Acoustic Emission Onset Time Detection via Histogram Distance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting the onset time of acoustic emission signals face challenges due to background noise and low signal-to-noise ratios, particularly in determining the adequate length of the time window, leading to low accuracy.
Innovation Solution
A method based on histogram distance is introduced, where the time-domain signal is divided into intervals by a sliding point, and relative frequency histograms are calculated to determine the onset time, maximizing the histogram distance, thus avoiding the need for a fixed time window and improving accuracy in low SNR conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the Akaike information criterion (AIC) method is used to detect onset time, then the detection can be performed with a fixed time window, but the accuracy is low when processing signals with severe background noise
Solution Approach 1:
The patent transforms the time-domain signal into a frequency-domain histogram representation, changing the parameter domain from time to frequency distribution. This transformation allows the method to capture signal characteristics that are invariant to time shifts and more robust to background noise, thereby improving detection accuracy in noisy environments without relying on fixed time window parameters
2Measurement precision
If the energy ratio (ER) method is used to detect onset time, then the detection can be performed, but the result is dependent on the length of the time window and not suitable for low signal-to-noise ratios
Solution Approach 1:
The patent segments the signal analysis into two distinct parts: the pre-event signal segment represented by histogram h1 and the post-event signal segment represented by histogram h2. By comparing these segmented histograms, the method eliminates the need for arbitrary time window selection and provides adaptability across different signal-to-noise ratio conditions, as the segmentation is driven by statistical properties rather than fixed temporal boundaries
3Ease of operation
If a fixed time window is used in onset time detection, then the detection process is simplified, but the accuracy decreases when the signal-to-noise ratio is low
Solution Approach 1:
The patent replaces the mechanical time-window-based detection approach with a statistical histogram-based approach. Instead of manually selecting and adjusting time window parameters, the method uses automatic histogram construction and comparison, substituting the mechanical parameter-tuning process with an automated statistical analysis that maintains both simplicity and high accuracy across varying signal conditions
Data Source
AI summary
The present invention discloses a method for onset time detection of a time-domain acoustic emission signal based on histogram distance. The method comprises the following steps: acquiring an acoustic emission signal; dividing the signal into two intervals with a sliding point k as the demarcation point; obtaining the relative frequency histograms of two adjacent intervals; obtaining histogram distance of the relative frequency histograms of two adjacent intervals; moving the sliding point k to the next element to obtain two new intervals and generating new histograms of the two new intervals and calculating the histogram distance of two new intervals; searching for the point which gives the maximum value of the histogram distances, and the corresponding time to this point is regarded as the onset time.