If you’ve ever fancied navigating the Death Star’s trench to target the thermal exhaust port, or dodge asteroids in the Millennium Falcon, geosteering might well be the droids you are looking for…
In the climax of A New Hope, Luke Skywalker is under pressure, navigating through the hostile environment of the Death Star while waiting to hit pay – the exposed exhaust port that will save the day. With only space meters to spare, he enters the trench, threading the needle, staying away from the boundaries, avoids an early exit and a costly sidetrack, hits the target and pulls out to complete the objectives before conditions deteriorate.
In The Empire Strikes Back, Han Solo expertly navigates the Millennium Falcon through a dangerous asteroid field. Using the ships sensors, real time displays, and his experience, he is able to optimally steer through a complex and narrow target zone, while avoiding the dangerous geological hazards all around him.
The ability to navigate complex trajectories, land out and stay in target zones, avoid geological hazards, and hit the target were as critical to Mr Skywalker and Solo as it is to the oil and gas industry. While they might have had the use of the Force, they also had use of advanced sensors and software to guide them.
Lucky for us, so do we! When it comes to geosteering, having access to the right downhole data and utilising the right software and techniques to interpret it, gives us the ability to successfully keep the bit within the target zone and navigate through complex geological formations to optimally place the well.
But first, what is “Geosteering”?
Geosteering is an advanced drilling technique used in the oil and gas industry to optimise the placement of a borehole in real time. Its primarily purpose is to ensure that the drill bit remains within a specific, often narrow, target zone within the geological formation to maximise reservoir contact (i.e. hitting as much pay zone as possible) and boost recovery. Typically this means drilling horizontal wells, where the well path tracks or targets specific formations or best reservoir properties, or can connect several separate reservoir packages. Traditionally this could be achieved with a basic gamma ray tool, monitoring its response and looking for formation markers letting you know when in or out of the reservoir package. More recent advances allow for LWD sensors to penetrate their reach far farther into the surrounding geology, opening up different geosteering techniques and options. The use of deep and ultra deep resistivity tools, for example, allows for resistivity measurements to penetrate up to 300ft form the well bore (more on this next time on #ToolTuesday)!
The benefits of geosteering are widely researched and documented, some key highlights are:
- Maximise Hydrocarbon Recovery: Keep the wellbore within the target formation, aim for most productive zones, connect multiple reservoir packages, and avoid exiting the reservoir early, geosteering typically leads to enhanced extraction of hydrocarbons.
- Avoid Hazards: Help avoid geological hazards and minimise drilling through non-productive formations (e.g. shale layers or baffles), problematic zones (rubble zones, unstable formations, faults), or away from subsurface features (oil water contacts).
Guevara et al., 2012, Janwadkar et al., 2012
How It Works
- Real-Time Data Collection: Using the real time data while drilling, LWD data such as gamma ray measurements, resistivity, density and neutron logs provide information about the surrounding rock and fluid characteristics.
- Interpretation and Modelling: This downhole data is fed into geosteering software that can model expected responses and interpret what the downhole environment is, or will be, further along the well path. Geosteering specialists can interpret this data and adjust the geological models to ‘fit’ the down hole measurements, and use the model to guide the drilling plans in real time.
- Adjustment and Monitoring: The model versus downhole tool responses is continually monitored to determine the position of the drill bit and sensors relative to the target zone. If the bit or the well path looks to be on course to exit the target zone prematurely, steering requirements are sent to the wellsite team who adjust the directional drilling systems and well path to optimise the trajectory.
While basic geosteering can be achieved by a solid understanding of the geology and a keen eye for pattern recognition, such as the LWD response when going up dip or down dip, or when crossing the same formation or geological horizon, more advanced geosteering often requires more advanced tools and data interpretation techniques.
A critical component of more advance geosteering is being able to predict what the LWD data should look like, ahead of drilling the well. Similarly, while in hole and geosteering, it is possible to predict what the geological environment is like by back calculating the sensor data to give likely scenarios. This is commonly known as Forward Modelling and Inversion.
Forward Modelling
Forward modelling = Use offset well or model to predict data
Forward modelling in geosteering is a predictive technique used to simulate how a well path will interact with the geological environment based on existing data and geological models. This approach allows operators to visualise the subsurface model, where exactly the drill bit and BHA are in the model, as well as anticipate expected formations and subsurface changes. Forward modelling enables making informed, proactive geosteering decisions in real time.
The primary goal of forward modelling is to reduce uncertainties and improve the accuracy of well placement in real-time drilling operations. By predicting the subsurface conditions ahead of the bit, operators can optimise the well path with the aim of maximising reservoir contact, and avoiding potential hazards. Forward modelling is a proactive strategy that leverages existing data to simulate and predict subsurface conditions, thereby enhancing decision-making and optimising the drilling process. This method is crucial in environments where precise well placement is critical for maximising hydrocarbon recovery.
How It Works
- Predictive Approach: Forward modelling begins with software that integrates various types of geological data such as offset well data, geological surfaces, log responses, well trajectories, rock types, fault, seismic data or mud logs to build a synthetic hypothesised geological model.
- Simulation: Using this data, the computational model simulates what the expected log responses should look like along a proposed well path as it encounters various formations. The model calculates the expected measurements and outputs (e.g. gamma ray, resistivity, density, neutron log response) as the hypothetical well path progresses through the geological strata.
- Comparison and Adjustment: When drilling, the simulated outputs are then compared with the real-time data collected from the LWD sensors. The model is continuously adjusted in real time to ‘fit’ the downhole tool responses. Should the downhole responses differ from the modelled responses, then adjustment of the model is required to better correlated the ‘expected vs actual’ log responses. This helps in planning and adjusting drilling operations to stay within the most productive zones. Often as geology is tricky, multiple models can be required.
Inversion
Inversion = Prediction of a model which adequately fits the observed data
The technology and calculations to perform inversion has been used by geophysicists for seismic interpretation for over 20 years now. In geosteering, forward modelling and inversion are two complementary methods used to understand and predict subsurface models, but they operate in opposing ways. Forward modelling takes a given model and predicts the response a downhole tool will produce in that model scenario, inversion does the opposite and uses real time data points from the in-hole LWD sensors and works “inversely” to generate a model that fits the actual observed measurement data. Inversion is particularly important to the latest generation of ultra deep azimuthal resistivity tools, which allow resistivity measurements in excess of 300ft away from the wellbore (a future #ToolTuesday post in the making!).
How It Works
- Real time data collected by downhole tools is pulsed to the surface and decoded by the surface systems.
- Due to the computational demands to perform inversion calculations, the data is typically sent to either processing computers, or increasingly, cloud computing is used to speed up the process.
- The data is then processed using various computational modelling methods (typically deterministic and stochastic models, with some hybrid systems in use) to derive realistic geological model(s) from the actual measurements, i.e. to predict the most likely geological setting that would result in the down hole measured LWD data. Inversion software interprets the physical properties of the subsurface by fitting computational models to the observed data.
- The output is typically a visualisation of the geological setting, rendered in such a way that allows the end user (geosteering and subsurface teams) to view a cross section of the wellbore.
How the inversion is calculated is generally split into either a deterministic or stochastic calculation method, however there are some variations that use a hybrid approach to resolving calculations. Deterministic and stochastic inversions are two methods used in geophysical data interpretation (i.e. not just geosteering), each with distinct approaches to model subsurface properties:
- Deterministic Inversion: This method produces a single best estimate of the subsurface properties based on the input data and predefined models. It assumes a clear relationship between the data and the model, leading to precise predictions of geological features. Deterministic inversion is often used when the geological structure is well understood and the model parameters are well constrained.
A potential downside of a deterministic inversion is that the output is heavily reliant on the input data. Without suitable offset data, the inversion calculation will struggle to accurately resolve the geological prediction. This could be considered as an approximate of survivorship bias, only known and expected geological models are considered in the calculator, with outliers excluded.
- Stochastic Inversion: Unlike deterministic methods, stochastic inversion acknowledges the uncertainty in subsurface properties by generating multiple models (or solutions) that are statistically probable, based on the input data. This method uses probabilistic techniques to account for the inherent uncertainties and variations in geological data, providing a range of possible outcomes rather than a single solution. It is particularly useful in complex geological settings where the subsurface properties are not well known or highly variable.
A stochastic inversion model is more independent (although not fully, some constraints are usually necessary such as limiting the number of layers to save on computing power) and will produce results that do not depend on offset well data to generate. Due to this, the inversion model provides a less human depended result, with the model based purely on the inversion algorithm and not down to any bias towards current reservoir understanding.
While each inversion type has pros and cons, each method has its applications depending on the level of detail and the uncertainty of the geological information available. It is wise to discuss the advantages and limitations of each method during tender stages with any potential geosteering vendor. Better yet, there is vendor independent modelling software that allows operators the ability to run their own stand alone models…trust but verify!
Technical advisory
When planning on geosteering it is worth considering the key objectives, and understanding what are ‘must have’ and what are a ‘nice to have’. For example, do you need ultra deep resistivity measurements with the longest reach if your sand body is a thin layer that is uniform across the field? Potentially not. If you have a massive tank reservoir with good depth and seismic control, then a shorter reading geosteering service may well be the most effective solution. Similarly, if the plan is to drill geometrically, with only minor adjustments to the inclination, then opting for a ‘3D’ real time inversion may bring limited to no additional value. Do you need, and are you ready to drill azimuthally? Often the options available when planning ultradeep resistivity inversions can seem daunting if this is the first time being exposed to the technology. Some considerations to include during tendering and well planning phase will pay off during operations:
- Geosteering requirements: What is the well or geosteering requirements? Will ‘basic’ LWD sensor data be enough to meet the objectives, or is deep or ultra deep resistivity a requirement? If not a requirement, is the data value a justification for increased BHA complexity and cost?
- Reservoir characteristics: Any known subsurface offsets can help understand the expected geosteering response for various tools and services. If the field has multiple penetrations and a solid seismic data set, then there may be less dependence on geosteering. If the resistivity contrasts between shale layers and reservoir sands is low, or bed boundaries are gradational then resistivity measurements will be less pronounced. A simple gamma ray or density neutron tool string may prove more useful in some circumstances. Conversely if the area has a high risk of faulting, then the use of deep reading tools may help locate the reservoir above or below the BHA if crossing a fault.
- Depth of detection: It is worth considering that published figures for how far geosteering tools can ‘see’ is often taken in ideal conditions. Don’t let headline figures convince you one service is superior to others, in reality, in non ideal conditions, the distance around the wellbore each service or tool will detect change is far lower…
- Modelling: Which brings us nicely to modelling, which should be a minimum requirement from any potential vendor to visualise the expected tool responses in a given scenario. Pre well modelling should be carried out as a proof of concept for your reservoir, showing how each tool, sensor or configuration is likely to respond. You may find that a given service responds earlier to a formation change or reservoir top, and gives you advanced warning to stop drilling or adjust the well angle to land out at the desired inclination. A point to note, basic modelling typically will be included to qualify a vendor to participate in a tender. Advanced modelling is a more time consuming and involved process and should be limited or reserved for genuine technical evaluations as there will often be a limit to how many models will be produced by the vendor.
- Inversion type: Depending on the well or field, certain inversions may be more suitable to your subsurface model. If multiple offset wells can be used during the modelling process, deterministic inversion may be suitable. However, where offsets are limited or absent, then stochastic inversion may be preferred as the answer product is less dependent on input data to generate a geological model. It’s important to note that this isn’t a mutually exclusive choice, and each method has its merits and flaws.
- Another tool, not the only tool in the box: Its worth keeping in mind that advanced geosteering techniques and tools are the latest tools that can be used to help deliver your well, however more conventional geosteering sensors also still exits. Non-azimuthal (conventional/bulk) logs as well as imaging tools also provide valuable near bore insight and aid in interpretation of the geology. Ideally a combination of tools and sensors should be used to provide a range of data to base geosteering decisions on.
- Dogleg limitations: While the idea of threading the needle and following an undulating reservoir contact or sweet pay zone might appeal, typically the type of wells geosteering is most valuable in are when a completion will be run in the well. In this case, dogleg limits need to be understood during the planning so that effective drilling strategies and dogleg limits are set. In addition, while being able to detect a reservoir around you is one benefit of geosteering, the bit has to be able to physically get there. If the BHA limits for dogleg capability are 4 °/100ft, but the reservoir is dropping away from us at 6°/100ft, then it is possible the well path will never catch up.
- Geosteering set up: With remote operations a regular activity in the industry, it is no longer required to have the geosteering team physically located in the same room as the subsurface team. So long as there is a stable and sufficient bandwidth to allow streaming of the real time data and inversion, geosteering can take place from other rooms, buildings, cities or even continents. Having said that, the value experienced geosteering personnel bring to any operation often goes beyond the basic deliverables. Personal experience, offset well knowledge, formation tendencies, lessons learned and face to face discussion to allow optimum well placement is a less measurable but no less beneficial element to consider. Fostering a collaborative ‘one team’ approach may be easier to cultivate if those involved are physically together.
- Data deliverables: It is important to understand what the answer product is when evaluating your vendors. What will be delivered to you, how will this be delivered, presented, and transferred. What data will you receive in real time (and memory) to allow you to actively geosteer? Will you be presented with predictions of model certainty (i.e. model has low confidence or high confidence), how many models are being run simultaneously, are you limited to the number of inversion layers? At the end of the well, is the inversion data yours to keep, or is the inversion image the only deliverable? Before agreeing on a service, it is prudent to discuss and agree what the required deliverables are, and what data will not be provided. The answers are not consistent across each vendor, be warned…
Conclusion
Geosteering is an ever-growing market; as oil & gas exploration and production develops increasingly more complex fields, more unconventional reservoirs, or harder to reach targets, and previously uneconomical wells become achievable by allowing more reservoir contact. The options of tools, services, processing and personnel that allow geosteering are higher than ever before. While this may be daunting at first glance, the reality is that geosteering simply gives you more options to hit the target, more of the time, and with a better understanding of your downhole environment. By employing varying levels of geosteering, despite the initial increased outlay, the benefits over the short and long term lifespan of a well can be drastically improved, with higher production and lower overall well cost.
As we have only touched on the surface of this topic, in the following weeks we will take a look at some more specific services on offer. For now, may the inversion be with you.
