Machine Learning Inversion Bandpass Filtering Borehole Stability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional machine learning methods are not well-suited for mapping physical responses to material properties in wellbore environments due to instability issues, such as non-one-to-one correspondences and incompatible data bandwidths between sensed and calculated data, leading to inaccurate material property identification.
Innovation Solution
Implementing a machine learning model with bandpass filtering to constrain its output to match the expected bandwidth of sensed data, ensuring a one-to-one mapping of physical responses to material properties, and using recursive processes and gradient calculations to refine the model's evaluations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If machine learning is used to map physical responses to material properties, then speed of material property identification is improved, but mapping stability deteriorates due to non-one-to-one correspondences and incompatible data bandwidths
Solution Approach 1:
The patent applies parameter changes by transforming the machine learning model's output through bandpass filtering to match the bandwidth characteristics of the sensed data. This involves adjusting the frequency parameters of the modeled physical responses to be compatible with the sensor measurements, thereby achieving stable one-to-one mapping while maintaining fast identification speeds.
Solution Approach 2:
The patent introduces bandpass filtering as an intermediary step between the machine learning model and the material property determination. This filtering mechanism acts as a mediator that reconciles the bandwidth incompatibility between computed and sensed data, enabling stable correspondence without sacrificing the speed advantage of machine learning.
2Productivity
If machine learning models are used for inversion, then productivity is improved, but measurement precision deteriorates due to incompatible data bandwidths
Solution Approach 1:
The patent modifies the output parameters of the machine learning model by applying bandpass filtering that aligns the frequency content with the sensed data bandwidth. This parameter adjustment ensures that the modeled physical responses are comparable to actual measurements, thereby maintaining high productivity while improving measurement precision.
Solution Approach 2:
The patent implements a feedback mechanism where the modeled data is compared with sensed data, and the bandpass filtering is applied iteratively to refine the model output. This feedback loop ensures that the material property determinations are both efficient and precise by continuously adjusting the model predictions to match observed measurements.
Data Source
AI summary
The present disclosure provides techniques to identify material properties from measured physical response data in ways that are more efficient than conventional numerical inversion. Systems and techniques of the present disclosure may use machine learning (ML) techniques to generate mappings that map specific physical responses to specific material properties based on knowledge of factors that may be inherent to certain types of sensing equipment. Data generated by a ML computer model may be constrained, altered, or filtered to more closely correspond to characteristics known to be associated with a certain type of sensing equipment. Such a constrained model may identify fewer possible results as compared to an unconstrained model, thereby, solving problems that may be encountered when using ML techniques to identify material properties from measured responses.