- Advanced Process Control
- Process Control
- Process Optimisation
- Operational Technology
- Mineral Processing
Advanced Process Control (APC) is one of the highest-value investments a mineral processing facility can make; delivering measurable improvements in throughput, recovery, and energy efficiency when implemented correctly. Yet according to McKinsey & Company, in some industrial companies fewer than 10 percent of installed APC systems are activated or properly optimised. Many more have never been reviewed since commissioning. The result is a significant and measurable gap between the potential value of APC and what sites are actually extracting from it, and in many cases, a system that is quietly sitting switched off.
This article examines Advanced Process Control failures in mineral processing with the four most common reasons APC systems underperform or stop running entirely, the warning signs that a system is degrading, and what a structured remediation approach looks like.aA
How widespread is Advanced Process Control failure in mineral processing?
The adoption gap in mineral processing remains significant. As recently as 2023, CIM Magazine observed that a significant majority of processing plants still rely on basic control strategies; a finding consistent with McKinsey & Company’s 2020 analysis showing that in some industrial companies, fewer than 10 percent of installed APC systems are activated or properly optimised. The problem is not a lack of available technology. It is the gap between what APC can deliver and what plants are actually extracting from it.
Of the sites that have invested in APC, a significant proportion are not running their systems at full capacity, or at all.
Four reasons Advanced Process Control systems fail
Model drift: when your APC no longer reflects your process
Most APC systems, particularly those using Model Predictive Control (MPC), are built on dynamic process models: mathematical representations of how a circuit responds to changes in its inputs. These models are calibrated at commissioning using historical process data and plant testing. At that point, they are accurate and the system performs as designed.
The problem is that mineral processing is not a static environment. Ore bodies change. Feed grades vary. Equipment wears over time. As these conditions shift, the gap between what the model predicts and what the plant actually does widens; a well-documented phenomenon known as Model Plant Mismatch. When this gap becomes significant, the APC makes control decisions based on an increasingly inaccurate picture of the process. Performance degrades. Operators notice the system behaving unexpectedly. And the APC gets switched off.
This is one of the most consistent patterns observed across processing sites. An APC installed on a flotation circuit or a SAG mill ends up disabled. The reason, when investigated, is almost always the same: the model has drifted from current operating conditions and there is no one on site with the expertise to retune it.
Instrumentation degradation: bad data leads to bad decisions
APC systems are only as good as the data they receive. A predictive controller fed inaccurate or unreliable data will make inaccurate control decisions, and those decisions will rapidly erode operator confidence in the system.
In many cases, instrumentation problems that existed before APC installation were not fully resolved or were addressed at the time and have since deteriorated. Field instruments go out of calibration. Sensors fail and are not replaced promptly. Signal quality degrades. The historian captures noise rather than meaningful process data.
The APC continues to operate, but its decisions become increasingly difficult to trust. Operators begin to override the system. Eventually, the path of least resistance is to switch it off and manage the process manually.
This failure mode is directly connected to a foundational problem: instrumentation readiness was not adequately established before the APC was installed or has not been maintained to the standard the APC requires to function correctly.
Operators don’t trust what they can’t see
Even a well-tuned APC running on accurate data will underperform if the operators responsible for the process do not understand or trust what it is doing. This is one of the most consistently underestimated challenges in APC deployment, and one of the most damaging.
When an APC operates as a black box, making control decisions that operators cannot see the reasoning behind, trust erodes quickly. Operators who cannot understand why a control action is being taken, will override it. If that override produces an improvement, they learn that the manual approach gets results. The APC gets bypassed more frequently. Eventually it stops being part of normal operations entirely.
The CIM Magazine has identified cultural scepticism among operations personnel as one of the biggest obstacles to APC deployment in mineral processing, noting that many metallurgists and process engineers arrive in operations without having received formal training on APC in their academic programmes. Without a foundational understanding of what APC does and why, operators are not equipped to engage with the system constructively.
Making APC decision logic visible, integrating it into the SCADA environment that operators work in daily, rather than running it as a separate system in a separate interface, is one of the most effective interventions for rebuilding operator trust and increasing the proportion of time the APC actually runs.
The expertise walked out the door
APC systems require ongoing technical ownership. The vendor or integrator who built the system understands it deeply at commissioning. If that knowledge is not transferred to site personnel, and if no internal capability is built to maintain, monitor, and retune the system, the APC is operating on borrowed time from day one.
When process conditions change, there is no one to update the model. When an instrument failure affects control performance, there is no one who understands the relationship between the faulty instrument and the APC’s behaviour well enough to diagnose it. When the project champion who drove the APC adoption moves to a different role, the institutional knowledge goes with them.
This is why a training pathway must be developed alongside the optimisation pathway, not as an afterthought. Documentation, structured e-learning, and operational procedures written in the language of the people running the facility are what sustain APC performance long after the implementation team has left the site.
Read more on APC readiness and what the audit process involves.
Warning signs your Advanced Process Control system is underperforming
The same identified four observable indicators of a suboptimised APC system – each assessable without specialist intervention. Each can be assessed without specialist intervention:
High variability in throughput and recovery despite APC being installed. If key process indicators are still oscillating significantly, the APC is not controlling to its potential, whether because it is running in a degraded state or because it is not running at all.
APC utilisation rate below 80 percent on controllers managing critical processes. A system that is in automatic mode less than 80 percent of the time is not delivering its potential value. Low utilisation typically reflects operator distrust, model degradation, or both.
No review of APC logic or setpoints. Control logic and constraint boundaries need to be reviewed as process conditions evolve. A system operating on parameters set at commissioning, with no review since, is almost certainly running on outdated assumptions.
No management system tracking instrumentation and sensor health. If there is no structured process for monitoring the health of the instruments feeding the APC, degradation will go undetected until it becomes a visible performance problem.
Beyond these, there are additional observable signals on the ground: controllers running permanently in manual mode, operators routinely overriding setpoints without logging the reason, and APC interfaces existing in a separate system from the SCADA environment operators use daily.
How to fix an underperforming Advanced Process Control system
The appropriate first step is an independent, structured assessment of the existing system. This is distinct from a pre-implementation APC Readiness Audit. It is a targeted diagnostic of why an installed system is not performing as expected.
This type of assessment considers model currency and the degree of Model Plant Mismatch; instrumentation health and signal quality across the critical measurement points; the reasons behind operator overrides and the frequency of manual intervention; knowledge availability on site for ongoing maintenance; and the level of integration between the APC and the broader control environment operators work within.
In some cases the path forward is relatively straightforward: the model requires retuning, specific instruments need attention, and operator interfaces need to be updated to make APC decision logic visible. In others, the foundational layers: instrumentation reliability, control loop performance, data infrastructure, need to be addressed before the APC can be meaningfully restarted.
Either way, the starting point is an accurate picture of where the system actually stands. An APC that is not running is not generating value. Understanding clearly why it is not running is the prerequisite to recovering that value.
Key takeaways
According to McKinsey & Company, in some industrial companies fewer than 10 percent of installed APC systems are activated or properly optimised. Sustained underperformance is an industry-wide pattern, not an isolated problem.
The four most common causes of APC failure are model drift from changing ore and process conditions, instrumentation degradation, operator disengagement due to poor system visibility, and loss of technical expertise after project handover.
Observable warning signs include high process variability despite APC being installed, utilisation rates below 80 percent, no regular review of control logic, and no management system tracking instrumentation health.
Making APC decision logic visible within the operator’s normal SCADA environment is one of the most effective interventions for rebuilding trust and increasing system uptime.
An independent diagnostic of a failing or underperforming APC – conducted on site, not just from a desk – is the appropriate starting point. The causes of underperformance cannot be reliably diagnosed from historian data alone.
Want to find out why your APC isn't delivering?
If your site has an APC system that is underperforming or has been switched off, Mipac has been engaged across multiple processing sites to diagnose and remediate failing APC deployments.
Or contact our team to arrange a discovery calls →
Interested in learning more about APC?
APC Readiness in Mineral Processing: What actually stops it from working
APC readiness in mineral processing goes well beyond selecting the right software. Mipac's Optimisation Team…
How to get your plant APC ready: a proven Readiness Audit framework
Is your plant APC ready? Advanced Process Control can boost throughput and stability - but…
9 questions to ask yourself before you start your APC journey
As a Process Manager, you know that advanced process control (APC) has the potential to…