Probabilistic Neural Network for CSEM Survey Design
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Reconnaissance controlled-source electromagnetic (CSEM) surveys lack effective design methods due to the absence of specific target information, requiring a new approach for survey design and interpretation that can evaluate and compare different designs on an expected value basis.
Innovation Solution
A method involving simulation of calibration and decision surveys using geologic and economic information to train a classifier algorithm, generating indicators of economic target presence, and calculating expected values for proposed survey designs based on probabilities derived from simulated results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional target-oriented CSEM survey design methods are used, then survey design can be guided by prior information about specific targets, but these methods are inapplicable when no specific target information is available (reconnaissance surveys)
Solution Approach 1:
The patent performs preliminary simulation of multiple possible survey designs before actual field deployment. By simulating calibration surveys and decision surveys with random target properties consistent with general geologic information, the method evaluates expected values and selects optimal survey designs in advance, without requiring specific target information.
Solution Approach 2:
The patent creates simulated survey data that copies the essential characteristics of real CSEM surveys. These simulations include synthetic electromagnetic responses based on general geologic models, allowing the development and testing of classification algorithms without accessing actual target data.
2Productivity
If reconnaissance CSEM surveys are conducted without specific target information, then large areas can be covered at reduced costs, but effective design methods are lacking and interpretation is difficult
Solution Approach 1:
The patent changes the approach from target-specific parameters to general geologic parameters. Instead of designing surveys around known target locations and properties, the method uses probability distributions of target properties and general geologic information to evaluate survey designs based on expected detection capability across large areas.
Solution Approach 2:
The patent implements a feedback loop where simulated survey results are used to train and validate classification algorithms. The simulated calibration surveys provide training data, and simulated decision surveys provide validation, allowing iterative optimization of the interpretation methodology.
3Measurement precision
If simulated calibration surveys are used to train classifier algorithms, then the ability to detect economic targets is improved, but the computational complexity and processing time increase
Solution Approach 1:
The patent performs the computationally intensive simulation and algorithm training in advance, before actual field surveys. By pre-generating simulated calibration surveys and training classifiers offline, the method avoids time-consuming processing during field operations, allowing rapid interpretation of actual survey data.
Solution Approach 2:
The patent uses simulated survey data as a copy substitute for actual survey data in the training phase. This allows the development of classification algorithms without requiring actual field data, enabling parallel processing and optimization without delaying field operations.
4Reliability
If multiple simulated surveys are run to evaluate different survey designs, then the expected value calculation becomes more accurate, but the computational resources and time required increase
Solution Approach 1:
The patent uses a sufficient number of simulated surveys to achieve statistically reliable expected value estimates without performing exhaustive simulations of all possible survey designs. By evaluating a representative sample of candidate designs with multiple simulations each, the method achieves adequate precision without prohibitive computational cost.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables the evaluation and comparison of different survey designs, optimizing the ability to detect economic targets while minimizing false positives, thereby reducing exploration costs and uncertainty.
Implementation Method 1
solving Maxwell's field equations to develop a sensitivity map database, each map giving an anomalous electromagnetic response at a central receiver location for a representative array of nearby target positions
Data Source
AI summary
Method for determining an expected value for a proposed reconnaissance electromagnetic (or any other type of geophysical) survey using a user-controlled source. The method requires only available geologic and economic information about the survey region. A series of calibration surveys are simulated with an assortment of resistive targets consistent with the known information. The calibration surveys are used to train pattern recognition software to assess the economic potential from anomalous resistivity maps. The calibrated classifier is then used on further simulated surveys of the area to generate probabilities that can be used in Value of Information theory to predict an expected value of a survey of the same design as the simulated surveys. The calibrated classifier technique can also be used to interpret actual CSEM survey results for economic potential.