1 Introduction

Wood-based materials, as a kind of environmental protection biomass material [1], were widely applied in furnishings field, such as doors and windows, fencing and flooring in terms of their flame retardancy, moisture resistance, insect prevention and so on properties [2,3,4,5,6,7,8]. However, for the applications in engineering field, metal-matrix composites are still mainly used due to its high hardness, stiffness, specific strength and low thermal expansion [9]. In order to expand the application of wood-based materials in engineering industry, magnesium oxide particle reinforced wood fiber-matrix composite has emerged, which inspired by a founding that claims particle-reinforced composites have advantage of easy manufacturing and low cost [10].

Magnesium oxide particle belongs to one kind of ceramic materials. It is difficult to cut ceramic materials due to its high hardness, leading to machine problems in surface quality [11,12,13]. Therefore, surface integrity of magnesium oxide particle reinforced wood fiber-matrix composite is still desired to investigate for expanding its application [14].

A series of surface integrity of particle-reinforced metal matrix composites was in focus by many researchers. Surface integrity of Silicon nitride particle-reinforced aluminium alloy using PCD tools was investigated through scanning electron microscope, it is concluded that the defect of machined surface made form particle-reinforced were mainly voids, scales, and scratches. In addition, an affected surface layer with a thickness of about 20–35 μm was also observed [15]. Cutting mechanism and damage behavior of particle-reinforced composites was studied by Yu, they found the failure mode and mechanism of machined surface [16]. Above studies, surface formation mechanism of particle-reinforced metal matrix composites is quite clearly. However, the surface formation about particle-reinforced wood-based composites was not clear, because the anisotropy of wood is compared to the isotropy of metal.

The effect of cutting parameters on surface quality still need to be addressed [17,18,19]. Silicon nitride particle-reinforced aluminium alloy machined by PCD tools in high-speed milling was investigated. Their results show that an increased cutting speed as well as feed rate can lead to a severe surface defect. It is also concluded that different superimposed effects, comprising particle pull-out as well as the subsurface voids within the surface layer results in the formation of deep voids [20]. The influence of cutting speed, feed per tooth and cutting depth on surface roughness of machined surface has been investigated. They founded that an increased feed per tooth and a decreased cutting speed led to an increasing surface roughness [21]. The optimal cutting parameters of cutting speed, feed and cutting depth when machining aluminium silicon carbide particulate metal matrix composite were investigated, they found cutting speed and cutting depth have the most contribution to cutting [22]. Although above studies concluded cutting speed and cutting depth have significant influence on surface integrity, the surface formation mechanism of under these cutting parameters still need to be revealed.

As mentioned above, surface quality of particle-reinforced composites has been investigated a lot, and applied in many engineering fields. But until now, as a new kind composite, MgO-particle-reinforced wood fiber-matrix composite, there is still no research about the machinability of this composite. In order to expand its application in engineering materials, it is desirable to investigate its surface integrity under different cutting conditions.

In this experimental study, particle-reinforced wood-based composite was selected as objective. Spiral up-milling tests were carried out to investigate surface formation mechanism using PCD tool. Surface defects were determined by images of scanning electronic microscope.

2 Material and Methods

2.1 Materials

MgO-particle-reinforced wood fiber-matrix composite, was used as milling object in this study. The raw material was supplied by HOMAG China Golden Field Ltd., China. The workpieces were rectangular blocks of 140 mm × 80 mm × 7 mm (length × width × thickness). which contained 55.0 wt.% wood fiber, 36.8 wt.% MgO, and 2.3 wt.% SiO2, as well as 5.9 wt.% some additives such as pigments. A representative microstructure of MgO-particle-reinforced wood fiber-matrix composite is shown in Fig. 1. Physical and mechanical properties of the workpieces are listed in Table 1.

Fig. 1
figure 1

SEM image of MgO-particle-reinforced wood fiber-matrix composite

Full size image
Table 1 Workpiece properties of MgO-particle-reinforced wood fiber-matrix composite
Full size table

In this experiment, the helix milling cutter with helix angle of 70°. The diameter of the tool with six cutting teeth was 140 mm. The blades were fabricated by injection moulding at high temperature (Kyocera Group Co. Ltd. Japan). Table 2 provide the tool parameters and mechanical properties of the cutting tools used in this study.

Table 2 Tool geometry and mechanical properties of PCD
Full size table

2.2 Experimental Set-up

Experimental set-up of machinability studies of MgO-particle-reinforced wood fiber-matrix composite was presented in Fig. 1. Spiral up-milling tests were performed on a MGK01 four-axis CNC machining center in Fig. 2a. As shown in Fig. 2b, dynamic cutting forces were monitored by a dynamometer (Kistler 9257B, Switzerland) with a charge amplifier (5070A, Switzerland) during the milling process. The sampling frequency of the recorder was less than the natural frequency of 3.5 kHz. Cutting force components in three orthogonal directions were measured. (Fig. 2b). The average resultant cutting force Fr was calculated from cutting force components in X, Y, and Z directions using the formula (1). A surface profiler was used to measure the surface roughness Ra, and the ACCTee software was used to calculate the surface roughness in Fig. 2c. The cut-off and sampling length were determined as 0.8 mm and 10 mm in this work. A scanning electronic microscope (SEM) was used to observe the machined surface topography.

$$Fr = \sqrt {\left( {Fx} \right)^{2} + \left( {Fy} \right)^{2} + \left( {Fz} \right)^{2} }$$
(1)
Fig. 2
figure 2

Experimental set-up: a machine center; b amplifier connected to PC; c surface profiler

Full size image

2.3 Calculation of Material Removal Rate and Experimental Design

In spiral up-milling, the cutting width b is periodically varying compared with a fixed cutting width in cylindrical milling. When the tooth of cutter just cut into the workpiece, b is very small, then gradually increases to the maximum, and finally decreases as tooth cut out. Figure 3 shows a schematic of chip section in spiral up-milling.

Fig. 3
figure 3

Schematic of chip volume in spiral cutting

Full size image

For one tooth, the unit width of chip is:

$$db = \frac{{\frac{D}{2}d\varphi }}{\sin \lambda } = \frac{D}{2\sin \lambda }d\varphi$$
(2)

where b is cutting width (mm), φ is the contact angle between the cutter and workpiece (°), λ is the spiral angle (°), D is the diameter of tool (mm).

Integral of db, the width of chip is:

$$b = \int_{0}^{\varphi } {\frac{D}{2\sin \lambda }d\varphi } = \frac{D \cdot \varphi }{{2\sin \lambda }}$$
(3)

For infinitesimal elements, the unit cross–section area (dA) of the chips cut by the length of cutter teeth can be expressed as:

$$dA = a \cdot db = Uz \cdot \sin \varphi \cdot \frac{D \cdot d\varphi }{{2\sin \lambda }} = \frac{D \cdot Uz}{{2\sin \lambda }} \cdot \sin \varphi \cdot d\varphi$$
(4)

where a is chip thickness (mm), Uz is feed per tooth (mm/Z).

The cross-sectional area of chips cut by one tooth is:

$$A = \int_{0}^{\varphi } {\frac{D \cdot Uz}{{2\sin \lambda }} \cdot \sin \varphi \cdot d\varphi } = \frac{D \cdot Uz}{{2\sin \lambda }}\left( {1 - \cos \varphi } \right)$$
(5)

The volume of chip cut per tooth is:

$$V = A \cdot \int_{0}^{\varphi } {\frac{D}{2} \cdot d\varphi } = \frac{{D^{2} \cdot Uz}}{4\sin \lambda } \cdot \varphi \left( {1 - \cos \varphi } \right)$$
(6)

The relation between contact angle, cutting depth and diameter of the tool is as follows:

$$\cos \varphi = 1 - \frac{2h}{D}$$
(7)

where the h is the cutting depth (mm).

Bring formula (7) into formula (6), the material removal rate is:

$$MRR = \frac{V \cdot U}{{Uz}} = \frac{{D^{2} \cdot U}}{4\sin \lambda } \cdot \varphi \left( {1 - \cos \varphi } \right) = \frac{D \cdot U \cdot h}{{2\sin \lambda }} \cdot \arccos \left( {1 - \frac{2h}{D}} \right)$$
(8)

where MRR is the material removal rate (mm3/s), U is the feed rate (mm/s).

In this work, different cutting speed and cutting depth with a constant feed rate were set as cutting parameters. The detailed experimental design for the research into the effect of cutting speed and cutting depth on cutting force and surface roughness is summarized in Table 3. It is clearly to be found that MRR has nothing to do with cutting speed according to formula (8).

Table 3 Parameters and levels of spiral up-milling test
Full size table

3 Results and Discussion

3.1 Analysis of Cutting Force

3.1.1 Effect of Cutting Parameters on Resultant Cutting Force

Figure 4 shows the trend of cutting force components as time increased. The first 0.25 s corresponded to a transient process before reaching the steady state cutting process. Following the approximately 1.5 s was the stable cutting process. After this, the tool was disengaged. An average of resultant cutting force during stable cutting process was calculated after filtered, and it was used for further analysis.

Fig. 4
figure 4

Typical cutting forced components in spiral up-milling

Full size image

Figure 5 shows the influences of cutting speed and cutting depth on resultant cutting force. Resultant cutting force decreased with the increase of cutting speed and increased with the increase of cutting depth. At the same cutting depth, when the cutting speed increased from 35 to 45 m/s, the resultant cutting force decreased about 10 N, 17 N and 45 N, respectively. However, at the same cutting speed, when the cutting depth increased from 0.5 to 1.5 mm, the resultant cutting force correspondingly increased by about 187 N, 152 N and 151 N. It can be easily concluded that the response of cutting depth to cutting force is higher than that of cutting speed in this experiment (at least in the investigated range of parameters). The growth of the resultant cutting force as cutting depth increases is likely related to the increase of MRR. The increase of cutting speed indicates the increase of spindle speed which leads to a lower feed per tooth, resulting in a lower resultant cutting force. In addition, Fig. 5 also shows the standard deviation of the resultant cutting force during each test (see the error bars). The increase of the standard deviation at large cutting depth (h = 1.5 mm) clearly indicates the occurrence of cutting force fluctuation.

Fig. 5
figure 5

Comparison of the resultant cutting force for various machining conditions

Full size image

3.1.2 Effect of Cutting Parameters on MRR

According to the Eq. (8), it is found that the MRR is a function of U, λ and h. But only cutting depth and cutting speed were investigated in this work. Therefore, the MRR is dependent of the cutting depth in this study. The characteristic of the MRR as a function of the cutting depth is shown in Fig. 6. The MRR firstly showed a slight upward trend as the cutting depth increased. When the continues increase in cutting depth, the MRR characteristic can be approximated by a linear function in the range of cutting depth (0.5 mm < h < 1.5 mm). This linear tendency of MRR confirmed the increasing resultant cutting force as the cutting depth increased.

Fig. 6
figure 6

Material removal rate as a function of the cutting depth

Full size image

3.2 Analysis of Surface Quality

3.2.1 Effect of Cutting Parameters on Surface Roughness

The influence of cutting speed and cutting depth on surface roughness can be seen in Fig. 7. It is remarkable that there is a similar tendency of the resultant cutting force as the cutting depth increases. Also, machined surface roughness decreased with the increase of cutting speed at the same cutting depth. The difference “L” in Fig. 7 illustrates that the machining efficiency can be improved by increasing both cutting depth and cutting speed, while a better surface roughness is obtained. It is of practical significance to improve the machining efficiency of materials for machining industry. Furthermore, if energy consumption is considered, the slight loss of surface roughness can be accepted by reducing the cutting speed. Because the surface roughness of V2 was close to V3.

Fig. 7
figure 7

Comparison of surface roughness for various machining conditions

Full size image

According to Figs. 5 and 7, the cutting speed had moderate effect on the resultant cutting force and surface roughness (at least in the investigated range of parameters), while the effect of cutting depth was more significant. And it is likely the different behaviors of materials deformation in cutting area in front of the tool tip under different cutting depths. Therefore, the cutting speed V2 = 40 m/s was selected for further analysis of surface topography. For this cutting speed, the machined surface topography at the cutting depth of 0.5 mm and 1.5 mm were observed to conjecture the behavior of materials deformation.

3.2.2 Binary Image Analysis of Surface Defects

In order to test the defect rate of machined surface, a viable method of particles size determination using binary image analysis was reported [23]. Specifically, this analysis mainly went through three stages in Fig. 8, including sample image loading, image binarization, and image morphological operation. At stage of images binarization, the threshold value was determined by the “graythresh” function to access the best possible threshold value for each image using MATLAB software. At the stage of morphological operation, image erosion was used to filter the binary image. By creating structural elements similar to the shape of defects on the machined surface, the defects are represented by black pixels and others are represented by white pixels.

Fig. 8
figure 8

An example of image analysis. a Original image; b Inverted binary image (threshold value calculated by the “graythresh” function); c Erosion of inverted binary image (disk radius R = 2)

Full size image

Figure 9a, b show the binarized image of machined surface under 0.5 mm cutting depth and 1.5 mm cutting depth, respectively. The defect rate of the machined surface was obtained by calculating the proportion of black pixels in the image. According to the calculation results, the defect rates of 0.5 mm cutting depth and 1.5 mm cutting depth are 1.09% and 11.34%, respectively. It can be clearly seen that the proportion of black pixels in image 9(a) was much more than that of image 9(b). It is also confirmed that the surface damage degree of machined surface under 1.5 mm cutting depth is far greater than that under 0.5 mm cutting depth. In other words, when the tool is at the same speed and time conditions, a depth of cut condition of 1.5 mm can result in far more material breakage in front of the tool tip than at a depth of cut of 0.5 mm. This is because the greater the depth of cut, the more material needs to be removed under the same conditions. Therefore, the degree of material deformation in front of the tool tip goes from mild to severe as the cutting depth increases.

Fig. 9
figure 9

Binary image after morphological operations of erosion (disk radius R = 2). a 1.09% defect rate at 0.5 mm cutting depth; b 11.34% defect rate at 1.5 mm cutting depth

Full size image

3.2.3 Surface Defects

Camera photos were taken account to evaluate the machined surface integrity in Fig. 10a, d. Some visible defects were observed in Fig. 10d, while Fig. 10a was not. It is no doubt that there is a quite different morphology of machined surface. To further explore this difference, a scanning electronic microscope (SEM) was utilized to assess the topography of machined surface (Fig. 10b, c, e, f). Defects, i.e., crack, pile-up and pits were observed on the machined surface at the cutting depth of 0.5 mm in Fig. 10b, c. In comparison to wood fiber composites without the addition of MgO particles, it has been noted that at small depths of cut, the damage to the machined surface is mainly due to debonding of the wood fibers from the matrix [24]. The formation mechanism of machined surface can be explained as follows. When MgO particles are added, a certain degree of cutting force pushes the hard MgO particles partially into the relatively soft wood matrix, causing cracks and pile-up on the machined surface, as well as pushing the MgO particles partially out of the wood matrix, causing the MgO particles to deboned, resulting in pits of the machined surface. This kind of material deformation, however, is implicitly characterized by plastic deformation, such as good re-formability, leading to a lower surface roughness and a higher surface quality.

Fig. 10
figure 10

a Camera photos of machined surface at the cutting speed V2 = 40 m/s and cutting depth h = 0.5 mm; b and c Defects of machined surface under the condition of (a); d Camera photos of machined surface at the cutting speed V2 = 40 m/s and cutting depth h = 1.5 mm; e and f Defects of machined surface under the condition of (d)

Full size image

By contrast, fracture of fibers and extensive flaking of wood fiber layers were observed on the machined surface at the cutting depth of 1.5 mm in Fig. 10e, f. When the cutting depth increased from 0.5 to 1.5 mm, the MRR increased from 297 to 1545 mm3/s. Namely, the volume of material removed in unit time under 1.5 mm cutting depth is about five times of that under 0.5 mm cutting depth. This sharp increase of material removal volume caused a higher resultant cutting force, which enough to tear wood fibers. Besides, wood fibers are not oriented, therefore, flaking of wood fibers was often observed when the wood fibers orientation is similar to the cutting direction. Otherwise, wood fibers fracture was often observed Fig. 10d. At large cutting depth, the addition of MgO has little effect on machined surface damage, as the larger cutting forces cause direct damage to the wood fiber matrix. This also indicates that at large cutting depths, the deformation behavior of the material tends towards brittle fracture, leading to severe surface damage.

4 Conclusion

In this paper, the surface integrity of MgO-particle-reinforced wood fiber-matrix composite during spiral up-milling was investigated. There are some results can be concluded as follows.

Surface integrity is significantly to be affected by cutting depth than by cutting speed. This is mainly because the MMR increases dramatically when the depth of cut is varied.

At different cutting depths, the deformation of the material in the cutting zone in front of tool tip is closely related to the surface integrity of the machined surface. At small cutting depths, the deformation is characterized by plastic deformation, which results in minor surface defects and better surface quality due to better re-formability. At large depths of cut, on the other hand, the deformation of the material tends to fracture, leading to severe surface damage and poorer machining quality.