The aim is to develop a torque sensor with artificial intelligence (AI) that detects inefficiencies and suggests preventive maintenance measures. This is intended to increase the longevity and sustainability of the sensor and the systems it monitors.
In contrast to static and wired measurements with force sensors, rotating measuring shafts with wireless signal transmission place higher demands on development. In addition to the mechanical design of the sensor, more complex questions about the interference-resistant, electromagnetic transmission of the sensor signals must also be taken into account. The manufacturing and adjustment processes are more complex due to the larger number of components.
The combination of torque measurement technology and AI-based data processing represents a significant challenge for the team and goes far beyond routine tasks. This includes modeling, data acquisition and preparation of training data for the AI.
Conventional systems often use a method of transmitting measurement signals from a rotating shaft based on electrical sliding contacts. However, this method has disadvantages as it leads to wear and losses and is maintenance-intensive.
In contrast, the project solution uses a contactless method of transmitting measurement signals from the rotating shaft, thus avoiding the disadvantages mentioned above.
Some competitors also offer contactless methods for torque measurement. However, these systems lack integrated sensor-based data processing based on artificial intelligence (AI). This enables detailed process analysis and provides valuable information on inefficiencies and preventive maintenance measures.
The aim is to develop a new type of system for contactless torque measurement on rotating shafts with real-time data acquisition, transmission and AI-supported data processing on a processor integrated in the sensor. A special electronic unit must be developed and suitable processor technologies selected to achieve the required space and energy efficiency. AI models must be designed for data processing and trained with simulated and measured data. It should be noted that the execution takes place on an embedded system with limited computing capacity and low energy consumption.
The interdisciplinary nature of the development work, including areas such as mechanical design, rotor dynamics, electronics, electromagnetics, embedded programming and AI models, carries an inherent risk. A product that is suitable for both industry and manufacturing must meet very high requirements and be robust against manufacturing and material tolerances as well as process variations.
The challenge is in particular to achieve robustness for embedded AI applications, an area in which there is no experience or reference to date. In addition, the implementation of complex AI models in compact and energy-efficient embedded systems is a demanding task. The risk is increased by the need to generate a sufficient amount of training data for the AI, which must also meet the requirements in terms of quality and variance.
Despite the significant risks, the project represents an opportunity to drive technological innovation. With the right team, robust design principles, intensive research and effective data and risk management, the challenges can be overcome.
This project has the potential to pave the way for future developments in this exciting field. It shows that with risk comes great opportunity.
Deutsch 
español 
Ελληνικά