PIANC Smart Rivers 2022

Reviewing Presentation documents

Topic:

Title:

Author(s):

Author 1

Name：

Affiliations：

Region：

Email：

Is corresponding author or not (one only)：

Author 2

Name：

Affiliations：

Region：

Email：

Is corresponding author or not (one only)：

Author 3

Name：

Affiliations：

Region：

Email：

Is corresponding author or not (one only)：

Author 4

Name：

Affiliations：

Region：

Email：

Is corresponding author or not (one only)：

Abstract : *(250~1000 words)

Maintaining hydraulic structures such as dams, penstocks, or water lock gates in operating conditions and optimizing their maintenance costs are key issues for energy production or river navigation. The ultimate objective is to know the real state of fatigue and damage of the structure and identify any related anomalies. Today, there is no standard tools for fatigue evaluation merging real data obtained with an embedded sensor network and a numerical model that converts in real time measured data into fatigue.

After 3 years of R&D collaboration between CNR and Morphosense in the maintenance of navigation lock gates or dam gates, this presentation exposes how our disruptive Live Digital Twin solution contributes to fatigue evaluation and more generally to global structural monitoring in dealing with fundamental issues of hydraulic structures: risk assessment, maintenance in operating conditions and maintenance costs optimization.

The proposed solution is based onto two parts.
The first one is a single cable hard-wired network of sensor nodes (synchronized tri-axis accelerometers, gyros, magnetometers, and a temp sensor) collecting tilts and vibrations at a spatial density adapted to the structure. An important feature is the interoperability of this sensor network: thanks to a specific interoperable sensor node, external sensors such as strain gauges, analog or digital sensors (weather station, wave radar, …) can be connected. API connectors allows connection to existing SCADA. Depending on the structure, the nodes can be welded, fixed with magnets, or strapped on structures.
The second one is a software solution for the Live Digital Twin design and processing. The Live Digital Twin is a digital representation of the real structure: measurements collected with the sensor network are used to continuously update the virtual representation of the structure in real time and a precise, up-to-date copy of certain properties and states of the structure, such as its shape, position, state and movements are provided. 
The proposed Digital Twin is a combination of three main elements.
First, a 3D simulation model with a specific tradeoff between complexity and simulation duration. This model needs to be sufficiently physics based to manage meaningful parameters, sufficiently accurate to manage application useful parameters and sufficiently quick to run in comparison with the observed phenomenon.  
Then an advanced strategy able to assimilate mass data, and finally a mechanism to adapt the model to the measurements by fitting parameters. The implementation of our Digital Twin is divided into three parts:  at first a calibration between the model and the measurement, then a sensitivity analysis to categorize and to correlate the parameters influencing the model with measurements, and finally a model order reduction (R.O.M.) 
The digital twin is then fed with load sources measurements (such as water level and temperature). To maintain data-model convergence, structural measurements acquired with the sensors network (such as static deformation or modal frequencies) are compared in real time with simulated outputs from the Digital Twin. Depending on the convergence, the model is dynamically updated, and virtual sensors are processed such as stress estimation and that regardless sensors positions. The Digital Twin then provides information in real-time relative to fatigue in the structure or any maintenance operations by targeting the pathologies measured and formalized via the model.

In the first part of the presentation, the historical context and the methodology applied to monitor the lock-gate using the sensor network and water level measurements will be introduced.
A relevant indicator combining the maximum deformation of the lock-gate and the water level will be presented in detail. The second part of the presentation will focus on the Live Digital Twin building, combining the data from the previous indicator, the 3D numerical model of the lock-gate and the continuous data flow of the sensor network. Operational implementation of the Digital Twin for 2 different lock gates will be detailed. The estimation of the fatigue and the service lifetime of the lock-gates with the Digital Twin is emphasized and the use for the maintenance plan optimization of the owner is explained.

As a conclusion, the approach will be generalized to other hydraulic structures or offshore wind turbines.

Key words: (max: 5)

Reference no.

Slides：

click to download