Refine
Year of publication
Document Type
- Conference Proceeding (9)
- Article (5)
- Master's Thesis (3)
- Book (1)
- Working Paper (1)
Institute
Language
- English (19) (remove)
Keywords
In this paper, we consider the question of data aggregation using the practical example of emissions data for economic activities for the sustainability assessment of regional bank clients. Given the current scarcity of company-specific emission data, an approximation relies on using available public data. These data are reported in different standards in different sources. To determine a mapping between the different standards, an adaptation to the Covariance Matrix Self-Adaptation Evolution Strategy is proposed. The obtained results show that high-quality mappings are found. Nevertheless, our approach is transferable to other data compatibility problems. These can be found in the merging of emissions data for other countries, or in bridging the gap between completely different data sets.
The usage of data gathered for Industry 4.0 and smart factory scenarios continues to be a problem for companies of all sizes. This is often the case because they aim to start with complicated and time-intensive Machine Learning scenarios. This work evaluates the Process Capability Analysis (PCA) as a pragmatic, easy and quick way of leveraging the gathered machine data from the production process. The area of application considered is injection molding. After describing all the required domain knowledge, the paper presents an approach for a continuous analysis of all parts produced. Applying PCA results in multiple key performance indicators that allow for fast and comprehensible process monitoring. The corresponding visualizations provide the quality department with a tool to efficiently choose where and when quality checks need to be performed. The presented case study indicates the benefit of analyzing whole process data instead of considering only selected production samples. The use of machine data enables additional insights to be drawn about process stability and the associated product quality.
Recent developments in the area of Natural Language Processing (NLP) increasingly allow for the extension of such techniques to hitherto unidentified areas of application. This paper deals with the application of state-of-the-art NLP techniques to the domain of Product Safety Risk Assessment (PSRA). PSRA is concerned with the quantification of the risks a user is exposed to during product use. The use case arises from an important process of maintaining due diligence towards the customers of the company OMICRON electronics GmbH.
The paper proposes an approach to evaluate the consistency of human-made risk assessments that are proposed by potentially changing expert panels. Along the stages of this NLP-based approach, multiple insights into the PSRA process allow for an improved understanding of the related risk distribution within the product portfolio of the company. The findings aim at making the current process more transparent as well as at automating repetitive tasks. The results of this paper can be regarded as a first step to support domain experts in the risk assessment process.
This paper analyses an electrical test tower of the OMCIRON electronics GmbH and evaluates whether a Predictive Maintenance (PdM) strategy can be implemented for the test towers. The company OMICRON electronics GmbH performs unit tests for its devices on test towers. Those tests consist of a multitude of subtests which all return a measurement value. Those results are tracked and stored in a database. The goal is to analyze the data of the test towers subtests and evaluate the possibility of implementing a predictive maintenance system in order to be able to predict the RUL and quantify the degradation of the test tower.
By assuming that the main degradation source are the relays of the test tower, a reliability modelling is performed which is the model-driven approach. The data-driven modelling process of the test tower consists of multiple steps. Firstly, the data is cleaned and compromised by removing redundances and optimizing for the best subtests where a subtest is rated as good if the trendability and monotonicity metric values are above a specific threshold. In a second step, the trend behaviours of the subtests are analyzed and ranked which illustrates that none of the subtests contained usable trend behaviour thus making an implementation of a PdM system impossible.
By using the ranking, the data-driven model is compared with the reliability model which shows that the assumption of the relays being the main error source is inaccurate.
An analysis of a possible anomaly detection model for a PdM is evaluated which shows that an anomaly detection is not possible for the test towers as well. The implementability of PdM for test towers and other OMICRON devices is discussed and followed up with proposals for future PdM implementations as well as additional analytical analyses that can be performed for the test towers.
With Cloud Computing and multi-core CPUs parallel computing resources are becoming more and more affordable and commonly available. Parallel programming should as well be easily accessible for everyone. Unfortunately, existing frameworks and systems are powerful but often very complex to use for anyone who lacks the knowledge about underlying concepts. This paper introduces a software framework and execution environment whose objective is to provide a system which should be easily usable for everyone who could benefit from parallel computing. Some real-world examples are presented with an explanation of all the steps that are necessary for computing in a parallel and distributed manner.
This master thesis investigates a Computational Intelligence-based method for solving PDEs. The proposed strategy formulates the residual of a PDE as a fitness function. The solution is approximated by a finite sum of Gauss kernels. An appropriate optimisation technique, in this case JADE, is deployed that searches for the best fitting parameters for these kernels. This field is fairly young, a comprehensive literature research reveals several past papers that investigate similar techniques.
To evaluate the performance of the solver, a comprehensive testbed is defined. It consists of 11 different Poisson equations. The solving time, the memory consumption and the approximation quality are compared to the state of the art open-source Finite Element solver NGSolve. The first experiment tests a serial JADE. The results are not as good as comparable work in the literature. Further, a strange behaviour is observed, where the fitness and the quality do not match. The second experiment implements a parallel JADE, which allows to make use of parallel hardware. This significantly speeds up the solving time. The third experiment implements a parallel JADE with adaptive kernels. It starts with one kernel and introduce more kernels along the solving process. A significant improvement is observed on one PDE, that is purposely built to be solvable. On all other testbed PDEs the quality-difference is not conclusive. The last experiment investigates the discrepancy between the fitness and the quality. Therefore, a new kernel is defined. This kernel inherits all features of the Gauss kernel and extends it with a sine function. As a result, the observed inconsistency between fitness and quality is mitigated.
The thesis closes with a proposal for further investigations. The concepts here should be reconsidered by using better performing optimisation algorithms from the literature, like CMA-ES. Beyond that, an adaptive scheme for the collocation points could be tested. Finally, the fitness function should be further examined.
Many test drives are carried out in the automotive environment. During these test drives many signals are recorded. The task of the test engineers is to find certain patterns (e.g. an emergency stop) in these long time series. Finding these interesting patterns is currently done with rule based processing. This procedure is very time consuming and requires a test engineer with expertise. In this thesis it is examined if the emerging field of machine learning can be used to support the engineers in this task. Active Learning, a subarea of machine learning, is used to train a classifier during the labeling process. Thereby it proposes similar windows to the already labeled ones. This saves the annotator time for searching or formulating rules for the problem. A data generator is worked out to replace the missing labeled data for tests. The custom performance measure “proportion of seen samples” is developed to make the success measurable. A modular software architecture is designed. With that, several combinations of Time Series Classification algorithms and query strategies are compared on artificial data. The results are verified on real datasets, which are open source available. The best performing, but computational intensive solution is an adapted RandOm Convolutional KErnel Transform (ROCKET). The custom query strategy “certainty sampling” shows the best results for highly imbalanced datasets.
ÖMG Conference 2019
(2019)
Stress testing is part of today’s bank risk management and often required by the governing regulatory authority. Performing such a stress test with stress scenarios derived from a distribution, instead of pre-defined expert scenarios, results in a systematic approach in which new severe scenarios can be discovered. The required scenario distribution is obtained from historical time series via a Vector-Autoregressive time series model. The worst-case search, i.e. finding the scenario yielding the most severe situation for the bank, can be stated as an optimization problem. The problem itself is a constrained optimization problem in a high-dimensional search space. The constraints are the box constraints on the scenario variables and the plausibility of a scenario.
The latter is expressed by an elliptic constraint. As the evaluation of the stress scenarios is performed with a simulation tool, the optimization problem can be seen as black-box optimization problem. Evolution Strategy, a well-known optimizer for black-box problems, is applied here. The necessary adaptations to the algorithm are explained and a set of different algorithm design choices are investigated. It is shown that a simple box constraint handling method, i.e. setting variables which violate a box constraint to the respective boundary of the feasible domain, in combination with a repair of implausible scenarios provides good results.