Sign up for free trial
Sign up and experience the power of Exebenus technology
Machine Learning Agents to prevent drilling incidents and increase performance
Sign up to experience the power of Exebenus Machine Learning
About product
Exebenus Spotter is a set of real-time machine learning applications (agents) to address critical drilling NPT and ILT issues. Targeted machine learning algorithms each work on specific problems, such as mechanical sticking, differential sticking, hole cleaning, washouts, mud losses, bit degradation, motor/tool failure, ROP optimization and others. If applied together, ML agents work collectively, interacting and supporting each other. The offered solution is an automated plug-and-play web application, WITSML vendor-agnostic, with no need for the models to be pre-trained on offset data and no back-office engineering support required.
During planning, our machine learning agents can be used for offset well analysis to optimize drilling parameters, evaluate risks and prepare ready-for-action contingency plans. This increases operational efficiency, reduces potential downtime and helps eliminate the hesitations that can add up to invisible lost time.
Real-time
ROP Optimization
ROP
Benefits
to any # of wells
Plug & Play on your WITSML
Key aspects
26 min is the approximate time taken to set up Exebenus Spotter on your next well.
The application consumes real-time or historical WITSML drilling data. Minimal configuration time, minimum human intervention and no data filtering or cleaning required. Built-in data pre-processor will automatically prepare the raw data needed for each ML agent. The output is displayed back within the same WITSML viewer you have been using, meaning no need to install any new software or dashboards on your screen.
Integration with your current workflows and real-time operation centers is seamless.
The rate of penetration (ROP) is a major contributor to drilling time and costs. Today, optimized ROP is achieved by adjusting the weight on bit (WOB), RPM, mud flow through the use of time consuming and complex simulation models. In addition uncontrollable factors such as bit dulling, buckling, vibration and strength formation can influence the ROP.
Exebenus Spotter Real-time ROP Optimizer is unique in its ability to decipher the relationship between the controllable and uncontrollable drilling parameter relationships. Our ROP machine learning agent provides immediate and reliable real-time advice on the combination of WOB, RPM and Mud Flow parameters 0.5 m ahead of the bit, leading increase in safety or ROP when possible.
Exebenus Spotter Stuck Pipe agents are designed to predict specific conditions related to pressure differentials, mechanical sticking, hole cleaning and wellbore geometry. Without any intervention, these conditions typically result in costly stuck pipe drilling incidents. Warnings and alarms are provided 30 minutes to 4 hours prior to potential events, giving rig crews sufficient time to take mitigating actions.
When used on historical real-time data as part of offset well analysis, the agents can identify unreported near misses and provide guidance for optimizing performance in the future.
At Exebenus, we develop an entire range of proactive Machine Learning agents to cover the full spectrum of drilling challenges, potential incidents and NPT/ILT causes. Clients have an option to pick-and-choose which agents they want to run depending on their operations risks and hazards, however we recommend to run all our agents together to ensure a full coverage and a great piece of mind.
Remember, running agents together means that they work collectively as a team of little virtual robots, cross-checking each other and thus delivering the most optimized and safe drilling operations.
How it works?
At Exebenus, we have chosen to develop targeted machine learning models rather than complex statistical models. Why this approach?
Complex models consume vast amounts of data, and take longer to set up, train and run. In contrast, our targeted models solve well-defined problems and deliver more accurate predictions. They use data that’s always available in real time on the rig, which means our agents can be used anytime, anywhere, easily.
Exebenus Spotter will automatically recognise the current rig status and current drilling operation type (tripping, reaming, drilling, hole cleaning, etc…) and will activate the appropriate ML agents that are relevant for the current operation, thus minimizing the amount of false positive alarms.
Self adapting algorithms
to any lithology, bit type, BHA and mud properties
Pre-trained physics-
informed targeted models
with rig status recognition
Built-in automatic data cleaning
and pre-processing
Sign up to experience the power of Exebenus Machine Learning
Agents portfolio
time
data
sticking
sticking
real-time
in – volumetric flow rate, Mud density in
(bit balling, vibration, buckling, bit degradation, ECD)
time
data
sticking
sticking
real-time
in – volumetric flow rate, Mud density in
(bit balling, vibration, buckling, bit degradation, ECD)
Value creation
92 of 100
Valid alarms
2.1 hours
Average prediction
heads-up time
100%
Accurately predicted
drill incidents
5.2MUSD
Average Cost saved per well
* based on average statistics from ~200 wells case study in Southeast Asia
Looking for supporting documentation?
Sign up for free trial
Sign up to experience the power of Exebenus Machine Learning