The problem
Oil and gas production data is noisy, incomplete, and expensive to interpret manually. Operators need to spot operational anomalies and understand likely production ranges without treating every model output as certainty.
Loading page
Preparing the content for you.
Production intelligence with uncertainty in view.
An AI-assisted well and reservoir intelligence platform for oil and gas operators. It validates production data, detects operational anomalies using statistical and machine-learning methods, forecasts production, reports uncertainty, and routes questionable findings to engineers for review.
Overview
WellSense focuses on the quality of operational decisions rather than raw prediction volume. Production readings are validated first, because anomaly detection and forecasting are only as useful as the data they consume.
Detected anomalies and forecasts are presented with uncertainty information and routed for engineering review. The platform is designed to support judgement, not to replace it.
Oil and gas production data is noisy, incomplete, and expensive to interpret manually. Operators need to spot operational anomalies and understand likely production ranges without treating every model output as certainty.
WellSense combines data validation, statistical and machine-learning anomaly detection, forecasting, and uncertainty reporting. Questionable findings are queued for engineer review instead of being silently accepted.
Architecture
A high-level conceptual view of the main stages in this system.
Production measurements and well metadata are collected and normalized.
Missing, duplicated, and out-of-range readings are flagged before modelling.
Statistical and machine-learning methods identify anomalies and project production.
Uncertainty is reported and questionable findings are routed to engineers.
Capabilities that are documented as part of the current implementation.
Flags gaps, duplicates, and implausible values before analysis.
Uses statistical checks and Isolation Forest to surface unusual behavior.
Random Forest models project production rates from historical signals.
Forecasts are presented with ranges rather than a single fixed number.
Questionable findings are routed to a person for confirmation.
Retrieval-augmented context supports interpretation of well behavior.
These items are planned and are not available yet. They are listed separately so completed functionality is never confused with future work.
A structured pilot with an operating partner and real production data.
Support for additional device protocols and telemetry formats.
An anonymized sample for exploring validation and review workflows.
Technology
The tools and technologies associated with the implementation.
Data processing, modelling, and service code.
API for production data and review workflows.
Model training and evaluation utilities.
Unsupervised detection of unusual production behavior.
Forecasting production from historical features.
Storage for well, production, and review records.
Typed database access and schema management.
Retrieval of operational context for interpretation.
Engineer-facing interface for findings and forecasts.
Repeatable local and deployment environments.
Challenges
The engineering constraints that shaped the design.
Sensor gaps and outliers can look like anomalies, so validation and confidence signals must come before detection.
Engineers will stop trusting the system if it raises too many low-value findings, which makes review quality central.
A forecast range must be presented clearly enough for an engineer to use without misreading it as certainty.
Outcomes
Qualitative outcomes from the design. Measured results are only reported where they have been formally evaluated.
Links
Explore the project through the destinations that exist. Missing destinations are labelled instead of being replaced with a placeholder URL.
No public link has been provided for this project yet.
No public link has been provided for this project yet.
No public link has been provided for this project yet.
No public link has been provided for this project yet.
0 of 4 external links are currently available. Unavailable links are shown deliberately rather than pointing to a placeholder destination.
Keep exploring
Browse the full portfolio or start a conversation about this project.