Introduction
Motivation of the study
The advancement of engineering materials has led to the emergence of difficult-to-cut materials designed to meet specific industry needs However, these materials pose significant challenges for machining, prompting increased focus from scientists and engineers worldwide This study aims to enhance the understanding of machining techniques for difficult-to-cut materials, supporting ongoing research and development in this critical area.
Dry machining offers both economic and ecological benefits, but it presents challenges when working with difficult-to-cut materials These challenges can result in poor surface quality and increased cutting temperatures, ultimately leading to a shorter tool life.
Using cutting fluids in flood form can effectively lower temperatures in the cutting zone and improve the longevity of cutting tools However, excessive use of these fluids can lead to increased costs and pose environmental challenges, as well as create issues for machine operators.
A new cooling technique, known as Minimum Quantity Lubrication (MQL), has been developed to replace the environmentally harmful flood cooling method, gaining traction in the manufacturing industry MQL utilizes minimal amounts of cutting fluids (50 - 500 mL/h) delivered as a mist under high air pressure (2 to 6 bar) directly to the cutting zone, making it an effective and eco-friendly solution The fluids used in MQL are required to be biodegradable, reinforcing its status as an economical and sustainable lubrication method Research has shown significant advancements in MQL technology, with studies identifying optimal MQL conditions and cutting parameters for improved surface roughness when processing AISI H13 using various lubricants Additionally, findings indicate that MQL can reduce cutting force and extend tool life significantly when using Al2O3/TiAlN-coated tools.
16-19] have also studied MQL to highlight its superiority when compared to both dry and flooded lubrication conditions
To further enhance the effective thermal conductivity of cutting fluid, a technological breakthrough was made by the use of miniscule solids particles (less than
In 1995, Choi introduced the concept of nanofluid or nanolubrication by incorporating nanoparticles (100 nm) into cutting fluids, paving the way for extensive research on the use of nanoparticles in minimum quantity lubrication machining This innovative approach enhances the thermal conductivity, density, and viscosity of cutting fluids, significantly improving heat transfer in the machining process These findings align with previous studies, confirming the effectiveness of nanoparticle-based cutting fluids in overcoming the limitations of traditional lubrication methods.
The incorporation of nanoparticles into cutting fluids significantly enhances their properties, leading to improved machining performance by reducing cutting force, enhancing surface quality, and extending tool life Research by Sharma et al demonstrated that Al2O3 nanoparticle-based cutting fluids under minimum quantity lubrication (MQL) reduced cutting force and tool wear while decreasing average surface roughness by approximately 38% compared to traditional lubricants Similar findings were reported by Sayuti et al., Hadi et al., and Vasu et al Furthermore, Sharma et al explored the innovative use of hybrid nanoparticle-enriched cutting fluids, which showed marked improvements in performance and tribological characteristics, including reductions in tool flank wear, nodal temperature, surface roughness, and cutting force by 12.29%, 5.79%, 20.28%, and 9.94%, respectively Additionally, Wang et al investigated the heat transfer capabilities of MQL with various nanoparticles in vegetable oil, confirming that nanofluids enhance lubrication properties.
3 morphology, showed excellent lubrication performances, and the expressed feature of the six nanofluids is presented in the following order: ZrO2 < CNTs < ND < MoS2 < SiO2 <
The incorporation of Al2O3 nanoparticles in lubrication has shown significant advantages, yet research on Al2O3 nano-lubrication remains limited Further investigation into the use of nanofluids in Minimum Quantity Lubrication (MQL) is essential to highlight the benefits of nano-lubrication in machining processes The adoption of nanofluids-assisted MQL is gaining traction among manufacturing industries due to its superior properties.
Reducing energy consumption in the modern manufacturing industry is crucial due to the depletion of global energy resources and the need to protect the environment In the context of machining tools, energy use can be addressed through two main strategies: enhancing machines and optimizing process parameters While machine enhancements may not always be practical, optimizing process parameters is a more feasible approach that leads to cost savings and requires less effort Consequently, many researchers are focusing on energy savings through the optimization of technological parameters in manufacturing processes.
Research focused on optimizing process parameters to conserve energy in metal processing is gaining significant traction in manufacturing industries Commonly utilized models for establishing the relationship between input parameters and technological responses include Response Surface Methodology (RSM), Artificial Neural Networks (ANN), the Kriging model, and Radial Basis Function (RBF) RSM stands out as the preferred choice among engineers due to its flexibility and reliability Additionally, the Kriging model effectively captures highly nonlinear relationships between process and response parameters This study employs both RSM and the Kriging model to accurately depict the interactions between these parameters.
In engineering, various methods are utilized to address optimization problems, particularly in metal cutting engineering Common techniques include Taguchi methods and grey relational analysis, which are widely applied to enhance performance and efficiency in manufacturing processes.
Traditional optimization techniques often fail to achieve a "real optimal" solution as they primarily focus on the design variables A popular method in simultaneous multi-response optimization is the desirability function approach, utilized by Mia and colleagues for multi-objective optimization across various cutting conditions, including dry, wet, cryogenic, and MQL cooling Recently, researchers have increasingly turned to revolutionary optimization algorithms, such as Adaptive Multi-Objective Genetic Algorithm (AMGA), Genetic Algorithm (GA), Particle Swarm Optimization (PSO), and Non-dominated Sorting Genetic Algorithm II (NSGA-II), for their efficiency, flexibility, and effectiveness in addressing multi-objective optimization challenges.
Recent advancements in evolutionary and bio-inspired multi-optimization technologies, such as NSGA-II and particle swarm optimization (PSO), have emerged as effective tools for optimization Unlike traditional gradient-based algorithms, which often get stuck in local optima, genetic algorithms (GA) and PSO are designed to identify global optimum points As a result, many researchers favor the use of evolutionary optimization algorithms for their robustness and efficiency in solving complex problems.
Machining difficult-to-cut materials presents significant challenges, particularly in enhancing their machinability The integration of nanofluids and multi-objective optimization is crucial for selecting optimal cutting conditions that minimize energy consumption while maximizing productivity and quality This study focuses on specific machining cases and materials, employing techniques like Minimum Quantity Lubrication (MQL) and nanofluid MQL to improve machinability By optimizing machining parameters, we can enhance key technical metrics such as surface roughness, cutting force, and material removal rate (MRR), ultimately promoting better tool durability, improved product quality, cost reduction, and a shift towards sustainable manufacturing practices.
Objective of the study
This research aims to enhance the machinability of difficult-to-cut materials by examining various cutting conditions, including dry machining, minimum quantity lubrication (MQL), and nanofluids, alongside different cutting parameters The study provides insights into optimizing input and output parameters, yielding significant improvements without the need for optimization algorithms Additionally, it contributes to the advancement of a sustainable and eco-friendly manufacturing industry The key findings of the study underscore the effectiveness of these approaches in improving machining processes.
1 Understand deeply about machining difficult-to-cut materials
2 Investigate the machining parameters and conditions under dry, MQL, and nanofluid MQL conditions
3 Optimization of cutting conditions and cutting parameters in machining difficult-to-cut materials
4 Applying innovative optimization algorithms to optimize muti-objectives of response parameters when machining difficult-to-cut materials
5 Assisting engineers, machine operators to choose suitable cutting conditions for each case when machining difficult-to-cut materials
6 Help scientists and engineers have a better overview of specific cutting energy as well as power consumption when machining difficult-to-cut materials
7 This study is a significant contribution to the environmentally-conscious machining.
Scope of the study
This study concentrates on two challenging materials to cut: hardened steels and super-alloys, which are part of a broader category that includes titanium alloys, metal matrix composites (MMCs), and ceramics Due to constraints in budget and time, the focus is narrowed to these specific materials.
This research focused on machining difficult-to-cut materials using three cutting conditions: dry, minimum quantity lubrication (MQL), and nanofluids The study intentionally excluded the flood cutting (wet cutting) method due to its associated disadvantages discussed in the first section.
This study focuses on milling, a prevalent machining process in the manufacturing industry It utilizes end milling cutters with diameters of 10 mm and 16 mm from renowned cutting tool manufacturers CMTEC (Taiwan) and Sandvik (Sweden) Additionally, the research incorporates slot milling, a commonly employed technique in milling operations.
This study focuses on the machining input parameters that significantly influence machining performance, specifically examining the effects of cutting fluids, workpieces, machining conditions, and cutting tools, while keeping other process parameters constant The key performance indicators evaluated include surface quality, cutting force, cutting temperature, specific cutting energy, power consumption, material removal rate, and tool wear, all of which are critical for assessing the output quality of the milling method utilized in this research.
Organization of the dissertation
The dissertation consists of 5 chapters The composition of the chapters in this dissertation is presented in Figure 1.1
In chapter 1 is an introduction to expressing motivation, objective, scope of research, and structure of the dissertation
In chapter 2, a brief background of the study is discussed The content covered in this chapter is machining difficult-to-cut materials, cooling and cutting fluids methods, and literature review
Chapter 3 exhibits the research method that was adopted in the dissertation The first section describes the design of experiment (DOE) utilized in this study The second part deals with the approximation models to render the relationship between the input and output variables of the dissertation The last section presents the optimization techniques used to find the optimal solution in each particular study
Chapter 4 presents four studies on difficult-to-cut materials The first study with the content is multi-objective optimization for surface roughness and cutting force under dry cutting conditions when milling on AISI H13 steel with 10 diameter cutting tool of the CMTEC company In the second study, also AISI H13 material is machined under three different cutting conditions: dry, MQL, nanofluids, various cutting speed, feed per tooth, and depth of cut By optimizing three process responses (surface roughness, cutting temperature, and cutting force), this study determines the solution for optimal cutting parameters and cooling conditions In the third study, multi-objective optimization for AISI H13 in terms of productivity, quality, and power consumption under nanofluid condition In the final research, the Inconel-800 superalloy is employed To increase the machinability of the difficult material, nanofluids are adopted A revolutionary optimization algorithm is employed to gain the best global solution for surface roughness, energy, and material removal rate
In chapter 5, the last chapter of the dissertation, summarizing conclusions drawn from researches and future work is also mentioned
Background
Machining difficult-to-cut materials
Materials science is a critical area of research, focusing on the development of innovative materials that meet the demands of various industries These materials, known as difficult-to-cut materials, exhibit exceptional mechanical and metallurgical properties, including superior strength-to-weight ratios, stiffness, toughness, heat capacity, thermal conductivity, hot hardness, corrosion and oxidation resistance, and fatigue resistance Difficult-to-cut materials encompass hardened steels, titanium alloys, super-alloys, metal matrix composites (MMCs), and ceramics This study specifically examines two of these challenging materials: hardened steel and super-alloys.
Hardened steels, a type of ferrous alloy, consist of various alloying elements and are recognized for their unique mechanical and metallurgical properties This distinction is primarily due to the heat treatment process applied to these steels, which enhances their strength and durability.
The carbon content in steel significantly influences its tensile and yield strength, as illustrated in Figure 2.1 An increase in carbon content corresponds to enhanced tensile and yield strength Additionally, incorporating other alloying elements, such as manganese, further improves the properties of steel, enhancing its ductility and performance in metal forming applications Hardened steel primarily consists of iron, with carbon levels ranging from 0.15% to 0.2%, along with trace amounts of other alloying elements.
Hardened steel is formed based on several heat treatments such as thermal treatment, cryogenic treatment, and surface hardening In that thermal treatment and
Cryogenic treatment is commonly applied to medium to high alloyed steel, while surface hardening is typically utilized for low alloy steels This process enhances the hardness of the surface without compromising the core's softness, making it an effective method for improving material durability and performance.
The heat treatment of steel results in a hardened state that exhibits distinct properties compared to conventional steel Key characteristics of hardened steel include high hardness, reduced ductility leading to brittleness, a high hardness-to-E-modulus ratio, and increased sensitivity to corrosion These unique properties are particularly significant for researchers exploring advancements in hard machining techniques.
Superalloys play a crucial role in modern industry due to their ability to retain mechanical properties under prolonged exposure to high temperatures Developed for specialized applications, these materials are essential in sectors like turbo-superchargers, aircraft turbine engines, gas turbines, rocket engines, and petroleum refineries Superalloys are categorized into three main types, highlighting their diverse applications and importance in advanced engineering.
10 concentration of metal in the composition, including Iron-based superalloys, Nickel- based superalloys, and Cobalt-based superalloys (Figure 2.2)
Superalloys are distinguished by their remarkable properties, including high fatigue resistance, high-temperature stability, and excellent creep resistance, while maintaining their chemical and mechanical integrity even at elevated temperatures Despite these advantages, they also present challenges such as poor thermal conductivity, significant hot hardness and strength, the tendency to form build-up edges (BUE), and adverse chemical reactions with cutting tools Consequently, machining superalloys can be quite difficult, necessitating the use of specialized support tools to improve their machinability.
The evolution of super-alloys is driven by the specific needs of various industries, with aerospace being the primary sector, accounting for 70% of their applications These materials are crucial for manufacturing fixtures and rotating components, particularly in gas turbines and jet engines, where safety and reliability are essential Super-alloys possess exceptional properties that make them indispensable in the aerospace industry, while they also find applications in medical, chemical, and structural fields, each contributing 10% to their overall use.
This study investigates Inconel 800, a nickel-based super alloy renowned for its high-grade properties It is widely utilized across various sectors, including aerospace, industrial furnaces, the automotive industry, and marine applications, due to its exceptional performance and durability.
Figure 2 3 Applications of superalloys materials in industry
2.1.2 Operations for machining difficult-to-cut materials
Hard machining is a process that involves machining materials with a hardness ranging from 40 to 70 HRC using geometrically defined cutter blades Traditionally, manufacturing involves multiple steps, including forming, annealing, rough turning, heat treatment, and grinding However, these extensive machining sequences lead to increased setup time and costs, particularly due to the grinding process, which also relies heavily on emulsifiable cutting fluids that pose environmental risks upon disposal As environmental concerns become increasingly important in the manufacturing industry, hard turning emerges as a viable alternative to grinding, offering a more efficient and less harmful solution It is essential, however, to ensure system rigidity during the hard machining process to achieve optimal results.
Figure 2 4 (a) Machining sequences of conventional manufacturing processes and (b) hard turning [60]
For effective hard turning, it is essential to ensure the system used provides adequate rigidity, while also selecting the right materials for cutting tools The materials chosen must exhibit key characteristics such as abrasion resistance, exceptional toughness, heat resistance, and stability of chemical and physical properties during high-temperature cutting processes.
Figure 2 5 Evaluation of the cost between (a) grinding and (b) hard turning [55] 2.1.2.2 Hard milling
Hard milling is a crucial hard machining technique widely used in the manufacturing industry, particularly in mold and die production, where complex surfaces pose challenges for finishing operations Traditionally, electrical-discharge machining (EDM) was employed for these tasks, but its limitations have led to the adoption of hard milling, which utilizes toolpath processing techniques generated by CAD/CAM software to overcome these challenges Common milling operations include shoulder milling, face milling, profile milling, groove or slot milling, and chamfer milling, with slot milling being the most prevalent A hard milling system comprises essential components such as the machine, cutting tools, and CAD/CAM systems, with cutting tools being the most critical element The cutting tools used in hard milling are categorized into solid carbide endmills, indexable carbide inserts, and ceramic indexable inserts.
Figure 2 6 Hard milling (Slot milling) [63]
Cooling and lubrication method
Green manufacturing is a global trend driven by climate change and rising waste levels It focuses on minimizing resource use, reducing pollution and waste, enhancing efficiency, and lowering emissions The primary goal of modern manufacturing is to limit environmental impact.
Cutting temperature poses a major challenge in machining, particularly with difficult-to-cut materials Effective cooling techniques are essential for managing these temperatures, which in turn improves machinability, enhances productivity, increases tool lifespan, and helps lower production costs These cooling methods can be categorized into conventional techniques and green (or clean) manufacturing practices.
Figure 2 7 Cooling/Lubrication techniques for metal cutting
Conventional cooling and lubrication methods, such as flood cooling, high-pressure coolant, mist cooling, and internal tool cooling, are often overlooked due to their negative environmental impact To address this issue, green manufacturing techniques like dry machining, also known as machining without cooling, and minimum quantity lubrication (MQL) have emerged as viable alternatives This study focuses on the effectiveness of dry machining and MQL as innovative cooling solutions.
Dry machining is an environmentally friendly processing method that eliminates the need for cutting fluids or lubrication, significantly reducing production costs This approach is ideal for the green industry, offering benefits such as zero pollution, easier handling and recycling, and cleaner chips By removing the costs associated with lubrication systems and cutting fluids, which currently account for approximately 7-17% of a finished product's cost, dry machining presents a more economical alternative Additionally, this method mitigates potential issues related to the use of cutting fluids, enhancing overall operational efficiency.
Figure 2 8 Advantages of dry machining [65]
The dry cutting method, which emerged in the 1990s, faces significant challenges due to heat generation during machining, particularly affecting the lifespan of cutting tools This issue is exacerbated when working with difficult-to-cut materials, leading to poor machinability in dry cutting processes Consequently, the strategic use of cutting fluids or lubrication, or the reduction of their usage, has become essential for advancing modern industry practices.
Minimum Quantity Lubrication (MQL) is an emerging cooling lubrication technique in the manufacturing industry, utilizing a minimal amount of cutting fluid (approximately 50 - 150 mL/hour) Often referred to as near-dry machining, MQL involves mixing small quantities of cutting fluids with high-pressure air to create a mist that is directed to the cutting zone This innovative method is recognized for being eco-friendly and cost-effective, making it a favorable choice in various manufacturing applications.
Dry machining may not be suitable or may yield low efficiency in certain applications In such instances, Minimum Quantity Lubrication (MQL) is frequently utilized, particularly for materials with unique mechanical properties or challenging machinability This approach also considers the longevity of the cutting tool and the quality of the finished surface.
Enhancing cutting fluids used in Minimum Quantity Lubrication (MQL) improves temperature control in the cutting zone, which significantly reduces tool wear Additionally, MQL with lubrication minimizes friction in the cutting area, leading to decreased cutting forces and energy consumption Utilizing high-pressure cutting fluids facilitates quicker and smoother chip removal, preventing chips from damaging the surface finish.
Figure 2 9 The cost in metal machining [66]
MQL delivery systems are categorized into two types: external and internal applications External applications consist of an ejector nozzle and a conventional nozzle, as illustrated in Figure 2.11 In contrast, internal applications feature two configurations: single channel and dual channel, with their principles depicted in Figure 2.12.
Figure 2 10 Classification of MQL systems
Figure 2 11 The principles of ejector nozzle (1) and conventional nozzle (2) of external application [70]
Figure 2 12 The principles of single channel and dual channel in internal application
MQL, or Minimum Quantity Lubrication, is an advanced cooling technique that utilizes minimal amounts of cutting fluid to enhance performance and maintain freshness The effectiveness of the MQL method heavily relies on the properties of the cutting fluid employed, making it a crucial element in achieving optimal results in machining processes.
When implementing the Minimum Quantity Lubrication (MQL) technique, it is essential to consider factors such as the type, wetting, and viscosity of the cutting fluid, as these significantly influence lubrication and cooling performance Additionally, key parameters including nozzle position, fluid flow, and air pressure must be properly adjusted for optimal system performance In this study, the author utilizes CT232, a commercially available lubricating oil in Taiwan, while maintaining constant settings for nozzle position, fluid flow, and air pressure, which will be elaborated upon in the subsequent section.
Nanofluid MQL refers to a minimum quantity lubrication method that employs a cutting fluid containing tiny particles, typically less than 100 nm, in a homogeneous suspension These nanoparticles, categorized into seven types, significantly influence the lubrication effectiveness when mixed with the cutting fluid Key characteristics such as particle size, volume concentration, shape, and temperature play crucial roles in determining the viscosity, thermal conductivity, and convection heat transfer properties of the nanofluid.
Table 2 1 Classification of some nanoparticles using as addition into cutting fluid/lubrication
Carbon and derivatives of carbon Graphene, diamond, SWCNT, MWNTs
Metals Sn, Fe, Bi, Cu, Ag, Ti, Ni, Co, Pd, Au
Metal oxide ZrO2, TiO2, Fe3O4, Al2O3, ZnO, CuO
Sulfides WS2, CuS, MoS2, NiMoO2S2
Rare earth compounds LaF3, CeO2, La(OH)3, Y2O3, CeBO3
Nanocomposites Cu/SiO2, Cu/graphene oxide, Al2O3/SiO2, serpentine/La(OH)3, Al2O3/TiO2
Others CaCO3, ZnAl2O4, Zeolite, ZrP, SiO2, PTFE,
Nanofluids are typically formulated using base fluids such as water, ethylene glycol, and oil The synthesis of these nanofluids predominantly employs two methods: the one-step method and the two-step method, with the latter being the more widely used approach Two common techniques for synthesizing nanofluids using the two-step method are illustrated in Figure 2.13, which includes microwave-assisted synthesis in Figure 2.13a and the direct mixing technique in Figure 2.13b.
Figure 2 13 Two different ways to synthesize nanofluid with two-step method [75]
Utilizing nanofluid minimum quantity lubricants (MQLs) in machining enhances thermal conductivity, while also improving viscosity and density of cutting fluid nanoparticles, which collectively contribute to better surface quality These nanoparticles facilitate wear and friction reduction through four distinct mechanisms.
(1) Tiny spherical nanoparticles which will probably roll between two frictional surfaces and convert the sliding friction into a coalition of rolling and sliding friction;
(2) Nanoparticles are prone to mesh with friction sets to generate a surface preventive layer;
(3) Nanoparticles settle down into the voids of the contact surfaces to create a physical tribo-layer that makes up for the loss of mass which is called “mending effect”;
(4) Large number of nanoparticles are capable to sustain compressive force evenly thereby reducing the compressive stress concentrations associated with high contact pressure
Figure 2 14 Four techniques of nanoparticles during machining under nanofluid MQL
Literature review
Green manufacturing encompasses not only the products themselves but also critical aspects such as health and safety, waste management, process efficiency, machining costs, and energy consumption.
22 this study, in addition to minimizing the waste discharged into the environment, the authors also focused on increasing energy efficiency in machining
Figure 2 15 Sustainability assessment in metal cutting [77]
Enhancing energy efficiency is crucial for reducing energy consumption and minimizing environmental impact in machining processes Energy consumption in machining is categorized into fixed and variable energy Fixed energy, which includes systems like hydraulics and cooling lubrication, remains constant regardless of machining conditions and is determined by the original specifications In contrast, variable energy can be optimized through specific technologies and practices, impacting the energy used by spindle and cutting materials during surface generation Additionally, energy consumption can be analyzed based on the machine tool's state, which includes three phases: basic, ready, and cutting.
In machining processes, the use of coolant is essential for dissipating heat generated during cutting Implementing nanofluid minimum quantity lubrication (MQL) presents an effective solution for promoting eco-friendly practices in the industry The adoption of MQL is increasingly vital due to its advantages, including reduced environmental impact, lower cutting temperatures, and extended tool lifespan This innovative cooling technique substitutes traditional lubrication methods by utilizing a minimal fluid quantity, typically ranging from 50 to 500 ml.
The use of optimal minimum quantity lubrication (MQL) conditions, particularly with vegetable oil synergy, significantly enhances specific cutting energy and surface roughness Research by Bashir et al explored the ideal flow rate of pulse jet MQL for milling hardened AISI 4140 steel, which has a hardness of 40 HRC Incorporating nanofluids with nanoparticles can double thermal conductivity, leading to improved cooling efficiency Additionally, nanofluids contribute to better surface roughness, reduced cutting force, lower friction, and increased tool wear resistance However, there is a lack of intensive research on the machining performance of hardened steel at varying hardness levels using nanofluids in MQL applications Consequently, this study aims to investigate the hard milling process under nanofluids MQL conditions.
In the pursuit of ecologically sustainable manufacturing, reducing cutting energy is crucial Energy consumption in machine tools can be analyzed through three primary levels: the machine itself, the spindle, and the overall process From the machine perspective, it is essential to evaluate the total energy efficiency of the equipment.
Achieving energy savings at the spindle level is impractical, as it is heavily reliant on specific motor performance At the process level, energy consumption reduction is influenced by factors such as material removal, chip formation, and surface generation Key process parameters, including cutting velocity (vc), feed per tooth (fz), depth of cut (ap), and material hardness, significantly impact energy efficiency However, selecting the appropriate parameters poses a challenge for operators aiming to enhance energy efficiency in manufacturing processes.
Process parameter optimization for energy savings in metal cutting has garnered significant attention from researchers Jang et al utilized an artificial neural network (ANN) and particle swarm optimization (PSO) to minimize specific cutting energy during the milling of SM45C steel under minimum quantity lubrication (MQL) conditions Nguyen et al improved energy consumption by 34.8% and enhanced the power factor by 28.8% through a combination of radial basis function models, Grey relational analysis (GRA), and principal component analysis In another study, Nguyen employed a Kriging model and archive-based micro-genetic algorithm (AMGA) to optimize multiple objectives, achieving lower specific cutting energy, improved surface quality, and higher material removal rates Park et al combined response surface methodology with the non-dominated sorting genetic algorithm-II (NSGA-II) to optimize energy efficiency in the hard turning of AISI 4140 steel Selaimia et al applied response surface methodology and the desirability approach to enhance material removal rates and reduce surface roughness when machining austenitic stainless steel in dry conditions Khan et al optimized energy consumption and surface quality in nanofluid MQL-assisted face milling, achieving a 20.7% reduction in energy use compared to traditional methods Overall, these studies highlight the effectiveness of various optimization techniques in reducing energy consumption and improving machining performance.
25 parameter optimization for cutting energy reduction in the MQL machining process is still of interest
Before optimizing the machining process, it is essential to model the relationship between process parameters and machining performances Commonly used methods include Response Surface Methodology (RSM) and second-order polynomial models due to their simplicity However, these approaches may not be effective for capturing highly nonlinear relationships between inputs and outputs In such cases, artificial neural networks (ANN) and Adaptive Neuro-Fuzzy Inference System (ANFIS) offer better correlation compared to traditional methods Additionally, the Kriging model is also capable of effectively handling highly nonlinear relationships between input and output parameters.
After modeling, the algorithm is utilized for optimization, which encompasses a variety of techniques applied across numerous fields, yielding significant results Optimization methods are categorized into conventional and advanced types For instance, the Taguchi method has been employed to optimize factors affecting energy consumption, although it has limitations as it only addresses discrete control factors and does not guarantee a "real optimal" solution Similarly, Grey relational analysis (GRA) falls short of providing a true optimal solution, focusing instead on identifying the best factor levels In contrast, advanced optimization algorithms like Particle Swarm Optimization (PSO), Non-dominated Sorting Genetic Algorithm (NSGA-II), and Archive based Micro Genetic Algorithm (AMGA) offer enhanced capabilities, including global search, handling both constrained and unconstrained objectives, and tackling complex problems For example, Jang et al integrated Artificial Neural Networks (ANN) with PSO to optimize specific cutting energy, though optimizing for a single objective function is often impractical in metal cutting applications Additionally, Nguyen utilized the Kriging model for further optimization efforts.
Recent studies have focused on optimizing dry milling processes to enhance machining energy, surface quality, and production rates Notably, Park et al employed a hybrid approach combining Response Surface Methodology (RSM) and Non-dominated Sorting Genetic Algorithm II (NSGA-II) to improve energy efficiency and identify optimal solutions for cutting energy This highlights the ongoing interest in modeling and optimizing energy efficiency within the nanofluid Minimum Quantity Lubrication (MQL) machining process.
The literature review highlights the importance of integrating hard milling, nanofluid minimum quantity lubrication (MQL), and optimization in machining technology for both practical applications and academic research Optimizing process parameters is crucial for enhancing the quality and productivity of hard milling, yet it remains a challenge Furthermore, a comprehensive analysis of energy consumption, productivity, and quality in hard milling has not been thoroughly explored This study aims to address this gap by focusing on the multi-objective optimization of the milling process for hard-to-cut materials using nanofluids to improve machining performance.
Table 2 2 MQL condition order to obtain high productivity, quality, and efficiency of energy consumption using various algorithms
Taguchi WEDM Evaluate the impact of process parameters on errors of holes
The application of GRA couple with RSM can reduce wear, coefficient of friction, and frictional force
The results of Taguchi method are good agreement with Weibull analysis with a
27 deviation of less than 2% in all the cases
GRA Milling GRA is a strong tool for multi-objective optimization in machining to optimize cutting parameters
Position errors in CNC machine
The improvements of R 2 for X, Y, Z axis are 1-2.6%, 1-14.3%, and 10.3- 71.1%, respectively
GA Robotics The application of GA optimization technique has improved higher results than previous studies
Hybrid social spider and GA optimization
Flexible job shop scheduling problem
Diminished 5.13%, 8.55%, 9.57%, and 9.74% when applying SSO in conjunction with GA for makespan time repeated SSO or the hybrid algorithm with ABC-GA, ABC, and
NSGA-II PMEDM The utilization of NSGA-II can be decided on the best optimal condition to obtain low surface roughness and high surface micro-hardness
The algorithm and functioning of fminsearch in the proposed PSO have
28 stabilized the system more quickly and effectively
PSO EDM Using PSO, the best optimal results to deposit the biomimetic HA-containing layer were determined
PSO EDM The application of PSO technique can increment the MRR and the quality of surface roughness to 14.89 and 15.94%
LS-SVL combine with PSO
Significant increases in modeling efficiency and precision are overcome by applying the proposed method
The confidence interval of the proposed model (HPSOSA) is 95%
Research Method
Design of experiment (DOE)
Design of Experiments (DOE) is a structured approach that identifies the relationships between input and output variables, helping to uncover causal links that inform the optimization of process parameters for improved outcomes.
Design of Experiments (DOE) involves defining the number of necessary experiments and strategically selecting variable combinations to minimize time and costs It effectively establishes the relationships between variables, enabling the generation of data for regression models and the resolution of optimization problems.
In my research, I focus on two specific Design of Experiments (DOE) methods: the orthogonal array and the Box-Behnken design These techniques are part of a broader range of DOE approaches, including parameter study, full factorial, fractional factorial, Latin hypercube, and optimal Latin hypercube.
Orthogonal array designs, commonly referred to as Taguchi tables, utilize a subset of the full fractional method to minimize the number of experiments while preserving orthogonality among variables and their interactions One significant advantage of orthogonal array designs is their capability to manage discrete variables effectively However, they do have a limitation in that they overlook the interactions between factors The experimental matrices available include L4, L6, L8, L9, L16, L27, L32, L64, L81, L128, and L256, corresponding to various numbers of experiments such as 4, 6, 8, 9, 16, and 27, among others.
The Box-Behnken design, developed in 1960 by George E P Box and Donald Behnken, is an efficient method for creating higher-order response surfaces This design allows researchers to achieve significant results with fewer experiments compared to traditional factorial techniques.
The Box-Behnken design, illustrated in Figure 3.2, features defect corner points within the design space, making it advantageous for simulations that do not yield results at these corners This design method is suitable for experiments involving between 3 and 21 factors, with each factor requiring three levels.
Approximate model function
To optimize performance criteria, various modeling techniques can be employed to analyze the impact of process parameters, including second-order polynomial models, Response Surface Methodology (RSM), and Kriging This study specifically focuses on the RSM and Kriging models for their effectiveness in capturing these influences.
RSM was presented in 1951 by two researchers named George E P Box and K
B Wilson RSM model is an integration of statistical and mathematical techniques that employs a second-degree polynomial model to modeling and optimizing [4, 37] In this model, the second-order is utilized to approximate the response as below:
(3.1) where (c0) is the constant, and (ci), (cii), and (cij) are coefficients (Xi) and (Xj) are the variables (ɛ) is the random error of the experiment
RSM model is usually carried out through the following [46]:
Step 1 Define the input process and the output process
Step 3 Render the relationship with the mathematical model
Step 4 Analysis to find out the influence of the input variables on the response parameters
Step 5 Experiment to check the suitability of the model If not satisfied, proceed with the adjustment
Step 6 Decide to accept or reject the model
RSM is the most popular model for demonstrate the relationship between the machining parameters and technological responses in machining because of its simple for utilization [4, 36, 37, 89]
The Kriging model, a statistical interpolation technique, was first introduced by South African statistician and mining engineer Danie G Krige, a pioneer in geostatistics This model was further developed by French mathematician Georges Matheron in the 1960s Various forms of the Kriging model exist, including simple Kriging, ordinary Kriging, Universal Kriging, and blind Kriging In our study, we utilize ordinary Kriging, which is the most widely used method for modeling the relationship between input and output variables.
Figure 3.3 illustrates one-dimensional data interpolation using the Kriging model, where red square points represent the initial data points The gray regions depict bell curve confidence intervals, while the dashed blue curve signifies a spline curve that intersects the initial data points but does not conform to the confidence intervals The red line, which aligns with the bell curve confidence intervals, effectively showcases the Kriging model's interpolation capability This model excels in producing an interpolated spatial representation and assessing the uncertainty associated with each point within the model [108].
Figure 3 3 The sample of one-dimensional data interpolation generated by the Kriging model [109]
The Kriging estimator provides an estimated value of Z(s0) at the point s0, based on observed values Z(s1), Z(s2), …, Z(sn) collected at n predetermined design points s1, s2, …, sn in Euclidean space Rd, through a linear combination of these observations.
The weight wi(s0) is find out to minimized the mean-square predictor error
Where γ indicates the (co) variances
is the matrix whose (i,j)th element is
It's necessary that the predictor is unbiased and the weights of the predictor requirement meet the constraint
By submitting a Lagrange multiplier -2λ and minimize (w, )= T w 2 T 2 ( T 1)
Two equations in the matrix form are derived by implementing stationary condition F=0
By answering Equation 3.7, the coefficient wi of the linear predictor of Equation 3.2 can be recognized as bellow:
The optimal weights for the Kriging model, as outlined in Equation 3.8, depend on the covariance relationships among the process parameter values The general function relationship of the Kriging model can be established as follows.
The correlation parameter, hj, quantifies the significance of process parameters j, while h is determined by the interval input points To assess the correlation function parameters, Kriging software and relevant literature utilize maximum likelihood estimators.
For each s0 point, an assessed value is computed utilizing Equations 3.2 and 3.8 The mean-squared forecast error is determined as
The Kriging model is considered a better approximation model than other approximations as RSM or ANN This model can perform better at nonlinear characteristics [112, 113] and power in experiment expenses [114].
Optimization
Particle Swarm Optimization (PSO) is a global optimization algorithm inspired by natural evolution, first introduced by Kennedy and Eberhart in 1995 Unlike gradient-based optimization methods that may become stuck in local optima, PSO effectively identifies the global optimum This algorithm leverages evolutionary computation principles and swarming theories, offering a cost-effective solution for optimization challenges.
The Particle Swarm Optimization (PSO) algorithm is extensively documented in existing literature, making further duplication unnecessary This work utilizes the PSO framework illustrated in Figure 3.4, with Isight 5.9 and Matlab R2015a software employed to enhance the multi-objective optimization process.
Figure 3 4 Algorithm for PSO 3.3.2 Non-dominated sorted genetic algorithm (NSGA-II)
The NSGA algorithm, initially proposed by Srinivas and Deb, presented complexities in calculations To address these challenges, Deb and colleagues introduced the improved NSGA-II in 1988, which incorporated several advancements to enhance the algorithm's convergence.
NSGA-II is a cutting-edge multi-objective optimization algorithm that is extensively utilized in machining optimization This algorithm is characterized by three key features: a rapid non-dominated sorting method, an efficient crowded distance estimation technique, and a straightforward crowded comparison operator The NSGA-II process can be outlined in a systematic step-by-step manner.
Step 1: Initialization of Non-dominated sorted genetic algorithm based on the problem range and constraint
Step 3: Combine population and evaluate with non-dominated sorting method Step 4: Generate population from the results from the last step
Step 5: Utilize selection, crossover, and polynomial mutation to produce the new population
Step 6: Recombination and selection to produce the new population in reliance on their rank
The steps of NSGA-II multi-objective algorithm can be seen more clearly through the flowchart shown in Figure 3.5
Figure 3 5 Flowchart of NSGA-II algorithm
Results and Discussion
Multi-objective optimization of surface roughness and cutting forces in hard
In this research, a stability lobe diagram of the machine tool is developed by means of CUTPRO software as is shown in Figure 4.1
Figure 4 1 Procedure of development of the stability lobe diagram
The input process parameters for the hard milling tests were determined using the stability lobe diagram illustrated in Figure 4.2, which delineates the spindle speed-dependent boundary between stable and unstable depths of cut This diagram was essential in identifying optimal cutting conditions for subsequent machining processes Consequently, the selected input parameter values for the hard milling process are detailed in Table 4.1.
Figure 4 2 Analytical stability lobe diagram for hard milling test
Table 4 1 Values of input parameters
Axial depth of cut, a [mm] 0.3 0.4 0.5 0.6
The Taguchi method is currently employed to design experiments for the hard milling of AISI H13 steel, with input variables and their values detailed in Tables 4.1 and 4.2 Additionally, Response Surface Methodology (RSM) is applied to formulate second-order experimental equations and facilitate multi-objective optimization for the hard milling process.
Hard milling trials were performed on AISI H13 steel workpieces measuring 200 mm × 100 mm × 40 mm with a hardness of 50 HRC The trials utilized CMTec's M520 ultra carbide end mills, featuring a diameter of 10 mm, a 35-degree helix angle, and a square type with four flutes.
The milled surfaces were measured at three different positions by a Surftest SJ-
In this study, the cutting force components in three directions were measured using a piezoelectric three-component Dynamometer Model 6423 (Lebow®) The resultant cutting force, essential for calculating the cutting force in hard milling, was determined using a specific equation.
After experimentation, the values of the Ra and the Ft were given in Table 4 2
Code of factors Value of factors Results
ANOVA for the response characteristics
Table 4 3 ANOVA results for Ra and Ft
Source DF SS MS F P % PC Remarks
ANOVA of model for Ra
ANOVA of model for Ft
The ANOVA results for Ra indicate that the cutting speed (v), feed rate (f), and depth of cut (a) are statistically significant factors, with v contributing the most at 35% of the total variation The feed rate and depth of cut follow closely, contributing 33.6% and 22.2%, respectively Additionally, the feed rate and depth of cut show significant contributions of approximately 38.57% and 39.21% in the Ft analysis The R² regression coefficients for Ra and Ft are 0.996 and 0.935, respectively, demonstrating a strong correlation between the empirical data and the statistical model.
The Ra and Ft values in the hard milling of AISI H13 steel are predicted using the RSM method combined with Taguchi experimental design, resulting in second-order equations derived from the experimental data.
Effect of the input variables on the Ra and Ft models
Figure 4 3 The plots of response surface for Ra
The relationship between feed rate (f) and cutting speed (V) significantly influences surface roughness (Ra), with a decrease in feed rate and an increase in cutting speed leading to a notable reduction in Ra Previous studies indicate that a built-up edge (BUE) forms at low cutting speeds, resulting in inferior surface quality Conversely, increasing the cutting speed eliminates the BUE, thereby enhancing the quality of the milled surface.
Figure 4.3b shows that an increase in the feed rate will lead to a rapid increases in
Ra This could be explained since the helicoidal movement of end mill cutter makes furrows on milled workpiece, as reported by many earlier researchers [120]
Figure 4.3c reveals that when the cutting speed is decreased and the depth of cut is increased, this will cause a considerable increase in the Ra This can be accepted
42 because the increasing speed of the depth of cut provides the chip load which in turn causes a boost in the cutting forces, as presented by many previous authors [121]
Figure 4 4 The plots of response surface for Ft
The interaction plots illustrate that the resultant cutting force significantly rises with increasing cutting parameters, such as feed rate (f) and depth of cut (a), while it decreases with a reduction in cutting speed (V).
The simultaneous optimization of both Ra and Ft has yielded optimal results for multiple objectives, as detailed in Table 4.4 The ideal input variables for achieving minimal Ra and Ft are identified as a cutting speed of 100 m/min, a feed rate of 0.015 mm/tooth, and a depth of cut of 0.44 mm The corresponding minimum values for Ra are also provided in the table.
The findings indicate that the Ft values are 0.206 àm and 66.58 N, respectively Additionally, the composite desirability achieved is 0.9473, which is nearly equal to 1 Therefore, it can be concluded that this multi-objective optimization is highly effective for the machining process.
Table 4 4 Results of multi-objective optimization for the Ra and Ft
This study utilized the Response Surface Methodology (RSM) with Taguchi experimental design to perform multi-objective optimization of Ra (surface roughness) and Ft (tool wear) during the hard milling of AISI H13 alloy steel at a hardness of 50 HRC The findings reveal significant insights into optimizing machining parameters for enhanced surface quality and tool longevity.
ANOVA results indicate that cutting speed, feed rate, and depth of cut significantly affect Ra at a 95% reliability level Among these parameters, cutting speed is the most influential, followed by feed rate and then depth of cut.
(2) Based on the results of ANOVA for Ft, it revealed that the given parameters (f and a) have a powerful effect on Ft at the reliability level 95%
The ideal machining parameters for achieving optimal surface roughness (Ra) and cutting force (Ft) were determined to be a cutting speed of 100 m/min, a feed rate of 0.0150 mm/tooth, and a depth of cut of 0.44 mm, resulting in Ra and Ft values of approximately 0.206 µm and 66.58 N, respectively.
Comparative machinability performance of AISI H13 steel under dry, MQL, and
The Design of Experiments (DOE) approach was utilized to identify key variables that significantly impact surface roughness, cutting force, and cutting temperature To achieve this, the Taguchi method and ANOVA were employed to optimize cooling conditions and cutting parameters aimed at minimizing these factors The L27 orthogonal array from Taguchi’s experimental design was used to structure the experiments, focusing on four input factors: cutting speed, feed rate, depth of cut, and cooling conditions, each with three levels as detailed in Table 4.5 These input factors and their levels were selected based on their effects on the output factors.
Responses Goal Optimum conditions Predicted values v f a
Table 4 5 The input factors and levels Level
The experimental setup, depicted in Figure 4.5, utilized an AISI H13 steel workpiece measuring 50 mm in width, 200 mm in length, and 100 mm in height, with a hardness of 45HRC The experiments employed Φ10 TiAlN coated end mill tools and utilized Minimum Quantity Lubrication (MQL) cutting conditions, specifically using cutting oil CT232 at a flow rate of 90 ml/h and an air pressure of 3 kg/cm² For the nanofluid-based MQL, aluminum oxide (Al2O3) nanoparticles, averaging 10-17 nm in diameter and at a concentration of 1 wt%, were chosen for their beneficial tribological properties A magnetic stirring device mixed the cutting oil and nanoparticles for 12 hours to ensure uniform dispersion The MQL nozzle was fixed at a 60-degree angle and 30 mm from the tool's relief face during all experiments, which were repeated three times to reduce experimental error The cooling conditions for machining are summarized in Table 4.6.
Figure 4 5 Setup for experiments Table 4 6 Machining cooling conditions Cutting tool End milling Φ10 TiAlN coated carbide
Al2O3 nanoparticle enhanced MQL MQL spray parameters Distance: 30 mm
Flow-rate of MQL oil: 90 ml/h 4.2.3 Results and Discussions
The purpose of this study is to find the optimal value of input factors for minimizing surface roughness, cutting force and cutting temperature Hence, the-smaller-
46 is-better type of Taguchi method signal-to-noise (S/N) ratio was selected as estimated according to the following equation:
Where: yi is the observed data, n is the number of experiments repeated
The L27 array was utilized to organize experiments involving four factors—cooling condition (C), cutting speed (v), depth of cut (d), and feed rate (f)—each assessed at three levels, indicated by "1," "2," and "3." The outcomes of the experiments, along with the signal-to-noise (S/N) ratio calculated using formula (4.4), are presented in Table 4.7.
Table 4 7 The result of the experiment and S/N ratios
Surface roughness is a crucial metric for assessing the output quality in metal machining In this study, the Mitutoyo SJ-401 surf-test instrument was used to measure the workpiece surface roughness The analysis revealed that the optimal conditions for minimizing surface roughness are a cooling condition at level two, a cutting speed at level three, a depth of cut at level one, and a feed rate at level one, corresponding to Experiment No 16 The findings indicate that the cooling condition is the most significant factor affecting surface roughness, followed by the feed rate.
Table 4 8 The mean of S/N response for surface roughness
Figure 4.6 illustrates the S/N response graph, which indicates that to achieve lower surface roughness, the optimal cutting parameters and cooling conditions involve the use of nanofluid-based MQL.
48 m/min for the cutting velocity, 0.2 mm for the depth of cut, and 0.01mm/tooth applied for the feed rate
The analysis of variance (ANOVA) results, as shown in Table 4.9, indicate that cooling condition is the most significant factor affecting surface roughness, contributing 60.11% to the total effect, followed by feed rate at 19.29% The P values for cooling condition, feed rate, and velocity are all less than 0.05, confirming their statistical significance With an R-Sq value of 86.20%, the study reveals that these input factors account for a substantial portion of the variability in surface roughness The research also highlights the effectiveness of nanofluid-based minimum quantity lubrication (MQL) in enhancing surface roughness compared to dry and MQL conditions Furthermore, optimal surface roughness is achieved with the minimum feed rate, minimum depth of cut, and maximum cutting speed.
The surface roughness of machined materials is significantly influenced by cooling conditions, followed closely by feed rate, cutting speed, and depth of cut Abrupt changes in roughness are primarily attributed to variations in cooling conditions, highlighting the effectiveness of nanofluids in enhancing surface quality due to their excellent anti-friction and anti-wear properties at the nano level Additionally, an increase in feed rate leads to a rise in surface roughness, primarily due to the impact of machining dynamics.
SN ratios for surface roughness
Signal-to-noise: Smaller is better
Research by Do and Hsu, as well as H Aouici et al., indicates that a higher feed rate results in deeper and larger furrows on machined surfaces Additionally, cutting speed and depth of cut have minimal impact on the overall surface finish.
Table 4 9 The ANOVA table for roughness
Source DF Adj SS Adj MS F-Value P-Value PC %
Cutting force is a crucial factor in designing machine-fixture-tool-work systems In this study, cutting force was measured using a three-component piezo-electric Kistler dynamometer The analysis revealed that the optimal conditions for the experiment are a cooling condition at the second level, cutting speed at the third level, depth of cut at the first level, and feed rate at the first level, corresponding to Experiment No 16 The findings indicate that the depth of cut is the most significant factor influencing cutting force, followed by the feed rate.
Table 4 10 The mean of S/N response for the cutting force
The S/N response graph in Figure 4.7 reveals that the optimal cutting parameters and cooling conditions for minimizing cutting force include the use of Nanofluid-based MQL for cooling, a cutting velocity of 80 m/min, a depth of cut of 0.2 mm, and a feed rate of 0.01 mm per tooth.
The analysis of variance (ANOVA) results in Table 4.11 indicate that the depth of cut is the most significant parameter affecting cutting force, contributing 44.58% to the total effect, followed by feed rate at 26.99% The cooling condition factor accounts for 20.39% of the total effect Statistically significant P values (