Primary food processing
General Information
Challenge, Value & Description
Performance, Access & Contact
1 – General Information
Partners
Sectors addressed
Application categories covered
Lifecycle level covered
Digital Engineering
Planning & Commissioning
Smart Production & Operations
Smart Logistics
Smart Maintenance
Customer Service
Circularity
AV. GERMANÍAS, 49, 46291 Benimodo, Spain
Geographical Scope
- Europe
2 – Challenge, Value & Description
Challenge
Water is intensively used in food processing, particularly in frozen food processing, from the blanching and cleaning of the ingredients that arrive from the fields to the cleaning of the production lines and the working areas (to meet regulatory and quality standards). The freezing tunnels that freeze the food, is also demanding assets in terms of water consumption. Water is a scarce resource in many agricultural regions across Europe, and an efficient use of water on these processes is a cornerstone to ensure the sustainability of food processing. Therefore, this use case pilot is closely related to sustainable manufacturing challenges.
Value
The Pilot focuses on the efficient use of water and the efficiency of the water installation, in particular:
- Water consumption observability through water meters and IoT enabling advanced detection capabilities and allowing the company to collect valuable training datasets for predictive models
- Water pump predictive maintenance focuses on the maintenance of the water pumps’ efficiency and availability by predicting water pump failures, anticipating maintenance needs. These actions lead to increased energy and water efficiency and availability, promoting a better use of water.
Description
Data from the sensors installed on the plant and its warehouse are collected from the Pilot and shared through a Data Space with the technical partner that develops AI models through a MLOps platform. On the MLOps platform the technical partner develops, trains and deploys the AI models, making them available for the Pilot to download the models from cloud to edge.
Data Value Chain Description
Infraestructure Elements
- Private Cloud
- (Open) Edge node
- Field Devices
3 – Performance, Access & Contact Info
Performance
Water consumption anomaly detection (i.e. due to water leakages, open hoaxes, or freezing): validation and testing of different sensors configuration, and rules and thresholds for detection models. The models developed supported the Pilot in enhancing the detection capabilities of maintenance managers, leading to a better efficiency in water consumption.
Water pump predictive maintenance: aims to reduce the probability of water supply failure and reduce water supply downtime. So far, prediction models, risk analysis and optimization models have been developed, tested and validated, but the experimentation is still ongoing.
Lessons Learned & Observations
- Operators experience and feedback is crucial to set models rules,
- Data availability is crucial to optimal modelling,
- The methodology proposed relies on continuous improvement methodologies to gather qualitative feedback and impact on KPIs.
Replication Potential & Feasibility Assessment
While the experimentation is still ongoing with models fine tuning, ARDO is evaluating the opportunity to scale the pilot to other plants of the group in other countries.