Periodic Reporting for period 1 - GO0D MAN (aGent Oriented Zero Defect Multi-stage mANufacturing)

Summary
At present, worldwide manufacturing industry is very varied. 100% quality controls are performed on the final product at the end of the production line and lately more and more distributed along multi-stage production lines for inspecting components and sub-systems...\n\nAt present, worldwide manufacturing industry is very varied. 100% quality controls are performed on the final product at the end of the production line and lately more and more distributed along multi-stage production lines for inspecting components and sub-systems. Furthermore, reliability tests are performed on statistical basis on production samples. At the same time all automated machines include sensors to monitor performance and run feedback control loops. In the meanwhile, simpler production environments continue to exist, where most operations are human-based including quality controls; a very limited numbers of sensors are used, hence data are lacking. Indeed, understanding and controlling the quality in complex multi-stage manufacturing systems goes beyond the capacity of the model driven approaches and requires the right quantity and quality of meaningful data, in order to extract relevant information and build-up useful knowledge. Therefore, the simplest production environments should evolve to acquire the minimum required amount of data while the advanced ones should optimize the use of the already available data for an effective ZDM. Moreover, today, data coming from quality controls are not used in real-time for production process control and vice versa: they are usually processed together as vertical silos in a multistage process, mainly off-line at the level of SPC.
The main idea of GO0D MAN project is to integrate and combine process and quality control for a multi ? stage manufacturing production into a distributed system architecture built on agent-based Cyber-Physical Systems (CPS) and smart inspection tools. This framework, based on agents associated to each process and to each product, supports the real-time data collection and defect diagnosis at single process level, as well as the inter-stage sharing and processing of information at global level using data mining techniques. The Zero Defect Manufacturing (ZDM) strategy proposed by the GO0D MAN project runs on this framework, at local and at global system level, and will be demonstrated in three relevant industrial cases with multi-stage production lines with different levels of automation and production rate.
Therefore, the objectives of GO0D MAN project are set to allow realization of a fully functional, replicable and therefore widely exploitable solution, employing multi-agent systems, smart on-line inspection tools and data analytics, for implementing a ZDM strategy.
Full success will consist in the demonstration in three relevant industrial cases, which represent more than 80% of the manufacturing sector.\n\nThis Innovation Actions (IA) primarily consists of activities directly aimed at producing and demonstrating improved manufacturing processes. For this purpose it includes prototyping, testing and demonstration in an operational environment.
The demonstration is foreseen in three different industrial sites, each separated in an autonomous testbed to enable a pragmatic and realistic deployment of the results. Each of the testbed is understood as an own eco-system, hence following the typical Plan-Do-Check-Act (PDCA) cycle in form of tasks within the work-package.
At the end of the first 18 months of the GO0DMAN project most of the Work Packages are in progress and some have been already finalized.
WP1 is concluded and successfully defined the Zero Defect Methodology that will be followed during the execution of the project. A further result of WP1 is a tool for the implementation of the developed management methodology.
WP2 is also concluded with the general definition of the MAS CPS architecture. The different types of agents and their interactions have been identified. The Multi agent system infrastructure has been then validated with a simplified demo of a electrical motor multi stage production line.
Different on line smart quality control systems have been developed in WP3, starting from the needs of the end users, such as: Tool for measuring geometrical features of bores, Burrs detection tool (patent application filed), Tool wear monitoring system, Gap&FFlush measurement tool (patent application filed), Hot air/steam leak detection tool, Fan-motor assembly vibration inspection tool
In WP4, the the knowledge management environment and the data analytics environment from a user interaction, structural perspective and deployment viewpoint have been defined. Then a prototype of the building blocks used to model the knowledge management system has been implemented as well as for the knowledge-based quality rules for adaptive decision-making support.
For the data analytics environment, the main technological innovation is related to a novel method for data-driven detection of anomalies (variations) as outlier clusters in a multidimensional data space. The data analytics framework has been extended in order to accommodate the requirements from use cases regarding the data analytics and to implement new interfaces with the other components of the GO0D MAN architecture, i.e. the MAS and the KM environment.
In the exploitation workpackage (WP9), a set of exploitable results has been already specified and initial Canvas Business model have been defined. An Innovation Shop has been set up.\n\nProgrammes and approaches to quality improvement have been adopted since the 1920s, when some of the first seeds of quality management were planted as the principles of scientific management swept through the industry sector. The last three decades have seen a dramatic progress towards the smart factory, which has ended in a real industrial revolution, today known as Industrie 4.0, defined as ?a collective term for technologies and concepts of value chain organization' which draws together Cyber-Physical Systems, the Internet of Things and the Internet of Services?. In this scenario, Zero Defect Manufacturing is highly complex, being at the frontier of new technologies, products and ways of production. Therefore, it is needed to develop advanced measurement systems, able to precisely track high product variation and to be fully aware of every instance of the production processes. The key challenge targets the processes of manufacturing 2.0 enterprises covering the management of high-level supply-chain management down to assets and inventories together with assembly lines and machinery that are dynamically monitored, configured and maintained. This is a pre-requisite for advanced production process management with zero defect strategy referencing on visibility, real-time tracking and self-healing product-process systems. The ambition of this project is to address all this complexity through the deployment of a zero defects manufacturing framework based on both intelligent agents and an advanced ICT infrastructure for data collection, transmission and mining. The latter also provides communication capabilities to the multi-agent system, which is organised in an architecture that combines distributed/local autonomous operations and centralised/global semi-autonomous/assisted operations, as described in the following paragraph.
The expected impacts of the GO0DMAN project are the following:
- Achievement of zero defects in a multi-stage production line:
- Reduction of production costs by 15%:
- Increased production flexibility. Higher production rates by 15%:
- Reduction of waste and scrap by 10%
- Wide adoption of the new strategies in the existing manufacturing systems
More information
aGent Oriented Zero Defect Multi-stage mANufacturing
Contacts
Not specified (see website if available)
Organisations
  • Not specified