Masafumi Sasaki

System Availability and Optimum Spare Units

The steady-state availability of a repairable system with cold standbys and nonzero replacement time is maximized under constraints of total cost and total weight. Likewise the cost can be minimized under constraints of steady-state availability and total weight. A new, more efficient algorithm is used for the constrained optimization. The problem is formulated as a nonlinear integer programming problem. Since the objective functions are monotone, it is easy to obtain optimal solutions. These new algorithms are natural extensions of the Lawler-Bell algorithm. Availability is adjusted by the number of spares allowed. Other measures of system goodness are considered, viz, failure rate, weight, price, mean repair time, mean repair cost, mean replacement time, and mean replacement cost of a unit.

Reliability of 3-state device systems with simultaneous failures

This work is concerned with computing the reliability of redundant systems of three-state (e.g., good, failed-open or failed-short) devices. Allowing for multiple failures, formulas are presented for calculating the reliability of a parallel-series system or a series-parallel system which is formed from three-state, non-DFM (dual failure-mode) devices, and which might require multiple paths to function. The words series and parallel are used in the layout-diagram sense. The formulation encompasses cases which have been analyzed by complex methods in numerous papers on the reliability of three-state DFM device systems. The effects of dominant failure-modes on the relationship between reliability of a series system and that of a parallel system, and on the relationship between the reliability of a parallel-series system and that of series-parallel system are presented. >

Improvement of Instantaneous Availabilities by Decreasing Delay Time

Many situations exist where a system cannot go to a repair station immediately when a unit fails. This paper models a 1-out-of-2:G hot standby system which is used intermittently. The system in Model I is a 2-unit parallel redundant system where only one user is allowed in the system. However, in Model II, two users are allowed to use the system. Model II uses two units independently (it is not a 2-unit parallel redundant system) when two users are in the system. Model I is equivalent to Model I when there is only one user in the system. Two problems are discussed: 1) the effect of a delay time on instantaneous system availability, 2) the possibility of trade-off between the delay time and the failure rate of a unit on instantaneous availability.

Optimal spare allocation problem based on the interval estimate of availability

Abstract The purpose of this paper is to discuss the optimal spare allocation when the distribution of failure-free operating time is known to be exponential and the only unknown quantity is a failure rate, which needs to be estimated from the results of a reliability test. In this paper, the failure rates of units in a system are treated as random variables, and their distributions are obtained using the Bayesian approach. Then the distribution of availability is derived. Finally, the optimal spare allocation problem is formulated using the interval estimate of availability.

Optimal Inventory Problem of a Repairable K-out-of-N:G System

The optimal inventory problem of a repairable K-out-of-N:G system is discussed. When a unit fails, it is repaired or discarded according to the extent of damage. Spare units are replenished at the replenishment times. The optimal inventory problem is to determine the number of spare units replenished at each replenishment time, which minimizes the total system cost constrained by system availability. This problem is formulated as a nonlinear integer programming problem. A new algorithm which is a modification of an algorithm proposed earlier by the authors is used for the problem.

Reliability of systems with two failure modes by using structure functions

Abstract This paper deals a three-state (normal, short-mode and open-mode failures) device network that is required to control the quantity of s-t flow. If the control policy is to pass the flow or to shut off the flow, i.e. the amount of s-t flow equals one, this network system is a conventional three-state system. In order to obtain the reliability of the network system, we propose a new structure function. Some examples are presented to show that using the structure function simplifies the reliability calculation.

Reliability analysis for a two-echelon repair system considering lateral resupply, return policy and transportation times

Abstract This paper considers a 2-echelon repair system. This system consists of identical items, one central depot and several bases where the items are operated. When an item fails in a base, the failed item is replaced with a spare item, if available, and is sent to the depot for repair. The purpose of this paper is to build models for evaluating the effectiveness of return policy and lateral resupply on reliability improvement of the items.

Optimal Selection Methods for Failure & Repair Rates Using Nomographs

This paper explains the optimal selection methods using nomographs to solve two essentially different problems. The one is the problem of unit level, and the other is the one of system level. The unit level assumes that the cost information as a function of failure rate ¿ and repair rate ¿ are empirically known. The paper presents a method, by which a nomograph is used to select easily the optimum pair from the infinitely many pairs (¿, ¿) of feasible solutions, to gain the required unit availability at minimum cost of this assumption. At the system level, the system is composed of n serial i-units which are selectable from a group provided for each i-unit (i = 1, 2, ..., n), several different repair plans are available for the unit. Each unit has a specific failure rate and associated cost, and each repair plan has a specific repair rate and associated cost. There are service personnel for each unit. The failure and repair rates are constant. The paper presents: 1) A method using nomographs to select the optimum pairs from the many pairs (¿i, ¿i) (i = 1, 2, ..., n) of feasible solutions, to gain the required system availability at the minimum system cost, 2) A method to select the optimum pairs, from the many pairs (¿i, ¿i (i = 1, 2, .

Availabilities for a Fixed Periodic System

Availability for a system with fixed periodic time intervals of demand and nondemand is defined for two models which might apply to vehicles, vessels, or aircraft. A specific operational readiness for the model is shown. Figures and a Table are presented in a Supplement as a result of the Monte Carlo simulation; the failure-time and repair-time distributions for the system are exponential, and Weibull or log-normal, respectively.