Artificial Neural Network on Tool Condition Monitoring in Hard Turning of AISI4140 Steel Using Carbide Tool

Ajitha Priyadarsini, S.; Rajeev, D.; S. Rai, Rajakumar; Nadar, Kannan Pauliah; Christopher Ezhil Singh, S.

doi:https://doi.org/10.1155/2023/2139906

Mathematical Problems in Engineering

On this page

Abstract Introduction Materials and Methods Results and Discussion Conclusions Data Availability Conflicts of Interest References Copyright Related Articles

Research Article | Open Access

Volume 2023 | Article ID 2139906 | https://doi.org/10.1155/2023/2139906

Artificial Neural Network on Tool Condition Monitoring in Hard Turning of AISI4140 Steel Using Carbide Tool

S. Ajitha Priyadarsini,¹D. Rajeev,²Rajakumar S. Rai,³Kannan Pauliah Nadar,⁴and S. Christopher Ezhil Singh⁵

Academic Editor: Afaq Ahmad

Received22 Jan 2023

Revised22 Apr 2023

Accepted19 May 2023

Published30 May 2023

Abstract

Hard turning has replaced conventional grinding in production processes in recent years as an emerging technique. Nowadays, coated carbide tools are replacing expensive CBN inserts in turning. Wear is a significant concern when turning with coated carbide; it immediately affects the acceptability of the machined surface, which causes machine downtime and loss due to wastage in machined parts. Online tool condition monitoring (TCM) is required to prevent such critical conditions. Hard turning differs from conventional turning in energy balance during metal cutting, resulting in greater thrust force; hence, the TCM model presented for conventional turning may not be suitable for hard turning. Hence, tool wear prediction for turning is projected based on thrust force using an artificial neural network (ANN). All of the tests were done using a design of experiments called full factorial design (FFD). The specimens were made of AISI 4140 steel that had been hardened to 47 HRC, and the inserts were made of coated carbide. The most impactful input features for wear, selected based on experimental outputs, were given to the neural network and trained. Tool wear is an estimated output from the training set that has been validated with satisfactory results for random conditions. The 5–10–1 network structure with the Levenberg–Marquardt (LM) learning algorithm, R² values of 0.996602 and 0.969437 for the training and testing data, and mean square error values of 0.000133152 and 0.004443 for the training and testing data, respectively, gave the best results. The MEP values of 0.575407 and 2.977617 are very low (5%). The LM learning algorithm-based ANN is good at predicting tool wear based on how well it predicts tool wear for both the testing set and the training set.

1. Introduction

Due to its speed and affordability, hard turning has recently become a new approach in the machining industry. Cylindrical grinding has been substituted with hard turning. Hard steels with values greater than 45 HRC are subject to the turning process [1, 2]. Better material removal rates, shorter work cycles, the absence of hazardous cutting fluids, and the ability to do both hard and soft turning on a similar machine are all benefits of this method [3, 4]. Nowadays, coated carbide tools that are not cost-effective are replacing CBN tools that were typically used for hard turning [5, 6]. The rate of tool wear is a concern for carbide tools. During the machining process, wear on the cutting tool is an inevitable occurrence. It happens when the cutting tool and the piece are being made rub against each other. A worn-out tool contributes to poor component quality [7], machine breakdown due to cutting tool failure that occurs suddenly, and an increase in production time and expense. 20% of the downtime is due to tool failure. The budget for replacing cutting tools’ production costs ranges from 3 to 12% [8, 9]. To avoid a tool failing suddenly due to wear, tool wear must be predicted and its condition monitored from the start of the process.

Dimla and Dimla, [10]; Siddhpura and Paurobally [11]; and Cheng et al. [12] have reviewed research about tool condition monitoring. The steps involved in TCM are feature extraction, feature selection, and model development for wear. Tool wear prediction relies heavily on cutting force [13]. Because there was more friction, the tool started to wear down. This also changed the static and dynamic parts of the three orthogonal cutting forces. Dynamometers were used to figure out how hard the tool was cutting. Zhang et al. [14] have suggested a TCM system that is based on a smart toolholder that can sense cutting force, torque, and vibration in all three directions. The data from the sensor signals was used to make an exact and reliable model based on certain features that were sensitive to wear. This model was then used to predict how long the tool would last. It is important to make a tool wear model because the characteristics picked need to be well related to wear [15]: The features of hard turning are very different from those of regular turning because of the way the cutting works. As a result, the characteristics utilized for the tool wear model in conventional turning cannot be used for hard turning, making it imperative to identify the most sensitive feature from the results of research [16].

Ozel and Karpat [17] have concluded that an analytic model describing the complex tool wear mechanism is not suitable for online tool condition monitoring. For online monitoring, artificial neural networks (ANN) are widely implemented. The ANN was utilized by Scheffer et al. [18] and Wang et al. [19] to estimate flank wear during hard turning with CBN tools. The ANN models the nonlinear dependencies among tool wear and sensor signals [13, 20]. ANN is based on the mapping between input values and output values. Eser et al. [21] created a universal dynamic surface roughness monitoring system for milling operations utilizing an ANN and RSM. The output from both models is satisfactory. In order to estimate the surface roughness when hard turning AISI D2 steel with ceramic cutting, Kara et al. [22] created an ANN model. For training the ANN, the backpropagation algorithm was utilized. The outcomes demonstrated that the ANN’s learning capacity was reasonably effective in estimating surface roughness.

The current scenario requires a well-designed model for estimating tool wear using less expensive carbide tools instead of costly ones like CBN. However, investigations have been conducted on the tool condition monitoring of carbide tools; most of them are limited to conventional turning. Features selected for model development in conventional turning may not be suited for hard turning. More predominantly, the features of hard turning are implicitly influencing the wear rate higher than compared to conventional turning. Thus, the establishment of a reliable wear-on tool model for turning using coated carbide tools, taking into account features significant to hard turning, was needed. This paved the way to pick the features of hard turning for further formulating the TCM model using ANN.

2. Materials and Methods

2.1. Work Piece, Cutting Tool, and Equipment

The AISI4140 alloy steel was heat-treated, and the hardness was measured as 47 ± 1 Rockwell hardness (HRC). The chemical composition is given in Table 1. The investigation’s cylindrical ingot had a 250 mm length and an 80 mm diameter. The durotomic carbide tool chosen as TH1500 grade is a titanium (C, N) + alumina cutting tool with a CVD coating. The nose had a 0.8 mm radius. It was stated that the cutting insert was ISO (CNMG120408). The cutting conditions used in the current study were as follows: back rack angle = −60, negative cutting edge inclination angle = −60, and major cutting-edge angle = 950. The tool holder utilized was a PCLNR 2525 M12 type. The studies were carried out in dry conditions using an industrial Kirloskar lathe with a 2.2 KW spindle power. Figure 1 shows the experimental configuration.

2.2. Experimental Designs

To conduct experiments systematically, the design of experiments (DOE) procedure is used. Full factorial design (FFD) is the foundation upon which all cutting conditions are framed. In FFD, tests are carried out at all level of each factor, while one factor is changed at a time. If “n” is the number of levels for the i^th component and “K” is the total number of factors, the required number of experiments to be run for FFD is n_k. The machining length “L” and the process parameters (V, f, and d) are selected as factors (k = 4). Table 2 lists the three tiers of criteria taken into account. According to FFD, a total of 81 (34) experimental conditions are created. The dynamometer’s answers are noted for each circumstance. Considered are three different machining lengths: 200 mm, 400 mm, and 600 mm. For ANN testing and training, this experimentally obtained data were utilized.

2.3. Tool Wear Measurement

The cutting tools are subjected to high forces, elevated temperatures, and sliding. All these conditions induce wear. Flank wear is the most prominent wear in hard turning. It is represented by Vb. The average wear height on the tool flank is used to compute the volumetric loss at the top of the tool edge, which is primarily due to abrasion. According to ISO 3685, the accepted wear limit is 0.3 mm [23]. After machining lengths of 200 mm, 400 mm, and 600 mm, respectively, the wear values for various cutting circumstances are measured. The flank wear values are measured using a toolmaker’s microscope. The suggested soft computing system training and testing uses the tool wear data that was gathered during various cutting situations.

2.4. Cutting Force Measurement

During experimentation, the force components acting on the cutting tool are detected by a (Kistler type 9257B) piezoelectric force dynamometer. The dynamometer mounting is demonstrated in Figure 2. The two thrust components are feed force (F_a) along the axial direction (x-direction) and thrust force (F_r) along the radial direction (y-direction). The tangential component is along the z-direction and is represented as a tangential force (F_t). By using a known weight, the dynamometer is calibrated. Signal acquisition was done using Dynaware software. An ANN was utilized to estimate tool wear using the mean force value that was collected during the experimentation. The mean value for the force signal denoted by f(t) over a time period is calculated using the relation given in the following equation:where T is the time period.

2.5. Tools Wear Modelling

2.5.1. Selection of Input Parameters for Model

The factors affecting tool wear during difficult turning are the input, and tool wear is the result of this process. Tool wear is investigated using analysis of variance (ANOVA) in relation to machining length and process variables. Suresh et al. [5] reported on a similar strategy. The variation of various cutting forces for tool wear as researched by Dimla and Lister [13] serves as the basis for the feature selection for the cutting force.

2.5.2. Model Development-ANN

Neural networks possess the ability to study the matching between input and output variables. Hence, it was regarded as one of the most highly sought-after modeling tools. ANN could be distinguished as a logical entity in which different nodal elements link to each other, thereby creating a stimulated network. Biological neurons are emulated by taking the input data. This is done by individual nodes in a particular network. The structure of an artificial neuron cell is demonstrated in Figure 3. This process is carried out by employing simple operations on the data. A feed-forward neural network topology’s fundamental building blocks are three layers, which are input, hidden, and output layers. The following equation’s relation is used to calculate the weighted sum of a neuron’s input.where x represents the input matrix, represents the weight matrix, and “b” represents the bias term.

The activation function of a neuron controls its output. The activation function receives the weighted total of the inputs, and the activation function is a logistic sigmoid function. The following equation specifies the output value from the logistic sigmoid function of the ANN model.

Based on the relationship shown in the following equation, input and output values are typically standardized between (0 and 1) before the training and testing procedure.

According to Erkan et al. [24]; where Nvi is the normalization value, V_max is the highest value of the input/output data, V_min is the minimum value of the input/output data, and N_i is the ith value.

In order to match the output values closely to the experimental values, the best algorithm and the ideal network are required. By continuously increasing the number of neurons in the hidden layer step by step (i.e., from three to thirteen), the ideal ANN design is discovered by trial and error. After building the ANN architecture, the choice of the appropriate learning algorithm is the factor that has the greatest impact on the effectiveness of ANN in practice [24]. Scaled conjugate gradient (SCG), Levenberg–Marquardt (LM), quasi-Newton backpropagation (BFGS), resilient backpropagation (RP), and conjugate gradient backpropagation (CGP) learning algorithms were employed as the training methods in this work. The algorithm uses the following procedure: The input features are continuously updated during each iteration (epoch).

The minimal value of mean square error (MSE) given in the convergence criteria or based on a maximum number of iterations determines when training is terminated. Figure 4 depicts the training process flowchart. Equation (3) provides the following formula for calculating the MSE:where , , , , and

Mean error percentage (MEP) and coefficient of correlation (R²) values were also used for comparison. MEP determines the mean error ratio among the predicted and experimental values. MEP is given by the relation.coefficient of correlation (R²) range varies from −1 to +1. R² close to +1 indicates positive linear correlation and towards −1 negative correlation. R² is calculated based on the relation

Based on the results of the ANOVA and cutting force analyses, the following features are selected in this study as inputs for model creation using artificial neural networks (ANN).(1)Cutting speed ()(2)Feed (f)(3)Depth of cut (a_p)(4)The mean value of forces in radial direction-direction (F_r)(5)Machining length (L)

As shown in Figure 5, a feed-forward multilayer neural network is employed to predict tool wear. As inputs, the process parameters , the average force in the radial direction (“F_r”), and L were given. Tool wear was calculated as the output. The momentum constant is assumed to be 0.5, and the learning rate is set at 0.01. The MSE target or the maximum number of iterations (epochs) is the basis for convergence. The maximum number of iterations allowed in this study is 1000, and the MSE target is set at 1 × 10⁻⁶. For ANN efficiency, determining the percentages of training and testing data is crucial. For the training and testing of the ANN in this study, 81 experimental data sets for tool wear prediction were prepared. The ratio chosen was 85%:15% for the training and test sets. 12 data points for the testing set and 69 data points for the training set were randomly chosen for this framework.

3. Results and Discussion

3.1. Tool Wear Analysis

The flank wear measurements lie between 0.06 and 0.493 mm. Figures 6(a) and 6(b) show photographs of the flank wear. With these cutting parameters ( = 170 m/min, f = 0.12 mm/rev, a_p = 0.6 mm, and L = 600 mm), the flank wear achieves a maximum value of 0.493 mm, indicating that the machining length, cutting speed, and depth of cut all affect tool wear. Abrasion is the primary cause of flank wear in the majority of cutting situations, and the value of flank wear is within the allowed range of 0.3 mm. This is comparable to the findings from Sahoo and Sahoo [25]. The rationale is that the coated carbide’s TiN coating reduces friction and serves as a diffusion barrier layer, which promotes material chemical stability. Coated carbide tools with Al₂O₃ help reduce the temperature at the tool interface. As a result, the coating on a carbide tool aids in maintaining hardness at higher cutting parameter values. The observation demonstrates the coated carbide insert’s capacity for hard turning.

(a)

(b)

3.2. Influence of Process Parameters

The total of squares, mean squares, degrees of freedom, probability, and F values are used in the ANOVA table to display the data for each component and interaction. If the value is 0.05, the response model is statistically significant. A factor is significant if it affects the result and its value is less than 0.05. Table 3 makes it evident that, following machining length, cutting speed has a significant role in determining tool wear. Both the cut and feed depths are significant. Suresh et al. [5] and Aouici et al. [2] both published results for the process parameter that were identical. From the ANOVA table, PCR shows that the machining length has more influence compared to other parameters.

3.3. Variation of Cutting Force with Wear

The variation in cutting force for the cutting condition ( = 70 m/min, f = 0.12 mm/rev, and d = 0.45 mm) is shown in Figures 7(a)–7(c). All three of the force’s F_a, F_r, and F_t fluctuations with wear are the subject of investigations. All wear patterns for the various turning force components point upward. Because the area of contact increases with wear, this pattern results. The thrust force is bigger than the tangential force and the feed force in severe tuning, so it should be kept in mind. The larger value of thrust force is caused by the harder turning’s shallower depth of cut and feed. The depth of cut value is smaller than the nose radius of the cutting tool.

(a)

(b)

(c)

So, the rounded tooltip is in contact during machining, which results in a severe ploughing effect. Another reason can be the spring-back effect of hardened material in hard turning, as investigated by Astakhov [26].

Spring back is represented by the following equation:

E value remains the same for both soft and hard materials. But the ultimate strength of hardened material is much higher compared to soft material, thus resulting in the high value of spring back for hard materials. This is the reason for higher contributions of power to the tool-workpiece interface during hard turning. This is not true with conventional turning.

The static force measured is used by researchers [13, 18] as an input feature for tool wear monitoring as it contains a lot of information regarding wear and is directly correlated with wear. It is evident from the variation of cutting force with wear in Figures 7(a)–7(c).

3.4. Prediction of Tool Wear with ANN

MATLAB software was used to create a computer algorithm for forecasting tool wear in hard turning utilizing carbide inserts in this investigation. The network’s input parameters are cutting speed, depth of cut, feed rate, mean force in radial direction, and machining length, and its output parameter is tool wear. Tables 4 and 5 show the hidden neurons and statistical characteristics of ANN models for five different learning techniques. As shown in Table 4, the prediction results for both the training and testing sets of the tool demonstrated that all of the algorithms produce very good predictions. All learning methods’ R² values for both training and testing sets are substantially closer to +1, indicating good tolerance. Furthermore, MSE and MEP were relatively low. The mean relative error for tool wear was found to be less than 5% over the training period, demonstrating that MEP values for the training and testing sets were below acceptable error limits (5%). In all learning methods, the 5–10–1 network structure produced the best results, indicating that it is the best network structure for forecasting tool wear. Table 5 shows the best performance of all algorithms. The LM learning algorithm produced the best results for predicting tool wear. Figures 8(a) and 8(b) show the regression curve between ANN-predicted values and experimental values for the training and testing data for the selected network. For tool wear prediction, the correlation coefficients (R² values) of the training and testing sets were 0.996602 and 0.969437, respectively. The MSE values for the training and testing data were 0.000133152 and 0.004443, respectively. The matching MEP values were also 0.575407 and 2.977617.

(a)

(b)

Figures 9 and 10 show comparisons of ANN predictions and experimental findings for training and testing sets of output parameters. The most remarkable feature here is that the prediction values for both the training and testing sets are quite close to the experimental values. The network’s forecasting performance for tool wear was notably satisfactory, as shown in Figures 9 and 10. This suggests that using five input characteristics as influencing factors for tool wear forecasts yields satisfactory results. Tool wear is calculated using the relationship described in equations (9) and (10) [24].where can be found bywhere E_i is the weighted total of the inputs and is derived using the equation in Table 6. The weight between the levels completes the data flow. Table 6 shows the weight and bias values.

In Table 7, the outputs of ANN are evaluated with the measured value for the testing data. The percentage error (ε) was calculated using the relation given in the following equation:where and are the experimental and projected (ANN) tool wear values, respectively. The expected tool wear has an average error of 5.92%. Due to fewer percentage errors in training and testing data, the created ANN model was very successful in predicting tool wear in hard turning utilizing coated carbide inserts.

4. Conclusions

In this work, the tool wear in hard turning of AISI 4140 alloy steel utilizing coated carbide inserts was experimentally evaluated at various combinations of cutting parameters based on the design of the experiments. For tool wear monitoring, force sensors are employed. An ANN-based wear estimator for coated carbide tools using features significant to hard turning has been proposed based on the investigation. The following points are concluded.(1)The flank wear values range from 0.06 to 0.493 mm. The flank wear images are demonstrated in Figures 6(a) and 6(b). For most of the cutting conditions, abrasion is the cause of flank wear, and the value of flank wear is within the prescribed limit of 0.3 mm. Hence, the cost-effective coated carbide tool is proven to be an effective alternative to CBN and ceramic tools, which are costlier.(2)ANOVA shows that all the machining lengths as well as process parameters are significant for tool wear, with machining length being the most influencing factor. The depth of cut is the next influencing factor, followed by feed. The percentage contribution ratio reveals that machining length, speed, feed, and depth of cut were taken as inputs for model development.(3)From the force analysis, flank wear raises the amplitude of force signals. This could be due to an increase in friction, which encourages amplitude growth. The thrust force along the radial direction “F_r” has been found to be substantially connected with wear when compared to “F_t” and “F_a” and is thus used as input for the tool wear model.(4)This research also focuses on the creation of an ANN model of a cutting process to forecast tool wear in hard turning. The ANN model for tool wear prediction was trained using experimental data with wear as an output. The performance of the ANN model was then evaluated by comparing the ANN predictions with experimental findings from the testing phase that were not used in the training stage. Five learning algorithms were employed to forecast tool wear by altering the hidden neurons; the LM learning algorithm for the network 5–10–1 produced the best ANN results.(5)For the selected network. It was found that the R² values are 0.996602 and 0.969437 for both training and testing data. The values are very much closer to +1, indicating a high correlation. The MSE values for training data are as small as 0.000133152 and 0.004443. The MEP is 0.575407, 2.977617, 0.014, and 2.19% for the testing data, which indicates very high prediction accuracy. Therefore, the developed ANN model can be considered for tool condition monitoring during hard turning.

Nomenclature

a_p:	Depth of cut (mm)
f:	Feed rate (mm/rev)
F_a:	Feed force (N)
F_r:	Thrust force (N)
F_t:	Tangential force (N)
HRC:	Rockwell hardness
:	Cutting speed (m/min)
:	Tool wear (mm)
L:	Machining length (mm)
MSE:	Mean square error.

Data Availability

We recognize it is not always possible to share research data publicly, for instance when individual privacy could be compromised, and in such instances data availability should still be available in the manuscript.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

References

S. R. Das, D. Dhupal, and A. Kumar, “Experimental investigation into machinability of hardened AISI 4140 steel using TiN coated ceramic tool,” Measurement, vol. 62, pp. 108–126, 2015.
View at: Publisher Site | Google Scholar
H. Aouici, M. A. Yallese, K. Chaoui, T. Mabrouki, and J. F. Rigal, “Analysis of surface roughness and cutting force components in hard turning with CBN tool: prediction model and cutting conditions optimization,” Measurement, vol. 45, no. 3, pp. 344–353, 2012.
View at: Publisher Site | Google Scholar
E. Aslan, “Experimental investigation of cutting tool performance in high speed cutting of hardened X210 Cr12 cold-work tool steel (62 HRC),” Materials & Design, vol. 26, no. 1, pp. 21–27, 2005.
View at: Publisher Site | Google Scholar
İ. Asiltürk and H. Akkus, “Determining the effect of cutting parameters on surface roughness in hard turning using the Taguchi method,” Measurement, vol. 44, no. 9, pp. 1697–1704, 2011.
View at: Publisher Site | Google Scholar
R. Suresh, S. Basavarajappa, and G. L. Samuel, “Some studies on hard turning of AISI 4340 steel using multilayer coated carbide tool,” Measurement, vol. 45, no. 7, pp. 1872–1884, 2012.
View at: Publisher Site | Google Scholar
G. Bartarya and S. K. Choudhury, “State of the art in hard turning,” International Journal of Machine Tools and Manufacture, vol. 53, no. 1, pp. 1–14, 2012.
View at: Publisher Site | Google Scholar
S. Pal, S. Heyns, B. H. Freyer, N. J. Theron, and S. K. Pal, “Tool wear monitoring and selection of optimum cutting conditions with progressive tool wear effect and input uncertainties,” Journal of Intelligent Manufacturing, vol. 22, no. 4, pp. 491–504, 2009.
View at: Publisher Site | Google Scholar
G. H. Lim, “Tool wear Monitoring in machine turning,” Journal of Materials Processing Technology, vol. 51, no. 1-4, pp. 25–36, 1995.
View at: Publisher Site | Google Scholar
M. Rizal, J. A. Ghani, M. Z. Nuawi, and C. H. C. Haron, “Online tool wear prediction system in the turning process using an adaptive neuro-fuzzy inference system,” Applied Soft Computing, vol. 13, no. 4, pp. 1960–1968, 2013.
View at: Publisher Site | Google Scholar
E. Dimla and Dimla, “Sensor signals for tool-wear monitoring in metal cutting operations-a review of methods,” International Journal of Machine Tools and Manufacture, vol. 40, no. 8, pp. 1073–1098, 2000b.
View at: Publisher Site | Google Scholar
A. Siddhpura and R. Paurobally, “A review of flank wear prediction methods for tool condition monitoring in a turning process,” The International Journal of Advanced Manufacturing Technology, vol. 65, no. 1-4, pp. 371–393, 2013.
View at: Publisher Site | Google Scholar
Y. Cheng, X. Gai, R. Guan, Y. Jin, M. Lu, and Y. Ding, “Tool wear intelligent monitoring techniques in cutting: a review,” Journal of Mechanical Science and Technology, vol. 37, no. 1, pp. 289–303, 2023.
View at: Publisher Site | Google Scholar
D. E. Dimla and P. M. Lister, “On-line metal cutting tool condition monitoring. I: force and vibration analyses,” International Journal of Machine Tools and Manufacture, vol. 40, no. 5, pp. 739–768, 2000a.
View at: Publisher Site | Google Scholar
P. Zhang, D. Gao, Y. Lu, Z. Ma, X. Wang, and X. Song, “Cutting tool wear monitoring based on a smart toolholder with embedded force and vibration sensors and an improved residual network,” Measurement, vol. 199, Article ID 111520, 2022.
View at: Publisher Site | Google Scholar
V. P. Astakhov and X. Xiao, “A methodology for practical cutting force evaluation based on the energy spent in the cutting system,” Machining Science and Technology, vol. 12, no. 3, pp. 325–347, 2008.
View at: Publisher Site | Google Scholar
D. Rajeev, D. Dinakaran, and S. C. E. Singh, “Artificial neural network based tool wear estimation on dry hard turning processes of AISI4140 steel using coated carbide tool,” Bulletin of the Polish Academy of Sciences, Technical Sciences, vol. 65, no. 4, pp. 553–559, 2017.
View at: Publisher Site | Google Scholar
T. Ozel and Y. Karpat, “Predictive modeling of surface roughness and tool wear in hard turning using regression and neural networks,” International Journal of Machine Tools and Manufacture, vol. 45, no. 4-5, pp. 467–479, 2005.
View at: Publisher Site | Google Scholar
C. Scheffer, H. Kratz, P. S. Heyns, and F. Klocke, “Development of a tool wear-monitoring system for hard turning,” International Journal of Machine Tools and Manufacture, vol. 43, no. 10, pp. 973–985, 2003.
View at: Publisher Site | Google Scholar
X. Wang, W. Wang, Y. Huang, N. Nguyen, and K. Krishnakumar, “Design of neural network-based estimator for tool wear modeling in hard turning,” Journal of Intelligent Manufacturing, vol. 19, no. 4, pp. 383–396, 2008.
View at: Publisher Site | Google Scholar
B. Sick, “On-line and indirect tool wear monitoring in turning with artificial neural networks: a review of more than a decade of research,” Mechanical Systems and Signal Processing, vol. 16, no. 4, pp. 487–546, 2002.
View at: Publisher Site | Google Scholar
A. Eser, E. Aşkar Ayyıldız, M. Ayyıldız, and F. Kara, “Artificial intelligence-based surface roughness estimation modelling for milling of AA6061 alloy,” Advances in Materials Science and Engineering, vol. 2021, Article ID 5576600, 10 pages, 2021.
View at: Publisher Site | Google Scholar
F. Kara, M. Karabatak, M. Ayyıldız, and E. Nas, “Effect of machinability, microstructure and hardness of deep cryogenic treatment in hard turning of AISI D2 steel with ceramic cutting,” Journal of Materials Research and Technology, vol. 9, no. 1, pp. 969–983, 2020.
View at: Publisher Site | Google Scholar
C. Scheffer, Development of a Wear Monitoring System for Turning Tools Using Artificial Intelligence, University of Pretoria, Pretoria, South Africa, 2002.
Ö. Erkan, B. Işık, A. Çiçek, and F. Kara, “Prediction of damage factor in end milling of glass fibre reinforced plastic composites using artificial neural network,” Applied Composite Materials, vol. 20, no. 4, pp. 517–536, 2013.
View at: Publisher Site | Google Scholar
A. K. Sahoo and B. Sahoo, “A comparative study on performance of multilayer coated and uncoated carbide inserts when turning AISI D2 steel under dry environment,” Measurement: Journal of the International Measurement Confederation, vol. 46, no. 4, pp. 2695–2704, 2013.
View at: Publisher Site | Google Scholar
V. P. Astakhov, “Machining of hard materials – definitions and industrial applications,” in Machining of Hard Materials, J. P. Davim, Ed., pp. 1–32, Springer, London, UK, 2011.
View at: Google Scholar

Copyright

Copyright © 2023 S. Ajitha Priyadarsini et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

PDF Download Citation

Download other formats

Order printed copies