Investigating the Process Parameter Optimization of High-Speed CNC Milling - Research Paper Example

?INVESTIGATION OF PROCESS PARAMETER OPTIMIZATION OF HIGH SPEED CNC MILLING FOR COMPOSITE MATERIALS USING COMBINED GENETIC ALGORITHAM AN ARTIFICALNEURALNETWORK Ph.d-Research proposal (1) Abstract: Machinability in terms of quality and productivity is of great concern in a competitive market. CNC milling performance is evaluated in terms of surface roughness, material removal rate, taper and facing. CNC machining is done initially in a number of passes an final end milling is done in a single pass. The most common machining parameters are cutting speed, feed, depth of cut. To have a balance between productivity and quality, the machining parameters need to be optimised. This research work intends to optimise the parameters of CNC milling process by the application of artificial neural networks and genetic algorithms. The parameters are optimised for improved surface finish and higher material removal rate. (2) Introduction: High speed CNC milling process needs to balance between productivity and quality of the end product. While quality is measured in terms of surface finish, the productivity is measured in terms of material removal rate. Good machinable materials acquire a smooth finish as they can be easily cut with less power. They also cause minimum damage to the tool. If the material has improved material properties then it’s machinablity becomes difficult. Hence improving machinability without sacrificing performance is a challenge. Predicting the optimal parameters for CNC milling is difficult as the milling process depends on several factors. The important factors relative to the material include the thermal conductivity, toughness, chemical properties and the microstructure of the material. The other important factors are the geometry of the cutting tool and the parameters of CNC milling process. (3) Objectives : The aim of this research work is to Find the important factors that characterise the best performance of the high speed CNC milling process. Using combined Genetic Algorithm and Artificial neutral networks techniques to optimise these high speed CNC milling parameters by identifying the correlation between the factors like feed, depth of cutting and cutting speed. Based on the correlation of these parameters, the optimal set is to be determined for perfect output parameters like surface roughness and material removal rate. (4) High speed CNC Milling : CNC milling is the most fundamental operation in industrial machining. According to Mike S. Lou et al (1999 ) “The quality of the surface plays a very important role in the performance of milling as a good-quality milled surface significantly improves fatigue strength, corrosion resistance, or creep life. Surface roughness also affects several functional attributes of parts, such as contact causing surface friction, wearing, light reflection, heat transmission, ability of distributing and holding a lubricant, coating, or resisting fatigue”. Speed of the milling depends on the tool size used. If the tool size is smaller, then to avoid breakage of the tool the spindle speed required will be more. For milling with using micro tools the speed range of the spindles may be up to 60,000 rpm ( Datron white paper, 2005). Thus high speed milling involves higher RPM rates with more feed rates and small step overs. The advantage of this high speed is that there is less heating of the parts because of less time for the heat to feedback. The generated heat is 40% due to the friction and 20% due to deformation. For better quality in machining, the low milling force and cooler tools can be used so that the vibration is less. A white paper of DATRON (2005) states that “The high spindle speed reduces the chip load to less than 0.005”. Such a low chip load significantly reduces the forces between the tool and the material. High-speed/low-force machining yields less heat, reduces tool deflection, and allows machining of thinner walled work pieces. This all results in cooler machining, superior surface and edge quality, better accuracy and, as a by-product (of low force), easier workholding — since modular vacuum tables can be employed for quick set up and job changeover (particularly with thin flat substrates).” The general rules of thumb in case of high speed milling red lining of the spindle should be avoided, when feed rate is higher reduce the step over, evacuate the work piece fast to remove heat. (5) Composite materials and Tools : Composite materials are being widely used in many industrial applications. Hence milling these materials has become a pre-requisite. For this research work, the composite material that has been chosen is duplex stainless steel DSS 2205. According to Ramesh et al (2011), “ Duplex stainless steel is a family of grades combining good corrosion resistance with high strength and ease of fabrication. Their physical properties are between those of the austenitic and ferrite stainless steel but tend to be closer to the ferrite and to carbon steel. The chloride pitting and crevice corrosion resistance of the duplex stainless steel is a function of chromium, molybdenum and nitrogen content.” The chemical composition of the duplex stainless steel (UNS number S31803) is as shown in Table.1. Table. 1. The chemical composition of duplex stainless Steel. Such a composition provides good strength, toughness and ductility. The DSS 2205 is a commonly used grade owing to its high strength, less weight and reduced wall thickness. Type of tools is another important factor that determines the performance of high speed milling. Though carbon and steel tools are commonly used in machining, their usage in high speed milling is not recommended. For high speed milling, cemented carbide (tungsten carbide cobalt) is preferred due to less wear and tear. These carbide tools are capable of withstanding high temperatures than steel tools. Also metal carbide tools ( hard metal tools) which are combination of tungsten carbide and cobalt are stronger to withstand the temperatures in high milling speeds. In this research work cemented carbide tool is to be used as the cutting tool. (6) Literature survey : Literature shows that many researchers have worked to optimise the performance of CNC milling process to balance between productivity and good product characteristics. The commonly desired product characteristics are form stability, dimensional accuracy and surface smoothness. Researchers have used mathematical tools like SPSS (Statistical Package for Social Science), MINITAB and neural networks for statistical analysis of the process parameters. The Taguchi method has been adopted by researchers to identify the most optimal surface roughness for a specific combination of cutting parameters (Yang and Chen, 2001). Ghani et al (2004) had used the same Taguchi technique to optimise the CNC milling parameters for carbide insert tool coated with TiN over hardened steel material. The parameters optimised include cutting speed, depth of cut and feed rate. Another research by Oktem et al (2005) has developed optimal cutting criteria for smooth surface using “Response Surface Methodology” (RSM) with genetic Algorithm. The research by Kopac and Krajnik (2007) has concentrated on flank milling parameter optimisation for Al –Alloy moulds. They adopted a Grey – Taguchi method combined with orthogonal Array and have used a Grey Relational Analysis (GRA) for optimising multi process parameters. The Principal Component Analysis (PCA) method has been used by Datta et al (2009) to calculate the correlation among the milling parameters. This method basically finds the uncorrelated factors called principal components. The principal component with highest accountability acts as the single objective function that can be optimised. The research by Sanjit Moshat et al (2010), has utilised the Taguchi method to optimise the milling parameters after application of principal Component Analysis (PCA) over the individual responses. Ramesh et al (2011) propose another method of multiple regression on an orthogonal array combined with ANNOVA (Analysis of variance) for identifying the turning parameters for machining of Duplex Stainless Steel with “CVD triangular carbide insert” . They optimise the turning parameters by correlating the cutting speed, depth of cut and feed rate. Thus most of the previous researches have used the Taguchi’s method for optimising the milling or machining parameters. The present research intends to concentrate on artificial neural networks and genetic algorithms to optimise the high speed CNC milling parameters. (7) Design of Experiment : As discussed earlier, this research work intends to use the composite material Duplex stainless steel DSS 2205 and cemented carbide tool as the cutting tool. The size of the insert has been proposed to be chosen based on the identification system - ANSI B212.4 – 1986. Among the non - contact and contact method of surface roughness measurement, this research intends to utilise the contact method by Mitutoyo SJ 201 instrument. Based on data from machining handbook, the maximum cutting speed of cemented carbide tool has been identified as 300 m/min. with the depth of cut of 0.5, 1.0, 1.5 mm fro a feed rate of 0.1 , 0.15, 0.2 mm/rev. Based on these a combination of these parameters has been planned as shown in Table 2. Cutting speed in m /min. Feed rate in mm / rev. Depth of cut in mm. 200 0.1 0.5 200 0.1 1 200 0.1 1.5 200 0.15 0.5 200 0.15 1 200 0.15 1.5 200 0.2 0.5 200 0.2 1 200 0.2 1.5 250 0.1 0.5 250 0.1 1 250 0.1 1.5 250 0.15 0.5 250 0.15 1 250 0.15 1.5 250 0.2 0.5 250 0.2 1 250 0.2 1.5 Table . 2. Combination of milling Parameters (input vector for optimiser) The technical data have been chosen as proposed by Ramesh et al (2011), “ X-axis (drive unit) , Measuring range: 12.5mm , Measuring speed: 0.25, 0.5mm/s (0.25mm/s: S-type) , Traversing direction: backwards , Detector range: 350?m (-200?mto+150?m) , Detecting method: skid measurement , Measuring force: 4mN , Stylus tip: diamond, 90°/5?mR , Skid radius of curvature: 40mm , Skid force: less than 400mN and the Detecting method: differential inductance , Power supply: rechargeable battery , DIMENSIONS (W?D?H) , Control unit: 307?165?94mm , Drive unit: 115?23?26mm , MASS Control unit: approx.0.3kg (SJ-201) Drive unit: 0.2kg.”. The work piece of proper dimension has to be chosen and then allowed fro milling in high speed CNC lathe. Machining is planned to be carried out in the 18 possible combinations of the milling parameters. After machining, the surface roughness is measured in Mitutoyo SJ 201. Chip analysis needs to be carried out on the chips collected. (8) Optimising Methodology : Genetic Algorithm and Bayes Classifier of artificial neural network have been proposed to optimise or classify the best set of input parameter combination (cutting speed, feed rate, depth of cut) so as to get the best surface smoothness. Geetha and Kamraj (2010) propose a novel method of markov blanket – embedded genetic algorithm for efficient classification of feature (vector) set. They claim that “ the embedded Markov blanket based memetic operators add or delete features (or genes) from a genetic algorithm (GA) solution so as to quickly improve the solution and fine-tune the search. Empirical results suggest that MBEGA is effective and efficient in eliminating irrelevant and redundant features based on both Markov blanket and predictive power in classifier model.” The proposed architecture for the optimiser design for this research work is as shown in Fig. 1. Fig. 1. The proposed optimiser logic flow. The logic of Markov Blanket Embedded Genetic Algorithm fine tunes the optimisation using Markov Blanket by adding or deleting the chromosomes and checking for the desired chip analysis output. The parent is considered as cutting speed, the next child or chromosome is considered as feed rate and the next child is taken as the depth of cut. The optimised feature set is obtained at the output of the GA optimising logic. This then can be is further classified by Bayes classifier. The algorithm is proposed to choose a specific vector combination (chromosome) of the milling parameters and if the output criteria are not met then, allow them through markov blanket operation (Lamarckian learning) to get improved chromosome. Evolutionary operators are performed based on crossover, selection and mutation. (9) Discussion : This research work intends to implement the concepts of genetic algorithm and neural networks to optimise the milling parameters like cutting speed, feed rate and depth of cut. For the chosen combinations, the optimised output is expected to be better than the usual Taguchi method of parameter optimisation. This method would be superior than the other methods. (13)Conclusion : Thus this research work intends to optimise the parameters of CNC milling process by the application of artificial neural networks and genetic algorithms. The parameters are optimised for improved surface finish and higher material removal rate. (14) References Dr. Mike S. Lou, Dr. Joseph C. Chen & Dr. Caleb M. Li, (1999), “Surface Roughness Prediction Technique For CNC End-Milling” , Journal of industrial technology, Volume 15, Number 1 - November 1998 to January 1999 Reviewed Article Datron Whitw Paper , (2005) High-Speed Machining with Micro Tooling, www.datrondynamics.com retrieved 15 th Jan 2012. M.Ramesh, R.P.Elvin, K.Palanikumar, K.Hemachandra Reddy , (2011), “Surface Roughness Optimization Of Machining Parameters In Machining Of Composite Materials” International Journal of Applied Research in Mechanical Engineering, Volume-1, Issue-1, 2011 Yang L.J. and Chen C.J.,( 2001).” A systematic approach for identifying optimum surface roughness performance in end-milling operations” , Journal of Industrial Technology, Vol 17, pp.1-8. Ghani J.A., Choudhury I.A. and Hassan H.H., (20040 . “Application of Taguchi method in the optimization of end milling parameters” , Journal of Material Processing Technology, Vol. 145, No. 1, pp. 84-92. Oktem H., Erzurumlu T. and Kurtaran H.,( 2005).” Application of response surface methodology in the optimization of cutting conditions for surface roughness” , Journal of Material Processing Technology, Vol. 170, No. 1-2, pp. 11-16. Kopac J. and Krajnik P. ( 2007).” Robust design of flank milling parameters based on grey-Taguchi method”, Journal of Material Processing Technology, Vol. 191, No. 1-3, pp. 400-403. Datta S., Nandi G., Bandyopadhyay A. and Pal P.K.,( 2009).” Application of PCA based hybrid Taguchi method for multi-criteria optimization of submerged arc weld: A case study” , International Journal of Advanced Manufacturing Technology, Vol. 45, No. 3-4, pp. 276-286. Sanjit Moshat, Saurav Datta, Asish Bandyopadhyay and Pradip Kumar Pal,(2010), “Optimization of CNC end milling process parameters using PCA-based Taguchi method”, International Journal of Engineering, Science and Technology Vol. 2, No. 1, 2010, pp. 92-102 Read More