Schlumberger Software: Building digital models of the subsurface
By Dr Chris Banks; Senior Geoscience Workflow Consultant
Developing an oil and gas field is a costly business, often running into billions of dollars. Drilling oil wells well is also an inherently risky business involving high pressures, H2S and flammable gas presenting risks to infrastructure, equipment, environment and life (e.g. consider the Macondo Well blow-out). The operators of oil and gas fields simply cannot afford to make mistakes. Software technology allows operators to simulate different development scenarios and make mistakes in the digital world in order for their development plans to be honed for the physical world.
When simulating development scenarios, oil companies build digital models of the subsurface. These models are divided into digital geocells (like Minecraft models or digital Lego bricks) with a size and number appropriate to the hypothesis/ question being asked. Geologists build “fine-scale” models consisting of millions of cells to answer questions such as, “where is the oil?”, “how much oil is there?”, “where should I drill?”. Reservoir Engineers typically build much coarser models because their models make many complex flow calculations to predict fluid production rates, pressures and compositions. Whereas the fine-scale models are fixed in time or “static”, reservoir engineers calibrate their coarse-scale “dynamic” models by running them back in time to calibrate against historical production records prior to running them forward to make production predictions.
Traditionally, geologists build the fine-scale models and then pass these to the reservoir engineers who “up-scale” them to be as coarse as their job requires. Upscaling was a necessary imposition due to computing limitations. Time is money and engineers simply cannot wait for simulations on millions cell models and so they reduce their resolution to provide timely answers.
Industry standard reservoir simulators have been around for c.30 years and computing infrastructure was a little different in the 80s. These 1st generation simulators have had continual development/improvement since their inception and have been proved throughout the oil industry. However, in a post-peak oil world with the easy oil gone, we are forced to explore and produce from areas with higher orders of geological complexity requiring many more cells than these classic simulators could accommodate.
The next generation of reservoir simulators are designed on modern computing infrastructure and are infinitely scalable with the number of licences and cores. This scalability, coupled with faster solvers, allows the engineers to do more simulations and test multiple development scenarios to find what is cost-effective, technically-effective and less risky. Next generation simulators (e.g. Schlumberger’s INTERSECT) permit testing complicated wells, thousands of wells per field, multi-stage completions and work overs, stimulation, hydraulic fracturing, multi-component fluids, multiphase fluids amongst many other possibilities.
Many millions of cells also enable greater representation of geological complexity because simulation can be run on the fine-scale model. Complex vertical and lateral facies heterogeneity, structural complexity in faulting/fracture networks and fine-scale changes in petrophysical properties (porosity and permeability) all effect the plumbing and performance of production and injection. Simulation at this scale gives greater realism to the simulations and so we can better predict outcomes of development.
Historical workflows involve reservoir engineers and geologists modelling their respective disciplines separated by an integration barrier called “upscaling”. In reality, the subsurface does not care about your scientific/engineering discipline and neither do the financial goals of the oil business. Discipline specific modelling was imposed by computing limitations that no longer exist.
An asset team should be able to give the business decision makers integrated answers, regardless of discipline. Geophysicists are now working more effectively with geologists since the incorporation of seismic interpretation tools into geomodeling suites. The loss of upscaling permits geologists and engineers to work within a single model. For the first time, seismic to simulation truly can be achieved within a single model.
Developing an oil and gas field is a costly business, often running into billions of dollars. Drilling oil wells well is also an inherently risky business involving high pressures, H2S and flammable gas presenting risks to infrastructure, equipment, environment and life (e.g. consider the Macondo Well blow-out). The operators of oil and gas fields simply cannot afford to make mistakes. Software technology allows operators to simulate different development scenarios and make mistakes in the digital world in order for their development plans to be honed for the physical world.
When simulating development scenarios, oil companies build digital models of the subsurface. These models are divided into digital geocells (like Minecraft models or digital Lego bricks) with a size and number appropriate to the hypothesis/ question being asked. Geologists build “fine-scale” models consisting of millions of cells to answer questions such as, “where is the oil?”, “how much oil is there?”, “where should I drill?”. Reservoir Engineers typically build much coarser models because their models make many complex flow calculations to predict fluid production rates, pressures and compositions. Whereas the fine-scale models are fixed in time or “static”, reservoir engineers calibrate their coarse-scale “dynamic” models by running them back in time to calibrate against historical production records prior to running them forward to make production predictions.
Traditionally, geologists build the fine-scale models and then pass these to the reservoir engineers who “up-scale” them to be as coarse as their job requires. Upscaling was a necessary imposition due to computing limitations. Time is money and engineers simply cannot wait for simulations on millions cell models and so they reduce their resolution to provide timely answers.
 |
| The difference between fine-scale and coarse-scale modelling |
Industry standard reservoir simulators have been around for c.30 years and computing infrastructure was a little different in the 80s. These 1st generation simulators have had continual development/improvement since their inception and have been proved throughout the oil industry. However, in a post-peak oil world with the easy oil gone, we are forced to explore and produce from areas with higher orders of geological complexity requiring many more cells than these classic simulators could accommodate.
The next generation of reservoir simulators are designed on modern computing infrastructure and are infinitely scalable with the number of licences and cores. This scalability, coupled with faster solvers, allows the engineers to do more simulations and test multiple development scenarios to find what is cost-effective, technically-effective and less risky. Next generation simulators (e.g. Schlumberger’s INTERSECT) permit testing complicated wells, thousands of wells per field, multi-stage completions and work overs, stimulation, hydraulic fracturing, multi-component fluids, multiphase fluids amongst many other possibilities.
 |
| Schlumberger's INTERSECT software - A next generation simulator |
Many millions of cells also enable greater representation of geological complexity because simulation can be run on the fine-scale model. Complex vertical and lateral facies heterogeneity, structural complexity in faulting/fracture networks and fine-scale changes in petrophysical properties (porosity and permeability) all effect the plumbing and performance of production and injection. Simulation at this scale gives greater realism to the simulations and so we can better predict outcomes of development.
Historical workflows involve reservoir engineers and geologists modelling their respective disciplines separated by an integration barrier called “upscaling”. In reality, the subsurface does not care about your scientific/engineering discipline and neither do the financial goals of the oil business. Discipline specific modelling was imposed by computing limitations that no longer exist.
An asset team should be able to give the business decision makers integrated answers, regardless of discipline. Geophysicists are now working more effectively with geologists since the incorporation of seismic interpretation tools into geomodeling suites. The loss of upscaling permits geologists and engineers to work within a single model. For the first time, seismic to simulation truly can be achieved within a single model.
*You can keep up to date with GERC activities on our social media channels*
Comments
Post a Comment