Perdue Foods - Driving Perdue's Data Transformation and Dive Into Industry 4.0

Name:

Mike Lindsey, Director of Manufacturing Systems
Amanda Hynous, Manufacturing Systems Development Manager
April Carver, Senior Systems Analyst
Harry Neumann, Senior Systems Analyst
Ryan Moore (vendor), VP of Data Science Solutions at Excelion Partners

Country: United States

Organization: Perdue Foods

Perdue Foods (“Perdue”) stands apart. In business for over 100 years, we are a fourth-generation, family-owned American food, and agriculture business.

Awards Categories:

  • Partner Acceleration

 

Business Challenge:

Perdue uses automated machines to debone breast meat. These machines have several key data points that drive throughput and yield. One of the key data points is loading efficiency. Basically, the product needs to be loaded onto the carrier to be processed by automation manually. Knowing how effective operations are with keeping the carriers fully loaded drives profitability. Other key items are downtime, error detection, and speed of the line.

Perdue Foods (“Perdue”) did not have a standardized system in place to capture and report on the machine efficiency, machine speed, or downtime data for the various WMDeboning machines in our plant operations. Sites independently report on efficiency metrics using various reports or manual methods which were inconsistent from site to site. What's more, Perdue could not currently systematically compare the machine performance across multiple sites due to process and system limitations.

 

Business Solution:

The key to solving the machine data visibility issue was to collect the data from the IIOT device, place it in a data lake, transform the data with Dataiku, and then build business visualizations that crossed many machines across multiple production plants located across the United States.

Provide Operations Management with a centralized performance reporting dashboard across all nine sites which operate 29 White Meat Deboning (WMDebone) machine/processing centers. The dashboard will provide the near-real-time status of each WMDeboning machine's throughput, downtime, and loading efficiency. It will allow for aggregating the entire WMDeboning process across the enterprise, while also allowing for drill-down visibility into each individual machine's performance.

We developed the solution with a partner who was experienced with Dataiku and was able to show Perdue how easy it was to use. This data analytics platform consists of many ecosystem partners, Dataiku being the primary for data ingestion and transformation. We have used some of the advanced visualizations to see insights we were not actually expecting. Things like when operations speed up the line past a certain point, machine faults, and failures increase. This value was extremely helpful in reducing downtime, saving Perdue from having to pay overtime to complete the day's production.

Business Area: Manufacturing

Use Case Stage: In Production

 

Value Generated:

With the help of our consultant partner, we architected a landscape utilizing Dataiku as the ingestion tool to gather the machine data from our local (on-prem) SQL database servers and store the data in a Snowflake RAW warehouse. We then leveraged the power of Dataiku to transform the data into usable consumer datasets, which we share with the users through our existing Power BI reporting structure.

image1.png

 

 

Value Brought by Dataiku:

By providing the business with meaningful data, the effort has resulted in a 1% reduction in downtime (equates to 6.8M more pounds) and an $850k reduction in labor.

Bringing Dataiku into Perdue for manufacturing showed the rest of Perdue, departments like Sales, Marketing, Procurement, Supply/Demand planning, as well as others that the tool could be used by all teams. We have now standardized our data analytics platform, and Dataiku is a major part of the ecosystem. We have now completed over seven different data analysis projects successfully saving Perdue over $2.3M in the first year of operations.

Value Type: Reduce cost

Value Range: Hundreds of thousands of $

Share:
Version history
Publication date:
01-08-2024 03:05 PM
Version history
Last update:
‎08-25-2022 05:02 PM
Updated by:
Contributors