Matthias Zech

Mr. Matthias Zech


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
30 September 2024 in Advances in Applied Energy
Verified
Journal Article
Open access
MZMatthias Zech
Matthias Zech
HTHendrik-Pieter Tetens
Hendrik-Pieter Tetens
JRJoseph Ranalli
Joseph Ranalli

1 November 2024 in Journal of Renewable and Sustainable Energy
Verified
Journal Article
Closed access
PDF file
JRJoseph Ranalli
Joseph Ranalli
MZMatthias Zech
Matthias Zech
HTHendrik-Pieter Tetens
Hendrik-Pieter Tetens
Data about behind-the-meter photovoltaics (PV) installations may be difficult to obtain for researchers. A number of investigators have considered deep learning as an attractive solution to this challenge, capable of directly identifying PV installations from aerial or satellite images. Deep learning models are well known to experience challenges when working with data from sources that they have never been exposed to. This study investigated whether generalizability can be improved by diversifying training data across available labeled data sources. We assessed the performance of models trained on all possible combinations of six different labeled datasets of aerial PV imagery, with a fixed number of total training images. Unfortunately, our results indicate that no combination of model training data achieved generalized performance that approaches models trained on data from a target data source. This implies that generalized ResNet models cannot be developed simply by modifying the configuration of the training data. Consequently, researchers should expect that some degree of data labeling is likely to be necessary when adapting these models to new applications, but our results do indicate that significant performance improvements are possible with only small (∼20%) introductions of target data. Future work may investigate alternative architectures, expanded training datasets, or ways to reduce the amount of labeled data necessary to adapt a model for a given application.

28 February 2024 in Applied Energy
Verified
Journal Article
Open access
MZMatthias Zech
Matthias Zech
LBLueder von Bremen
Lueder von Bremen

8 September 2021 in Energies
Verified
Journal Article
Open access
PDF file
JDJan Dasenbrock
Jan Dasenbrock
APAdam Pluta
Adam Pluta
MZMatthias Zech
Matthias Zech
WMWided Medjroubi
Wided Medjroubi
Energy system modeling is essential in analyzing present and future system configurations motivated by the energy transition. Energy models need various input data sets at different scales, including detailed information about energy generation and transport infrastructure. However, accessing such data sets is not straightforward and often restricted, especially for energy infrastructure data. We present a detection model for the automatic recognition of pipeline pathways using a Convolutional Neural Network (CNN) to address this lack of energy infrastructure data sets. The model was trained with historical low-resolution satellite images of the construction phase of British gas transport pipelines, made with the Landsat 5 Thematic Mapper instrument. The satellite images have been automatically labeled with the help of high-resolution pipeline route data provided by the respective Transmission System Operator (TSO). We have used data augmentation on the training data and trained our model with four different initial learning rates. The models trained with the different learning rates have been validated with 5-fold cross-validation using the Intersection over Union (IoU) metric. We show that our model can reliably identify pipeline pathways despite the comparably low resolution of the used satellite images. Further, we have successfully tested the model’s capability in other geographic regions by deploying satellite images of the NEL pipeline in Northern Germany.

1 August 2024 in Monthly Weather Review
Verified
Journal Article
Closed access
PDF file
MZMatthias Zech
Matthias Zech
LBLueder von Bremen
Lueder von Bremen
Dynamical numerical weather prediction has remarkably improved over the last decades. Yet postprocessing techniques are needed to calibrate forecasts which are based on statistical and machine learning techniques. With recent advances in the derivation of year-round, large-scale atmospheric circulations, or weather regimes, the question arises of whether this information can be valuable within forecast postprocessing methods. This paper investigates this by proposing a bias correction scheme to integrate the atmospheric circulation state derived from empirical orthogonal functions, referred to as weather patterns, for deterministic short-term, near-surface temperature forecasts based on least absolute shrinkage and selection operator (LASSO) regression. We propose a computational study which first evaluates different weather pattern definitions (spatial domain) to improve temperature forecasts in Europe. As a bias could be associated with the weather pattern at the model initialization time or at the realization time of the forecast, both variants are tested in this study. We show that forecasted weather patterns with the identical spatial domain as the forecast show best skill reaching mean-square-error skill improvements of up to 3% (day ahead) or 1% (week ahead). Only considering land surface improvements in Europe, improvements of 4%–6% for day-ahead forecasts and 1%–5% for week-ahead forecasts are observable. We believe that this study not only introduces a simple yet effective tool to reduce bias in temperature forecasts but also contributes to the active discussion of how valuable weather patterns are and how to use them within forecast calibration techniques.

15 September 2025 in Solar Energy
Verified
Journal Article
Open access
MZMatthias Zech
Matthias Zech
AHAnnette Hammer
Annette Hammer
LBLueder von Bremen
Lueder von Bremen