Woensdag 27 oktober om 11:00 uur:
Prepared for a datadriven future?
Mischa Beckers (HZ University of Applied Sciences)
A data driven and risk based approach often go hand in hand when dealing with smart maintenance and operations. Using an asset management framework including functional decomposition, FME(C)A, criticality analysis etc. may lead to a set of components that need (immediate and) most attention. Analytics helps in getting insight in current status of components, relating behavior to observed consequences and is preferably used for predicting future behavior and status. However, what does that insight actually tell you, is the observed relation really significant and how to interpret a prediction model? Both the availability and the amount and quality of data have an effect on this. Next to that, the objective of using Analytics, such a binary or multiclass classification (failure or failure classes), regression (lifetime or failure time) or probability estimation (survival analysis) is dependent on the type of data that is available.
By means of several real-life examples we unravel some of the underlying basic principles and best-practices. Among these examples are:
- RIOBASE, prediction of failure and classification of damages for sewer pumping stations and piping using both sensor and process data and inspection robot videos/images
- COMPI, remaining useful lifetime prediction for valves in process industry using both process, maintenance and sensor data
- CAMPIONE, survival analysis for predicting useful lifetime prediction for heat pumps in buildings when only process data is available
- AURTUB, classification of erosion categories for wind turbine blades using 3D laser scans
Donderdag 28 oktober om 14.15 uur:
Electric motors in the circular economy
Kurt Stockman (Universiteit Gent)
The presentation starts with a Life Cycle Analysis of an electric motor, starting from the manufacturing of the machine over the motor sizing, down to the disposal and recycling phase. This analysis shows the critical aspects that should be taken into account when optimal performance of the machine in terms of reliability, livespan and energy efficiency. In the second part of the presentation, the impact of maintenance is discussed. Modern sensor systems for condition monitoring are flooding the market and provide valuable data on the actual performance of the motor system. The question arises how this data should be used from the perspective of the motor maintenance. Based on the data analysis and the initial motor sizing, several scenarios are discussed to improve the motor system performance in terms of energy consumption and reliability. The proposed strategy is illustrated on an industrial use case in the chemical industry. The results presented in this presentation are the outcome of an Interreg project on Circular Maintenance in the chemical industry. The outlines of this project will also be highlighted in this presentation.
Praktijklab Corrosie & Isolatie
Donderdag 28 oktober om 14.15 uur:
Practice-oriented research into new methods for corrosion management and monitoring
Jeroen Tacq (Sirris) & Leendert Schouten (KicMPi)
KIC|MPI (Kennis en innovatiecentrum Maintenance Procesindustrie), Scalda, the Antwerp Maritime Academy and Sirris are developing infrastructure where new corrosion management solutions can be developed, tested and demonstrated in realistic but controlled circumstances. These new ‘CorrosionLabs’, are supported by the EU under the Interreg program Vlaanderen-Nederland. Our goal is to provide the industry with a relevant test-bed where real industrial circumstances can be created in a controlled way. Innovative solutions can be developed and tested, including coating systems, insulation methods and choice of materials. The infrastructure will also be used to investigate how industry 4.0 solutions can change the way in which assets are being maintained and managed. Innovative corrosion and moisture monitoring solutions will be investigated and their performance evaluated together with industrial partners. This includes evaluating minimum detection levels and response rate in relation to the risk presented by monitored parameters. Data driven corrosion management can reduce uncertainty as well as operating costs, but only if there is certainty about sensor performance. Evaluation in a controlled environment can help to get better insight in sensor reliability. In addition to providing a test-bed, the CorrosionLabs will also be able to evaluate corrosion prevention measures like coatings, process water conditioning, new materials and new insulation techniques in a short timeframe by accelerating the degradation processes. It is critical that the accelerated test conditions are sufficiently representative for in-field conditions. This presents an important challenge. Within the CorrosionLabs, new solutions to this challenge can be explored. Data driven corrosion management is more than selecting appropriate sensors. It’s also about where these sensors need to be deployed and what you do with the data. Within the SOCORRO project, Artificial Intelligence is used to develop a corrosion management framework for water-submerged or water-containing structures. Water parameters are linked to a corrosion risk and accumulated risk can be used for asset management. On the long run, the goal of a data driven corrosion management approach should be automatic work order generation. The steps we are taking towards this with the approach in the SOCORRO project will be explained and opportunities for future developments discussed.