Digital twins enable the monitoring, analysis, and optimization of processes throughout the entire product lifecycle. Additionally, predictions and decisions can often be made independently, often with the help of artificial intelligence. Through automated data flow, the physical entity and digital entity continuously influence each other.
The use of digital twins allows for the control of processes during operation with the assistance of the model. This enables numerous new application possibilities for products and systems, such as early anomaly detection for predictive maintenance and quality assurance, or better achievement of a target state through planning and optimization during and before operation.
The “System Modeling, Digital Twins” research group uses data-based models in addition to purely physically based models. Neural networks are trained using physical models and reinforcement learning (an AI method), which can then control a complex system. The use of physical models allows any number of data sets to be generated for training the neural network.
Intellibake - A Digital Twin Revolutionizes the Baking Process
Researchers at HSLU – Technik & Architektur are developing an AI-based controller for ovens that could potentially replace traditional temperature control methods. With additional sensors in the oven and a model of the baking process, this new controller aims to monitor and control the condition of the food being cooked.
Visit project
