In recent times, both geopolitical challenges and the need to counteract climate change have led to an increase in generated renewable energy as well as an increased demand for clean electrical energy. The resulting variability of electricity production and demand as well as an overall demand increase, put additional stress on the existing grid infrastructure. This leads to strongly increased maintenance demands for distribution system operators (DSOs). Today, condition monitoring is used to address these challenges. Researchers have already explored solutions for monitoring critical assets like switchgear and circuit breakers. However, with a shrinking knowledgeable technical workforce and increasing maintenance requirements, mere monitoring is insufficient. Already today, DSOs ask for actionable recommendations, optimization strategies, and prioritization methods to manage the growing task backlog effectively. In this paper we propose a vision of a grid-level cognitive assistance system that translates the outcome of diagnosis and prognosis systems into actionable work tasks for the grid operator. The solution is highly interdisciplinary and based on empirical studies of real-world requirements. We also describe the related work relevant to the multi-disciplinary aspects and summarize the research gaps that need to be closed over the next years.

Enterprise Modelling and Information Systems Architectures – International Journal of Conceptual Modeling

As AI models become more and more common in business and even in our daily lives, it is important to understand what the carbon impact of these models is. Recent papers have shown that this impact can be quite great, i.e., the training of a single high-end model can result in emissions of more than 500t of CO2eq. In this paper we discuss the factors that influence the carbon footprint of AI models, explore what impact different decisions have, and show how the footprint can be reduced. We also examine the footprint of different models to give a guideline on how urgent action is for different organizations.

As deep learning AI becomes more and more common in business and even in our daily lives, it is important to understand what the carbon impact of this type of software is. Recent papers have shown that it can be quite great, i.e., the training of a single high-end model can result in emissions of more than 500t of CO2eq. In this article we describe a life-cycle-focused framework to estimate the carbon drivers of a new deep learning model. We experimentally verify some claims in the literature and provide suggestions on how to reduce the carbon footprint of a deep learning-based offering. The article should enable developers and managers to make informed and meaningful decisions to minimize their ML projects’ sustainability impact.

In this paper, we describe a machine learning approach for predicting machine health indicators with a large time horizon into the future. The approach uses state-of-the-art neural network architectures for sequence modelling and can incorporate numerical-sensor and categorical data using entity embeddings. Moreover, we describe an unsupervised labelling approach where classes are generated using continuous sensor values in the training data and a clustering algorithm. To validate our approach, we performed an ablation study to verify the effectiveness of each of our model’s components. In this context, we show that entity embeddings can be used to generate effective features from categorical inputs, that state-of-the-art models, while originally developed for a different set of problems, can nonetheless be transferred to perform industrial asset health classification and provide a performance boost over simpler networks that have been traditionally used, such as relatively shallow recurrent or convolutional networks. Taken together, we present a machine health monitoring system that can accurately generate asset health predictions. This system can incorporate both numerical and categorical information, the current state-of-the-art for sequence modelling, and generate labels in an unsupervised fashion when explicit labels are unavailable.