Use Case 1.1 Energy Price Prediction

Powerplant facilities are constantly faced with ramp up or ramp down decisions due to prioritized renewable energies. Future electricity prices must be available several days ahead, as coal power plants are inert. Electricity prices fluctuate due to geopolitics, unplanned outages, and an increasing share of renewables. Competitors offer just blank predictions without any model description or explainability.

Use Case 1.2 Operator Support / Anomaly Detection

AI-based anomaly detection models have the potential to uncover unusual plant situations that could easily go unnoticed. A common limitation of anomaly detection is the lack of explainability of AI black box models and the lack of known historic anomalies (labeled data) when building these models. This use case aims to provide more robust and comprehensive anomaly detection systems for plant operators, with the help of interactive XAI-based anomaly explainers.

Use Case 2.1 Maintenance Ticket Analysis

This use case focuses on operators who are responsible for searching and identifying Technical Reports (TR) from all parts of metal manufacturing plants. A semantic analyzer engine allows TR to be analyzed for their meaning and not just for the presence of certain keywords. By using large language models, we can improve the accuracy of the results and provide more relevant information to users.

Use Case 3.1 Use Life Estimation

Electronic manufacturing contains many inspection steps, like after the Printed Circuit Board (PCB) assembly or after the final assembly of the product, to guarantee high-quality products. The goal of this use case is to connect process variations identified by visual inspection with lifetime performance using AI and ML, specifically deep learning. This aims to improve defect detection performance, minimize false calls and non-conforming products, gain new insights into process optimization, and reveal design changes to enhance product lifetime.

Use Case 3.2 Visual Product Inspection on Electronics

To guarantee high-quality products, electronics manufacturing contains many inspection steps, like after PCB assembly (using machines from MEK) or after the final assembly of the product. The inspection types include checks for presence, alignment, or damage. Some are performed manually or by algorithms that do not use deep learning, resulting in the slip of non-conforming products, high false-positive rates, and lowered traceability of defects. Automating these inspections using state-of-the-art explainable deep learning algorithms is vital to support a scalable and sustainable electronics industry. This will allow for quicker introduction of new products, higher quality inspections, and shorter design cycles to improve existing production processes.

Use Case 4.1 Flotation

This use case covers the flotation process, which separates valuable minerals from the mined ore’s overburden, and focuses on how optimizing this process through AI and machine learning forecasting enables operators to proactively control the process and improve the production of valuable minerals, leading to economic benefits and operational efficiency.

Use Case 5.1 Vibration Monitoring

The EXPLAIN project partner Södra employs hundreds of vibration sensors on assets like motors, gearboxes, pumps, and fans. By analyzing vibration monitoring data, they aim to predict equipment failures and irregularities, allowing maintenance professionals to identify patterns and address potential challenges proactively. This process includes detailed event information and user feedback to enhance the accuracy of machine learning models for more informative alerts.

Use Case 5.2 Kappa Sensor

This use case focuses on improving the forecasting of the Kappa value, a critical quality indicator in the pulping process, to maintain high paper quality. By using advanced process control (APC) and predictive models, operators can anticipate deviations and make proactive adjustments. This approach enhances the stability and efficiency of the pulping process, directly impacting the quality of the final paper product.