Cookie Consent by PrivacyPolicies.com
Blog >
Understanding Different Levels of Quality Assurance (QA) Automation

Understanding Different Levels of Quality Assurance (QA) Automation

Please fill the following form to read the post.

Blog >
Understanding Different Levels of Quality Assurance (QA) Automation

February 3, 2022

Understanding Different Levels of Quality Assurance (QA) Automation

'This market study was conducted together with ABI Research''

Today, QA automation is a major enabler for improved profit margins and competitive positioning. Together with ABI Research, Relimetrics developed a maturity model to help manufacturers define the different capabilities that exist in this space. Using the Relimetrics' Quality Automation Maturity Model, manufacturers can assess where they are in their digital maturity and easily identify ways to add scale and accuracy to their processes.

In Relimetris' QA Automation Maturity Model (shown in Figure 1), there are five different levels concerning the automation of the quality inspection process, ranging from humans performing all tasks related to QA to fully-automatic QA inspection. Automation allied with AI means that each item is inspected to a greater degree of accuracy than humans performing often tedious work. Furthermore, the incorporation of AI and machine learning means that the inspection process can consistently adapt to new circumstances and tolerance levels. One item of note is understanding how manufacturers are distributed today along the five stages. Most (~20%) are in the very beginning of their journey today, and very few manufacturers (~7%) are operating at level 5.

Currently, the industry average Maturity Level is 2.8 out of 5.

Relimetrics guarantees to raise this average to 4.5 for any production facility within 3 months.

Do you want to be one of them? Speak to Our Automation Experts.

In Relimetrics' QA Automation Maturity Model, the two key considerations are how inspection information is collected and the extent that technological solutions can adapt to new circumstances (e.g., production variability) independently with no downtime. Different levels of QA automation is defined below:

Level 1: Inspection data is collected by staff members on the factory floor and stored in manual records. In some cases, the staff collects the data with a mobile device (laptop, tablet, smartphone) and creates/edits digital records.

Level 2: Cameras capture images of parts and products being produced, which are fed to software applications for traceability. At this level, staff members, who used to collect inspection data, now conduct inspections on the acquired images.

Level 3: Cameras are now integrated with machine vision software; the integrated offering is not only capturing images, but also is performing inspections for highly repetitive parts and products being produced. The images are stored in third-party software applications. An element of human inspection is still required to weed out false detections and to decide on which parts and products are to be scraped or reworked.

Level 4: Cameras are now integrated with AI-based machine vision software, with computing either at the camera level or at the edge. The software component becomes more sophisticated and can adapt to changes in the production environment or products. An element of human inspection is still required to weed out false detections that are significantly less with respect to Level 3 and to decide on which parts and products are to be scraped or reworked.

Level 5: QA automation is fully digitized and automated, and there is continuous learning in-line. There is no human involvement. The camera is integrated with AI-based machine vision software, which identifies problem items and instructs the Programmable Logic Controller (PLC) to scrap the item or send it to a rework station. As the sophistication of the technology increases, staff members are only involved in assessing marginal cases and can focus their time on root cause analysis and making improvements.

Oops! Something went wrong while submitting the form.