Probabilistic Look-Up Tables for Drilling Automation Uncertainty
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automation solutions in oil and gas exploration face challenges due to the uniqueness of each well, varying rig equipment, and the reluctance of stakeholders to share data, leading to difficulties in developing reliable and transparent control algorithms that can handle uncertainties in sensed data.
Innovation Solution
A method utilizing look-up tables representing all models in an automation control architecture, storing data as conditional probability tables or distributions from multiple stakeholders, and converting them into visually displayed look-up tables to handle uncertainties and facilitate data integration and control algorithm development.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If automation control algorithms are developed for each unique well and rig configuration, then control reliability and safety are improved, but development time and complexity increase significantly
Solution Approach 1:
The patent segments the control algorithm into modular components: a probabilistic data processing layer that handles uncertainties independently, and a control decision layer. This segmentation allows the probabilistic handling module to be developed and validated separately, then applied across multiple wells without redeveloping the entire algorithm, thus improving reliability while reducing development time.
Solution Approach 2:
The patent changes the parameter representation from deterministic values to probabilistic distributions. By representing sensed data as probability distributions and using look-up tables that map input distributions to output distributions, the system handles uncertainties systematically. This parameter transformation allows the same algorithm structure to handle varying well conditions and sensor qualities without requiring complete redevelopment for each case.
2Measurement precision
If strict control on hardware is implemented for algorithm execution, then measurement precision and control accuracy are improved, but adaptability to different rigs decreases
Solution Approach 1:
The patent transforms precise but rigid hardware requirements into flexible probabilistic parameter representations. Instead of requiring exact sensor specifications, the system represents sensor outputs as probability distributions with varying means and standard deviations. This allows the same control algorithm to adapt to different sensor qualities and rig configurations while maintaining control accuracy through the probabilistic framework.
Solution Approach 2:
The patent creates a universal probabilistic data processing module that can handle multiple sensor types, rig configurations, and uncertainty sources through a unified interface. This module serves multiple functions: validating incoming data, characterizing uncertainties, propagating probability distributions through the control model, and generating control decisions. The universal design eliminates the need for separate hardware control specifications for each rig type.
3Reliability
If comprehensive data integration from all stakeholders is performed, then control algorithm completeness is improved, but data integration complexity and time requirements increase
Solution Approach 1:
The patent transforms the data integration problem by changing from integrating raw data values to integrating probability distribution parameters. Each stakeholder contributes data represented as probability distributions (with mean and standard deviation), and the system combines these distributions mathematically rather than manually reconciling conflicting values. This parameter-based integration reduces complexity while maintaining completeness.
Solution Approach 2:
The patent introduces a probabilistic data processing module as an intermediary between raw stakeholder data and the control algorithm. This intermediary module standardizes incoming data from multiple sources into a unified probabilistic format, handles validation and uncertainty characterization, and presents processed distributions to the control logic. This intermediary layer simplifies the integration process by providing a consistent interface that abstracts the complexity of multi-source data reconciliation.
Data Source
AI summary
A method, system and computer program product for utilizing look-up tables representing all models in an automation control architecture to independently handle uncertainties in sensed data. Data is stored in a form of conditional probability tables (CPTs) or conditional probability distributions (CPDs), where the data comes from an operator, a service provider, a drilling contractor and an equipment manufacturer. Models of the drilling process domains, such as wellbore hydraulics, drill bit/rock interactions, torque and drag modeling, vibration modeling and drilling machinery operation, are received. Data is extracted from these models into the CPTs or CPDs. The CPTs or CPDs are converted to look-up tables. Data in the look-up tables are then visually displayed in graphical form. As a result, real-time troubleshooting of drilling operations occurs in an efficient manner.