ITP Aero – Improving the Broaching Process Using ML
ITP Aero
Sector: Industry
Business Case
It seeks to improve both the capacity (variability reduction) and quality of broached products at ITP Aero through data-driven modelling.
Objectives
Process capability improvement (reduction of dimensional variability) and quality improvement (reduction of allowances) through the use of ML-based correctors for the entire broaching line (4 machines) in Zamudio.
Use case
Identify all parameters likely to affect the process. Build software to assist in the recording and storage of information. Model it. Put a 24/7 support system into production for the Zamudio factory.
Infrastructure
On Premise
Technology
Machine learning and deep learning
Data
Time series Private dataset
Resources
IT staff (2), process engineers (3), manufacturing technologists (2), data engineers (1) and data scientists (2).
Difficulties and learning
Dirt in manually entered data. Necessity to automate data logging.
KPIs (business impact and metrics of the model)
Number of concessions per piece. Cp, Cpk, Pp, Ppk.
Funding
–
Collaborators, Partners
No
