Mines have access to more information about the performance of their tailings storage facilities (TSFs) than ever before. But as the number of monitoring instruments and volume of data increase, the challenge is shifting from collecting information to knowing what to do with it: what matters, how to act on it and how to demonstrate what happened afterwards.
Insight Terra CEO Alastair Bovim says the company has seen some sites move from 10 monitoring instruments five years ago to thousands of instruments generating millions of data points a day.
“More data should give operators greater visibility of risk, but it also creates complexity,” explains Bovim. “On a single site, you may have multiple different technologies or vendors in use. If that information remains siloed, the people responsible for the facility still have to piece together what is happening and decide what requires action.”
For Rohit Prabhu, Senior Virtual Twin Specialist at Dassault Systèmes’ GEOVIA, the challenge is also about maintaining continuity between a monitoring event, the response and its eventual resolution. “The full history should remain available against the instrument, including participants, actions, timestamps, comments, approvals and attachments,” says Prabhu. “This is important for audit and historical review.”
This was one of the challenges explored during From Data to Accountability: The Next Era of Tailings Performance Management, a webinar hosted by Dassault Systèmes’ GEOVIA and environmental risk intelligence company Insight Terra on 21 July.
The discussion examined how the tailings monitoring challenge is evolving as mines deal with growing volumes of geotechnical, environmental and operational data, alongside increasing expectations around continuous performance monitoring, governance and evidence-based conformance with the Global Industry Standard on Tailings Management (GISTM).
From collecting data to understanding what it means
Modern monitoring systems can provide near real-time information on the conditions and performance of a tailings facility. Data may come from different sensor types, vendors and systems. An unusual reading may indicate a change in the facility, but it could also point to a problem with the instrument itself, a communications failure or another data-quality issue.
This makes the quality and context of the underlying data increasingly important. Before information can reliably support engineering decisions, automated workflows or AI-assisted analysis, it needs to be brought together, checked and verified.
To address this, Insight Terra and GEOVIA have brought together their respective monitoring and engineering capabilities in a joint tailings solution. Insight Terra ingests data from multiple sources to create a trusted data foundation for monitoring and analysis. This includes helping to identify gaps or anomalies in the data and distinguish operational issues, such as a sensor that has stopped transmitting, from changes that may require closer engineering attention.
“GISTM demands continuous performance monitoring and evidence-based conformance,” says Bovim. “For a tailings owner, data gaps are simply no longer an option.”
The Insight Terra platform creates an immutable record of the underlying monitoring data, helping to preserve a transparent and auditable history of the information on which subsequent decisions and actions are based.
What happens after a warning?
If monitoring identifies a warning or anomaly, the important questions quickly become practical: Who has seen it? Who is responsible for investigating it? What action is required? Has that action been completed, and has the issue been resolved?
This creates a clearer chain of accountability, and a defensible record of how an issue was handled. It also reduces the reliance on information spread across separate spreadsheets, emails, reports and systems.
Connecting performance monitoring with GISTM evidence
Continuous performance monitoring and conformance management are closely connected, but the information needed to demonstrate compliance may sit across different teams and systems. Bringing the relevant requirements, evidence, actions and approvals together can give operators a clearer view of where they stand.
“At enterprise level, executives can navigate from a geospatial view to a particular mine site or TSF and review its GISTM conformance status,” he says. “Requirements can be filtered and interrogated to understand why a protocol conformance status is met, partially met or not met, and who owns the requirement.”
This does not automate compliance or remove the need for engineering judgement. It can, however, make it easier to connect monitoring information with responsibilities, supporting evidence and the actions required to address gaps.
When design intent meets real-world performance
Live monitoring data can also be integrated with the engineering context of the facility. “GEOVIA models how the facility should behave, based on the design intent, and Insight Terra monitors how it actually behaves,” says Bovim. “Together, that creates a living virtual twin.”
This allows engineers to view real-world monitoring information in the context of the facility’s design rather than treating sensor readings as isolated data points. The virtual twin can also be used to model future performance, helping engineers explore how the facility may behave under changing conditions.
Tailings facilities do not remain static, and neither do their monitoring requirements or Trigger Action Response Plans (TARPs). The ability to compare observed behaviour with engineering expectations helps teams identify where closer investigation is required.
AI needs reliable information to work with
AI adds another layer to this increasingly complex environment, but its usefulness depends heavily on the information and knowledge it can access.
GEOVIA’s AI-powered Virtual Companion can interrogate an integrated knowledge base, retrieve information and identify its source, summarise risks, and help users plan and track actions. For example, a user could ask what action should be taken when an instrument was not responding. The AI Virtual Companion draws on information held in the knowledge base to identify the relevant response plan and assist in creating and assigning the tasks required to address the issue.
The value, says Bovim, is not in replacing the engineer. It is in reducing the time experts spend finding and organising information so that they can focus on interpreting it and making decisions.
Ultimately, what matters is whether operators can trust the data they are receiving, understand what it is telling them, act when something changes and maintain a clear record of what happened.
“More data does not make a tailings facility safer on its own,” says Bovim. “The value comes from turning that data into information people can trust and act on, and being able to trace the decisions and actions that follow.”
