AI-Driven Operational Efficiency for Energy Leader

Challenge

A multinational, vertically integrated energy operator was motivated to shift to a risk-based approach for its offshore platforms. While the operator has developed an effective monitoring methodology with a central surveillance team and a wide set of tools, they saw an opportunity for increased operational and production efficiency with AI.

While the existing monitoring approach was effective, it was time-intensive and inefficient due to siloed systems. The central surveillance team used multiple systems to monitor asset health, including data dashboards, sensor analyses, maintenance logs, and case management systems. Similarly, when on-site engineers received an alert from the central surveillance team, they had to navigate different systems to understand alerts and related sensors and evaluate individual sensor trends. Moreover, the existing systems flooded central surveillance with false alerts, compounding the time and effort required for risk identification and triage.

As the global energy operator expanded this monitoring approach, they recognized the need for a more holistic and risk-based approach that provides them with an accurate view of process system risk, enabling them to do the right maintenance at the right time, and allocate precious engineering resources to the highest-risk systems and other high-value activities.

Additionally, in anticipation of forthcoming government regulations on carbon emissions taxation, the company sought a digital solution to decrease power and fuel consumption, lower carbon emissions, adhere to regulatory requirements, and improve operational efficiency.

Approach

In 2022, the energy operator chose to partner with C3 AI to implement both the C3 AI Reliability and C3 AI Process Optimization applications to drive operational efficiency for one of its offshore platforms with advanced AI. The operator evaluated tools from different providers, and C3 AI was selected because of its ability to rapidly deploy and scale AI-enabled applications, and its domain expertise in the energy sector.

The initial deployment was focused on two turbine-driven gas compression trains. In 16 weeks, the C3 AI team successfully unified and integrated 3 years of historical data and live data, and developed 20 ML models in C3 AI Reliability to deliver near-real-time interpretable and predictive alerts for asset anomalies. As well, the team configured an advanced optimizer in C3 AI Process Optimization to drive enhanced fuel efficiency for the company.

With C3 AI Reliability, the global energy operator reduced alert noise by 99% and alert triage time by 90%. With the time and effort saved by C3 AI Reliability, the on-site team can focus on higher-value initiatives such as fuel gas optimization. With C3 AI Process Optimization, the company can save up to $4.7 million in annual carbon tax for just one platform.

About the Company

• Multinational oil and gas corporation
• $200+ billion annual revenue
• 20+ manufacturing sites globally

Project Objectives

• Integrate and unify data from disparate data sources (e.g., sensor data, work orders, vibrational data)
• Apply machine learning algorithms to predict asset health in near real-time, and generate interpretable alerts to allocate precious engineering resources
• Leverage AI and machine learning technology to optimize production processes, driving energy efficiency of platform equipment

Project Highlights

• 16 weeks from kickoff to pre-production
• 2 enterprise applications
• 3 years, 100M+ rows of historical data
• Live data integrated in 3 weeks, with >1.2B rows ingested to date
• C3 AI Reliability – 20 ML models developed to detect equipment failure and anomalies
• C3 AI Process Optimization – 10 iterations of an advanced optimizer