SAving Energy by Learning and ImproviNG
logic-based optimization models
logic-based optimization models
Learn more
What is SAELING?
SAELING is an innovative research project that aims to develop cutting-edge artificial intelligence (AI) methods for optimizing energy consumption in industrial settings. Our goal is to make manufacturing processes more sustainable by reducing energy waste, carbon emissions, and costs.
With the collaboration of leading research institutions, industry partners, and experts from Austria and beyond, we are pushing the boundaries of AI-based optimization techniques to make industrial manufacturing more efficient and environmentally friendly.
With the collaboration of leading research institutions, industry partners, and experts from Austria and beyond, we are pushing the boundaries of AI-based optimization techniques to make industrial manufacturing more efficient and environmentally friendly.
The SAELING Use Case
In the metalworking industry, a typical band saw machine consumes around 8.4 MWh per year. Voestalpine provides services for approximately 2,500 metalworking machines. Optimizing the power consumption of these machines is challenging because many factors must be considered. The specification of sufficiently accurate physical models of such machines with reasonable effort is impossible. To mitigate the modeling complexity, SAELING will integrate machine learning methods with scheduling in multi-objective optimization problems. SAELING will find energy- and resource-saving strategies in the industrial use case of metalworking.
Overview
Optimization of energy consumption in the metal processing industry
High energy consumption for sawing, grinding, milling
A typical band saw machine consumes ~8.4 MWh per year
Voestalpine services around 2,500 such machines
(∑ ~21 GWh per year, equivalent to approximately 5,000 average households!)
Status quo: Energy consumption not optimized
No sufficiently accurate physical models available
SAELING will combine Machine Learning with Multi-Objective Optimization.
High energy consumption for sawing, grinding, milling
A typical band saw machine consumes ~8.4 MWh per year
Voestalpine services around 2,500 such machines
(∑ ~21 GWh per year, equivalent to approximately 5,000 average households!)
Status quo: Energy consumption not optimized
No sufficiently accurate physical models available
SAELING will combine Machine Learning with Multi-Objective Optimization.
Our Research Goals
Goal 1: Reduction of energy consumption
The most crucial practical goal is the reduction of energy consumption and the trade-off with other objectives like the wear of tools, labor costs, or customer satisfaction. An additional goal is optimizing the total cost of ownership.
Goal 2: Uncertain multi-objective decision making
SAELING will develop methods and algorithms for optimizing decision-making, integrating the scheduling of tasks, and control machine settings. This optimization must consider the uncertainty of the empirically learned decision models. The optimization problem is multi-objective because there may be solutions that cannot be compared. E.g., it may be difficult for the management to decide on a general level the importance of reducing energy consumption versus maximizing customer satisfaction (e.g., meeting deadlines).
Goal 3: Transfer learning and decision-focused learning
SAELING aims to learn accurate energy consumption prediction models, first at the individual machine level; then across heterogenous machines incl. the ability to make inferences for a machine for settings outside of its common operations (transfer learning); and finally, ML models tuned specifically for the optimization/scheduling problem at hand (decision-focused learning), in a way that takes into account not only the individual errors of the predictions, but also their interaction effect on the scheduling and the quality of the total schedule.
