Condensed Damage Index Factors for Structural Health Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Structural Health Monitoring (SHM) systems face challenges in accurately evaluating damage in structures due to the need for a large number of training records and inputs for neural networks, which can be cost-prohibitive and lead to unreliable damage prediction.
Innovation Solution
A method and apparatus that determine and select condensed damage index factors from signal information before and after damage, correlating them with measured dimensions to identify suitable input parameters for algorithmic evaluation, reducing the number of inputs required for accurate damage assessment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a large number of training records are used to train the neural network for damage evaluation, then the reliability of damage prediction is improved, but the cost and complexity of the system increases significantly
Solution Approach 1:
The patent transforms the raw signal data into a reduced set of damage index parameters through mathematical processing. By changing the form of the data from complex time-domain signals to simplified frequency-domain features and damage indices, the system achieves reliable damage prediction with fewer training records, thereby reducing system complexity while maintaining accuracy
Solution Approach 2:
The patent extracts essential damage-related features from complex structural health monitoring signals by calculating damage indices based on frequency shifts between healthy and damaged states. This extraction process isolates the most relevant information for damage detection, enabling reliable predictions with reduced data dimensions and fewer training records
2Measurement precision
If many input factors are provided to the neural network, then comprehensive damage evaluation is achieved, but the number of training records required increases making the system cost-prohibitive
Solution Approach 1:
The patent applies parameter transformation by converting complex signal data into a small set of meaningful damage indices through frequency analysis. This parameter reduction maintains comprehensive damage evaluation capability while significantly decreasing the quantity of training data needed, making the system cost-effective
Solution Approach 2:
The patent segments the complex damage evaluation problem into distinct components: signal acquisition, frequency analysis, damage index calculation, and neural network classification. This segmentation allows each component to be optimized independently, achieving accurate damage evaluation with minimal training records by focusing computational resources on the most discriminative features
3Ease of manufacture
If a small number of training records are used, then the system cost is reduced, but the reliability and accuracy of damage prediction deteriorates
Solution Approach 1:
The patent applies mathematical transformations to convert raw signals into condensed damage indices that capture essential damage information in a compact form. This parameter condensation enables the system to achieve reliable damage prediction with far fewer training records, reducing system cost while maintaining accuracy
Solution Approach 2:
The patent performs preliminary signal processing and damage index calculation before the neural network training phase. By pre-processing the signals to extract meaningful damage features, the system reduces the information burden on the neural network, enabling effective training with minimal records and lowering overall system cost
Data Source
AI summary
A method for creating at least one input parameter for an algorithmic system to evaluate damage in a structure may include: (a) Determining a plurality of damage index factors using first signal information relating to a first signal transmitted through the structure before the damage is imposed, and second signal information relating to a second signal transmitted through the structure after the damage is imposed. (b) determining a plurality of condensed damage index factors using the plurality of damage index factors. (c) Correlating selected of the condensed damage index factors with selected measured dimensions relating to the damage to determine a correlation index for selected combinations of the condensed damage index factors and the dimensions. (d) Selecting the at least one input parameter from among the selected condensed damage index factors having a correlation index meeting at least one predetermined criterion.