A special issue article on the topic of Fully Integrated and System-Optimized Electronic Solutions on Solar Modules has been published in the journal “Prograss in Photovoltaics”. It deals with research results on so-called AC-PV modules and focuses on methods for further simplification and cost reduction in the field of photovoltaic electrical energy generation. The article is available here: http://doi.org/10.1002/pip.3909 

A second paper was written by Mr. Jan-Philipp Roche as part of his research on modeling electrical components and networks using machine learning. Existing so-called black-box modeling approaches in machine learning typically suffer from a fixed combination of input and output functions. A new approach to reconstruct missing variables in time series is presented. This also makes it possible to change the input and output combination of a trained neural network. The aim is to improve the modeling of electrical structures under aspects of electromagnetic compatibility. The article entitled: “Using Autoencoders and Automatic Differentiation to Reconstruct Missing Variables in a Set of Time Series” has been published by Springer Nature Computer Science and is available here: https://link.springer.com/article/10.1007/s42979-025-03798-5

A special issue article on the topic of Fully Integrated and System-Optimized Electronic Solutions on Solar Modules has been published in the journal “Prograss in Photovoltaics”. It deals with research results on so-called AC-PV modules and focuses on methods for further simplification and cost reduction in the field of photovoltaic electrical energy generation. The article is available here: http://doi.org/10.1002/pip.3909 

A second paper was written by Mr. Jan-Philipp Roche as part of his research on modeling electrical components and networks using machine learning. Existing so-called black-box modeling approaches in machine learning typically suffer from a fixed combination of input and output functions. A new approach to reconstruct missing variables in time series is presented. This also makes it possible to change the input and output combination of a trained neural network. The aim is to improve the modeling of electrical structures under aspects of electromagnetic compatibility. The article entitled: “Using Autoencoders and Automatic Differentiation to Reconstruct Missing Variables in a Set of Time Series” has been published by Springer Nature Computer Science and is available here: https://link.springer.com/article/10.1007/s42979-025-03798-5