AI-Augmented N-Value Prediction for Sparse Ground Investigation Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for identifying N-values in ground characteristics face challenges due to limited drilling data, leading to inaccurate pile design and increased costs and delays, particularly when empirical judgments are relied upon for undrilled points.
Innovation Solution
A method and apparatus using artificial intelligence-based data augmentation to predict N-values by generating hypothetical points radially around actual locations, incorporating circular augmentation to enhance data sets and improve accuracy through machine learning techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If limited drilling data is used due to time and cost restrictions, then drilling investigation speed is improved, but N-value prediction accuracy deteriorates
Solution Approach 1:
The patent creates synthetic copies of actual drilling data points by generating hypothetical N-values at multiple radial distances from each actual measurement point. This data augmentation technique replicates the limited actual data into numerous virtual data points, effectively increasing the dataset size without requiring additional physical drilling operations, thus maintaining high prediction accuracy while working with limited field data
Solution Approach 2:
The patent transforms the limited 2D spatial drilling data into a 3D augmented dataset by adding radial distance dimensions. From each actual drilling point, multiple hypothetical points are generated at different radial distances, creating a multi-dimensional data structure that enriches the information content and improves model training effectiveness despite the limited number of actual measurements
2Measurement precision
If more drilling data is collected to improve N-value prediction accuracy, then measurement precision is improved, but time and cost increase
Solution Approach 1:
Instead of conducting additional physical drilling operations to gather more data, the patent creates virtual copies of existing data points through mathematical modeling. Each actual drilling measurement generates multiple hypothetical data points at different radial distances, multiplying the information value from each actual measurement without requiring additional time or resources for field operations
Solution Approach 2:
The patent performs preliminary data augmentation during the drilling investigation phase by generating hypothetical N-values before final prediction is needed. This preliminary creation of augmented data allows the machine learning model to be trained on a comprehensive dataset early in the process, eliminating the need for extensive post-drilling data collection or iterative field investigations
3Device complexity
If empirical judgment is used for undrilled points, then device complexity is reduced, but N-value prediction accuracy deteriorates
Solution Approach 1:
The patent replaces engineer empirical judgment with machine learning models trained on augmented data. The system creates virtual training data by copying and transforming actual drilling measurements into multiple hypothetical data points, which then serve as the foundation for training accurate prediction models that automatically replace subjective human judgment with objective, data-driven algorithms
Solution Approach 2:
The patent substitutes the mechanical process of physical drilling and manual engineering judgment with an automated information processing system. Instead of relying on engineers to manually interpret limited drilling data and make empirical predictions, the system automatically generates augmented data and trains machine learning models to produce accurate N-value predictions for undrilled points
Data Source
Figure 1
Figure 2
Figure 3~4A
AI summary
The N-value prediction apparatus according to the present invention includes a hypothetical learning data augmentation unit, based on an actual N-value measured at an actual location according to the Standard Penetration Test through drilling investigation, generating at least one of hypothetical N-values corresponding to a preset hypothetical point based on the actual location, an N-value prediction model learning unit learning ground characteristic data corresponding to each of the actual location and the hypothetical point by artificial intelligence, the ground characteristic data including the actual N-value and the hypothetical N-values, and an N-value prediction result calculation unit predicting an N-value at an arbitrary prediction target point by using an N-value prediction model generated by artificial intelligence learning executed in the N-value prediction model learning unit.