International Journal of Advanced Mechanical Engineering. ISSN 2250-3234 Volume 8, Number 1 (2018), pp. 61-69 Research India Publications http://www.ripublication.com Tool Life, Force and Surface Roughness Prediction by Variable Cutting Parameters for Coated and Uncoated Tool Kuldip A. Baviskar 1, U.S. Chavan 2 and Deepak Kulkarni 3 1 Mechanical Engineering Department, Vishwakarma Institute of Technology, Pune, 2 Mechanical Engineering Department, Vishwakarma Institute of Technology,Pune,411037, India. 3 Engineering Manager, Kalyani Studio, Pune 411006, India. Abstract This paper presents the obtained mathematical models of Feed Force, Cutting Force, Thrust Force, Surface Roughness and Tool life during turning operation for uncoated and coated tool insert as a function of processing parameters cutting speed, feed and depth of cut. The turning operation has been performed on the CNC Lathe machine using tungsten carbide (uncoated and coated with Al2O3) tool inserts and the work piece material is AISI 4340. Process influencing parameters considered are in the range between A = 150 and 300 m/min, B = 0.16 and 0.32 mm/rev and C = 0.5 and 1.5 mm. Cutting forces are measured using computerized experimental setup with KISTLER s three components piezoelectric dynamometer, surface roughness is measured with MITITOYO s portable surface roughness tester and tool life is measured with measuring flank wear by microscope. Experiments are performed as per the first order three factorial experimental plan for coated and uncoated tool insert. Mathematical modeling of cutting forces, surface roughness and tool life have been prepared with regression by MINITAB 18 and with ANN by MATLAB 18. Keywords: Mathematical Model, Cutting Forces, Surface Roughness, Tool Life ANN. I. INTRODUCTION In modern manufacturing industries it is not enough to improve productivity but also the quality of component or product in terms of dimensional accuracy and surface finish by selecting proper tool-work piece, cutting parameters, coating material, Life of tools for inventory plan for undisturbed production. It is very inconvenient to
62 Kuldip A. Baviskar, U.S. Chavan and Deepak Kulkarni predict the quality, productivity, life of tool, cutting parameters, coating materials, forces based on mere experience. Also to find all above parameters before operation, it is very costly, time consuming and not feasible with experimentation so there is a need of some other tools which is cost effective and time saving to predict the parameters. Low cost manufacturing along with superior quality products in less time span is the goal of modern industry. There has been a few data collection constraints exists which go hand in hand with machining effecting market like parameters for maximum efficiency, production and cutting economic conditions. (3) To determine the machining results, tool wear and tool strength, it is necessary to have data about the cutting forces magnitude in turning process as function of the parameters and conditions of tool treatment (1). Forecasting tool life by machining process models can help gaining you a competitive edge (5). In this global competition of low cost and high quality solution in the market, companies are forced to optimize their processes but constraints faced by the engineers in accessing important machining parameters such as Taylor and Kienzle has led to difficulty in defining production parameters (3). These defined cutting parameters should ensure that the work piece should maintains the stricter tolerance and enhance part quality (5). It has been always a preferable option to use analytical tools for process optimization over costly trial and error method. Dynamic models of cutting processes provide the ability to apply large combination of process parameter to improve tool life and predict stable cutting conditions. One such popular model widely used for predicting machining performance by researcher is Artificial Neural Networks (ANN) model. In this research, the experiments are performed on the CNC Lathe machine with the KISTLER S Piezoelectric Dynamometer setup to get the three directional forces generated on the tool insert. After getting those results, surface finish and tool wear has been measured by using MITUTOYO s Surface roughness tester and Microscope respectively. II.OBJECTIVE To develop a Mathematical Model by regression and ANN for predicting Tool Life, Cutting Forces and Surface Roughness Values when turning AISI 4340 material rod using uncoated and coated tungsten carbide tipped tool under various cutting conditions. To propose better mathematical model to predict Tool Life, Cutting Forces and Surface Roughness Values. To Study the effect of coating on performance of tool insert III.SCOPE OF WORK Perform experimental runs as per experimental plan for uncoated and coated tool insert. Develop a regression model for predicting Tool Forces, Surface Roughness and Tool Life by using Minitab software.
Tool Life, Force and Surface Roughness Prediction by Variable Cutting 63 Develop an Artificial Neural Network (ANN) model for predicting Tool Forces, Surface Roughness and Tool Life by using MATLAB 18 Compare the Predicted results from regression model and ANN with Experimental results and propose accurate method for predicting results. IV.EXPERIMENTAL SETUP (a) Workpiece The workpiece material selected in this thesis is AISI 4340 which is generally used to manufacture crankshafts in industry. The workpiece used for the experiments are having 100 mm length and 50 mm diameter as shown in fig. (b) Cutting Tool Insert: Here in this thesis the tool insert which used is made by Carmet tools. Tool insert used are of Tungsten carbide material and coated with Al2O3 material by CVD coating method as shown in figure. is (c) Force Measurement: KISTLER S Piezoelectric Dynamometer attached with the base plate is used to measure the forces generated on the tool insert. Dynamometer is fixed in the turret on which tool holder is mounted and it is connected with the computer through the 3 component signal s amplifier and A/D Card represented in schematic view of dynamometer setup.
64 Kuldip A. Baviskar, U.S. Chavan and Deepak Kulkarni When the tool gets in contact with the workpiece, the electric signal from dynamometer are send to signal amplifier which amplifies the signals and transfers amplified signals to A/D Card which converts AC signals to DC signals. These DC signals are read into the KISTLER s Dynoware software which are displayed on the computer monitor. The individual graphs of three directional forces have been generated in Dynoware for each run separately. (d) Surface Roughness Measurement: MITUTOYO s portable surface roughness tester is very sensitive measuring device with the stylus which detects the profile of the measuring surface. This stylus is main sensitive part of the tester which should not fall on the floor or hit by any means. the (e) Tool Life Measurement: Tool life has been measured by applying criteria for rejection of tool insert after flank wear of 0.3 mm. (f) Experimental Plan: The experiments have been carried out using full factorial design of experiment in order to achieve tool life, surface roughness and tool forces generated on tool insert. From the literature survey it has been understood that there are three important parameters which affects the tool life, surface roughness and tool forces most. So in this thesis three cutting parameters are considered as an influencing
Tool Life, Force and Surface Roughness Prediction by Variable Cutting 65 factor which are cutting speed, feed and depth of cut. Each factor is having two levels that is low and high. Levels Factors Name Low High Unit A Cutting Speed 150 300 m/min B Feed 0.16 0.32 mm/rev C Depth of cut 0.5 1.5 mm Number of Factors (k) = 3 Number of levels (L) = 2 Number of experiments = Lk = 23 = 8 (g) Experimental Procedure: Workpiece is positioned in the chuck of CNC Lathe machine and tool along with Piezoelectric Dynamometer is mounted on the turret. A cable from Piezoelectric Dynamometer is connected to the three component Amplifier. Procedure stated above for Piezoelectric dynamometer is carried out to measure tool forces. Program is fed into the CNC Lathe with the help of keys. Start the CNC machine and simultaneously start data acquisition on Dynoware software so that software will plot actual data of generated forces with respect to time. Same procedure is conducted for all experiments with different set of tool and workpiece. Surface finish of the machined workpieces are measured with the help of MITUTOYO s surface roughness tester by following the procedure stated above for surface roughness measurement. Tool flank wear of the tool inserts after every machining are being measured by the microscope and time required to worn out the inserts are measured by going through the steps which are mentioned in tool wear measurement. Results for forces (Fx, Fy, Fz), Surface Roughness (Ra) and Tool Life (T) are obtained through experiments. V. MATHEMATICAL REGRESSION MODEL: In this research, regression method is applied to develop mathematical model to predict the tool forces, surface roughness and tool life. The regression model is generated on the MINITAB 18 software of student s version and analyzed the models with ANOVA technique. The individual mathematical models were obtained for Fx (Feed Force), Fy (Cutting Force), Fz (Thrust Force), Ra (Surface Roughness) and TL
66 Kuldip A. Baviskar, U.S. Chavan and Deepak Kulkarni (Tool Life) for uncoated and coated tool insert separately. The effect of input parameters on output are also observed. ANN models are generated in MATLAB 18 student s version which consists of processors called Neurons that are linked by weighted interconnections. It can easily predict the output results from huge and complicated data base. It develops analytical model to solve problem of prediction, diagnosis and decision-making. It includes learning data as an experimental result for preparation of model and generated model can predict output for any number of variation of input parameter.
Tool Life, Force and Surface Roughness Prediction by Variable Cutting 67 VI. EXPERIMENTAL RESULTS: The following results were obtained after going through each experiment for uncoated tungsten carbide tool and for coated tungsten carbide tool. (a) Model Effect (b) Coating Effect:
68 Kuldip A. Baviskar, U.S. Chavan and Deepak Kulkarni VII. CONCLUSION: Following are the conclusions made after going through above research and experiments - Individual effects of process parameters on response are studied. It is observed that the feed plays significant role in feed force, cutting speed plays significant role in cutting force, depth of cut plays significant role in thrust force and feed plays significant role in surface roughness value ANN model proved to provide a better prediction with overall error within 10%. Also it provides lesser time for calculating once model is prepared. Surface quality of workpiece surface after machining with the coated tool is good compared with the uncoated tool. Tool Life of the coated tool is higher than that of uncoated tool with noticeable surface finish. Even though ANN is predicting results effectively, percentage error for regression model is less than 15% so it proves that mathematical models are extremely useful in machining processes for predicting results. VIII. REFERENCES [1] M. M. J. Patil, Investigation in Tool Life of Coated and Uncoated Carbide Tools in Turning., 2nd National conference, vol. second, pp. 452-456, 2010. [2] J. G. a. K. Mohandas, Tool Life Prediction Model Using Regression and Artificial Neural Network Analysis, International Science Press, vol. 3, pp. 9-16, 2010. [3] M. R. S. Y. a. M. S. S. Amir Mahyar Khorasani, Tool Life Prediction in Face Milling Machining of 7075 Al by Using Artificial Neural Networks (ANN) and Taguchi Design of Experiment (DOE), International Journal of Engineering and Techinology, vol. 3, no. 1, February 2011. [4] J. E. F. d. O. Noemia Gomes de Mattos de Mesquita, Life Prediction of Cutting Tool by the Workpiece Cutting Condition, Advance Materials research, vol. 223, pp. 554-563, 2011.
Tool Life, Force and Surface Roughness Prediction by Variable Cutting 69 [5] S. S. &. Ravikumar, Experimental analysis and mathematical modeling of optimized cutting parameters in microturning, Indian journal of Engineering & Material Science, vol. 21, pp. 397-408, August 2014. [6] Machining forces and Merchant s Circle Diagram (MCD), in Mechanics of Machining, vol. 2, Khargpur, IIT, 2015, pp. 22-36. [7] K. V. a. P. P. C. Ramesh kannan, Analysis the Residual Stress of cutting tool insert in turning of Mild Steel A Review, International Journal of Advanced Science and Technology, vol. 90, pp. 49-60, 2016.
70 Kuldip A. Baviskar, U.S. Chavan and Deepak Kulkarni