Eva Dierkes

Eva Dierkes

Auto-generated profile, based on publicly available information

Eva Dierkes

Center for Industrial Mathematics University of Bremen Germany

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 June 2023 in Chaos: An Interdisciplinary Journal of Nonlinear Science
Verified
Journal Article
Open access
PDF file
EDEva Dierkes
Eva Dierkes
COChristian Offen
Christian Offen
SOSina Ober-Bloebaum
Sina Ober-Bloebaum
KFKathrin Flaßkamp
Kathrin Flaßkamp
Recently, Hamiltonian neural networks (HNNs) have been introduced to incorporate prior physical knowledge when learning the dynamical equations of Hamiltonian systems. Hereby, the symplectic system structure is preserved despite the data-driven modeling approach. However, preserving symmetries requires additional attention. In this research, we enhance HNN with a Lie algebra framework to detect and embed symmetries in the neural network. This approach allows us to simultaneously learn the symmetry group action and the total energy of the system. As illustrating examples, a pendulum on a cart and a two-body problem from astrodynamics are considered.

25 January 2021 in Proceedings in Applied Mathematics and Mechanics
Verified
Journal Article
Open access
PDF file
EDEva Dierkes
Eva Dierkes
FJFrancesca Jung
Francesca Jung
CBChristof Büskens
Christof Büskens
The digital revolution, especially in the field of manufacturing, has great potential to change the economy sustainably. In this work, the development of methods for predictive maintenance and condition monitoring is a central focus. Data‐based models for drive technology in automotive production are investigated in order to generate adequate models, to make statements about the life cycle of the drives, or to suggest system modifications, such as the adjustment of weights. One task in this field is to detect anomalies or disturbances in given engine data. Since this collected data is variable and noisy, the detection of faults is non‐trivial. In this context two different methods for anomaly detection are studied. First, a model based on statistical analyses and second, a machine learning model is evaluated.

14 December 2021 in Proceedings in Applied Mathematics and Mechanics
Verified
Journal Article
Open access
PDF file
EDEva Dierkes
Eva Dierkes
KFKathrin Flaßkamp
Kathrin Flaßkamp
The great success of machine learning in image processing and related fields also motivated its application to dynamical system identification. In particular, neural networks are trained to learn equations of motion and thus provide an alternative to first‐principle modeling. While these black‐box algorithms are quite flexible regarding the system structure, they often have difficulties in learning basic physics laws which are intrinsic system properties. Recently, Hamiltonian neural networks (HNN) were introduced to explicitly learn the total energy of a system in order to overcome the lack of physical rules. However, Hamiltonian systems often also contain other structures such as symmetries in terms of further invariances. In this contribution, we extend HNN such they respect system invariances in addition to the Hamiltonian. The proposed extension leads to a trade off between energy conservation and conservation of invariance properties, which we investigate exemplarily.

24 April 2023 in arXiv
Verified
Preprint
Open access
PDF file
EDEva Dierkes
Eva Dierkes
COChristian Offen
Christian Offen
SOSina Ober-Blöbaum
Sina Ober-Blöbaum
KFKathrin Flaßkamp
Kathrin Flaßkamp
Recently, Hamiltonian neural networks (HNN) have been introduced to incorporate prior physical knowledge when learning the dynamical equations of Hamiltonian systems. Hereby, the symplectic system structure is preserved despite the data-driven modeling approach. However, preserving symmetries requires additional attention. In this research, we enhance HNN with a Lie algebra framework to detect and embed symmetries in the neural network. This approach allows to simultaneously learn the symmetry group action and the total energy of the system. As illustrating examples, a pendulum on a cart and a two-body problem from astrodynamics are considered.

30 December 2024 in Proceedings in Applied Mathematics and Mechanics
Verified
Journal Article
Open access
EDEva Dierkes
Eva Dierkes
LKLars Kappertz
Lars Kappertz
VSViacheslav Solovievskyi
Viacheslav Solovievskyi
AHAnnika Hackenberg
Annika Hackenberg
CBChristof Büskens
Christof Büskens
This work introduces an optimally controlled forecast‐based energy management system (EMS) that integrates energy demand and production forecast with optimal control techniques to optimize the management of renewable energy sources. Although controlling the energy generation is not feasible for weather‐dependent power plants, it is possible to predict the amount of energy produced. Self‐sufficiency can be improved by regulating individual devices and storage units based on anticipated energy production and demand. For electricity tariffs much larger than feed‐in tariffs, such demand‐side management can notably improve the profitability of (private) renewable energy systems. The EMS is based on long‐term predictions that enable high‐level control decisions and specific controls for the next controlling steps on a low level. Weather forecasts and forecasts of the power demand and state behavior of each device are considered to evaluate both cost function and constraints for the next prediction horizon. The system's performance is evaluated through a case study, demonstrating its ability to control the energy distribution successfully. The proposed EMS presents a promising approach to smart energy management that can contribute to a more sustainable future.