Dieffenbacher’s EVORIS platform is a digital solution that aims to help manufacturers produce smarter and more sustainably.
EVORIS uses artificial intelligence (AI) which optimises board quality while simultaneously lowering production costs, avoiding rejects and reducing the use of wood, glue and other expensive raw materials.
Traditionally, a sample board is cut and sent to a laboratory for analysis once per shift. Measuring quality parameters this way can take from several hours to days, sometimes forcing long delays in the control and production circuit.
With the EVORIS quality prediction app, manufacturers can have a continuous, real-time prediction of board quality parameters during production.
According to Dieffenbacher, manufacturers can view quality parameters and laboratory measurements on the EVORIS dashboard. All relevant quality parameters are displayed live and up to one month retrospectively. Quality deviations prompt warnings, which allow operators and technologists to identify their causes and make corrections faster.
Also, the AI-supported system learns and improves itself independently as more laboratory data flows automatically into the application. The result is increasingly accurate quality predictions.
In the near future, a simulation of changed production parameters will be able to produce quality predictions without the need to change production. Trained models will automatically adapt to changing production conditions and products. Manufacturers will be able to speed up production and work with lower tolerances by moving closer to quality limits.
In combination with the new AI-based anomaly detection, deviations in the production process can be detected faster. This avoids downtime and helps to achieve consistent quality. Meanwhile, the anomaly detection can be used for different processes at the same time, which enables more accurate diagnoses.
Dieffenbacher reported that more EVORIS enhancements are coming. The Quality Prediction app will deploy AI algorithm ensembles to analyze quality parameters in parallel, improving prediction accuracy. Operators will also be able to view the accuracy of each potential production model.
Another app being developed will track particle size, allowing manufacturers to measure and analyze particle size and distribution online. This information will enhance the quality prediction AI algorithm.