Smarter maintenance, stronger performance: the first year of Aspen Mtell® at Ok Tedi
Provision of Asset Management System
- Papua New Guinea
- Aspen Mtell®
- Mining
Background
In 2023, Ok Tedi Mining Limited (OTML) selected Aspen Mtell® as the preferred solution to strengthen asset reliability across its copper and gold operations in Papua New Guinea. Mipac was engaged to lead the implementation, following a detailed evaluation of available systems and a long-standing relationship with OTML through major automation and improvement projects.
The initial deployment focused on Mtell’s predictive analytics capabilities — using machine learning to detect early signs of degradation leading to equipment failure and provide timely, actionable insights for maintenance teams.
Image of Mtell System Manager – configuration screen for connecting PI historian data to the predictive maintenance system
Scope of the initial rollout
The first phase of implementation began in May 2024 and focused on ten critical assets in the crushing and grinding circuit. These assets were selected for their impact on production and history of unplanned maintenance.
The delivery was split into two parts:
- AspenTech, supported by Mipac, implemented the first five assets
- Mipac led the deployment of the remaining five
The project included:
- Installation of Mtell® software
- Integration with sensor time-series and maintenance data provided by OTML
- Development and tuning of Mtell’s intelligent agents for each asset
- Ongoing support and review of system alerts in collaboration with site personnel
Agent probability trend – when training the Mtell agent, the green trend represents the probability of failure at any given time, by default it alerts when above 50% or probability of 0.5
Results after 12 months
After one year of operation, Mtell® has delivered significant benefits to Ok Tedi’s maintenance and operations teams and OTML is looking to renew for another three years. These benefits include:
Earlier identification of issues
The system successfully identified signs of equipment degradation ahead of failure, allowing maintenance teams to intervene before downtime occurred.
Improved reliability of high-impact assets
Several incidents were avoided through early alerts, reducing the risk of production losses in critical circuits.
Reduced unplanned maintenance
With better visibility into equipment condition, site teams were able to plan interventions more effectively and avoid reactive repairs.
Enhanced decision-making
The system’s insights supported more focused discussions between operations and maintenance teams, helping to prioritise work and allocate resources more efficiently.
Mipac worked closely with site personnel to review alerts, refine system performance, and embed the use of Mtell® into day-to-day workflows.
Excel spreadsheet for reviewing agents and alerts on assets over time
Mtell® Asset Rollout
In June 2025, OTML issued a new purchase order for the rollout of Mtell® across an additional 38 assets, expanding the system into the flotation plant and power generation areas. This broader deployment reflects OTML’s confidence in Mtell’s performance and the value it delivers to site.
Mipac’s new scope includes:
1. Project Initiation
- Kick-off meeting to align teams, schedule, and expectations
- Workshop to identify gaps in instrumentation or data availability
2. Design Phase
- Review of OTML’s historical work orders, sensor data, and Enterprise Asset Management (EAM) hierarchy
- Data quality validation and mapping sensor tags to relevant assets
- Importing alarm limits and failure history into Aspen Mtell®
- Grouping similar equipment for streamlined monitoring (e.g. pump trains)
- Use of P&IDs and process flow diagrams to define logical groupings
- Identification of ON/OFF conditions and asset operating states
3. Build Phase
- Configuring templates and agent types:
- Hidden failure agents
- Anomaly detection agents
- Rule-based and failure library agents
- Iterative testing and validation with OTML subject matter experts
4. Deploy & Go Live
- Real-time deployment of agents to monitor live data streams
- Agents continuously learning and retraining to alert operators to emerging failure patterns
- Final project close-out with issue tracking and system support
Mipac will continue to provide support and optimisation services for the next 12 months to ensure the system remains accurate, relevant and embedded in operational practices.
System architecture diagram of the Mtell system and the historian data source at Ok Tedi
Outcomes
While project outcomes will continue to develop over time, the Mtell® deployment has positioned OTML to:
- Detect asset degradation well before failure beyond the capabilities of existing OTML systems
- Automate insights into abnormal operating behaviour
- Improve condition-based maintenance scheduling
- Build a digital foundation for scaling predictive maintenance across the site
Alert manager – an image of what a user would see on the Mtell Alert web page, listing a specific agent and tags that agent monitors and probability of failure
Mipac's role
Mipac acted as the integration and deployment partner, bridging the gap between OTML’s data sources and Aspen Mtell’s machine learning models. This included deep data preparation, stakeholder workshops and hands-on engineering expertise to ensure a robust and effective outcome tailored to the site’s operational context and history.
In summary
The first year of Mtell® implementation at Ok Tedi has demonstrated the potential of predictive analytics to improve equipment reliability, reduce unplanned downtime and support more proactive maintenance strategies.
For OTML, the expansion of Mtell® represents a shift toward long-term, data-driven asset management. For Mipac, it reinforces the importance of combining the right technology with deep site knowledge and ongoing support to deliver sustainable outcomes.
The AspenTech Mtell® system
Aspen Mtell® software is a powerful equipment analytics solution that accurately predicts an anomaly in monitored equipment and its related process, indicating precisely when and how an event is likely to occur. This solution leverages machine learning to identify and alert on early patterns of degradation which lead to reduction in useful life. The resulting output provides insights that can assist in the containment of risk, supporting a strategic transformation program, which can drive net results with an increase in production output and a reduction in expenses relating to equipment ownership, maintenance costs and overall energy consumption.
The Mtell® software executes very precise pattern recognition of changes in streaming signals produced from sensors on and around the equipment being monitored. This permits detection of extremely early onset of equipment degradation through multi-variate differences across all the streaming signals and through temporal distinctions: tiny changes in the streaming signals, offset by time.
Visit the AspenTech website for more information.
Want to know more about how Aspen Mtell® can optimise your complex mineral processing operation?
Aspen Mtell goes beyond traditional monitoring by becoming a proactive guardian of your assets. It deploys intelligent “agents” that continuously monitor and analyse sensor data from your critical equipment. These agents leverage various methodologies to predict potential failures before they occur. Read more about AspenTech Mtell.
