Quenching in a fluid is a complex process. There are several different heat transfer mechanisms that may be occurring at the same time, with the heat transfer coefficients changes as a function of position (x, y, z) and surface temperature on the same part. This is further complicated by having multiple different parts in the same load. Agitation, racking of the parts and the quench tank design all play a role in the resultant properties and distortion of a given part. Further complicating this problem, is that there are multiple methods to measure quenching performance. In this paper, we will be describing an agitation apparatus used at Quaker Houghton for determining heat transfer coefficients as a function of agitation and surface temperature. The probe used is the ISO 9950 (ASTM D6200) Inconel probe, and the heat transfer coefficients are determined by an inverse method provided by the SmartQuench Integra software by RISE/ivf. The apparatus is examined using Computational Fluid Dynamics (CFD), and the calculated flow is compared to the measured fluid flow.

The vapor film collapse that occurs in the quenching process is complicated and unstable, these affects the heat treatment quality and its distortion. Particularly in mass production, where production costs are taken into consideration, products are often packaged in group load setting, so it is very important to know the deformation variation and distribution within that process. In order to incorporate it into the MBD technology required these days, it is necessary to predict the quality of heat treatment by CAE, shorten the product development period. However, in the past day, in order to formulate the vapor film collapse on a simulation, it was necessary to perform a very large amount of computational fluid dynamics calculation (CFD), because of a problem in computer resources and the model of vapor film collapse. In addition, this phenomenon has an complexity behavior of the phenomenon in iterative processing in mass production, which also complicates the calculation. The vapor film collapse phenomenon was visualized by using cellular automaton simulation its include the phenomena of “Vapor film thickness”, “Flow disturbance”, “Surface step of workpiece”, “the pressure for vapor film”. In this study, the Markovian property of vapor film surface vibration was clarified, and the heat treatment deformation instability of ring-shaped parts due to it was predicted.

The analysis of cooling curves obtained by immersing a probe in the quench medium has been widely used since its availability. For instance, methods described in standards such as ISO 9950 and ASTM D 6482 refer to the use of an Inconel 600 specimen which is quenched to obtain the cooling curve of a given fluid; however, spray quenching is being mostly used in induction hardening processes. In this work, the quenching characteristics of a PAG polymer at 6 and 12 % concentration were determined and compared with water as a baseline. The fluid was heated at 30 °C, while the solution flow rate was set at 90 L/min; two different quenching rings were designed and used in a laboratory-scale setting. Also, the fluid flow in the quench rings was simulated through Computational Fluid Dynamics (CFD), to obtain flow patterns inside the quenching devices. From the results obtained, the cooling rate curves showed no vapor phase, and the maximum cooling rate was found to be higher in one of the quench ring designs. The design of the quench ring device has a significant influence on the quenching characteristics of the quenchant, mainly at medium and low temperatures of the cooling rate curve.

Manufacturers regularly employ finite-element (FE) process modeling tools for the simulation of heat treatment applications, such as quenching. These tools may utilize thermal, mechanical and microstructural calculations in the analysis of part distortion and residual stresses. Heat treatment modeling workflows are challenged by the requirement for user-provided heat transfer boundary conditions, which vary based on part geometry and process parameters. Representative Heat Transfer Coefficients (HTCs) are typically reversed-engineered using experimental thermocouple data, thermal simulations and inverse optimization methods. This paper will present ‘state of the art’ developments integrating computational fluid dynamics (CFD) capabilities into the heat treat modeling environment of the DEFORM system. It will describe how CFD and thermal modeling of a quench medium is being coupled with deformation and heat transfer modeling of a part through the use of CFD-calculated, local heat transfer boundary conditions. Studies verifying the implemented CFD methods against published literature will be summarized. Application examples will show how residual stress and distortion in parts, during single-part or batch gas quenching, is made possible by coupled CFD and thermo-mechanical process modeling tools.

Prof. Tatsuo Inoue passed away on September 23, 2023, at the age of 83. He held a professorship at Kyoto University from 1983 to 2003 and made significant contributions to the theory of heat treatment simulation, which is now widely used. His theory was reported at an international conference in Linkoping, Sweden in 1984. Fundamental equations in his theory cover metallurgical coupling effects caused by changes due to phase transformation, temperature, and inelastic stress/strain as well as carbon diffusion during the carburizing process. Prof. Inoue designated these effects as “metallothermo- mechanical coupling”. Software applying his theory was presented at ASM International’s 1st International Conference on Quenching and the Control of Distortion in 1992, where its advanced nature was recognized. In 1994, Prof. Inoue published a paper on the application of heat treatment simulation to the quenching of Japanese swords, revealing changes in temperature, curving, microstructure, and stress/strain in their model during the traditional quenching process. In 2017, he published “The Science of Japanese Swords” with Sumihira Manabe, a swordsmith, to communicate his specific achievements to the general public.

Quenching is one of the primary processes to improve mechanical properties in steels, particularly hardness. Quenching is well established for different geometries of individually treated steel components; while in-steam quenching of large diameter continuously cast steel bar has several specific features which are difficult and costly to experimentally optimize. The end-quench Jominy test has been used extensively to study the hardenability of different steel grades. Different numerical, analytical, and empirical models have been developed to simulate the Jominy process and to understand quenching of steels. However, it is not straight forward to translate experimental data from Jominy test on instream quenched large diameter continuously cast products. Therefore, in this work, coupled thermal, mechanical, and metallurgical models were used to simulate the end-quench Jominy test and in-stream quenched industrial round billets with a goal to obtain similarity of experimental structure and properties for both quenched products. For this purpose, finite element analysis (FEA) was employed using the software FORGE (by Transvalor). Used thermophysical properties were generated by JMATPro software. The evolution of microstructure during quenching and resulting hardness were simulated for AISI 4130, and AISI 4140 steel grades. The cooling rates at different positions in the Jominy bar were determined by simulation and compared to experimental. After verification and validation, the FEA simulation was utilized to predict different phases and hardness at different conditions in industry produced round billets. Additionally, relations between Jominy positions and radial positions in the billet were established allowing us to predict structure and properties in inline quenched continuously cast bar having different diameters.

Heat Treatment represents one of the largest challenges for component risk management. Traditional metallurgical test methods do not meet AIAG/VDA Defect detection criteria for safety-critical components and can represent significant overhead costs. Newer non-destructive methods are difficult to implement with substantial upfront costs and must be integrated as 100% inspection to impact PFMEA detection ratings, which can introduce a throughput constraint. Production controls and automated escalation are imperative to minimizing risk. On the development side, it is impractical to physically evaluate all combinations of product/process variation, or even test specification limits. Consequently, designs which met requirements in validation may experience degraded functionality in production due to ‘normal’ process variation that cannot be eliminated, or inevitable differences between early development and production scale processes. With the accelerated pace of innovation seen in the automotive industry, use of FEA simulation to evaluate part sensitivities is essential to identify and optimize design/process, reducing risk. Increased confidence must be achieved in test and data processing methodology through robust implementation which often requires substantial investment in time and data analysis, which can be streamlined through machine learning.

From concept and design to production readiness, introducing a new product or service can be expensive. Because of the lower cost of smart technologies, it is now possible to validate designs with less rounds of prototyping and testing. This leads to savings in the design phase, improved product quality, and less materials and energy being used to bring the product or service to market. The sum total of these savings’ lower startup costs.

As Computational Fluid Dynamics (CFD) methods evolve and mature, more engineering problems are being solved through computer simulation to reduce reliance on the costly and time-consuming experimental methods. This trend is also occurring in the gear manufacturing industry, where an increasing number of quality issues related to the oil quenching process are being investigated by CFD methods to find solutions. However, while the CFD theory and numerical methods have made significant advancements, gaps still exist between the academic research and industrial applications. In the case of the oil quenching processes, the prospect of using CFD methods to visualize and study the oil flow pattern in the gear quenching tank is promising yet challenging. The obstacle to simulating the oil quenching process using CFD methods lies not in the numerical method itself for solving the Navier-Stokes equation, but in building a computer simulation model that encompasses all the geometrical details of the quenching tank, fixtures, centrifugal pumps, and gears, including all the gear teeth. This task is particularly challenging for Finite Volume Method (FVM) CFD solvers, as the computation mesh could take days or weeks to build. In this research, a new solution method based on Smoothed Particle Hydrodynamics (SPH) is introduced to simulate the oil flow in the gear quenching tank. Since SPH is a mesh-free Lagrangian method, it not only greatly simplifies the mesh generation task for building the computational models but also handles the complex physics of the free surface flow and fluid-structure interaction with great ease. In addition, the oil flow in the gear quenching tank usually is driven by centrifugal pumps whose dynamics can be simulated directly in SPH methods, as opposed to FVM methods which require complicated moving mesh computation.

Heat-treatment simulation is a powerful tool for gear design and process troubleshooting, but many times the predicted gear distortion is difficult to compare to physical gear measurements and to required specification charts or measurements. To help ease this burden, two software programs are utilized to provide powerful gear analyses to heat-treatment simulation results. This paper briefly describes the software used, DANTE and Integrated Gear Design (IGD), and presents a simple case study. The stress and deformation from the heat treatment of a small gear made of SAE 10B22 are predicted using DANTE. The distorted gear geometry is then imported into IGD and the predicted distortion is compared to the actual measurements of the gear.

Heat treatment simulation has progressed to the point where commercial software is widely available, and validations of simulation functions using experimental results have played a big role in getting here. For this reason, the author presents a number of validation cases and explains what relatively simple experiments can reveal about the complex phenomena of heat treating. In the case of validating basic functions, such as heat transfer and phase transformation, the author uses experimental results of the inverse hardening of quenched steel cylinders. When validating software at the stage where stress and strain functions are added, the author uses measurement data corresponding to length and diameter changes and residual stress distributions in normally quenched steel cylinders. Other cases presented include the validation of curving in long specimens cooled unevenly and the validation of distortions and residual stresses in carburized and quenched, induction hardened and nitrided steels.

A physics-based software model is being developed to predict the nitriding and ferritic nitrocarburizing (FNC) performance of quenched and tempered steels with tempered martensitic microstructure. The microstructure of the nitrided and FNC steels is comprised of a white compound layer of nitrides (ε and γ’) and carbides below the surface with a hardened diffusion zone (i.e., case) that is rich in nitrogen and carbon. The composition of the compound layer is predicted using computational thermodynamics to develop alloy specific nitriding potential KN and carburizing potential KC phase diagrams. The thickness of the compound layer is predicted using parabolic kinetics. The diffusion in the tempered martensite case is modeled using diffusion with a reaction. Diffusion paths are also developed on these potential diagrams. These model predictions are compared with experimental results.

Commercially, carbon steels are induction heated at heating rates on the order of 100 to 1,000 °C·s -1 for surface hardening. The high precision DIL 805L dilatometer employs induction heating and is often used to study transformation characteristics and prepare test specimens for metallurgical analysis. However, heating the commonly used 4 mm diameter by 10 mm long specimens at rates above 50 °C·s -1 results in non-linear heating rates during transformation to austenite and large transient temperature variations along the specimen length. These limitations in heating rate and variances from ideal uniform heating can lead to inaccurate characterization of the transformation behavior compared to commercial induction hardening practices. In this study it is shown that changing the specimen design to a thin wall tube allows faster heating rates up to 600 °C·s -1 and modifies the pattern of temperature variations within the test sample. The response of selected specimen geometries to induction heating in the dilatometer is characterized by modelling and tests using multiple thermocouples are used to verify the models. It is demonstrated that the use of properly designed tubular test specimens can aid in more accurately establishing transformation characteristics during commercial induction hardening.

Heat treaters are adopting more and more Industry 4.0 techniques and solution packages to improve production processes and product quality. Proper specification, measurement, and control of heat-treating atmospheres are always critical to achieving the desired metallurgical and microstructural results. The combination of atmosphere measurements and other furnace operating parameters (e.g., furnace temperature and pressure) can provide a better view of the whole production. Thermodynamic calculations and field experiences can be integrated into the smart solution to provide process engineers more capabilities to manage and optimize production. In this article, our recent research and development work on smart solutions for the heat-treating industry will be presented and discussed.

AISI 52100 is a high carbon alloy steel typically used in bearings. One hardening heat treatment method for AISI 52100 is austempering, in which the steel is heated to above austenitizing temperature, cooled to just above martensite starting (Ms) temperature in quench media (typically molten salt), held at that temperature until the transformation to bainite is completed and then cooled further to room temperature. Different austempering temperatures and holding times will develop different bainite percentages in the steel and result in different mechanical properties. In the present work, the bainitic transformation kinetics of AISI 52100 were investigated through experiments and simulation. Molten salt austempering trials of AISI 52100 were conducted at selected austempering temperatures and holding times. The austempered samples were characterized and the bainitic transformation kinetics were analyzed by Avrami equations using measured hardness data. The CHTE quench probe was used to measure the cooling curves in the molten salt from austenitizing temperature to the selected austempering temperatures. The heat transfer coefficient (HTC) was calculated with the measured cooling rates and used to calculate the bainitic transformation kinetics via DANTE software. The experimental results were compared with the calculated results and they had good agreement.

In this paper, we study the energy absorption of metamaterials composed of unit cells whose special geometry makes the cross-sectional area and the volume of the bodies generated from them constant (for the same enclosing box dimensions). After a parametric description of such special geometries, we analyzed by finite element analysis the deformation of the metamaterials we have designed during compression. We 3D printed the designed metamaterials from plastic to subject them to real compression. The results of the finite element analysis were compared with the real compaction results. Then, for each test specimen, we plotted its compaction curve. By fitting a polynomial to the compaction curves and integrating it (area under the curve), the energy absorption of the samples can be obtained. As a result of these investigations, we drew a conclusion about the relationship between energy absorption and cell number.

Computer simulations are increasingly being used in the automotive industry to evaluate the state of stress in cylinder blocks during casting and heat treat processes. With recent advancements, it is now possible to model casting and quenching processes as well as residual stress and high cycle fatigue. However, calculating the final stress in cylinder blocks requires the integration of several software tools with different meshing topologies, numerical methods, data structures, and post-processing capabilities. The intent of this research is to develop an integrated virtual engineering environment that combines casting simulation, computational fluid dynamics, and finite element methods in order to simulate the manufacturing process from the beginning of casting, through water quenching heat treatment, to engine dynamometer testing. The computational environment is built on three CAE tools, Magmasoft, AVL Fire, and Abaqus, and required considerable amounts of research and development to validate each numerical method and the tools that facilitate data exchange between them.

The cooling history of carburized heat-treated gears plays a significant role in developing microstructure, hardness, and residual stress in the tooth that influences the fatigue performance of the gear. Evaluating gear carburizing heat treatment should include a microstructure and hardened depth evaluation. This can be done on an actual part or with a test piece. The best practice for a test piece is to use a section size that closely approximates the cooling rate at the gear flank of the actual gear. This study furthers work already presented showing the correct test piece size that should be used for different gear modules (tooth thicknesses). Metallurgical comparisons between test pieces, actual gears, and FEA simulations are shown.

The knowledge of the thermal boundary conditions helps to understand the heat transfer phenomena that takes place during heat treatment processes. Heat Transfer Coefficients (HTC) describe the heat exchange between the surface of an object and the surrounding medium. The Fireworks Algorithm (FWA) method was used on near-surface temperature-time cooling curve data obtained with the so-called Tensi multithermocouple 12.5 mm diameter x 45 mm Inconel 600 probe. The fitness function to be minimized by a Fireworks Algorithm (FWA) approach is defined by the deviation of the measured and calculated cooling curves. The FWA algorithm was parallelized and implemented on a Graphics Processing Unit architecture. This paper describes the FWA methodology used to compare and differentiate the potential quenching properties of a series of vegetable oils, including cottonseed, peanut, canola, coconut, palm, sunflower, corn, and soybean oil, versus a typical accelerated petroleum oil quenchant.

ASTM D6200 is a standard test method to evaluate cooling characteristics of quench oils. The test produces six discrete numbers representing the cooling characteristics: three temporal scales (time to cool to 600°C, 400°C, and 200°C), two cooling rates (max cooling rate and cooling rate at 300°C), and one temperature scale (at max cooling rate). One of the main purposes of ASTM D6200 is to monitor the oil quality to ensure gears are properly quenched. The current standard only includes specifications for gear quenching oil and its applications are limited to physical testing. The intent of this research is to explore the possibility of broadening the support for more quenchants and extending applications to virtual engineering. This research includes two parts. The first part is the development of a systematic method to identify the characteristic points of a cooling curve. The second part is the construction of an analytical cooling curve based on the characteristic points. The analytical cooling curve is a mathematical function of temperature versus time that can provide temperature at any given time in the quenching process. In addition, the curve is differentiable to provide the cooling rate information at any given time as well.