Detecting Quality Losses – Reducing Low Weight Complaints

The Challenge

During product changeovers, the chocolate tank empties, slightly lowering the dosage. This can lead to low-weight products and customer complaints. Operators must manually increase rotor speed, but it's hard to verify if they do it.

‍

The Goal

• Detect emptying events >15 min
• Verify rotor speed adjustments
• Enable traceability
• Support daily ops with a dashboard

‍

The Approach

• The team applied Aggregations in TrendMiner to capture process dynamics, calculating rotor range and maximum tank levels over 20-minute moving windows. This provided a clearer view of how emptying unfolded beyond raw signal noise.
• Using Value Based Searches the system automatically detected good and bad emptying periods of at least 15 minutes by combining level conditions (max > 60 & current < 50) with rotor activity (range > 0 / range = 0).
• With Event Contextualization in TrendMiner,  these periods were labeled as successful or failed emptying events and directly compared across historical signals, highlighting differences in operating conditions.
• A dedicated Dashboard then brought everything together—event counters, historical trends, and real-time tiles of key variables—into a single monitoring view.

Insight: With these TM features, you gain a clear, contextualized view of operator actions during emptying, improving quality control and operational efficiency.

‍

‍

The Results

‍

‍

The Takeaway

Real-time monitoring improved traceability, reduced complaints, and ensured consistent dosing during critical periods.

‍

‍