Exploring the intricacies of machine learning-based optimization of electric discharge machining on squeeze cast TiB2/AA6061 composites: Insights from morphological, and microstructural aspects in the surface structure analysis of recast layer formation and worn-out analysis

Kumar, Raman; Channi, Arvinder Singh; Kaur, Rupinder; Sharma, Shubham; Grewal, Jasmaninder Singh; Singh, Sehijpal; Verma, Amit; Haber, Rodolfo

doi:10.1016/j.jmrt.2023.09.127

Cited by 21 publications

(10 citation statements)

References 87 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Level one is the system design, which looks into the basic factors that cause the desired output . The emphasized second level is the parameter design, which intends to minimize variation from the target by establishing very pertinent optimal parameters covering all the basic factors . The third level is the tolerance design, intended for the needed level of variation by identifying the sensitive elements affecting the product quality .…”

Section: Methodsmentioning

confidence: 99%

“… 52 The emphasized second level is the parameter design, which intends to minimize variation from the target by establishing very pertinent optimal parameters covering all the basic factors. 53 The third level is the tolerance design, intended for the needed level of variation by identifying the sensitive elements affecting the product quality. 54 As a result, the technique aids in determining which process parameters are crucial to improving product quality, which parameters should be modified, the level at which process criteria should operate, and what needs to be improved and rigorously controlled.…”

Section: Methodsmentioning

confidence: 99%

“…The median absolute error (MedAE), eq , is similar to MAE but uses the median instead of the mean to calculate the average absolute differences between predicted and actual values . It is less sensitive to outliers in the data compared to mean-based metrics like MAE or RMSE. , RMSE = ∑ i = 1 N ( A N i − A V i ) 2 N MSE = 1 N ∑ false( A V i − A N i false) 2 r = n ( ∑ x y ) − ( ∑ x ) ( ∑ y ) [ n ∑ x 2 − ( ∑ x ) 2 ] [ n ∑ y 2 − ( ∑ y ) 2 ] MAE = 1 N ∑ | A V i − AN | …”

Section: Machine Learning: Random Forest Regressionmentioning

confidence: 99%

Section: Machine Learning: Random Forest Regressionmentioning

confidence: 99%

See 3 more Smart Citations

Plain-Woven Areca Sheath Fiber-Reinforced Epoxy Composites: The Influence of the Fiber Fraction on Physical and Mechanical Features and Responses of the Tribo System and Machine Learning Modeling

Subramanyam,

Kotikula,

Bennehalli

et al. 2024

ACS Omega

Self Cite

View full text Add to dashboard Cite

Recent studies focus on enhancing the mechanical features of natural fiber composites to replace synthetic fibers that are highly useful in the building, automotive, and packing industries. The novelty of the work is that the woven areca sheath fiber (ASF) with different fiber fraction epoxy composites has been fabricated and tested for its tribological responses on three-body abrasion wear testing machines along with its mechanical features. The impact of the fiber fraction on various features is examined. The study also revolves around the development and validation of a machine learning predictive model using the random forest (RF) algorithm, aimed at forecasting two critical performance parameters: the specific wear rate (SWR) and the coefficient of friction (COF). The void fraction is observed to vary between 0.261 and 3.8% as the fiber fraction is incremented. The hardness of the mat rises progressively from 40.23 to 84.26 HRB. A fair ascent in the tensile strength and its modulus is also observed. Even though a short descent in flexural strength and its modulus is seen for 0 to 12 wt % composite specimens, they incrementally raised to the finest values of 52.84 and 2860 MPa, respectively, pertinent to the 48 wt % fiber-loaded specimen. A progressive rise in the ILSS and impact strength is perceptible. The wear behavior of the specimens is reported. The worn surface morphology is studied to understand the interface of the ASF with the epoxy matrix. The RF model exhibited outstanding predictive prowess, as evidenced by high R-squared values coupled with low mean-square error and mean absolute error metrics. Rigorous statistical validation employing paired t tests confirmed the model's suitability, revealing no significant disparities between predicted and actual values for both the SWR and COF.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Machine Learning: Random Forest Regressionmentioning

confidence: 99%

Section: Machine Learning: Random Forest Regressionmentioning

confidence: 99%

See 2 more Smart Citations

Plain-Woven Areca Sheath Fiber-Reinforced Epoxy Composites: The Influence of the Fiber Fraction on Physical and Mechanical Features and Responses of the Tribo System and Machine Learning Modeling

Subramanyam,

Kotikula,

Bennehalli

et al. 2024

ACS Omega

Self Cite

View full text Add to dashboard Cite

show abstract

“…As a result, a distinct surface metamorphic layer is formed, as depicted in Fig. 13, comprising a loose layer, recast layer, and heat-affected layer [27]. Post the wire-cutting procedure, specimens were meticulously prepared by affixing them with A and B glue in a 5:3 volume ratio.…”

Section: Metallographic Structurementioning

confidence: 99%

Experimental investigations on WEDM performance characteristics of FGH99 powder metallurgy superalloy

Qian,

Zhao,

et al. 2024

Preprint

View full text Add to dashboard Cite

In response to challenges posed by conventional machining methods, an exploratory experimental investigation was conducted to assess the effectiveness of wire electric discharge machining (WEDM) in fabricating turbine disc slots from powder metallurgy superalloy FGH99. The present study conducted a single-factor experiment on WEDM using FGH99 alloys, with the aim of investigating the impact of different WEDM parameters on the material removal rate, surface roughness, and surface morphology. Additionally, the analysis encompassed the characterization of both the heat-affected zone and plastic deformation zone on the workpiece surface. Results show that there is a positive correlation between surface roughness and variables such as pulse width, pulse gap, servo voltage, and wire tension. Conversely, wire speed exhibits an inverse relationship with surface roughness. Notably, the impact of pulse width on surface roughness surpasses that of other parameters. Metallographic examination reveals that under the selected process parameters, while ensuring the efficiency of material removal, superior machining quality was achieved. The absence of conspicuous recast layers on the workpiece surface, with an average thickness of the plastic deformation layer measured at 3.97 µm, indicating favorable processing outcomes. The aforementioned findings provide valuable experimental insights into the WEDM process of FGH99 alloys and offer foundational engineering data for the fabrication of turbine disc slots made from this material.

show abstract

Material extrusion‐based additive manufacturing of sandwiched acrylonitrile butadiene styrene/glass fibers composites: Machine learning approach to model tensile stress

Kumar,

Kumar

2024

Polymer Composites

View full text Add to dashboard Cite

Additive manufacturing (AM), also recognized as 3D printing, has gained significant attention in various industries for its potential to revolutionize production processes. One critical aspect of AM is ensuring the quality and performance of printed parts, particularly concerning mechanical properties like tensile stress. In the present work, the effect of process variables on 3D‐printed acrylonitrile butadiene styrene (ABS)/Glass fiber composite materials was explored. The machine learning approach, classification and regression trees (CART) algorithm, was used to predict tensile stress in ABS/Glass fiber composite materials based on predictor variables such as layer thickness, nozzle temperature, bed temperature, and infill density. The objective is to develop an accurate and interpretable model that captures the relationships between these variables and tensile stress. The model is evaluated using performance metrics such as R‐squared, mean absolute deviation (MAD), and root mean squared error (RMSE) on both training and test datasets. From results the highest tensile stress of 39 MPa was achieved at nozzle temperature of 250°C, bed temperature of 80°C and infill density at 60%. The CART model predicts the most influencing parameter as infill density followed by nozzle temperature and bed temperature.Highlights Novel material is manufactured by sandwiching glass fiber in ABS layers. FDM process for novel material is optimized based upon process parameters. SEM analysis reported good interlayer adhesion with few micro‐porosities. Modeling of tensile stress by classification and regression tree algorithm. CART model predicts significant impact of infill density on tensile stress.

show abstract

Exploring the intricacies of machine learning-based optimization of electric discharge machining on squeeze cast TiB2/AA6061 composites: Insights from morphological, and microstructural aspects in the surface structure analysis of recast layer formation and worn-out analysis

Cited by 21 publications

References 87 publications

Plain-Woven Areca Sheath Fiber-Reinforced Epoxy Composites: The Influence of the Fiber Fraction on Physical and Mechanical Features and Responses of the Tribo System and Machine Learning Modeling

Plain-Woven Areca Sheath Fiber-Reinforced Epoxy Composites: The Influence of the Fiber Fraction on Physical and Mechanical Features and Responses of the Tribo System and Machine Learning Modeling

Experimental investigations on WEDM performance characteristics of FGH99 powder metallurgy superalloy

Material extrusion‐based additive manufacturing of sandwiched acrylonitrile butadiene styrene/glass fibers composites: Machine learning approach to model tensile stress

Contact Info

Product

Resources

About