Understand the problem first
We start from the client's engineering question, available data, constraints and required decision—not from a predefined software solution.
Physics + AI for faster engineering decisions
From the surface to the repository.
Physics where accuracy matters. AI where speed matters.
Numerical simulation plus AI and surrogates: rigorous insight, lower cost, faster turnaround. We use AI to extend physics, not replace it.
High-fidelity physicsAI accelerationFast engineering decisionsCost reduction
About us
About SOROCE
SOROCE is a French engineering and technology company specializing in subsurface engineering, geomechanics, cement testing design, advanced numerical simulation and scientific computing.
We work across CO₂ storage, geothermal energy, underground energy storage, geological disposal and other complex subsurface applications, combining engineering expertise with high-fidelity simulation, experimental and testing approaches, data analysis and AI-enabled computational methods.
Our activities span engineering consulting, R&D, testing methodology, and the development of customized computational tools and software.
How we help our clients
SOROCE helps organizations solve complex engineering problems faster, reduce uncertainty and make better-informed technical decisions.
Our clients often face challenging physics, limited or uncertain data, computationally expensive simulations, complex testing requirements, and projects requiring specialized expertise. SOROCE provides a flexible and pragmatic way to access these capabilities without having to build them entirely in-house.
Use the right level of modelling and testing
From targeted test design and rapid engineering screening to high-fidelity multiphysics simulation, we adapt the approach to what the project actually requires.
Accelerate when it creates value
Physics-based simulation can be combined with AI, surrogate models and automation to explore more scenarios, quantify uncertainty and reduce turnaround and computational cost.
Build reusable value
When relevant, a one-off engineering study can evolve into an automated workflow, surrogate model or customized engineering tool that the client can reuse.
Work flexibly
SOROCE can address a focused technical question, design a testing program, reinforce an existing engineering team, collaborate on R&D, develop a proof of concept, or build a customized computational solution.
Our core value
Complex physics. Practical solutions. Flexible collaboration.
From a specific engineering challenge to testing, simulation and reusable computational solutions, SOROCE helps clients move from problem → evidence → simulation → understanding → decision.
- Numerical simulation
- Geomechanics & multiphysics
- AI & surrogate modeling
- Uncertainty & optimization
- Engineering decision support
The SOROCE portfolio
From subsurface physics to faster engineering decisions.
CO₂ storage
Plume & pressure modelling, injectivity, caprock integrity, fault reactivation, legacy-well screening and history matching.
Subsurface geomechanics
Stress, deformation, poromechanics, fracture mechanics, fault stability and 4D geomechanical modelling.
Geothermal energy
Reservoir and geomechanical modelling, coupled thermo-hydro-mechanical processes and thermal cycling.
Underground energy storage
Hydrogen storage, salt-cavern modelling, containment assessment and coupled THM analysis.
Geological disposal
Thermo-hydro-mechanical modelling, geological barriers and long-term subsurface evolution for nuclear waste storage.
Numerical simulation
High-fidelity multiphysics modelling of complex subsurface systems using advanced scientific computing.
Physics + AI
Combine validated physics-based simulations with AI, surrogate and reduced-order models for rapid engineering predictions.
Uncertainty & optimization
Sensitivity analysis, uncertainty quantification, scenario screening and engineering optimization.
Engineering decision support
Transform complex simulations into practical, rapid and repeatable workflows for engineering decisions.
Evidence in practice
Selected technical studies.
Each case states the engineering question, the method and the decision it supports. The figures are representative. Client studies use project geometry, data and claim boundaries.
The method
Physics first. AI to go further, faster.
High-fidelity models remain the foundation. Simulation results and engineering data can then train surrogate or reduced-order models so large scenario sets, sensitivities, uncertainty cases and screening studies become tractable.
- 01
Physics models
Governing equations and coupled processes for the decision at hand.
- 02
High-fidelity simulation
Validated numerical evidence: pressure, flow, stress, integrity, thermal response.
- 03
Data
Simulation campaigns and engineering measurements assembled for reuse.
- 04
AI / surrogates
Reduced-order models that evaluate many cases once the physics basis is sound.
- 05
Engineering decisions
Sensitivity, UQ, optimization, history matching and operating envelopes.
- More scenarios
- Faster turnaround
- Lower computational cost
- Better-informed decisions
Why SOROCE
Engineering physics. Then acceleration.
SOROCE sits between traditional numerical consulting and scientific software: specialist depth, validated physics, and AI used only to make that physics faster and more useful for the decision.
Physics-driven
Advanced numerical simulation for complex subsurface problems: geomechanics, flow, cement integrity and coupled physical processes.
AI-accelerated
Surrogate models, reduced-order approaches and data-driven methods turn computationally intensive campaigns into rapid engineering tools — they do not replace the physics model.
Validation-first
Models and workflows are benchmarked before engineering use. Validated capabilities are stated separately from R&D and work still under development.
Decision-oriented
Go beyond a single run. Use the model for scenario screening, sensitivity analysis, uncertainty quantification, optimization and a defensible next step.
Solvers are chosen for the problem, not the brochure: analytical methods, Python, GEOS, OpenFOAM, Code_Aster, etc. — without platform lock-in. Clients buy engineering outcomes, not a software stack.
Progressive engagement
Start with evidence. Scale with confidence.
Two contractually separate phases reduce the risk of evaluating a new specialist partner while keeping the first result technically useful.
Feasibility & Validation
A paid, bounded study built around a representative problem, independent references and quantitative validation.
- Useful technical result
- Quality and delivery evidence
- No obligation to continue
Client decisionGO / NO-GOTechnical quality · validation · delivery · collaboration
Industrial Assessment
After approval, the complete study uses real geometry, properties, operating history and industrial scenarios. AI acceleration is used only when the physics basis justifies it.
- Client-specific model
- Calibration, sensitivities and screening
- Engineering recommendations
The percentages are indicative. Phase 1 reduces the cost and risk of evaluating SOROCE; it is not a discounted fraction of the full industrial project.
Start with one decision
Bring the question. We will define the evidence.
A short brief is enough to start: the decision, the data you have, and the claim you need to support.