RiaStone (RST) is part of “Visabeira Industria” a sub-holding of the “Visabeira Group” conglomerate. RST manufactures the IKEA worldwide supply of “Dinera”, “Fargrik” and “Flitighet” tableware families, being these products fabricated through an innovative Industrial ceramics production process: tableware automated single firing.
Riastone is working towards improving its Overall Production Effectiveness (OPE) KPI; from ~92%, to 99%. This requires new approaches to production, promoting better and innovative defect management and production control methods, which are consistent with the implementation of ZDM processes, namely in-line inspection technologies, and integration of tools for autonomous, automatic, smart system decision making.
In order to achieve the required improvement goals, RST is applying several systems-level strategies consisting in the integration of new inline inspections systems and AQL control loops into the current production line.
In parallel new in-line automation components will be implemented that will automatically remove detected defective parts from the production lines; enabling the introduction of recycling and re-using of raw materials into the production line for a new defect-free production cycle.
Additionally, RiaStone will expand and leverage the outcomes of a H2020 project (BOOST4.0) which will integrate the BOOST4.0 Big Data processing platforms and the newly implemented Qu4lity ZDM Autonomous Quality Control Loops.
In the process of using the Quality AQL real-time inspection of intermediate tableware products RiaStone will:
- Perform Artificial Vision “Submillimeter inspection” of Tableware after the ceramics pressing and fettling stage
- Perform Artificial Vision “Surface inspection” of Tableware after the glazing application stage
- Use Edge computing technologies to process computer vision imaging collected from the inspected tableware
- Through ML computing, compare processed imaging originated in the AQL Loops against an in-memory defects database consisting of 10000+ annotated defects pictures
- Issue a decision of conformity or non-conformity after Artificial Vision analyses computing.
- Instruct control gate systems for each piece of tableware regarding the Go-NoGo decision at tableware piece level.
With these base systems in mind RiaStone implemented the Qu4lity ZDM -AQL loops in the following Business Processes:
Business process RiaStone_1- Isostatic Process AQL:
In Business Process 1 Iso-Static pressing is used to evenly form the main ceramic structure, due to the necessary delicate balance of several physical variables (density/Temperature/Pressure) in this stage, quality deviations can occur that cause geometry and quality defects.
Business process RiaStone 2 - Glazing Process AQL:
In Business Process 2 glazing is applied to the already pressed Tableware, due to the exact Glaze Viscosity needed in this stage, glaze viscosity variations, will originate excessive, or deficient glaze coverings, which will cause quality problems such as the formation of glaze drips and glaze bead formations,
| Country: | PT |
Using the opportunities brought by the Qu4lity project, RiaStone with the collaboration of Synesis and IntraSoft, built a commercial grade ZDM implementation scenario, which brings to the ceramics industry the ability to implement Autonomous Quality Loops, which will add new approaches to production, promoting better and innovative defect management and production control methods, consistent with the integration of Zero defect Manufacturing processes, these being namely: in-line inspection technologies, and integration of ICT tools for autonomous, automatic, smart system decision taking
The RiaStone Qu4lity Pilot enables human involvement at various levels and production stages.
Human operators inspect, monitor, and view all created datasets. During the process of model training, the operator guides the ML algorithm through the image inspection training process.
This is performed after the image acquisition in the production line, through joint algorithm/human expert labeling of the images (new defect classification, conformant/non-conformant product), this allowing for the training of the algorithm in the existing defects dataset, and at evolving the algorithm and pushing the new (improved) model to the production line AQL platform
In the RiaStone Qu4lity Pilot the ZDM-AQL is implemented in a modular architecture, which includes both in-factory data processing, Edge processing Systems, Cloud processing systems, and Machine Learning processing Services.
The RiaStone Qu4lity Pilot goal is to recognize, detect, and reconfigure the production process parameters as soon as a failure is detected in real-time.
This process is based in the collected data, advanced analytics, machine learning image inspection methods
The data acquired through computer vision, is processed through machine learning algorithms and compared to an existing database of ~10000+ images already noted by human operators
Inspection results are fed into the synesis-consortium machine control platform that decides necessary changes to the machine parameters driving production in both Business Processes (1&2)
The ZDM-Autononous Quality Solutions are used as systems that perform tasks in an autonomous/automated way, requiring the intervention of an operator only when an operational tie-breaker is needed. When that is the case, the operator has to analyse the incident and provide for a solution to the AQL System, interacting with it via an HMI interface.