Using Ultrasound for Monitoring Metal Moving Tools in the Food Processing Industry and Implementing AI Condition-based Maintenance Planning
Embedded Files
What?
What?
ASCALIA (Croatia) and ONCONTROL (Portugal) aim to implement smart maintenance planning in the food processing sector by combining monitoring of ultrasounds created by metal moving parts of the food processors and AI-based condition-based maintenance. Employing a broad ultrasound range sensor for real-time monitoring of the food processors, insights into moving metal parts integrity status will be provided, enabling timely and efficient maintenance planning aiming at zero pollution of processed food by metal foreign objects (MFB). This innovative, more accurate, and precise maintenance planning will significantly increase food safety and consequently reduce recalls and market withdrawals.
The new solution will utilise Artificial Intelligence (AI), Machine Learning (ML), ultrasonic detection, and the Internet of Things (IoT) technologies. Metal parts that are damaged emit different ultrasonic spectra compared to those that are undamaged, but the noisy environment in food production facilities can hinder the detection of these changes. By employing a broad-range, sensitive sensor, it is possible to effectively detect and monitor signal changes within a quieter ultrasonic spectral range. Integrating AI and ML with track and trace principles makes it possible to pinpoint the most critical maintenance moments to prevent MFB pollution (AI Condition-based Maintenance Planning). IoT will be leveraged to efficiently gather data, analyse it, and make informed decisions.
Why?
Why?
It has been estimated that the presence of metallic foreign bodies (MFB) in food may lead to serious health risks, ranging from 1-5%. The USA experienced 25 instances of food recall in 2022, leading to the disposal of millions of kilograms of food. Despite advancements in detecting MFB, working conditions in the food industry still limit 100% detection. Combining Oncontrol's efficient ultrasound detection with Ascalia's AI-powered traceability platform offers an effective solution to plan maintenance and prevent the creation of MFB by food processors. This innovative approach addresses key challenges in the modern food industry and enhances food safety.
Main development targets
Main development targets
Spectrum identification, hardware implementation, and testing
Spectrum identification, hardware implementation, and testing
Study and compare the collected signals for the identification of the ultrasound spectrum range;
Integrate ultrasound sensors with sensitivity in the identified range;
Develop food-grade retrofit hardware setup;
Test the hardware in real-world conditions.
AI software development, model training, and validation
AI software development, model training, and validation
Clean and pre-process the collected data;
Design different models, incorporating expert knowledge and feature engineering;
Train the models, fine-tune them, and perform initial internal validations;
Validate the models using industry-standard metrics.
User interface development and production optimisation
User interface development and production optimisation
Desenvolvimento da interface do utilizador e otimização da produção
Desenvolver uma interface básica e de fácil utilização com o envolvimento precoce do utilizador final e aperfeiçoá-la iterativamente;
Otimizar o processamento de dados e o desempenho dos algoritmos.
Partners
Partners
SharpSense project has received funding from HAMAG-BICRO and ANI through Eureka and the Eurostars programme, which is co-funded by the European Commission as part of the European Partnership on Innovative SMEs.
This project is a result of the project 19009 - SharpSense supported by Programa Regional do Centro (COMPETE 2030), under the PORTUGAL 2030 Partnership Agreement, through the European Regional Development Fund (ERDF)
Page updated
Google Sites
Report abuse