Process optimisation in an aluminium company
How an early warning model built on data measured before the production run started cut the repetition caused by quality defects by 23 percent. Prospero does not name the manufacturer.
Catching the defect before the process starts
A flow process cannot be stopped once it is running, so a defect found at the end of it is a whole run wasted. The work moved the check to the only point where it could still change the outcome.
At a glance
- Sector
- Manufacturing
- Solution
- DetectX® Business Monitoring
- Customer
- Not published
- Location and date
- Not published
The story, as Prospero publishes it
Starting situation
Defects found at the end of a process that cannot be interrupted
- The manufacturer was experiencing high costs for quality defects that were only identified at the end of the production process of aluminium blocks. The production was a flow process that could not be interrupted once started. The finished product was inspected for defects at the end of the process, and any items that did not meet the quality standard had to be discarded and returned to the start of the manufacturing process.
Project
An early warning model on data measured before the run
- Prospero created an early warning model with DetectX® Business Monitoring on the basis of available input data, which are measured before the production process is started.
Results
Less repetition, less waste, less energy
- Reduction of process repetition due to lack of quality by 23 percent, resulting in significant savings. Production time was saved and wastage reduced, which, combined with reduced energy costs, resulted in significant cost savings. The live page gives no figure for any of those three savings beyond the 23 percent.
What the model returned
One figure, what the model reads, and what the source leaves unmeasured. The 23 percent is the only measurement on the source page, and it is attached to no named institution.
Published result
23% less repetition
The reduction in process repetition caused by a lack of quality, as Prospero states it. No baseline, period or measuring party is published.
Before the run, not after
The model works from input data measured before the production process starts, which is what makes an early warning possible in a flow process that cannot be stopped.
Three savings, no figures
The source reports saved production time, reduced wastage and lower energy costs alongside the 23 percent, and attaches a figure to none of the three.
There is no customer quotation on this page because the source has none, and no customer name because the source gives none. A precise percentage attached to an unnamed institution is the weakest kind of claim on this site, and the page would rather say so than dress it up.
Where to look next
The sector page for work outside financial services, the platform the model was built on, and the two other manufacturing studies.
- Beyond financial services
- Where DetectX® is applied outside banking and insurance.
- DetectX® platform
- The predictive analytics layer an early warning model is built on.
- Early detection of delivery delays
- The same module, reading delivery and customer data instead of production inputs.
- Optimisation of production planning
- A sales forecast model feeding production planning, on the same platform.