Considering the increasing manufacturing plant size and complexity, unassisted maintenance management is becoming a losing strategy. Intense and global competition forces industries toward the maximal exploitation of their available resources, including equipment. However, the almost uninterrupted use of machine and robots degrades their performances and leads to an increased insurgence of breakages and failures. In turn this leads to higher costs due both to machine unavailability and expensive repair interventions.
The best solution to this problem would be to know at every given moment the real deterioration status of each piece of equipment. This knowledge would allow to select and schedule the best moment for maintenance, right before a component breaks or leads to unacceptable performance for the system/production chain of which it is part of.
Reliable information about the real component remaining lifetime can only be achieved by combining different techniques (trend analysis, components modelling, simulation, …) while the determination of the best maintenance schedule must rely on the correct assessment of the impact that each component has on the whole system as well as the compatibility with company’s production deadlines.
The PROGRAMS main objective is thus the optimal management of a factory maintenance activities, through the maximization of operating life of production systems and the minimization of maintenance related LCC costs. In order to achieve these results, the proposed approach relies on a Smart Objects Technology (named Smart Control System) deployed at factory level and on the concept of cloud-enabled Central Application Layer, able to improve overall business effectiveness with respect to the following perspectives:
- Increasing Availability (A) and then Overall Equipment Effectiveness (OEE) through increasing of MTBF.
- Continuously monitoring the criticality of system components by performing/updating the FMECA analysis.
- Providing alternative methods to estimate the condition and the Remaining Useful Life (RUL) of equipment components. (including physical-based models of the critical components and a data driven approach)
- Determining an optimal strategies and policies for the maintenance activities of each equipment component.
- Integrating the maintenance and production activities into a single schedule that optimizes the overall performances.
- Providing robust and customizable data analysis services.
- Allowing the company staff to actively participate to the maintenance management by an intra factory information service.
The PROGRAMS project aims at reaching its intended objective by developing the following novel technologies:
- An easy to install hardware system to collect process data and exploit them at machine level (WP2).
- Several alternative RUL computation methods (including a model-based prognostics method and a data driven approach) for the smart evaluation of equipment conditions (WP2-WP3)
- A support tool to speed up the FMECA deployment (WP4)
- A tool for reliability parameters estimation of components based on maintenance reports (WP4)
- A novel DSS tool for optimal maintenance strategy determination (WP4).
- A tool for merging maintenance and production schedules with integrated ERP support (WP4)
- A platform to share maintenance and machines related information between involved personnel (WP5).
The proposed approach will have an important impact especially on the SME sector, where the need for flexibility, production efficiency, and strategies optimization analysis, often meets resource limitation, localized knowledge gap, and lack of available tools.
The main benefits of the project outcomes are expected along the following lines:
Increased Availability and Overall Equipment Effectiveness
Continuous monitoring of system components criticality
Creation of physical-based models of the components which have a higher criticality level
Selection of optimal strategies for the maintenance activities
Integration of maintenance and production activities minimizing overall LCC
Provision of robust and customizable data analysis services
Development of an Intra Factory Information Service
The overall concept behind this project is based on 3 key layers:
Physical Layer: predictive maintenance techniques relied on data monitoring
Cloud Layer: allowing affordable time and space for elaboration techniques
ICT Layer: ICT infrastructure and protocols which are the backbone for the other two layers
|Total budget - Public funding:||5 995 272,00 Euro - 5 995 272,00 Euro|
|Call topic:||Novel design and predictive maintenance technologies for increased operating life of production systems (FoF.2017.09)|
This is a set of Specific Objectives and Research & Innovation Objectives that is subject to a consultation in preparation of the Made In Europe Partnership. For more guidance about the consultation, please see www.effra.eu/made-in-europe-state-play.
- Inherent Availability (+5%)
- Schedule Adherence (+5%)
- Reaction time (-30%)
- Difference between real and virtual position signals (5-10%)
- Rate of successful parameter adaptions (80%+)
- Anomaly detection effectiveness (90%+)
- FMECA completeness (+30%)
- LCC (-25%÷30% on the medium-long term)
- Effectiveness of maintenance optimisation (-20% )
PROGRAMS contributed to shorten the bridge from the data collected on the factory lavel to its application at all the levels in the factory ecosystem, including the process parameter adaption, the computation of components status, the identification of best maintenance strategies and the creation of an integrated maintenance/production schedule. Developed solutions now require extensive testing to verify the degree of improvement that they can provide. Among the KPIs that will be used for such validation, there are:
PROGRAMS succesfully deployed a fast and close loop starting from the acquisition of data from the sensors located at factory level, their exploitation to evaluate the condition of the components of a production equipment and finally the use of such information to modify the processing paramers by adapting them to current equipment status. Developed solutions require the acquisition of data over a longer timespan to better train the artificial intellgence and digital twins based methods. Among the KPIs that will be used for such validation, there are:
PROGRAMS deployed a FMECA constrained Maintenance Decision Support System to identify which are the critical components of a production equipment and the most profitable maintenance stategies and policies to be used for each of them. Developed solutions require an accurate comparison of the output of the AI driven optimization and the real cost assocaited to maintenance management. Among the KPIs that will be used for such validation, there are:
- Managing data relative to all maintenance strategies (including PdM)
- Evaluating the cost associated to different maintenance strategies and policies
- Select and allocate the optimal maintenance Strategies (considering PdM) and Policies for each factory/machine assets (that minimise the overall cost)
- Allowing the seamless transfer of information from all factory levels
- Allowing easy PdM solutions deployment and exploitation
- Optimizing integration of PdM based maintenance with production activities
- Gathering and sharing maintenance information at all factory levels
PROGRAMS aims at developing a HW/SW suite of solutions capable of:
PROGRAMS solution will allow SMEs to access the benefits of Predicitive Maintenance wilth limited costs.
- Confident and robust
- Easy to integrate
- Quick to deploy
A survey among the customers of PROGRAMS industrial partners confirms that Predictive Maintenance practice will be effectively exploited ONLY IF it is:
In addition Predicitive maintenance solutions must be flexible enough to allow different objectives, like the avoidance of sudden failure or the limitation of performace degradation.
- Lack of pre-existing maintenance data: Industry 4.0 is only slightly improving the deployment of tools for collecting data
- Difficult data synchronization: existing data is saved into tens of different formats and with different sampling frequencies
- Lack of sensors data relative to equipment fault status: equipment is never purposefully left to reach such a degraded status and, even then, faults happens only few times a year (so there is an high chance of never seeing faults during project duration).
Several challenges limit the succesfull application of Predictive Maintenance in factories:
- analysis of real sensor data coupled with physically-based models allows to understand the real status of system components
- optimal maintenance strategies allows to perform maintenance before components crash
- precise determination of components RUL will allow to replace/repair components before critical downtime
- smart maintenance strategies allow to minimize maintenance impact on the production activities
- repair or replace of components before crash will lead to less unexpected components breakage
- rationalized maintenance crews deployments will reduce the number of unavailable spare parts or maintenance teams.
- shared maintenance information will reduce time for reacting to unplanned outages.
- Smart RULE algorithms will allow to exploit a component up to its full useful life
- tuning of processing parameters allows to use equipment for a longer time
- Demonstrate that, when correctly applied, the LCC of machines and components actually decreases when using PdM based technology, while the machine availability increases.
- Provide an easy to install solution able to detect the deterioration of performance (at component, machine and system level) and manage maintenance accordingly.
- Implement SotA security layers, showing that the security of maintenance data transmission methods is on par with widespread and accepted communication protocols, like the ones used for e-commerce.
PROGRAMS solution will allow to gain a “10% increased in-service efficiency through reduced failure rates, downtime due to repair, unplanned plant or production system outages and extension of component life.”•reduced failure rates:
All these aspects will reduce the costs related to maintenance activities, thus increasing the sustainability of the production process.
PROGRAMS solutions will allow a more widespread adoption of predictive maintenance as a result of the demonstration of more accurate, secure and trustworthy techniques at component, machine and system level. In fact one of the biggest obstacles is getting people to change long-held maintenance practices. PROGRAMS will:
Precise determination of the RUL of components will allow to replace them before their status degrades production equipment performances beyond unacceptable levels.
Reducing failure rates and production equipment unavailability will improve factory productivity.
- Component failures will be reduced by optimal maintenance strategies and analysis of real sensor data coupled with physically-based models .
- Maintenance efficiency will be increased by precisely pinpointing the components whose deterioration is or will impairing the system performance.
- PROGRAMS will allow to identify design problems or faulty components batches by comparing initial FMECA analysis with updated reliability information.
- A plant wide network will constantly update employees on maintenance activities and components degraded performances.
- The automatic tuning of control parameters will mitigate the impact of a damaged/worn components until maintenance can restore them to their optimal conditions.
- Production equipment operators
- Maintenance operators
- Data scientists
- Maintenance managers
- Software developers
PROGRAMS solution will increase accident mitigation capability”
Predictive maintenance requires different skills and thus new professional figures will have to to be trained:
PROGRAMS interoperability at platform level is granted by the choice of a widely shared communication approach: JSON files over HTTP protocol. Common modules architectures and data formats for file exchange reinforce the PROGRAMS interoperable approach.
A Common Authentication System based on user credentials is shared by all PROGRAMS modules.
NIST guidelines are being followed to manage Users access.
PROGRAMS project aims at integrating its platform with a number of legacy system, like ERPs or Life Data Analysis tools. Since it is not possible to cover all commercial tools interfaces, the consortium decided instead to provide open interfaces (based on JSON files over HTTP protocol exchange) for all its modules.
More than 230 existing Standards have been collected and analysed by the PROGRAMS consortium. As a result, 10 of them have been considered extremely important (labelled as ‘Mandatory”) to foster exploitability of project results. These identified Standards refers mainly to ICT-related topics such as cyber security, data communication/exchange protocols, data mining and control SW.
The project has been very recently established and thus no information about standardization is currently available.
This information will be added by the end of the second year of the project (2018-09).
No new Standards will be developed within PROGRAMS due to the very long time requested for the concerned procedures at ISO level committee and to the limitation of the PROGRAMS solutions that new (unused) standards would impose. Anyway if gaps between already existing standards and standardization needs will be found, appropriate strategies to handle them will be defined, including the identification of potential stakeholders and of a timescale for potential standardization.
Visit Twitter Page
Go to project forum
- No published items
- | FIDIA SPA (Coördinator)
- | AURRENAK S COOP
- | Budapest University of Technology and Economics
- | CASP
- | CE.SI CENTRO STUDI INDUSTRIALI DI TADDEI SIMONA MARIA EC SAS
- | CICERO HELLAS SA
- | DEBBACHE-LAGIOS EE (Gizelis Robotics)
- | IK4-IDEKO
- | LMS - University of Patras
- | RHEINISCH-WESTFAELISCHE TECHNISCHE HOCHSCHULE AACHEN - WZL
- | SAVVY DATA SYSTEMS SL
- | University of Brescia
- | We Plus S.r.l.