Skip to main content

VerSatilE plug-and-play platform enabling remote pREdictive mainteNAnce

Periodic Reporting for period 1 - SERENA (VerSatilE plug-and-play platform enabling remote pREdictive mainteNAnce)

Periodo di rendicontazione: 2017-10-01 al 2019-03-31

The growing complexity of modern engineering systems and manufacturing processes is an obstacle to implementing Intelligent Manufacturing Systems (IMS) and keeping these systems operating at high levels of reliability. The number of sensors and the amount of data gathered on the factory floor increases constantly, while there are hidden resources, 85% of data and information are unstructured and 42% of all transactions (sending and receiving information) are still based on paper. This opens the vision of truly connected production processes where all machinery data are accessible allowing easier maintenance in case of unexpected events. Physical maintenance issues can cause costly disruptions in the manufacturing process. With predictive analytics, however, repairs and maintenance tasks can be prioritized and allocated to planned outages based on real-time probabilities of various future failures.

The strategy of predictive maintenance saves time and money and helps minimize costly production down times. Additionally, the use of remote predictive maintenance enables the minimization of the production dead time and the cost for spare parts and suppliers. Last but not least, predictive maintenance techniques such as vibration and thermal monitoring along with Reliability techniques such as Failure Modes and Effects Analysis (FMEA) and Root Cause Failure Analysis (RCFA) will result in bottom-line savings through early detection.

Considering the above, SERENA will provide a bridge for transferring the latest R&D results in predictive maintenance towards inherently different industrial sectors (white goods, metrology engineering, elevators and steel parts production) considering the needs for versatility, transferability, remote monitoring & control, by providing:
- advanced IoT systems and smart devices for collecting data from different resources (robots, machines, welding guns, PLCs, external sensors etc.) and cloud-based remote management of these data
- versatile remote platform for predictive maintenance activities & augmented reality AR based operator local maintenance personnel support,
- advanced artificial intelligence methods for predictive maintenance involving advanced data analytics,
- plug-and-play cloud-based communication framework being applied in a variety of machines and plants.
"In the period from M01 to M18, the work performed can be summarized as follows:

- SERENA use cases and requirements: The project requirements catalog has been defined. The use cases and system requirements have utilized the pilot cases’ definition but are provided as a generalized set of requirements so as to increase the target of SERENA system. Within the work carried out for the pilots the industrial partners identified the as-is situation, potential for improvement and the to-be scenarios incorporating SERENA results.
- SERENA pilot cases test-beds: Based on the requirements stemming from the ""SERENA use cases and requirements"", the designs of the test-beds, the definition of the scenarios to be undertaken as well as the specification of hardware and software requirements have been carried out. Towards validating at an early stage, the SERENA proposed reference architecture, COMAU has designed and implemented a robotic test-bed, collected, and provided the data from it as well as defined the failure modes to be investigated. Additionally, the designs of the other use cases have been defined. Finally, the individual test-beds have been connected to the SERENA cloud in order to provide machine data from the edge up to the cloud.
- SERENA platform design and architecture: A common approach has been conceived to address all project industrial pilots. The SERENA architecture is designed in a way that allows for modularity and configurability thus facilitating the application of the SERENA platform in different industrial environments. The reference architecture is instantiated in each use case.
- SERENA technical developments:
1. A prototype implementation of the SERENA system has been completed in November 2018, covering a basic workflow from acquiring data from the robotic test-bed, namely robot box, up to the scheduling of maintenance operations based on predictive analytics.
2. A robot test-bed namely robot box, implemented by COMAU
3. Gateway components have been tested and integrated, namely the IPT databox and the commercial gateways from DELL, transferring data from the edge to the cloud
4. Edge analytics converting raw data to ""smart data"", such as calculating the RMS value from the current values of a motor.
5. The cloud infrastructure and platform
6. The integration of the MIMOSA schema to the platform architecture
7. The creation of JSON-LD schema per case and based on the MIMOSA schema
8. A generic self-adjusted analytics service has been implemented which will be tested and refined through the different use cases up to its full validation in the final demonstrator test-bed.
8. A scheduling component has been implemented using the predictive analytics result to create a new maintenance aware schedule, based on a multi-criteria decision making framework.
9. A real time visualization service providing status updated of the monitored equipment and based on the acquired data.
10. An AR platform for providing instruction to maintenance personnel.
11. The definition and design of the security middle-ware for the cloud platform.

Based on the first validation of the SERENA integrated approach, the SERENA approach is customized for each individual case following a two step integration approach:
1. Bridge the edge equipment to the SERENA cloud system, following the specific architecture and conventions. [by April 2019]
2. Integrate additional services to each individual demonstrator based on its requirements [by September 2020]"
"SERENA platform will enable the remote predictive maintenance activities by significantly reducing the downtime, production stoppages and the cost, by providing different methods for predicting failures, including advanced data analytics, hybrid physics-based and data driven approaches, as well as planning/scheduling of repairing activities. Moreover, SERENA targets the reduction of the possibility of failures and accidents within the production lines considering the contribution of SERENA solutions:
# Early detection of failures through AI condition based maintenance;
# Preventive actions and planning/scheduling of repairing/maintenance activities based on the importance and the estimation of RUL of components;
# Increased accuracy of prediction by introducing hybrid models considering physics-based and data-driven approaches;
# Increase awareness of the production status and potential failures.
In conclusion, SERENA innovations can be summarized as follows and based on two different perspectives:
1. From the analytics perspective, SERENA proposes an innovative analytics service, bridging data driven and physic based approaches, through the evaluation of different machine learning techniques with the capability of self-assessment.
2. From the platform point of view, SERENA proposes a scalable and resilient hybrid cloud platform, decoupled of specific technologies and protocols, tested in 5 different industrial use cases."
SERENA early validation in November 2018 architecture