Adaptive Hybrid Borehole Wave Processing for Slowness Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing acoustic logging methods face challenges in accurately processing dispersive acoustic waves in complex borehole environments, as model-based approaches are biased by uncontrolled parameters and data-driven methods produce inaccurate results with low-quality data, leading to inconsistent slowness curve quality.
Innovation Solution
An adaptive hybrid processing technique that automatically determines and alternates between model-based and data-driven methods using supervised, physics-based machine learning to assess data quality and derive scale factors for hybrid processing, combining waveform data features and model constraints to improve slowness curve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If model-based processing is used to extract body wave slowness from dispersion waves, then a smooth slowness curve can be generated even with poor data quality, but the curve becomes strongly biased due to unaccounted factors such as anisotropy and irregular borehole shape
Solution Approach 1:
The patent introduces a machine learning model as an intermediary that learns the complex relationship between dispersion wave data and body wave slowness from training data. This intermediary captures the effects of anisotropy, irregular borehole shape, and other unaccounted factors without requiring explicit physical models, thereby resolving the contradiction between smoothness and accuracy by learning from data patterns rather than relying on simplified assumptions
Solution Approach 2:
The patent transforms the problem from directly modeling the complex physical relationships to learning parameter mappings from data. By changing from a physics-based parameter model to a data-driven parameter learning approach, the system can adapt to varying borehole conditions and produce accurate slowness estimates without being constrained by incomplete physical understanding
2Measurement precision
If data-driven processing is used to extract body wave slowness from dispersion waves, then accurate values can be generated from high quality data, but the slowness curve becomes very inaccurate as data quality decreases
Solution Approach 1:
The patent employs a dynamic approach where the machine learning model adapts its behavior based on input data quality. The system can switch between relying more on data-driven predictions when data quality is high and incorporating more regularization or alternative processing when data quality deteriorates, thereby maintaining both accuracy and reliability across varying data conditions
Solution Approach 2:
The system incorporates feedback mechanisms where the quality of input dispersion wave data is assessed and used to adjust the processing approach. By monitoring data quality metrics and adjusting the model's confidence or switching to alternative processing strategies when quality is poor, the system maintains reliable and accurate slowness curves across varying data conditions
3Reliability
If solely model-based or data-driven approaches are used with lower quality sensors or sensors in complex environments, then variable consistency in quality of curves is produced, but using a hybrid approach can improve consistency
Solution Approach 1:
The patent merges model-based and data-driven processing approaches into a unified hybrid system. The machine learning model combines the advantages of both approaches: the physical constraints and smoothness from model-based methods and the accuracy from data-driven methods. This integration resolves the contradiction by achieving consistent reliability while managing complexity through a unified processing framework
Data Source
AI summary
Systems and methods are provided for determining a formation body wave slowness from an acoustic wave. Waveform data is determined by logging tool measuring the acoustic wave. Wave features are determined from the waveform data and a model is applied to the wave features to determine data-driven scale factors The data-driven scale factors can be used to determine a body wave slowness within a surrounding borehole environment and the body wave slowness can be used to determine formation characteristics of the borehole environment.