Mud Gas Analysis for Fluid Type Prediction in Geosteering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current geosteering methods face challenges in accurately distinguishing between different types of hydrocarbon fluids, such as oil, gas condensate, and gas, due to insufficient contrast in resistivity-based measurements, which can lead to invalid assumptions about fluid composition and compartmentalization in reservoirs.
Innovation Solution
A method involving the measurement of mud gas properties at the surface of a wellbore, using pre-trained machine-learning models to predict fluid properties such as fluid type, composition, and gas-oil ratio, and re-training these models based on comparisons with measured properties to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If resistivity-based measurements are used to visualize geological structure, then real-time information for well placement is available, but sufficient contrast to distinguish between different types of hydrocarbon fluids is not achieved
Solution Approach 1:
The patent introduces mud gas analysis as an intermediary measurement method between the drill bit and surface laboratory analysis. The mud gas samples carry fluid information from the reservoir to the surface, where GC/MS instruments analyze the gas composition to distinguish between different hydrocarbon fluid types, thereby mediating the information transfer without direct contact between the measurement tool and the formation fluids
Solution Approach 2:
The patent replaces the mechanical/resistivity-based measurement system with a chemical analysis system. Instead of using resistivity logs that rely on electrical properties, the system uses gas chromatography and mass spectrometry to analyze the chemical composition of mud gas, substituting a chemical detection mechanism for the electrical measurement approach
2Measurement precision
If downhole fluid samples are collected and analyzed in the fluid laboratory, then reliable fluid information is obtained, but the process is time-consuming, taking weeks to months
Solution Approach 1:
The patent performs preliminary fluid analysis by collecting and analyzing mud gas samples continuously during the drilling process. The GC/MS instruments at the surface analyze the gas composition in real-time or near real-time, providing preliminary fluid characterization data before the well is completed, thereby avoiding the need for time-consuming post-drilling laboratory analysis
Solution Approach 2:
The patent skips the traditional multi-step process of coring, retrieving, and transporting formation samples to the laboratory by directly analyzing mud gas that has already been brought to the surface with the drilling fluid. This rushes through the fluid analysis process by eliminating intermediate steps and providing results within hours rather than weeks or months
3Loss of information
If a pilot well is drilled to characterize the fluid, then fluid composition information is obtained, but the assumption that fluid composition remains unchanged may be invalid when faults and barriers create compartmentalization
Solution Approach 1:
The patent implements continuous fluid characterization by analyzing mud gas composition throughout the drilling process rather than relying on a single pilot well measurement. The continuous monitoring provides ongoing verification of fluid composition, allowing detection of changes that may indicate fault intersections or compartmentalization, thereby maintaining the reliability of fluid information throughout the well trajectory
Data Source
AI summary
A method for predicting fluid properties includes measuring one or more measured fluid properties of a mud gas at a surface of a wellbore. The method also includes measuring one or more mud gas properties of the mud gas at the surface of the wellbore. The method also includes predicting one or more first predicted fluid properties using one or more pre-trained machine-learning (ML) models. The one or more first predicted fluid properties are predicted based at least partially upon the one or more mud gas properties. The method also includes comparing the one or more measured fluid properties to the one or more first predicted fluid properties. The method also includes re-training the one or more pre-trained ML models to produce one or more re-trained ML models in response to the comparison.