Dynamic benchmarks for operations evaluation in an aluminium ingot mill

New internal benchmarks are developed as performance metrics for operational activities in the Ingot Mill of an aluminium smelter. A holistic model for performance of complex organisations is first proposed and used to identify the need for such measures for both completeness and as practical tools for improvement management. These benchmarks reflect the dynamic performance levels of homogeneous resource groups of the organisation: the Equipment system, the Process, and the Workforce. They are developed using a wide variety of shop-floor information including soft data with different levels of information deficiencies. In order to systematically combine the diverse forms of shop-floor infonnation, and embed the maximum knowledge possible in the benchmarks, a new system of information semantics is proposed for identifying the different types of information in a numerical quantity. Using these information semantics, a general method for soft reliability analysis is then developed and used to compute the Equipment benchmark. To compute the Process, Workforce, and overall Plant benchmarks, an information fusion technique is employed which enables different fonns of factor interdependencies to be modelled in the synthesis process. In general, the computational procedures represent a new syncretic decision theoretic which addresses several real-world complexities that present obstacles to information synthesis by the classical approaches to decision analysis. While the procedures are specifically designed for the Ingot Mill, they can be adapted to the needs of different sites, such as assembly line operations, to enable simplified top-down improvement management.