Start small, think big with autonomous operations
August 10, 2020
If you think autonomous operations just means driverless trucks and unmanned drills—think again.
Yes, enabling equipment and machinery to move about performing tasks with no human intervention is part of it. But the potential of autonomous operations is so much more. And contrary to popular belief, you do NOT have to invest in a ‘big bang’ project to leverage autonomous operations. There are a number of ways to adopt autonomous operations in smaller, focused ways and still realize tremendous value. Instead of equipment, think about autonomous processes.
Focus on high risk, high impact
The guidance we offer our clients is to start by focusing on process areas of either high risk or high impact on business value. Look at those areas within the operation where you frequently see a delay in decision-making or where transferring information from one area to another is slow and holds up subsequent processes.
For example, consider a common risk area like tailings storage. You can automate the monitoring and management process for your tailings storage facility, using machine learning models to continuously track the risk based on data collected from the site and satellite data. If an increase in risk is detected, the recommended action can be automatically generated and an alert sent to the supervisor to execute the action. Depending on the level of risk, even the response action itself could be automated, such as sending out a drone to conduct a detailed survey of the risk area.
Short-term planning is another example where you can quickly gain value from automation. Companies tend to plan a shift schedule with a weekly—or at best 24-hour—time horizon. However, as soon as the shift begins the plan is often no longer entirely actionable. Perhaps a piece of equipment is unavailable, or a shovel isn’t where it’s needed. Whatever the issue, you’d be working off a shift plan that is no longer up to date or detailed.
Instead, suppose you collect near-real-time data from the site, and automate how that data is shared in order to accelerate decision making, adjust the plan and avoid operational delays. We worked with one client recently to assist them to adjust their plans dynamically within a given shift facilitating short interval control. It helps our client check the plan as it’s being executed and, if it’s not on target, adjust the plan to define the preferred alternative actions.
There are many areas where miners can automate, in addition to site excavation and production. For example, with another client we are developing machine learning algorithms to automate geological modelling based on drill hole logs, assay and survey data. Following traditional methods, geologists can spend up to six months building a model. Machine learning has the potential to reduce that time to a matter of days.
<<< Start >>>
Contrary to popular belief, you do NOT have to invest in a ‘big bang’ project to leverage autonomous operations.
<<< End >>>
Autonomous operations: Not all or nothing
It’s important to remember that there are degrees of autonomous operations, and that you don’t need end-to-end automation across the entire operation to realize value. Start by understanding your process framework—where automating a process could reduce delay or downtime, improve and accelerate access to insights, decision-making, lower risk, etc.
Once the priority processes are identified, it is necessary to consider the data requirements and future workflow for autonomous process and decision making-autonomy. To create an intelligent data foundation, a data model and fit for purpose technology architecture is recommended, allowing for legacy and new implementations to facilitate future proofing as more processes become integrated.
Ensuring solid data foundations (governance, data quality and architecture) can provide the platform for automating future high-value processes and the ability to scale across your enterprise, accelerating your speed to value and increasing returns.
Then you can apply diagnostic and predictive analytics and scenario modeling, leveraging machine learning, artificial intelligence (AI) and robotic process automation, to build autonomous maturity across the operation. These technologies allow you to gain reliable situational awareness, the ability to model options for what actions result in the best outcomes and automate transfer of data, work orders and processes—all based on access to data.
Autonomous operations don’t have to be all or nothing. It can start with one process, then another, and so on—with each step adding value along the way to support your broader performance goals.
