Mechanical strength estimation of ultrafine-grained magnesium implant by neural-based predictive machine learning
Graphical abstract
Introduction
Mg-based alloys are remarkably lightweight metals with a density that is close to that of natural bone and much less than that of conventional Ti-based implants [1]. They have devoted considerable attention owing to their similar mechanical response to those of bone [2], nevertheless, the widespread applications of these biodegradable alloys are still hindered because of their extreme degradation levels and the resultant loss in mechanical strength.
More recently, numerous reports have investigated the potential role of SPD methods in fabricating UFG Mg alloys for fracture fixation and orthopedic implants. Besides, different mathematical modeling approaches were developed, however, some drawbacks should be addressed through new manufacturing strategies and modeling methods [3]. In view of the assorted parameters of the SPD procedures with nonlinear interlinkages, a precise estimation of the material properties dictates plenty of complex interactions that are not commercially viable [4]. Therefore, mechanical properties modeling of the orthopedic implants, for instance, UFG Mg alloys provides the basis for analyzing and adjusting the biomechanical behavior under machine learning (ML) perspective [5].
Despite the importance of developing biomechanical findings based on the ML algorithms, most previous reports have concentrated only on the manufacturing processes and descriptions of the processed alloys [6]. Therefore, this work gives for the first time, a comparative framework between neural-based models (ANFIS and SVM) and genetic-based approaches (GEP and GEP), to forecast the mechanical properties of the severely deformed biodegradable Mg alloy using limited datasets. It should be noted that this is a newborn field of research with lots of challenges and open questions which will be further developed in future work to address accuracy, scalability, and process efficiency.
Section snippets
Material and methods
The ZK60 Mg alloy was acquired as an extruded bar, followed by machining into the desired shape (cylindrical). The details of the manufacturing process are given elsewhere [4]. Fig. 1a-d shows a schematic illustration of the PTCAP tools and process, the experimental output, as well as the dimension of the tensile test specimens. Here, ANFIS was employed and adjusted membership functions and the relevant factors towards the target data sets using the MATLAB software [7].
The ANFIS structure was
Results and discussions
In the present work, the parameters of ANFIS and SVM were learned from the training dataset, where the split was adjusted to 50/50 for training and testing [11]. The mechanical features of the severely deformed biodegradable ZK60 Mg alloy extracted from our previous findings [4] compared to the literature values [12], [13], [14], [15], [16], are tabulated in Table S1 (Supplementary data 1). From this table, ZK60-3 processed alloy exhibited higher mechanical strength than those of the
Conclusions
Four different genetic- and neural-based models were developed to predict the mechanical properties of the severely deformed biodegradable ZK60 Mg alloy. The predictions from the ANFIS and SVM models showed a satisfactory level of forecasting of the mechanical response of the UFG Mg-based implant, with an |R| higher than 0.8 for all simulations. By comparing the genetic-based models with neural-based models, it was found that the GEP model possessed the lowest error values and it was more
CRediT authorship contribution statement
Maohua Li: Writing - original draft, Investigation. Mohsen Mesbah: Conceptualization, Methodology. Alireza Fallahpour: Software, Validation. Bahman Nasiri-Tabrizi: Methodology, Data curation, Writing - original draft. Baoyu Liu: Supervision, Writing - review & editing.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References (16)
- et al.
Improved biodegradable magnesium alloys through advanced solidification processing
Scr. Mater.
(2020) - et al.
Parallel tubular channel angular pressing (PTCAP) as a new severe plastic deformation method for cylindrical tubes
Mater. Lett.
(2012) - et al.
Characterization of indigenously coated biodegradable magnesium alloy primed through novel additive manufacturing assisted investment casting
Mater. Lett.
(2020) - et al.
An expert system of price forecasting for used cars using adaptive neuro-fuzzy inference
Expert Syst. Appl.
(2009) - et al.
Microstructure, texture evolution, mechanical properties and corrosion behavior of ECAP processed ZK60 magnesium alloy for biodegradable applications
J. Mech. Behav. Biomed.
(2014) - et al.
ZK60 alloy processed by ECAP: Microstructural, physical and mechanical characterization
Mater. Sci. Eng. A
(2014) - et al.
Microstructure and mechanical properties of ZK60 alloy processed by two-step equal channel angular pressing
J. Alloys Compd.
(2010) - et al.
Improvement of mechanical properties of magnesium alloy ZK60 by integrated extrusion and equal channel angular pressing
Acta Mater.
(2011)
Cited by (9)
Ultrafine-grained Mg alloy: Preparation, properties, design strategy
2024, Journal of Materials Research and TechnologyInterpretable machine learning-based analysis of mechanical properties of extruded Mg-Al-Zn-Mn-Ca-Y alloys
2023, Journal of Alloys and CompoundsApplication of logistic regression, support vector machine and random forest on the effects of titanium dioxide nanoparticles using macroalgae in treatment of certain risk factors associated with kidney injuries
2023, Environmental ResearchCitation Excerpt :Numerous health-promoting properties of marine algae, such as anti-inflammatory, antibacterial, and anticancer properties, have been discovered via the evaluation of their bioactivities in vitro and in vivo (Salamati et al., 2020; Kamyab et al., 2018b; Cheah et al., 2018). Furthermore, marine algae are a rich source of prebiotics, which are complex carbohydrates (Lopez-Santamarina et al., 2020; Mosleh-Shirazi et al., 2022; Kamyab et al., 2014c). Additionally, marine algae are a good source of lipids, dietary fiber, omega-3 fatty acids, minerals, proteins, essential amino acids, polysaccharides, and vitamins A, B, C, and E. Studies showing the anti-inflammatory and antioxidant actions of sulfated polysaccharides (PLS) have led to the conclusion that these compounds are extremely advantageous physiologically (Bilal and Iqbal, 2019; Tavakkoli et al., 2022b; Kamyab et al., 2016b; Putri et al., 2011).
Machine learning-based design of biodegradable Mg alloys for load-bearing implants
2023, Materials and DesignCitation Excerpt :In the field of Mg, both corrosion current density and compressive yield strength of Mg-Ca-Zn alloys were predicted by ANN with an architecture of 4-hidden nodes [17]. Mechanical strength of ultrafine-grained ZK60 implant was estimated by neural-based models such as adaptive neuro-fuzzy inference system, radial basis function and multilayer perceptron [7,8]. It was also found through feed forward neural networks with back propagation algorithm that CO2 and the composition of the buffer system played a crucial role in Mg corrosion [25].
Application of machine learning in key properties of medical metal materials
2024, Chinese Journal of Tissue Engineering Research