US US10430725B2

An analysis of U.S. Patent US10430725B2 reveals a system for optimizing oil and gas production using machine learning. As of April 26, 2026, there is no indication of any litigation involving this patent in the 2026 dockets of the U.S. Court of Appeals for the Federal Circuit (CAFC).

Summary of U.S. Patent US10430725B2

Title: Petroleum analytics learning machine system with machine learning analytics applications for upstream and midstream oil and gas industry

Assignee: The original assignee was AKW Analytics Inc. However, records indicate a reassignment on March 27, 2020, to the inventors: Kressner, Arthur; Wu, Leon L.; Anderson, Roger N.; and Xie, Boyi.

Inventors:

• Roger N. Anderson
• Boyi Xie
• Leon L. Wu
• Arthur Kressner

Filing Date: January 18, 2017

Issue Date: October 1, 2019

Abstract:
The patent describes a "Petroleum Analytics Learning Machine (PALM)" system. This system uses machine learning for the analysis of upstream and midstream oil and gas operations. The goal is to optimize exploration, production, and gathering from oil and natural gas fields to maximize output while minimizing costs. The system processes normalized data to find correlations and identify a machine-learned ranking of "importance weights" for various attributes. It employs unique combinations of machine learning and statistical algorithms for predictive and prescriptive optimization. The system also classifies unstructured textual data to identify patterns correlated with optimal production, aiming to capture the dynamics of one or more wells.

Plain-Language Overview of Independent Claims

U.S. Patent US10430725B2 has one independent claim.

Claim 1: This claim outlines a method for optimizing oil and gas field operations. The core of the method involves:

1. Data Collection: Gathering a wide range of structured and unstructured data. This includes geological, geophysical, drilling, hydraulic fracturing (fracking), and production data. The data comes from various sources in real-time and from historical records.
2. Data Processing: The collected data is "cleaned" to remove noise, normalized, and stored in a central database.
3. Machine Learning Analysis: A "Petroleum Analytics Learning Machine (PALM) system," which is a computer-based system, processes the cleaned data. It uses machine learning to identify which factors (attributes) are most important for production. It does this by finding clusters of correlations and assigning "Importance Weights" to each attribute. These weights are then combined with the specific attributes of a given well to identify patterns that can enhance production.
4. Predictive and Prescriptive Optimization: The system uses a unique mix of machine learning models (like support vector machines, decision trees, neural networks, etc.) to make predictions and recommend actions.
5. Unstructured Data Analysis: It analyzes text-based data (like reports and logs) to find additional patterns related to optimal production.
6. User Interface and Action: The system displays its analyses and recommendations on a graphical user interface. These recommendations can be automatically communicated to field systems to guide operations like drilling and fracking in real-time, effectively creating a self-driving or autopilot-like system to improve future production based on detected trends.