From Ad-Hoc to Routine: Rethinking Production Surveillance and Field Optimisation

Production surveillance and optimisation are fundamental to every operating oilfield. They are the mechanisms by which engineers maintain field performance, protect well reliability, and identify where to optimise production. Yet in many operations, these activities struggle to keep pace with the realities of modern fields.

As asset bases grow, systems become more interconnected, and operational complexity increases. Surveillance and optimisation can quietly shift from proactive engineering disciplines into reactive tasks, often overshadowed by day-to-day operational issues.

This is not a failure of engineering capability. It is a consequence of workflows that have not evolved at the same pace as the systems they are meant to manage.

The Reality of Production Surveillance Today

In many fields, production surveillance still relies heavily on manual review of well and network trends. Engineers scan time-series data, respond to alerts triggered by predefined thresholds, and use experience to decide what looks abnormal or worth investigating.

This approach works when asset counts are low and conditions change slowly. However, as fields mature and interactions between wells, gathering networks, and facilities increase, the workload grows quickly.

Engineers are required to monitor more wells, more variables, and more interdependencies - yet the time available to do so remains fixed. As a result, surveillance becomes selective. Attention is focused on known problem wells, recent events, or the loudest alarms, while more subtle changes can go unnoticed.

Over time, surveillance shifts from systematic to reactive.

Why Optimisation Becomes Hard to Sustain

When surveillance is time-intensive, optimisation is pushed down the priority list. Setpoint changes are made well-by-well, based on local observations, and broader system interactions are considered only when problems arise.

Many optimisation workflows are also built on static rules or assumptions that take time to adjust as conditions change. Updating these rules requires manual effort, validation, and confidence that changes will not introduce risk.

The result is that optimisation tends to lag behind reality. Engineers know there is more value to be unlocked, but the effort required to find and assess opportunities limits how often optimisation can be performed.

The Structural Limitation of Traditional Approaches

Traditional surveillance and optimisation approaches place the burden of interpretation on the engineer. Every trend, alert, and event must be mentally evaluated, often across multiple tools.

This does not scale well. As the system grows, so does the cognitive load.

The key limitation is not data availability; it's the lack of structure in how data is translated into engineering insight.

Compare the difference between traditional and modern approaches to field optimisation here

A Different Way to Think About Surveillance

Modern approaches start by rethinking what surveillance is meant to achieve.

Rather than asking engineers to review raw trends, the goal becomes ensuring that every meaningful event or change in system behaviour is interpreted consistently and systematically.

This is where dynamic, adaptive algorithms play a role. Instead of relying on fixed thresholds, these algorithms learn how wells and networks normally behave and continuously interpret new data in that context.

As conditions change, the interpretation adapts. This allows the system to distinguish between noise and meaningful changes in behaviour, across many wells, without requiring manual tuning.

Importantly, this does not remove engineering judgement. It standardises the first step of interpretation, ensuring that nothing important is missed.

Shifting Optimisation from Ad Hoc to Routine

When surveillance becomes systematic, optimisation can become routine.

Rather than being triggered by individual investigations, optimisation analysis can be scheduled to run regularly, using current field conditions. In this model, engineers are not searching for opportunities - they are reviewing them.

Setpoint optimisation is evaluated at a system level, considering inflow, lift, network behaviour, and facility constraints together.

Opportunities are assessed in context and prioritised based on impact and feasibility.

And engineers are able to apply their expertise where it matters most: evaluating trade-offs, validating recommendations, and deciding when and how to act.

What This Means for Production Engineers

The practical impact on the engineer’s day is significant.

Instead of spending hours scanning trends and responding to alerts, engineers begin with a prioritised view of the field. They see where behaviour has changed, where optimisation opportunities exist, and how these relate to nearby wells and network conditions.

Time is no longer spent searching for what has changed. It is spent deciding what to do - allowing engineers to manage larger asset bases more effectively, without sacrificing rigour or control.

Why This Matters in Operational Reality

Production surveillance and optimisation are easily deprioritised when operations are busy. Wells trip, facilities go offline, and urgent issues demand attention.

Without structured, repeatable workflows, optimisation becomes something teams “get to when they can.” Over time, this leads to lost production and missed opportunities that are difficult to quantify.

By embedding surveillance and optimisation into automated, scheduled workflows, these activities become part of normal operations rather than optional tasks.

Transforming the Discipline, Not Replacing It

The shift described here is not about replacing engineers or automating decisions prematurely. It is about changing how engineering effort is applied.

By using adaptive algorithms to interpret system behaviour and present prioritised opportunities, engineers are supported, not sidelined. Their judgement remains central, but it is applied more consistently and at greater scale.

A Foundation for the Future

As oilfields become more complex and operational margins tighten, the need for scalable, systematic optimisation will only increase.

Rethinking surveillance and optimisation workflows is not about adopting new tools for their own sake. It is about ensuring that production engineers can continue to do what they do best - understand systems, manage risk, and improve performance - without being overwhelmed by data and manual effort.

Production surveillance and optimisation will always be core requirements. The opportunity now is to modernise how they are delivered.

---

At SIG Machine Learning, we help production teams start the day with a prioritised set of field-level events and optimisation opportunities - so engineers spend less time scanning trends and more time utilising their expertise to make decisions that improve reliability and performance. Contact us to find out more.