As, the meteoric rise of artificial intelligence and machine learning continues, it becomes difficult to stop, take an analogue breath, and recognise what we've achieved. Back in 2022, we predicted "the future is closer and smarter than you might think". Now, firmly within that digital future, we remain at the cutting edge of predictive maintenance with more than one eye on things to come.
What is predictive maintenance and how has Wood deployed it?
Predictive maintenance is the proactive strategy of using data analysis, sensors and machine learning to predict when equipment is likely to fail or require servicing before the failure actually occurs. Wood's systems can send alerts to clients that their assets have been running at hazardous operating levels, increasing the risk of failure. This enables the operators to make necessary changes.
To date Wood has employed artificial intelligence and machine learning in several projects, including:
Pipeline corrosion
Our client had a large network of around 1,200 pipelines of various ages. They needed to assess the risk of a leak in their pipelines in order to prioritise intelligent pigging. Leak information from around 200 of these pipelines that had failed previously was investigated and mapped together to build a machine learning model. This examined data patterns to discern what caused the failures. Once the model was built, it was deployed to the other thousand pipelines. This enabled our client to prioritise the maintenance activities required.
Dew point breach
A gas plant was tripping due to a dew point breach. Initially, we tried to model the production fluid that was coming into the plant by using traditional process models to predict when the breach was occurring and if we could identify a cause. No matter what we did, nor how much we adjusted the physics-based model, we were unable to get a good match. So, we deployed machine learning algorithms which were able to understand what was causing the failure. With these models we were also able to interrogate the model to understand which part or variable in the data had the biggest impact in the prediction, and this is where we could combine our domain knowledge with machine learning models to get to the root cause and extract more value.
Gas-lift optimisation
A production system consists of multiple wells and pipelines feeding into a central facility. When a well lacks enough pressure to lift the fluid column, gas lift is used to assist. Determining the optimal gas allocation for each well pad to maximise production with limited gas is a complex real-time challenge. Existing simulation models provide useful guidance, but they require significant computation, which is often impractical. By combining deep reinforcement learning with simulation models, we created an AI solution that delivered advisory or closed-loop control with a minimum of effort.
Offshore asset failure detection
On this project, some wellheads were failing, while others were not. In a bid to understand the root cause of these failures, we examined manufacturing data and installation information. We then combined this with other critical operating data such as how many days each well was in service, the operating pressure and temperture ranges. The goal was to find a pattern through machine learning that helped build an advance warning system.
The road to maintenance 5.0
The term Maintenance 5.0 represents the shift from purely predictive strategies to a human-centric, sustainable and resilient approach to asset reliability. Today, Wood is firmly on this path shown by the examples above and through solutions like maintAI. Both combine advanced analytics with decades of domain expertise to ensure your maintenance schedules run smartly and final decisions remain with human beings. By embedding explainable AI into workflows and focusing on resilience and sustainability, Wood is not just adopting Maintenance 5.0 principles, it is helping define them for the energy and industrial sectors.