Machining of Hard Materials

Machining is most widely used to transform the material into the product of desired shape and size by the mechanism of removing excess material. Machining involves group of processes, wherein the excess material is removed from the work specimen in sequential steps with the help of cutting tools (either single point or multi-point). It is to be noted that machining with a single-point cutting tool uses well-defined tool geometry (i.e. cutting edges (honed, sharp, chamfered) possessing different faces (rake, flank, etc.)), whereas grinding process uses abrasive wheel with multi-point micro-cutting edges having undefined geometry [1‐3].

Over the years, machining industries are continuously striving to manufacture the parts at reduced cost and improved quality. This can be achieved by selecting appropriate set of tool–work materials and effective modelling and optimization of the process. Optimized grades of high-speed steel (HSS) are used to be treated as ultimate tool material till the 1930s [1]. However, American metalworking industry had shown three-time improvement in productivity with the use of same machines and manpower during the period 1939–1945.

Planning and conducting experiments is the key in effective monitoring of system, which leads to success in manufacturing. The traditional approach of experimental study (i.e. one factor at a time, OFAT) requires more number of experiments and consequently consumes more resources. Moreover, the interpretations and analysis that can be made from the experimental data are also limited. Design of experiments (DOE) is a statistical tool, which uses well-planned set of experiments to collect the input–output data. Further, DOE can be used to analyse the experimental data, establish input–output relations, and optimize the process. Figure 3.1 shows the general steps followed in designing a statistical-based experiment.

In the Mid of 1950s, artificial intelligence was emerged to solve practical problems in engineering domain by using tools, developed based on human intelligence. Genetic algorithm ‘GA’, artificial neural network ‘ANN’, and fuzzy logic are some AI-based soft computing tools used to predict and assist in the control of manufacturing processes. Today, huge money is spent throughout the globe on the development of AI technology to assist manufacturing industries.

In real-life engineering problems, conducting practical experiments and collecting experimental data for analysis and evaluation in order to attain optimal solutions are difficult as compared to data-driven optimization of mathematical functions. In particular, the numerical modelling and simulation process yield solutions and the duration may vary from few seconds to hours depending on the complexity of problems to be solved. Moreover, the solution obtained may or may not be the global optimal solution. Numerical modelling and simulation task can only predict the outputs for set of inputs and needs many try-error runs, which may not yield optimal solutions. On the other hand, optimization tools are capable to locate the global solutions with very less computational efforts, iterations, and time.