Sebastian Lang

Last Name

Lang

First Name

Sebastian

ORCID

0000-0002-1668-879X

Organisational unit

09706 - Bambach, Markus / Bambach, Markus

Show all metadata

Search Results

Publications 1 - 10 of 11

Training efficient and compensating fast: Data augmentation for thermal error compensation models of machine tools
Lang, Sebastian; Lampert, Nico; Mayr, Josef; et al. (2024)
SIG : Thermal Issues - Proceedings
The increasing importance of sustainability in manufacturing has created a trilemma situation between resource efficiency, productivity and precision. As the relevance of thermal errors increases with the shift to less cooling-reliant manufacturing approaches, compensation models for thermal errors have been proposed as a possible solution to the described trilemma. Since those data-driven compensation models are heavily reliant on both data amount and data quality, this paper aims at investigating different approaches of preprocessing and potentially augmenting the input data for compensation models. This allows for a higher sampling frequency than that of the employed measurement cycle and makes more, as well as additional synthetic, data points available for training. The employed ARX model can reduce the volumetric error by around two thirds. The use of data augmentation represents an increase in volumetric modelling accuracy from 48.2% to 65.8 % without requiring any additional measurement effort.
Kalman filter-driven state observer for thermal error compensation in machine tool digital twins
Lang, Sebastian; Talleri, Sofia; Mayr, Josef; et al. (2024)
Manufacturing Letters
Sustainable reduction of thermal errors during production is the key challenge in modern high-precision manufacturing. Numerical compensation models provide an energy-efficient solution, but in the case of data-driven models, high-quality experimental data must be time-consuming and expensive to produce, negatively impacting overall productivity. Furthermore, robustness concerns arise in the case of new operating conditions, which were not contained in the training data. This paper presents a novel use of a Kalman filter together with model order reduced finite element models to observe the entire thermal state, which allows the subsequent solution of the mechanical model and computation of the thermal errors in real-time without requiring any training data but instead purely based on the physical system model. The effectiveness of this approach is evaluated using experiments on a thermal test bench with 16 out of 40 temperature sensors employed for observation and demonstrated on a 5-axis machine tool (MT) with 13 out of 25 temperature sensors used. Due to the combination of the reduced order model and Kalman filter these 13 temperature sensors are sufficient to represent a MT mesh of more than 350’000 elements. The entire temperature profile of the thermal test bench is reconstructed to achieve a root mean square error (RMSE) of the unobserved temperature sensors of only 2.7 °C, which accounts for more than 83% of all temperature variations and 1.3 °C for the 5-axis MT. For the thermal error of the thermal test bench, the RMSE could be reduced from 67.4μm to 33.3μm, corresponding to a reduction of 52.7 %. This could be achieved without the need for experimental data for model calibration, in a real-time capable physics-based model.
Thermally compensated 5-axis machine tools evaluated with impeller machining tests
Zimmermann, Nico; Müller, Elija; Lang, Sebastian; et al. (2023)
CIRP Journal of Manufacturing Science and Technology
Sustainable precision manufacturing requires a transformation from resource-based to intelligence-based reduction strategies for thermal errors. Current approaches such as machine cooling and air-conditioning of shop floors are highly energy intensive. Therefore, this paper presents and evaluates a comprehensive thermal error compensation strategy for 5-axis machine tools considering thermal errors of linear and rotary axes, as well as a wide variety of thermal load cases. The method encompasses an automatic characterization of the thermal behaviour of 5-axis machine tools as well as an automated setup and adaption of the data-driven compensation models to realize high robustness to changing thermal boundary conditions. The applied on-machine measurement cycle identifies 15 axis-specific thermal errors using a touch trigger probe and a precision sphere. To demonstrate the universal application and long-term robustness of the presented compensation strategy, it is evaluated on two different 5-axis machine tools using long-term experiments between 700 h and 900 h. The peak-to-peak values of the volumetric thermal errors at the considered working space positions are reduced from 76 µm to 20 µm and from 84 µm to 33 µm, respectively. This corresponds to a reduction of 74% and 61%. The corresponding root mean squared errors are reduced by 84% and 65%. Finally, the effectiveness of thermally compensated 5-axis machine tools is analysed during simultaneous 5-axis milling by manufacturing two uncompensated and two compensated impellers. The self-learning thermal error compensation reduces the maximum root mean squared error of the impeller blades up to 73% from 32 µm to 9 µm for temperature variations of about 10 °C. Thus, thermally compensated 5-axis machine tools increase the process capability in fluctuating ambient temperatures. Consequently, the self-learning thermal error compensation enables a significant increase in accuracy without requiring prior knowledge of the thermal machine tool behaviour. This provides a significant step towards more sustainable precision manufacturing.
The impact of self-learning thermal error compensation models on the accuracy of 4-axis thermal test pieces
Zimmermann, Nico; Lang, Sebastian; Mayr, Josef; et al. (2022)
Self-learning thermal error compensation models are a cost-effective and robust approach to reduce thermal errors of machine tools. This paper evaluates a self-learning thermal error compensation strategy under cutting conditions with and without metal working fluid using thermal test pieces mapping the most relevant thermal errors of a rotary table. The experimental results indicate that the self-learning thermal error compensation models reduce the deviations of the thermal test pieces for both cutting conditions between 63% and 73%. Consequently, the machining results on the thermal test pieces verify the ability of the self-learning thermal compensation models to reduce robustly thermal errors of machine tools.
Validating real time compensation: A thermal test piece for 5-axis machine tools to separate thermal errors in Z-direction
Zimmermann, Nico; Lang, Sebastian; Mayr, Josef; et al. (2024)
Precision Engineering
The accuracy of 5-axis machine tools is a key factor to manufacture multi-axes machined workpieces. However, thermal deformations of the machine structure often cause significant deviations at the tool centre point. To evaluate the impact of thermal errors on the accuracy of 5-axis machine tools under machining conditions, suitable thermal test pieces are required. Therefore, this paper introduces a 5-axis thermal test piece which is based on the thermal test piece for rotary axes being part of the standard ISO 10791-10:2022. The new thermal test piece identifies ten instead of five thermal errors during eight time steps. The additional features enable an error separation of the table-related thermal error and the thermal error pointing consistently in Z-direction. This error separation is especially important to analyse the accuracy of five-axis milling operations. In total, the developed 5-axis thermal test piece analyses five location, two length and three orientation errors without requiring a kinematic model of the machine tool. The conducted experiments comprise thermal load cases without and with metal working fluid and show a clear agreement between the thermal location errors and the dominant thermal orientation errors of the thermal test piece and an on-machine measurement cycle. On the other hand, the analysed thermal length errors of the thermal test piece indicate a clear effect of the thermal material expansion compared to the thermal positioning errors of the analysed machine tool. Furthermore, the analysed thermal load cases of the linear and rotary axes conducted without and with metal working fluid result in a different thermal behaviour of the machine tool depending on the use of metal working fluid. Finally, the developed 5-axis thermal test piece is used to evaluate a thermally compensated 5-axis machine tool under machining conditions for thermal load cases without and with metal working fluid. The self-learning thermal error compensation reduces the mean of the thermal location errors at the thermal test piece by up to 71%.
Thermal Error Compensation Models Utilizing the Power Consumption of Machine Tools
Lang, Sebastian; Zimmermann, Nico; Mayr, Josef; et al. (2023)
3rd International Conference on Thermal Issues in Machine Tools (ICTIMT2023)
Thermal errors are among the most significant contributors to deviations of products manufactured on modern machine tools (MTs). Reducing them is typically achieved through either design adaptation, active cooling of the MT and its environment, or compensation using measurements or model-based predictions. Model-based compensation strategies promise to have the lowest environmental footprint by far. In general, a compensation model needs to be accurate, robust to changing boundary conditions and must require only minimal experimental efforts as this reduces the productivity of the MT. Model inputs such as temperature measurements or the power consumption of various components, can be used to predict the thermal errors. The temperature inputs require additional sensors, effort and cost for the MT manufacturer to install and ensure up-time while the power consumption could be logged and are typically provided from the control system anyway. Adaptive compensation models are created using four different sets of inputs consisting of 13 temperature sensors and 7 power measurements. While the best results were obtained with all 20 inputs, the 7 energy recordings give similar results as the 13 temperature sensors if the environmental temperature is considered. The volumetric RMSE was reduced by 72% and the maximal error from 32.75 µm to 9.5 µm. ARX models proved to be suitable and even outperform more complex model structures such as LSTM and especially those without time dependency such as feed forward neural networks.
Influences of metalworking fluid on the thermal errors of a 4-axis thermal test piece
Lang, Sebastian; Zimmermann, Nico; Mayr, Josef; et al. (2022)
Thermally induced errors of machine tools can cause up to 75% of geometric errors on machined workpieces and are therefore a crucial precision defining factor on any produced part, which in turn has further effects on the performance of the produced part during its life cycle. In the machining process, metal working fluid fulfils many tasks and ensures beneficial cutting properties. It can have a significant impact on the thermal behaviour of machine tools. A thermal test piece is produced to identify the thermal errors of a rotary axis of a 5-axis machine tool under a thermal load consisting of rotating the C-axis for four hours and subsequently idling for four hours, once with metalworking fluid active during the C-axis movement and once without the use of any metalworking fluid. This test piece is produced in accordance to ISO/CD 10791-10:2020 in such a way that it allows for separating the thermal location error contributions of the main machine axes, and to infer on the effect of the thermal errors on the finished product. The metalworking fluid has a significant effect on the thermal behaviour of the analysed machine tool. In the analysed load, the thermal errors are reduced, when metalworking fluid is active due to reduced and homogenised temperatures. However, this cooling can also lead to a sign change of the thermal error. The asymmetric MWF drainage system design in the Y-direction leads to thermal errors that could not be observed if no metalworking fluid is used as otherwise the Y-direction behaves thermosymmetrically.
Model Uncertainty Estimation for Thermal Error Compensation in Machine Tools
Rhiner, Lenny; Lang, Sebastian; Mayr, Josef; et al. (2024)
SIG : Thermal Issues - Proceedings
This paper introduces a method to compensate for thermal errors in machine tools (MT) using LSTM neural networks, with a focus on addressing prediction uncertainties. It presents the application of Monte Carlo Dropout (MC-Dropout) to estimate the uncertainty of LSTM predictions using data generated in a simulated MT environment. MC-Dropout offers a practical, computationally efficient method to allow for effective thermal error compensation without repeated on-machine measurements. Incorporating uncertainty estimates can enhance decision-making, thus allowing more autonomous machine operations and improve the selection of training data for machine learning models, leading to greater overall prediction accuracy.
Adaptive input selection for thermal error compensation models
Zimmermann, Nico; Lang, Sebastian; Blaser, Philip; et al. (2020)
CIRP Annals
The presented method selects optimal inputs for compensation models based on the Thermal Adaptive Learning Control methodology. The number of inputs and the individual inputs for each considered thermal error are automatically adapted. The intelligent combination of k-means clustering and Time Series Cluster Kernel enables the approach to handle time series of thermal error measurements with missing data due to operational reasons. The results show that the adaptive sensor selection approach, tested on a 5-axis-machine tool, significantly increases the robustness of the used compensation model. The productivity loss due to on-machine measurements is reduced by approximately 40 percent. © 2020 CIRP. Published by Elsevier Ltd. All rights reserved.
Sensor placement utilizing a digital twin for thermal error compensation of machine tools
Lang, Sebastian; Zorzini, Mario; Scholze, Stephan; et al. (2025)
Journal of Manufacturing Systems
Thermal errors in machine tools significantly impact precision and, therefore, productivity. Mitigating these errors often results in a trade-off between energy efficiency and accuracy. While data-driven compensation models show promise in addressing this challenge and achieving sustainable precision, their effectiveness hinges on the careful selection and placement of sensors as model inputs. This paper introduces a novel temperature sensor positioning method for thermal error compensation that leverages a digital twin framework to virtually determine ideal sensor positions and their effects on the compensation model. By accurately identifying temperature-sensitive points, our approach improves compensation accuracy and reduces the number of sensors required, thus enhancing both model robustness and operational efficiency. For choosing this set not only one simulation model is used but an ensemble with varying boundary conditions and thus model properties. Validation results show that the proposed method outperforms traditional, manually determined sensor placement strategies, providing a scalable solution for adaptable, energy-efficient thermal management in precision manufacturing. The selected sensor set based on a hybrid singular value decomposition and Least Absolute Shrinkage and Selection Operator approach yields a more robust compensation using only 7 instead of the manually chosen 22 temperature sensors. The thermal error reduction ranges from 77%-94% using simulated data with a corresponding reduction of 75%-85% achieved on the physical machine.

Publications 1 - 10 of 11

Sebastian Lang

Last Name

First Name

ORCID

Organisational unit

Filters

Search Results