Research on Quantitative Identification of Pipeline Defects Based on Wavelet Analysis
The most important means in oil and gas transportation is pipeline transportation. Corrosion and defects of oil and gas pipelines will bring huge economic losses and environmental pollution to our country, especially the current project of 4,000 km long-term gas-to-east gas transmission project in China, which costs nearly 20 billion US dollars. Intelligent pipeline inspection devices with independent intellectual property rights are still urgent. To this end, the measurement and control center of Shenyang University of Technology has developed a high-precision pipeline magnetic flux leakage online detection system in China, which produces magnetic flux leakage at the defects of the metal pipe wall after magnetization. . This paper uses wavelet transform as a tool to analyze the corrosion defect data and establish a quantitative analysis model to provide a theoretical basis for intelligent detection of defects.
1 Defect magnetic flux leakage signal detection Due to the complicated conditions of on-site detection, the journal of Shenyang University of Technology for natural corrosion defects is complex and diverse. In the experimental stage, it is necessary to simulate on-site sample defects as much as possible. By measuring the produced defects, accurate sample data is obtained, and the characteristics of the magnetic flux leakage signal of the detected defect are extracted and analyzed by wavelet transform, so as to effectively establish a mathematical model for defect discrimination.
Experimental steel plate thickness: 10 mm, a series of defects were made on the steel plate according to the depth h, mm, 70 mm, and 90 mm. The 10 probes placed on the board are placed side by side with a length of about 20 cm. The probes are routed every 2 cm, the pulling speed is lm / s, and they are sampled every 1.5 mm, each composed of a radial signal and an axial signal. Taking the information measured by one of the probes, the main research object is the axial flux leakage component. Part of the magnetic flux leakage signal is shown in Figure 1.
2 Defect signal analysis based on wavelet transform 2.1 Wavelet transform and its modulo extremal wavelet transform refer to taking a function called wavelet as displacement b, and then making inner product with the signal f to be analyzed at different scales a Among them: conjugate.
As a wavelet function, the admissible condition C = d must be satisfied, so there is an inverse transformation where W (a, b) is the wavelet transformation of the function f. Under scale, if there is a point (a) that makes 0, it is called this point (a) Is the local extreme point. If for any point b in a certain neighborhood of b, there is W) the modulus maximum point of wavelet transform, called W) is the modulus maximum value of wavelet transform. The modulo minimum has a similar definition.
2. Wavelet analysis of abrupt signals Traditional signal analysis mainly uses Fourier transform, but it is a global transform. The wavelet transform analysis method has the ability to characterize the local characteristics of the signal in both time and frequency domains, and is suitable for detecting transient anomalies entrained in normal signals. Defect magnetic flux leakage signals are mostly abrupt signals, so wavelet transform is selected as a tool for analysis and research.
The sudden point of the signal in the wavelet transform domain often corresponds to the extreme point of the wavelet transform module, so after the measured signal (Figure 1) is subjected to wavelet transform (Figure 2), the maximum and minimum values ​​of the wavelet transform module are used to determine the signal The position of the mutation point [6, 7]. At the same time, the choice of wavelet function has a great influence on the signal analysis. In this paper, the Haar wavelet with good regularity, orthogonality, symmetry and simple calculation is selected.
2. 3 feature extraction In order to quantify the size of the identified defect, the waveform of the measured signal after wavelet transformation is selected (see Figure 3): x is the peak-to-peak spacing, mm x is the peak-to-valley height, T x is the waveform Area (overcast is area / peak-to-valley height, mm x is area / peak-to-peak spacing, T.
Use the wavelet transform in Matlab to find each feature value 3 Quantitative identification of defect size 3.1 Basic theory of multiple regression analysis Defects A feature parameter is related to multiple feature quantities (variables), and multiple feature quantities and defects Between characteristic parameters Xie Yanhong et al: Research on quantitative identification of pipeline defects based on wavelet analysis Note: In the table, y is the actual length of the defect and h is the actual depth of the defect.
The relationship is generally non-linear, and the solution to this problem can use multiple nonlinear regression methods. Through variable substitution, the nonlinear relationship can be transformed into a linear relationship [9]. The basic steps are as follows: 1) Assume the regression model is Y = 2) Use least squares estimation to calculate the parameters ^ 3) The significance of the regression equation (model) Performance test and error analysis 4) Get a model that meets the actual requirements.
3.2 Multiple regression analysis method to quantitatively identify defect size 3. 2. 1 Relationship between defect length and feature quantity The analysis of Table 1 can assume that the regression model of length y with respect to feature quantity is transformed z 3, then the above modeling For the linear regression model, Matlab linear regression function regress () is used for regression [10], and the regression model complex correlation coefficient R-reduced residual plot and confidence interval are shown in Figure 4. R is very close to 1, indicating that the degree of fitting of the model is quite high, and the significance probability value p is much smaller than the test level = 0.05, and the residuals are relatively small, so the regression model is meaningful.
3.2.2 The relationship model between defect depth and feature quantity is analyzed in Table 1, and the linear regression function regress () is used to obtain the regression model complex correlation coefficient R graph as shown in Figure 5. R is closer to 1, indicating that the model fits better. The significance probability p is much smaller than the test level = 0.05, and the residuals are relatively small, so the regression model is meaningful.
4 Conclusion Using wavelet transform as a research tool to extract the characteristics of the magnetic flux leakage signal, using multiple regression analysis method, a mathematical model for identifying the length and depth of the defect is obtained. After residual analysis and significance test, the fitting degree is ideal. And through the experiment of Shenyang University of Technology, the correctness of the regression model is verified, and the accuracy is within the allowable range of error. Of course, there is still a lot of work to be done to apply the results of the test stage to the field.
[1] Wang Gongli, Wang Li. Oil and gas pipeline technology status and development trend [J].
[2] Liu Xiuqing, Gao Songwei, Yang Lijian. Research and application of ANSYS in the detection of pipeline magnetic flux leakage method [J]. Journal of Shenyang University of Technology, 2001, 23 [3] Yang Lijian, Wang Yumei, Feng Haiying. Intelligent pipeline magnetic flux leakage detection device [4] Yang Fusheng. Engineering Analysis and Application of Wavelet Transform [M]. Beijing: Science Press, 2000.
[5] Cui Jintai. Introduction to Wavelet Analysis [M]. Xi'an: Xi'an Jiaotong University [8] Hu Changhua, Zhou Tao, Xia Qibing, etc. System analysis and design wavelet analysis based on MATLAB [M]. Xi'an: Xidian University Press, 2002.
[9] He Xiaoqun, Liu Wenqing. Applied regression analysis [M]. Beijing: Renmin University of China Press, 2001.
[10] Chen Guiming, Qi Hongyu, Pan Wei. MATLAB Mathematical Statistics (6. X) [M]. Beijing: Science Press, 2002.
Controlling the PWM circuit to control the welding current and voltage can realize the welding sequence control. At the same time, it also realizes the closed-loop control of the spot welding system through the real-time detection of the welding current and the net voltage.
Li Chaoyang. The principle and interface technology of single chip microcomputer [M]. Beijing: Beijing University of Aeronautics and Astronautics Press, 1999.
Li Aiwen. Modern inverter technology and its application [M]. Beijing: Wang Fusheng from science. Research on Microcomputer Controlled Hanging Inverter Secondary Rectifier Spot Welding Tongs Zhang Caiyun. Resistance welding technology and its application equipment [J]. Electronic process Geng Zheng. Development of resistance welding equipment and control research [J]. Welding, 1998 [6] Zeng Hongzhi. Research progress of resistance spot welding process and quality control methods [8] Xu Aiqing. Intel 16-bit single-chip microcomputer [M]. Beijing: Beijing University of Aeronautics and Astronautics Press, 2002.
Xie Yanhong et al: Research on Quantitative Identification of Pipeline Defects Based on Wavelet Analysis
It is no doubt that it would be more convenient and safer to use a slicer tool than knives, as far as the effectiveness the slicer tools perform better on slicing works, which have more tangent uniform and edge with the equal pieces. Especially we could finish the works in a few seconds, it is very efficient and easy to clean, and the slicer tools are safe in dishwasher.
Fruit Slicer,Cheese Slicer,Apple Slicer,Apple Cutter
Fortary Industry&Trading Co., Ltd. , https://www.kitchenset.de