Damage Localization in Pressure Vessel by Guided Waves Based on Convolution Neural Network Approach

2020 ◽  
Vol 142 (6) ◽  
Author(s):  
Chaojie Hu ◽  
Bin Yang ◽  
Jianjun Yan ◽  
Yanxun Xiang ◽  
Shaoping Zhou ◽  
...  
Keyword(s):  
Neural Network ◽  
Pressure Vessel ◽  
Guided Waves ◽  
Pressure Vessels ◽  
Guided Wave ◽  
Damage Localization ◽  
Accuracy Rate ◽  
Neural Network Approach ◽  
Monitoring Process ◽  
Fully Connected

Abstract This paper investigates the damage localization in a pressure vessel using guided wave-based structural health monitoring (SHM) technology. An online SHM system was developed to automatically select the guided wave propagating path and collect the generated signals during the monitoring process. Deep learning approach was employed to train the convolutional neural network (CNN) model by the guided wave datasets. Two piezo-electric ceramic transducers (PZT) arrays were designed to verify the anti-interference ability and robustness of the CNN model. Results indicate that the CNN model with seven convolution layers, three pooling layers, one fully connected layer, and one Softmax layer could locate the damage with 100% accuracy rate without overfitting. This method has good anti-interference ability in vibration or PZTs failure condition, and the anti-interference ability increases with increasing of PZT numbers. The trained CNN model can locate damage with high accuracy, and it has great potential to be applied in damage localization of pressure vessels.

Defect Localization for Pressure Vessel Based on Circumferential Guided Waves: An Experimental Study

2018 ◽  
Author(s):  
Shuangmiao Zhai ◽  
Chaofeng Chen ◽  
Gangyi Hu ◽  
Shaoping Zhou
Keyword(s):  
Pressure Vessel ◽  
Guided Waves ◽  
Pressure Vessels ◽  
Guided Wave ◽  
Normal Operation ◽  
Propagation Characteristics ◽  
Dense Array ◽  
Imaging Algorithm ◽  
Sparse Array ◽  
Damage Imaging

Pressure vessels are normally employed under extreme environments with high temperature and high pressure. Inevitably, the defects like crack and corrosion that easily occur in the equipment and can significantly influence the normal operation. Guided wave-based method is a cost-effective means to measure the utility of pressure vessel. In this paper, finite element (FE) simulation is used to explore the propagation characteristics of circumferential guided waves in pressure vessel. Based on the propagation characteristics, the experiments with different configurations of piezoelectric transducers (PETs), which contain a sparse array and a dense array, have been conducted on pressure vessel respectively. Different imaging methods, including discrete ellipse imaging algorithm and probability damage imaging algorithm have been applied to locate the defect based on the configurations above. Furthermore, a multi-channel ultrasonic guided wave detection system has been set up for pressure vessel inspection. The experimental results show that the sparse array with the discrete ellipse imaging algorithm can locate the defect effectively. The imaging results based on probability damage imaging algorithm show that the dense array presents the better localization result.

Propagation of Guided Waves in Pressure Vessel and its Application for Damage Detection

2012 ◽  
Author(s):  
Xuewei Sun ◽  
Fucai Li ◽  
Jinfu Wang ◽  
Guang Meng ◽  
Limin Zhou
Keyword(s):  
Structural Health Monitoring ◽  
Health Monitoring ◽  
Pressure Vessel ◽  
Guided Waves ◽  
Pressure Vessels ◽  
Guided Wave ◽  
Special Equipment ◽  
Element Analysis ◽  
Structural Health ◽  
Monitoring Technique

Pressure vessel is a kind of special equipment with explosion and leakage dangerous. Therefore, structural health monitoring (SHM) techniques for pressure vessel should ensure the safe operation of this kind of equipments and is becoming more crucial in petroleum, chemical, and relative industries. Guided wave-based structural health monitoring technique can be an appropriate method for real-time and online non-destructive damage monitoring technique. In recent years, applications of guided wave-based structural health monitoring techniques are mainly limited in simple structures, such as plates and tubes. Relatively few research papers focused on the application of this technique in large and complex structures like pressure vessels. Propagation characteristics of guided waves in pressure vessel are investigated in this study. Dispersion curves calculated by using numerical methods for longitudinal, circumferential, and torsional modes are presented. On the basis of comprehensive analysis of the guided waves dispersion and experimental waveforms, the parameters of the excitation wave are therefore optimized. In order to overcome the difficulties to identify the damage characteristics of signal, the layout scheme of sensor network is designed and optimized in this paper to simplify the waveform. Furthermore, both finite element analysis (FEA) and experiment methods are employed to investigate the propagation of elastic guided waves in a standard pressure vessel.

3D-3 Classification of Defects for Guided Waves Inspected Pipes by a Neural Network Approach

2007 ◽  
Author(s):  
G. Acciani ◽  
G. Brunetti ◽  
G. Fornarelli ◽  
F. Bertoncini ◽  
M. Raugi ◽  
...  
Keyword(s):  
Neural Network ◽  
Guided Waves ◽  
Network Approach ◽  
Neural Network Approach

scholarly journals A Customized Deep Neural Network Approach to Investigate Travel Mode Choice with Interpretable Utility Information

2020 ◽  
Vol 2020 ◽  
pp. 1-11
Author(s):  
Zhengchao Zhang ◽  
Congyuan Ji ◽  
Yineng Wang ◽  
Yanni Yang
Keyword(s):  
Neural Network ◽  
Learning Community ◽  
Mode Choice ◽  
Feature Representation ◽  
Classification Problems ◽  
Neural Network Approach ◽  
Nonlinear Fitting ◽  
Linear Utility ◽  
Utility Models ◽  
Fully Connected

Discrete choice modeling of travel modes is an essential part of traffic planning and management. Thus far, this field has been dominated by multinomial logit (MNL) models with a linear utility specification. However, deep neural networks (DNNs), owing to their powerful capacity of nonlinear fitting, are now rapidly replacing these models. This is because, by using DNNs, mode choice can be assimilated with the classification problems within the machine learning community. This article proposes a newly designed DNN framework for traffic mode choice in the style of two hidden layers. First, a local-connected layer automatically extracts an effective utility specification from the available data, and then, a fully connected layer augments the feature representation. Validated by a practical city-wide multimodal traffic dataset in Beijing, our model significantly outperforms the random utility models and simple fully connected neural network in terms of the prediction accuracy. Besides the comparison of the predictive power, we also present the interpretability of the proposed model.

Burst Pressure Prediction of Cylindrical Vessels Using Artificial Neural Network

2020 ◽  
Vol 142 (3) ◽  
Author(s):  
Abolfazl Zolfaghari ◽  
Moein Izadi
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Pressure Vessel ◽  
Pressure Vessels ◽  
Machine Learning Techniques ◽  
Burst Pressure ◽  
Outer Diameter ◽  
Wide Range ◽  
Artificial Neural ◽  
Predictor Model

Abstract Pressure vessel plays an important role in wide range of applications to store gas or liquid substances. In order to design a pressure vessel safely, one of the main factors which has to be considered is selection of proper burst pressure perdition criterion. Due to large range of available materials in manufacturing of the vessels under different working conditions, several criteria to forecast burst pressure of the vessels have been developed and used by designers. Choosing the most proper criterion based on working condition and the material is a vital task to meet design requirements because inappropriate criterion may lead to unsafe vessel or over design. This issue makes not only pressure vessel design more complex but also maintenance planning, especially for designers who do not have enough experience, is a challenging task. Therefore, lack of a burst pressure predictor model, which is able to determine the pressure more accurately for wide range of materials and applications, has been remained unsolved. To evaluate machine learning techniques in prediction of burst pressure of pressure vessels, in this paper, a new model based on artificial neural network (ANN) has been proposed and developed. Input parameters of the model include internal and outer diameter, thickness, ultimate and yield strength; output is burst pressure. The obtained results showed that the constructed model has a good potential to be used as more applicable model compared to current models in design of pressure vessels.

A Study on Defect Tomographic Imaging in Pressure Vessel With Liquid Medium

2020 ◽  
Vol 142 (4) ◽  
Author(s):  
Gangyi Hu ◽  
Chaofeng Chen ◽  
Shaoping Zhou ◽  
Shuangmiao Zhai
Keyword(s):  
Pressure Vessel ◽  
Reconstruction Algorithm ◽  
Pressure Vessels ◽  
Guided Wave ◽  
Liquid Level ◽  
Matching Method ◽  
Imaging Results ◽  
Guided Wave Propagation ◽  
Financial Consequences ◽  
Tomography Method

Abstract Pressure vessels are widely utilized in many areas of industrial production and daily life for medium storage, which causes performance degradation in pressure vessels, such as crack and corrosion, and lead to serious safety and financial consequences. Reconstruction Algorithm for the Probabilistic Inspection of Damage (RAPID) is a kind of guided wave-based tomography method which is suitable to evaluate structure integrity of pressure vessels. In this article, the effect of liquid level on guided wave propagation and imaging results of RAPID algorithm is investigated, and an optimal baseline matching method based on amplitude variance is proposed to improve the imaging accuracy of RAPID algorithm with liquid-contained condition. The attenuation effect of liquid on guided wave amplitude is investigated. The damage signals are matched with baseline signals recorded at different liquid levels, and the effect of liquid on RAPID algorithm is discussed based on the results. The experiment of image reconstruction for pressure vessel using the optimal baseline matching method based RAPID algorithm is conducted as well. The experimental results show that the optimal baseline matching method can effectively select the best baseline signal, and the reconstructed images can accurately locate the defects on pressure vessels with considering the change of liquid level.

scholarly journals EVALUATION ON MECHANICAL FRACTURE OF PWR PRESSURE VESSEL AND MODELING BASED ON NEURAL NETWORK

2016 ◽  
Vol 18 (2) ◽  
pp. 87 ◽  
Author(s):  
Mike Susmikanti ◽  
Roziq Himawan ◽  
Abdul Hafid ◽  
Entin Hartini
Keyword(s):  
Neural Network ◽  
Stress Intensity Factor ◽  
Stress Intensity ◽  
Intensity Factor ◽  
Pressure Vessel ◽  
Fracture Analysis ◽  
J Integral ◽  
Network Approach ◽  
Neural Network Approach ◽  
Mechanical Fracture

ABSTRACT EVALUATION ON MECHANICAL FRACTURE OF PWR PRESSURE VESSEL AND MODELING BASED ON NEURAL NETWORK. The important component of the PWR is a pressure vessel. The material resistance in the pressure vessel needs to be evaluated. One way of evaluation is by the mechanical fracture analysis. The modeling needs to know the phenomena of the analysis result in general. A number of researches have been completed on the calculation of mechanical fracture in the pressure vessel with an internal load. The mechanical fracture was modeled using a neural network approach. In relation to the material resistance of the pressure vessel, which is used in PWR AP1000, the material must be evaluated because of the effect of the load. The modeling is needed to predict the effect of the load. The aim of this study is to evaluate the material resistance through mechanical fracture analysis because of the influence load on the pressure vessel on PWR AP1000. The material, which was observed, is SA 508. This analysis consists of the calculation of stress intensity factor and J-integral with some load at the crack propagation position. The fracture mechanic was analyzed by finite element simulation. The result of Stress Intensity factor and J-Integral was compared with fracture toughness to know the durability of the material. The modeling of  J-Integral and Stress Intensity Factor were obtained for some load based on neural network approach. Keywords: Material resistance, mechanical fracture, neural network, PWR, pressure vessel, crack propagation.   ABSTRAK EVALUASI FRAKTUR MEKANIK PADA BEJANA TEKAN PWR DAN PEMODELAN BERBASIS NEURAL NETWORK. Komponen penting dari PWR adalah  bejana tekan. Ketahanan bahan di bejana tekan perlu dievaluasi. Salah satu cara adalah dengan analisis fraktur mekanik. Pemodelan diperlukan untuk mengetahui fenomena hasil analisis pada umumnya. Terdapat penelitian untuk perhitungan fraktur mekanik dalam bejana tekan dengan beban internal. Penelitian lain adalah hasil dari fraktur mekanik dimodelkan menggunakan pendekatan jaringan syaraf. Sehubungan dengan ketahanan material dari bejana tekan yang digunakan dalam PWR AP1000, bahan harus dievaluasi karena efek dari beban. Pemodelan diperlukan untuk memprediksi pengaruh beban pada bahan dalam bejana tekan. Tujuan dari penelitian ini adalah untuk mengevaluasi ketahanan material melalui analisis fraktur mekanik karena pengaruh beban pada bejana tekan. Bahan yang diamati, adalah SA 508. Analisis ini terdiri dari perhitungan faktor intensitas tegangan dan J-integral dengan beberapa beban pada posisi perambatan retak. Fraktur mekanik dianalisis dengan metode elemen hingga. Hasil faktor intensitas tegangan dan J-Integral dibandingkan dengan ketangguhan patah untuk mengetahui daya tahan material. Pemodelan J-Integral dan faktor intensitas stres diperoleh untuk beberapa beban berdasarkan  jaringan saraf. Kata kunci: Ketahanan bahan, teknik patahan,  jaringan syaraf,  PWR,  bejana tekan, perambatan retak. 

Guided Wave Imaging Based on Fully Connected Neural Network for Quantitative Corrosion Assessment

2021 ◽  
Author(s):  
Xiaocen Wang ◽  
Min Lin ◽  
Junkai Tong ◽  
Lin Liang ◽  
Jian Li ◽  
...  
Keyword(s):  
Neural Network ◽  
Real Time ◽  
Guided Wave ◽  
Scientific Management ◽  
Imaging Method ◽  
Online Testing ◽  
Real Time Imaging ◽  
Detection Technology ◽  
Fully Connected ◽  
Mean Square Errors

Abstract Corrosion can affect the reliability of materials, which has attracted the attention of the industry. Corrosion detection and quantitative analysis are particularly important for scientific management and decision-making. In this paper, the imaging method based ultrasonic guided wave (UGW) detection technology and fully connected neural network (FCNN) is proposed to realize real-time imaging of corrosion damages. The imaging method contains offline training and online testing. Offline training aims to establish the relationship between detection signals and velocity maps and it is accelerated by adaptive moment estimation (Adam) algorithm. In the process of online testing, the trained model can be called directly to realize real-time imaging, that is, the detection signals are fed into the model and the network will predict the velocity maps. Finally, the velocity maps are converted to thickness maps according to the dispersion curves. Numerical experimental results show that the mean square errors (mses) are respectively 9.08 × 10−4, 2.47 × 10−3 and 2.59 × 10−3 in training, validation and testing. Compared with irregular corrosion damages, the imaging method has better imaging quality for circular corrosion damages.

Sign in / Sign up

Export Citation Format

Share Document