Manufacturing Process Research Articles

Surface Treatment of Additively Manufactured Polyetheretherketone (PEEK) by Centrifugal Disc Finishing Process: Identification of the Key Parameters.

Polyetheretherketone is a promising material for implants due to its good mechanical properties and excellent biocompatibility. Its accessibility to a wide range of applications is facilitated by the ability to process it with an easy-to-use manufacturing process such as fused filament fabrication. The elimination of disadvantages associated with the manufacturing process, such as a poor surface quality, is a main challenge to deal with. As part of the mass finishing process, centrifugal disc finishing has demonstrated good results in surface optimization, making it a promising candidate for the post-processing of additively manufactured parts. The objective of this study is to identify the key parameters of the centrifugal disc finishing process on the waviness of additively manufactured PEEK specimens, which has not been investigated previously. The waviness of the specimen was investigated by means of confocal laser scanning microscopy (CLSM), while weight loss was additionally tracked. Six parameters were investigated: type, amount and speed of media, use of compound, amount of water and time. Type of media, time and speed were found to significantly influence waviness reduction and weight loss. Surface electron microscopy images demonstrated the additional effects of deburring and corner rounding. Results on previous studies with specimens made of metal showed similar results. Further investigation is required to optimize waviness reduction and polish parts in a second post-processing step.

Influences of different parameters on selected properties of gears for robot-like systems.

For a drive unit for axes of robots and robot-like systems (RLS) usually a motor-gearbox arrangement is chosen due to its high-power density. The combination of a high-ratio gearbox and a high-speed electric motor ensures a very compact and efficient design of the drive train. The transmission properties primarily determine the properties of the axes and the whole robot system. Robots and RLS use various types of high-ratio precision gearboxes based on different operating principles. Due to the different operating principles, it is difficult to describe comparable properties across all different types. In addition, there are many influences on the properties which significantly determine their shapes and values. These influencing parameters are insufficiently documented and are often poorly accessible for profound comparability and further consideration. In this paper, an overview of the properties of robot gearboxes is given. Based on these properties, different robot gearboxes can be systematically evaluated and compared to one another. The properties are influenced by various design, operating or manufacturing factors such as the gearbox size, the operating torque and speed or the manufacturing process. In a further step, these influences on the most relevant properties, efficiency and stiffness, are determined and systematically evaluated. This evaluation is based on the specification data of various robot gearbox manufacturers. The properties efficiency and stiffness show a dependency on the gearbox size, the operating torque, speed as well as the ambient temperature and on the transmission ratio. The shown procedure can also be adapted to other properties.

Improving the product of strength and ductility and corrosion resistance by adding He shielding gas in the CMT additive manufacturing process of Ti6Al4V

The cold metal transfer additive manufacturing (CMTAM) of titanium alloys presents a promising solution for the efficient fabrication of large titanium alloy armor components. In this study, different proportions of He gas were added to the welding shielding gas used in the CMTAM of Ti6Al4V, aiming to improve forming quality, the product of strength and ductility and corrosion resistance. Simultaneously, the influence of different He contents on the microstructure, mechanical properties, and corrosion behavior of CMTAM of Ti6Al4V was explored. The research revealed that during the process of CMTAM of Ti6Al4V, as the proportion of He as the shielding gas increased, the arc voltage increases, and the arc becomes more stable. Meanwhile, the wettability of the weld pool improves, resulting in a more spread out weld pool. The fabrication precision of the built material also improved. Microstructural observations indicate that the prior-β grain size gradually increased with the increase in He proportion, the heat input for additive manufacturing also increased，and the martensitic α’ transformed into a basketweave α+β structure. The length and width of the α phase increased, the texture intensity decreased, and the Schmid factor increased. Tensile tests were conducted on samples with different He proportions showed that as the He proportion increased, within the component resulted in a noteworthy enhancement in the product of strength and ductility increase exceeding 50 %. This suggests that He as the shielding gas addition can effectively optimize the strength-plasticity balance in additive manufactured titanium alloys without the need for post-processing heat treatment. Electrochemical tests were conducted in a 3.5 wt% NaCl solution revealed that the addition of He gas significantly improved the corrosion resistance of Ti6Al4V. Specifically, the corrosion resistance was optimal when the shielding gas composition was 50 %He+50 %Ar, far surpassing that of cast Ti6Al4V.

Controlling Mechanical Properties of Medical-Grade Scaffolds through Electrospinning Parameter Selection.

Electrospinning (ES) is a versatile process mode for creating fibrous materials with various structures that have broad applications ranging from regenerative medicine to tissue engineering and surgical mesh implants. The recent commercialization of this technology for implant use has driven the use of resorbable electrospun products. Resorbable electrospun meshes offer great promise as temporary implants that can utilize the layer upon layer method of additive manufacturing to incorporate porosity as a function of process parameters into a scaffold structure. The interconnected porosity and feature size known to ES have previously been observed to hold great potential for simulating the natural cellular environment of soft tissue. This microstructure, proper degradation kinetics, and mechanical properties combine to provide the design basis for artificial tissue structures that could aid in not only wound healing but also true tissue engineering and regenerative medicine. While current advancement in the field is understood to be limited by material properties, the importance of optimizing mechanical properties with currently available materials should not be overlooked. This work investigated the process parameter effects and interactions that control the structure-property relationship for a range of medical-grade aliphatic polyester materials with a range of intrinsic properties. An ε-caprolactone homopolymer (PCL), l-lactide homopolymer (PLLA), and Lactoflex, a copolymer with intermediate properties relative to the homopolymers, were characterized before, during, and after the additive manufacturing process. The interacting effects of process parameters, distance to collector, and dispensing rate were shown to produce variable-density, nonwoven scaffold structures. The resorbable mesh scaffolds of PLLA, PCL, and Lactoflex demonstrated a broad range of mechanical properties (approximately 1-10 MPa ultimate tensile strength and 5-390 MPa tensile modulus). Postprocessing of scaffolds demonstrated removal of solvents and preservation of micrometer-sized features. Resorbable polymers and electrospinning can produce scaffold materials with excellent features and offer tremendous potential in the field of implantable resorbable devices.

Flt3 ligand augments immune responses to soluble PD1-based DNA vaccine via expansion of type 1 conventional DCs

DNA vaccines are prospective for their efficient manufacturing process, but their immunogenicity is limited as they cannot efficiently induce CD8+ T cell responses. A promising approach is to induce cross-presentation by targeting antigens to DCs. Flt3L can expand the number of type 1 conventional DCs and thereby improve cross-presentation. In this study, we first constructed a DNA vaccine expressing soluble PD1 and found that the therapeutic effect of targeting DCs with only the sPD1 vaccine was limited. When combined the vaccine with Flt3L, the anti-tumor effect was significantly enhanced. Considering the complexity of tumors and that a single method may not be able to activate a large number of effective CD8+ T cells, we combined different drugs and the vaccine with Flt3L based on the characteristics of different tumors. In 4T1 model, we reduced Tregs through cyclophosphamide. In Panc02 model, we increased activated DCs by using aCD40. Both strategies triggered strong CD8+ T cell responses and significantly improved the therapeutic effect. Our study provides important support for the clinical exploration of DC-targeted DNA vaccines in combination with Flt3L.

Damage analysis of solid propellants with default defects based on macro-microscopic approach

Solid Rocket Motors (SRMs) rely on solid propellants, directly impacting the energy capacity of launch vehicles. High energy density material (HEDM) propellant exhibits superior energy density owing to the incorporation of a substantial quantity of energetic particles. However, the elevated particle fill ratio makes the formation of pores more prevalent during the manufacturing process of propellants. These pores, representing default defects, are indispensable factors in the investigation of mechanical properties of propellants. This study investigated default defects effects on HEDM propellants, analyzing both microscopic and macroscopic perspectives. Utilizing micro-CT imaging, the microstructure and default defects were characterized. Subsequently, using the cohesive zone model (CZM), microscopic damage evolution was depicted with the representative volume element (RVE) model, allowing in-depth analysis. Furthermore, a damage constitutive model, based on microscopic damage, was developed, with its softening function calibrated using microscopic damage evolution characteristics. Implemented into ABAQUS via user-defined material subroutine (UMAT), this model enabled the prediction of the mechanical response of HEDM propellants with different defects. The results demonstrate that the developed model accurately predicts the mechanical responses of HEDM propellants with different defects, showcasing the potential of macro-microscopic approaches in analyzing propellants with default defects.

Amphiphilic silica monomers induced superhydrophilic and flexible silica aerogels for radiative cooling and atmospheric water harvesting

Silica aerogels have attracted extensive attention owing to their exceptional porous structures and crucial applications for sustainable environments. However, the inherent fragility and complex manufacturing process of silica aerogels present a challenge for their prolonged applications. In this study, an amphiphilic silica monomer, triethoxysilane end-capped poly(ethylene oxide) (PEO-TES), is designed and synthesized, which can co-gelled with methyltrimethoxysilane in water and gives birth to flexible and superhydrophilic silica aerogels via ambient pressure drying (APD). PEO-TES plays a dual role as both co-monomers and surfactants, effectively stabilizing the sol–gel transition, adjusting the porous structure of hydrogels, and imparting superhydrophilicity. The silica aerogel has low density (0.147 g cm−3), zero water contact angle, high elasticity (80 % compression), high solar reflectance (0.888), and high IR emissivity (0.967). These features make them robust all-day passive radiative cooling materials with 10.5 °C sub-ambient cooling capacity and atmospheric water harvesting capacities for both vapor (a water harvesting rate of 0.041 kg m−2h−1) and fog (harvesting capacity of 2.9 g g−1). The strategy provides a simple APD method for synthesizing flexible and superhydrophilic silica aerogels with controllable porous structures, passive cooling, and water harvesting performance, which may help address global climate change and water scarcity challenges.

Eddy Current Sensor Probe Design for Subsurface Defect Detection in Additive Manufacturing.

Pore and crack formation in parts produced by additive manufacturing (AM) processes, such as laser powder bed fusion, is one of the issues associated with AM technology. Surface and subsurface cracks and pores are induced during the printing process, undermining the printed part durability. In-situ detection of defects will enable the real-time or intermittent control of the process, resulting in higher product quality. In this paper, a new eddy current-based probe design is proposed to detect these defects in parts with various defects that mimic pores and cracks in additively manufactured parts. Electromagnetic finite element analyses were carried out to optimize the probe geometry, followed by fabricating a prototype. Artificial defects were seeded in stainless steel plates to assess the feasibility of detecting various flaws with different widths and lengths. The smallest defect detected had a 0.17 mm radius for blind holes and a 0.43 mm notch with a 5 mm length. All the defects were 0.5 mm from the surface, and the probe was placed on the back surface of the defects. The surface roughness of the tested samples was less than 2 µm. The results show promise for detecting defects, indicating a potential application in AM.

DETERMINATION OF CHANGES IN MECHANICAL PROPERTIES OF METAL OF FEEDWATER PIPELINES BENDS AND “HOT” STEAM OF THE SECOND POWER UNIT OF PIVDENNOUKRAINSK NPP FOR THE SECOND 100,000 h OF OPERATION

An analysis of the results of periodic testing of hardness of metal of pipelines bends of power unit No. 2 of Pivdennoukrainsk Nuclear Power Plant (PNPP) after ~100,000 and ~200,000 h of operation was carried out. It is shown that during the operation of feedwater and “hot” steam pipelines during the second 100,000 h of operation, the aging of the metal is observed, which is manifested in a partial, at least in their surface layers, uneven decrease in the strength of the metal. An increased rate of aging of the metal of pipeline bends is observed in areas where the metal was subjected to the greatest plastic deformation during the pipe manufacturing process (compressed and stretched sides of the bends). The critical parameter among the normalized mechanical properties of bending metal is the tensile strength, which is the first to reach a critical level when extrapolating the obtained data. While maintaining the rate of change in the mechanical properties of the metal at the level of the second 100,000 h of operation for post-project operation, the projected residual resource of metal bends of feedwater pipelines and "hot" steam is ~590,000 and ~380,000 h, respectively.

Molecular dynamics study of the groove-forming mechanism of mechanically scratched grating multilayer aluminum films

For a long time, the forming quality of large area diffraction gratings is restricted by the one-time coating quality and uniformity of large thickness and large area aluminum films. Large thickness coatings, low accuracy of large depth grooves and diamond tool wear problems were the main reasons for the failure of the test to inscribe large area diffraction gratings. Therefore, the use of layered coating process can effectively ensure the uniformity and consistency of large-area and large-thickness grating blanks coated with aluminum film, and improve the high precision of large depth diffraction grating slot type. In order to reveal the mechanism of the groove-forming process of multilayer aluminum films, the microstructure and mechanical properties of different aluminum films were observed. In this article, the removal behavior and mechanism of nanomechanically scratched multilayer aluminum film materials are investigated using a molecular dynamics approach, and the characteristics of multilayer aluminum films scratched into grooves are analyzed in terms of surface morphology, scratching force, structural evolution, dislocation distribution, von Mises stress and shear strain. The results indicate that in the scratching simulation process, multilayer aluminum film has a higher number of removed atoms and a larger atomic displacement compared to single layer aluminum film, which is conducive to precise shape control of the groove. The average tangential force and normal force of multilayer aluminum film show a decreasing trend compared to single layer aluminum film, which helps to scratch grooves and reduce tool wear. The total length of the dislocation lines of the multilayer aluminum film is large, and the distribution area is large, which improves the precision of scratching the grooves. Therefore, an appropriate increase in the number of aluminum film layers has a precise control effect on the groove shape. This study has certain reference significance for analyzing the complex grooving deformation law of grating aluminum film, improving the diffraction efficiency of grating aluminum film, and further improving the manufacturing process of grating aluminum film.

A Comprehensive Review on the Biogenic Amines in Cheeses: Their Origin, Chemical Characteristics, Hazard and Reduction Strategies.

Most of the biogenic amines are naturally found in fermented foods as a consequence of amino acid decarboxylation. Their formation is ascribable to microorganisms (starters, contaminants and autochthonous) present in the food matrix. The concentration of these molecules is important for food security reasons, as they are involved in food poisoning illnesses. The most frequent amines found in foods are histamine, putrescine, cadaverine, tyramine, tryptamine, phenylethylamine, spermine and spermidine. One of the most risk-prone foods are cheeses, mostly ripened ones, which could easily accumulate amines due to their peculiar manufacturing process and ripening. Cheeses represent a pivotal food in our diet, providing for nutrients such as amino acids, calcium, vitamins and others; thus, since they are widely consumed, it is important to evaluate the presence of toxic molecules to avoid consumers' poisoning. This review aimed to gather general information on the role of biogenic amines, their formation, the health issues and the microorganisms and processes that produce/reduce them, with a focus on their content in different types of cheese (from soft to hard cheeses) and the biotic and abiotic factors that influence their formation or reduction and concentration. Finally, a multivariate analysis was performed on the biogenic amine content, derived from data available in the literature, to obtain more information about the factors influencing their presence in cheeses.

Investigation of Effects of Spreading Parameters on Powder Bed Quality in Additive Manufacturing

Powder bed-based additive manufacturing processes such as laser powder bed fusion, binder jetting, and electron beam melting are commonly utilized in various critical areas such as medical, aviation, and energy. Common to all these operations, the powders are first spread onto the built platform in a layer-by-layer fashion and selectively fused or bound with a suitable method. The quality of the process depends on several process parameters, including how the powders are spread onto the built platform. Powder spreading is an important step in these operations and can affect various process outputs. In this study, the powder spreading is numerically investigated using the discrete element method to determine the effects of the layer thickness, rotation, and transition velocities selected as parameters with a powder spreader roller. The relationship between powder spreading parameters and powder volume packing fraction was investigated using the 'Hertz-Mindlin and Johnson-Kendall-Robert’s (JKR) contact model considering the interactions between particle-particle and particle-built plate, has been added to the numerical model. A Design of Experiment combined with analysis of variance (ANOVA) was utilized to gain a broader understanding on the relationship between process parameters and green density and dynamic angle of repose.

A Data-Driven Approach for Leveraging Inline and Offline Data to Determine the Causes of Monoclonal Antibody Productivity Reduction in the Commercial-Scale Cell Culture Process.

The monoclonal antibody (mAb) manufacturing process comes with high profits and high costs, and thus mAb productivity is of vital importance. However, many factors can impact the cell culture process, and lead to mAb productivity reduction. Nowadays, the biopharma industry is actively employing manufacturing information systems, which enable the integration of both online data and offline data. Although the volume of data is large, related data mining studies for mAb productivity improvement are rare. Therefore, a data-driven approach is proposed in this study to leverage both the inline and offline data of the cell culture process to discover the causes of mAb productivity reduction. The approach consists of four steps, namely data preprocessing, phase division, feature extraction and fusion, and cluster comparing. First, data quality issues are solved during the data preprocessing step. Next, the inline data are divided into several phases based on the moving window k-nearest neighbor method. Then, the inline data features are extracted via functional data analysis and combined with the offline data features. Finally, the causes of mAb productivity reduction are identified using the contrasting clusters via the principal component analysis method. A commercial-scale cell culture process case study is provided in this research to verify the effectiveness of the approach. Data from 35 batches were collected, and each batch contained nine inline variables and seven offline variables. The causes of mAb productivity reduction were identified to be the lack of nutrients, and recommended actions were taken according to the result, which was subsequently proven by six validation batches.

Revolutionizing Grain and Particle Size Measurement in Metals: The Role of Sensor-Assisted Metallographic Image Analysis.

Metallographic image analysis is vital in the field of metal science due to its potential to automate the sensing process for grain and particle size estimation. To ensure the good quality and reliability of metal products, analysis of the integrity of metallic components is required. In contemporary manufacturing processes, microscopic analysis is a crucial step, mainly when complex systems like gearboxes, turbines, or engines are assembled using various components from multiple suppliers. A final product's quality, durability, and lifespan are determined via the quality analysis of properties of a material with proper tolerances. A flaw in a single component can cause the breakdown of the entire finished product. To ensure the good quality of a material, micro-structural analysis is necessary, which includes the routine measurement of inclusions. The particle and grain sizes of particulate samples are the most crucial physical characteristics of metals. Their measurement is routinely conducted across various industries, and they are frequently considered essential parameters in the creation of many products. This paper discusses the role of sensors in enhancing the accuracy and efficiency of metallographic image analysis, as well as the challenges and limitations associated with this technology. The paper also highlights the potential applications of sensor-assisted metallographic image analysis in various industries, such as aerospace, automotive, and construction. The paper concludes by identifying future research directions for this emerging field, including the development of more sophisticated algorithms for grain and particle size estimation, the integration of multiple sensors for more accurate measurements, and the exploration of new sensing modalities for metallographic image analysis.

A Systematic Review of Modeling and Simulation for Precision Diamond Wire Sawing of Monocrystalline Silicon.

Precision processing of monocrystalline silicon presents significant challenges due to its unique crystal structure and chemical properties. Effective modeling and simulation are essential for advancing the understanding of the manufacturing process, optimizing design, and refining production parameters to enhance product quality and performance. This review provides a comprehensive analysis of the modeling and simulation techniques applied in the precision machining of monocrystalline silicon using diamond wire sawing. Firstly, the principles of mathematical analytical model, molecular dynamics, and finite element methods as they relate to monocrystalline silicon processing are outlined. Subsequently, the review explores how mathematical analytical models address force-related issues in this context. Molecular dynamics simulations provide valuable insights into atomic-scale processes, including subsurface damage and stress distribution. The finite element method is utilized to investigate temperature variations and abrasive wear during wire cutting. Furthermore, similarities, differences, and complementarities among these three modeling approaches are examined. Finally, future directions for applying these models to precision machining of monocrystalline silicon are discussed.

Development of a sustainable polymer composite: enhancing properties of waste plastic with ground tire rubber reinforcement

The growing interest in utilizing waste for composite development has prompted investigations across social, economic, and environmental domains. This study focuses on utilizing recyclable waste plastic materials and micro-sized ground tire rubber (GTR) of varying sizes (600 μm, 300 μm, and 150 μm) to create a polymer matrix composite (PMC). Employing a thermal blending technique, the manufacturing process adjusts the composition ratios of polymer and GTR from 90:10 to 50:50. Subsequently, Energy Dispersive Spectroscopy (EDS) is utilized to analyze the PMC composition, while Fourier-transform infrared spectroscopy confirms functional group and chemical structure. The study demonstrates significant improvements in various properties upon adding GTR to High-Density Polyethylene (HDPE) and Low-Density Polyethylene (LDPE) composite materials. For HDPE-GTR composites, the melting temperature (Tm) ranged from 120.29 °C to 138.53 °C, crystallization temperature (Tc) from 102.84 °C to 127.14 °C, and enthalpy of melting (ΔHm) from 70.96 to 139.67 J g−1. Crystallinity (Xc) varied from 48.43% to 52.96%. In LDPE-GTR composites, Tm ranged from 106.08 °C to 129.57 °C, Tc from 90.27 °C to 112.20 °C, ΔHm from 75.59 to 142.53 J g−1, and Xc from 51.59% to 54.05%. Moreover, mechanical properties such as tensile strength, flexural modulus, and impact strength exhibited enhancements with GTR addition to the polymer matrix. These findings underscore the potential of sustainable waste utilization, advancing environmentally friendly and resource-efficient composite materials.

Utilization of Fibrous Mat Residues from Upholstered Furniture as Sustainable Fillers in Plywood Production.

Nonwoven upholstery fabric is a waste product which is mainly generated during upholstered furniture production. The polyester composition makes it problematic to recycle and reuse this product. This study examined the manufacturing process of nonwoven fabric-reinforced plywood composites and their selected mechanical and physical properties. Nonwoven fabric was integrated between veneers bound with urea-formaldehyde resin to improve standard layered composites' mechanical and physical properties. Several board variants were produced, differing in the position of the nonwoven layers in the composite structure. The composites were evaluated for modulus of rupture (MOR), modulus of elasticity (MOE), internal bond, and screw withdrawal resistance, among others. The results showed that the addition of nonwoven fabric significantly improved some properties, like internal bond and screw withdrawal resistance. Variants with strategically placed nonwoven layers showed the highest performance increases. The results underscore the potential of nonwoven fabric as an effective reinforcing material, offering a path to developing high-performance plywood composites suitable for demanding applications. Another environmental advantage is that the nonwoven fabric waste used in the tested plywood production has not been subjected to burning or landfilling but, through its incorporation into plywood structure, has positively contributed to the Carbon Capture and Storage (CCS) policy. The findings advocate for a circular economy approach, in which industrial waste is effectively repurposed, contributing to the development of green materials in the wood-based composite industry.

Linoleum waste as PLA filler for components cost reduction: Effects on the thermal and mechanical behavior

Polylactic acid (PLA) is a biodegradable polymer from renewable resources with mechanical properties comparable to traditional polymers, but with a higher cost. A solution to this issue is the production of bio-based composites to partially replace the PLA matrix with industrial wastes characterized by a zero-cost, e.g., linoleum, to also valorize them in a circular economy perspective. Linoleum heterogeneous nature deriving from the simultaneous presence of lignocellulosic and inorganic fillers and oil/rosin binders, made the evaluation of matrix/filler compatibilization strategies necessary. Two approaches were considered, one from the filler perspective with NaOH and silane treatments, and the other one from the matrix perspective by adding a chain extender (C.E.). The first approach marginally improved tensile stiffness (by 1.6 %) compared to neat PLA but caused a significant decrease of 32.8 % in strength. Considering this, the costs and disposal of the chemicals and the increased environmental impact of the process, this approach was discarded. One the contrary, the introduction of C.E. does not modify the manufacturing process and increases tensile stiffness and elongation at break of 7.2 % and 415.5 % compared to neat PLA with a tolerable reduction in strength, i.e., 16.6 %, thus being a suitable way to exploit linoleum as zero-cost filler.

DETERMINING THE PERFORMANCE OF A HEAT EXCHANGER USED IN THE FOOD INDUSTRY

The study of heat transfer in the food industry is as important as in any other industry. Food products are constantly subjected to heat treatments. For this reason, a sunflower food oil, which undergoes heat treatment in the manufacturing process, was chosen for this study. The study was carried out on a laboratory plant and aimed to determine the heat transfer coefficients for three sets of determinations: partial heat transfer coefficient inside - where the warm sunflower oil circulates, partial heat transfer coefficient in the annular space - where cold water circulates and overall heat transfer coefficient, in a tube-in-tube heat exchanger made of copper. The obtained values of heat transfer coefficients were compared with literature data for the same type of food oil. This type of heat exchanger, tube-in-tube, is specific to the food industry. Countercurrent circulation of flows through the appliance was established from the outset, as this type of circulation is preferred from the point of view of determining the performance of heat exchange appliances. Efficiency values were also established for the three sets of experimental determinations. For the calculations, the values for the temperatures of the fluids, both at the inlet and outlet of the apparatus and the values for the volume flow rates of the fluids were required.

Manufacturing time estimation for offer pricing: A machine learning application in a French metallurgy industry

- In today's market where many companies are competing for the same opportunities, a quick, accurate, and reliable estimation of the offers' prices is essential for the suppliers. This allows them to answer to more solicitations and to remain competitive. Many suppliers use the manufacturing time as key parameter to define they offers' prices. However, in Make-To-Order (MTO) and Engineer-To-Order (ETO) contexts, with the increasing number of demands, the huge diversity of the products and the manufacturing process complexity, estimating the manufacturing time is challenging. To face this issue, this article proposes a new approach for the manufacturing time estimation using machine learning. Its originality relies on the exploitation of structured tabular data, unstructured textual data and domain knowledge. It also proposes two strategies that allow to choose the best estimation model for each product family, while minimizing the number of estimation models and considering product families with fewer available data. All the proposals have been applied to a real industrial case in a French small metallurgy industry, and the results have shown their applicability and effectiveness. Among the plethora of machine learning techniques implemented, CatBoost, which is a member of the family of Gradient Boosted Decision Trees (GBDT) ensemble techniques, provided the best results.