Community Research and Development Information Service - CORDIS

H2020

INFRALERT Report Summary

Project ID: 636496
Funded under: H2020-EU.3.4.

Periodic Reporting for period 1 - INFRALERT (LINEAR INFRASTRUCTURE EFFICIENCY IMPROVEMENT BY AUTOMATED LEARNING AND OPTIMISED PREDICTIVE MAINTENANCE TECHNIQUES)

Reporting period: 2015-05-01 to 2016-10-31

Summary of the context and overall objectives of the project

The condition of the land transport infrastructure has a big societal and economic relevance, since constraints result in disruptions of service. The demand for surface transport will significantly increase in the next years. Given budget restrictions, a substantial enlargement of the road/rail network in the next decades is doubtful. Besides, the aging infrastructure will require more maintenance interventions which infer normal traffic operation. Therefore, the only way to increase infrastructure capacity for the increased transportation demand is to optimise the performance of the existing infrastructure. This is precisely the goal tackled by INFRALERT.
INFRALERT aims to develop an expert-based information system to support and automate infrastructure management from measurement to maintenance. This includes the collection, storage and analysis of inspection data, the determination of maintenance tasks necessary to keep the performance of the infrastructure system in optimal condition, and the optimal planning of interventions.
The major challenges of INFRALERT are:
- Developing information technologies and standard procedures applicable to linear transport systems in general.
- Developing expert-based toolkits built on artificial intelligence and optimization techniques to support decision making in maintenance planning, renewal and new construction
- Integrating all previous models and tools in a cloud-based framework compatible with existing asset management systems.

The overall goal of INFRALERT is to improve the operability and functionality of linear asset transport infrastructures based on large-scale automated predicting, determining and planning of maintenance and renewal interventions in order to support decision making.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

The main achievement of WP1 is the definition of requirements for the whole expert-based Infrastructure Management System (eIMS) as well as for the several sub-components. External and internal analyses have been carried out taking into account the drawbacks of existing tools, from this functional and technical requirements were derived. The positive and negative factors have been identified at external and internal level, and the outcome of this analysis has been concluded in a SWOT matrix for the project.
In WP2 the definition of abstract data types describing the linear transport infrastructure with its relevant features and parameters is given, using the universal concepts of data ontologies. An important aspect in infrastructure management is the localisation of data and how data are being associated to measurements. The methodologies and database architectures for efficient localisation have been selected and implemented.
In the first period for WP3 the purpose was to set the framework for the methodologies for asset condition assessment and prediction. The framework also includes the hierarchy of condition information and the selection of relevant indicators. Starting from a functional description of linear asset and a description of degradation mechanisms, corresponding models are then described. The condition information provided by WP3 is the essential asset knowledge needed for the further WPs on Asset management and Decision support.
Work performed in WP4, which deals with the management of maintenance alerts, covers the development of several machine learning procedures to generate alerts from infrastructure conditions. In a first step, the concept of asset alert levels has been introduced, which relates historical and current asset features, maintenance interventions and condition assessment and predictions in order to conclude an assets' severity level. Based on this concept, different algorithms for the automated recognition of alerts have been implemented and validated using a real-world scenario. A fusion model combining the different machine learning methods has been selected.
Focus of the developments in WP5 is on a probabilistic approach to the calculation of RAMS and LCC parameters. The data necessary to perform RAMS and LCC analysis have been identified, which can be divided in two types: data concerning failures and repairs and data concerning costs. The methodology to derive statistical measures useful for failure rate analysis has been implemented on component and system level.
An important effort in this first period for WP6 was dedicated to specify the theoretical background of condition- and risk-based intervention planning and setting a framework for smart operation and maintenance decision support. The defined framework includes robust optimisation techniques to tackle uncertainties coming from several sources of input data - asset condition, interventions to be applied, LCC and RAMS parameters. For the application of this framework the respective mathematical models have been defined.
The definition of the overall eIMS solution, which is part of the ongoing work in WP7, resulted in a detailled list of actors, use cases and functionalities. After evaluations of requirements coming from WP1, important design decisions for the further implementations have been made. First prototype screens and wireframes have already been presented. Security considerations regarding eIMS framework are described based on risk assessments, use of existing standards and WP1 requirement analysis.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

The main aspects in which INFRALERT will make significant progress beyond the state of art are the following:
- Develop an expert-based Infrastructure Management System (eIMS) with AI features like asset nowcasting and forecasting, machine-learning based alert management, RAMS&LCC analysis and smart decision support tools.
- Develop efficient data management in the eIMS (cloud-based) supporting flexibility, adaptability and evolutionary infrastructure maintenance.

The overall economic impact will be an increase in the performance of transport infrastructures, keeping pace with the growing demand for traffic. This is shown in two dimensions: (1) an improvement in the cost-effectiveness of infrastructure operation and maintenance, with respect to both the recurrent costs resulting from maintenance interventions to be executed as well as to the overall life cycle costs of the assets coming from a long-term assessment and evaluation, and (2) a guarantee to transport capacity, which is necessary for the infrastructure system to perform as required, by enhancing both the reliability and the availability of the networks.

Related to environmental and social impact there will be (3) an optimised energy and resource consumption, reduction of pollution (due to greenhouse gas (GHG) and noise emission), and (4) an improved service quality in the transport of goods and people, higher comfort and safety for passengers and less traffic disruptions and delays; better conditions for working crew because of a more organised planning of maintenance tasks.

The main outcomes of INFRALERT will be:
- Ensuring service reliability and safety by minimising incidences and failures of decaying assets.
- Keeping and increasing the infrastructure availability by optimising operational maintenance interventions and strategic long-term planning decisions on new construction.
- Ensuring the operability under traffic disruptions due to interventions.

Related information

Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top