Thesis Artifacts for: Forecasting Power Consumption of Manufacturing Industries Using Neural Networks

Thesis Artifacts for: Forecasting Power Consumption of Manufacturing Industries Using Neural Networks A detailed description can be found in the README.md. The corressponding thesis can be found here. Abstract The reduction of power consumption should be reached for many ecologic and economic reasons. Since a large part of the power is consumed by the industrial sector, we propose to increase the energy efficiency and applying the DevOps approach. Forecasting the power consumption of manufacturing industries can increase the energy efficiency and also allows using anomaly detection systems and predictive maintenance for manufacturing industries. In this work, we designed multiple different model variations, using different neural networks, forecasting methods and features for the forecasting of three industrial power consumers. We also evaluated our different model variations with the aim to find the model variation with the highest accuracy. By providing the evaluations, we offer conclusions about different forecasting methods, features used and different neural networks for forecasting the power consumption of industrial power consumers. We also propose a forecasting microservice as a base for an anomaly detection and predictive maintenance