Fraunhofer Institute for Industrial Mathematics ITWM
Fraunhofer Institute for Industrial Mathematics ITWM
Restlaufzeitvorhersage für Filter: Angepasste Learning-Methoden ermöglichen die Vorhersage der Standzeit von Filtersystemen während des Betriebs.
© iStockphoto
Restlaufzeitvorhersage für Filter: Angepasste Learning-Methoden ermöglichen die Vorhersage der Standzeit von Filtersystemen während des Betriebs.

Online Prediction of Filter Service Life

Using problem-adapted learning methods, the remaining operating time can be predicted with reduced data requirements.

© iStockphoto

Predictive Maintenance for Filtration

Online Service Life Predictions for Filter Systems

When should a filter be replaced to avoid failures while also maximizing its service life? Using model-based methods for predictive maintenance, we forecast the remaining useful life of filters.

Filter devices and filter elements must be taken into account in maintenance planning to minimize downtime. At the same time, operational safety must be ensured. Relying solely on differential pressure measurements therefore only delivers reliable results in exceptional cases. Since it merely represents a snapshot in time, usable service life is often »wasted« in this way. A more effective approach is to use the history of operating conditions to predict service life.

Depending on the application area, practical implementation presents various challenges. Unlike standardized tests on filter test rigs, not all information about the contamination to be removed is typically available during operation, such as concentration and particle size and shape distribution. In mobile applications in agriculture and construction in particular, seasonal and geographic variations in operating conditions add further complexity. In addition, available sensor technology imposes limitations, meaning that remaining useful life prediction is often based on only a few measured variables.

Remaining Service Life Prediction Through Modeling and Simulation

Here, we draw on our many years of experience in modeling and simulating filtration processes. The relationships between fluid volumetric flow, concentration, separation, and the increase in differential pressure follow well-defined principles. Accordingly, a wide range of suitable models is available for different application areas. With the appropriate model selection, the amount of data required to update the service life prediction can be significantly reduced.

Privacy warning

With the click on the play button an external video from www.youtube.com is loaded and started. Your data is possible transferred and stored to third party. Do not start the video if you disagree. Find more about the youtube privacy statement under the following link: https://policies.google.com/privacy

The service life of a hydraulic filter element is predicted based on measured differential pressure as well as oil and ambient temperature. The prediction of the scaled pressure differential (smooth curve) takes into account the fluctuating measurement data recorded up to that point. The vertical line marks the current point in time.

Our approach has been established within industrial projects and has been successfully applied, among other areas, to the demanding field of hydraulic filters for agricultural machinery. Thanks to the high flexibility of the approach, it can also be transferred to other fields such as air filtration or filter regeneration.

Related Links: