Advanced Quality Evaluation – Why Quality Engineers Love Both Cats and Foxes
11 Nov 2020 - Industry 4.0, Artificial Intelligence, Production
A quality management system should be reliable and fast, like the cat in the Aesop fable, where he perfected one trick and climbs a tree in order to escape from the dogs. At the same time the solution has to be intelligent and versatile – just like the fox who analyzes numerous strategies in order to find the best one. Quality experts in steel and aluminum industry often face similar challenges – they have to ensure the quality and react quickly, flexibly and sometimes creatively to unforeseen quality problems. With its concepts of QIs (Quality Indicators) and QPS (Quality Process Snapshots) PSImetals Quality combines the best characteristics of both fox and cat and by this helps the user to ensure an optimal and fast quality evaluation in order to “escape the hounds”!
The quality of a product is always ensured by the knowledge of experts. The exact execution of the production steps is as important as the intelligent assignment of the product to the corresponding order. As the American author Max Lucado said: “Nobody can do everything, but everybody can do something very well”. For example, the quality engineer knows an expert in the production environment who is familiar with every single detail of the lines or someone who has an excellent understanding of the needs of individual customers. Added to this is the deep technological knowledge that helps the quality engineer to understand the actions of the other players.
Establishing the Basis for a Quality Decision
In the age of Industry 4.0 this also applies to automation. In order to have a basis for a sophisticated quality decision, the relevant quality data must be stored and assigned in a suitable form. This is done in the so-called QPS (Quality Process Snapshot) which is a data container where all quality relevant data linked to this special material and process step is stored, including a reference to production time. This data naturally includes measured data on material properties such as measurement curves, data from surface inspection systems or even raw data from the automation layer, but also process data including primary data input and output - which can refer to both time and material coordinates.
The QPS are fully versioned and have their own data structure in the PSImetals Factory Model which makes the data traceable and trustable.
In case of e.g. a customer claim or a product audit, the complete process is transparent and no data modification remains undetected. Hence, QPS can be used in various ways: for decisions on material usage, for data on only one specific material along the production chain or for performing SPC (Statistical Process Control). When the totality of all snapshots is considered, the data is also suitable for advanced quality predictions based on Deep Learning or other Machine Learning techniques.
Bringing the Knowledge of the Quality Engineer Into the System
But how does the knowledge of the quality engineer find its way into the system? This is done by the definition of so-called QIs (Quality Indicators), which are also part of the QPS data set. A QI is a measure that represents a quality value of a certain material. It is based on the dependencies between the data of the production process including its setup, the input data and the properties of the output material. Accordingly, the QIs and test results are used by the system to make quality decisions to determine whether the material meets the actual quality objectives planned for that stage of production.
However, not only data from the associated snapshot can be processed – also data from previous process steps of this material and from previous materials in this process step (or any other data) can be considered.
The Journey of the Material
Once a quality problem has been identified, the journey of the material is of course not over. Depending on the condition of the materials and also depending on the current situation of the environmental conditions, there are various possibilities for automated reaction. You could take a look in your order book and search for any other order where the material you just produced meets the targets. Changing the outstanding production steps or production parameters could also help. The decision depends not only on the technical data, but also on knowledge of the order book and strategy, e.g. cost optimization, fulfillment rate or throughput.
Unfortunately in the fable the fox was caught by the dogs because he had too many tricks and had to think too long about which one was the best for escaping. In our industry there doesn't have to be so much drama. PSImetals Quality combines the best qualities of the cat and the fox and offers an intelligent, flexible and above all fast solution for advanced quality evaluation, so that you can take a deep breath and enjoy the silence of a smooth production!
Do you want to know more? Follow our Enjoy the Sound of Silence campaign!
Enjoy the Sound of Silence Webinar Series:
What is your opinion on this topic?
Dr. Joachim Gnauk
Competence Center Quality PSI Metals GmbH
In more than 20 years of working in R&D and quality assurance in and for the steel and aluminum industry, Joachim has gained a deep understanding of the production and control processes in the industry. Since joining PSI Metals in 2018, Joachim has led the Competence Center Quality to new shores by initiating activities to integrate machine learning, artificial intelligence and other innovative techniques into the PSImetals Quality Standard. At home he spends a lot of time with his family to make music together.
+49 211 60219 -368
jgnauk@psi.de
