Creep-feed grinding of tungsten carbide using small diameter electroplated diamond wheels

This paper, by Z. Shi, H. Attia, D. Chellan and T. Wang, is concerned with an experimental investigation into creep-feed grinding of tungsten carbide using electroplated diamond wheels smaller than 25.4 mm in diameter. This work was motivated by applying plated diamond wheels to the grinding of deep slots or grooves in tungsten carbide components accessible only by small wheels. The objective is to explore the feasibility to grind tungsten carbides at large depths of cut, reasonable removal rates, and wheel life with small plated diamond wheels and to obtain some practical process parameters. Straight surface grinding experiments were conducted on tungsten carbide bars over a wide range of grinding conditions with waterbased grinding fluid. Grinding power, forces, workpiece surface roughness, and radial wheel wear were measured. Depths of cut as high as 4 mm corresponding to a specific material removal rate of 3.2 mm2/s were achieved with a wheel of 19 mm in diameter. Experimental results for specific energy, forces, and surface roughness are discussed in detail. It was shown that small plated diamond wheels are capable of grinding deep slots or grooves in tungsten carbides efficiently.

Tungsten carbide materials possess extraordinary physical and mechanical properties such as high hardness and rigidity, high compressive strength, and excellent resistance to wear and heat. They are commonly used as materials of cutting tools, punch-and-die sets, and wear resistance components. Due to their extremely high hardness, tungsten carbides are generally machined by grinding with diamond wheels [1]. Both resin bond and single layer plated diamond wheels have been used for grinding tungsten carbides, with the resin bond wheels mainly for plain or simple workpiece profiles [2 - 7] and the plated wheels for grinding of both simple and complex profiles [7 - 9]. Despite of being very brittle, tungsten carbides were found to exhibit extensive plastic deformation during grinding

as revealed by flow-typed chips, large compressive residual stresses in ground surfaces, and high energy consumed on the plowing of diamond grains against workpieces [1, 2]. Typical wheel speeds with resin bond wheels

reported were in the range of 15 to 30 m/s [3 - 5]. Creep-feed grinding of tungsten carbides with resin bond wheels showed benefits long time ago [6, 7]. A cost saving of about 83% was achieved in creep-feed grinding of solid tungsten carbide gear cutting hobs as compared with conventional pendulum grinding [6]. It should be noted that profile grinding with wheels that need truing such as vitrified or resin bond wheels is usually limited by the complexity of the ground profiles. If the corresponding wheel axial profile is dependent on the wheel diameter, as in the case of helical flute grinding, the workpiece profile cannot be ground precisely to the geometry without truing the wheel profile based on the wheel diameter, which becomes smaller and smaller due to the removal of wheel material in truing operations. Electroplated superabrasive wheels, either with diamond or cubic boron nitride (CBN) abrasives, are manufactured with a single layer of abrasive grains held on the wheel hub by an electroplated nickel bond. Unlike other types of wheels, plated wheels are not periodically trued or dressed and are more suitable for profile grinding. Grinding of tungsten carbides with plated diamond wheels was introduced long before the invention of Cubic Boron Nitride (CBN) abrasives [8] and was successfully applied to the grinding of various tungsten carbide components [9]. However, there was very little work reported concerning the grinding performance of tungsten carbides with plated diamond wheels as compared with the grinding of metallic materials with plated CBN wheels. The limited amount of literature suggested that the wheel speeds are around 20 m/s, and workspeeds from 0.2 to 0.4 mm/s in creep-feed grinding of tungsten carbide with plated diamond wheels [7 - 9].

Performing grinding operations in multi-tasking machining centres is one of the recent increasing trends for producing components with complex geometric features [10]. Plated wheels are more attractive in this regard since they do not need truing/dressing and can be programmed as milling cutters. Simulated grinding of nickel alloy turbine blade slots using small profiled plated wheels on machining centres showed very promising prospects [11].

Results

As mentioned above, grinding tests were conducted only after the initial wheel run-in. The measured radial wheel wear profile at the end of wheel run-in is presented in Fig 3, which shows an average wear depth of about 40 μm. In addition to the wear depth, this profile also reveals the progressive smoothing trend of the wheel surface with continued grinding as indicated by the larger peak-to-valley value of 40 μm produced by the reference sections of the wheel width as compared to the smaller value of about 30 μm produced by the section used for grinding. Wheel wear measurements following an accumulated volumetric material removal per unit width of 480 mm2 at the end of tests indicated that there was almost no measurable change the in wear depth and peak-to-valley values. Therefore, any change in grinding power, forces, and surface roughness as presented below can be attributed to the effects of grinding parameters. Optical microscope observations during grinding intervals and at the end of tests revealed that diamond grains were worn down by grain fracture and grain pullouts. There was virtually no wear flat contrary to what was found in the grinding of ceramics with plated diamond wheel [13]. This wheel wear mechanism was further confirmed by the wheel topography scanned on the used wheel at the end of tests as shown in Fig 4. Net power, horizontal and vertical forces for all the grinding conditions are presented in Fig 5 as plots versus workspeeds for each depth of cut. In every case, the power and forces tended to increase with workspeeds and depths of cut. However, power increased much faster with the larger depth of cut and slower workspeeds than with the corresponding smaller depth of cut and faster workspeeds of equal material removal rates. It can also be seen in Fig 5 that the vertical force was bigger than the horizontal force with the smaller depth of cut of 2 mm. This relationship was, however, reversed with the larger depth of cut of 4 mm. Measured surface roughness are summarised in Fig 6 including the arithmetic average value Ra measured after each grinding pass and after three spark-out passes. It can be seen from this figure that surface roughness fuctuates around 3.5 μm within a relatively small range. There was no apparent trend on the effects of depths of cut, workspeeds, and spark-out passes on the Ra value.