.png)
Case Study
Delivering automated, network-level optimisation of well fields for a global energy pioneer.
Combining human experience and AI to deliver
autonomous well optimisation and control for Santos
Overview
Global energy company Santos Ltd is enhancing its internal engineering expertise with artificial intelligence software to move towards its goal of autonomous operations.
Collaborating with Kelvin and SIG Machine Learning they have worked together to deliver improved operating efficiency and increased production across Santos’ upstream production systems.
Challenge
Improve overall asset management in a complex and growing system
Santos operates a growing fleet of Coal Seam Gas wells in its Gladstone LNG Roma asset.
The high volume and growing complexity of wellsites created new operational challenges for Santos. Traditional methods of controlling and monitoring wellsites were reaching their limits in terms of efficiency and sustainability.
Outcomes at a Glance
95%
1%
2%
>>
>>
Reduction in time spent on manual tasks
Increase in production
Reduction in energy usage
Enhanced safety by reducing the number of trips to wellsites
Reduced onsite equipment failures through pump off reduction
As Santos grew its number of assets, there was a real shortage of engineering time to optimise these assets. Production Engineers were each trying to maintain and optimise a growing fleet of wells to meet production goals.
Given the large number of wells, only the top 5% to 10% of wells were being assessed for optimisation on a regular basis. Along with a shortage of engineering time, there was also a challenge to balance objectives including:
Improving productivity,
Achieving profitability,
Maintaining safety, and
Reducing emissions to contribute to net zero goals.
Solution
SmartPCP Project
Kelvin and SIG Machine Learning delivered the SmartPCP Project – a software application optimised for CSG operations. The software translates vast volumes of data into actionable insights and automated responses.
The project is a tailored solution for Progressive Cavity Pump (PCP) systems. It integrates with Santos’ existing control systems and has successfully scaled across production operations. The SmartPCP software system is engineered to be self-regulating and self-learning over time, continually honing its capability to drive operational excellence. Further, the application is able to help achieve the operational objectives of Santos, without requiring more engineering time.
Impact
Production up, costs down
The SmartPCP Project has eliminated the challenges of managing a high number of wells each day as a 24-hour operation.
Santos was able to achieve 98% accuracy from SmartPCP recommendations, and a 95% reduction in time spent on manual tasks. Automating these routine tasks resulted in reduced costs and human error.
The improved efficiencies from streamlining workflows saw a 70% time-saving for engineers involved in surveillance and optimisation of well fields. This allowed a switch from manual monitoring to strategic tasks, and enabling well-informed decisions to be made quicker.
Use of advanced analytics to automate optimisation decisions to maximise output gave Santos a 1% increase in production. Coordinated wellsites operations are also delivering better production performance.
The project has also been able to deliver a 2% reduction in energy consumption across numerous wellsites. Ensuring assets operate more efficiently has reduced workovers and the fugitive emissions that result from them.
By minimising the need for human intervention in hazardous areas has also also resulted in enhanced safety.
Keys to Success
Start small, build trust, prove value, and then rapidly scale
Autonomous well management was introduced gradually, with every optimisation control change presented initially as a recommendation to be approved or rejected by engineers.
When trust in the system was established, production moved from “human-in-the-loop” recommendations to automated actions through closed-loop control.
Production engineer tools made it possible to automate app workflows to optimise every well, instead of only a handful of wells.
The SIG Machine Learning model enabled optimisation of CSG wells based on their individual characteristics while contributing to overall field production increase, and reduced emissions.
Santos commenced an initial pilot on 20 wells with the Kelvin system connected to real-time streaming well data. Within weeks, an additional 65 wells had been deployed that further validated the applicability, accuracy and results.
Based on the initial pilot success, Santos moved forward rapidly to scale to standardise “the best engineer on their best day” production optimisation strategies across over 500 wells. This roll-out further validated the application's capability to convert recommendations into automated control decisions by “closing-the-loop” back to the Santos control network.
"We believe we are a “world first”’ in using a cloud-based AI model to optimise wells. SIGs capability in both machine learning and engineering have ensured a successful trial project where others have failed due to not understanding both.
We are now moving toward a large scale roll out and look forward to partnering with SIG and realising $Ms of value in our business."
Engineering Team Leader
Conclusion
Santos Autonomous Operations
Santos is continuing to make strong progress on its path to autonomous operations. By working with Kelvin and SIG Machine Learning to deliver applications that optimise assets, Santos can continue to balance the objectives of improving productivity, achieving profitability, maintaining safety, and reducing emissions as its assets grow.