Defect Recognition in Concrete Ultrasonic Detection Based on Wavelet Packet Transform and Stochastic Configuration Networks

Zhao, Jinhui; Hu, Tianyu; Zheng, Ruifang; Ba, Penghui; Mei, Congli; Zhang, Qichun

doi:10.1109/access.2021.3049448

Cited by 24 publications

(17 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The response surface optimization design method enables horizontal optimization analysis to be continuously carried out for multiple influencing factors. It can overcome the defects of the control variable method and the orthogonal test method, which can only optimize the design and analysis of each isolated point [ 29 , 30 ]. At the same time, within a certain range, it has the advantages of fewer test time, an accurate fitting equation, good prediction performance and can fully consider the interaction between different factors.…”

Section: Response Optimization Design Methodsmentioning

confidence: 99%

Research on optimization of perforation parameters for formation fractures based on response surface optimization method

et al. 2021

View full text Add to dashboard Cite

For staged multi-cluster fracturing, methods for controlling perforation friction to adjust the flow distribution of each cluster can effectively promote the uniform extension of multiple fractures but lacks a fast and quantitative optimization method for different perforation parameters of each cluster. By establishing a numerical model of single-stage three-cluster flow-limited fracturing under stress-seepage coupling, and based on the response surface optimization method, fully considering the impact of perforation parameters interaction among three perforation clusters, according to the regression equation fitted under the global response, the rapid optimization of perforation parameters of segmented multi-cluster fracturing model is realized. The results show that: in determining the three factors of the study, it is found that there is an obvious interaction between the number of intermediate cluster perforations and the number of cluster perforations on both sides, the number of cluster perforations on both sides and the diameter of intermediate cluster perforations, the response surface optimization method gives the optimal perforation parameter combination of three clusters of fractures under global response; When the perforation parameters were combined before optimization, the fracture length difference was 32.550m, and the intermediate perforation cluster evolved into invalid perforation cluster, when the perforation parameters were combined after optimization, the fracture length difference was 0.528m, the three perforation clusters spread uniformly, and there are no invalid clusters. At the same time, the regression equation under the response is optimized before and after the comparison between the predicted value of the equation and the actual simulation value. It is found that the estimated deviation rate of the equation before optimization is 1.2%, and the estimated deviation rate after optimization is 0.4%. The estimated deviation rates are all less, and the response regression equation based on the response surface optimization method can quickly optimize the perforation parameters. The response surface optimization method is suitable for the multi parameter optimization research of formation fracturing which is often affected by many geological and engineering factors. Combining with the engineering practice and integrating more factors to optimize the hydraulic fracturing parameters, it is of great significance to improve the success rate of hydraulic fracturing application.

show abstract

Section: Response Optimization Design Methodsmentioning

confidence: 99%

Research on optimization of perforation parameters for formation fractures based on response surface optimization method

et al. 2021

View full text Add to dashboard Cite

show abstract

“…The continuous improvement of computing power has made data-driven deep learning methods appear on the stage of history, bringing about a series of excellent algorithms. Deep learning-based methods show superior performance over traditional algorithms in many computer vision fields ( Hu et al, 2021 ; Liu et al, 2021b ; Teng et al, 2021 ; Zhao et al, 2021 ). The current mainstream object detection methods can be classified into Anchor-based methods and Anchor-free methods according to whether candidate regions are generated, with the main difference being whether target regression and classification are performed by predefined anchor frames.…”

Section: Related Workmentioning

confidence: 99%

Lightweight and efficient neural network with SPSA attention for wheat ear detection

Dong

Liu

Kang

et al. 2022

PeerJ Computer Science

View full text Add to dashboard Cite

Advancements in deep neural networks have made remarkable leap-forwards in crop detection. However, the detection of wheat ears is an important yet challenging task due to the complex background, dense targets, and overlaps between wheat ears. Currently, many detectors have made significant progress in improving detection accuracy. However, some of them are not able to make a good balance between computational cost and precision to meet the needs of deployment in real world. To address these issues, a lightweight and efficient wheat ear detector with Shuffle Polarized Self-Attention (SPSA) is proposed in this paper. Specifically, we first utilize a lightweight backbone network with asymmetric convolution for effective feature extraction. Next, SPSA attention is given to adaptively select focused positions and produce a more discriminative representation of the features. This strategy introduces polarized self-attention to spatial dimension and channel dimension and adopts Shuffle Units to combine those two types of attention mechanisms effectively. Finally, the TanhExp activation function is adopted to accelerate the inference speed and reduce the training time, and CIOU loss is used as the border regression loss function to enhance the detection ability of occlusion and overlaps between targets. Experimental results on the Global Wheat Head Detection dataset show that our method achieves superior detection performance compared with other state-of-the-art approaches.

show abstract

“…Daubechies (db) wavelet function has good compact support, approximate symmetry, and smoothness [ 29 ]. Different orders of db wavelet decomposition have been applied to detection signals, and db15 is still selected by the trial-and-error method in WPT sub-algorithm to decompose and reconstruct the ultrasonic detection signals [ 12 , 24 ].…”

Section: Algorithm Descriptions and Basic Stepsmentioning

confidence: 99%

“…Hu [ 11 ] used a genetic algorithm optimized back propagation neural network to identify the ultrasonic signals of concrete defects and the average recognition accuracy is 91%. Zhao [ 12 ] used the stochastic configuration network to identify the ultrasonic detection signals of concrete defects and the accuracy is around 95%. Their experimental data were collected from concrete blocks of C30 class containing hole defects using the ultrasonic testing system [ 13 ].…”

Section: Introductionmentioning

confidence: 99%

A Wavelet Packet Transform and Convolutional Neural Network Method Based Ultrasonic Detection Signals Recognition of Concrete

Zhao

Zhang

2022

Sensors

Self Cite

View full text Add to dashboard Cite

This paper proposes a new intelligent recognition method for concrete ultrasonic detection based on wavelet packet transform and a convolutional neural network (CNN). To validate the proposed data-based method, a case study is presented where the K-fold cross-validation was adopted to produce the performance analysis and classification experiments. Moreover, three evaluation indicators, precision, recall, and F-score, are calculated for analyzing the classification performance of the trained models. As a result, the obtained four-classifying CNN reaches more than 99% detection accuracy while the lowest recognition accuracy is not less than 92.5% on the testing dataset for the six-classifying CNN model. Compared with the existing stochastic configuration network (SCN) models, the presented method achieves the design objective with better recognition performance. The calculation results of the six-classifying and five-classifying models and related research clearly indicate the remaining challenging tasks for intelligent recognition algorithms in extracting features and classifying mass data from various concrete defects precisely and efficiently.

show abstract

Defect Recognition in Concrete Ultrasonic Detection Based on Wavelet Packet Transform and Stochastic Configuration Networks

Cited by 24 publications

References 25 publications

Research on optimization of perforation parameters for formation fractures based on response surface optimization method

Research on optimization of perforation parameters for formation fractures based on response surface optimization method

Lightweight and efficient neural network with SPSA attention for wheat ear detection

A Wavelet Packet Transform and Convolutional Neural Network Method Based Ultrasonic Detection Signals Recognition of Concrete

Contact Info

Product

Resources

About