Pipesim steady-state multiphase flow simulator
Production optimization over the complete lifecycle
The new generation in multiphase flow simulation to overcome fluid flow challenges and optimize production.
Physics-informed AI
Simulation-quality insight with the speed and flexibility of AI
Overview
Optimize production and process systems in real-time
Optimizing complex production and process systems requires accuracy in real-time. Traditional oil and gas operations rely on expertise limitations, siloed working, and slow, non-continuous optimization to operate assets. This leads to excess carbon emissions, process inefficiency, and new net zero technologies that are economically unfeasible. AI offers promise however, training data availability is typically very limited. The non-linear behavior of systems is difficult to capture, and contextual understanding is required to address previously unseen conditions.
Physics-informed AI addresses these challenges, by performing predictive, real-time optimization for a wide spectrum of complex systems across upstream, midstream, and downstream operations.
Applications
The solution delivers scientifically consistent results, captures system dynamics, delivers interpretable results, and typically solves orders of magnitude faster than physics-only simulators.
Working in partnership with Geminus AI, our physics-informed AI model building service delivers models that can be accessed through advisory applications, enabling operators to enhance profitability and reduce carbon footprint, at scale.
Benefits
Physics-informed AI optimizes complex oil and gas production and process systems in real time across a wide range of scenarios.
Proof of value has been clearly established in a wide range of case references, including methanol injection optimization, liquid slug mitigation, electric submersible pump (ESP) power optimization, natural gas liquids (NGL) processing, and gas-oil separation plant (GOSP) optimization. Results include:
Features
Industrial AI at scale
Highly flexible and portable— applicable to a myriad of use cases.
Agnostic to source and suitable even when data sets are sparse.
Accuracy that builds trust
Multi-fidelity hybrid models that maximize the accuracy of model estimates.
Models adapt as systems and operating conditions change.
Explainable AI, with uncertainty quantified.
High Speed ROI
Models in hours, not months.
From concept to delivery in weeks, with easy deployment.
Repeatable techniques, speed model creation, and re-training.
Helpful links
About SLB