Nghiên cứu xác định chế độ cắt tối ưu theo hàm mục tiêu chất lượng bề mặt, đảm bảo năng suất cắt khi gia công thép SUS304 trên máy tiện CNC TT TIENG ANH

MINISTRY OF INDUSTRY AND TRADE HANOI UNIVERSITY OF INDUSTRY TRAN VIET HOI DETERMINATION OF OPTIMAL CUTTING PARAMETERS TO IMPROVE SURFACE INTEGRITY, ENSURING MACHINING PRODUCTIVITY IN TURNING SUS304 ON CNC LATHE Major: Mechanical Engineering Code: 9.52.01.03 SUMMARY OF DESERTATION IN TECHNICAL DOCTOR THESIS Hanoi, 2022 This desertation has been completed at: HANOI UNIVERSITY OF INDUSTRY Scientific supervisors: Assoc Prof Dr Pham Van Bong Prof Dr Tran Van Dich Reviewer 1: Reviewer 2: Reviewer 3: The desertation was defended at the Doctoral Evaluating Council at University level, held at Hanoi University of Industry at …., date… 2022 The desertation can be found at: - The library of Hanoi University of Industry - Vietnam National Library INTRODUCTION The importance of the topic Austenitic stainless steel has good mechanical, physical properties, high hardness, good corrosion resistance and heat resistance, so it is widely used However, austenitic steel is considered a difficult material to machine due to its high tensile strength, low thermal conductivity, high cutting force leading to high work hardening, increased tool wear rate, poor surface quality and low machining productivity After the processing procedure, surface quality is an important criterion to evaluate the quality of the workpiece, the corrosion resistance and the fatigue strength of the workpiece Residual stress and surface roughness are evaluated as two important criteria Residual stresses are generated in the processing procedure due to heat generation, mechanical deformation and changes in material organization The surface after machining with residual compressive stress will be beneficial to limit crack propagation and increase fatigue strength, whereas tensile residual stress will adversely affect the above problem In practice, the measurement, processing of measurement results and modeling of residual stresses are very complex In manufacturing, the machining process's efficiency is evaluated by improving quality, reducing costs, and increasing productivity So optimizing the machining process is the goal and the challenge of manufacturing With the development of science and technology, new approaches have been deployed to solve optimization problems for accuracy and fast processing speed in finding optimal results Research on the characteristics and machinability of stainless steel to improve surface integrity is a topic that has received the attention of many researchers before Still, the study and publication mainly evaluate machining quality or accuracy by evaluating surface roughness and microhardness In contrast, for the part after machining, the criterion of residual stress plays a vital role because this even determines the fatigue strength and cracks formed on the part's surface There are few studies and publications on analyzing the influence of the machining process on residual stress Solve the multiobjective optimization problem of important surface integrity such as surface roughness and residual stress when turning austenitic stainless steel SUS304 based on advanced algorithms The above issues are guidelines for the author to choose the topic: “Determination of optimal cutting parameters to improve surface integrity, ensuring machining productivity in turning SUS304 on CNC lathe” The aim, objective and scope of the study 2.1 The aim of the study This research aims to research, evaluate the influence and determine the relationship between the cutting parameters and some typical output factors of the turning process Moreover, the research was conducted to develop and solve optimization problems when processing stainless steel to improve the efficiency of the machining process 2.2 The objective and scope of the study - Research object: Research and evaluate the influence of input cutting parameters (V, f, t) on the machining process in turning SUS304 on CNC lathe - Research scope: Study to determine the relationship between cutting speed (V), feedrate (f), depth of cut (t) to surface roughness, microhardness, and residual stress Research methodology - Research on the theory of the cutting process as a basis for initial assessment sets and orientations for experimental research - Experimental research to get data of some indicators Apply software to calculate, process, evaluate the influence of cutting parameters, determine regression functions and solve optimization problems Scientific and practical significance - Scientific significance: Research is the basis for establishing cutting parameters when turning stainless steel on CNC machines and is the basis for optimization to improve the surface quality and machining productivity - Practical significance: Research results can be applied in production with products made from stainless steel, and at the same time as documents for research at universities Dissertation structure: The thesis is presented in four chapters: Chapter 1: Overview of stainless steel machining Chapter 2: Research to evaluate the influence of cutting parameters to surface integrity Chapter 3: Experiment to determine the effect of cutting parameters on surface integrity in turning SUS304 Chapter 4: Optimizing cutting parameters to improve surface integrity in turning SUS304 New contributions of the thesis - Develop experimental model, measure, calculate output criteria and analyze and evaluate the influence of cutting parameters on surface roughness, microhardness, surface residual stress - Apply the Response surface methodology (RSM) and experimental design Box-Behnken (BBD) to develop mathematical models of the relationship between cutting parameters with surface roughness, microhardness and residual stress - Applying Pareto optimal solution based on Bat algorithm (BA) to solve a multi-objective optimization problem to determine optimal cutting parameters to improve surface integrity CHAPTER 1: OVERVIEW OF STAINLESS STEEL MACHINING 1.1 Overview of stainless steel The addition of Stainless steels alloying elements in stainless steel affects its mechanical and Basic families Derived families physical properties Changing the Ferritic Martensitic Austentitic Duplex PH chromium content and Ferit/Austenit Martensitic adding other elements Semisuch as Nickel and Austentitic Molybdenum leads to a Austentitic change in stainless Figure 1.1 Types of Stainless steels stainless steel's mechanical, physical and anti-corrosion properties The change leads to the formation of 05 stainless steel groups (Figure 1.1), including Austenitic, Ferrite, Duplex, Precipitation hardening 1.2 Austenit stainless steel Austenitic stainless steel has a minimum Nickel and Chromium content of 7% and 16%, respectively, a Carbon content of ≤ 0.08%, and a few other elements Austenitic steels are divided into two groups: Standard group (Type 300), where Nickel is the austenitic stabilizer with a sufficient amount of Chromium and Nickel; Nitrogen can also be used to increase strength, in which SUS304 is the most popular grade stainless steel due to its excellent formability and weldability, is non-magnetic, has a much greater coefficient of thermal expansion and lower thermal conductivity than other grades Manganese group (Type 200), which adds a significant amount of Manganese, usually with higher levels than Nitrogen 1.3 Machinability of austenitic stainless steels The machinability of material is evaluated through some criteria such as size, surface finish quality, energy consumption, chip formation, wear, and tool life Austenitic steel has high tensile strength and low thermal conductivity (Table 1.1), non-transformed steel, so it cannot harden but tends to increase cold hardening It is considered a more difficult material to work with than carbon steel Table 1.1 Physical properties of the materials Tensile Thermal Elongation Grade strength conductivity (%) ( MPa) (W/mK) SUS304 515 40 16 C45 450 21 58 1.4 Research situation on stainless steel processing 1.4.1 Overseas studies Technological parameters affecting surface quality are of interest to many researchers The conducted and published studies show that the feedrate and cutting speed influence the surface roughness, as shown in the publications: M Batista researched based on SOM to evaluate chip shrinkage when turning Dry Titanium The results show that the chip shrinkage coefficient is more significant when turning structural carbon steels due to the low thermal conductivity Xinxin Zhang and et al studying high-speed stainless steel milling, show that feedrate is the most important factor affecting surface roughness Ra Lakhdar Bouzid studied the optimization of tool wear in turning SUS304 using the desirability function approach (DFA) Franko Puh studied the optimization of cutting parameters when turning with quality combinatorial properties using gray relation analysis (GRA) Residual stress in high-speed milling of aluminum alloy 6061T651 with finite element analysis by author YB Guo et al The results show that the residual stress in the infeed direction is tensile near the surface and rapidly becomes compressive at a depth of 20-25µm DW Wu in the study shows that the hardness of the material directly and significantly affects the value of residual stress caused by machining and identifies other hard steel processing methods to process ductile steel, machined surface of ductile steel without any phase transition Selecting the cutting parameters for processing procedure is one of the stages determining product quality and processing productivity In the recent trend, researchers have focused on developing new algorithms to optimise the processing procedure, ensuring many different goals Many publications have shown the effectiveness of applying new algorithms to solve optimization problems such as: Authors Poornima and Sukumar research on optimizing the cutting parameters inturning SUS40 materials using response surface methodology (RSM) and the genetic algorithm (GA) N Ahmad studied and compared optimally the surface roughness when machining SUS1045 steel using GA and particle swarm algorithm (PSO) The results obtained from the study show that the predicted values according to the RSM method are 99.3% Meanwhile, PSO obtained the lowest surface roughness when compared with Taguchi and GA methods 1.4.2 Previous studies in Vietnam In Vietnam, studies on the effect of cutting parameters on surface quality have received the attention of researchers However, the studies mainly evaluated the influence of cutting parameters on surface roughness, tool wear, cutting force such as: Research by Nguyen Tien Dung in turning SUS304 steel, evaluated the influence of (V, f, t) to surface roughness Ra The results show that the feed-rate is the most influential parameter Author Le Thi Hoai Thu, studies the machining accuracy when turning high ductile materials to evaluate the influence of the cutting parameters on the parameter Ra In the doctoral thesis of Nguyen Chi Cong, he assessed the influence of the cutting parameters on the roughness Ra, tool wear and cutting force in turning SUS304, applying analytical methods to solve the problem and find the set of tools The optimal cutting parameters when turning are V = 42m/min, f=0,08mm/rev, t=0,6mm CONCLUSION OF CHAPTER In order to improve the surface quality of the part and the efficiency of the machining process, especially when processing materials with high ductility and strength such as SUS304, it is necessary to consider the problems encountered when processing, through an overview study found that: - Austenitic stainless steel in which SUS304 is one of the difficult materials to process Machinability (technology in machining) and efficiency of the machining process are assessed through the quality of the part surface after machining, the wear mechanism and the tool life - Studies in Vietnam and other countries related to the influence of (V, f, t) on surface quality when machining stainless steel, techniques and tools applied to optimize processing procedure are very diverse However, research on the influence cutting parameters on surface layer residual stress has not been paid much attention Studies show that the determination of surface quality criteria includes: surface roughness (criteria for determining product quality), micro hardness (characteristic criteria for corrosion resistance), application Residual stress (main criterion affecting fatigue strength) in turning SUS304 steel on CNC lathe is an important and necessary research direction CHAPTER 2: RESEARCH TO EVALUATE THE INFLUENCE OF CUTTING PARAMETERS TO SURFACE INTEGRITY 2.1 Topography of surfaces 2.1.1 Surface parameter: The surface roughness value ( Ra ) is determined by equation 2.1 as follows: L Ra  y( x ) d ( x ) L o (2.1) where: Ra is the average order compared to with the mean line, L is the standard length for evaluation, y ( x ) which, is the rough profile 2.1.2 Influence of cutting parameters on surface roughness Surface roughness is influenced by many factors such as: cutting parameters, phenomena occurring in the processing procedure, tool geometry parameters, workpiece characteristics (Figure 2.1) In which the influence of cutting speed ( V ), feedrate ( f ), depth of cut ( t ) has been received the most attention Cutting tools properties Tool material Machining parmeters Process Tool shape kinematics Runout errors Cooling fluid Depth of cut Stepover Nose radius Workpiece diameter Workpiece Workpiece length hardness Workpiece properties Feedrate Tool angle Cutting speed SURFACE ROUGHNESS Accelerations Chip formation Friction Cutting Cutting force phenomena variation Figure 2.1 Cutting parameters affecting Ra 2.2 Microhardness Microhardness is one of the important parameters of surface quality and is used to evaluate the effect on the workability and service life of the workpiece Some studies have shown that surface hardening will increase the fatigue strength of the part by about 20%, increase the wear resistance by to times However, if the surface is too hard, it will reduce the fatigue strength of the part 2.3 Residual stress The compressive residual stress on the surface can increase the fatigue strength of the part by 50% and reduce it by 30% when the surface has tensile residual stress Three sources generate residual stresses during machining: heat generated during cutting, mechanical deformation, and organizational change.The main techniques for measuring residual stress include: non-destructive, semi-destructive, and destructive depending on the test conditions and the sample to be measured Among them, X-ray diffraction is one of the best methods for determining residual stress XRD data analysis methods to determine strain in materials such as Scherrer, Williamson-Hall, strain size histogram (SSP), Warren-Averbach method In which WilliamsonHall is evaluated as a straightforward analytical method based on the half-peak width of the FWHM diffraction From the X-ray diffraction pattern, the width of diffraction peaks βhkl is determined by the width due to the change in crystal size βL and the width due to microscopic deformation βε according to the formula: βhkl = βL + βε (2.7) in which βhkl is the total diffraction width, βL is the width due to crystal size and βε is the width due to strain Peak width due to crystal size k change is calculated from formula:  L  (2.8) L cos  where: - Wavelength (0.15405 nm); L - Crystal size (nm);  : diffraction angle (°/ rad); k : 0.94 Similarly, the XRD peak width due to deformation is determined by the formula: βε = 4εtanθ, with  is deformation k 4 sin   Substituting into formula (2.7) we get:  hkl  (2.10) L cos  cos  Multiplying both sides by cos  , k 4 sin   cos  we get: cos  hkl  cos  L cos  cos  Draw a line (  hkl cos ) based on (4sin  ) in which degrees strain ( ) is the slope and the intersection with the vertical axis is ( k ) From there, we can calculate the crystal size L L 11 Figure 3.4 Experimental workpiece drawing 3.3.3 Cutting tools Research using specialized chips for processing stainless steel of Sandvik brand, symbol DCMT 11 T3 04 - MF 2220 coated with CVD Ti (C, N) + Al2O3 + TiN 3.3.4 Instrumentation * Roughness meter: Figure 3.7 Roughness meter Mitutoyo * Vickers hardness meter * X-ray machine: Figure 3.8 Microhardness meter Figure 3.9 X-ray machine 3.4 Experimental determination of some characteristics of surface integrity in turning SUS304 3.4.1 Experimental sequence Step 1: Turn a thin layer with t= 0.1mm throughout the machining length to eliminate residual errors, deviations in non-parallelism between 12 the machine centerline, and the longitudinal displacement of the table Step 2: Turn all 15 surfaces on the workpiece according to the defined cutting parameters After turning, take part and clean the surface Step 3: Measure surface roughness and microhardness on 15 samples, at each surface measured at three positions 120 apart, take the average value Step 4: Take X-ray diffraction to determine residual stress 3.4.2 Regression function and influence of cutting parameters on surface roughness Table 3.5 Experimental design and surface roughness measurement results Sample 10 11 12 13 14 15 V f t Ra (m/min) 290 260 260 230 230 260 260 260 260 230 290 290 230 230 290 (mm/rev) 0,2 0,14 0,14 0,2 0,14 0,08 0,2 0,14 0,08 0,2 0,14 0,08 0,14 0,08 0,14 (mm) 0,25 0,25 0,25 0,5 0,1 0,5 0,1 0,25 0,1 0,25 0,1 0,25 0,5 0,25 0,5 (µm) 1,58 0,73 0,73 1,72 0,93 0,45 1,55 0,73 0,44 1,66 0,87 0,48 0,85 0,64 1,02 Analysis of variance (ANOVA) to determine the significance level of the input parameters and their contribution to the output The model was considered significant if the P-value < 0.05 Using Minitab 18 software, we get the results of ANOVA analysis in Table 3.6 Table 3.6 ANOVA for surface roughness Ra Source DF Model V f t V2 f2 t2 V*f 1 1 1 Seq SS 2,83234 0,08893 2,45459 0,00763 0,08038 0,17579 0,00770 0,00114 Cont (%) 99,49% 3,12% 86,22% 0,27% 2,82% 6,18% 0,27% 0,04% Adj SS 2,83234 0,00109 2,01253 0,00356 0,05363 0,19252 0,01057 0,00086 Adj MS 0,31470 0,00109 2,01253 0,00356 0,05363 0,19252 0,01057 0,00086 F value 109,05 0,38 697,39 1,23 18,58 66,71 3,66 0,30 P value 0,000 0,566 0,000 0,317 0,008 0,000 0,114 0,609 13 Source DF V*t f *t Error Lack of fit Pure error Total 1 14 Seq SS 0,01384 0,00235 0,01443 0,01443 0,00000 2,84677 Cont (%) 0,49% 0,08% 0,51% 0,51% 0,00% 100,0% Adj SS 0,01590 0,00235 0,01443 0,01443 0,00000 Adj MS 0,01590 0,00235 0,00289 0,00481 0,00000 F value 5,51 0,81 For the effect of each parameter: feedrate has the greatest influence (86,22%) ompared to the influence contribution of the cutting speed (3,12%) and the depth of cut has no significant effect (0,27%) Mutual interactions are relatively small Quadratic effect: f2 has the largest contribution (6,18%), followed by V2(2,82%) and t2(0,27%) In addition, P-value of the cutting speed (0,566) and of the depth of cut (0,317) show that V, t does not exhibit any statistically significant level to Ra Observing the Pareto F-value analysis chart in Figure 3.10, we can see that the main cause of surface roughness is the Figure 3.10 Pareto analysis chart of the influence of parameters on Ra feedrate (20% of the causes) The remaining parameters have little or no effect on the surface roughness The separate influence of each parameter on the surface roughness (Figure 3.11), it can be seen that: When V is low Figure 3.11 Main effects plot on Ra (230m/min), this is due to the possibility of stye formation in the speed region large roughness value, when V is at an average of 260 m/min without the influence of the tool stye, Ra will be the smallest When the cutting speed is up to P value 0,066 0,408 14 Surface Plot of Ra vs V t Surface Plot of Ra vs f V Hold V f 0,1 Hold Values t 0,3 290 m/min, due to the influence of temperature and chip deformation, Surface Plot of Ra vs f t the surface roughness Ra tends to increase but not much The depth of cut a b has little influence on Ra ) ) Figure 3.12 shows the simultaneous relationship of each pair of cutting parameters to Figure 3.12 3D plotsceffect of V,f,t on Ra the surface roughness ) From experimental data, using Minitab 18 software, calculating the model's coefficients, the results of the regression function representing the relationship between the cutting parameters and the surface roughness are: Ra  12,11  0,0818V 11,57 f  3,69 t  0,000149V 1,0 1,5 Ra (µm) 0,9 Ra (µm) 1,0 0,8 0,5 0,7 0,20 40 270 255 f (mm/vg ) 0,10 285 Hold 0,10 Values 0,25 V 260 40 0,15 255 V (m/ph) V (m/ph) 270 0,40 285 t (mm) 0,55 1,5 Ra (µm) 1,0 0,5 0,55 0,20 0,40 0,15 0,10 f (mm/vg ) 0,25 t (mm) 0,10  64,68 f 1,460 t  0,0079Vf  0,01002Vt  2,27 ft (3.10) R  99,49% 3.4.2 Regression function and the influence of cutting parameters on the microhardness Table 3.7 Experimental design and microhardness measurement results Sample 10 11 12 13 14 15 V f t (m/min) 290 260 260 230 230 260 260 260 260 230 290 290 230 230 290 (mm/rev) 0,2 0,14 0,14 0,2 0,14 0,08 0,2 0,14 0,08 0,2 0,14 0,08 0,14 0,08 0,14 (mm) 0,25 0,25 0,25 0,5 0,1 0,5 0,1 0,25 0,1 0,25 0,1 0,25 0,5 0,25 0,5 HV0,025 348 329,5 329,5 441 332,5 336 402 329,5 309 438 316 312,5 392,5 335 324,5 15 Table 3.8 is the result of ANOVA analysis with micro-hardness showing that in throughs, the feedrate f has the greatest influence on the microhardness after machining with a contribution of 42.92% then to the cutting speed V with 33.51% and depth of cut t with 5.82% Table 3.8 ANOVA for microhardness Source DF Model V f t V2 f2 t2 V*f V*t f*t Error Lack of fit Pure error Total 1 1 1 1 14 Seq SS 26701 9175,5 11750,8 1592,4 34,8 2498,5 132,3 916,9 277,9 322,1 676,4 676,4 27377,4 Cont (%) 97,53% 33,51% 42,92% 5,82% 0,13% 9,13% 0,48% 3,35% 1,01% 1,18% 2,47% 2,47% 0,00% 100,00% Adj SS 26701 5074,6 7501,3 713,9 51,8 1849,7 38,8 980,7 428,5 322,1 676,4 676,4 Adj MS 2966,78 5074,56 7501,29 713,85 51,84 1849,69 38,8 980,66 428,46 322,12 135,27 225,45 F value 21,93 37,51 55,45 5,28 0,38 13,67 0,29 7,25 3,17 2,38 P value 0,002 0,002 0,001 0,07 0,563 0,014 0,615 0,043 0,135 0,183 F-value analysis in Figure 3.16 shows that the ranking order of effects on HV is as follows: f has the largest effect, followed by V and f2 In Figure 3.17 it can be seen that when the cutting speed changes from 230m/min to 260m/min, the surface Figure 3.16 Pareto analysis chart of the microhardness decreases influence of parameters on HV drastically Because when the cutting speed is increased, a layer of metal in the cutting zone is melted, so the bonding force between the metal elements is reduced, the friction between the tool and the chip is reduced, the cutting force is reduced, so the hardening rate is reduced Similarly, when the feedrate increases at a high level from 0.14mm/rev to 0.2mm/rev, the microhardness increases very rapidly 16 Surface Plot of HV vs f V Surface Plot of HV vs t V Hold Values t 0,3 and reaches its highest value when the feedrate is maximum 450 HV 380 400 360 350 HV 340 Surface Plot of HV vs f t 300 320 Hold Values 0,55 0,40 V 260 0,20 40 255 V (m/ph) 0,15 270 0,10 285 f (mm/vg ) 40 255 V (m/ph) 270 0,25 285 0,10 a ) b ) 375 HV Figure 3.17 Main effects plot on HV t (mm) 350 325 300 0,55 Figure 3.18 shows the influence of each pair of cutting c Figure 3.18 3D plots effect of V,f,t parameters on the microhardness on)HV Using calculation software to find regression function and reliability, we get the following results 0,20 0,40 0,15 f (mm/vg ) 0,25 0,10 t (mm) 0,10 HV  428 1,56V  1248 f  543t  0,00462V  6340 f  88 t  8,47Vf 1, 645Vt  842 ft (3.11) R  97,53% 3.4.3 Regression function and the influence of cutting parameters on residual stress Residual stress cannot be determined directly from measuring devices, but must be processed and intermediately calculated according to the following process Step 1: Using MDI Jade 6.5 software to read X-ray diffraction data (*.raw format), the diffraction peaks were normalized by the PseudoVoigt function with default values to read the half-peak width results  hkl (FWHM) at the diffraction peaks Step 2: Calculate diffraction peak width Step 3: Calculated  from the first-order interpolation function passing through points of equation 2.14, the results are presented in Table 3.9 Step 4: Calculate residual stress from formula 2.13 as a function of  with constants elastic modulus E=200 GPa and Poisson coefficient   0,293 17 Table 3.9 Determination of residual stress calculation data Sample 10 11 12 13 14 15 2θ (o) [hkl] 43,531 50,766 74,490 43,796 51,007 74,773 43,796 51,007 74,773 43,823 51,022 74,743 43,763 50,975 74,747 43,866 51,067 74,814 43,685 50,894 74,564 43,796 51,007 74,773 43,767 50,981 74,759 38,033 44,266 74,767 43,797 51,010 74,632 43,792 51,012 74,749 43,802 50,997 74,733 43,840 51,023 74,741 43,939 51,138 74,859 111 200 220 111 200 220 111 200 220 111 200 220 111 200 220 111 200 220 111 200 220 111 200 220 111 200 220 111 200 220 111 200 220 111 200 220 111 200 220 111 200 220 111 200 220 FWHM (o ) 0,380 0,756 0,884 0,401 0,803 0,796 0,401 0,803 0,796 0,409 0,769 0,976 0,405 0,842 0,829 0,406 0,770 0,817 0,426 0,830 0,998 0,401 0,803 0,796 0,398 0,796 0,801 0,404 0,479 0,904 0,443 0,994 0,968 0,445 0,703 0,866 0,385 0,804 0,945 0,410 0,873 0,854 0,384 0,732 0,794 βcosθ (rad) 0,00594 0,01182 0,0122 0,00629 0,01255 0,01095 0,00629 0,01255 0,01095 0,00642 0,01201 0,01347 0,00636 0,01317 0,01141 0,00637 0,01202 0,01124 0,00671 0,01299 0,01379 0,00629 0,01255 0,01095 0,00624 0,01244 0,01102 0,00646 0,00757 0,01246 0,00699 0,01558 0,01336 0,00702 0,01096 0,01193 0,00602 0,01257 0,01303 0,00644 0,01366 0,01176 0,006 0,01142 0,01092 4sinθ 1,483 1,715 2,421 1,492 1,722 2,429 1,492 1,722 2,429 1,493 1,723 2,428 1,491 1,721 2,428 1,494 1,724 2,430 1,488 1,719 2,423 1,492 1,722 2,429 1,491 1,721 2,428 1,303 1,507 2,429 1,492 1,722 2,425 1,492 1,722 2,428 1,492 1,722 2,428 1,493 1,723 2,428 1,496 1,726 2,431 ε % 0,533 0,327 0,327 0,625 0,354 0,372 0,607 0,327 0,343 0,534 0,450 0,433 0,589 0,372 0,386 18 Calculation results to determine residual stress of the samples studied using the Williamson-Hall method are given in Table 3.10 Table 3.10 Experimental design and calculation results of residual stress V Mẫu (m/min) 290 260 260 230 230 260 260 260 260 230 290 290 230 230 290 10 11 12 13 14 15 f t  (mm/rev) 0,2 0,14 0,14 0,2 0,14 0,08 0,2 0,14 0,08 0,2 0,14 0,08 0,14 0,08 0,14 (mm) 0,25 0,25 0,25 0,5 0,1 0,5 0,1 0,25 0,1 0,25 0,1 0,25 0,5 0,25 0,5 MPa 201,6 125,9 125,9 240,8 136,3 143,1 233,3 125,9 131,7 204,5 172,5 166,7 226,5 143,2 148,3 ANOVA on the influence of cutting parameters on surface residual stress in table 3.11, shows that the influence ratio of cutting speed, feedrate and depth of cut are 3.98%, 41.64 % and 4.90%, respectively Table 3.11 ANOVA for residual stress Source DF Model V f t V2 f2 t2 V*f V*t f*t Error Lack of fit Pure error Total 1 1 1 1 14 Seq SS 22657 984,6 10297,3 1211 1620 3376,4 1709,7 97 2731,4 629,4 2072,7 2072,7 24729,7 Cont (%) 91,62% 3,98% 41,64% 4,90% 6,55% 13,65% 6,91% 0,39% 11,04% 2,55% 8,38% 8,38% 0,00% 100,0% Adj SS 22657 196,5 4976,1 70,1 1638,4 2598,1 1162,7 78,7 3241,7 629,4 2072,7 2072,7 Adj MS 2517,45 196,47 4976,09 70,05 1638,43 2598,06 1162,71 78,69 3241,69 629,45 414,53 690,89 F value 6,07 0,47 12 0,17 3,95 6,27 2,8 0,19 7,82 1,52 P value 0,031 0,522 0,018 0,698 0,104 0,054 0,155 0,681 0,038 0,273 Pareto diagram (Figure 3.22) shows the order of effects of factors 19 on residual stress, factors and interactions including (f), (V*t), (f2) represent the level of significance for the output variable, where the feedrate is the parameter that has the greatest influence on the Figure 3.22 Pareto analysis chart of the residual stress influence of parameters on  In Figure 3.23 it can be seen that the residual stress is greatest when the feedrate is at its highest, while the residual stress value decreases when the feedrade is reduced, the depth of cut and the Surface Plot of σ vs V t cutting speed are average Surface Plot of σ vs f V The influence of (V, f) Figure 3.23 Main effects plot on  on σ is shown in Figure 3.24a When V is medium, f is small, σ is low Figure Surface Plot of σ vs t f 3.24b shows a small residual stress when V and t are medium Figure a b 3.24c shows the effect of ) ) (f, t) on σ, showing that at medium t and small f, the residual surface stress is small c effect of Figure 3.24 3D plots A polynomial equations of V,f,t on) second-order was built to show the relationship between residual stress and cutting parameters described by equation (3.12): Hold Values t 0,3 200 210 σ (MPa) 180 σ (MPa) 180 160 150 140 120 0,20 40 255 V (m/ph) 40 0,15 270 f (mm/vg ) 0,10 285 255 V (m/ph) 277 270 0,4 0,40 ,40 285 0,55 210 180 σ (MPa) 150 120 0,55 0,,10 f (mm/vg ) 0,40 0,15 0,25 0,20 0,10 t (mm) 0,10 0,25Hold Values t (mVm)260 Hold Value f 0,14 20   1559 11,99V  665 f 1066 t  0,0260V  7514 f  484 t  2,40Vf  4,52Vt 1177 ft (3.12) R  91,62% CONCLUSION CHAPTER 3: From the orientation in Chapter 2, Chapter uses experiments according to the Box-Behnken design with 15 experiments Using advanced measuring instruments and methods to consider and evaluate the influence of (V, f, t) on surface roughness, microhardness, and residual stress in turning SUS304, The results show that: - Cutting parameters affect surface roughness, microhardness, residual stress In which the feed rate has the most influence on surface roughness, microhardness, residual stress - The relationship between cutting parameters and output criteria is determined by a polynomial function of second-order (equations 3.10, 3.11, 3.12) with a confidence level of 90% or more The above results help technologists choose reasonable cutting parameters and form the basis for the construction and solution of the optimal problem CHAPTER 4: OPTIMIZING CUTTING PARAMETERS TO IMPROVE SURFACE INTEGRITY IN TURNING SUS304 4.1 Develop optimization model 4.1.1 Overview of Optimization Tools machining process and Techniques optimization Conventional Non-Conventional Techniques Techniques The techniques for Meta-Heuristic Design of Mathematical solving optimization Search experiment Iterative Search problems can be Swarm Dynamic Linear Non-Linear Evolutionary Programming Programming Programming Intelligence algorithm (EA) -based -based -based (SI) divided into two Algorithms Algorithms Algorithms categories: traditional Genetic Simulated Tabu Algorithm Annealing Taguchi Factorial Response Surface Seach (GA) (SA) and advanced methods, Method- Design- Methodology based based (RSM) as shown in Figure 4.1 Ant colony Particle Firefly Cuckoo Bat optimization swarm algorithms search algorithm Many overview studies (ACO) (PSO) (FA) (CS) (BA) Figure 4.1 Optimization Tools and Techniques 21 on algorithms applied in different fields, Bat algorithm (BA) is evaluated as the newer algorithm, more straightforward mathematical structure and more effective (more robust) than PSO 4.1.2 Bat Algorithm Each virtual Bat in the dimensional search space is defined by position and speed in the BA algorithm The frequency, velocity, and position parameters after the iterations are updated as follows: fi  f  ( f max  f ) (4.1) vik  vik 1  (w ik 1  w * ) fi (4.2) w ik  w ik 1  vik (4.3) Where:  0,1 is a uniformly distributed random vector, w* is the best global position (solution) in the Bat population,  fmin , fmax  the frequency interval depends on the domain size of the field of consideration, initially, Each Bat will randomly pick a frequency in the frequency range under consideration 4.1.3 Pareto multi-objective optimization solution Vilfredo Pareto proposed the Pareto optimal solution in the 19th century If no other solution can improve at least one objective without weakening any other, then Pareto superiority and optimality are appropriately used for the multi-objective problem 4.2 Application of Bat algorithm to optimize single-objectives The mathematical model of the minimum roughness singleobjective optimization problem is: Objective function Ra  f1 ( x)  f1 ( x )  12,11  0,0818V 11,57 f  3,69 t  0.000149V  64,68 f 1,460 t  0,0079Vf  0,01002Vt  2, 27 ft The parameters (Table 4.1) are encoded by the position of a virtual Bat and the Bat algorithm flowchart (Figure 4.2) 22 Start Table 4.1 Parameters of BA Parameters Value 0,8 Loundness, A Pulse rate, r 0,8 Minimize frequency, f Maximize frequency, f max Number of iteration, t 300 BA population, n 100 Define the input data: Q, k, kmax, vi, fi, Ai, ri Random Tạo ngẫu nhiên generate trọng số wi wi(Vi, fi, ti) Build objective function Ra(wi) Calculate and find Ra(best) with wbest Generate new solutions and update by eq (4.1) to (4.3) No Rand(0,1)>ri Yes Update parameters by eq.(4.4) The result of the objective function Ra achieved through the algorithm has a value of 0.427 corresponding to the optimal parameter value: V = 262.242 m/min; f = 0.08 mm/rev; t = 0.302 mm Calculate and find Ra(new-best) with wnew-best No Rand(0,1)>ri and Ra(new-best) < Ra(best) Đ Update Rabest=Ranew-best, wbest=wnew-best k=k+1 Increase ri & decrease Ai by eq (4.5) Yes k < kmax No End Figure 4.2 Flowchart of BA 4.3 Multi-objective optimization to improve surface integrity The simultaneous optimization problem of two criteria is minimum surface roughness and minimum residual stress Objective function F ( x)   f1 ( x), f2 ( x)   f1 ( x )  12,11  0,0818V 11,57 f  3,69 t  0.000149V    2  64,68 f 1,460 t  0,0079Vf  0,01002Vt  2,27 ft     f ( x )  1559 11,99V  665 f 1066 t  0,0260V   7514 f  484 t  2,40Vf  4,52Vt 1177 ft    Table 4.4 Parameters of MOBA Parameters Loundness, A Pulse rate, r Minimize frequency, f Value 0,8 0,8 Number points of Pareto, N 1000 Parameters Number of iteration, t BA population, n Maximize frequency, f max Value 1000 100 23 The problem results for 10 points with the best value of the objective function Table 4.4 Optimal solutions archived by MOBA No 10 V (m/min) 252,779 261,006 258,689 257,074 254,277 259,661 256,727 257,665 255,833 262,242 f (mm/rev) 0,1 0,08 0,088 0,092 0,098 0,085 0,093 0,09 0,095 0,08 t (mm) 0,201 0,258 0,235 0,226 0,21 0,243 0,223 0,229 0,219 0,302 Ra (µm) 0,516 0,43 0,453 0,468 0,5 0,443 0,472 0,461 0,481 0,427  ( MPa) 117,987 124,112 120,366 119,189 118,099 121,630 118,929 119,616 118,571 126,941 Select a set of parameters to experiment with the same conditions as when conducting the initial experiment to verify the optimal parameters results obtained from the optimal problem The obtained results show the error between the actual confirmatory test with a minimal computation of approximately 2% Table 4.5 Results of confirmation test Experimental result Predicted value Experimental value Ra (µm) 0,461 0,472 119,616 121,658  ( MPa) Error 2,40% 1,70% As can be seen, the maximum machining productivity achieved is Q = 6.335,8mm3/min, the surface roughness Ra = 0.427 μm, residual stress is  = 126,941MPa corresponding to the optimal cutting parameters (V, f, t) are (262.242; 0.08; 0.302), respectively CONCLUSION CHAPTER 4: The research results determine a mathematical function representing the relationship between the cutting parameters and some output factors of the machining process Applying the Bat algorithm (BA) to solve the problem of optimizing specific results is as follows: - Solving the single-objective optimization problem of surface roughness to find the optimal set of technological parameters is V = 24 262.242 m/min, f = 0.08 mm/rev, t = 0.302 mm and surface roughness value the smallest is Ra = 0.427 μm - Solving the multi-objective optimization problem of surface roughness and residual stress, applying Pareto optimal solution to determine the optimal set of solutions in the Pareto domain The result is an optimal set of results The values of the objective function obtained are: Roughness Ra  0,427 μm, Residual stress   117,987 MPa, machining productivity: Q = 6335.8 mm3/min CONCLUSIONS AND RECOMMENDATIONS Conclusion Based on theoretical research combined with experiments to clarify the influence of cutting parameters on some output indicators of turning SUS304 Regressions model has been identified, showing the relationship between cutting parameters with surface roughness, residual stress, microhardness (Eq 3.10, 3.11, 3.12) with high reliability of 99.49%, 91.62%, and 97.53%, respectively The Bat algorithm (BA) is applied to solve the multi-objective optimization problem to find the Pareto multi-objective optimal boundary region The results found an optimal set of results The obtained objective function values are: Ra  0,427 μm;   117,987 MPa, Machining productiviy with optimum surface integrity value: Q  6.335,8 mm3/min Recommendations Stainless steel is high strength and toughness when processing is always tricky The research to optimize the machining process involves many issues, from making workpieces, designing and manufacturing cutting tools, selecting processing equipment, and appropriate cutting parameters Therefore, it is necessary to have the cooperation of scientists for further research Further research To develop and complete the research on machining materials with high conductivity in Vietnam with difficult to machine The following research should expand the influence of cutting tool geometry parameters, smoothness/cooling on machinability, wear and tool life, and surface integrity when machining austenitic stainless steel SCIENTIFIC PUBLICATION Tran Viet Hoi, Pham Van Bong, Tran Van Dich: “Application of the bat algorithm (BA) to determine the optimal surface roughness in machining SUS304 on CNC lathes Proceedings of the 5th National Conference on Science and Technology on Mechanical Engineering VCME 2018, 2018 Viet-Hoi Tran, Van-Bong Pham, Van-Dich Tran: “Study of the Mechanisms of Chip Formation in Turning of 304 Austenitic Stainless Steel” Proceedings of the 2nd Annual International Conference on Material, Machines and Methods for Sustainable Development (MMMS2020), Part of the Lecture Notes in Mechanical Engineering book series (LNME), pp 138-146, 2021 (Scopus) https://doi.org/10.1007/978-3-030-69610-8_18 Viet-Hoi Tran, Van-Bong Pham, Van-Dich Tran: “Modeling of the Effect of Cutting Parameters on Surface Residual Stress When Turning of 304 Austenitic Stainless Steel” Proceedings of the 2nd Annual International Conference on Material, Machines and Methods for Sustainable Development (MMMS2020), Part of the Lecture Notes in Mechanical Engineering book series (LNME), pp 177-183, 2021 (Scopus) https://doi.org/10.1007/978-3-030-69610-8_23 Bong Pham Van, Hoi Tran Viet: “Application of Bat algorithm for Improvement of Surface Integrity in Turning of SUS304 Austenitic Stainless Steel” Journal of the Korean Society for Precision Engineering, Vol 38, No 4, pp.237-244, 2021 (Scopus, Q4) https://doi.org/10.7736/JKSPE.021.003 Pham Van Bong, Tran Viet Hoi, Tran Van Dich: “Modeling of the influence of cutting parameters on the microhardness in turning SUS304" Science and Technology Journal of Hanoi University of Industry, Vol 57 – No (6/2021), pp.75-79, 2021 ... the influence of input cutting parameters (V, f, t) on the machining process in turning SUS304 on CNC lathe - Research scope: Study to determine the relationship between cutting speed (V), feedrate... influence of cutting parameters to surface integrity Chapter 3: Experiment to determine the effect of cutting parameters on surface integrity in turning SUS304 Chapter 4: Optimizing cutting parameters... cutting parameters on surface quality have received the attention of researchers However, the studies mainly evaluated the influence of cutting parameters on surface roughness, tool wear, cutting