Skip to main content

ON LINE QUALITY OPTIMISATION OF MOULDED PRODUCTS THROUGH ADAPTIVE FEED FORWARD CONTROL OF EXTRUSION PROCESSES WITH INTELLIGENT MONITORING SYSTEMS

Objective


The project result is a software system able to tune process parameters on-line, based on the feedback from process monitoring procedures and on an extensive analysis of pre-collected data from past production batches. The system is able to smoothly self-adapt to low-rate changes of operating conditions by periodically executing an incremental learning procedure.

The anticipated benefits are the reduction in waste of raw materials, the reduction in waste and reworking of products, increase in productivity, render product quality more stable and more uniform. In addition, the system is aimed at improving reactions to deviations and changing situations as well as identifying best process settings thereby reducing time to production.
This project aims at developing, on the top of existing process control systems, an intelligent monitoring and on-line optimisation tool for extrusion processes, allowing to:

- Reduce waste of raw materials and products and to reduce reworking.
- Increase product quality and productivity
- React more quickly to deviations from optimal process conditions.
- Reduce time-to-production of new products, supporting process settings.

Specifically, the project will develop an hybrid control system, based on a mix of Knowledge Bases Systems, Neural Networks and Databases technology, able to:

- Intelligently monitor the process, providing prognostic information concerning in process product quality and predicting anomalous process operating conditions.
- On-line optimise the process, diagnosing anomalous process operating conditions and performing calculations for on-line process parameter tuning.
- Simulate process behaviour for give process parameters and check new process parameter settings for consistency.
- Continuously display on-line product/process information.
- Help in off-line evaluating the result of previous product-batches.
- Smoothly self-adapt to low-rate changes of the operating conditions by periodically executing an incremental learning procedure.

The project will ensure transferability of the results to other, continuous and semi-continuous processes from many diverse Industrial Sectors, such as chemical, pulp and paper, food, metal working, water treatment, etc.

Funding Scheme

CSC - Cost-sharing contracts

Coordinator

Pirelli SpA
Address
Viale Sarca 222
20126 Milano
Italy

Participants (5)

ALCATEL
Austria
Address
Ruhnergasse 1-7
A1210 Wien
BICC plc
United Kingdom
Address
Quantum House Maylands Avenue
HP2 4SJ Hemel Hempstead
BRAINWARE GmbH
Germany
Address
Gustav-meyer-allee 25
1000 Berlin
PGCC Technologie
France
Address
31 Avenue Du Général Leclerc
92340 Bourg-la-reine
Pomini SpA
Italy
Address
Via L. Da Vinci 20
21053 Castellanza Varese