Sensor development, intelligent data logging, mining and analysis, model updating and statistical methods to predict the performance of aging infrastructures


The problem addressed

Sensor development, intelligent data logging, mining and analysis, model updating and statistical methods to predict the performance of aging infrastructures Safety is a crucial issue for operators and regulators in many sectors such as the oil & gas industry, the chemical industry as well as for the transportation infrastructure. Aging and degradation can increase the probability of failures, possibly leading to accidents with severe consequences such as human fatalities, impacts on environment, bad company reputation and business interruption. The expected outcome of this project will be a significant step towards Condition-Based Management which is based on monitoring and forecasting the integrity of structures and is the most effective way to safeguard structural integrity while reducing maintenance costs and maximizing the “up-time” of structures. The predictive risk assessment and management system will provide a real-time identification of degradation states, prediction of remaining useful life (RUL) and dynamic risk-informed ranking of risks, continuously updated based on the predictions of RUL.

A focus of the project is the question, how data fusion and trend analysis techniques using big data can improve the accuracy of predictions and the control of aging risk. This is linked to the integration of these approaches into efficient maintenance management systems.

The benefit of the successful integration of prognostics and health management (PHM) and probabilistic safety assessment (PSA) is twofold: on one side, it will provide the risk-informed basis for decision making to prevent accidents and control abnormal events; on the other side, it allows the development of a risk-prioritized predictive maintenance management system that will improve safety while reducing the overall maintenance costs.


Keywords: sensors maintenance data fusion

Research questions

The project will analyze how data fusion and trend analysis techniques using big data can improve the accuracy of predictions and the control of aging risk. This is linked to the integration of these approaches into efficient maintenance management systems.

Scientific disciplines: chemistry/chemical engineering

Expected outputs

The project will develop:
  • a validated methodology for PHM, with the associated algorithms for detection, diagnosis and prognosis

  • a methodology for performing an update of a finite element model of a structure with measured structural data, in the context of structural health monitoring

  • a methodology for integrated risk assessment and management, and the predictive maintenance management system, with the algorithms for prioritization of the risks and for identification of the portfolio of optimal maintenance activities
  • a method to prioritize critical structures and determine the weak spots that have to be monitored
  • two case studies to validate the methods developed, one concerning load bearing structures and the other the oil and gas industry.

Workplan

The project is built around 7 main work packages. The first part addresses the development of sensors for corrosion detection, the second work package deals appropriate tools for data acquisition, especially in the context of big data. Work package 3 uses the monitoring data for model updating techniques with a focus on Bayesian methods. In the fourth work package, data mining techniques to extract important information from the data as well as models for trend analysis are developed. In the fifth work package, those models are used to determine the system reliability as well as the evolution of the performance over the service life combined with optimization techniques for monitoring and maintenance planning. The final work package deals with a validation of the developed methods for practical reference cases.

Participating researchers

Jörg Unger (BAM, Germany) — project coordinator

Piero Baraldi (Politecnico di Milano, Italy)

Francesco Maio (Politecnico di Milano, Italy)

Johan Reinders (TNO, The Netherlands)

Annika Radermacher (BAM, Germany)

Funding organizations

BAM (Germany)

INAIL (Italy)

TNO (The Netherlands)

More details

Duration 2017-10 to 2020-10
Contact email joerg.unger@bam.de
More information

Information last updated on 2019-05-13.

Return to the list of projects.