Paper Technology International 2020 - Journal - Page 66

PAPERTECHNOLOGYINTERNATIONAL
eliminated. It used to be common practice for the administrator or a
This analysis is further simplified and offers an even greater
trained operator to look at a very large number of defect images and
cost savings potential when the WIS is combined with the WBM.
manually assign them to different classifier classes. In the future,
Then, a defect detection of the Inspection system is triggering the
this lengthy process of manual classifier training will be carried out
web break monitoring cameras alongside the machine back to the
headbox. Following the defect through the entire production process by AI algorithms - better and faster than humans can do.
reveals the position where it occurred and the defect origin.
System operators can now concentrate on their core
tasks: The Production of Paper. The amount of data in the images
Tissue: a different substrate, but (almost) the same process
To reduce web breaks and - once again - costs in the further evaluated by the algorithm is also significantly higher than with
manual processing, increasing the classification accuracy.
processing of tissue products, the Converting Control System
It is important that hundreds of features are calculated for
(CCS) detects surface defects logged by the inspection system
each defect recognized by the inspection system.
on the Tissue machine. With double-width tissue machines, an
additional middle-cut detection, inspecting the
quality of the edges along the cut is done, and the
production data is stored for later processing and
retrieval.
During converting, several reels are
combined and processed to form a multi-layer
product. The converted jumbo-reels are not always
complete reels. Several reels of different lengths
per layer can be processed. Here, the CCS enables
synchronization of the defect positions on the
different layers; a layer merger has to combine
defects from multiple reels into a multi-layer defect
database first. The layer-independent defect
position to control the speed of the converting line
is then transmitted by the CCS. Thus, web breaks
and equipment damage will be drastically reduced
in tissue converting lines.
The speed in converting is automatically
reduced whenever critical defects occur. This
solution leads to significant cost savings due to
avoidable reprocessing of the converting line after
a web break, allowing for processing significantly
more tissue and further increasing your yield.
Paper Master: Software makes the difference
and connects the dots
A WBM or WIS cannot achieve the level of
optimization without the software that connects all
Figure 3: The Converting Control System (CCS) helps to reduce
the bits and pieces of information gathered from
web breaks.
sensors and cameras. PAPER MASTER 4.0 is the
first browser-based platform for surface analysis
to access the data of the ISRA system from anywhere, whether it
Thanks to these complex features, each defect can be
be directly at the production line, in a meeting room, or on the road,
described individually. This new, AI-based classifier is a step towards
by simply using mobile devices such as a tablet or a smartphone.
fully automated surface control. All this complex technology is
PAPER MASTER 4.0 combines all data from the ISRA inspection
hidden to the operator, running in parallel with the goal to make life
and web break monitoring systems, the slitter and winder control,
easier for the user.
with all the software modules, in our paper quality management
Make invisible data visible for predictive maintenance
system on one browser-based platform. A new, modern, and intuitive
Another powerful software tool of Paper Master 4.0,
user interface makes data access even easier.
E
PROMI Live, enables you to easily and clearly monitor quality data
PAPER MASTER 4.0 introduces a new dimension in the
of your surface and the manufacturing process. In addition, it offers
automated classification of defects and their assignment to defect
information on the status of the systems in operation, the most
classes and relieves the operator in the daily work while at the same
important quality data, and the production performance.
time classification is secure.
This solution is able to highlight possible causes for
The use of modern methods for artificial intelligence and
emerging problems and prevent faults through targeted intervention
deep learning reduces the optimization times of defect classification.
– enabling predictive maintenance across the whole production
The artificial intelligence methods, automatically assign the detected
process. In addition, EPROMI Live also supports dashboards that
defects to individual classes. The usual, time-consuming, and
display and compare quality data from the ISRA inspection system
error-prone training of the classifier by the operator is completely
with other data from the manufacturing process (such as speed,
moisture, grammage, etc.).
64

Paperturn flipbook