Real-Time Frackability Prediction Using Surface Logging Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Hydraulic fracturing planning in wells is inefficient due to reliance on well log data, which does not provide information about formation fluids, and existing methods are time-consuming and costly, requiring alternative approaches for real-time decision-making.
Innovation Solution
A computer-implemented method using surface logging data, including drilling parameters and mud log lithology, to predict formation frackability in real-time through data quality control, drift analysis, and predictive modeling, allowing for interactive determination of fracking intervals and visualization of frackability probabilities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If well log data is used for hydraulic fracturing planning, then formation porosity and fluid saturation can be measured, but the process is time-consuming and costly
Solution Approach 1:
The system performs preliminary analysis by using drilling parameters and surface logging data that are already collected during the drilling process. The predictive model is trained beforehand on historical data, enabling real-time frackability assessment without waiting for separate well logging operations. This preliminary preparation of data and models eliminates the time-consuming sequential process of traditional well logging followed by analysis.
Solution Approach 2:
The system creates a predictive copy of the formation frackability characteristics by training machine learning models on historical well log data and drilling parameters. Once trained, this digital model can predict frackability using only surface logging data, effectively creating a virtual representation that replaces the need for physical well logging operations while preserving the essential formation characteristics.
2Reliability
If well log data is used for hydraulic fracturing planning, then formation geomechanical properties can be obtained, but additional costs are incurred
Solution Approach 1:
The system makes drilling parameters and surface logging data serve multiple functions: they are used for real-time frackability prediction, formation characterization, and hydraulic fracturing planning. By training predictive models on this multi-purpose data, the system eliminates the need for separate, costly well logging operations while maintaining reliable formation property assessment through the same data collected during routine drilling.
Solution Approach 2:
The system enables the drilling operation itself to provide the data needed for fracturing planning. Drilling parameters and surface logging data collected during the drilling process automatically become the input for frackability prediction, making the drilling operation self-sufficient for generating planning data without requiring additional service operations like well logging.
3Measurement precision
If traditional well logging methods are used, then formation properties can be assessed, but real-time decision-making is not enabled
Solution Approach 1:
The predictive model is trained beforehand on historical well log data and drilling parameters, creating a ready-to-use assessment tool. During drilling operations, the pre-trained model continuously processes incoming surface logging data in real-time, enabling immediate frackability assessment without the delays associated with traditional well logging workflows that require separate operations and manual analysis.
Solution Approach 2:
The system replaces the mechanical well logging process with a computational predictive model. Instead of using physical logging tools to measure formation properties, the system uses machine learning algorithms that process drilling parameters and surface logging data computationally, enabling real-time predictions that substitute the slow mechanical well logging process with fast electronic computation.
Data Source
AI summary
Systems and methods include a computer-implemented method for determining frackability probabilities. Surface logging data of a well being drilled for fracking is accessed. Quality control (QC) and quality assurance (QA) are performed on the surface logging data. Drift analysis is performed on the surface logging data. The surface logging data is prepared for predictive model processing. A predictive model is executed using the prepared surface logging data. A visualization is generated based on executing the predictive model, including information provided for different drilling depths. Fracking intervals are interactively determined using user inputs in the visualization. Frackability probabilities are determined for the fracking intervals and updating the visualization to include a visualization of the fracking intervals.