The application of a digital twin aggregate for an LNG liquefaction process

Dino DiMattia's picture
Dino DiMattia, STP - Process Engineering & Loss Prevention Advisor, ExxonMobil
Rupesh Parbhoo's picture
Rupesh Parbhoo, Process Engineer, ExxonMobil
Comments: 0

Process surveillance is a key aspect in achieving optimal process efficiency of the liquefaction process, mitigating the risk of unplanned downtime.  Enhanced digitalisation through the application of a digital twin aggregate provides a platform for predictive analytics that integrates process, integrity and safety virtual representations of physical processes to maximise production and reduce the risk of downtime events (Figure 1).  

Figure 1 - Objectives of a Digital Twin.

The digital twin (twin) is a virtual model of the components that make up a physical process.  The twin’s ecosystem creates a setting that can improve operating efficiencies, prevent downtime, reduce maintenance costs, and create a lower risk environment throughout the lifecycle of the facility.  The twin uses field data in a virtual representation of equipment and processes to communicate real-time performance data to both operations and remote engineering locations.  Through the continuous monitoring of performance data the twin can identify trends and quickly apply past learnings to models that can predict future performance of the liquefaction process.

The twin is a predictive analytical vehicle that connects two spaces, the material and the virtual, in order to make better decisions through insights which are difficult to generate in a timely manner with traditional surveillance programs.  Process and integrity surveillance is an intensive human effort that requires the gathering and categorisation of data to allow analysis key performance indicators (KPIs).  Conversely, a twin continuously collects sensor and inputted data, applying innovative analytics and self-learning to gain insights about system performance and operation, perpetually tuning models for greater accuracy, providing oversight of the physical components of the process. 

The twin can acquire knowledge on a continuous basis, through the application of machine learning (ML) algorithms. This allows the twin to optimise process performance and reduce risk through proactive analysis of failure mechanisms and predicting future process conditions that can lead to premature failures or process conditions that are outside the normal operating envelope of process components.  The time-based demands of tasks, such as data gathering and analysis for process and integrity surveillance, can be reduced or even eliminated to provide expert analysis of process anomalies and remedies.  The twin acts as a continuous source of knowledge transfer, ensuring the boundaries (operational and safety) set in the design phase of projects are not violated by current or predicted future operating scenarios. 

Applying an enhanced digital strategy to the liquefaction process can complement current tactics through additional functionality and powerful on-line analytical tools, within a surveillance framework, that encompasses the following aspects:

  • Real-time simulation modelling and historical data analytics of liquefaction process
  • Early advice/recommendation based on changes to mixed refrigerant (MR) composition
  • Liquefaction efficiency analysis (e.g. compressors, heat exchangers)
  • Predict impact on process efficiency and LNG production rates

The flexibility of the digital twin is ideal to accommodate and adapt to the various liquefaction processes that exist today and to future modifications as these processes evolve.  Employing an enhanced digitalisation strategy helps facilitate real-time process optimisation and provides a platform for value-added surveillance capabilities improving operator and engineering situation awareness (Figure 2).

Figure 2 Liquefaction System - Digital Twin Functions.

The ecosystem of a digital twin incorporates both integrity and consequence analysis algorithms to form a comprehensive aggregate for each piece of equipment within the LNG plant.  Each twin resides within a honeycomb environment seamlessly linked and learning from past operations, while conducting real-time analytics and predicting future performance, allowing for better decision making (Figure 3).

Figure 3 Representation of a Digital Twin Aggregate.

A comprehensive digital twin encompasses surveillance on a process, integrity and safety basis, providing a powerful predictive tool that can complement existing surveillance programs to enhance decision making and reduce downtime.  The main attributes of a digital twin surveillance program includes:

  • Application of real time data to process and integrity based models
  • Sharing results between models to provide enhanced situation awareness 
  • Predictive what-if capabilities that allows engineers and operations to support decision making
  • Assessment of risk both from an operational and safety perspective 

The application of ML algorithms provides value to both existing surveillance programs and process control systems.  These algorithms provide insights from the facility’s operating history, bringing invaluable knowledge and understanding to optimise and troubleshoot current operations.  The continuous surveillance of operations and equipment integrity provides fast and accurate feedback to enhance key day-to-day operating decisions provide insightful feedback to help optimise LNG production.

Renowned for its quality, breadth and expertise, the Gastech Conference is one not to be missed! Register now to hear the exclusive insights from the world-class leaders.