Prediction Of Material Removal Rate Research Articles

Deep ensemble learning for material removal rate prediction in chemical mechanical planarization with pad surface topography

Chemical mechanical planarization (CMP) stands as a critical process in semiconductor manufacturing, necessitating precise prediction of material removal rate (MRR) to achieve optimal global planarization of material surface step heights. A notable hurdle in MRR prediction lies in the nonlinear pad-wafer interaction stemming from variations in the polishing pad surface. Despite numerous extant studies examining the asperity of the pad surface and the wafer-pad contact, aspects such as slurry lubrication and micro-topographical structure of the pad surface remain unaddressed. This study firstly proposes a deep ensemble learning-based approach for MRR prediction, leveraging the intricate interactions between the overall features of the pad surface and the polishing pressure. We experimentally generated diverse pad surfaces by controlling the process variables governing pad surface topography. The Multi-TabNet model, constructed by independently training multiple TabNet instances with distinct initialization and hyperparameters, furnishes a framework for MRR prediction across a broad spectrum of pad surface features, encompassing pores and cores. The results demonstrate a root mean square error of 80 Å/min, with the error margin reduced by up to 70 % compared to predictions utilizing a single parameter of the pad surface. Furthermore, it exhibits an error improvement exceeding 17 % relative to the existing method solely considering asperity. This study not only imparts insights for addressing MRR variations contingent on pad condition but also facilitates dependable material removal even on variable pad surfaces due to conditioner wear.

Real-time prediction of material removal rate for advanced process control of chemical mechanical polishing

Polishing torque holds significance in monitoring the chemical mechanical polishing (CMP) process due to its close correlation with material removal. This study introduces a new model-based technique for estimating the material removal rate (MRR) using in-process data from CMP machines. The proposed method employs either the sequential least squares method or Kalman filter for real-time state estimation. Real-time estimation of MRR enables material removal control without relying on conventional endpoint detection (EDP) technologies. The accuracy of the proposed approach is validated through oxide CMP experiments, demonstrating precise estimation of the center-slowed MRR profile towards the end of the pad life.

Effect of Al7075 and activated carbon reinforced composite on optimizing WEDM responses

This paper presents wire cut electrical discharge machining (WEDM) response characteristics of Aluminium 7075 (Al7075) reinforced with powdered activated carbon (PAC) composite. In recent days WEDM has become a significant machining process in targeting its benefits of contributing improved material removal rate (MRR) and low surface roughness (SR). This is due the rising need for intricate, accurate, and superior structural components, the WEDM process emerges as a formidable alternative to traditional machine tools. In this work Pulse-on time (Ton), pulse-off time (Toff), discharge current (IA) and servo speed rate (SS) are the variables to be given as input and machining responses such as MRR and SR are studied. From Analysis of Variance (ANOVA) study it is found that discharge current and servo speed is the significant parameters. The optimal desirability condition is obtained with input parameters Ip: 2000 mA; Ton: 8.9 μs; Toff: 25 μs and SS: 150 rpm for the precision machining. The optimum response parameters are found as MRR 10.46 mm3/min and SR 3.32 μm. Results shows that the model designed for the prediction of MRR produces an above 98.27% and the prediction of SR is above 97.17%. The error percentage among the experimental and predicted MRR and SR were estimated. Additionally confirmatory test is performed with optimal results achieved from response surface methodology (RSM) and desirability technique. Metallurgical tests like electron backscatter diffraction analysis (EBSD) and microstructure are conducted to confirm the surface properties and atomic force morphology (AFM) analysis is applied to clarify the structural features of machined composites. The results revealed that the variation of hard deflection is caused by depression of eroded materials on the top layers of machined surface.

Micro-feature fabrication on SS304 stainless steel using WEDM process

ABSTRACT Fabricating micro-pillars is a tough procedure as these features are highly debile. Wire Electro-Discharge Machining (WEDM) is a popular micromachining process to fabricate complex features on electrical conducting materials. SS304 stainless steel is commonly used in manufacturing of heating and cooling devices. A simple method of micro-pillars array fabrication on SS304 Stainless Steel material using WEDM was used through perpendicular crossed wire movement, keeping specified amount of step-over distance to form pillars. An analytical model for material removal rate prediction was developed using input parameters like spark duration, off time, voltage and current. Also, the influence of the process parameters was studied on depth and kerf width produced. Voltage and pulse-on time were found to be the major influencing parameters. Micro-featured surfaces fabricated for minimum kerf width and maximum feature depth were analysed using SEM and EDS along with the wettability study. The fabricated surfaces remained hydrophobic even after 100s.

A Novel Physically Guided Data Fusion Prediction Model for Micro-EDM Drilling

Accurate prediction of Electro-Discharge Machining (EDM) results is crucial for industrial applications, aiming to achieve high-performance and cost-efficient machining. However, both the current physical model and the standard Artificial Neural Network (ANN) model exhibit inherent limitations, failing to fully meet the accurate requirements for predicting EDM machining results. In addition, Micro-EDM Drilling can lead to the distortion of the macroscopic shape of machining pits under different input conditions, rendering the use of only the volume of machining pits as the evaluation index insufficient to express the complete morphological information. In this study, we propose a novel hybrid prediction model that combines the strengths of both physical and data-driven models to simultaneously predict Material Removal Rate (MRR) and shape parameters. Our experiment demonstrates that the hybrid model achieves a maximum prediction error of 4.92% for MRR and 5.28% for shape parameters, showcasing excellent prediction accuracy and stability compared to the physical model and the standard ANN model.

Modelling and simulation of MRR in electro-chemical surface grinding of Al-SiC-Gr using hybrid analytical approach

Electro-Chemical Surface Grinding (ECSG) is a hybrid machining process which utilizes synergic actions of three sub-processes viz. electrochemical dissolution, grinding action of abrasive particles and erosive action of electrolyte flow in inter electrode gap. Since, occurrence of turbulence in machining zone significantly affects the machining performance, it is required to analyze the turbulence phenomenon in ECSG process. The exact solution and Numerical solution–based hybrid analytical approach has been used to predict the Material Removal Rate (MRR) in ECSG of Al-SiC-Gr. In this model, the streaming potential of electrolytic action, abrasive action of grinding wheel and shear force of electrolyte flow between inter electrode gap have been considered. Finite Element Method based Numerical modelling has been carried out to analyze the electrolyte flow in machining zone in order to evaluate the MRR due to erosive action. It was observed that the application of shear force on oxide layer rises with the increase in flow rate of electrolyte. Confirmation experiment has been performed to validate the modelling results. Validation results show a minor error of 8.21% in prediction of MRR using analytical as well as finite element model.

Implementation of robotic electrochemical machining in freeform surface machining with material removal rate prediction using different machine learning algorithms

Robots are used as machine tools via some advanced properties such as versatility, adaptation and mobility with other control systems. These advantages can be the solution to overcome the encountered problems in electrochemical machining (ECM). In this manner, the usability of a new machining method that is Robotic electrochemical machining (RECM) in freeform surface machining is investigated. Firstly, ECM parameters and their effects on the material removal rate (MRR) is discussed and compared with the literature. Experimental results showed that MRR variation for RECM acts similar characteristics with conventional ECM processes. The increase in voltage, electrolyte conductivity and feed rate increase MRR. The highest effect on MRR for both materials is voltage, electrolyte conductivity, and feed rate, respectively. Secondly, three different machine learning models are designed as Gaussian Process Regression (GPR), regression tree and deep neural network (DNN) models and the relationships between their actual and predicted values are discussed. The performance of the DNN model is better than other models with a correlation of determination values equal for Inconel 718 and AISI 1040. Furthermore, the DNN model with four input parameters outperforms developed models such as GPR and DTR with optimized hyperparameters regarding the R2, MSE, RMSE, and MAE. The change in height in the Z-axis and the surface roughness of the machined surfaces are also measured. The results show that the decrease in the Z axis is homogeneous, that is, uniform machining occurs and also the surface roughness has improved. Therefore, RECM can be a solution to machine free-form surfaces for different materials with high predictability.

Influence of wall-slip on material removal in abrasive flow machining

Abrasive flow machining (AFM) process, as a widely known technique for finishing internal surfaces of complex parts, can significantly improve the performance and reliability of components in numerous industries. Prediction of material removal rate (MRR) is a vital task in AFM process because it significantly affects the efficiency and cost-effectiveness of the process. In-depth studies about the influence of input parameters on material removal in AFM process, however, are limited. In this paper, the mechanism of material removal was investigated, and a MRR modelling was then established in which viscoelastic extrusion modelling with wall slip conditions was employed in AFM process. A novel fixture with gradient shear flow channel was designed. MRR obtained by simulations and experiment showed a reasonable agreement. Finally, parameters, including abrasive media pressure, velocity and shape geometry of flow channel, were identified as key factors affecting MRR in AFM process. Established computational fluid dynamics (CFD) analysis and MRR modelling can be a powerful tool to determine MRR in future AFM process.

Prediction of material removal rate of SS304 alloy in electrochemical micromachining using machine learning

Electrochemical micromachining (ECMM) is a non-traditional machining process used in machining microfeatures that are being used in the fabrication of MEMS, finishing surgical tools, implants, and cooling channels. Machine learning (ML) algorithms can be used to predict the output of a machining process based on the given input parameters. In this Work, ML is interpreted into ECMM where, different ML models were built based on Decision tree regression, Random Forest regression, and Support vector regression algorithms to find the relationship between the interactive effects of the type of electrolyte, voltage, feed rate, and duty cycle on Material Removal Rate (MRR). The work also aims to find the best-fit algorithm and approach toward the prediction of MRR of AISI SS304 alloy when machined in ECMM using Regression Analysis. The results showed that higher efficiency was attained in the Decision tree regression model with a training accuracy of 99.87% and testing accuracy of 99.75% and decision tree was found to be the best fit algorithm for the ECMM process for the single response parameter.

Material removal rate prediction and surface quality study for ultrasonic vibration polishing of monocrystalline silicon

Material removal rate prediction and surface quality study for ultrasonic vibration polishing of monocrystalline silicon

Performance prediction and parameter optimization of alumina-titanium carbide ceramic micro-EDM hole machining process based on XGBoost

Aiming at the difficulties in setting process parameters and the low accuracy of process performance prediction in electrical discharge machining (EDM) hole machining of alumina ceramics, a novel EDM process performance prediction method based on eXtreme Gradient Boosting (XGBoost) is proposed in this study. The independent variables selected include processing polarity (P), interelectrode voltage (V), peak current (I), pulse frequency (F), pulse width (T), and tool electrode and workpiece electrode gap (W), while the dependent variables are tool electrode length loss (ELW), tool electrode volume loss (EVW), and material removal rate (MRR). An L18 (21 × 35) orthogonal test is designed to obtain training samples for model development. The accuracy of the prediction results for ELW, EVW, and MRR by the XGBoost model is found to be 1.7%, 2.5%, and 9.4%, respectively. These results show a significant improvement compared to the linear regression model, with an improvement of 41.7%, 62.5%, and 19.4%, respectively. Furthermore, when compared to the support vector regression (SVR) model, the XGBoost model also shows improvement of 23.6%, 5.9%, and 12.0% for ELW, EVW, and MRR prediction accuracy, respectively. These findings suggest that the proposed XGBoost-based method is effective and accurate in predicting the performance of alumina ceramic EDM processes. In addition, the third generation nondominated sorting genetic algorithm (NSGA-III) is utilized to optimize the process parameters with both single objective and multi-objective approaches. The optimal parameters for ELW, EVW, and MRR, including the selection of machining polarity, interelectrode voltage, peak current, pulse frequency, pulse width, and tool electrode and workpiece electrode gap are obtained through the Pareto frontier of the three comprehensive optima. The average relative errors between the experimental results and the optimized results are found to be 6.46%, 10.45%, and 9.58% for ELW, EVW, and MRR, respectively, indicating the accuracy and effectiveness of the optimized results.

Material Removal Rate Prediction Based on Broad Echo State Learning System for Magnetically Driven Internal Finishing

Material removal rate (MRR) modeling is crucial for process and quality control in abrasive finishing processes. In this paper, a novel broad echo state learning system (BESLS) is developed to predict MRR in the magnetically driven internal finishing process based on the force signal. This BESLS model employs reservoirs with echo state properties to replace the enhancement layer nodes of the conventional broad learning system. Three representative BESLS architectures are designed and experiments are conducted to validate the developed models. The effect of feature-fusion approaches on prediction accuracy is investigated. Results demonstrate that the prediction error of MRR is only 6.86% for the proposed BESLS model. The model has one reservoir with variable nodes (BESLS-III) and fuses extracted force features and process parameters as input. It also shows remarkable computational efficiency with a testing time of 0.134 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\rm s$</tex-math></inline-formula> , making it a potential candidate for online process control.

Semi-Supervised Deep Kernel Active Learning for Material Removal Rate Prediction in Chemical Mechanical Planarization.

The material removal rate (MRR) is an important variable but difficult to measure in the chemical-mechanical planarization (CMP) process. Most data-based virtual metrology (VM) methods ignore the large number of unlabeled samples, resulting in a waste of information. In this paper, the semi-supervised deep kernel active learning (SSDKAL) model is proposed. Clustering-based phase partition and phase-matching algorithms are used for the initial feature extraction, and a deep network is used to replace the kernel of Gaussian process regression so as to extract hidden deep features. Semi-supervised regression and active learning sample selection strategies are applied to make full use of information on the unlabeled samples. The experimental results of the CMP process dataset validate the effectiveness of the proposed method. Compared with supervised regression and co-training-based semi-supervised regression algorithms, the proposed model has a lower mean square error with different labeled sample proportions. Compared with other frameworks proposed in the literature, such as physics-based VM models, Gaussian-process-based regression models, and stacking models, the proposed method achieves better prediction results without using all the labeled samples.

Investigation of Wear and Wire Electrical Discharge Machining Characteristics of Al-Mg-MoS2 Composites Using Response Surface Method

Abstract Aluminum metal matrix composites remain sophisticated materials with a broad series of applications because of their high strength-to-weight ratio, better resistance to wear and corrosion, improved high-temperature zone applications, and reasonable machinability. Furthermore, when manufactured through the powder metallurgy process, the tribological and machinability properties of aluminum-magnesium–based composites are found to be improved with the inclusion of molybdenum disulfide reinforcement. The response surface methodology technique was used for estimating the wear loss and coefficient of friction to study the tribological property. Also, it is employed in the wire electrical discharge machining (WEDM) technique, to predict material removal rate, surface roughness, and overcut. The projected values in both the wear and WEDM parameters are well within the minimal error value, according to the perturbation plot, and sufficient validation has been performed. The created model reveals that at 15 N, 2 m/s, 750 m, and 15 A, 90 µs, and 45 V, better tribology and WEDM characteristics are acquired.

Parametric optimization and prediction of MRR and surface roughness of titanium mixed EDM for Inconel 718 using RSM and fuzzy logic

Powder mixed electric discharge machining (PMEDM) was developed to improve the machining rate and surface quality of conventional EDM. In this study, PMEDM was performed on Inconel 718 by adding titanium particles to the dielectric fluid. The input parameters selected for the experiment were powder concentration, pulse current, gap voltage, pulse on time and pulse off time and the effects of the input parameters were investigated on the material removal rate (MRR) and the surface roughness of the sample. Response surface methodology (RSM) was used to construct the design of experiment. Analysis of variance (ANOVA) was carried out to find out the significant input parameters for each response. Further, Mamdani based fuzzy logic was developed using the experimental data and used to predict the responses in optimized condition. The multi-objective optimization was performed using the RSM desirability function approach to maximize the MRR and minimize the surface roughness. From ANOVA, it is observed that powder concentration, pulse current, gap voltage and pulse on time are the significant input parameters for MRR and surface roughness. The fuzzy result is compared with the experimental data and found the accuracy of prediction is 89.21 % for MRR and 91.23 % for surface roughness. The optimized setting of input parameters is obtained as powder concentration = 8 g/l, pulse current = 9.5 A, gap voltage = 60 V, pulse on time = 150 µs and pulse off time = 20 µs and the corresponding optimum value of MRR and surface roughness are 16.623 mm 3 /min and 3.71 µm respectively. The SEM result indicates that the width of the surface crack is found more in the optimized sample compared to the sample having the least surface roughness. From the EDX analysis, it is found that a very small amount of copper and titanium particles were included on the surface during the machining.

ANN and RSM based predictive model development and EDM process parameters optimization on AISI 304 stainless steel

The current research was conducted in an Electrical Discharge Machining (EDM) on AISI 304 stainless steel. The experimental study was carried out using full factorial design methods, with three input parameters chosen for this experiment: pulse time (T), discharge current (C), and gap voltage (V), and the responses being material removal rate (MRR) and surface roughness (SR). Artificial neural network (ANN) and response surface methodology (RSM) models are the two different types of predicting models for MRR and SR that have been established. It has been found that both models have good accuracy, albeit ANN predicted model is more accurate. The maximum percentage of error in the MRR prediction was 6.16 for the ANN model and 10.45 for the RSM model, respectively Also, the maximum percentage of error in the SR prediction was 5.32 and 16.18 for the ANN and RSM models, respectively. The purpose of this paper is to increase the MRR as well as decrease the SR and equivalent optimum machining parameters which is a very important aspect of the present research. For this purpose, a multi objective function for MRR and SR has been developed based on the RSM method and finally, it is optimised. It has been found that 9 A, 200 s, and 45 V, respectively, are the optimal machining parameters for obtaining discharge current (C), pulse on time (T), and gap voltage (V). Future attempts may use more advanced artificial intelligence methods like ANN-GA, ANN-PSO, etc.

In-process belt-image-based material removal rate monitoring for abrasive belt grinding using CatBoost algorithm

A reliable material removal rate (MRR) prediction method significantly optimizes the grinding surface quality and improves the processing efficiency for robotic abrasive belt grinding. Using worn-belt image features to predict MRR is a direct and reliable method; however, this method is rarely reported at present. This paper proposes an MRR prediction method for Inconel 718 grinding based on the abrasive belt image analysis and categorical boosting (CatBoost) algorithm. During belt grinding, four wear types of abrasive belts, namely fracture, adhesion, rubbing wear, and fall-off, are identified and analyzed. Under various grinding parameters, the experimental MRR rapidly decreases at first, then in a gradual manner. For an effective evaluation of belt wear severity, cutting grain area ratio, color moments, and texture features are extracted from belt images. MRR and abrasive belt image features are strongly correlated after normalization. All image features are taken into account for MRR prediction model training. Verification experiments indicate that the predicted data is in good agreement with the experimental data. The maximum absolute error, mean absolute error, root mean square error, and determination coefficient of the MRR prediction model are 0.17 μm, 0.4 μm, 0.2 μm, and 99.42%, respectively, which are superior to those of other popular machine learning algorithms. In this study, we present a comprehensive understanding of the relationship between MRR and abrasive belt characteristics, as well as demonstrate the feasibility of accurately predicting MRR using the CatBoost algorithm.

EXPERIMENTAL INVESTIGATION AND OPTIMIZATION OF WIRE EDM PARAMETERS FOR SURFACE ROUGHNESS, MATERIAL REMOVAL RATE AND SURFACE TOPOGRAPHY IN MACHINING OF AISI D3

Wire electric discharge Machining is a crucial alternative machining technique. As a consequence of its superior precision, finish, machinability of any hard material, and ability to form complex shapes, demand for it continues to rise. In this research, we apply the Taguchi Method in conjunction with the Grey relational analysis approach to determine the optimal parameters in WEDM for the machining of AISI D3. WEDM machines allow for cutting by adjusting process parameters such as pulse on time, pulse off time, servo voltage, wire feed rate, current, and more to get the desired results in terms of response variables like surface roughness and material removal rate. Grey Relational Analysis is used for optimization of all these parameters with multi-response features. Best WEDM parameters are determined by studying grey relational grade. Specifically, the findings demonstrate its improved surface roughness and material removal rate prediction skills and application to such industrial WEDM cutting, which leads to the effective selection of machining settings for improved Machining quality

Material removal rate prediction in chemical mechanical planarization with conditional probabilistic autoencoder and stacking ensemble learning

Material removal rate prediction in chemical mechanical planarization with conditional probabilistic autoencoder and stacking ensemble learning

Material Removal Rate Prediction for Sapphire Double-Sided CMP Based on RSM-SVM

As a crucial substrate material for optoelectronic materials, sapphire has important applications in both military and civilian fields. In order to achieve the final processing quality of sapphire substrate materials, double-sided chemical mechanical polishing (DS-CMP) is a necessary process, which is also a guarantee for the preparation of high-end LED chips. In this article, the sapphire DS-CMP processing plan based on the Box-Behnken design is obtained and experimented. Then, a hybrid approach of response surface method (RSM) and support vector machines (SVM) algorithm is established as the material removal rate (MRR) prediction model for sapphire DS-CMP. Furthermore, the material removal process of sapphire DS-CMP, the influence of response variables on the MRR of sapphire DS-CMP, and the prediction results of RSM-SVM on sapphire DS-CMP are analyzed respectively. From the experimental results, the maximum MRR obtained is 387.59 nm min−1, which is more than 6 times the reported MRR of single-sided CMP under similar process parameters. The mean square error of predicted value through RSM-SVM is basically around ±10% of the experimental value, which possess satisfied validity for the MRR prediction of sapphire DS-CMP. Finally, both top and bottom surface quality of sapphire wafers after DS-CMP processing was investigated.