Projects / Programmes
E-maintenance of electro-mechanical drives: prognostics and health management solutions under non-stationary operating conditions
| Code |
Science |
Field |
Subfield |
| 2.06.01 |
Engineering sciences and technologies |
Systems and cybernetics |
Control systems technology |
| Code |
Science |
Field |
| P170 |
Natural sciences and mathematics |
Computer science, numerical analysis, systems, control |
| Code |
Science |
Field |
| 2.02 |
Engineering and Technology |
Electrical engineering, Electronic engineering, Information engineering |
E-maintenance; condition monitoring; fault diagnosis; prognostics and health management; rotatioanal machines and drives; signal processing
Researchers (14)
Organisations (3)
Abstract
Rotational machines and drives represent the most ubiquitous item of equipment present in almost all branches of industry, power engineering and transport. As such they contribute significantly to the maintenance costs which still represent sizeable share in the EU economy estimated to 450 billion EUR/year of which about 70 billion EUR/year are estimated to be wasted through ineffective maintenance. Downsizing the cost is becoming a must in order to keep the companies globally competitive. This requires better use of IT support and integration of "information islands" within the enterprise via concept of e-maintenance.
Prognostics and health management (PHM), with on-line condition monitoring (CM), are key decision-support technologies in e-maintenance system. In spite of significant advances in enabling technologies, no massive use of CM (and PHM in particular) in industry has been witnessed so far. According to a review done in USA, only in 1% of the machinery the condition is monitored automatically all the time!
In spite of obvious need for 'general purpose' PHM solutions that would be applicable to a broader range of operating machines the customers are still rather sustained to implement them, primarily due to relatively high cost of implementation and cost of ownership. The problems originate in
(i) - overwhelming and hence costly design cycle;
(ii) - use of conventional CM methodologies incapable to handle non-stationary operational conditions in a simple and efficient way,
(iii) - lacking prognostic abilities, and
(iv) - costly interoperability with existing enterprise resouce planning (ERP) and maintenance management systems (MMS).
Fuelled by these challenges, the project will conduct basic and applied research leading to the following anticipated major results:
1. A TRL5 prototype of a versatile and low cost CM&PHM module (in the sequel referred to as CM&PHM) that will be suitable for implementation on a broad range of already operating machinery as well as an embedded module being part of newly manufactured machines. It will perform inference on asset condition and communicate the results across the e-maintenance system.
2. New algorithms for condition assessment and prognostics of the remaining useful life robust to incomplete prior information about the operating load and speed, external disturbances and other prior data from components manufacturers. These are the heart of CM&PHM. We will stress on a class of machinery manufactured in series. The promising novel Bayesian fusion concept, will be able to merge information from on-line sensors, tribological analysis and past life cycles in the final prediction.
3. Novel interfaces to e-maintenance systems supporting tools for machinery life cycle management. The information about the machines' condition becomes instantaneously available through the pre-existing ERP/MMS IT infrastructure thus allowing optimal maintenance and production planning. Furthermore, timely CM information reduces the risk of unscheduled production interruptions.
4. The prototype versions of the system will be validated on laboratory test rig and demo industrial installation, already agreed with the company supporting the project.
Project will be realised by academic partners with top rate expertise in CM and PHM, embedded systems design and tribology and a leading manufacturers of electrical motors.
The immediate envisaged impact of the project:
(i) the oririginal equipment manufacturers in Slovenia will get the opportunity to advance their products with low cost embedded PHM solutions, hence coming up with a new generation of self-aware machinery, highly competitive on the global market. Domel has vision of 1 mil.EC motors per year with embedded PHM by 2020.
(ii) by installing the solutions on existing machines, their owners can expect near zero downtimes, minimal maintenance resources and h
Significance for science
The expected research accomplishments that will genuinly contribute to the field of PHM are the following:
· A powerful method for threshold selection in condition monitoring of mechanical drives under nonstationary operating conditions. No systematic solution to this difficult problem is known so far. The anticipated solution will require only one design parameter and will be insensitive to fluctuating conditions. This will have direct impact on speeding up the design phase and reducing the design costs.
· Novel Bayesian framework for prediction of the remaining useful life is believed to be a qualitatively novel approach in the area of PHM. Fusion of mixed data types like run-to-failure data records, results of maintenance inspections, reliability data available from the component manufacturers and tribological parameters will be possible in a unified manner.
· A data driven model that describes the relationship between tribological parameters, vibrational patterns and RUL is hoped to be found from experimental data. Very likely we will adopt the Gaussian process model, a non-parametric model that requires no priors.
· The proposed cyber-physical system (CM&PHM with accompanying algorithms and ERP integration module) represents a link between the management and the production level. Such a system will yield vital information about the current condition of critical industrial asset, thus enabling the optimisation of production, maintenance and management processes. The information about the machines' condition becomes instantaneously available through the pre-existing ERP/MMS systems thus allowing optimal maintenance and production planning. This will lead to significant reduction of risk of unscheduled downtimes.
· A comprehensive subset od data obtained through experiments on various scales will be made public for other researchers. Data collected under naturally evolving faults, and monitored tribological processes are holly grail for the community and of great help for the researchers to assess their PHM ideas.
Significance for the country
a. Significance for original equipment manufacturers (OEMs). OEM, especially in machining and metalworking sector, are expected to be main beneficiaries in exploitation of the developed prototypes. With 50.000 employees and 6.4 billion € revenues this sector is regarded as being near the top of recession-resistant industries. Companies are strongly committed to investments into more efficient technologies. From figures mentioned, it is clear that there is a large potential market for deploying machines with automated self-diagnosis. Such machines are highly attractive for end-users since full e-maintenance functionality is available at almost no cost of maintenance and radically downsized cost of ownership. Domel as a world leader in the sector of vacuum cleaner motors has strong vision of extending their current EC motors with embedded PHM functionality. With such a novel family of smart products they are likely to be pioneers in this highly challenging domain.
b. Significance for end-users in industry. Apart of few exceptions the general level of maintenance practice in Slovenia is still rather poor. Rotational machines with on-line condition monitoring are very rare. Several industrial partners from the Competence Centre for Advanced Control Technology, a manufacturer of rolling bearings and a water distribution company have already endorsed their interest to assess the prototype solutions.
c. Development of a new enterprise. The project is expected to bring original knowledge and new skills in the domain of e-maintenance which form a good ground for an innovative spin-out company. The portfolio of solutions is believed to assure high competitiveness with respect to CM solution providers. Actually, PHM solution providers today are rare and concentrated on equipment on top of cost rank.
d. Importance for the country. In case of successfully completed project the ambition is to proceed with realisation of the industrial prototype of CM&PHM and market uptake. In that case the entire value chain, from development to production, is Slovenian. The software modules are authentic (copyright) and auxiliary outsourced software is open source.
e. Contribution on the EU scale. The project contributes to the objectives on a wider EU room, mainly through "Leadership in enabling and industrial technologies" work programme. The project is complient with several Horizon 2020 initiatives as for example:
• FoF 1-2014, Cyber-Physical Systems since the proposed concept belongs to the greater class of smart sensor technologies, smart system design and embedded systems.
• SPIRE 1 – 2014: Integrated Process Control. The proposed CM&PHM system allows accurate and real-time estimation of machines' condition enabling more reliable and sustainable industrial operations.
Most important scientific results
Interim report,
final report
Most important socioeconomically and culturally relevant results
Interim report,
final report
