The Challenge
A manufacturing facility (anonymized for confidentiality) relied on reactive maintenance for its knife replacement process. Unexpected downtime due to dull knives was frequent and costly, disrupting production and impacting overall efficiency. There was no system in place to predict when knives would need replacement.
The Problem
The facility faced several key challenges:
- Unexpected component replacements (knives) lead to unplanned production halts.
- A complete lack of insight into remaining knife life, makes timely replacement planning impossible.
- No method for calculating the percentage of knife life remaining before replacement.
- Reactive maintenance, leading to rushed replacements and potential safety concerns.
The Solution: Listening to the Knives – A Proactive Approach
To address these challenges, a novel sound-based predictive maintenance solution was implemented. This marked a significant shift from their previous reactive approach. The new system:
- Gathers information by sound, capturing the acoustic signature of the knives during operation.
- Utilizes a learned AI/ML model trained to recognize varying levels of knife sharpness based on sound.
- Predicts when a component replacement (knife) should occur.
- Provides a “compliment” – a calculation of the percentage of knife life remaining before replacement is needed.
Results & Impact: From Reactive to Proactive
The transition from reactive to predictive maintenance yielded significant improvements:
- Eliminating unplanned downtime due to dull knives (or a substantial reduction – quantify if possible).
- Accurate prediction of remaining knife life, enabling proactive replacement scheduling.
- Optimized knife inventory management, ensuring knives were available when needed.
- Improved safety through planned maintenance and reduced rushed replacements.
By leveraging sound analysis and AI/ML, the facility gained valuable insights into the condition of its cutting knives, something it lacked entirely before. This enabled a complete shift from reactive to predictive maintenance, improved efficiency, reduced downtime, and enhanced safety.
The “compliment” feature, providing a percentage of remaining knife life, further enhances planning and resource allocation. This case study highlights the transformative power of sound-based predictive maintenance to optimize critical processes and minimize disruptions in manufacturing environments, especially when moving from a purely reactive maintenance model.
