ProjectDynamisk Underhåll och Schemaläggning för Tåg

Business Areas
Dynamisk Underhåll och Schemaläggning för Tåg

Doing the Right Maintenance at the Right Time - that is what the DUST project is about.

Today's rail vehicles contains numerous sensors and advanced computers measuring the status and condition of train components. In addition, railside condition monitoring equipment is becoming more and more common. In the DUST project we investigate the complete chain of maintenance information flow, from sensor data to optimized maintenance plans. The companies participating in the project are Green Cargo, Euromaint and Bombardier Transportation.

 Vision

The project proposes a new integrated and flexible maintenance concept,with the following properties:

  • It connects the information flow all the way from the low level sensors up to the high level operations and planning.
  • It takes the uncertainties in the process into appropriate consideration.
  • It allows for dynamic replanning on short notice in the light of new indications and events.

The flow of information is shematically shown in the figure above. The bottom represents the large amount of inputs and sensors measuring the condition and state of the system. As we move upwards in the figure, this raw information is gradually refined into more high level information, like indications of service needs, diagnoses of detectedproblems, and prognoses of remaining operation time until service is required. This information in turn is input to planning of operation and maintenance of different parts of the system, and finally these plans have to be synthesized into the global plan for the operation. The small triangles represent that each information processing typically uses several information sources and compiles them into one or a few output values (high fan-in, low fan-out). Finally all information has converged into the main plan.

Publications
Number of items: 7.

Bohlin, Markus and Holst, Anders and Ekman, Jan and Sellin, Ola and Lindström, Björn and Larsen, Stefan (2012) Statistical Anomaly Detection for Train Fleets. In: Proceedings of the 21st Innovative Applications of Artificial Intelligence Conference, 2012. (Submitted)

Bohlin, Markus and Holst, Anders and Ekman, Jan and Sellin, Ola and Lindström, Björn and Larsen, Stefan (2012) Statistical Anomaly Detection for Train Fleets. In: Proceedings of the 21st Innovative Applications of Artificial Intelligence Conference.

Bohlin, Markus and Ekman, Jan and Holst, Anders (2011) The opportunistic replacement and inspection problem for components with a stochastic life time. [SICS Report]

Ekman, Jan and Holst, Anders and Jonasson, Arndt (2011) Condition based maintenance of trains doors. [SICS Report]

Holst, Anders (2011) Analys och prediktion av slitage på hjulprofiler och bromsbelägg på tåg. [SICS Report]

Aronsson, Martin and Bohlin, Markus and Doganay, Kivanc and Holst, Anders and Kjellqvist, Tommy and Östlund, Stefan (2010) An Integrated Adaptive Maintenance Concept. In: International Conference on Condition Monitoring and Diagnosis 2010, 6-11 September 2010, Tokyo, Japan.

Doganay, Kivanc and Bohlin, Markus (2010) Maintenance plan optimization for a train fleet. In: 12th International Conference on Computer System Design and Operation in Railways and other Transit Systems (COMPRAIL 2010), 31 August - 2 September 2010, Beijing, China.

This list was generated on Wed Dec 13 12:43:22 2017 CET.