Innovative concept of severe plastic deformation manufacturing system to enhance classic equal channel angular drawing method: a preliminary study for flexible manufacturing systems looking to Industry 4.0

Equal channel angular drawing (ECAD) represents the most successful severe plastic deformation (SPD) technique for continuous industrial manufacturing of longer wires, with a constant cross section, characterized by smaller diameters and better mechanical properties (i.e. high strains and hardness) mainly attributed to the grain size refinement. In this paper an advanced innovative concept to impose SPD, on commercial 1370 pure aluminium wires (Al 99.7%), is proposed to improve the flexibility of the classic manufacturing method of ECAD by controlling and regulating process parameters in real time to obtain several combinations of mechanical properties and to increase manufacture productivity. This paper provides a preliminary analysis of mechanical and microstructural changes occurring during ECAD process and, preserving the principle of the ECAD method, describes an innovative concept of plastic deformation showing the potential improvements to practice. The proposed manufacturing system has been validated by finite element analysis (FEA) implementing a flow stress empirical model that includes the influence of the grain size change, for the material behaviour and two customized user-subroutines for predicting grain refinement and hardness variation. The study demonstrates the possibility to renew the classic industrial techniques within an Industry 4.0 ecosystem.


Introduction
Industry 4.0 (I4.0), also considered the fourth industrial revolution, refers to the trend of industrial automation that integrates some new production technologies, to improve working conditions, create new business models and increase the productivity and production quality of plants. Some authors focused their attention on the smart operators in I4.0 [1] and on smart factory [2]. Ahuett-Garza carried out a wide analysis of the trends of habilitating technologies for I4.0 and Smart manufacturing [3]. Stock et al. considered I4.0 as an enabler for a sustainable progress, developing a qualitative assessment of the paradigm ecological and social potential [4], while Tao et al. were mainly focused on simulation and, in particular, on digital twin-driven product design and manufacturing [5]. I4.0 is strictly related to the concept of smart factory, which is based on three main pillars: i. Smart production: new or reshaped manufacturing technologies that create interaction between all the elements present in the production system and collaboration between operator, machines and tools; ii. Smart service: all the intelligent and Internet of things (IoT), cyber-physical systems (CPSs) and Internet of services (IoS) techniques that allow systems to be integrated and to connect, in a collaborative way, companies (supplier-customer) with each other and with external structures (roads, hubs, waste management, etc.); iii. Smart energy: careful control of energy consumption, by developing efficient systems able to reduce energy waste.
The synergy between I4.0 and sustainable manufacturing can reform the current consumption of natural resources and dangerous emissions, implementing sustainable initiatives in monitoring and controlling greenhouse effect, carbon emissions and energy consumption, reducing waste and cost and increasing the prosperity of the environment, economy and society in its complex.
Bonilla et al. [6] published a scenario-based analysis of the impact and challenges to investigate I4.0 and its sustainability implications; a more detailed analysis was carried out by Oláh et al. a couple of years later [7]. On the other hand, other researchers pointed out some criticisms in the implementation of I4.0 paradigm to increase sustainability: Díaz-Ramírez et al. discussed an environmental assessment of electrochemical energy storage device manufacturing [8], Duflou et al. depicted frame on energy and resource efficient manufacturing [9], Narwane et al. introduced their view about the barriers in sustainable I4.0, focusing on the footwear industry [10]. Smart manufacturing represents the core of the new industrial stage, improving manufacturing processes in terms of automation (i.e. robots, machine-tomachine communication (M2M)), vertical integration (i.e. programmable logic controllers (PLC); enterprise resource planning (ERP)), traceability (i.e. raw materials, final products), flexibility (i.e. additive manufacturing (AM), flexible lines), virtualization (i.e. simulation and modelling, artificial intelligence (AI)) and energy management [11][12][13]. For this reason, the fourth industrial revolution is rooted in the concept of smart manufacturing, considering it the beginning and the main purpose of I4.0. Despite the growth of advanced and smart technologies is one of the relevant trends in both industrial and academic research fields, the investigation of barriers related to the implementation of I4.0 represents a central point of view for a number of researchers all over the world [14][15][16][17][18]. Looking at the manufacturing process, with a particular interest on the production of fine wires, ECAD technique represents a traditional and most successful SPD method for continuous production of ultra-fine-grained material with increased strength for the grain size reduction [19][20][21][22]. In particular, during ECAD manufacture a large plastic deformation is imparted to the workpiece (typically a round bar) by drawing the material through a die consisting of two channels of equal cross section (Fig. 1).
In the perspective of the fourth industrial revolution, focusing on the industrial automation, the conventional ECAD manufacture can be moved on the trending idea of I4.0 by integrating the predictive capability of the numerical simulations with emerging concepts of renewing the traditional manufacturing methods. Hence, with the aim to give a further impulse to the implementation of I4.0 technologies in industrial production, overcoming part of the literature investigated barriers, the present work describes a preliminary analysis of an innovative concept of SPD system. By preserving the principle of ECAD method, an innovative procedure is proposed to restructure the classic manufacturing method of ECAD usefulness for continuous production In particular, experimental tests of the conventional ECAD process, on commercial 1370 pure aluminium (99.7% Al) wires, were firstly conducted and analysed. Hence, a corresponding finite element (FE) model of the process was developed including an additive strengthening equation that considers the influence of the grain refinement and dislocation density on the material flow stress behaviour, and two user-subroutines for the grain size and hardness change prediction. Once validated the effectiveness of the proposed numerical tool and the robustness of the material model under SPD condition, by comparison with experimental data, they were both used to study and analyse the new SPD system looking at I4.0. After a FEA, the rule of the traditional ECAD die was renewed through an advanced technology able to significantly increase the flexibility of the process improving its productivity (i.e. customized components with different mechanical and microstructural properties) and drastically reducing the waste usually generated for manufacturing the ECAD die by the classic machining operations (turning or milling).

Material and methods
Commercial 1370 pure aluminium rods (Table 1) with an initial diameter of 9.50 mm were firstly manufactured by a multiple-pass cold drawing process (15 passes at room temperature with a drawing speed of 25 m/sec and wet lubrication conditions-by mineral oil-able to maintain a process temperature of 75 ± 5 °C) till to reach a wire of 2 mm of diameter with a section reduction of 95.6% and a total drawing strain of 3.11.
Then, on the same manufacturing line, an ECAD process, with an inner die angle Φ = 140° and outer die angle   ψ = π-Φ, was performed at room temperature preserving the cross section of the wire (Fig. 3a). By an MTS Criterion Model 45 testing machine, were performed uniaxial tensile tests at room temperature on the aluminium initial rods, drawn wires and ECAD manufactured wires: three replicas per sample were carried out, for a total of nine tests. The results reported in Fig. 3b and c show a sequential material strengthening, due to the single contribution of drawing manufacture and ECAD process, with a corresponding total true strain reduction of 64% [23]. The flow stress rise was mainly induced by dislocation density increase [24, 25] and Hall-Petch (H-P) strengthening effect [26,27] due to the grain refinement ( Fig. 4), showing the importance of the imposed plastic deformation in controlling the mechanical performance of the manufactured material.
Afterwards, 9 specimens (three aluminium rods, three drawn wires and three ECAD processed wires) were mounted into a resin holder for the grain size analysis and the hardness measurement of the cross section. Mechanical polishing followed by etching (Keller's reagent: 92 ml of distilled water, 6 ml of nitric acid, 2 ml of hydrochloric acid, 2 ml hydrofluoric acid) was conducted before the metallographic analysis. The cross sections of the specimens were investigated by an optical microscope for microstructural analysis, while the micro-hardness measurement (HV 0.01 ) was performed by an instrumented micro-nano-indenter. Figure 4 shows the microstructure of the initial rod, multipass drawn wire and ECAD manufactured wire, all characterized by grains with the same equiaxial shape (grain boundaries represented by black and thicker lines) but different size. In fact, an average grain size of about 58 µm was found for the initial rod, with a reduction of 81% when analysing the cross section of the multi-pass drawn wire featured by a size of 11 µm, and a final reduction to 6 µm on the final wire after applied the ECAD manufacture.
The experimental outcomes highlight the effectiveness of the ECAD, by imposing an intense plastic deformation on the material, in achieving microstructural evolution which results in grain size reduction, for the dynamic recrystallization [28], and dislocation density increases, both leading to a material strengthening improvement. In fact, considering the experimental outcomes (material flow stress and grain size change) and the additive strengthening equation (Eq. 1), it is possible to study the influence of the grain size and dislocations on the material strengthening.
Following the additive strengthening model (Eq. 1), the material flow stress at a particular strain σ(ε) can be considered as influenced by three aspects: the frictional stress, the grain size change and the dislocation density evolution.
(1) where σ 0 is the frictional stress (a constant which includes contributions from solutes and particles but not from dislocations), σ gs is the contribution from the grain size-related strengthening and σ disloc is the contribution from the dislocation-related strengthening.
σ gs can be written as: where k 1 is a constant and d is the average grain size. By this term the material behaviour is influenced by the dynamic recrystallization that significantly modifies the microstructure of the material (i.e. grain refinement, Fig. 4) resulting in material strengthening ( Fig. 3b and 3c) according to the H-P effect.
σ disloc can be written as: where M is the Taylor factor, α is a coefficient, µ is the elastic shear modulus, b is the length of the Burger vector and ρ is the dislocation density. By Eqs. 2 and 3, the additive strengthening equation can be written as: All the constants characterizing the material behaviour are listed in Table 2.
By considering the strengthening equation (Eq. 4) for a reached strain of 0.02 (Fig. 3b) and the experimental grain size change (Fig. 4), it is possible to analyse the dislocation density evolution during manufacturing (Table 3). Table 3 shows as both the drawn and the ECAD procedures have an evident influence on the dislocation density variation. In particular, comparing the initial rod firstly with the multipass drawn wire and then with the ECAD manufactured wire a respective increase of about 4 and 6 times was found for the dislocation density ρ. Now, considering Eq. 4, and the data listed in Tables 2 and 3, it is possible to evaluate the single contribution of the grain size and dislocation density on the material strengthening increase, as reported in Table 4.
It is evident the influence of both the grain size and the dislocation density on the material flow stress behaviour, together making up 90% of the material strengthening. Furthermore, Table 4 shows as the more significative part of the material strengthening is represented by the dislocation strengthening stress σ disloc with an incidence of about 80%. This analysis allows to have a clear idea concerning the single contribution of the grain size and dislocations on the material strengthening during both the investigated manufacturing processes.
Concerning the micro-hardness, ten indentations per sample (three aluminium rods, three drawn wires and three ECAD processed wires) were performed, for a total of 90 tests, and the average value was measured (Fig. 5) resulting in a homogeneous micro-hardness distribution along the whole cross section: a reduced micro-hardness measurement variation was found for each sample (Fig. 5c). As a consequence of the grain size reduction, a material hardness enhancement was observed (according to the H-P effect [29]) with a previous increase from 39 to 53 HV 0.01 after the multi-pass drawing, and a further increase of 9 HV 0.01 after performing ECAD process.

Numerical model
The commercial FE software SFTC DEFORM-3DTM has been used to simulate the ECAD manufacture of commercial 1370 pure aluminium drawn wire. The workpiece was modelled as a plastic object with 52,000 isoparametric tetrahedral elements, while for the die a rigid model with 60,000 elements was considered.
The additive strengthening equation, based on the single contribution of the frictional stress, the grain size change and the dislocation density evolution, was implemented for the material flow stress (Eq. 4). All the material parameters are listed in Table 2. A regression approach, by considering the tensile tests of both the multi-pass drawn and ECAD manufactured wires, was implemented to determine the value of the dislocation density as a function of the strain [30], Eq. 5.
The physical events influencing the mechanical properties of the material were predicted by two customized usersubroutines implementing a continuous dynamic recrystallization (CDRX) model for grain size reduction [24, 33,34] and H-P relation for the hardness change. Due to the nature of the investigated material, CDRX represents the main physics metallurgical phenomenon [35][36][37], and hence, a continuum mechanical model was implemented for predicting the grain size, Eq. 6.
where d is the recrystallized grain size, d 0 the initial grain size, d f the saturation grain size, and k X and C X are parameters describing the recrystallization evolution with increasing plastic deformation. The McCauley brackets < > indicate that recrystallization phenomena will occur when the effective strain ε p eff will reach the threshold value ε p c . For the ECAD process strain ε p eff was considered the model developed by Y. Iwahashi et al. [38] (Eq. 7).
where Φ and ψ are the ECAD die angles (Fig. 1) and N pass is the number of ECAD passes. While the parameters k X and c X and the critical strain leading to CDRX were set 3.8, 2 and 0.1 according to [35], respectively. Finally, the hardness variation was calculated as an inverse function of the recrystallized grain size, according to the H-P equation (Eq. 8): where C 0 and C 1 are two material constants while d represents the average grain size. The value of C 0 and C 1 , were determined through the previously measured values of the material hardness and grain size of both initial aluminium rods and drawn wires and were set equal to 28.2 and 82.2, respectively.

Fe validation
To validate the robustness of the FE model and the user-subroutines, the predicted hardness change and grain refinement, due to ECAD process, were compared with the previously shown experimental data. According to the experimental outcomes (final grain size and hardness, respectively, of 6 µm and 62 HV 0.01 ), the numerical results (Fig. 6) show an acceptable difference with a predicted microstructure of 5 µm and a corresponding hardness of 65 HV 0.01 . This deviation results from the accuracy of both experimental measurements and numerical procedure (empirical laws and equations, setting and calibration of the numerical constants). The stable and uniform predicted data reported in Fig. 6 confirm the robustness of the customized FE model and user-subroutines in predicting the effects due to the SPD induced by ECAD procedure. In detail, when the effective strain ε p eff reaches a threshold value, recrystallization phenomena occur in the deformed material resulting in a new fine-grained structure that leads to mechanical and microstructural changes influencing the material behaviour (Eq. 4).
In agreement with the experimental evidences, the deformation imposed by ECAD die leads in the material an average grain size reduction of about 50% which corresponds to a hardness enhancement for the inverse function of H-P relation (Eq. 8), validating the influence of the imposed strain in controlling the mechanical properties modification and the performance of the manufactured components. Validated the robustness of the material empirical law and of the two user-subroutines, and considering Iwahashi's strain principle [38], it was possible to analyse the innovative manufacturing method (Fig. 2), to impose SPD in continuous manufacturing processes, more flexible than conventional ECAD die and able to impart on the sample the same equivalent shear strain necessary to achieve grain refined structure for the recrystallization phenomena.
After simulated both the traditional ECAD technique and the innovative SPD method, implementing the previously validated numerical tool, the FE analysis (Fig. 7) demonstrated the possibility of the proposed SPD technique to reach and overcome the strain obtained by classic ECAD approach and to cover a wide range of imposed severe strain by setting the parameters h and d (Fig. 7b). Defining the ratio h/d = z, the imposed strain is a direct function of z: the increase of z results in higher imposed process strain due to the reduction of the inner angle Ф of the ECAD die (Eq. 7). The innovative method preserves the criterion of the conventional equal channel angular process, but renovates its appliance, by proposing a new concept in the manufacturing process able to improve the suppleness of the process and its control giving a potential and advanced contribution towards smart manufacturing and Industry 4.0.
Moreover, the FE model allowed to investigate the deformation homogeneity by strain analysis. In detail, the degree of the strain homogeneity can be evaluated by Eq. 9 [39]: where Ci is the strain inhomogeneity index, Max ε is the maximum equivalent strain, Min ε is the minimum equivalent strain and Avg ε is the average equivalent strain: a lower value of the index means better degree of strain homogeneity.
After simulated both the investigated procedures, implementing the same process conditions (Fig. 8a), the strain inhomogeneity index Ci was determined along the cross section of the wires by the slicing tool available in DEFORM 3DTM. The measurements were taken on three different positions of both the manufactured wires (C i1 , C i2 , C i3 - Fig. 8b) and the average value was considered. A lower value of the index was found for the proposed innovative SPD method (Ci = 0.56) showing a better degree of strain homogeneity when compared to the conventional ECAD process (Ci = 0.72). The developed SPD system looks at the old problems of ECAD method, representing a progress to the manufacturing techniques and an advanced solution for improving mechanical material performances. In fact, traditional ECAD die operates as a rigid system offering only one single shape and not allowing the possibility to regulate the imposed strain at varying the material properties (i.e. ferrous-nonferrous material, chemical composition, temperature, hardness, strength) and its geometry: every single ECAD process requires a (9) C i = Max − Min Avg Fig. 6 Numerical simulation: a grain size (µm) and b hardness (HV 0.01 ) variation during ECAD process. c experimental and numerical comparison of the hardness along the cross section specific designed die (i.e. cross-sectional diameter, inner and outer angles, etc.) manufactured by classic milling and turning machining with a major energy consumption and waste production for chip generation. Moreover, the friction between the sample and the ECAD die, during the draw, not only reduces the regular flow of the process, moving the manufacture towards the material failure (evident increase of the strain among the two curves- Fig. 8b), but increases the material temperature leading to the grain grow phenomena instead of the expected grain refinement. Furthermore, it is not possible to control and set the lubrication conditions directly into the die: making inner channels for lubricant/ coolant access could create edges that affect the surface integrity of the drawn material and its mechanical performance for the presence of external defects. The innovative technology represents an important advance in continuous manufacturing processes, giving to the classic ECAD method the flexibility to renew the process in the Industry 4.0 perspective. In fact, the proposed concept, based on two rotating parts, improves the flow of the material during the process, reducing the friction and the possibility of sample failure, and allows to manufacture samples with different diameters by using the same set-up, Fig. 8. Moreover, the new system offers the possibility to control the lubrication conditions at the tool-sample contact zone reducing the process temperature (raising for the high induced strain) and making to occur the recrystallization phenomena that lead to the grain size reduction. The possibility to set different diameters (Fig. 9) and to regulate the parameters "h" and "d" (Fig. 7b) allows to control the imposed strain during the process and therefore to obtain different materials microstructures (i.e. grain size).
Hence, at varying the above cited parameters, it is possible to get different combinations of mechanical properties (i.e. material strength, hardness change, threshold stress for crack evolution) increasing the manufacture productivity: materials with different strength, hardness and processability characteristics. Finally, interconnecting the advanced system with a FE method tool, able to predict aspects as microstructural changes [24], hardness variation, material strength, deformation evolution, process temperatures, etc., it is possible to control and monitor the process sharing the collected data in real-time, creating a proper digital twin. This approach represents an intelligent system able to take decisions in an automated way, setting manufacturing process conditions (i.e. lubrication, drawn-speed, imposed strain) without human involvement and making the operator a supervisor instead of a problem solver.  In this paper a preliminary analysis of an advanced innovative SPD technique, for continuous production of aluminium wires, was proposed. In particular, experimental tests of the conventional ECAD process were firstly conducted, showing in the manufactured wires an evident increase of the material strengthening, a grain size reduction of the 90% and a corresponding hardness increase of the 60%. After a comprehensive analysis concerning the influence of the grain size and the dislocations on the increase of the material strengthening, the experimental outcomes were used to validate the robustness of a predictive FE model including an additive strengthening equation that considers the influence of the grain refinement and dislocation density on the material flow stress behaviour, and two customized user-subroutines for the grain size and hardness change prediction. Preserving ECAD principle and implementing the previously validated numerical tool, it was illustrated, by a FEA, a new technology for SPD method able to improve the flexibility of the conventional ECAD process and its control. The analysis showed the possibility to imprint to the sample the same equivalent strain, usually imposed by ECAD die, leading to microstructural changes and mechanical properties variation. Moreover, the innovative set-up allowed to regulate on line the imposed strain at varying the material properties and its geometry, getting different combinations of mechanical properties in terms of strengthening, hardening and processability. A detailed analysis of the strain distribution, along the cross section of the manufactured wires, showed for the innovative SPD method a better degree of strain homogeneity when compared to the conventional ECAD process. Then it was highlighted the possibility of the shown approach to manufacture wires with different diameters by using the same set-up overcoming the rigid set-up of the conventional ECAD die designed only for one geometry and not allowing to regulate the inner angle (Fig. 1). Considerations on the friction showed as the new system lets to control the lubro-cooling flow directly at the tool-sample contact zone resulting in a better lubrication and manufacturing conditions: regular flow of the wire, reduced risk of the manufacture to move towards material failure and reduced material temperature increase. The key elements emerging from this study demonstrate a new contribution to practice, steering the classic manufacturing processes to Industry 4.0 by i. adopting new technologies (FE method, flexible lines) to improve productivity and allow smart process, ii. fitting out manufacturing plant with IoT to network humans and technologies, and iii. implementing software algorithms (AI) to collect and analyse data thus providing automated choices and avoiding human intervention.
Author contributions All authors contributed to the study conception and design. Material preparation, data collection, numerical modeling and analysis were performed by SC. All authors wrote the first draft of the manuscript and commented on previous versions of the manuscript. LF reviewed, edited and supervised the manuscript. All authors read and approved the final manuscript.
Funding Open access funding provided by Università della Calabria within the CRUI-CARE Agreement.

Conflict of interest
The authors have no conflicts of interest to declare that are relevant to the content of this article.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.