Automation & Robotics

ML and AI create value for plant digital twin technologies

Training new operators is one of the most pressing needs of the industry

The manufacturing and process industries are challenged by volatility in the market and have a huge pressure to exercise control on their maintenance and operations spending. The industry has seen a huge benefit in leveraging predictive technologies powered by machine learning (ML) and artificial intelligence (AI) for bringing in maximum efficiency and taking proactive decisions leading to tangible savings and reduced costs.

It is estimated that in the Middle East & Africa (MEA) spending on AI will grow 24.7 per cent to total $1.2 billion, spending on Big Data analytics will grow at 8.1 per cent to reach $3 billion, and spending on public cloud services will grow 27.3 per cent to surpass $6.8 billion in 2022. Currently, the MEA industrial sector is well positioned to embrace Industry 4.0 technologies.

Rana Atif Rehman

Rana Atif Rehman

One of the most pressing needs of the industry is training new operators in simulated environments to get them well versed with plant operations, that is, process modelling, process simulation and optimisation. The Operator Training Station (OTS), an off-line simulation system, has been supporting this need for decades. However, organizations are now realising the importance of unutilised data lying in their systems. This unutilised data can be used with the Digital Twin to achieve more accurate models for simulation.

Hybrid Digital Twin solutions combines First Principal Thermodynamic Process Models (that is, domain expertise of process contains decades of knowledge of mass and energy balance, heat transfer relations etc.) along with artificial intelligence creating all-inclusive and precise models for all stages of asset lifespan. This feature is vital for understanding and predicting the safe and efficient operations for select processes. The assets along with its systems, that is, ICSS, LIMS, etc. are becoming more complex with the highest desirable efficiency in place and lowest O&M costs. 

The Digital Twin provides insight to process, reliability, and operations teams to:

• Visualise the real-time plant behaviour based on models fetched from operational data and first-principles constraints;

The industry has seen a huge benefit in leveraging predictive technologies powered by ML

The industry has seen a huge benefit in leveraging predictive technologies powered by ML

• These highly reliable models support enhanced decision-making for real-time optimisation and savings;

• ML and AI makes models retuning easy for ever changing asset operating / process conditions.

With the Digital Twin solution, ML leverages simulation or plant operations data to develop an enhanced model.

The Digital Twin creates virtual copies of physical locations, plant processes, business processes, and assets. With the utilisation of AI, the Digital Twin also allows plant operators to obtain value from existing plant data which can be leveraged to drive improvement across various operations.

 

These models use operational data for plant equipment and/or system processes, even if First Principal Thermodynamic models are not available. The scalability of the solution is not limited due to lack of data science knowledge or expertise.

The two technologies capture what is happening in the plant and, with the data insights provided by them, plant operators can reach better operational decisions, and achieve enhanced operational excellence.

The industry is continuously supported by Emerson in terms of state-of-the-art control systems, field instrumentation / control valves etc. Emerson is a leader for introducing technologies of the future for safe and reliable plants. It is one of the leading OEMs for software and platforms for industry, supporting the digitalisation roadmaps of leading MEA O&G upstream, midstream, and downstream industries along with chemical, petrochemical, water, power and others.