Early warning decision support system for the management of underwater scour risk for road and railway bridges

Lead Research Organisation: University of Strathclyde
Department Name: Civil and Environmental Engineering

Abstract

PARTNERS- Transport Scotland (TS), Network Rail (NR), SEPA, Arup.
CHALLENGE- Flood induced scour is by far the leading cause of bridge failures, resulting in loss of lives, traffic disruption and significant economic losses. TS is responsible for 1,567 bridges, 8% of these currently classified as needing scour protection measures and costing £2m per annum in routine inspections. Similarly, NR Scotland Route includes 1,750 structures susceptible to scour requiring an annual inspection spend of £0.4m. Inspections are expensive and time consuming, and often the information collected is qualitative and subjective.
OBJECTIVE- To develop a prototype of a Decision Support System (DSS) for scour risk management for NR's and TS's bridges, which enhances users' safety, while minimizing traffic disruptions and transport agencies' operational cost. The system will estimate, and continuously update, the present and future scour failure risk using (i) real-time information from monitoring of scour depth (ii) rainfall forecast and (iii) river flow characteristics.
INNOVATION- Sensor and communication technologies offer the possibility to assess in real time the scour depth at critical bridge locations; yet monitoring an entire infrastructure network is not economically sustainable. A way to overcome this limitation is to install monitoring systems on a limited number of critical locations and use a probabilistic approach to extend this information to the entire asset. In the occurrence of a flood, monitoring observations are used to give the real-time best estimate of bridge failure probability. The sensing technology proposed is a prototype of smart probe with integrated electromagnetic sensors designed to detect changes in the dielectric permittivity of the surrounding bridge foundation, designed and calibrated to detect scour and sediment deposition in various soil types.
KEY ACTIVITIES
- Collect data from past scour inspection on NR's and TS's inspection records; define form vulnerability model based on the fragility curve method.
- Define flood hazard models based on SEPA rainfall and river flow datasets, and accounting for CCRA climate change prediction.
- Model correlation in hazard and vulnerability among different bridges, using Bayesian networks.
- Collect data from TS pilot scour monitoring system.
- Implement algorithms for updating fragility curves and state variables of the Bayesian network based on real-time information from monitoring system.
- Develop a prototype DSS that continuously updates bridge failure risk using real-time information from pilot monitoring system and rainfall forecast.
- Produce an exploitation plan.
DELIVERABLES- The BRIDGE VULNERABILITY MODEL, the FLOOD HAZARD MODEL and the SCOUR HAZARD MODEL will provide NR and TS with quantitative up-to-date information on the probability of scour failure of each bridge. Before a flood event, this will allow prioritising inspections according to risk and, when necessary, to introduce precautionary bridge closures or traffic management. This framework can be equally used to simulate the impact of potential extreme weather events on travel agencies' network, and to plan in advance appropriate emergency procedures and countermeasures. The TS PILOT SCOUR MONITORING SYSTEM will provide validation to a new technology for direct assessment of underwater scour, which can complement traditional NR's and TS's river-bank visual assessments. The DECISION SUPPORT SYSTEM will provide information for long term prediction of future scour risk to drive strategic maintenance, repair and rehabilitation actions. The same DSS can be used as an early warning system that automatically takes action when the risk estimated exceeds a given threshold. The EXPLOITATION PLAN will set a roadmap to transfer the new technology and knowledge to other infrastructure systems and geographical areas.
DURATION- 12 months
COST- £94,763.02

Planned Impact

The objective of the project is to develop a prototype of a Decision Support System (DSS) for scour risk management for road and railways bridges, which enhances the safety of users, while minimizing traffic disruptions and operational costs. The system will estimate, and continuously update, the present and future failure risk due to scour using: (i) real-time information from monitoring of scour depth, (ii) rainfall forecasts and (iii) river flow characteristics.

The anticipated benefits of the system are:

1. Direct assessment of the risk of scour bridge failure. The system will provide NR and TS with quantitative up-to-date information on the probability of scour failure of each bridge. Before a flood event, this will give an opportunity to prioritise inspections according to risk and, when necessary, to introduce precautionary bridge closures or traffic management. Similarly, immediately after an event, this will help to identify the critical bridges requiring river-bank or underwater inspections.

2. Support to strategic MR&R. Information for strategic long-term prediction of future scour risk to drive strategic maintenance, repair and rehabilitation (MR&R) actions. Particularly, the information from the DSS will complement the Scour Risk Rating assessed through the Level 2 process as in Procedure EX2502 or BD97/12, providing a formal quantitative risk evaluation, rather than a mere classification. This will facilitate risk ranking and strategic action prioritisation on the bridge stock.

3. Scenario planning for extreme events. The proposed risk analysis framework can be equally used to simulate the impact of potential extreme weather events on transport agencies' road networks, and to plan in advance appropriate emergency procedures and countermeasures.

4. Early warning. As risk assessment is carried out in real-time, based on the measurements from on-line monitoring systems, the same DSS can be used as an early warning system that automatically takes action when the estimated risk exceeds a given threshold. For example, the system could, in principle, control a traffic light which automatically restricts access to the bridge when the prediction of scour level is estimated to be critical for the safety of the users.

5. General method. Although the proposed method is developed for scour risk, its scope is much broader, and can be equally applied to any engineering problem where different data sources are integrated to optimize the operation of infrastructure.
 
Description Scour is the removal of river bed material due to flowing water, and it is the leading cause of bridge failure worldwide. This project delivered models and tools for improving the management of bridge scour risk by integrating into a single framework the existing knowledge in probabilistic risk analysis, sensor technology and decision making.
We aimed at delivering a Decision Support System that uses data from scour monitoring systems to achieve a more confined estimate of bridge scour risk. These outcomes are then used to inform a decision model thus supporting transport agencies' decision schemes. The project was broken down into three phases.
The first phase was to define the bridge scour capacity by developing a vulnerability model. Scour vulnerability assessment methodologies, used by Transport Scotland and Network Rail, were analysed to build scour fragility functions.
The second phase was to define the scour demand. Starting from a flood model (involving SEPA), the total scour depth is estimated by a scour hazard model based on Bayesian Networks. The model can predict the scour depth at any bridge component, even if it is unmonitored, by exploiting observation from scour monitoring systems installed on limited numbers of bridges.
The third and final phase was to develop a decision model. Transport Scotland's and Network Rail's decision procedures follow a "visual" decision process (to close or not the bridge) based on water level markers installed on bridges. The decision model then uses the outcomes from the second phase, based on direct measurement of scour depth, to update the scour threshold that triggers bridge closures thus supporting transport agencies' decision frameworks.
Exploitation Route The impact achieved in this project primarily relates to UK transport agencies. The key outcomes of the Decision Support System are the scour hazard model and the decision model.
The former model provides Network Rail and Transport Scotland with up-to-date quantitative information on the probability of scour failure of each bridge. Their current evaluation practice is based on routine inspection. Inspections are expensive, time consuming and in normal conditions, are repeated every three years. The risk analysis framework gives them the opportunity to prioritise inspections according to risk and, when necessary, to introduce precautionary bridge closures or traffic management.

Transport Scotland's and Network Rail's decision framework are currently based on water level markers only. The proposed decision model uses Bayesian Network's outcomes and observations of the pilot scour monitoring system to inform decision about bridge scour management. In particular, the scour depth is used as quantity to trigger actions. The outcomes present an increase of the scour threshold that could help transport agencies in reducing the times that bridges might be closed unnecessarily as a precautionary action. The increasing of the scour threshold will help the transport agencies in significantly reducing operational and bridge downtime costs.
As a result of these findings, the Decision Support System can be used as an early warning system that automatically takes action when the scour exceeds a given threshold. For instance, the system could control a traffic light, which automatically restricts access to the bridge when the estimated scour level is close to the threshold. Such an automatic system can be applied to any bridge of Transport Scotland and Network Rail, even if not directly monitored with scour sensors.
Sectors Environment,Transport

 
Description There are 220,000 underline rail bridges in Europe, of which 19,000 in the UK, whereas the road bridges are 350,000 in Europe and 30,000 in the UK. Network Rail owns and operates around the underline bridges nationally: approximately 8,700 of these structures are held within a National Scour Database and the projected spend on scour protective works from 2014-2019 is in the region of £27m. For the Scotland route only, 1750 railway bridges are routinely inspected for scour and, of those, 58 are considered to be at high risk. Transport Scotland is responsible for the Scottish trunk road network including 1,567 bridges or culverts over water. The value of these assets is currently £6.41bn, including the Forth Road Bridge. Of those, around 8% are currently classified as needing detailed consideration, including possible monitoring and scour protection measures, and Transport Scotland is currently aware of about £3.5m of known scour repairs and scour resilience works to carry out. Looking at the three scales (i.e., Europe, the UK and Scotland), there are similar figures: 35-40% of bridges are susceptible to scour, of which 8-10% are considered at high risk. Transportation agencies' current evaluation practice is based on routine inspections, which are typically expensive (the average inspection cost is estimated as £3,800 per bridge) and time consuming. Furthermore, failure to identify critical bridges in time, and undertake preventive actions, may result is extensive repair costs. Following the January 2014 flood event, the A82 Kiachnish bridge alone required £320,000 of works to repair the scour damage, works which could have been partially avoided through preventive maintenance and more accurate knowledge of the bridge's prior condition. A current market study on SHM systems found that, because of the ageing of civil infrastructure, SHM in civil infrastructure, that was valued at 530.2 million $ in 2015 (in the global market), will grow at a Compound Annual Growth Rate of 24.16% between 2016 and 2022 and will dominate the SHM market by 2022 accounting for about 38% of this market. 27% of Europe's bridges have spans of more than 25m. One third of these are railway bridges, 67% of which are more than 60 years old. Two thirds of Europe's bridges are highway bridges, and about 50% of these are more than 50 years old. There is therefore the need of risk management tools that help agencies in managing their assets and taking decisions. Implementation of the Decision Support System will significantly reduce the risk of failures, increase the safety of the users, and minimize bridge closures and their indirect socio-economic impact on the community. The early warning system will aid agencies in prioritizing inspection according to risk and, when necessary, to introduce precautionary bridge closures or traffic management based on early warning system feedback. Particularly, the monitoring solution can potentially replace the need to put people/divers into the rivers to visually inspect structures, reducing the inherent dangers that this activity poses. In addition, transportation agencies will be able to provide government with timely, clear advice and information on scour location where the technology is in place. The Decision Supprt System addresses the needs of transport agencies and is developed specifically for the transport infrastructure. However, it can be also applied into bridge management in the broader sense or into general risk analysis and decision support system to other engineering problem where different data sources are integrated to optimize the operation of infrastructure. Currently, there is not such a system on the market, that combine in a risk assessment framework, information from structural health monitoring to address the management of bridges under extreme flood events.
Sector Environment,Transport
 
Description Partecipation in CEN/TC250 "Structural Eurocodes", WG2 "Existing Structures", European Committee for Standardization (CEN)
Geographic Reach Europe 
Policy Influence Type Participation in a advisory committee
 
Description Partecipation in Technical Committee "Bridge Health Monitoring", International Association for Bridge Maintenance and Safety (IABMAS)
Geographic Reach Multiple continents/international 
Policy Influence Type Participation in a advisory committee
 
Description Partecipation in Workgroups UNI-CIS-GL-6 "Structural Health Monitoring" and UNI-CIS-GL-2 "Existing Structures", Italian Standardization Agency (UNI)
Geographic Reach National 
Policy Influence Type Participation in a advisory committee
 
Title Decision Support System 
Description Decision Support System for bridge scour management that exploits data from scour monitoring systems to achieve a more confined estimate of bridge scour risk. These outcomes are then used to inform a decision model thus supporting transport agencies' decision schemes. The Bayesian Network of the Decision Support System provides updated information on bridge scour risk based on a direct scour measurement. These outcomes are used to update the scour threshold that triggers bridge closures thus supporting transport agencies' decision frameworks. The Decision Support System is the first case where a risk assessment framework combines information from scour monitoring to address the bridge management under flood events. 
Type Of Material Computer model/algorithm 
Year Produced 2019 
Provided To Others? Yes  
Impact The Decision Support System allows decreasing inspection costs and unnecessary bridge closures; it can be also used as an early warning system that automatically takes action when the risk exceeds a given threshold. 
 
Title SHM-based Bayesian Network for scour estimation 
Description This probabilistic tool is the first application of Bayesian Networks to bridge scour risk management, and the first implemented case where updating of the network is based on real-time information from a monitoring system. Sensor technology offers the possibility to assess in real-time the scour depth at critical bridge locations, yet monitoring scour in every pier of an entire bridge infrastructure network is not economically feasible for transport agencies. Installing scour monitoring systems at critical locations and using a probabilistic approach overcome this limitation. The developed Bayesian Network is a probabilistic algorithm able to estimates the scour depth using data from the monitored scour depth and then extends this information to the entire asset. 
Type Of Material Computer model/algorithm 
Year Produced 2018 
Provided To Others? Yes  
Impact The probabilistic algorithm provides transport agencies with quantitative up-to-date information on the probability of scour failure of each bridge, even if it is unmonitored. It complements their scour risk rating, providing a formal quantitative risk evaluation, rather than a classification. 
 
Title Scour database 
Description Database of the scour depths measured and recorded by the pilot scour monitoring system installed on the A76 bridge on River Nith. 
Type Of Material Database/Collection of data 
Year Produced 2018 
Provided To Others? No  
Impact Direct measurement of the scour depth is difficuilt, the pilot scour monitoring system provides a direct and accurate measurement, in contrast to monitoring systems based on sonars and smart sensing bars were proposed in the past. They provide an undirect scour measurement and often requires expensive installation and maintenance. Interpretation of data can be time-consuming and difficult. The databese is also representing a valuable source of information for strategic long-term prediction of future scour risk and for calibrating scour progression models. It can drive Network Rail and Transport Scotland strategic maintenance, repair and rehabilitation actions. 
 
Description Arup 
Organisation Arup Group
Country United Kingdom 
Sector Private 
PI Contribution During the NERC project, we have been directly involved in every deliverable and model developed. We brought our expertise in Bayesian statistic, sensors & structural health monitoring, intelligent infrastructure, safety & resilience and computational modelling.
Collaborator Contribution Arup brought their experience in design and management of structural health monitoring systems, and they contributed providing technical advice on the design of the pilot installation and the decision support system. Arup will be involved in the Bridge Vulnerability Model, the Scour Hazard Model and the Decision support system.
Impact The Bridge Vulnerability Model, the Scour Hazard Model and the Decision support system
Start Year 2017
 
Description Heriot-Watt University 
Organisation Heriot-Watt University
Country United Kingdom 
Sector Academic/University 
PI Contribution During the NERC project, we have been directly involved in every deliverable and model developed. We furthermore designed and developed the pilot scour monitoring system installed on the A76 bridge on the River Nith. We brought our expertise in Bayesian statistic, sensors & structural health monitoring, intelligent infrastructure, safety & resilience and computational modelling.
Collaborator Contribution Heriot-watt University brought their experience in quantitative/probabilistic risk assessment of infrastructure systems and computational modelling. Heriot-watt University has been involved in the Scour Hazard Model and the Decision support system.
Impact The Scour Hazard Model and the Decision support system
Start Year 2017
 
Description Network Rail 
Organisation Network Rail Ltd
Country United Kingdom 
Sector Private 
PI Contribution During the NERC project, we have been directly involved in every deliverable and model developed. We brought our expertise in Bayesian statistic, sensors & structural health monitoring, intelligent infrastructure, safety & resilience and computational modelling.
Collaborator Contribution Network Rail provided access to examination report records and other historic information; technical mentoring from their engineers; trial sites for implementation of the scour risk model. Scotland Route office at Network Rail has been involved in the Bridge Vulnerability Model, the Scour Hazard Model and the Decision support system.
Impact The Bridge Vulnerability Model, the Scour Hazard Model and the Decision support system.
Start Year 2017
 
Description SEPA 
Organisation Scottish Environment Protection Agency
Country United Kingdom 
Sector Charity/Non Profit 
PI Contribution During the NERC project, we have been directly involved in every deliverable and model developed. We brought our expertise in Bayesian statistic, sensors & structural health monitoring, intelligent infrastructure, safety & resilience and computational modelling.
Collaborator Contribution SEPA provided access to data on Scotland's rainfall and river flows, as well as to their sediment transport model in order to enhance the predictive capacity of constriction scour depth. SEPA has been involved in the Flood Hazard Model.
Impact Flood Hazard Model
Start Year 2017
 
Description Transport Scotland 
Organisation Transport Scotland
Country United Kingdom 
Sector Public 
PI Contribution During the NERC project, we have been directly involved in every deliverable and model developed. We furthermore designed and developed the pilot scour monitoring system installed on the A76 bridge on the River Nith. We brought our expertise in Bayesian statistic, sensors & structural health monitoring, intelligent infrastructure, safety & resilience and computational modelling.
Collaborator Contribution Transport Scotland provided access to bridge inspection and related data; in-kind contribution of staff time to provide technical advice and site-specific expert knowledge; management and maintenance of the pilot scour monitoring system currently being installed on the A76 bridge on the River Nith. Transport Scotland contributed in the development of Bridge Vulnerability Model, Scour Hazard Model and the Decision support system.
Impact Bridge Vulnerability Model, Scour Hazard Model, Decision support system and the Dataset of scour depth collected during system operation.
Start Year 2017
 
Description Structural Health Monitoring course 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Local
Primary Audience Postgraduate students
Results and Impact Last year, we launched a new class in Structural Health Monitoring for Civil Engineers at the University of Strathclyde. In the course, we teach the students how to analyse and interpret data by engaging them with the analysis of real-world case studies. This year we presented the project to the students by showing them the developed pilot monitoring system and some of the outcomes achieved. The attendance of the course is around 45 students.
Year(s) Of Engagement Activity 2018,2019