INTERNATIONAL JOURNAL OF MECHANICAL ENGINEERING AND TECHNOLOGY (IJMET)

INTERNATIONAL JOURNAL OF MECHANICAL ENGINEERING AND TECHNOLOGY (IJMET) International Journal of Mechanical Engineering and Technology (IJMET), ISSN 0976 6340(Print), ISSN 0976 6340 (Print) ISSN 0976 6359 (Online) Volume 5, Issue 7, July (2014), pp. 125-133 IAEME: www.iaeme.com/ijmet.asp Journal Impact Factor (2014): 7.5377 (Calculated by GISI) www.jifactor.com IJMET I A E M E PREDICTION OF VIBRATIONS, CUTTING FORCE OF SINGLE POINT CUTTING TOOL BY USING ARTIFICIAL NEURAL NETWORK IN TURNING Prof. L. B. Raut 1, Prof. Matin Amin Shaikh 2 1, 2 (Department of Mechanical Engineering, SVERI s College of Engineering Pandharpur, India) ABSTRACT In this paper the objective of this work is to develop a model to simulate the vibrational effects of rotating machine parts on the single point cutting tool and cutting force acting on single point cutting tool in turning. In this paper experimental studies were performed on turning process & vibration is measured with the help of accelerometer along with a device called as Fast Fourier Transformer (FFT) Analyzer and cutting force is measured with the help of Tool dynamometer. The vibration of single point cutting tool is sensed by accelerometer located on the tool-post of lathe machine. The accelerometer will send the sensed vibration to FFT Analyzer which can be convert the sensed data by using accelerometer shown in PC such as frequency, Amplitude, displacement & so on and cutting force is sensed by strain gauges which are compacted in tool post. The sensed force will send to dynamometer, it displays the cutting force. The obtained experimental data given to an Artificial Neural Network (ANN) in Matlab, with the help of experimental data ANN is to be trained. And by using ANN can predict the vibrations and cutting force by changing parameters of turning such as spindle speed, feed & depth of cut. This model of ANN can be predict vibrations of single point cutting tool and cutting force acting on single point cutting tool to avoid the failure of cutting tool. Keywords: Vibrations, Cutting Force, Cutting Tool, Turning, ANN, Prediction. 1. INTRODUCTION Much emphasis has been placed upon vibrations in machine tools during recent years because many people have recognized that accuracy, surface finish and, last but not least, production costs are considerably influenced by them. Today an arsenal of sophisticated instruments is available for the investigation of machine tool vibration. Cutting tool have always vibrated and will continue to do 125

so. We strive to measure these vibrations and keep them at or below a tolerable level. While higher cutting speeds generally contribute to an improvement of the surface finish obtained, but they increase the vibrations of machine & often excite components of the machine tool at their natural frequency. The exciting force is trying to cause vibration of cutting tool. If the vibrations are increased then the failure of cutting tool occurred. The failure of cutting tool results wastage of time, money etc... In metal cutting operation our goals to increasing productivity, reliability and quality of work piece. So through prediction of vibration of cutting tool by using some developed artificial neural network. In this work strive to predict vibrations and cutting force coming on cutting tool at the time of turning and keep them at or below a particular level to avoid the cutting tool failure. 2. EXPERIMENTAL SET UP & INSTRUMENTATION There are many parameters which affect the vibrations & cutting force of cutting tool. In this experimental study, the structural parameters for the machine tool variables are constant for every experiment and also all the experiments have been completed on the same machine tool.similarly, cutting tool parameters are constant because the cutting tool used has the same characteristics. Also the cutting parameters have been reduced to three to simplify matters. Variable cutting conditions have been selected such as listed in Table 1. In this study 31 different cutting conditions have been considered. The whole work was done on Conventional lathe machine. The work piece material used was EN-8 and the tool used was TiN coated carbide insert. En8 has a hardness of 35 HRC. It is mainly used for engine shafts, studs, connecting rods, dynamo and motor shafts etc. the workpiece material in tests was selected to represent the major group of workpiece materials used in the industry. The specimen was cylindrical bar with 40mm diameter. After removing the surface imperfections on the workpiece the 31 different cutting conditions applied and respective vibration & cutting force are measured with the help of accelerometer-fft Analyzer and Tool dynamometer respectively. Fig. 1: Actual photograph of Test rig & Top view of Test rig with FFT Analyzer & Tool dynamometer This present paper presents an experimental study to investigate the effects of cutting parameters like spindle speed, feed and depth of cut on surface finish on EN-8. In this work EN-8 material used as workpiece and Orthogonal machining process is used. So, only two forces acting on 126

the cutting tool Axial feed force or thrust force (Fa) acting in horizontal plane parallel to the axis of workpiece and Cutting force or tangential force (Ft) acting in vertical plane and tangent to the work surface. And the resultant force or total force (R) acting on the cutting tool can be calculated by using below formula. The orthogonal cutting method is as shown in Fig. 2. Fig. 2: Orthogonal machining The cutting conditions was carried out without coolant and totally 31 experiments were performed according to full factorial design. The photograph of experimental test rig is shown in Fig. 1. The vibration & cutting force parameter generally depend on the manufacturing conditions like feed, depth of cut, cutting speed, machine tool and cutting tool rigidity etc. In this study three main cutting parameters, feed, cutting speed and depth of cut was used. Three cutting parameters for each factor were used because the considered variables are multi-level variables and their outcome effects are not linear. Table no. 1 shows the full experimental data. The Vibration graph shown by the FFT Analyzer at Spindle speed 30rpm, feed 0.85mm/revol, depth of cut 0.5mm and at spindle speed 72 rpm, feed 0.85mm/revol, depth of cut 0.8mm is shown in Fig. 3. Fig. 3 Vibration graph at Spindle speed 30rpm, Feed 0.85mm/revol, Depth of cut 0.5mm and at Spindle speed 72rpm, Feed 0.85mm/revol, Depth of cut 0.8mm 127

Sr. No. Spindle Speed (rpm) Table no. 1: All data to design ANN Feed Depth of Vibrations (mm/revol. cut (mm) (RMS) ) Cutting Force (kn) 1 30 0.85 0.1 295.6 44.9 2 30 0.85 0.25 303.8 48.5 3 30 0.85 0.5 306 60 4 30 0.85 0.6 315.1 62.2 5 30 0.85 0.75 322.9 64.9 6 30 0.85 1.2 400.6 71.7 7 47 0.85 0.25 523 40.5 8 47 0.85 0.5 538 55.5 9 47 0.85 0.75 543 72.9 10 72 0.85 0.25 571 39 11 72 0.85 0.4 608 49.8 12 72 0.85 0.5 787 56.5 13 72 0.85 0.6 828 61.5 14 72 0.85 0.75 941 72.1 15 72 0.85 0.8 974 74.7 16 110 0.85 0.25 706 28.6 17 110 0.85 0.5 820 41.2 18 110 0.85 0.75 880 49.4 19 196 0.85 0.25 546 19.2 20 196 0.85 0.4 844 25.5 21 196 0.85 0.5 889 28.6 22 196 0.85 0.7 1067 36.4 23 196 0.85 0.75 1099 38.3 24 310 0.85 0.25 614 13.4 25 310 0.85 0.5 981 27.9 26 310 0.85 0.75 1463 39.6 27 473 0.85 0.25 700 10.3 28 473 0.85 0.5 851 19.2 29 473 0.85 0.75 1001 25.6 30 733 0.85 0.25 819 11.7 31 733 0.85 0.5 1122 24.4 3. ARTIFICIAL NEURAL NETWORK Artificial neural networks are information processing systems, and since their inception, they have been used in several areas of engineering applications. ANNs have been trained to solve nonlinear and complex problems that are not modeled mathematically. ANNs eliminate the limitations of the classical approaches by extracting the desired information using the input data. Applying ANN to 128

International Journal of Mechanical ISSN 0976 6359(Online), Volume 5, a system needs sufficient input and output data instead of a mathematical equation. Furthermore it can continuously retrain for new dataa during the operation, thus it can adapt to changes in the system. Artificial Neural Networks are non-lineabrain. They are powerful tools for modeling, especially when the underlying data relationship is mapping structures based on the function of the human unknown. ANNs can identify and learn correlated patterns between input data sets and corresponding target values. After training, ANNs can be used to predict the outcome of new independent input data. ANNs imitate the learning process of the human brain and can process problems involving non- to predict linear and complex data. In this work, artificial neural network model have been developed vibrations & cutting force in the machining of EN8 material. 4. PROCEDURE FOR PREDICTION Engineering and Technology (IJMET), ISSN 0976 6340(Print), Issue 7, July (2014), pp. 125-133 IAEME The experiment data is divided in to test data set. Test data is used to check the behavior of the ANN model created to fit the sample of 31; preferred ratio selected is 9:22. The training data to train the network is shown in Table no. 2, as well as test data is shown in Table no. 3. Next the number of nodes in hidden layers is being taken 2. The Levenberg-Marquardt training algorithm was found to be the best fit for application because it can reduce the MSE to a significantly small value and can provide better accuracy of prediction. The transfer function, training function, learning function and performance functions used in this study are tansig, trainlm, learngdm and MSE respectively. So a network of 3 input nodes, 2 hidden nodes and 1 output node is created, so 3-2-1 network is structured. So neural network model with feed forward back propagation algorithm and Levenberg-Marqudt approximation algorithm was trained with data collected for the experiment. The neural network has training windoww is shown in Fig.4. The effectiveness of ANN model is fully depends on the trial and error process. The regression graph shown by the modeled network is shown below Fig.5. Fig. 4: ANN training tool Fig. 5: Regression graph 129

After training the network considering above explained all parameters, the network is test with test data. The graph between the actual and predicted values has also been plotted and from the graph it is clear that the actual and predicted results come to a very close value. 5. RESULT & DISCUSSION After training the network the results shows that the training data and the predicted training data has come to a very close value. The graph shows the result of the training data of the actual value with the predicted value. For training data & testing data, actual value & predicted value of cutting force are compared shown in Fig.6 and testing data & training data, actual value & predicted value of vibrations are compared shown in Fig.7. The test data has predicted values are shown in Table no.4. and the errors obtained in this model because of weights required training the ANN model & these weights NNTOOL taken randomly, this is trail error method we can t change the weights. Sr. No. Spindle Speed (rpm) Table no.2: Training data for ANN Depth Feed Vibrations of cut (mm/revol.) (RMS) (mm) Cutting Force (kn) 1 30 0.85 0.1 295.6 44.9 2 30 0.85 0.25 303.8 48.5 3 30 0.85 0.5 306 60 4 30 0.85 0.6 315.1 62.2 5 30 0.85 0.75 322.9 64.9 6 30 0.85 1.2 400.6 71.7 7 47 0.85 0.25 523 40.5 8 47 0.85 0.5 538 55.5 9 47 0.85 0.75 543 72.9 10 72 0.85 0.25 571 39 11 72 0.85 0.4 608 49.8 12 72 0.85 0.5 787 56.5 13 72 0.85 0.6 828 61.5 14 72 0.85 0.75 941 72.1 15 72 0.85 0.8 974 74.7 16 110 0.85 0.25 706 28.6 17 110 0.85 0.5 820 41.2 18 110 0.85 0.75 880 49.4 19 196 0.85 0.25 546 19.2 20 196 0.85 0.4 844 25.5 21 196 0.85 0.5 889 28.6 22 196 0.85 0.7 1067 36.4 130

Table no. 3: Testing data for ANN Sr. No. Spindle Speed (rpm) Feed (mm/revol.) Depth of cut (mm) Vibrations (RMS) Cutting Force (kn) 1 196 0.85 0.75 1099 38.3 2 310 0.85 0.25 614 13.4 3 310 0.85 0.5 981 27.9 4 310 0.85 0.75 1463 39.6 5 473 0.85 0.25 700 10.3 6 473 0.85 0.5 851 19.2 7 473 0.85 0.75 1001 25.6 8 733 0.85 0.25 819 11.7 9 733 0.85 0.5 1122 24.4 The error is calculated using absolute percent error given by the relation, 100 45 80 40 70 35 60 Cutting Force (kn) 30 25 20 15 Cutting Force (kn) 50 40 30 10 20 5 10 0 0 1 2 3 4 5 6 7 8 9 1 3 5 7 9 11 13 15 17 19 21 No. of Readings No. of Readings Experimental Predicted Experimental Predicted Fig. 6: Comparison of Experimental& Predicted Cutting force of Train data& Test data respectively 131

1600 1400 1200 1000 RMS 800 600 400 200 1200 1000 800 RMS 600 400 200 Fig. 7: Comparison of Experimental & Predicted Vibrations of Train data& Test data respectively Sr. No. 0 1 2 3 4 5 6 7 8 9 No. of Readings Experimental Predicted Table no. 4: Error between Experimental values & Predicted values Vibration (RMS) Cutting Force (kn) Experimental Predicted % Error Experimental Predicted % Error 1 1099 1073 2.41 38.3 40.1 4.5 2 614 614 0 13.4 13.56 1.2 3 981 981 9 27.9 26.85 3.93 4 1463 1463 3.76 39.6 37.21 6.41 5 700 700 2.94 10.3 10.16 1.38 6 851 851 0 19.2 19.5 1.5 7 1001 1001 0 25.6 26.12 2 8 819 810 1.11 11.7 10.3 13.6 9 1122 1105 1.54 24.4 23.7 2.95 0 1 3 5 7 9 11 13 15 17 19 21 No. of Readings Experimental Predicted 6. CONCLUSION From the results it can be easily seen that the minimum error obtained for the predicted value of test data. This study concludes that the model of ANN can be predict the vibrations & cutting force of single point cutting tool at any three parameters such as spindle speed, feed & depth of cut. And this predicted value is nearly equal to actual value of vibrations & cutting force respectively. So with the help of ANN model we can easily predict the vibrations & cutting force of single point cutting tool without any experiment. And effectiveness of ANN model can be improved by modifying the number of layers and nodes in the hidden layers of the ANN network structure. 132

REFERENCES [1] D. J. Kim, B. M. Kim has present Application of neural network and FEM for metal forming process.sep.-1999. [2] W. S. Lin, B. Y. Lee, C. L. Wu has presented Modeling the surface roughness and cutting force for turning.april-1999. [3] Thomas M. Beauchamp Y, Youssef A. Y. Masounave J. Effect of tool vibrations on surface roughness during lathe dry turning process. [4] F. M. Longbottom and F. D. Lanbam Cutting temperature measurement while machining. [5] Bahaa Ibraheem Kazem, Nihad F.H. Zangana, A Neural Network based real time controller for turning process. ISSN 1995-6665, Sep.-2007. [6] Alex Sohn, Lucas Lamonds and Ken Garrard, Modeling of Vibration in single- point diamond turning 2006. [7] Dennis H. Shreve, Introduction to vibration technology IRD Mechanalysis, Inc. Columbus, Ohio 43229, November 1994. [8] Ashvin P. Yajnik, Vibration in machine tools. [9] L. Håkansson, I. Claesson and P.-O. H. Sturesson, Adaptive Feedback Control of Machine- Tool Vibration based on The Filtered-x LMS-algorithm. [10] A. Bhattacharyya, Metal cutting theory & practice, New central book agency (P) ltd. 8/1 Chintamoni das lane, Kolkata 700 009 India. [11] Amitabh Ghosh, Ashok Kumar Mallik, Manufacturing Science Affiliated East-West press private limited New Delhi. [12] G.R.Nagpal, Machine tool engineering Khanna publishers. [13] Serope Kalpakjian, Steven R. Schmid, Manufacturing Processes PEARSON. [14] Bhandari V.B, Design of Machine Elements, Tata McGraw-Hill Publishing Company Limited New Delhi, 2004, pp234-237. [15] Ghosh and Mallik, Theory of Mechanism and Machines, East-West Press Private Limited New Delhi India pp. 290-293. [16] Vinay Babu Gada, Janakinandan Nookala, Suresh Babu G, The Impact of Cutting Conditions on Cutting Forces and Chatter Length for Steel and Aluminum April 2013. [17] Tejinder pal singh, Jagtar singh, Jatinder madan and Gurmeet kaur, Effects of Cutting Tool Parameters on Surface Roughness, International Journal of Mechanical Engineering & Technology (IJMET), Volume 1, Issue 1, 2010, pp. 182-189, ISSN Print: 0976 6340, ISSN Online: 0976 6359. [18] Ajeet Kumar Rai, Richa Dubey, Shalini Yadav and Vivek Sachan, Turning Parameters Optimization for Surface Roughness by Taguchi Method, International Journal of Mechanical Engineering & Technology (IJMET), Volume 4, Issue 3, 2013, pp. 203-211, ISSN Print: 0976 6340, ISSN Online: 0976 6359. [19] A. Hemantha Kumar, Krishnaiah.G and V.Diwakar Reddy, ANN Based Prediction of Surface Roughness in Turning, International Journal of Mechanical Engineering & Technology (IJMET), Volume 3, Issue 2, 2012, pp. 162-170, ISSN Print: 0976 6340, ISSN Online: 0976 6359. [20] Nitin Sharma, Shahzad Ahmad, Zahid A. Khan and Arshad Noor Siddiquee, Optimization of Cutting Parameters for Surface Roughness in Turning, International Journal of Advanced Research in Engineering & Technology (IJARET), Volume 3, Issue 1, 2012, pp. 86-96, ISSN Print: 0976-6480, ISSN Online: 0976-6499. 133