Statistically Validated and Optimized Tabu Search Estimation of Cutting Tool Life in Turning

In machining, tool life is the measure that gives manufacturers fundamental indications on useful tool utilization. In this work, investigated is the influence of turning parameters such as spindle speed, depth of cut, and tool feed on tool life. Literature-reported studies typically utilize Taguchi orthogonal array method for experimental design while tool-wear estimates utilized mathematical expressions by means ofregression analysis. In this work, the authors combine a tabu search (TS) algorithm with the traditional regression analysis (REG) to introduce a combined (TS-REG) scheme that minimizes a weighted sum of the outputs that represent different measures of turning quality. The aim in this paper is to show that the proposed novel combination TS-REG reaches better fitted models in minimal time as compared with other statistical methods. Tool life estimates are plotted versus different turning process parameters. The results that represent optimized tool life of the process are compared to literature reported experimental values. Good correlations confirm desirable efficiency and the ability of the TS-REG method in estimating tool life in turning. In order to determine statistical significance decisions, both p-values and mean absolute percentage error (MAPE: mean of the relative absolute error) were employed. The TS-REG method aids in determining the optimal set of parameters for any combination of the weighting factors resulting in improved tool life.