Author: Tom Martin – JetSoft CEO
Recently I had the task of sourcing some information for a proposal we are working on.
Due to the varied deployments, we have now delivered, doing this mostly consists of searching the back catalogue of closely related jobs we’ve done and pulling material from them.
During this search I found, at the proverbial back of the drawer, a proposal that detailed a roadmap to full automated inspection. This proposal was for a large international aerospace defence company and designs a four-phase approach to achieving a ‘lights out’ (their term) inspection process. I’d largely forgotten this ‘whitepaper’ but rereading I was reminded how good it was, and felt it was too good to sit in the drawer. So, I thought I’d redact the sensitive information and share it.
The full paper can be download from the link at the bottom of this article.
The Four-Phase Roadmap
The roadmap is divided into four distinct, yet interlinked phases, each designed to bring incremental improvements while driving the overall goal of “lights-out” inspection. Each phase adds value and contributes to the larger vision of automation.
In general, these four phases are designed to:
The whitepaper certainly champions our solutions, but looking past that, I think the roadmap itself presents a realistic and valuable approach to achieving fully automated inspection.
Key Takeaways
Two critical points emphasized in the whitepaper, which are often overlooked in industry discussions, are:
I encourage you to take a closer look at the whitepaper and let me know your thoughts.
Progress on the Roadmap
While the company this proposal was written for didn’t end up taking us up on the proposal, we have forged ahead with other customers on this roadmap (somewhat inadvertently). Including very recently completing the first installation of our virtual machine engine, therefore completing the 3rd phase of this plan.
- Improve “right first time” and reduce waste through understanding the causes of defects and driving process changes
- Aid faster iterative continuous improvement
- Reduce variable manufacturing costs and increase yield by making intelligent decisions on the direction of components following each manufacturing process via virtual assessment
- Expose data in a form that can be used as part of a digital twin
- Introduce automation in a controlled and measured way to the testing department, reducing the human requirement for inspection, and delivering a faster more consistent evaluation
The whitepaper certainly champions our solutions, but looking past that, I think the roadmap itself presents a realistic and valuable approach to achieving fully automated inspection.
Key Takeaways
Two critical points emphasized in the whitepaper, which are often overlooked in industry discussions, are:
- The separation of indication detection and part sentencing. In the paper defining these separate process as Automatic Indication Detection (AID) and Automatic Part Sentencing (APS) respectively. The distinction between, finding something, and evaluating it is a difference that I think is often overlooked when automatic defect detection (and its many names and guises) is discussed.
- The importance of supplementary data when evaluating. For an automatic system to be as good as inspector, it must have all the information that a human uses to make an evaluation. Under the manufacturing framework that the paper is written, this is the manufacturing process data.
I encourage you to take a closer look at the whitepaper and let me know your thoughts.
Progress on the Roadmap
While the company this proposal was written for didn’t end up taking us up on the proposal, we have forged ahead with other customers on this roadmap (somewhat inadvertently). Including very recently completing the first installation of our virtual machine engine, therefore completing the 3rd phase of this plan.
| whitepaper_-_roadmap_to_full_automated_inspection.pdf |
