Sadanori Shingai

Brief paper: Fault location using digraph and inverse direction search with application

A fault location method for large-scale plants is described. Fault location is executed using the following procedure: first, a digraph which indicates the failure propagation network of a plant is drawn using nodes corresponding to the devices of the plant or failure modes of the devices. Arrows which correspond to the direction of failure propagation between adjacent nodes are drawn in. Second, some nodes are chosen as candidates of the failure origin by back-tracing, using the arrows, starting from the nodes which indicate abnormal states such as rapidly rising pressure. Third, the candidates are screened by using failure propagation probabilities between adjacent nodes, failure propagation time between adjacent nodes, and back-tracing, starting from the nodes which indicate normal states. Finally, the failure propagation probabilities and the failure rates of the devices are used to evaluate the priority ranking among the screened candidates. This method is applied to pump and evaporation plants.

Brief paper: Failure propagating simulation and nonfailure paths search in network systems

This paper describes a method for identifying failure propagation paths in networks. It also assesses the risks associated with failure propagation and identifies the nonfailure paths remaining in the network after failure propagation has occurred. The method employs the following procedure: (1) selection of elements with likelihood of failure; (2) determination of failure source; (3) drawing of failure propagating network; (4) simulation of failure propagation; (5) risk assesment; and (6) search for nonfailure paths.

A New Technique to Optimize System Reliability

This paper describes an algorithm for solving reliability optimization problems formulated as nonlinear binary programming problems with multiple-choice constraints. These constraints stand for restrictions in which only one variable is assigned to each subset making up the set; thus, they are expressed by equations whose r.h.s. is unity. Different types of methods for achieving high reliability (an increase in component reliability, parallel redundancy, standby redundancy, etc.) can be easily used simultaneously as design alternatives for each subsystem. In order to solve the problem effectively, the Lawler & Bell algorithm is improved by introducing a new lexicographic enumeration order which always satisfies the multiple-choice constraints. The function for obtaining feasible solutions which give first ~ L-th minimum values of the objective function is added to the algorithm in order to make it more useful for decision making. After a numerical example assists in understanding the algorithm, the computational efficiency is compared with that of the Lawler & Bell algorithm.

A Method of Rapid Markov Reliability Calculation

Our reliability calculation method for a Markov state-transition graph enables a rapid computation by finding and cutting (removing) non-effective edges (NEEs). An NEE is an edge in a Markov state transition graph, the cut of which has little effect on the reliability calculation. NEEs can be found only by checking in a small graph, given the assumption that component failure rate is far smaller than component repair rate. NEEs are found and cut until the Markov graph is separated into two subgraphs. One subgraph is usually very small compared with the original, and the reliability can be approximately calculated on this small subgraph of the Markov graph. Proof and numerical examples are presented.

A Probability Bound Estimation Method in Markov Reliability Analysis

In many practical systems, the uncertainty of component failure/repair rates results in uncertainty of system failure probability. Concerning a repairable system, uncertainty is evaluated as a probability bound in the Markov process. In practical analysis, the Laplace transform has the advantage of relatively less computing time than that of a numerical method, eg, Runge Kutta. This paper proposes an algorithm for evaluating this uncertainty using the Laplace transform method. This algorithm assumes the Johnson SB distribution for system-failure probability. Then, the mean and the variance of system-failure probability are obtained using Newtons method and an integral form for calculating parametric differentiation. Finally, the probability bounds are obtained by applying the conventional moment-matching method. A tutorial example is presented at the end of this paper.

An Algorithm for Obtaining Simplified Prime Implicant Sets in Fault-Tree and Event-Tree Analysis

Some fault trees or event trees use NOT logic in expressing an accident sequence. The number of prime implicant sets (PIS), a generalization of a minimal cut set and includes logical NOT as well as AND and OR, is apt to exceed general computer capacity. In order to deal with this problem, this paper assumes that component failure probability is far smaller than component success probability. Three steps are developed to delete certain sets of basic events from the fault tree of an accident sequence so as to obtain upper side approximate probability calculations and this with the number of PISs enumerated reduced to a practical level in most cases. Asymptotic convergence is proved and numerical examples are provided.

An Improved Moment-Matching Algorithm for Evaluating Top-Event Probability Bounds

This paper proposes improvements in the momentmatching algorithm developed by Apostolakis & Lee (A&L algorithm). The moment-matching method in the A&L algorithm evaluates reliability uncertainty by: 1) assuming that system reliability is governed by the Johnson SB distribution; and 2) calculating reliability bounds by using the moment-matching method, given the reliability mean value and variance. The proposed (T.S.S.) algorithm deletes the singular points and the strict peak in the numerical integration by variable transformation in step 2 above, and uses Newtons method in the convergence calculation. As a result, an increase in calculation accuracy and reduction of computation time are realized. Some numerical examples demonstrate the effect of the T.S.S. algorithm.

Corrections: A Probability Bound Estimation Method in Markov Reliability Analysis

In many practical systems, the uncertainty of component failure/repair rates results in uncertainty of system failure probability. Concerning a repairable system, uncertainty is evaluated as a probability bound in the Markov process. In practical analysis, the Laplace transform has the advantage of relatively less computing time than that of a numerical method, eg, Runge Kutta. This paper proposes an algorithm for evaluating this uncertainty using the Laplace transform method. This algorithm assumes the Johnson SB distribution for system-failure próbability. Then, the mean and the variance of system-failure probability are obtained using Newtons method and an integral form for calculating parametric differentiation. Finally, the probability bounds are obtained by applying the conventional moment-matching method. A tutorial example is presented at the end of this paper.