occur, but all the event sequences leading to the occurrence of a minimal cut set do not always occur without their detection. Using the detectability index of a basic event, it is evaluated whether the critical state of a basic event is latent where its occurrence can cause the system failure. Based on the number/probability of critical states that can be detected by the detectability of a basic event, a new measure of its importance in contributing to the reduction of system failure is proposed. A simple illustrative example shows these properties.

Maintenance Data Management System

Abstract For the effective management of manufacturing facilities throughout their life cycle, it is important to collect maintenance data and make use of them for operation and maintenance planning or design of new machines. However, it is seldom that the maintenance data are properly recorded and utilized in a systematic way. To solve these problems, we have proposed a maintenance data management system. In this paper, two major modules of the system are discussed. The first one is a malfunction data collection system which can navigate users to input malfunction cases in a proper format. The second one is a feedback data generation system which induces common causes implied in the malfunction cases by means of the attribute-oriented induction algorithm. An experimental system is applied to malfunction cases of machine tools for demonstrating its effectiveness.