Quality prediction and driver analysis for french fry production

Quality prediction and driver analysis for french fry production

Challenges 

  

The french fries producer was facing difficulties in tracking fries back to potatoes. The fries manufacturing process comprised of 21 steps, which resulted in various trackability problems. The management wanted to understand the different factors that drive the fries’ quality and optimizing system settings based on potatoes. 

  

Solutions 

  

Neal Analytics pulled data from various sensors and created unit traceability across a 21-step process using Azure Data Factory. We enabled real-time Power BI visualization of data for real-time, shop-floor decision-making. Machine learning algorithms were deployed to predict fries’ length and increase batch quality for increased yield.    

  Power BI dashboard

Results 

  

Implementing machine learning helped identify key drivers of fry length, which determines how much output is worth. This has resulted in increased fries’ quality and line yield. Real-time Power BI visualization has helped identify and intercept the quality issues, which resulted in higher manufacturing line yield.