1 Introduction

To improve the performance of an engineering component, it is required to improve surface properties without sacrificing the bulk material properties. Surface engineering deals with the improvement in surface properties without changing bulk material properties. Modified surface acts as a protective coating over the bulk of component [1]. As per ASME definition, “treatment of the surface and near-surface regions of a material, to allow the surface to perform functions that are distinct from those functions demanded from the bulk of the material, is termed as surface engineering [2]”. Due to wear, component needs frequent repair and replacement which costs money and causes downtime of equipment [1]. Hence, wear of a component can be considered as a critical factor influencing the product service life. It has been studied that the wear can be classified based on its material failure mechanism like abrasive wear, adhesive wear, corrosion and surface fatigue [3].

Various researchers have discussed on hardfacing in terms of method of deposition, weld consumables, base metal and summarized their results [4,5,6]. Coronado et al. [7] performed hardfacing with flux-cored arc welding (FCAW) and shielded metal arc welding (SMAW) and found that hardfacing with three layers was more wear resistant. The prevention of the material from wear and corrosion can be achieved by adding some strong binding agents which implies that the properties of components will be enhanced [4]. As per an investigation [8] of abrasive wear behaviour of different hardfacing electrodes deposited on grey cast iron used for coal crushing unit’s top bearing plate, it was reported that the wear rate can be reduced by 50% to that of commercial hardfacing electrode when niobium and molybdenum are added as a carbide forming element and as a matrix hardening agent, respectively. Chakraborty et al. [9] investigated the properties of three nickel base Ni–Cr–B–Si hardfacing alloys deposited on austenitic stainless steel. It was concluded that in the case of NiCr-C deposit, large primary borides and carbides resist deformation and result in high wear resistance by increasing the frictional forces. Results showed that beyond a certain hardness, any increase in hardness does not increase the wear resistance of the deposition, i.e. NiCr–B will be a better choice rather than NiCr–C in case, where wear resistance and friction coefficients are the only considerations. Martinez et al. [10] tested flat tip and hemispherical tip pin under different loading condition. The results showed that under high speed and light load (10 N), the wear of the flat pin is more than that of the hemispherical one, while it is opposite for heavy load (40 N). John et al. [11] performed a simulation of stress, which leads to the wear in pin-on-disc apparatus; when the pin starts wearing out, the maximum value of equivalent stresses is increased. On the other hand, when pressure increases, the corresponding maximum deformations also increases. Basavarajappa et al. [12] studied the wear behaviour of glass epoxy polymer composites using SiC and graphite particles as a reinforcement material. The weight was investigated using Taguchi’s technique. The results showed that the reinforcement particles improve the wear resistance of the material. Nickel-based coatings can be utilized in applications where wear resistance combined with oxidation or hot corrosion resistance is desirable [13]. They are mainly employed in chemical industry, hot working punches, glass mould industry, fan blades and mud purging elements in cement factories [14]. After their use in nuclear industry, Co-based hardfacing alloys are sources of highly radioactive 60Co isotope. Hence, carbide-containing Ni-based alloy with similar properties is developed. Ni-based alloy with Cr and Mo has been selected as hardfacing alloys due to high hardness and wear resistance.

Finite element analysis has been used in pin wear tester and in the study of wear to model the phenomena at widely different length scales. Khot and Borah [15] suggested that finite element method (FEM)-based numerical modelling of machine operation with appropriate wear models will enable estimation of wear. A priori accurate modelling of the state of stress and strain is essential, since the accuracy of calculated wear values is dependent on the state of stress and strain in the components and the process parameters like applied load, sliding speed, sliding distance and the reinforcement. Applied load has the highest influence on the wear of the composites. Variables like applied load and sliding distance have more noticeable effects rather than the sliding speed. Benabdallah and Olender [16] made use of finite element (FE) software package ANSYS to perform the simulation of wear profile of pin made of polyoxymethylene (POM), in sliding contact with rotating steel disc. The FE simulation predicted the pressure distribution at the contact zone which was used to map with experimental wear profile, and the results were in good agreement. Archard’s wear model is a simple phenomenological model which assumes a linear relationship between volume of material removal V, for given sliding distance S, an applied normal load FN and hardness (normal load over projected area) of softer material H. A proportionality constant in Archard’s law, the wear coefficient (K), characterizes the wear resistance of the material, and hence, it becomes an important parameter for predicting wear and the effective life span of any general tribosystem [17].

The number of runs required for a full factorial design increases geometrically. Fractional factorial design is efficient and reduces the time significantly. However, the fractional design may not contain the best design point. The Taguchi’s design can be further simplified by expending the application of the traditional experimental designs to the use of orthogonal array. Taguchi method is a statistical method to improve the quality of manufactured goods, and more recently, it is also applied to engineering, biotechnology, marketing and advertising [18]. Taguchi is a technique for design of experiments (DOE), which is based on well-defined guidelines. This method uses set of arrays called orthogonal arrays, and these arrays are ordered in such a way that minimum number of experiments can give the full information about the experiment’s factors which will affect the response parameters. On the basis of specific number of variables and levels, there are many standard arrays available like L4, L9, L16, L27, etc. [19]. Sahoo [20] employed a Taguchi orthogonal array to optimize four coating process parameters (viz. bath temperature, concentration of nickel source solution, concentration of reducing agent and annealing temperature) with respect to wear behaviour of electroless Ni–P coatings sliding against steel and observed that annealing temperature and bath temperature have the most significant influence in controlling wear characteristics.

It has been observed that the researchers have strived to understand the effect of various parameters on wear characteristics in order to improve materials’ surfaces with more wear resistance, through simulation and experimental studies.

Thus, this paper focuses to carry out experimental investigation to analyse the effect of % titanium (Ti) reinforcement weight on wear resistance along with supplementary simulation analysis to promote the understanding of effects of factors.

2 Experimental Method

2.1 Work Material

The technical data and composition of electrode used are shown in Tables 1 and 2.

Table 1 Technical data of electrode
Full size table
Table 2 Composition of electrode
Full size table

2.2 Preparation of Specimen

In the present investigation, mild steel was selected as base material. Hardfacing of Ni–Cr–Mo electrode was performed by manual metal arc welding. 100 A current was selected as per electrode’s specification. Welding was performed with direct current reverse polarity (electrode as positive and workpiece as negative).

Measurement of mass of specimen was performed before and after hardfacing in order to calculate the mass difference. Further, titanium in varying amount (2, 3 and 4 wt% weight of hardfacing) was added as a reinforcement material. After that, samples of 1 cm × 1 cm were cut from the hardfaced samples and the samples were brazed on pin to carry out wear testing. The brazed pins are shown in Fig. 1.

Fig. 1
figure 1

Pins made form hardfaced material

Full size image

2.3 Description of Experimental Set-up

An ASTM G99-95 standard [21] was used for dry sliding test. The tests were conducted on pin-on-disc apparatus. Ducom wear friction monitor was used for wear testing as shown in Fig. 2. The material of disc was EN-31 hardened to 60 HRC. The diameter and thickness were 165 mm and 8 mm, respectively. The pin of 10 mm × 40 mm was used to conduct the wear test. The machine has an LVDT sensor having 10 mm ball diameter. Before performing the test, pin and disc were cleaned in acetone to remove any dirt.

Fig. 2
figure 2

Pin-on-disc apparatus used in the present research

Full size image

All the samples were polished with 600-grade abrasive paper to get flat surface for accurate readings, and tests were conducted at room temperature (25 °C) under dry condition. Loading on pin was done by placing counter weight, which formed frictional contact between pin and disc. The disc was rotated at defined speed (rpm) value, while the pin was constrained in all direction resulting in a wear track, as shown in Fig. 3. The wear readings were taken in terms of microns. The measurement of mass of pin samples was performed after each experiment with the help of electronic weighing machine having least count of 0.001 gms.

Fig. 3
figure 3

Wear track on pin-on-disc machine

Full size image

2.4 Experimental Conditions

Experimental runs were designed according to Taguchi’s orthogonal array L9. The selected process parameters for the study were three, namely reinforcement weight %, applied load and rotational speed (rpm), with three levels for each parameter. The selected process parameters are shown in Table 3.

Table 3 Process parameters selected in the present investigation
Full size table

All testing parameter and test condition were kept constant to analyse effect of reinforcement weight (%), applied load (kg) and rotational speed (rpm). Table 4 shows the arrangement of process parameters according to Taguchi’s L9 orthogonal array. Measurement of initial and final mass was done to calculate loss of volume in each sample.

Table 4 Process parameter in L9 orthogonal array
Full size table

3 Finite Element Analysis

3.1 Governing Equation

Archard’s wear law was used for wear calculation. According to Archard’s law, wear (W) is the function of applied load (F), sliding distance (L) and hardness of the material (H). Wear coefficient (K) is dependent on experimental conditions and type of material, and hence, it is determined by the experimental procedure. Equation 1 shows Archard’s law. Equations 2 and 4 are the modified form of the Archard equation [22] where sliding distance (L) is replaced by sliding velocity (V) and experimental time (t); further, velocity can be expressed in terms of wear track diameter (D) and rotational speed (N). The simulations were carried out by finite element software ANSYS.

$$ W = \frac{K}{H} \times F \times L $$
(1)
$$ W = \frac{K}{H} \times F \times V \times t $$
(2)
$$ W = \frac{K}{H} \times F \times \frac{ \pi \times D \times N }{60} \times t $$
(3)

3.2 3D Model of Pin-on-Disc Tribometer

The geometry of pin-on-disc tribometer was created in design modeller, considering the inner and outer diameter of discs to be 80 mm 110 mm, respectively. And, the diameter and height of pin were taken as 10 mm and 30 mm, respectively. The pin was considered as a flexible body and the disc as a rigid body. The material of disc and pin was assigned as EN-31 hardened steel and Ni–Cr–Mo hardfacing alloy, respectively. Material properties such as density, Young’s modulus, Poison’s ratio and tensile strength for both the materials were updated accordingly. Figure 4 shows the 3D model of pin-on-disc tribometer.

Fig. 4
figure 4

3D model of tribometer

Full size image

3.3 Contact Region

Under the connection group object, contact regions appear as child objects which control various parameters of a contact pair. Frictional contact was chosen as the type of contact with asymmetric behaviour. The value of friction coefficient was calculated experimentally using pin-on-disc apparatus. Normal stiffness was selected as manual with stiffness factor to be 0.1. The details of frictional contact are shown in Fig. 5. Augmented Lagrange algorithm was chosen as it removes the geometrical penetration over the iterations. Nodal normal to target detection method was used because it was required during the implementation of Archard wear model.

Fig. 5
figure 5

Details of contact region between pin and disc

Full size image

Pin was considered as contact body with bottom surface of the pin as contact surface, while disc was considered as target body having the top surface of the disc as target surface. Figures 6 and 7 show target and contact surfaces.

Fig. 6
figure 6

Target body view

Full size image
Fig. 7
figure 7

Contact body view

Full size image

3.4 Meshing

FEA model was optimized by altering number of nodes and elements with consideration of solution time by attempting different options: edge sizing, body sizing, multizone and face meshing. The number of nodes and elements for selected meshing is shown in Table 5. Four-nodded quadrilateral elements (Conta174, Target170) were used in contact-pair meshing. Figure 8 represents overall meshing of the 3D model of pin-on-disc tribosystem.

Table 5 Number of nodes and elements in tribosystem
Full size table
Fig. 8
figure 8

Overall meshing of pin-on-disc 3D model

Full size image

3.5 Applied Load and Boundary Conditions

Load of 9.81 N (1 kg) was applied on top face of the pin. Cylindrical support was given to the pin because of which the motion of the pin was constrained in all direction except in axial direction, and the disc was rotating about its centre. The rotation about z direction was applied to the joint which was specified under the connection tab. Figure 9 shows the boundary conditions and load applied in the present simulation. Initial time step of 1 × 10−4 s was applied with zero rotation of disc. After the initial step, rotation of the disc was in small increments. Due to extended solution run time and limited resources, the simulation time was taken as 1 s. Figure 10 shows the time step size used for the simulation.

Fig. 9
figure 9

Applied load and boundary conditions

Full size image
Fig. 10
figure 10

Step size used in the present simulation

Full size image

3.6 Command Prompt

Contact surface wear can be simulated by defining a wear model (TB, WEAR); however, one can also use a user-defined wear model (USERWEAR subroutine). Archard wear model defines the rate of wear as a function of contact pressure sliding velocity and material hardness. The wear model was activated by TB command as shown in Fig. 11.

Fig. 11
figure 11

Commands used in simulation

Full size image

The values of wear coefficient (K) were taken from experimental runs. Calculated mean values corresponding to a particular type of material were calculated, i.e. for 2, 3, 4 wt% reinforcement weight and unreinforced material were used for the analysis. The values of exponent of contact pressure (m) and exponent of sliding velocity (n) were taken as 0.40 and 1.3, respectively.

3.7 Post-Processing

Post-processing section of the present analysis consisted of contact pressure (MPa), contact depth (mm), equivalent stress (MPa) and volume loss due to wear (mm3). NLHIST command was used for the activation of result tracker menu. Simulation procedure was repeated for all nine experiments mentioned in L9 orthogonal array. The minimum and maximum wear (microns) during the experiments were recorded. Further, two more simulations were carried out with unreinforced Ni–Cr–Mo alloy as a pin material.

4 Results and Discussion

The difference between the experimental and simulation data has been recorded and is represented in Table 6.

Table 6 Comparison of experimental and simulation data
Full size table

The experimental data of 2 wt% titanium-reinforced samples under the different loading conditions are compared in Fig. 12. A drastic increase in the volume loss can be observed with the increase in applied load and rotational speed. The wear values under the 1, 2 and 3 kg applied loads are 56.56, 490.22 and 724.14 microns, respectively. The average coefficient of friction evaluated at 1, 2 and 3 kg are 0.354, 0.371 and 0.276, respectively.

Fig. 12
figure 12

Comparison of wear of 2 wt% Ti under specified parameters

Full size image

The experimental data of 3 wt% titanium-reinforced samples under different loading condition are compared in Fig. 13. It is observed that initially volume loss increases rapidly with the increase in applied load. The wear values under 1, 2 and 3 kg of applied load are 52.69, 182.18 and 219.90 microns, respectively. The average coefficient of friction evaluated at 1, 2 and 3 kg are 0.3037, 0.354 and 0.226, respectively.

Fig. 13
figure 13

Comparison of wear of 3 wt% Ti under specified parameter

Full size image

The experimental data of 4 wt% titanium reinforcement are compared in Fig. 14. It can be observed that volume loss at 1 kg load is more than that at 2 kg load. However, there is an increase in rotational speed for 1 kg load. So, in the case of 4 wt% titanium reinforcement, applied load and rotational speed have major effects on volume loss of substrate. Therefore, a drastic increase in volume loss is observed at 3 kg load. The wear values for 1, 2 and 3 kg applied load are 139.43, 78.37 and 777.37 microns, respectively. The average coefficient of friction evaluated at 1, 2 and 3 kg are 0.220, 0.2533 and 0.331, respectively. Table 7 represents S/N (signal-to-noise) ratio and volume loss due to wear for 9 designed runs with % reinforcement variation.

Fig. 14
figure 14

Comparison of wear of 4 wt% Ti under specified parameters

Full size image
Table 7 S/N ratio and volume loss due to wear
Full size table

The experimental data of unreinforced Ni–Cr–Mo hardfacing samples are compared in Fig. 15. A drastic increase in wear with the increase in applied load for the same rotational speed is observed. The wear values under 1 and 3 kg applied load are 111.65 and 556.94 microns, respectively. The average coefficient of friction evaluated at 1 and 3 kg are 0.434 and 0.351, respectively.

Fig. 15
figure 15

Comparison of wear of unreinforced hardfacing under specified parameters

Full size image

4.1 Effect of Reinforcement % on Wear

Figure 16a shows the effect of reinforcement weight % on volume loss occurred. Table 8 shows that delta value is the minimum for reinforcement weight % among the three selected parameters; therefore, it has been assigned rank 3. It implies that the reinforcement weight % has lesser effect on volume loss due to wear, compared to the other two chosen parameters.

Fig. 16
figure 16

Effect of control factor on volume loss

Full size image
Table 8 Response table for S/N ratio for smaller is better
Full size table

4.2 Effect of Applied Load on Wear

Figure 16b shows the effect of applied load on wear. Table 8 shows that the delta value is the maximum for applied load among three selected parameters; hence, it has been assigned rank 1. It signifies that applied load has highest influence, among the selected parameters on volume loss due to wear.

4.3 Effect of Rotational Speed on Wear

Figure 16c shows the effect of rotational speed on wear. Table 8 shows that the delta value for rotational speed lies between applied load and % reinforcement weight, due to which it has been assigned 2nd rank out of 3 parameters. It means that the rotational speed has moderate effect on volume loss due to wear compared to other two chosen parameters.

MINITAB 18 software has been used for the investigation of various parameters such as response for signal-to-noise ratio, delta values and rank of each parameter which are presented in Table 8. It can be seen that the most effective parameter is applied load having maximum delta value 20.546 followed by the rotational speed and reinforcement weight % as 10.155 and 6.802, respectively.

4.4 ANOVA for S/N Ratio of Volume Loss

ANOVA (analysis of variance) has also been implemented to understand the impact of selected parameters on volume loss due to wear. Table 9 shows ANOVA for S/N ratio of volume loss with degrees of freedom (DF), adjusted mean squares (adj MS), adjusted sums of squares (adj SS), F value (F), and p value (p). Table 10 shows model summary with R-sq (R2), R-sq (adj) (adjusted R2), R-sq (pred) (predicted R2) and S, where S represents how far the data values fall from the fitted values.

Table 9 Analysis of variance for S/N ratio
Full size table
Table 10 Model summary
Full size table

Investigation of calculated F values and p values for all control factors signify that applied load has high impact, whereas % reinforcement weight has low impact and rotational speed has moderate impact on volume loss due to wear. Also, it is noteworthy to observe in Table 9 that p value for applied load is nearer to 0.05 which also indicates that applied load has high influence on volume loss due to wear.

4.5 Regression Equation

An equation has been generated from available experimental data using MINITAB 18 software. Equation 4 represents volume loss of wear as a function of three parameters—% reinforcement weight, applied load and rotational speed with calculated coefficient.

Table 11 shows the comparison of experimental and theoretical volume loss, calculated by regression equation. Experimental results show minimum wear of 52.69 microns and volume loss of 1.802 mm3 for 3% titanium reinforcement weight—1 kg applied load—400 rpm rotational speed. For the said case, the value of calculated volume loss using the following regression equation is 1.27 mm3.

$$ {\text{VL}} = - \left( {19.5} \right) - \left( {2.70} \right) {\text{RW}} + \left( {14.87} \right) {\text{AL}} + \left( {3.50} \right) {\text{RS}} $$
(4)

where VL volume loss (mm3), RW reinforcement weight (%), AL applied load (kg) and RS rotational speed (rpm).

Table 11 Comparison of experimental and theoretical volume loss
Full size table

5 Conclusions

In the present investigation, Taguchi L9 orthogonal array method was used for deciding the experimental runs and subsequent analysis of collected data. The applied load, rotational speed and composition (variation in % titanium reinforcement weight) were considered as process parameters to study the wear behaviour. Finite element analysis was performed on a 3D model of pin-on-disc apparatus. The results from the experimental work and the simulation work were compared. The observations are as follows:

  • In the case of 2 wt% Ti reinforcement, a drastic increase in volume loss is observed with the increase in the applied load and rotational speed.

  • In the case of 3 wt% Ti at 3 kg load, volume loss decreases after 850 s. This may be due to work hardening as the sample is subjected to repeated loading for long time.

  • In the case of 4 wt% Ti reinforcement, applied load and rotational speed have major effects on the volume loss of substrate.

  • In the case of unreinforced Ni–Cr–Mo hardfacing, drastic increase in wear with the increase in applied load for same rotational speed is observed.

  • ANOVA for S/N ratio shows that applied load has maximum F value and minimum p value which means that applied load has a significant effect on the volume loss due to wear.

  • Regression equation calculation results show minimum wear for the fourth experiment.