Observed and Un-Observed Covariate Effects on Baseline Hazard Rate - Case study: Jajarm Bauxite Mine

In the mining industry, different departments undertake the control task of a system. The failure data from different sections in a variety of formats are collected in different times and operation conditions and these heterogeneity problems will result in the collected data and the data bank. While the studies that are done so far, two items of time and environmental effects have been examined and in fact, the data are assumed homogeneously. However, a part of heterogeneity by the help of observed covariate is omitted, but the problem starts where entering all environmental conditions is not possible. Since some factors such as the effect of management decisions, different brands, operator skills cannot be recorded or they cannot be quantifiable; and in spite of being observable, some are not recorded or not classified. These categories of environmental factors that lead to heterogeneity in the data are categorized as the unobserved covariate. For this purpose, in this paper, a coherent approach was proposed for system performance analysis based on the reliability index with the impact of environmental conditions. Environmental conditions are considered in two observed categories, which include factors such as road conditions, weather conditions, rock types, and unobserved ones that include heterogeneity. The proposed approach was evaluated with data from the mining fleet of Jajarm Bauxite mine consisting of 5 trucks. The significant difference between the results of the inclusion of heterogeneity with its lack of consideration indicates the importance of the impact of unobserved covariates on determining the system's behavior.