US 10,832,152 B2 · Wet well prediction using real time data

Abstract

Embodiments of systems and methods for predicting when a dry oil producing well is expected to transition to a wet oil producing well, the system configured to generate predictions from multiple databases of information. A first database is designed to store real time data about an active producing well and a second database can be a historical production information database, such as an Online Transactions Processing (OLTP) database. The data from both databases can be normalized, compared, and correlated to create a predictive model that accurately predicts when a particular well is expected to transition from a dry oil producing well to a wet oil producing well. The system and method can be configured to use many different techniques including an autoregressive integrated moving average equation to create a predictive model based off the normalized data from the multiple databases.

Claims18

Dependent claims

A method for predicting when a dry oil producing well is expected to transition to a wet oil producing well, the method comprising:

gathering and storing real-time data from at least one well relating to water cut and storing it in real-time in a database configured to store real-time data;
gathering and storing historical online transaction processing data from the at least one well and storing the historical online transaction processing data in an online transaction processing database;
combining and normalizing the real-time data and the historical online transaction processing data from both databases;
removing outliers from the combined and normalized data;
converting the combined data to time series data;
building a predictive model using an autoregressive integrated moving average for the at least one well that predicts when the at least one well will be producing hydrocarbons with water content greater than one percent based on the converted data;
and
identifying whether the at least one well should be further tested for water and determining when the well will change from dry to wet based on the predictive model.

5 more elements

A system for predicting when a dry oil producing well is expected to transition to a wet oil producing well, the system comprising:

a real-time database configured to store real-time data from at least one well relating to water cut;
an online transaction processing database configured to gather and store historical data from the at least one well;
a computer system configured with a processor and memory, the computer configured to combine and normalize data from the real-time and online transaction processing databases to generate combined and normalized data, remove outliers from the combined and normalized data, convert the combined data to time series data, and build a predictive model using an autoregressive integrated moving average for the at least one well that predicts when the at least one well will be producing hydrocarbons with water content greater than one percent based on the converted data, and identify whether the at least one well should be further tested for water and determining when the well will change from dry to wet based on the predictive model.

Dependencies are inferred from the claim text.

Wet well prediction using real time data

Life of this patent

Wet well prediction using real time data

Life of this patent

Abstract

Claims18