On the evaluation of grinding wheel surface condition under different process parameters via in-machine monitoring

doi:10.21203/rs.3.rs-2713493/v1

Download PDF

Research Article

On the evaluation of grinding wheel surface condition under different process parameters via in-machine monitoring

https://doi.org/10.21203/rs.3.rs-2713493/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

The role of the grinding wheel is significant in all grinding operations. Two main indicators related to the process behaviour are the ground workpiece surface roughness and the consumed amount of energy. Both of these factors are affected by grinding wheel topography condition. Thus, an in-deep analysis of the wheel surface condition will extend the understanding of both process behaviour and wheel performance. This condition can be described by the micro-wear mechanisms such as flattening of abrasive grits (wear flat) and adhesion of chips between grits and pores (wear by deposition or loading) in conjunction with the arrangement of abrasive grains. In this sense, in this work a tool for in-machine analysis of wheel surface features such as attritious wear, loading and the number of grains involved in the cutting action is developed. The device is then used in experimental tests in order to validate its ability to analyse the wheel-process behaviour. Researchers have deeply studied the influence of the wheel wear condition on the grinding process, but there is still a lack of knowledge on the effect that both grinding and dressing processes have on the wheel wear condition evolution. The developed tool is also employed for that purpose. Results obtained with the device are consistent with previous works and literature. Therefore, it can be used for a comprehensive understanding of process behaviour during grinding with different wheel-workpiece material combinations and different process conditions, as well as for the analysis and optimization of new abrasives and bonds performances.

Grinding

Dressing

Wear

Image processing

In production of critical parts, the grinding process is usually located following the material-conversion processes such as forming, machining and heat treatment. Then, the grinding process aims to remove hard stock while meeting specifications for geometrical tolerances and surface topography [1]. Therefore, the grinding operation need to ensure a good material removal and the production of precision surfaces. The grinding wheel is a commonly used tool in industry for manufacturing highly finished products. However, the quality progressively deteriorates in the course of grinding, as the wheel wears out.

Dimensional tolerances are directly affected by the wear mechanisms such as bond fracture, grain pull-out, and grain macro-fracture, also referred to as radial wear. Meanwhile, the loss of the surface quality relies mainly on the condition of the wheel surface.

Another characteristic that distinguishes the grinding process from other machining processes is its high energy requirement per unit volume of material removal, known as specific energy. This aspect is mainly due to the grinding wheel topography [2]. Moreover, this specific energy is increased as grinding wheel grains become blunt and as a result, the workpiece surface integrity may be damaged. This phenomenon is known as grinding burn and it can cause cracking, changes in residual stress or the formation of re-hardening zones.

In this way, the surface roughness of ground part and the specific energy, which are two main indicators of grinding behaviour, are both influenced by grinding wheel topography condition. Furthermore, during the grinding process, the condition of the grinding wheel topography continuously changes due to wear mechanisms. The wheel-workpiece material combination, wheel specification, and grinding and dressing parameters all affect the variation of grinding wheel topography condition in the course of grinding.

As the grinding wheel wears out, it loses its cutting ability and the extent of material removal stages varies [3]. The process is then dominated by rubbing and ploughing mechanisms, along with a reduction in the cutting stage. Therefore, knowing the condition of the grinding wheel surface helps in understanding the behaviour of the grinding process. This condition can be well described by means of wear flat, wear by deposition (loading percentage) and the number of active cutting edges. The former is related to the attritious wear referred to the dulling of the abrasive grains as a result of mechanical abrasion, corrosion and diffusion process described by Malkin [4]. This cause an increase in friction between grits and workpiece surface, resulting in a rise in contact temperature. Malkin [4] showed that it also leads to higher grinding forces and lower cutting ability. Another micro-wear mechanism is the wear by deposition, which involves adhesions of workpiece material on the grits and in the pores. This can also increase the contact temperature since it could influence the chip removal mechanism and the cooling effect of the lubricant [5]. Finally, the number of active cutting edges refers to the number of cutting edges that takes part on the cutting action.

Monitoring wheel condition is important as it directly affect the grinding process. Previous studies have analysed the effect of grinding wheel condition on the process variables such as surface roughness, energy consumption, and force generation ( [2], [4], [5], [6], [7], [8], [9], [10]). The overall conclusion is that attritious wear causes a rise in the power consumption, process forces and temperature. Additionally, depending on the process, initial workpiece roughness could experience an improve with the early increment of wear flat, and as the wheel wears out, the number of active grains increases.

There are some previous works where methodologies for direct and indirect measuring of the grinding wheel condition have been developed [11]. Indirect measurements are based on real-time sensor information, such as spindle power or acoustic emission signals, and are more suitable for practical applications. However, they only provide indirect values of the wheel condition and do not allow for a comprehensive understanding of the wear mechanisms and process behaviour. Direct measurement methods use machine vision and image processing to monitor different features of the grinding wheel condition and enable to relate changes in the process behaviour to individual wear mechanisms ( [5], [7], [9] [10] ). Liu [5] presented a tool for measuring loading. Additionally, Puerto [9] measured wear flat and based on the worn grains estimated the number of active grain. Lee [7] estimated the condition of a cBN grinding wheel by means of statistical analysis of numerical values obtained through image processing. The author measured the wear by means of grinding passes but did not related it to any individual micro-wear mechanism. Then Ludwig [10] developed a tool for measuring and evaluating both loading and wear flat area during different stages. Despite the efforts of numerous authors, a tool and methodology that analyses the entire micro-wear mechanisms simultaneously remain to be done.

There are also large number of authors analysing the effect of the grinding wheel condition in the process variables. However, there is a lack of scientific knowledge on the influence of both grinding and dressing conditions into the wear evolution of the grinding wheel surface and hence, into the stability of the process. Klocke [12] analysed the influence of the dressing process on wheel radial wear during grinding and find out that the lower overlapping ratio, the higher the radial wear. Then Prinz [13] did similar analysis of the dressing process with wheel grain size. Despite the efforts, the analysis of micro-wear mechanisms were out of scope. In the work done by Godino [6] the effect of grain microstructure in wear flat evolution is analysed. Besides, the author give a comparison between the two grit types in terms of specific energy and friction coefficient.

Taking all this facts into account, in the current paper a novel complete method is presented for analysing the wheel surface condition. As a starting point, a previous work [14] is taken where the first modules of the tool were developed which were able to quantify the percentage of the abrasive grain concentration, to detect grain pull out and to follow each abrasive grain size on different dressing stages. In this work a new module for quantifying micro-wear mechanisms and number of active grains is added, resulting in complete tool for measuring wheel surface condition. With this tool, a comprehensive understanding of process behaviour during grinding with different wheel-workpiece material combination and different process conditions is possible. In the grinding wheel manufacturing industry the device developed in this work represents a possibility for in-depth analysis of the behaviour and performance of new types of abrasives and binders under different process conditions. In order to validate the tool three grinding experiments batches are performed. The effect of both grinding and dressing processes on the wheel wear condition evolution is analysed through this tool. The influence of dressing traverse rate, wheel speed and direction and grain size on wear evolution are analysed. Finally, the paper aims to link grinding process behaviour to the grinding wheel topography condition as the amount of removed material increases.

First of all, the wheel condition monitoring tool is described followed by the three batches of experimental test in where surface roughness and power consumption have been measured. The results are then discussed.

The main goal of the presented condition monitoring tool is to analyse the characteristics of an abrasive surface such as micro-wear and grain arrangement. Regarding the wheel condition monitoring analysis, a software tool is developed in MATLAB. Figure 1 shows the input-output block diagram of the tool, which is divided into four modules (M) to characterize abrasive surfaces. In this work a conventional grinding wheel is analysed but it is not limited to this type of tools. It could also be employed in belts, stones or dressers. A Godino et al [14] presented the modules M1 to M3 in the online analysis of a resin bonded diamond wheel. The authors developed a methodology for quantifying the percentage of abrasive grain concentration, detecting the grain pull out and tracking each abrasive grain size through dressing stages. In this work, a novel module is added that enables the quantification of the percentage of wear flats and loading, as well as the amount of active grain based on wear, resulting in complete tool for measuring wheel surface condition. Identifying the real amount of grains that take part in the action of cutting is a complex task. In order to simplify this, some authors ( [4], [9] ) have achieved satisfactory results relating the number of grains that present wear to the number of grains that have been active in the grinding process.

The developed module, M4, operate using binary segmentation of 2D optical images to detect the relevant wheel condition-related variables in the surface.

According to the detection of wear flat areas, reflections of flat surfaces under direct illumination are parallel to the emitted light, which makes it possible to identify and isolate them from other reflections. In this work an optical microscope Dino-Lite with a coaxial light tap N3C-A is employed for this purpose. The tap ensures that the emitted light from the camera is parallel to the centre axis, as shown in Fig. 2 focusing on an abrasive surface. With a strategy of reduced light, the bright areas on the image are related to the flats surfaces of the wheel. Thus, wear flat areas are noticeable as small bright areas on the image whereas other features are obscured.

However, the adhered material as wear by deposition (loading) does not necessarily have to be flat. It may appear flat due to tribochemical effects as explained by Godino [6], but in this cases, it would be taken as wear flat. Therefore, for the estimation of loading wear, an elevated light strategy is employed. To isolate loading wear from other features, colour differences are mainly relied upon. Therefore, working directly with colour images is of utmost importance.

Once the strategies are defined, image processing methods are used to extract the surface condition-related variables.

The image processing used in the software consist of binary segmentation of colour images. Unlike other relate works, ( [5], [9], [10] ) where the first step is to convert the image into grey scale, in this work colour images are directly analysed. In return of an increment of calibration time, this approach improves in versatility and accuracy, which is essential for determining wear by deposition. In this way, every different bond-abrasive combination is possible, no matter if they have similar tones. In a grayscale binary segmentation, a single threshold is established between worn and non-worn surfaces, which could lead to the loss of some worn pixels or the inclusion of non-worn pixels. In contrast, the presented software defines upper and lower thresholds for each of the three components that form a pixel in the HSV colour model, as shown in Fig. 3.

Every pixel in HSV model is an 1x3 array where each component corresponds to the H (hue), S (saturation) and V (value) value of the represented colour, respectively. So for a pixel to count as worn surface, the three values must be between the corresponding limits. Attending to the figure, only the Pixel 1 would count as worn surface pixel.

Once the binary segmentation scanning is completed, the module enables the optimization of the image using filters and filling functions. After the optimization, the software provides the percentages of worn surface in terms of wear flats and wear by deposition, as well as the number of active grains based on wear. Figure 4 and Fig. 5 shows the interface after four measurements of a corundum vitrified wheel using a strategy of reduced and elevated light, respectively. Additionally, it is also possible to monitor the wear evolution in different stages.

In this research, a corundum vitrified wheel is dressed before grinding an F5220 steel until excessive wheel wear is detected through different process variables such as grinding power and surface roughness. In order to validate the capacities of the developed tool different test are carried out with different process parameters.

The main objectives of the experimental tests are to analyse the effectiveness of the presented measuring method. The tool is used to analyse the influence of the process conditions on wear evolution and the effect of excessive wear on the behaviour of grinding process.

The tests were performed in a horizontal machining centre, DS630 DANOBAT, equipped with a high-speed adapted for grinding wheels. Hardened steel F5220 was ground with a 4.8% emulsion coolant. Accumulative grinding tests were carried out in each batch to analyse wear evolution. The grinding wheel was only dressed before starting the test in each batch, meaning that the difference between the first and last passes was the grinding wheel’s surface condition. Therefore, the changes in the measured variables are only related to wear. Grinding power was measured during the complete test and surface roughness was analysed after a defined number of passes. Experimental equipment is summarized in ¡Error! No se encuentra el origen de la referencia.. Additionally, images of the surface were taken at intervals determined based on significant changes in the process variables.

Characterization of grinding wheel condition was accomplished using the tool presented in the previous section. Images were captured during the tests using an optical microscope at a magnification of 50x inside the machine. The grinding wheel was automatically positioned for image capture, and two images were taken around the entire circumference of the wheel in each position at 0º, 90º, 180º and 270º. In total 8 images are analysed at each stage. The setup for wheel surface measurement and image acquisition is shown in Fig. 6.

Table 1

Experimental equipment
Machine and equipment
Grinding machine	DANOBAT DS-630 high speed horizontal machining centre
Rotary dresser	Single point diamond dresser from Tyrolit
Data acquisition card	National Instruments USB-4432
Roughness measurement device	Mitutoyo SJ-310 roughness tester
Workpiece	Hardened Steel HRC 60–62 (F5220)
Optical microscope	Dino-Lite Edge AM7517MZT

In order to study the effectiveness of the wheel surface monitoring tool with different dressing parameters, the effect of dressing on the micro-wear evolution of grinding wheels and, therefore, on the grinding process, was analysed through the first batch. There are some works that analyse the influence of the dressing process on the wheel radial wear [12], but despite the efforts, none of which evaluate the wear related to topography. For this purpose, three different industrial dressing parameters are chosen, as shown in ¡Error! No se encuentra el origen de la referencia.. The variation in the dressing traverse feed allows to study the influence of the overlapping ratio U_d on the wear evolution in the subsequent grinding process.

Table 2

Process conditions in batch 1
Test	Grinding wheel	Dressing			Grinding
Test	Grinding wheel	\({V}_{s}\) [m/s]	\({f}_{ad}\) [mm/rev]	\({a}_{ed}\) [mm]	\({V}_{s}\) [m/s]	\({V}_{w}\) [mm/min]	\({a}_{e}\) [mm]
1	MA 100IJ8V489	35	0.3	0.015	35	5000	0.024
2	MA 100IJ8V489	35	0.3	0.015	35	5000	0.024

Then a second batch was performed to evaluate the presented tool with different wheel speeds. The main goal is to analyse effect of the wheel speed on the condition of the wheel surface and its evolution during the grinding process. For that, two grinding conditions were tested (Test 3 and Test 4), with the single difference being the wheel speed. The process parameters are shown in ¡Error! No se encuentra el origen de la referencia.. During the test, both material removal rate and aggressiveness were maintained. Aggressiveness is a dimensionless number suitable for characterizing abrasive processes demonstrated in previous works [15].

Table 3

Process conditions in batch 2
Test	Grinding wheel	Dressing			Grinding
Test	Grinding wheel	\({V}_{s}\) [m/s]	\({f}_{ad}\) [mm/rev]	\({a}_{ed}\) [mm]	\({V}_{s}\) [m/s]	\({Q{\prime }}_{w}\) [mm²/s]	\(aggr\) [-]
3	MA 100IJ8V489	25	0.2	0.015	25	1.25	27
4	MA 100IJ8V489	35	0.2	0.015	35	1.3	25
5	MA 100IJ8V489	35 (up)	0.15	0.015	35 (up)	3	46
6	MA 100IJ8V489	35 (down)	0.15	0.015	35 (down)	3	46

Still in this batch, another two additional test (Test 5 and Test 6) were performed in order to analyse the effect of wheel speed direction on wear evolution. For this purpose, the same grinding parameters were employed in down-grinding and up-grinding strategies. The employed conditions are displayed in ¡Error! No se encuentra el origen de la referencia..

Finally, in order to assess the tool with two different wheel specifications, the effect of grain size has on the wear evolution and grinding process was analysed in the last batch. While Printz [13] analysed the effect of grain size on grinding wheel radial wear, micro-wear mechanisms were not evaluated. To this end four additional test are carried out in this batch, in which two different wheels are used with the only variation being the grain size. Additionally, the behaviour of each tool under two industrial grinding parameters (roughing and finishing) is analysed. The tests information is summarized in ¡Error! No se encuentra el origen de la referencia.. Before each test, the grinding wheel was dressed with a wheel speed of 35 m/s, a traverse rate of 0.15 mm/rev and an infeed pass of 0.015 mm. Each test was ended when visual marks of grinding burn were detected.

Table 4

Process conditions in batch 3
Test	Grinding wheel	Grinding conditions
Test	Grinding wheel	\({V}_{s}\) [m/s]	\({V}_{w}\) [mm/min]	\({a}_{e}\) [mm]	\({Q{\prime }}_{w}\) [mm²/s]
7	46H12V489	35	9000	0.010	1.5
8	46H12V489	35	12000	0.015	3
9	80H12V489	35	9000	0.010	1.5
10	80H12V489	35	12000	0.015	3

4.1. First test batch

First of all, the results obtained from the use of the wheel surface monitoring tool in the first batch are analysed. Figure 7 (a) displays the evolution of wear flat areas with specific amount of removed material. A linear trend of wear flat increment can be seen in both tests of the batch, which is consistent with the findings from previous works on this subject ( [2], [4], [6], [9] ). The initial wear flat percentage value is also in accordance with previous studies with the exception of the value obtained by Godino [6], where the reason for such a difference is explained there.

The effect of the dressing traverse rate in wear flat evolution is analysed, and it can be seen that with small specific removed material, the difference in the generated wear flat area among tests is insignificant. However, at 960 mm³/mm of removed material, attritious wear in the test with the lower traverse rate is 52–58% higher and visual marks of grinding burn are observed, leading to the test being stopped. Dressing with a higher traverse rate, a higher effective surface roughness is formed on the grinding wheel with less cutting edges but higher protrusion [16]. In this work, this effect can be seen by the results obtained from the tool in the Fig. 7 (b) where the estimations of active grain for Test 1 and 2 are displayed.

In previous studies ( [4], [9] ) the increase in the number of active grains with the specific removed material was discussed. These studies demonstrated that as the grains become blunt, their protrusion is reduced and as a result, grains of lower layers begin to make contact. Moreover, as wear flat area increases, the increase of the normal force [4] will cause a growth in the real contact length due to wheel deformation. Using the presented tool, this trend can be observed in Fig. 7 (b). It can be seen that as the amount of specific removed material increases, the number of active grains also increases. The results shown in the figure are consistent with those obtained by Malkin [4] and Shi [8].

Figure 7 (c) shows loading evolution during the tests. Test 1 (\({f}_{ad}\)=0.3) and Test 2 (\({f}_{ad}\)=0.2) exhibit an initial increase until a transition point is reached, after which, a steady behaviour with fluctuations is observed. The order in which test were performed is the main reason for this higher initial value in test 2. In this test, the wheel presented higher amount of chips in deeper pores, originated in previous tests that the dressing process was not capable to eliminate. A higher number of passes would be needed. However, as further demonstrated, this chips in deeper pores has no effect in the grinding process.

With regard to process behaviour, according to literature [3], smaller dressing traverse rate lead to smoother ground surface but to a higher grinding forces and specific energy. This can be evidenced in Fig. 8 (a) and (b) where the monitored process variables are shown. Although the initial ground surface is smoother with the smaller traverse rate, the worse wear evolution (shown in Fig. 7 (a)) causes that from 760 mm³/mm onwards the roughness obtained in the test 1 to be better than that obtained in test 2. In fact, the power consumption is lower (as shown through the specific energy in Fig. 8 (a)) and so does the temperature in the contact zone, slowing down the wear generation.

In order to link the process behaviour to grinding wheel surface condition, Fig. 9 illustrates the surface roughness and specific energy with the corresponding amount of wear. On the one hand, Fig. 9 (a) and (b) show the effect of the wear by deposition (loading). It can be observed that the same wear percentage resulted in different values of surface roughness and specific energy. Therefore, the observed differences suggest that this type of wear has a lesser impact on process behaviour compared to other wear mechanisms under the specific combinations of wheel-workpiece material and process conditions.

On the other hand, the effect of wear flat is shown in Fig. 9 (c) and (d). According to literature, blunt grains change the extent of material removal stages [3]. This results in an increase of the ploughing and rubbing stages where most of the energy goes in form of plastic deformation, with the subsequent reduce of cutting stage. As a result, there is an increase in energy consumption accompanied by a rise in the normal force [4]. The consequence is that the ridges generated in the ploughing stages are of greater amplitude. In roughness values of this magnitudes this can lead to an increase of the measured value [9]. Attending to results, this increase of both surface roughness and specific energy with the percentage of blunt grains can be seen by the measurements obtained from the developed tool.

4.2. Second test batch

In this section, results from the second batch are discussed where the effect of wheel speed is analysed. First, results obtained with the tool are displayed in Fig. 10. Figure 10 (a) shows the evolution of wear flats area with specific amount of material removed. In both test, a linear trend of wear flat increment is noticeable. According to the effect of wheel speed on a grinding process with the same material removal rate and aggressiveness, the process with the lower wheel speed (25 m/s) presents an improved wear evolution (up to 5,5 times).

Tawakoli [17] analysed the expansion of grinding wheels. According to the model, a greater grinding wheel speed results in a greater radial expansion due to centrifugal forces. The author demonstrated how this expansion affected the geometry of the grinding contact zone. The real contact length between wheel and workpiece is increased, and thus, the number of grains involved in the cutting action is expected to be higher. The grinding condition monitoring tool is able to observe this trend, as shown in Fig. 10 (b).

Taking into account the little to none effect loading area observed in the previous batch, this type of wear is not analysed in this tests.

In Fig. 11 results related to process behaviour are shown. The monitored process variables present insignificant differences between tests. This is expected due to having employed the same material removal rate and aggressiveness according to Badger [15].

However, considering process behaviour and grinding wheel condition, Fig. 12 displays a linear increase trend of both surface roughness and specific energy wit wear flat. The process with the lower speed generates 20–35% higher Ra value and consumes 35–40% more energy for the same wear flat value. This implies that wear flat has an influence in the specific energy and surface roughness but there are other factors such as process kinematics that should be also considered.

Finally, apart from the effect of wheel speed magnitude, the speed direction is also analysed through Test 5 and Test 6 from ¡Error! No se encuentra el origen de la referencia.. The results are shown in Fig. 13, which indicates a linear increasing trend with similar slope for both. Nevertheless, the values in up grinding tend to be 2,3 times higher. According to Rowe [3], up grinding is less aggressive to grain, and rubbing and ploughing stages continues for a greater extent, leading to faster blunting of the grains. Contrary to this, in down grinding, the first contact is more aggressive and tends to keep the wheel sharper. This effect can be observed with monitoring tool developed in this work.

4.3. Third test batch

In this section, results from last batch are analysed. The results are obtained using the wheel condition monitoring tool for two wheels of different grain size (mesh number 46 and 80) under two process conditions each; roughing conditions (Q3) and finishing conditions (Q1.5). These conditions are summarized in ¡Error! No se encuentra el origen de la referencia.. In Fig. 14 measurements of the wheel condition are shown. According to Fig. 14 (a), in all four cases, the wear flat presents a linear increase trend with the specific removed material. The wheel with a grain size 80 in finishing conditions presents the less wear during the first 75 mm³/mm and shows the most favourable evolution with grinding passes, up to 57% more favourable than the second. Following, the grinding wheel with grain size 46 under finishing conditions and the wheel with grain size 80 under roughing conditions present similar wear evolution and up to 45% more favourable than the last case. Overall, it has been observed that both bigger grain size and roughing conditions generate more attritious wear.

It is known that increasing the grain size decreases the number of cutting edges [16]. Attending to Fig. 14 (b), this trend can be seen with the results obtained in this work with the tool. The wheel with smaller grain size presents a higher value of active grains during grinding.

Considering the little to none effect wear by deposition showed in the previous batch, this type of wear is not analysed in this tests.

Analysing the process behaviour, on the one hand, linear increments in the specific energy are observed in all four tests as removed material increases. T8 consumes the least amount of energy (39 J/mm³) with 75 mm³/mm of specific removed material, but it also generates the highest amount of wear, as shown in Fig. 15 (a), and thus had the highest increase in specific energy with grinding passes. Grinding burn is detected in this test with 1276 mm³/mm and a specific energy of 85 J/mm³. T10 and T7 have a similar increasing trend, but T10 consumes 35% less energy. Grinding burn is detected in T10 with 976 mm³/mm and 56 J/mm³ of specific energy, while in T7 it is observed with 1650 mm³/mm and a specific energy of 81 J/mm³. Finally, the lower trend of increase is observed in T9. In this test grinding burn is the later detected with 2300 mm³/mm specific removed material and 86 J/mm³ of specific energy.

According to the surface roughness Ra from Fig. 15 (b), it can be seen that the best surface roughness is obtained with the small grain wheel under finishing conditions, as expected, whereas the worse value is obtained with the big grain wheel under roughing conditions. Once again, the other two test presented similar results. The reason for this behaviour is likely to be found in the maximum undeformed chip thickness value \({h}_{cu\_max}\). This parameter has been modelled in other works [3] as:

\({h}_{cu\_max}=2·\frac{1}{C·r}·\frac{{V}_{w}}{{V}_{s}}·\sqrt{\frac{{a}_{e}}{{d}_{e}}}\) [1]

Where \(C\) is the number of active cutting edges and \(r\) is the size factor of the grains. In this work \(C\) is shown in Fig. 14 (b) and \(r\)is unknown but the same for both wheels. Finally, \({d}_{eq}\) is the equivalent diameter, in surface grinding equals the wheel diameter.

This parameter gives an idea of how deep the grains penetrate into the workpiece. Therefore, the higher the thickness, the deeper the grains penetrate into the workpiece, resulting in a greater peaks generated on the surface. Figure 16 displays the approximation of the maximum chip thickness multiplied by the size factor \(r\), along with the specific removed material. As observed, T9 generates the lowest surface roughness, and it also the process with the smallest maximum chip thickness. This means that in this process, the grains penetrate less deeply into the workpiece surface and the generated peaks on the surface are of smaller amplitude. On the other hand, T8 is the process with greater maximum chip thickness, resulting in the higher roughness values.

In Fig. 17 roughness and specific energy variation with wear flat is shown for all four test. Linear increments are observed in both variables with the wear flat. Thus, wheel condition directly influences the surface roughness and the energy consumption and should be included in any energy or surface modelling. However, the differences between test makes clear that there are other factors that has to be taken into account such as wheel grain size and process conditions.

Finally wheel behaviour under different process conditions is discussed. On the one hand, with regard to the roughing conditions, it can be seen that the wheel with grain size 46 has removed 24% more material before grinding burn is observed compared to the wheel with a smaller grain size, but with a higher surface roughness. Since surface roughness is less relevant in roughing steps, the wheel with a bigger grain size is more suitable for this process. On the other hand, under finishing conditions, the wheel with smaller grain size removes 29% more material until grinding burn is detected and achieved a smoother surface with the same wear level as compared to the bigger grain size wheel. Moreover, it presents an improved wear evolution. Thus, for finishing steps, the wheel with a smaller grain size is more appropriate.

From the results obtained on this work the following can be concluded:

A complete methodology is developed to in-machine analysis of wheel surface features. The module that is added in this works enables monitoring attritious wear, wear by deposition and estimation of the number of active grains based on wear. The information of process behaviour obtained through this tool is consistent with grinding literature. Thus, he tool can be used for a comprehensive understanding of process behaviour during grinding with various wheel-workpiece material combinations and different process conditions, as well as for the analysis and optimization of new abrasives and new bonds performances.
With regard to process behaviour analysis, the results obtained using the grinding wheel surface monitoring tool show that both surface roughness and specific energy tend to increase with wear flat in all combinations wheel-workpiece material and process conditions analysed in this work. The results also indicate that as the amount of removed material increases, the number of active grains also increases.
According to wear by deposition, it can be assumed that it has little to none effect on process variables compared to attritious wear, given the employed wheel specifications and workpiece material.
Finally, analysing the influence of process parameters on wheel surface condition, it was found that a higher dressing traverse rate, lower wheel speed and smaller grain size lead to a less pronounced wear flat evolution. In contrast, a lower dressing traverse rate, higher wheel speed, and smaller grain size resulted in a higher number of grains participating in the cutting action.

Results will be useful to extent process understanding and to analyse both process behaviour and grinding wheel performance.

Funding The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Competing interests The authors declare no competing interests.

Author Contributions All authors contributed to the research and the reviewing of the paper. The first draft of the manuscript was written by Arkaitz Muñoz.

Hashimoto F, Yamaguchi H, Krajnik P, Wegener K et al (2016) Abrasive fne-fnishing technology. CIRP Ann Manuf Techn 65(2):597–620
Guo C, Shi Z, Attia H, Mcintosh D (2007) Power and wheel wear for grinding nickel alloy with plated CBN wheels. CIRP Ann - Manuf Technol. https://doi.org/10.1016/j.cirp.2007.05.079
Rowe WB (2013) Principles of modern grinding technology. William Andrew
Malkin S, Cook NH (1971) The wear of grinding wheels: part 1— attritious wear. J Manuf Sci Eng 93(4)
Liu C-S, Ou Y-J (2020) Grinding wheel loading evaluation by using acoustic emission signals and digital image processing. Sensors 20(15):4092
Godino L, Pombo I, Sanchez JA, Alvarez J (2018) On the development and evolution of wear flats in microcrystalline sintered alumina grinding wheels. J Manuf Process 32:494–505. https:// doi.org/10.1016/j.jmapro.2018.03.023
Lee ET, Fan Z, Sencer B (2021) Estimation of cBN grinding wheel condition using image sensor. Procedia Manufact 53:286–292
Shi Z, Malkin S (2006) Wear of electroplated CBN grinding wheels. J Manuf Sci Eng-Transact ASME 128(1):110–118
P. Puerto (2013), Desarrollo de sistemas de control inteligente para incrementar la productividad de procesos de rectificado y asegurar la integridad superficial de la pieza - Dissertation, Mondragon Unibertsitatea.
Azarhoushang, B., & Ludwig, S. (2019). In-process grinding wheel wear evaluation using digital image processing. International Journal of Abrasive Technology, 9(2), 99–112.
Kalpana K, Arunachalam N (2018) A digital twin for grinding wheel - an information sharing platform for sustainable grinding process. J Manuf Sci Eng 141(2). https://doi.org/10.1115/1. 4042076
F. Klocke, J. Thiermann and P. Mattfeld (2015) Influence of the dressing process on grinding wheel wear, Prod. Eng. Res. Devel., 9 (5): 563-568
Prinz S, Trauth D, Mattfeld P, Klocke F (2018) Infuence of diferent grinding wheel and dressing roller specifcations on grinding wheel wear. Prod Eng 12:441–448
Leire Godino, Arkaitz Muñoz, Iñigo Pombo, Jose Antonio Sánchez & David Barrenetxea (2022) In-machine data acquisition for evaluating the conditioning efficiency of resin-bonded super-abrasive grinding wheels, International Journal of Computer Integrated Manufacturing, DOI: 10.1080/0951192X.2022.2095673
J. Badger, R. Draˇzumeric, P. Krajnik (2021), Application of the dimensionless Aggressiveness number in abrasive processes, Proced. CIRP 102 361–368, https://doi.org/10.1016/J.PROCIR.2021.09.062
Klocke F (2009) Manufacturing processes 2 Cutting
Tawakoli T, Reinecke H, Vesali A (2012) An experimental study on the dynamic behavior of grinding wheels in high efficiency deep grinding. 5th CIRP conference on high performance cutting 2012, Procedia CIRP 1, p.p. 382–387
A. Muñoz, L. Godino, J. A. Sánchez, I. Pombo and I. Tolosa, “Resin Bonded Diamond grinding wheels conditioning using SiC rotary dresser,” IOP Conf. Series: Materials Science and Engineering , vol. 1193, 2021.
I. Marinescu, M. Hitchiner, E. Uhlmann, Rowe W.B and Inasaki I., Handbook of machining with grinding wheels, CRC Press, 2006.
S. Malkin and C. Guo, Grinding technology: theory and application of machining with abrasives, Industrial Press Inc., 1986.

Download PDF

Version 1

posted

You are reading this latest preprint version

On the evaluation of grinding wheel surface condition under different process parameters via in-machine monitoring

Status:

Version 1

Abstract

Figures

1 Introduction

2 Grinding Wheel Condition Monitoring

3 Experimental Setup

4 Results And Discussion

4.1. First test batch

4.2. Second test batch

4.3. Third test batch

5 Conclusions

Declarations

References

Status:

Version 1