Programme

Physics-informed neural networks (PINNs) have become increasingly popular in several engineering domains due to their efficiency in addressing real-world problems characterized by noisy data and occasionally incomplete physics information. In the realm of PINNs, automatic differentiation plays a pivotal role in evaluating differential operators without succumbing to discretization errors. Furthermore, a multi-task learning approach is defined to concurrently accommodate observed data fitting and adherence to the fundamental governing principles of physics.

In this context, we present the application of PINNs to heat treatment operations, with a specific focus on handling realistic conditions that are challenging for traditional computational methods. The estimation of unknown thermal boundary conditions on surfaces is facilitated by predicting them through sparse temperature measurements. We illustrate the applicability of PINNs through several industrial applications related to quenching, underscoring the effectiveness of neural networks in tackling complex heat transfer problems encountered in industrial settings.

The heat treatment industry has long utilized automation technologies—such as PID control, setpoint programming, and supervisory systems like SCADA—to enhance efficiency and consistency. Today, we stand at the threshold of a new transformation: moving from conventional automation to intelligent autonomous systems powered by Artificial Intelligence (AI), Machine Learning (ML), and Machine Teaching (MT).

This paper examines how the integration of AI, ML, MT, and advanced thermal-loop components is revolutionizing heat treatment processes. Utilizing the Edge Process Management (EPM), we showcase how real-time data analytics and edge solutions are optimizing thermal processes and significantly reducing energy consumption. The EPM platform facilitates immediate, localized decision-making, improving response times and data security compared to cloud-based systems.

Machine Learning algorithms enable predictive maintenance and adaptive control, enhancing process stability and efficiency. Simultaneously, Machine Teaching (MT) allows industry experts to embed their knowledge directly into AI models, accelerating the development of intelligent systems. By uniting these technologies, we achieve unprecedented levels of process optimization.

Through case studies and experimental results, we highlight substantial improvements in energy efficiency, process consistency, and overall operational sustainability. This advancement marks a critical step in the industry's evolution—from standard automation to intelligent, self-optimizing systems.

The VariQuench process is a stepped quenching process during case hardening. The components are rapidly quenched to a surface temperature around martensite start temperature of the carburized surface. Afterwards, a slow cooling rate is applied. This can significantly improve the properties of case-hardened gears. The cause of the improved performance of components due to the stepped quenching is currently still under debate. An optimised stress state but also more tolerant microstructural properties are discussed. To summarize, optimisation and transfer to different component geometries is actually only possible through an extensive trial-and-error process.

Dilatometer tests and accompanying metallographic investigations were used to investigate self-tempering effects in the case-hardening steel 20MnCr5 (SAE 5120) at two carburized states to determine microstructural effects. During stepped quenching the retained austenite stabilization, the martensite tetragonality together with transition carbides precipitation from martensite were identified as significant microstructural effects. They all depend on the step temperature or the cooling rate afterwards.

The identified self-tempering effects were modelled and transferred to component simulations. The required parameters for the models were determined by fitting to the dilatometer results. The models and its parameters were included in a user subroutine UMAT for the FE-software Abaqus. The microstructure and residual stresses of gears for pulsator tests were calculated using the FE-simulation. A reference heat treatment with direct, non-stepped quenching and two VariQuench variants were simulated. The simulation results obtained are discussed in comparison with experimental results from component tests applying these quenching conditions. The temperature evolution within the gears, the hardness and the residual stresses where determined experimentally after heat treatment.

The technology of Water Spray Quenching has been established in industry for many years, e.g. in steel production for the quenching of strips or for semi-finished products. When using this process, water-mist is formed in nozzles and accelerated with high velocities towards the components to be quenched. Water spray quenching provides very high cooling rates. Additionally, by variation of the rate of water-flow and air-flow, the quench intensity can be varied in a wide range.

However, for the heat treatment of complex shaped serial parts, such as gear-wheels, this process could yet not be applied successfully. So far this was not possible, since the water spray could not reach into the center of the heat treat loads consisting of multiple layers since the outer parts of the load shield the inner parts from the effect of the water spray.

With a new combination of vacuum heat treatment in single layers with subsequent Water Spray Quenching, it is now possible to apply Water Spray Quenching for serial parts.

An industrial-scale test rig has been developed and used to investigate the following topics:

• cooling curves,
• heat transfer coefficients,
• core hardness – values,
• CHD – values,
• surface quality after quenching and
• the effect of a variation of quench intensity.

It was demonstrated that the process offers a wide range of quench intensities up to very high cooling rates. Even when using low alloyed case hardening steels, high core hardness values were successfully reached.

Spray quenching is an important technique used to quench parts after induction hardening. It has the advantages of rapid cooling of work pieces, and a repeatable process through the control of spray pattern, quenchant velocity, and pressure. There have been numerous efforts to characterize the heat transfer coefficient as a function of surface temperature in the literature. One method has been to use standard probes used to produce cooling curves for immersion quenchants. However, due to the high surface heat transfer, steep gradients can occur, underestimating the heat transfer coefficient. An examination of these thermal gradients as a function of data acquisition rate and material was accomplished, to determine the proper probe geometry to measure and calculate the resultant heat transfer coefficient was determined.

According to DIN EN 10052, quenching is defined as “cooling a workpiece at a greater speed than in still air”.

The quenching process and the quenching product in use are therefore of great importance for the success of the heat treatment process, affecting the resulting material structure and mechanical properties of the heat-treated parts, as well as the height of residual stresses and dimensional and shape changes.

The proper choice of the quenching product, its adaptation to the range of parts and material specifications but furthermore to the existing equipment and process technology is the precondition for reliable and consistent hardening results, reaching the defined requirements for the parts and its subsequent function.

In the last years, low-pressure carburizing (LPC) with subsequent oil quenching has become more and more important in industrial applications as a direct alternative to high-pressure gas quenching processes.

The paper provides an overview of the requirements for the quenching oils and its necessary properties resulting from the LPC-Process. Special attention will be drawn to the possibilities of pressure adjustment to influence the quenching effectiveness and the heat transfer properties. In addition, practical applications are presented.

Vacuum oil quenching has become a key element for heat treating large metal parts, offering distinct advantages over traditional quenching methods such as water or gas quenching. These conventional methods often encounter challenges related to cooling uniformity and distortion, especially when applied to large or complex shaped components. Vacuum oil quenching can address these issues by providing precise control of cooling rates and minimizing the risk of thermal stress, cracking or distortion in large parts. Large parts often require deep carburizing depths, which stretches furnace capabilities and vacuum carburizing process knowledge to the limit.

A key advantage of vacuum oil quenching is its ability to treat large parts such as turbine blades, gears or crankshafts, which are difficult to treat with traditional quenching methods. The system ensures uniform cooling, which is essential to achieve the desired mechanical properties, such as hardness, toughness and fatigue resistance. In addition, the vacuum oil quenching process allows the processing of high-performance alloys and large parts with complex geometries, ensuring their structural integrity.

Beyond its technical advantages, the vacuum oil quenching system also improves operational efficiency by reducing cycle times and energy consumption compared to conventional methods. In accordance with the FIT55 policy. Its ability to increase production throughput while maintaining the quality of heat-treated parts makes this technology a valuable solution for sectors such as aerospace, automotive and heavy machinery manufacturing.