Optimize Unconventional Well and Frac Design with Fast AI Models that Honor Data and Physics
Unique Modeling Technology
Physics + Machine Learning
Fraceon combines physics based well/pad models and state-of-the-art machine learning techniques into a seamless full field model
Run multi objective optimization to optimize well and frac design variables against many objectives like oil production, NPV, etc.
Fast Model Creation
Build and validate predictive models in a matter of days that match all your historical production and fracking data
Easily and quickly update models with latest data, so that the models are always reliable and can be used for operational decision making
Run probabilistic forecasts to estimate P10-P50-P90 production and recovery in minutes for any well, pad or full field
Frac Design Insights
Understand which well and frac design variables (e.g. numbers of stages, proppant concentration, etc.) have the most impact on production
How it Works
In traditional reservoir management, various types of predictive models have been applied over the years for quantitative decision making. Such models range from the very simple analytical models (type-curves, etc.) to the very complex reservoir simulation models. Simulation models attempt to model detailed behavior of reservoir physics and integrate all kinds of measured data. However, many issues such as the significant time and effort required to build and calibrate these models, computational complexity, etc. generally make their practical application difficult.
Tachyus’ Data Physics is the amalgamation of the state-of-the-art in machine learning and reservoir physics. These models can be created as efficiently as machine learning models, integrate all kinds of data, and can be evaluated orders of magnitude faster than full scale simulation models, and since they reservoir physics, they have long term predictive capacity and provide physically realistic results. Data Physics based Fraceon combines many well known machine learning models like Neural Networks, Support Vector Machines, etc. with semi-analytical and numerical well and pad models. Users can go from data preparation and modeling to decisions in a matter of days.