Lorenzo Bortoloni

Task 4.4 Reduced Order Models Layers for digital twins

Partners involved

In this Task, ITAINNOVA has participated as Task leader adapting and developing the Twinkle library. Besides, SCHNEIDER and LIBRA have participated integrating the library in the final IoT platform. Finally, SUBTERRA has supported providing input to build the geotechnical Digital Twins.

The end-users involved in this task are the ones that will have a Real-Time Digital Twin at the end of the project, i.e. the end-users: MARINI, TAPO and TITANIA.

Objectives and outcomes

The main objective of this Task is to adapt the existing library Twinkle (developed by ITAINNOVA) to build the real-time digital twins. The library will be adapted for integration in the Dig_IT IoT platform as a layer to evaluate the resulting real-time digital twins, enabling the visualization of real-time risk maps in the DSS. These realtime digital twins will be the base for process control and quality assessment during operation.

This task plays a very important role within WP4, as the Reduced Order Models of geotechnical and fluid dynamic simulations developed in tasks 4.2 and 4.3 are built based on the library developed in this task.

What has been done in Task 4.4?

In this Task, the existing Twinkle library was adapted to generate real-time digital twins for geotechnical and fluid dynamics simulations. The resulting models stand for virtual sensors that can predict in real time the geotechnical or fluid-dynamics risks, mainly related to safety (slope stability, pollutants concentration in air and water). An exhaustive and automatized pre and post processing workflow was developed and implemented in the final tool according to each end-user requirements and needs. A user manual was generated in order to ensure the understanding regarding the use of this tool.

Finally, the exploitation of the generated ROMs is highly improved using a graphical user interface (GUI), which allows for fast evaluations using interactive sliders, real time tendencies evaluation with graphs, plotting of contour predictions etc. ITAINNOVA has a web GUI created in VOILA, which has been successfully adapted by LIBRA for platform implementation.

The deliverables related to this task are D4.4 ‘Python library and documentation for building digital twins from engineering simulation tools (v1.0)’, which was already submitted in M21, and D4.5 ‘Python library and documentation for building digital twins from engineering simulation tools (final)’, which is a final version of the previous deliverable and will be submitted at the end of the project.

Software and tools used

For this task, Python scripting was used in order to adapt and implement the Twinkle library into the IoT platform, to develop the post-processing capabilities according to each use-case, and to adapt the GUI.

GUI to display the real-time digital-twin for air quality and ventilation assessment in Kemi mine.
Image of the Kemi operaiting area under study, indicating the plane where the digital twin is being evaluated.
GUI to display the real-time digital-twin for geotechnical stability assessment of the tailing pond of La Parrilla mine.
Image of the La Parrilla area under study, indicating the plane where the digital twin is being evaluated.

Task 3.4 Big Data Optimisation and Analysis

Partners involved

The partners involved in Task 3.4 are BRUNEL, ICCS, LIBRA, and SUBTERRA for Task 3.9. Moreover, TITANIA has also been involved as end user, along with LA PARRILLA for Task 3.9.

Objectives and outcomes

The main goal of Task 3.4 is to develop a human assistance tool for early warning hazard prediction in a sustainable digital mine of the future. The tool is based on data fusion and Long Short-Term Memory Recurrent Neural Network (LSTM-RNN), which stems from the big data optimisation concept which involves complex applications with elements such as predictive models, statistical algorithms, and what-if analysis in a mine operation practice.

 

A Human Decision Tool within the Dig_IT Decision Support System (DSS) is developed to alert the mining operator if any slope instability occurs within the open pit mine. The tool will be able to predict at least 16 minutes before the real occurrence to give some time for the mining operator to make strategic operation decision-making.

The outcome of this task will incorporate real-time data exchange between Brunel (Data Model) and ICCS (Data Warehouse). Then, the data will be translated to the Decision Support System (LIBRA) for displaying it in the interactive Human Decision Tool.  

The Human Decision Tool, which is part of the Dig_IT IoT Middleware layer, will: 
 
  • Construct a residual signal and compare it with a predefined threshold in generating alarms in the presence of bad data resulting from limited communication capacity. 
  • Perform fault detection and fault-tolerant control (FTC) under communication constraints feature/event detection, classification, and prediction in large-scale data environments in supervised and unsupervised modes.
  • Perform efficient evolutionary methods in preprocessing to seamlessly incorporate data denoising and data fusion across the entire early warning system, thereby augmenting the system’s overall robustness, observability, and accuracy.
  • Perform feature engineering techniques to understand the nature of the data and use short time series analysis to establish a model for detecting, identifying and predicting the events/failures.
  • Develop an early warning event prediction model that combines statistical time series modelling, relevant domain knowledge and intelligent search technique.
  • Apply the hazard prediction model using the real-time Dig_IT Decision Support System dashboard.
 

Software and tools used

The Fault Tolerant Control software package is built on Python 3 with standard libraries such as NumPy, Pandas, Matplotlib, Seaborn, Keras, and Sklearn. A decent machine with a good CPU/GPU with the installed Python libraries can be used to train the model. The model output from the prediction model will be updated in the Decision Support System (DSS) via the MQTT protocol that has been established between the CPU and the Dig_IT Platform.

Task 4.3 Fluid Dynamics digital twins design and development: Risk Maps

Participants

The technical development has been carried out entirely by ITAINNOVA.

The end-users have supported ITAINNOVA in the definition of the activities according to their interests and needs; MARINI and TAPOJARVI regarding the air quality models, and TITANIA regarding the water quality model. 

Objective and Outcomes

The purpose of this task is to build fluid dynamic RT-DTs for air and water quality. To that end, a CFD methodology is developed and validated with experimental data. Finally, this methodology is used to simulate a large number of different scenarios (a  design of experiments, DoE). The CFD results are then postprocessed with the library developed in T4.4 to build the virtual sensors of each mine (real time digital twins). Those virtual sensors will be integrated in the DecisionSupport System (DSS) of the Dig-IT IoT platform to allow the end-user to see in real time the risk maps on the air/water quality.

The main goal of this WP is to develop RT-DTs of engineering assets, geotechnical and fluid dynamic processes, so this task contributes directly to the achievement of this goal.

Task 5.6 will directly use the outcomes of this task, as the final ROM for each use-case will be integrated and displayed in DSS of the final IoT platform.

What has been done in Task 4.3?

On the one hand, regarding the air quality models from MARINI and KEMI, a CFD methodology has been developed and validated with experimental measurements performed in the real use-cases. In this way, a specific solver for each use case has been developed using the open-source CFD software OpenFOAM, including the appropriate modifications in the code according to the requirements of each mine, regarding heat transfer and buoyancy phenomena, pollutants transport, etc. After that, a Design of Experiments has been designed and run, covering the range of the input variables according to the end-users requirements. Finally, the results of all these simulations have been used in order to build the virtual sensor (ROM), which will be finally implemented in the IoT platform. 

On the other hand, regarding the water quality analysis from TITANIA, a data-based model has been developed in order to predict in real time the relevant pollutants (suspended solids and nickel) released to the environment. For such task, historical data was used in order to generate and train the prediction model generated through Python scripting. Likewise, the resulting model will be embedded in the DSS of the IoT platform to provide real time information on the prediction of the pollutants in the mine water streams. 

For this task, open-source code OpenFOAM was used in order to perform the CFD simulations. Besides, Python scripting was used to automatize the pre and post processing activities of the workflow, and to develop the data-based model from the TITANIA use-case. The library developed in T4.4 has been used to generate the ROMs.
CO2 concentration prediction in the operating area of Kemi mine (CFD results).

Ventilation flow path in the operating area of Kemi mine  (CFD results), when the rocks loading is starting and the stope is blocked by the rocks.

Task 4.2 Geo-spatial attributes digital twins design and development

Partners involved

SUBTERRA has participated and lead T4.2 along with MARINI MARMITITANIA and LA PARRILLA, which have been involved as end users.

Objectives and outcomes

The main goal of Task 4.2 is to develop geotechnical digital twins in the end users locations. Task 4.3 will use the outcomes of this task, by using outputs of La Parrilla numerical model, to develop the ROMs for the digital twin.

What has been done in Task 4.2?

In this task, three digital twins have been developed: La Parrilla, Titania and Marini Marmi.

The calculation model of La Parrilla has been finalised, using the FLAC 3D software. For this model, the geometries obtained from the drone flights have been taken into account, as well as the data from the installed sensors described in the previous WP2.

In a first step, the actual 3D model of the tailings pond has been represented, using the photogrammetric restitution of the images obtained by the UAS technique. This first approximation serves to compare the geometries in the different flights performed, as well as to detect the appearance of cracks or other elements that could compromise the integrity of the structure.

The first step in the development of the numerical model is the elaboration of the geometry. The meshing of the elements is a complex process, but it requires maximum detail so that the subsequent calculations are not affected due to an incorrect selection of the elements that make up the tailings pond. In the elaboration of the geometry of the model, all the constructive elements present have been taken into account, as well as the materials used for the waterproofing of the ground to avoid the infiltration of percolating water into the subsoil by means of the application of a geomembrane.

The results shown in Figure 1 correspond to the current situation of the tailings pond, where a groundwater level is located below the dam, and a non-evolving pore pressure due to the cessation of activity at the La Parrilla mine.

Figure 1. Numerical model results for current scenario in La Parrilla Tailing Storage Facility

The numerical model for Titania follows a similar methodology to that presented for the La Parrilla model, although in this case the behaviour of a slope in the open pit is simulated. The photogrammetric restoration of the different flights has been carried out and the geometries have been compared. These data have been taken into account for the geometry. Figure 2 shows the geometry obtained in the most recent UAS flight (November 2022). Figure 6 shows the area to be calculated, where the greatest instabilities occur. The geometry that FLAC 3D will use for the calculations has been obtained.

Figure 2. 3D model generated using UAS images performed in the last visit to Titania.

The Marini Marmi model represents the update on the progress of the underground works. At Subterra, this progress is received periodically from Marini Marmi Team, and is represented in a 3D model created with Leapfrog. In this model, the progress of the underground galleries can be observed, as well as their crossing through complex zones such as faults, or low resistance materials such as clays. Figure 3 shows an image of the 3D model representing the excavated galleries. The sections that are not represented in 3D show the progress planned but not executed yet.

Figure 3. 3D model of Marini underground galleries excavated.

10th Technical Meeting

On the 7-10th of November, the 10th Technical Meeting of Dig_IT took place at the Titania AS Tellnes Mine in Egersund, Norway. The partners presented the progress made during the period, and discussed any challenges and issues that had arisen in their tasks.

During the meeting, the participants had the opportunity to visit the Tellnes Mine, where they learned about the management and production processes of the mine, as well as the measures taken to ensure safety and sustainability. They also discussed the installation of sensors in the mine to improve the monitoring and control of the production process.

Overall, the meeting was a great success, and the partners left with a deeper understanding of the challenges and opportunities facing the Dig_IT project. They also had the opportunity to network and build relationships with other professionals in the industry.

 

9th Technical Meeting

The 9th Technical Meeting of Dig_IT was conducted virtually on September 27th. Each Work Package (WP) leader presented the progress made during the previous period. The participating partners provided comments and raised questions for each WP.

Furthermore, they reviewed a list of unresolved issues, analyzed the potential risks for upcoming tasks, and deliberated on potential management strategies and solutions.