The global marketplace is highly competitive and organisations, who want to survive long-term, have to continuously improve, change and adapt in response to market demands. These improvements should focus on cost cutting, increasing productivity levels and quality and guaranteeing deliveries in order to satisfy customers. Total Productive Maintenance (TPM) is one method, which can be used to achieve these goals. TPM is a change management approach that involves employees from both production and maintenance departments. The purpose is to eliminate major production losses by introducing a program of continuous and systematic improvements to production equipment. TPM should be developed and expanded to embrace the whole organisation and all employees should be involved in the process as members of improvement teams. This paper gives a short description of the development of TPM and the TPM implementation process. Findings are reported from a case study in which one of the authors had the possibility of following and guiding a company through their TPM implementation. The implementation process takes several years and the research has focused on the initial three years. The study demonstrates that driving forces, obstacles and difficulties often are dependent on the organisation, its managers and the individual employees.

This paper considers an aperiodic preventive maintenance (PM) model for repairable systems, in which the time intervals between two consecutive preventive maintenances are unequal. To propose such an aperiodic PM model, we assume that each PM reduces the current hazard rate by a certain amount which depends on the number of PMs performed previously. If the system fails between PMs, the minimal repair is performed and the hazard rate remains unchanged after the repair. We give the exact expressions for the hazard rate function for the aperiodic PM model. Based on the proposed aperiodic PM model, we suggest the maximum likelihood method to estimate the parameters characterizing the model and apply the method to the case of Weibull distribution. Numerical examples for estimating the parameters are presented for the purpose of illustration.

In a recent paper Iskandar ＆ Sandoh (1999) studied an opportunity-based age replacement policy for a system which has a warranty period (0,S]. When the system fails at age x $\leq$ S a minimal repair is performed. If an opportunity occurs to the system at age x, S $\leq$ x $\leq$ T, we take the opportunity with probability p to preventively replace the system, while we conduct a corrective .replacement when its fails in (S,T). Finally, if its age reaches T, we perform a preventive replacement, Under this policy the design variable is T. For the case when opportunities occur according to a homogeneous Poisson process, the long-run average cost of this policy was formulated and studied analytically by Iskandar & Sandoh (1999). The same problem is here analysed by using a graphical technique based on scaled TTT-transforms. This technique gives, among other things, excellent possibilities for different types of sensitivity analysis. We also extend the discussion to the situation when we have to estimate T based on times to failure.

In a resent paper, Na and Kim(2000) develop a family of test statistics for testing whether or not the mean residual life changes its trend based on complete data and show that the new tests perform better than previously known tests. In this paper, we extend their tests to the randomly censored data. The asymptotic normality of the test statistics is established. Monte Carlo simulations are conducted to compare our tests with a previously known test by the power of tests.

We consider a system whose inherent life follows an Erlang distribution, which is subject to two heterogeneous random shocks. Minor shocks arrive according to a renewal process and each causes the system to fail independently with a certain probability. A major shock whose interarrival times follow an Erlang distribution causes the system to fail with probability one. The Laplace transform of the distribution of the time to system failure is derived in a functional form of the Laplace transform of the interarrival time distribution of minor shocks. An algorithm is given for the computation of the moments of the time to system failure.

This research considers a system which has an ultimate terminal event such as death, critical failure, bankruptcy together with a certain indicative events (temporary malfunction, special treatment, kind of defaults) that frequently occurs before the terminal event comes to the system. Some investigation of a model for the corresponding bivariate data of the system have been done with an explanation of the situation in terms of two continuous variables instead of continuous-discrete variables and some other properties. Also an analysis has been carried out to evaluate the effect of intermediate observation of occurrence of indicative event so that the result can be used for a possible suggestion of an intermediate observing schedule.

One of the most important quantities in reliability theory is the expected number of renewals of a system during a given interval. This quantity, the renewal function, is used to determine the optimal preventive maintenance policy and to estimate the cost of a warranty. In this paper we study a parametric approach for a renewal function. The simulation study is presented to compare the relative performance of the introduced estimators of a renewal function. And we show that the proposed parametric estimator performs well.

Accelerated testing consists of a variety of test methods for shortening the life of products or hastening the degradation of their performance. This paper presents practical, modern statistical methods for evaluating the reliability of Nickel-Cadmium batteries at their design temperature of 2$0^{\circ}C$ by accelerated life test. Batteries have been life tested at three high temperature conditions, 50, 60, 7$0^{\circ}C$, respectively to yield failures quickly. The failures have been observed and judged by means of charge and discharge current integration. Analyses of life data from those conditions resulted in the Weibull distribution, which has been verified on the ground of the Kolmogorov-smirnov test and the pairwise t-test. Life data are modeled according to the Arrhenius life-temperature relationship. The mean life of tested batteries is assessed at about 590 cycles, and the activation energy of this chemical reaction is concluded to be 0.39eV as results. This study provides procedures for estimating the reliability of batteries in a short period, which has little been possible in domestic industries. The results can be applied in many fields such as proof testing, acceptance testing, and estimating assurance periods.