Final Report Summary - SELSUS (Health Monitoring and Life-Long Capability Management for SELf-SUStaining Manufacturing Systems)
Executive Summary:
Today’s production equipment must enable mass customization, has to enable to adapt various environmental impacts, and fluctuating production volume. Additionally, the cost pressure increased significantly, due to increasing suppliers. Along with this, to enable availability of production equipment and by that establish a high level of ability to deliver within supply-chain considering low level of inventory and buffer stock, just ensuring relative high level of reliability and availability at production level is not sufficient.
The European Union significantly supports industry, suppliers, research institutes and universities in facing this challenge to manage the switch to digitalization of the production, comprising component, equipment, and entire line and considering the entire life-cycle starting from planning phase, down to ramp-up, production, and re-use.
With a new and advanced machines, fixtures and tools, comprising extended sensor capabilities, smart materials and ICT self-diagnosis and self-awareness will be enabled within SelSus project and will also prove self-healing capabilities
Additionally, distributed diagnostic and predictive and renovation models will be an integrated part of smart devices, enabling prognosis failure modes, component degradations, and by that predict future downtimes of devices.
These goals are reflected in the work packages and milestones, focusing on conversing vision into real working machines and production lines. As one of the key objectives within SelSus, data will be gathered, transformed into information and knowledge, and by that may offer an added value to supplier and end-to-end-user using self-diagnosing and self-healing devices.
Sensor networks, built up of a variety of sensor nodes, will add the capability for distributed analysis, interoperable and delay-tolerant communication.
Integrated models and methodologies for predictive maintenance will enable constant life-long assessment, by forecasting degradation and deterioration trends of single components and equipment. Additionally, machine learning and data mining techniques, combined with discrete material flow simulation models will enable decision models, on how to anticipate unforeseen future malfunctions and downtimes.
These methodologies will be proofed and demonstrated on three different levels, which are device level, equipment level, and factory, respective line level.
As scheduled, one integrated task will be, to disseminate publishable results. Additionally, to website, Linked-In account, flyers, etc., the consortium organized a Smart Factory Workshop to directly communicate goals and results to industry and universities.
Project Context and Objectives:
The main goals for the phase reported, being achieved, were to get the project started; aligning all partner and create a commitment to the overall project aims. Therefore, and as a first step, the elicitation of industrial requirements, use cases and demonstrators through bilateral meetings with the different project partners were of upmost importance. Based on the gathered industrial requirements and use-cases the overall concepts and architecture for the predictive maintenance environment were developed and described in the SelSus Architecture Concept Document. The relevant data for condition monitoring of smart devices and machines as well as the specification and the development of a conceptual framework for such devices were defined as the basis for further work in different WPs.
To achieve this, the analysis model structure and SelComp architecture has been evaluated and updated based on implementation decisions and feedback from demonstrators. To ensure developments address real business motivations, guidelines for new SelSus business models, considering major IPR aspects, has been defined.
As base for further developments, the common data model has been implemented into welding controls, linear axis and jigs and fixtures, enhanced with interface with system level. These have been equipped with WSNs (Wireless Sensor Networks) accessing Sensor SelComp template and a, so called, Dynamic Modular Software Reconfiguration, to increase flexibility to operate the sensing capabilities at shop-floor level. Additionally, a HMI using mobile application to increase the flexibility in data presentation, able to show sensor data from Sensor SelComp and recognize via WSN using QR Code at shop-floor level.
Furthermore, a structure of Sensor SelComp was created and reported based on several brainstorming sessions between all the partners. This structure was used to further develop a prototype to validate all the specified requirements, where sensors were successfully integrated, being able to monitor, access and manipulate the information, not only at the Component level (Sensor SelComp), but also at the System Level (Cloud). Further related investigations concerning the internal structure of SelComps, the connection to clouds and the interaction with human were performed. Regarding the topic of degradation prediction of wear-out components Bayesian network models were developed already. This includes the use of condition-based monitoring methods, identifying relationships between components failure and wear-out, revising and adapting methods for condition monitoring as well as feeding the models into the project’s diagnostic models. Based on information about the healthiness of linear axis a first model for diagnosis was developed. To demonstrate the interaction between sensors and Bayesian networks, an application that continuously measures spectra of permanently changing colors and queries a Bayesian network for the probabilities of 16 color states and visualizes the results in on-line mode was developed.
Based on use cases and scenarios for the decision support for renovation and repair strategy enables a better understanding what the decision support business process is and what the impact and solution strategies in case of real or expected downs would be.
Using the common data model for life-cycle knowledge is suitable to represent expert knowledge from the design phase of the machine tool and from the design phase of production systems as well as the lifelong gathered feedback in the usage phase. In addition to this, an enhancement of the FMEA modelling procedure including the way to model it in a common COTS-Tool has been derived.
Project Results:
After the identification of the foreground IP and the ownership, the analysis of the results of the workshop took place. HWH as the responsible task leader collected all information and compiled a first draft of an output matrix. This matrix has the purpose to list the exploitable results together with the following information for each result:
Result ID - Result owner - Result description - Link to WP - TRL M24 - Target TRL M48
1 - ADP - Univariate fault detection module for sensor signal Processing - WP3 - TRL 4 /TRL 6
2 - ADP - Multivariate time trace shape analysis for fault detection - WP 2 - TRL 4 - TRL 7
3 - ELX - Innovative strategies and tools to monitor aged manufacturing systems (SelSus approach) - all WP
4 - ELX - Reliable decision support system in maintenance operations to fast react to production systems` criticalities – all WP
5 - FRD - Sensor derivation from machine FMEA and CDA FMEA - WP 3
6 - FRD - Sealant dispensing cell knowledge
7 - FHG - FMEA for machinery with combination of components-FMEAs (Methodology) - WP6.2 - TRL3 - TRL6
8 - FHG - Enhanced FMEA (Influences, Sensor, Methodology) - WP6.2 - TRL 3- TRL 5
9 - FHG - App: Applications for mobile devices to interact with SelSus backbone - WP3/5 - TRL3 - TRL6
10 - FHG - App: Mobile diagnostic combining AR and BN - WP3/4/5 - TRL2 - TRL6
11 - FHG - App: Context-specific data acquisition using mobile sensors and AR for interactive data provision - WP3/5 - TRL4 – TRL6
12 - FHG - Degradation guarantees realistic lifetime predictions - WP4 - TRL3 - TRL5
13 - FHG - Context-sensitive maintenance planning - WP5 - TRL 2 - TRL 4
14 - FHG - Simulation/ Software/ Methodology: Application for combined use of BN and DES for enhanced prediction
especially using degradation - WP4/ 5 - TRL 2 - TRL 6
15 - GMX - SW library (XML parser, DLLs, xphs, protocols) - WP5 - TRL 4 - TRL 5
16 - HIS - Dynamic resource monitor (power signatures)
17 - HIS - Generating machine FMEA from Component FMEAs
18 - HUG - Diagnostics with BN - WP4 - TRL 4 - TRL 7
19 - HUG - Diagnostics models - WP4 - TRL 4 - TRL 7
20 - HUG - Web interface for device monitoring - WP4 - TRL 4 - TRL 4
21 - HUG - Integration between BN and DES for enhance health / performance / prediction - WP4 - TRL 2 - TRL 4
22 - HUG - Degradation with BN - WP4
23 - HUG - Enhanced HUGIN software (based on "background" knowledge / enable unoffrera/ fct. GDL - WP4 - TRL 3 - TRL 8
24 - HUG - HUGIN API on android - WP4 - TRL 8 - TRL 8
25 - HWH - Maintenance support system (software for defect detection) - WP 2
26 - HWH - Software library for the detection of faults in welding process - WP 2
27 - HWH - App for "limping-home" mode for welding controls - WP 2
28 - HWH - SelComps architecture - WP 2
29 - IEF - Self-sustainable linear axis - WP 2
30 - IEF - Lifetime and degradation calculation (linear axis) - WP 2
31 - IEF - Proactive maintenance scheduling (linear axis) - WP 2
32 - INO - Networked RF sensors utilising multiple transducers and GPRS modem transmission
33 - ISR - Dynamic reconfiguration of sensor SelComp - WP3 - TRL4 - TRL7
34 - ISR - Conceptual framework (Sensor SelComp) for sensor integration/processing (sensor fault detection) - WP3 - TRL5 - TRL8
35 - ISR - VFK enabler as a Sensor Cloud solution - WP3 - TRL1 - TRL6
36 - LBU - Knowledge engineering method for OOBN in manufacturing system diagnostics
37 - LBU - SelComp interfaces and architecture
38 - LBU - Reference architecture for predictive maintenance of modular production systems
39 - LBU - FMEA and engineering knowledge-based Bayesian diagnostic model generation (methodology and tools)
40 - LBU - Autonomous modelling methodology
41 - LBU - FMEA-based methodology to build BN models
42 - MTC - DSS architecture for predictive maintenance
43 - MTC - SelComp design methodology
44 - MTC - Methodology to design maintenance and repair DSS
45 - MTC - Use of smart materials for self-healing
46 - MTC - Two way communication with industrial robots
47 - MTC - Connectivity for remote sensor data storage
48 - MTC - Framework / methodology to use ID for predictive maintenance decision-support
49 - UNT - Fault aware and disconnection aware communication framework (FDASS) provides critical communication
mechanism for future smart manufacturing environments
50 - UNT - Realtime and historical (integrated) monitoring and notification dashboard for distributed sensor readings
and physical topology (People | devices) including cloud interface)
51 - UNT - FDASS infrastructure (sensor opportunistic network clouds) can be used for wellbeing monitoring of staff
on the manufacturing floor and improvement of working condition
52 - XET - SelComp cloud service
53 - XET - SelComp cloud gateway
54 - XET - Integration to MES on cloud platforms
55 - XET - Measurement and alarm data extension in XETICS lean MES
56 - XET - Measurement and alarm data extension for android app in XETICS lean MES
57 - XET - (AML) common data model for diagnostics and maintenance
58 - UNOT - Network for distributed systems
Potential Impact:
SelSus show an impact on different levels and industries.
It has been implemented into the industry, e.g. at Electrolux or Ford. Also it upgrades pruduct in the machine industry, e.g. Harms&Wende or IEF Werner.
SelSus made it possible for ICT provider to offer new products, e.g. Hugin, Xetics, or Gamax.
Additionally, for research institutes and universities SelSus has built a bridge from theoretical frameworks to real applications and also enabled new and adavanced research activities.
These has been pubished via
- Two flyers
- Homepage
- Social media
- youtube
- Scientific publications
- Magazin publications
- Trade fairs
- Smart Factory Workshops
- University lectures
- etc.
List of Websites:
www.selsus.eu
Fraunhofer IPA
Martin Kasperczyk
martin.kasperczyk@ipa.fraunhofer.de
Today’s production equipment must enable mass customization, has to enable to adapt various environmental impacts, and fluctuating production volume. Additionally, the cost pressure increased significantly, due to increasing suppliers. Along with this, to enable availability of production equipment and by that establish a high level of ability to deliver within supply-chain considering low level of inventory and buffer stock, just ensuring relative high level of reliability and availability at production level is not sufficient.
The European Union significantly supports industry, suppliers, research institutes and universities in facing this challenge to manage the switch to digitalization of the production, comprising component, equipment, and entire line and considering the entire life-cycle starting from planning phase, down to ramp-up, production, and re-use.
With a new and advanced machines, fixtures and tools, comprising extended sensor capabilities, smart materials and ICT self-diagnosis and self-awareness will be enabled within SelSus project and will also prove self-healing capabilities
Additionally, distributed diagnostic and predictive and renovation models will be an integrated part of smart devices, enabling prognosis failure modes, component degradations, and by that predict future downtimes of devices.
These goals are reflected in the work packages and milestones, focusing on conversing vision into real working machines and production lines. As one of the key objectives within SelSus, data will be gathered, transformed into information and knowledge, and by that may offer an added value to supplier and end-to-end-user using self-diagnosing and self-healing devices.
Sensor networks, built up of a variety of sensor nodes, will add the capability for distributed analysis, interoperable and delay-tolerant communication.
Integrated models and methodologies for predictive maintenance will enable constant life-long assessment, by forecasting degradation and deterioration trends of single components and equipment. Additionally, machine learning and data mining techniques, combined with discrete material flow simulation models will enable decision models, on how to anticipate unforeseen future malfunctions and downtimes.
These methodologies will be proofed and demonstrated on three different levels, which are device level, equipment level, and factory, respective line level.
As scheduled, one integrated task will be, to disseminate publishable results. Additionally, to website, Linked-In account, flyers, etc., the consortium organized a Smart Factory Workshop to directly communicate goals and results to industry and universities.
Project Context and Objectives:
The main goals for the phase reported, being achieved, were to get the project started; aligning all partner and create a commitment to the overall project aims. Therefore, and as a first step, the elicitation of industrial requirements, use cases and demonstrators through bilateral meetings with the different project partners were of upmost importance. Based on the gathered industrial requirements and use-cases the overall concepts and architecture for the predictive maintenance environment were developed and described in the SelSus Architecture Concept Document. The relevant data for condition monitoring of smart devices and machines as well as the specification and the development of a conceptual framework for such devices were defined as the basis for further work in different WPs.
To achieve this, the analysis model structure and SelComp architecture has been evaluated and updated based on implementation decisions and feedback from demonstrators. To ensure developments address real business motivations, guidelines for new SelSus business models, considering major IPR aspects, has been defined.
As base for further developments, the common data model has been implemented into welding controls, linear axis and jigs and fixtures, enhanced with interface with system level. These have been equipped with WSNs (Wireless Sensor Networks) accessing Sensor SelComp template and a, so called, Dynamic Modular Software Reconfiguration, to increase flexibility to operate the sensing capabilities at shop-floor level. Additionally, a HMI using mobile application to increase the flexibility in data presentation, able to show sensor data from Sensor SelComp and recognize via WSN using QR Code at shop-floor level.
Furthermore, a structure of Sensor SelComp was created and reported based on several brainstorming sessions between all the partners. This structure was used to further develop a prototype to validate all the specified requirements, where sensors were successfully integrated, being able to monitor, access and manipulate the information, not only at the Component level (Sensor SelComp), but also at the System Level (Cloud). Further related investigations concerning the internal structure of SelComps, the connection to clouds and the interaction with human were performed. Regarding the topic of degradation prediction of wear-out components Bayesian network models were developed already. This includes the use of condition-based monitoring methods, identifying relationships between components failure and wear-out, revising and adapting methods for condition monitoring as well as feeding the models into the project’s diagnostic models. Based on information about the healthiness of linear axis a first model for diagnosis was developed. To demonstrate the interaction between sensors and Bayesian networks, an application that continuously measures spectra of permanently changing colors and queries a Bayesian network for the probabilities of 16 color states and visualizes the results in on-line mode was developed.
Based on use cases and scenarios for the decision support for renovation and repair strategy enables a better understanding what the decision support business process is and what the impact and solution strategies in case of real or expected downs would be.
Using the common data model for life-cycle knowledge is suitable to represent expert knowledge from the design phase of the machine tool and from the design phase of production systems as well as the lifelong gathered feedback in the usage phase. In addition to this, an enhancement of the FMEA modelling procedure including the way to model it in a common COTS-Tool has been derived.
Project Results:
After the identification of the foreground IP and the ownership, the analysis of the results of the workshop took place. HWH as the responsible task leader collected all information and compiled a first draft of an output matrix. This matrix has the purpose to list the exploitable results together with the following information for each result:
Result ID - Result owner - Result description - Link to WP - TRL M24 - Target TRL M48
1 - ADP - Univariate fault detection module for sensor signal Processing - WP3 - TRL 4 /TRL 6
2 - ADP - Multivariate time trace shape analysis for fault detection - WP 2 - TRL 4 - TRL 7
3 - ELX - Innovative strategies and tools to monitor aged manufacturing systems (SelSus approach) - all WP
4 - ELX - Reliable decision support system in maintenance operations to fast react to production systems` criticalities – all WP
5 - FRD - Sensor derivation from machine FMEA and CDA FMEA - WP 3
6 - FRD - Sealant dispensing cell knowledge
7 - FHG - FMEA for machinery with combination of components-FMEAs (Methodology) - WP6.2 - TRL3 - TRL6
8 - FHG - Enhanced FMEA (Influences, Sensor, Methodology) - WP6.2 - TRL 3- TRL 5
9 - FHG - App: Applications for mobile devices to interact with SelSus backbone - WP3/5 - TRL3 - TRL6
10 - FHG - App: Mobile diagnostic combining AR and BN - WP3/4/5 - TRL2 - TRL6
11 - FHG - App: Context-specific data acquisition using mobile sensors and AR for interactive data provision - WP3/5 - TRL4 – TRL6
12 - FHG - Degradation guarantees realistic lifetime predictions - WP4 - TRL3 - TRL5
13 - FHG - Context-sensitive maintenance planning - WP5 - TRL 2 - TRL 4
14 - FHG - Simulation/ Software/ Methodology: Application for combined use of BN and DES for enhanced prediction
especially using degradation - WP4/ 5 - TRL 2 - TRL 6
15 - GMX - SW library (XML parser, DLLs, xphs, protocols) - WP5 - TRL 4 - TRL 5
16 - HIS - Dynamic resource monitor (power signatures)
17 - HIS - Generating machine FMEA from Component FMEAs
18 - HUG - Diagnostics with BN - WP4 - TRL 4 - TRL 7
19 - HUG - Diagnostics models - WP4 - TRL 4 - TRL 7
20 - HUG - Web interface for device monitoring - WP4 - TRL 4 - TRL 4
21 - HUG - Integration between BN and DES for enhance health / performance / prediction - WP4 - TRL 2 - TRL 4
22 - HUG - Degradation with BN - WP4
23 - HUG - Enhanced HUGIN software (based on "background" knowledge / enable unoffrera/ fct. GDL - WP4 - TRL 3 - TRL 8
24 - HUG - HUGIN API on android - WP4 - TRL 8 - TRL 8
25 - HWH - Maintenance support system (software for defect detection) - WP 2
26 - HWH - Software library for the detection of faults in welding process - WP 2
27 - HWH - App for "limping-home" mode for welding controls - WP 2
28 - HWH - SelComps architecture - WP 2
29 - IEF - Self-sustainable linear axis - WP 2
30 - IEF - Lifetime and degradation calculation (linear axis) - WP 2
31 - IEF - Proactive maintenance scheduling (linear axis) - WP 2
32 - INO - Networked RF sensors utilising multiple transducers and GPRS modem transmission
33 - ISR - Dynamic reconfiguration of sensor SelComp - WP3 - TRL4 - TRL7
34 - ISR - Conceptual framework (Sensor SelComp) for sensor integration/processing (sensor fault detection) - WP3 - TRL5 - TRL8
35 - ISR - VFK enabler as a Sensor Cloud solution - WP3 - TRL1 - TRL6
36 - LBU - Knowledge engineering method for OOBN in manufacturing system diagnostics
37 - LBU - SelComp interfaces and architecture
38 - LBU - Reference architecture for predictive maintenance of modular production systems
39 - LBU - FMEA and engineering knowledge-based Bayesian diagnostic model generation (methodology and tools)
40 - LBU - Autonomous modelling methodology
41 - LBU - FMEA-based methodology to build BN models
42 - MTC - DSS architecture for predictive maintenance
43 - MTC - SelComp design methodology
44 - MTC - Methodology to design maintenance and repair DSS
45 - MTC - Use of smart materials for self-healing
46 - MTC - Two way communication with industrial robots
47 - MTC - Connectivity for remote sensor data storage
48 - MTC - Framework / methodology to use ID for predictive maintenance decision-support
49 - UNT - Fault aware and disconnection aware communication framework (FDASS) provides critical communication
mechanism for future smart manufacturing environments
50 - UNT - Realtime and historical (integrated) monitoring and notification dashboard for distributed sensor readings
and physical topology (People | devices) including cloud interface)
51 - UNT - FDASS infrastructure (sensor opportunistic network clouds) can be used for wellbeing monitoring of staff
on the manufacturing floor and improvement of working condition
52 - XET - SelComp cloud service
53 - XET - SelComp cloud gateway
54 - XET - Integration to MES on cloud platforms
55 - XET - Measurement and alarm data extension in XETICS lean MES
56 - XET - Measurement and alarm data extension for android app in XETICS lean MES
57 - XET - (AML) common data model for diagnostics and maintenance
58 - UNOT - Network for distributed systems
Potential Impact:
SelSus show an impact on different levels and industries.
It has been implemented into the industry, e.g. at Electrolux or Ford. Also it upgrades pruduct in the machine industry, e.g. Harms&Wende or IEF Werner.
SelSus made it possible for ICT provider to offer new products, e.g. Hugin, Xetics, or Gamax.
Additionally, for research institutes and universities SelSus has built a bridge from theoretical frameworks to real applications and also enabled new and adavanced research activities.
These has been pubished via
- Two flyers
- Homepage
- Social media
- youtube
- Scientific publications
- Magazin publications
- Trade fairs
- Smart Factory Workshops
- University lectures
- etc.
List of Websites:
www.selsus.eu
Fraunhofer IPA
Martin Kasperczyk
martin.kasperczyk@ipa.fraunhofer.de
