Data pattern recognition reveals the ‘TRUTH’ in complex processes

Situation:

Modern production is complex and the amount of production data available is increasing rapidly. However, data is not readily available to everyone that needs it.

Challenges:

  • Take full advantage of valuable production data (or risk falling behind competitors).
  • Identify root causes within complex processes.
  • Ensure correct data is used for optimizing production.

Success:

An ideal solution will help to…

  • Enable better and faster decisions.
  • Identify production deviations at an early stage.
  • Find patterns and root causes in a big data environment.

Solution:

Powerful and easy to use analysis tools that enable you to dive deep into your production data and generate valuable insights about complex processes.

A SHORT CASE STORY:

A plant is losing money on one of its batch processes. Plant management want to know the cause of the problem…

Operations Management: “The problem is caused by operators not strictly following the temperature profiles during the first ‘React’ operation.”

Operators: “No, the problem is something else.”

Erik Jansson
Erik JanssonSenior Specialist Consultant, Manufacturing Information
Tel: +46 (0)70 647 31 54

Question: Is batch cost related to heat cycle time?

Question: Do good batches have a common temperature profile?

Analysis: Use pattern recognition function to identify occurrences of the pattern in historical data.

Answer: Not all good batches fit the pattern.

Answer: Not all batches that fit the pattern are good.

Conclusion: Whether or not the operators are following the prescribed temperature profile… (1) the profile does not give consistently good batches; (2) the profile is not represented in all good batches.