Periodic Reporting for period 3 - PREVIEW (PREdictiVe system to recommend Injection mold sEtup in Wireless sensor networks)
Reporting period: 2017-01-01 to 2017-12-31
In the last years, a constant demand of diversity and personalisation in the plastic sector industry has led to a change of the manufacturing paradigm. This has entailed a globalized production model in which a mould is used in different types of injection machines. Moreover, short productions demand a high degree of flexibility, leading towards a high number of mould changes in plastic moulding factories, causing downtimes and scraps. The scraps during the setup phase represent up to 40% of the total scraps. A reduction of the mould setup time for each machine and a reduction of the reject rates are highly demanded within this context.
The use of statistical process control systems based on data acquisition from both sensors located inside the cavity of the mould and process parameters have tried to reach the required productivity improvement. Nevertheless, the reality is that currently less than 3% of new moulds are equipped with sensors and a change of production machine implies starting a new setup process. Sensor manufacturers have pointed out that one of the main causes of the converters’ refusal to use sensors is due to the wiring, proprietary interfaces and external devices, increasing the cost of the solutions and hinders the daily operations. This entails to low acceptance rates.
Overall objectives of the project:
PREVIEW project aims at developing a Cyber Physical System (CPS) for plastic injection manufacturing processes monitoring, control and optimization by incorporating several innovative and cutting edge technological solutions: advanced Artificial Intelligence and Machine Learning techniques, robust Industrial wireless communication, Internet of Things (IoT) and wireless indoor localization. PREVIEW is a middleware solution that facilitates easy, ubiquitous, holistic and fast sharing of product and process information across the entire injection production process. In general, PREVIEW will boost the use of a process control systems by reducing the number of visible components, simplifying its installation and improving and optimizing the production process. PREVIEW overall objectives are:
- To reduce mould setup time by developing a holistic methodology for predicting the optimal machine parameters configuration
- To optimize production control by applying AI and Machine learning techniques to analyse real time production data
- To develop a compact wireless-capable data acquisition system for injection moulds and machines
- To improve operational flexibility by implementing robust industrial wireless communication
- To seamlessly provide process information by using an indoor real time location system able to deliver specific information to a determined profile operator
- Finalization and integration of the hardware module in charge of the adaptation, amplification and digitalization of pressure and temperature signals from the cavity, as well as collection of machine parameters of an injection machine (WP2)
- Development of single-hop and multi-hop communications algorithms for robust data transmission, implemented on real hardware and extensively tested in real conditions (WP3)
- Implementation and test of the location-based content delivery system in real conditions, showing a suitable accuracy for the targeted application.
- Implementation of the Advanced Predictive system including production control and mould setup predictive functionalities
- Integration of the system as a whole
Main exploitable results achieved so far:
- A compact and cost-effective data acquisition module to monitor the mould cavity in real time compatible with most commercial sensors which avoid the use of external signal amplifiers.
- A robust wireless protocol implemented on real hardware for a soft real time transmission of production data able to operate in harsh wireless channel conditions and with high level of packet loss.
- An indoor positioning system based on wireless sensor network which includes a deployment tool and a mobile service running on commercial portable devices.
- An Advance Predictive system for production control and predictive mould setup
- An integrated system including all the above listed components tested in real conditions
PREVIEW project is proposing progress beyond the state of the art according to the following technological challenges:
- Data acquisition systems miniaturization and versatility
- Secure and efficient large-scale wireless monitoring
- Application-tailored data representations to deal with Wireless Monitoring Networks (WMNs
- Mould setup predictive system
- Production control and optimization
- Location-based detection
Expected potential impact:
The plastic and metal transforming manufacturing small and medium enterprises need innovative solutions that can improve their competitiveness against often low-wage competitors in other parts fo the world by offering higher added value products, while being more efficient in terms of energy and raw material consumption. PREVIEW is the next step in the manufacturing monitoring devices sector, incorporating the latest hardware and software solutions and bringing them into the industrial environment. PREVIEW offers an all-in-one standalone solution incorporating the data acquisition and the decision making Artificial Intelligence expert system. In this context, the expected impact of PREVIEW is:
• A reduction of the mould setup time in different machines by 50% through the use of a novel predictive system.
• Reduction of scraps and energy consumption by 20%
• An increase of productivity and flexibility by 30% with fast response against changes or new customer demands.
• Fostering a holistic view of the productive process by implementing a robust industrial wireless communication system capable of operating collaboratively across the manufacturing value chain.
• Reducing complexity of process monitoring equipment by developing a novel data acquisition system to collect the data from the in-mould cavity pressure and temperature sensors as well to collect data from the injection machine. Thanks to this thee mould and machine become IoT-capable, allowing for interoperability at intra-plant level with other devices