CN114199992A - Oil storage tank wall corrosion detection method and system - Google Patents

Oil storage tank wall corrosion detection method and system Download PDF

Info

Publication number
CN114199992A
CN114199992A CN202111462839.3A CN202111462839A CN114199992A CN 114199992 A CN114199992 A CN 114199992A CN 202111462839 A CN202111462839 A CN 202111462839A CN 114199992 A CN114199992 A CN 114199992A
Authority
CN
China
Prior art keywords
signal
magnetic
ultrasonic
information
ultrasonic signal
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202111462839.3A
Other languages
Chinese (zh)
Other versions
CN114199992B (en
Inventor
王维斌
杨玉锋
张强
林明春
高晞光
刘硕
许琛琛
王春明
张希祥
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
China Oil and Gas Pipeline Network Corp
National Pipe Network Group North Pipeline Co Ltd
Original Assignee
China Oil and Gas Pipeline Network Corp
National Pipe Network Group North Pipeline Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by China Oil and Gas Pipeline Network Corp, National Pipe Network Group North Pipeline Co Ltd filed Critical China Oil and Gas Pipeline Network Corp
Priority to CN202111462839.3A priority Critical patent/CN114199992B/en
Publication of CN114199992A publication Critical patent/CN114199992A/en
Application granted granted Critical
Publication of CN114199992B publication Critical patent/CN114199992B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N27/00—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means
    • G01N27/72—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means by investigating magnetic variables
    • G01N27/82—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means by investigating magnetic variables for investigating the presence of flaws
    • G01N27/83—Investigating or analysing materials by the use of electric, electrochemical, or magnetic means by investigating magnetic variables for investigating the presence of flaws by investigating stray magnetic fields
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N29/00—Investigating or analysing materials by the use of ultrasonic, sonic or infrasonic waves; Visualisation of the interior of objects by transmitting ultrasonic or sonic waves through the object
    • G01N29/04—Analysing solids
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2291/00—Indexing codes associated with group G01N29/00
    • G01N2291/02—Indexing codes associated with the analysed material
    • G01N2291/023—Solids
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2291/00—Indexing codes associated with group G01N29/00
    • G01N2291/02—Indexing codes associated with the analysed material
    • G01N2291/028—Material parameters
    • G01N2291/0289—Internal structure, e.g. defects, grain size, texture
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/08—Feature extraction

Landscapes

  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Chemical & Material Sciences (AREA)
  • Computational Linguistics (AREA)
  • Biomedical Technology (AREA)
  • Data Mining & Analysis (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Biophysics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Evolutionary Computation (AREA)
  • Pathology (AREA)
  • Artificial Intelligence (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • Immunology (AREA)
  • Acoustics & Sound (AREA)
  • Chemical Kinetics & Catalysis (AREA)
  • Electrochemistry (AREA)
  • Investigating Or Analyzing Materials By The Use Of Magnetic Means (AREA)
  • Investigating Or Analyzing Materials By The Use Of Ultrasonic Waves (AREA)

Abstract

The application discloses a method and a system for detecting corrosion of a tank wall of an oil storage tank, wherein the method comprises the following steps: acquiring information of a newly-thrown oil storage tank wall by obtaining a first magnetic leakage signal and a first ultrasonic signal, wherein the first magnetic leakage signal and the first ultrasonic signal both comprise acquisition position information; acquiring a second magnetic leakage signal and a second ultrasonic signal, wherein the second magnetic leakage signal and the second ultrasonic signal are acquired information of the tank wall of the oil storage tank to be detected, and the acquired position information of the first magnetic leakage signal and the second magnetic leakage signal and the acquired position information of the first ultrasonic signal and the acquired position information of the second ultrasonic signal have correspondences; preprocessing all acquired magnetic flux leakage signals and ultrasonic signals; dividing according to a preset step length to obtain N training samples and N test samples; acquiring a multi-type signal cyclic convolution neural network; obtaining abnormal sample information; and determining the corrosion position of the tank wall of the oil storage tank. The method solves the technical problems that in the prior art, the number of corrosion samples of the tank wall is small, the types are various, and online detection is difficult.

Description

Oil storage tank wall corrosion detection method and system
Technical Field
The application relates to the field of artificial intelligence, in particular to a method and a system for detecting corrosion of a tank wall of an oil storage tank.
Background
The national economy is rapidly developed, along with the increasing demand of energy, petroleum is one of important energy, is a dangerous product, causes great pollution to the ecological environment once leakage occurs, generates huge economic loss, has harsh storage and transportation conditions, and is mainly stored in storage facilities which mainly comprise tank walls, a tank bottom, a tank top, a floating plate and the like. Because the tank wall is in contact with the petroleum for a long time, the tank wall is easily damaged under the influence of factors such as corrosion and the like, and the petroleum is leaked. The more commonly used methods for detecting defects in can walls include can opening detection, inspection by inspection on the case, and on-line detection. The defects of large workload, easy petroleum waste and no practicability; the inspection method is to obtain the defect condition of the tank wall of the storage tank through direct observation of human eyes, and has the defects that the detection precision is very limited, some tiny defects cannot be found, and hidden dangers are easily left; the online detection method is characterized in that the petroleum in the storage tank does not need to be emptied, the storage tank does not need to stop working, the defect data of the wall of the storage tank is obtained in real time in an online mode, and the corrosion condition of the wall of the storage tank is judged on the premise of not opening the tank in an online detection mode. Therefore, on-line detection is currently the most common method for tank wall defect detection.
In the process of implementing the technical solution in the embodiment of the present application, the inventor of the present application finds that the above-mentioned technology has at least the following technical problems:
the technical problems of small quantity of corrosion samples of the tank wall, various types and difficulty in online detection exist in the prior art.
Disclosure of Invention
The application aims to provide a method and a system for detecting corrosion of a tank wall of an oil storage tank, which are used for solving the technical problems that in the prior art, the corrosion samples of the tank wall are few in quantity and various in types, and online detection is difficult.
In view of the above problems, the embodiments of the present application provide a method and a system for detecting corrosion of a tank wall of an oil storage tank.
In a first aspect, the present application provides a method for detecting corrosion of a tank wall of an oil storage tank, the method being implemented by a system for detecting corrosion of a tank wall of an oil storage tank, wherein the method comprises: acquiring a first magnetic leakage signal and a first ultrasonic signal which are acquired information of a newly thrown oil storage tank wall and comprise acquisition position information; acquiring a second magnetic leakage signal and a second ultrasonic signal, wherein the second magnetic leakage signal and the second ultrasonic signal are acquired information of the tank wall of the oil storage tank to be detected, and the acquired position information in the first magnetic leakage signal and the second magnetic leakage signal and the first ultrasonic signal and the second ultrasonic signal have correspondences; preprocessing all acquired magnetic flux leakage signals and ultrasonic signals; dividing the preprocessed magnetic leakage signal and ultrasonic signal according to a preset step length to obtain N training samples and N test samples, wherein N, N are positive integers; training a neural network model through the N training samples to obtain a multi-type signal cyclic convolution neural network; inputting the n test samples into the multi-type signal cyclic convolution neural network to obtain abnormal sample information; and determining the corrosion position of the tank wall of the oil storage tank according to the abnormal sample information.
In another aspect, the present application further provides a system for detecting corrosion of a wall of a storage tank, which is used for performing the method for detecting corrosion of a wall of a storage tank according to the first aspect, wherein the system comprises: a first obtaining unit: the first obtaining unit is used for obtaining a first magnetic leakage signal and a first ultrasonic signal, wherein the first magnetic leakage signal and the first ultrasonic signal are information collected by newly throwing into the tank wall of the oil storage tank and comprise collected position information; a second obtaining unit: the second obtaining unit is used for obtaining a second magnetic leakage signal and a second ultrasonic signal, the second magnetic leakage signal and the second ultrasonic signal are collected information of the tank wall of the oil storage tank to be detected, and the collected position information in the first magnetic leakage signal and the second magnetic leakage signal and the collected position information in the first ultrasonic signal and the second ultrasonic signal have correspondences; a first execution unit: the first execution unit is used for preprocessing all acquired magnetic leakage signals and ultrasonic signals; a third obtaining unit: the third obtaining unit is configured to segment the magnetic leakage signal and the ultrasonic signal after the preprocessing according to a preset step length to obtain N training samples and N test samples, where N, N are positive integers; a fourth obtaining unit: the fourth obtaining unit is used for training the neural network model through the N training samples to obtain a multi-type signal cyclic convolution neural network; a fifth obtaining unit: the fifth obtaining unit is configured to input the n test samples into the multi-type signal cyclic convolution neural network to obtain abnormal sample information; a first determination unit: the first determining unit is used for determining the corrosion position of the tank wall of the oil storage tank according to the abnormal sample information.
In a third aspect, an embodiment of the present application further provides a system for detecting corrosion of a tank wall of an oil storage tank, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method according to the first aspect when executing the program.
One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:
1. acquiring a first magnetic leakage signal and a first ultrasonic signal which are acquired information of a newly thrown oil storage tank wall and comprise acquisition position information; acquiring a second magnetic leakage signal and a second ultrasonic signal, wherein the second magnetic leakage signal and the second ultrasonic signal are acquired information of the wall of the oil tank to be detected, and the acquired position information in the first magnetic leakage signal and the second magnetic leakage signal and the first ultrasonic signal and the second ultrasonic signal have correspondence; preprocessing all acquired magnetic flux leakage signals and ultrasonic signals; dividing the preprocessed magnetic leakage signal and ultrasonic signal according to a preset step length to obtain N training samples and N test samples, wherein N, N are positive integers; training a neural network model through the N training samples to obtain a multi-type signal cyclic convolution neural network; inputting the n test samples into the multi-type signal cyclic convolution neural network to obtain abnormal sample information; and determining the corrosion position of the tank wall of the oil storage tank according to the abnormal sample information. The method achieves the technical effects of intelligently training the multi-type signal cyclic convolution neural network based on an unsupervised training mode, thereby detecting the corrosion condition of the tank wall, improving the information utilization rate and simultaneously improving the corrosion detection accuracy.
2. Through utilizing magnetic leakage signal and ultrasonic signal, establish based on unsupervised study the polymorphic type signal circulation convolution neural network to improve information reuse rate, and then reach the technological effect of reinforcing oil storage tank corrosion detection degree of accuracy.
The foregoing description is only an overview of the technical solutions of the present application, and the present application can be implemented according to the content of the description in order to make the technical means of the present application more clearly understood, and the following detailed description of the present application is given in order to make the above and other objects, features, and advantages of the present application more clearly understandable.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only exemplary, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
FIG. 1 is a schematic flow chart of a method for detecting corrosion of a tank wall of an oil storage tank according to an embodiment of the present disclosure;
fig. 2 is a schematic flow chart illustrating a process of acquiring a magnetic flux leakage signal and an ultrasonic signal according to the moving direction and the moving step length information of the instrument based on the first mark position in the oil storage tank wall corrosion detection method according to the embodiment of the present application;
fig. 3 is a schematic flow chart of obtaining the n test samples according to the second leakage magnetic signal set and the second ultrasonic signal set in the method for detecting corrosion of the tank wall of the oil storage tank according to the embodiment of the present application;
FIG. 4 is a schematic flow chart of a cyclic convolution neural network for obtaining the multi-type signals in the method for detecting corrosion of the tank wall of the oil storage tank according to the embodiment of the present application;
FIG. 5 is a schematic structural diagram of a corrosion detection system for a tank wall of an oil storage tank according to an embodiment of the present disclosure;
fig. 6 is a schematic structural diagram of an exemplary electronic device according to an embodiment of the present application.
Description of reference numerals:
a first obtaining unit 11, a second obtaining unit 12, a first executing unit 13, a third obtaining unit 14, a fourth obtaining unit 15, a fifth obtaining unit 16, a first determining unit 17, a bus 300, a receiver 301, a processor 302, a transmitter 303, a memory 304, and a bus interface 305.
Detailed Description
The embodiment of the application provides a method and a system for detecting corrosion of a tank wall of an oil storage tank, and solves the technical problems that in the prior art, the number of corrosion samples of the tank wall is small, the types are various, and online detection is difficult. The method achieves the technical effects of intelligently training the multi-type signal cyclic convolution neural network based on an unsupervised training mode, thereby detecting the corrosion condition of the tank wall, improving the information utilization rate and simultaneously improving the corrosion detection accuracy.
In the following, the technical solutions in the embodiments of the present application will be clearly and completely described with reference to the accompanying drawings, and it is to be understood that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present application without making any creative effort, shall fall within the protection scope of the present application. It should be further noted that, for the convenience of description, only some but not all of the elements relevant to the present application are shown in the drawings.
Summary of the application
The national economy is rapidly developed, along with the increasing demand of energy, petroleum is one of important energy, is a dangerous product, causes great pollution to the ecological environment once leakage occurs, generates huge economic loss, has harsh storage and transportation conditions, and is mainly stored in storage facilities which mainly comprise tank walls, a tank bottom, a tank top, a floating plate and the like. Because the tank wall is in contact with the petroleum for a long time, the tank wall is easily damaged under the influence of factors such as corrosion and the like, and the petroleum is leaked. The more commonly used methods for detecting defects in can walls include can opening detection, inspection by inspection on the case, and on-line detection. The defects of large workload, easy petroleum waste and no practicability; the inspection method is to obtain the defect condition of the tank wall of the storage tank through direct observation of human eyes, and has the defects that the detection precision is very limited, some tiny defects cannot be found, and hidden dangers are easily left; the online detection method is characterized in that the petroleum in the storage tank does not need to be emptied, the storage tank does not need to stop working, the defect data of the wall of the storage tank is obtained in real time in an online mode, and the corrosion condition of the wall of the storage tank is judged on the premise of not opening the tank in an online detection mode. Therefore, on-line detection is currently the most common method for tank wall defect detection.
The technical problems of small quantity of corrosion samples of the tank wall, various types and difficulty in online detection exist in the prior art.
In view of the above technical problems, the technical solution provided by the present application has the following general idea:
the application provides a corrosion detection method for a tank wall of an oil storage tank, which is applied to a corrosion detection system for the tank wall of the oil storage tank, wherein the method comprises the following steps: acquiring a first magnetic leakage signal and a first ultrasonic signal which are acquired information of a newly thrown oil storage tank wall and comprise acquisition position information; acquiring a second magnetic leakage signal and a second ultrasonic signal, wherein the second magnetic leakage signal and the second ultrasonic signal are acquired information of the tank wall of the oil storage tank to be detected, and the acquired position information in the first magnetic leakage signal and the second magnetic leakage signal and the first ultrasonic signal and the second ultrasonic signal have correspondences; preprocessing all acquired magnetic flux leakage signals and ultrasonic signals; dividing the preprocessed magnetic leakage signal and ultrasonic signal according to a preset step length to obtain N training samples and N test samples, wherein N, N are positive integers; training a neural network model through the N training samples to obtain a multi-type signal cyclic convolution neural network; inputting the n test samples into the multi-type signal cyclic convolution neural network to obtain abnormal sample information; and determining the corrosion position of the tank wall of the oil storage tank according to the abnormal sample information.
Having thus described the general principles of the present application, various non-limiting embodiments thereof will now be described in detail with reference to the accompanying drawings.
Example one
Referring to fig. 1, an embodiment of the present application provides a method for detecting corrosion of a tank wall of an oil storage tank, where the method is applied to a system for detecting corrosion of a tank wall of an oil storage tank, and the method specifically includes the following steps:
step S100: acquiring a first magnetic leakage signal and a first ultrasonic signal, wherein the first magnetic leakage signal and the first ultrasonic signal are information acquired by newly throwing into the tank wall of the oil storage tank and comprise acquisition position information;
particularly, the oil storage tank wall corrosion detection method is applied to the oil storage tank wall corrosion detection system, and can intelligently train a multi-type signal cyclic convolution neural network based on an unsupervised training mode, so that the corrosion condition of the tank wall is intelligently detected, the information utilization rate is improved, and the corrosion detection accuracy is improved. The oil storage tank is a storage tank for storing oil and mainly comprises a tank wall, a tank bottom, a tank top, a floating plate and the like. Because the oil storage tank is contacted with oil for a long time, the tank wall is easy to be damaged due to the influence of corrosion and the like, and oil leakage occurs.
The first magnetic leakage signal is acquired for a specific tank wall position of any newly-used oil storage tank. Similarly, the first ultrasonic signal is acquired for a specific tank wall position of any newly-used oil storage tank. The leakage flux is magnetic field energy of a magnetic source leaked in air (space) through a specific magnetic circuit, the magnetic field of the magnet is closed inside and does not show magnetism to the outside, and the magnetic field is generated after a magnetic pole is formed to the outside. Ultrasound, i.e., ultrasound, is a mechanical wave with an extremely short wavelength, generally shorter than 2cm (centimeters) in air, and must propagate through a medium, and cannot exist in a vacuum (e.g., space). In addition, the method emphasizes that signal acquisition is carried out on a newly-used oil storage tank, the obtained signal is a standard tank wall signal with a complete and defect-free tank wall, and the first leakage magnetic signal and the first ultrasonic signal are acquired at the same tank wall position. Finally, the technical effect of obtaining the magnetic flux leakage signal and the ultrasonic signal of the wall of the oil storage tank in the complete state is achieved.
Step S200: acquiring a second magnetic leakage signal and a second ultrasonic signal, wherein the second magnetic leakage signal and the second ultrasonic signal are acquired information of the tank wall of the oil storage tank to be detected, and the acquired position information in the first magnetic leakage signal and the second magnetic leakage signal and the first ultrasonic signal and the second ultrasonic signal have correspondences;
particularly, the second magnetic leakage signal is the magnetic leakage signal information of examining the oil storage tank wall and gathering, and is in examine the gathering position on the oil storage tank wall and the first magnetic leakage signal and corresponding each other in the new oil storage tank wall gathering position that puts into use. Similarly, the second ultrasonic signal is signal information acquired from the tank wall of the oil storage tank under test at a position corresponding to the acquisition position of the tank wall of the oil storage tank newly put into use. The oil storage tank signal to be detected is acquired at the position corresponding to the wall of the newly-input oil storage tank, the influence of the acquisition position on the signal acquisition result is controlled to be the same, so that the wall signal of the oil storage tank to be detected and the wall signal of the newly-input standard tank are compared, the control variable is reached, and the technical effect of the accuracy of the test analysis result is ensured.
Step S300: preprocessing all acquired magnetic flux leakage signals and ultrasonic signals;
step S400: dividing the preprocessed magnetic leakage signal and ultrasonic signal according to a preset step length to obtain N training samples and N test samples, wherein N, N are positive integers;
specifically, the method comprises the steps of preprocessing a first magnetic leakage signal, a first ultrasonic signal, a second magnetic leakage signal and a second ultrasonic signal which are acquired at the corresponding positions of the tank wall of the newly-thrown oil storage tank and the tank wall of the oil storage tank to be detected in sequence, so that the acquired signals are effectively available.
Further, the preprocessed magnetic leakage signal and ultrasonic signal are segmented according to a preset step length, and therefore a plurality of sample data are obtained. A plurality of samples formed after a first leakage magnetic signal and a first ultrasonic signal acquired from the wall of a newly-thrown oil storage tank are divided are used as training samples, namely the N training samples; and a plurality of samples formed after the second leakage magnetic signal and the second ultrasonic signal acquired from the wall of the oil storage tank are divided are used as test samples, namely the n test samples. Wherein N and N are both positive integers. The preset step length refers to a preset step length threshold value of the oil storage tank wall corrosion detection system based on signal conditions and other comprehensive analysis. By segmenting all signals, the technical effect of obtaining model training data and test data is achieved.
Step S500: training a neural network model through the N training samples to obtain a multi-type signal cyclic convolution neural network;
specifically, the multi-type signal cyclic convolution neural network is obtained through training based on an unsupervised training mode through the N training samples formed after signal segmentation is acquired based on the tank wall of a newly-thrown oil storage tank. The multi-type signal cyclic convolution neural network comprises a magnetic leakage signal cyclic convolution neural network and an ultrasonic signal cyclic convolution neural network. The Convolutional Neural Network (CNN) is a high-precision data classification algorithm, and is a deep feedforward neural network with the characteristics of local connection, weight sharing and the like. Unsupervised learning is an efficient data classification algorithm, and statistical rules in the unsupervised training data can be rapidly acquired. The multi-type signal cyclic convolution neural network is obtained through unsupervised learning training, corrosion samples are not needed to serve as training samples, and the problem that prediction is difficult due to the fact that the number of the corrosion samples is small and the types of the corrosion samples are various in actual tank wall signals is solved.
Step S600: inputting the n test samples into the multi-type signal cyclic convolution neural network to obtain abnormal sample information;
step S700: and determining the corrosion position of the tank wall of the oil storage tank according to the abnormal sample information.
Specifically, the n test samples are respectively input into the multi-type signal cyclic convolution neural network, and the abnormal information in the n test samples can be obtained through intelligent contrast analysis of the multi-type signal cyclic convolution neural network, so that the abnormal sample information is formed. Further, based on the abnormal sample information, the position information of corrosion occurring on the tank wall of the oil storage tank to be detected can be determined. Based on an unsupervised training mode, the multi-type signal cyclic convolution neural network is intelligently trained, so that the corrosion condition of the tank wall is detected, the technical effects of improving the information utilization rate and improving the corrosion detection accuracy are achieved.
Further, as shown in fig. 2, step S100 in the embodiment of the present application further includes:
step S110: obtaining the size information of the oil storage tank;
step S120: calculating to obtain signal acquisition parameters according to the size information of the oil storage tank, wherein the signal acquisition parameters comprise an acquisition starting position, an instrument moving direction and moving step length information;
step S130: obtaining a first mark position according to the acquisition starting position;
step S140: and acquiring and obtaining a magnetic leakage signal and an ultrasonic signal according to the moving direction and the moving step length information of the instrument based on the first mark position.
Specifically, data such as an acquisition initial position, an instrument moving direction, an instrument moving step length and the like in the signal acquisition process are calculated through various parameter information of the size of the oil storage tank. And marking the acquisition initial position, namely the first marked position, and further acquiring magnetic flux leakage signals and ultrasonic signals based on the moving direction and the moving step length information of the instrument to enable the positions to correspond to the signals. The signals are acquired based on a specific direction and a specific step length, and the position information corresponding to each signal can be obtained through calculation by combining the size of the oil storage tank, so that a foundation is laid for subsequent corrosion positioning.
Further, step S300 in the embodiment of the present application further includes:
step S310: all magnetic leakage signals and ultrasonic signals obtained by collection are preprocessed, and the preprocessing comprises the following steps: baseline correction, anomaly detection, loss recovery.
Specifically, the preprocessing of the first magnetic leakage signal, the first ultrasonic signal, the second magnetic leakage signal and the second ultrasonic signal which are acquired from the signals at the corresponding positions of the tank wall of the newly-thrown oil storage tank and the tank wall of the oil storage tank to be detected comprises the steps of baseline correction, abnormal detection and deletion recovery processing of the signals. The baseline correction refers to the correction of signal baselines, so that the baselines of all collected signals are consistent, signal comparison is facilitated, and meanwhile, accurate later positioning is guaranteed; the abnormality detection means that each signal point is intelligently detected through calculation and analysis, and the abnormal signal points are removed; the missing recovery refers to the recovery processing of the missing signal points by adopting a cubic spline algorithm. Through preprocessing, a basis is laid for subsequent signal segmentation and use.
Further, as shown in fig. 3, step S400 in the embodiment of the present application further includes:
step S410: obtaining a first magnetic leakage signal set and a first ultrasonic signal set according to the first magnetic leakage signal and the first ultrasonic signal;
step S420: obtaining the N training samples according to the first leakage magnetic signal set and the first ultrasonic signal set;
step S430: obtaining a second magnetic leakage signal set and a second ultrasonic signal set according to the second magnetic leakage signal and the second ultrasonic signal;
step S440: and obtaining the n test samples according to the second leakage magnetic signal set and the second ultrasonic signal set.
Specifically, the preprocessed first leakage magnetic signal and the preprocessed first ultrasonic signal are segmented according to a preset step length. Wherein, the signal sample that first magnetic leakage signal formed after cutting apart according to the preset step constitutes first magnetic leakage signal set, the signal sample that first ultrasonic signal formed after cutting apart according to the preset step constitutes first ultrasonic signal set. Furthermore, the N training samples are formed by the first leakage magnetic signal set and all the signal samples in the first ultrasonic signal set, that is, the signal acquired by the newly used oil tank wall is divided and then used as the training sample. And on the same principle, dividing the second magnetic leakage signal and the second ultrasonic signal according to a preset step length to obtain a second magnetic leakage signal set and a second ultrasonic signal set, and further forming n test samples by the second magnetic leakage signal set and the second ultrasonic signal set, namely dividing the signal collected by the wall of the oil storage tank to be detected to serve as the test samples. Wherein N and N are both positive integers. By dividing the signals, the acquisition of the model training sample and the acquisition target of the tank wall signal sample of the oil storage tank to be detected are realized.
Further, as shown in fig. 4, step S500 in the embodiment of the present application further includes:
step S510: step a: respectively inputting the first leakage magnetic signal sets in the N training samples into a first convolution feature extraction network to obtain first convolution feature information;
step S520: step b: respectively inputting the first ultrasonic signal sets in the N training samples into a second convolution feature extraction network to obtain second convolution feature information;
step S530: step c: taking the first convolution feature information as a label of a second convolution feature extraction network, and performing ultrasonic extraction network parameter optimization on the second convolution feature extraction network by using a random gradient descent method through a minimum loss function to obtain a second updated convolution feature extraction network;
step S540: step d: obtaining second updated convolution characteristic information output by a second updated convolution characteristic extraction network;
step S550: step e: taking the second updated convolution characteristic information as a label of the first convolution characteristic extraction network, and optimizing the magnetic flux leakage extraction network parameters of the first convolution characteristic extraction network by using a random gradient descent through a minimized loss function to obtain a first updated convolution characteristic extraction network;
step S560: step f: and e, repeating the steps c to e until the ultrasonic extraction network parameters and the magnetic flux leakage extraction network parameters reach preset requirements, and obtaining the multi-type signal cyclic convolution neural network.
Specifically, based on the N training samples, all divided leakage signal samples in the first leakage signal set are sequentially input into a first convolution feature extraction network, and the first convolution feature information is obtained through extraction by the first convolution feature extraction network. In the same method, the second convolution characteristic information corresponding to the first ultrasonic signal set is extracted and obtained based on a second convolution characteristic extraction network.
And taking the first convolution feature information as a label of a second convolution feature extraction network, performing ultrasonic extraction network parameter optimization on the second convolution feature extraction network by using a random gradient descent method through a minimum loss function to obtain the optimized second updated convolution feature extraction network, and further obtaining second updated convolution feature information output by the second updated convolution feature extraction network. Among them, the random gradient descent sgd (stochastic gradient device) is often used for learning linear classifiers under convex loss functions such as support vector machines and logistic regression. After the loss function is calculated for each data, the gradient updating parameter is solved, and the calculation speed is extremely high. In the same method, the second updated convolution characteristic information is used as a label of the first convolution characteristic extraction network, and the first updated convolution characteristic extraction network is intelligently obtained, so that the first updated convolution characteristic information output by the first updated convolution characteristic extraction network is obtained. And repeating the steps until the ultrasonic extraction network parameters and the magnetic flux leakage extraction network parameters reach preset requirements, and updating the network parameters to obtain the multi-type signal cyclic convolution neural network. The preset requirement refers to a cutoff updating condition set after the system is comprehensively analyzed.
Through utilizing magnetic leakage signal and ultrasonic signal, establish based on unsupervised study the polymorphic type signal circulation convolution neural network to improve information reuse rate, and then reach the technological effect of reinforcing oil storage tank corrosion detection degree of accuracy.
Further, step S510 in the embodiment of the present application further includes:
step S511: obtaining a first training sample matrix according to the N training samples;
step S512: performing convolution processing on the convolution layer of the first convolution feature extraction network to obtain a convolution output matrix;
step S513: activating the convolution output matrix through an activation function;
step S514: and inputting the activated convolution matrix into a pooling layer of the first convolution feature extraction network for down-sampling, and outputting through a full connection layer to obtain the first convolution feature information.
Specifically, the first training sample matrix is any sample matrix obtained based on N training samples, and the convolution output matrix is obtained through convolution processing of convolution layers of the first convolution feature extraction network. For example, a training sample dm is obtained by convolving a p × q matrix with a step size of 1 through an α × α convolution kernel, as follows:
Figure BDA0003389264690000151
Figure BDA0003389264690000152
wherein the content of the first and second substances,
Figure BDA0003389264690000153
for convolution operations, matrices
Figure BDA0003389264690000154
The output matrix is convolved.
Further utilizing the activation function:
Figure BDA0003389264690000155
the convolution output matrix is activated through an activation function, and the nonlinearity of the algorithm is enhanced.
And finally, inputting the activated convolution matrix into a pooling layer of the first convolution feature extraction network, and outputting the convolution matrix after down-sampling by using a filter in the pooling layer, namely the first convolution feature information.
Further, step S600 in the embodiment of the present application further includes:
step S610: respectively inputting a second magnetic leakage signal set and a second ultrasonic signal set in the n test samples into the multi-type signal training convolutional neural network to respectively obtain a magnetic leakage signal characteristic vector and an ultrasonic signal characteristic vector;
step S620: according to the magnetic leakage signal characteristic vector and the ultrasonic signal characteristic vector, passing through a formula
Figure BDA0003389264690000161
The characteristic error is obtained by calculation,wherein, the FM magnetic leakage signal characteristic vector, the Fu ultrasonic signal characteristic vector and the E are characteristic errors;
step S630: obtaining a preset abnormal threshold value which is the maximum error value in all training samples;
step S640: and when the characteristic error is not less than the preset abnormal threshold value, determining the characteristic error as the abnormal sample information.
Specifically, a second leakage magnetic signal set and a second ultrasonic signal set in the n test samples are respectively input into the multi-type signal training convolutional neural network, so as to obtain a feature vector corresponding to the leakage magnetic signal and a feature vector corresponding to the ultrasonic signal.
Further, by the following formula:
Figure BDA0003389264690000162
and calculating the characteristic error of the magnetic leakage signal characteristic vector and the ultrasonic signal characteristic vector. Wherein FM is a magnetic leakage signal characteristic vector, Fu is an ultrasonic signal characteristic vector, and E is a characteristic error.
The preset abnormal threshold refers to the maximum error value in all training samples. And judging whether the corresponding characteristic error of each sample is abnormal or not according to the threshold value. When the characteristic error is greater than or equal to a preset abnormal threshold value, determining the characteristic error as abnormal sample information; and when the characteristic error is smaller than a preset abnormal threshold value, determining the characteristic error as normal sample information.
Further, step S310 in the embodiment of the present application further includes:
step S311: according to the formula
Figure BDA0003389264690000163
Performing baseline correction on all magnetic leakage signals and ultrasonic signals, wherein P is the number of channels of the acquired signals, k is the number of sampling counting points, and oijOriginal value, o, for the ith count point position for the jth channelij' is the correction value of the j channel at the ith counting point position, and s is the median of all channels;
step S312: carrying out anomaly detection on all magnetic flux leakage signals and ultrasonic signals based on a 3 sigma criterion to obtain abnormal points, and removing the abnormal points;
step S313: and performing missing point recovery on the magnetic flux leakage signal and the ultrasonic signal after the abnormal points are removed.
Specifically, all the acquired signals are preprocessed, and the baseline correction of the signals is performed first, according to the following formula:
Figure BDA0003389264690000171
wherein, P is the channel number of the collected signals; k is the number of sampling counting points; oijThe original value of the ith counting point position of the jth channel; oij' is the correction value of the j channel at the ith counting point position; s is the median of all channels.
Further, the signal is subjected to anomaly detection and rejection based on a 3 sigma criterion.
When in use
Figure BDA0003389264690000172
When the system is used, the signal point is considered as an abnormal point, and the system intelligently removes the abnormal point.
And finally, performing missing point recovery on the magnetic flux leakage signal and the ultrasonic signal after the abnormal points are removed by adopting a cubic spline algorithm.
To sum up, the method for detecting corrosion of the tank wall of the oil storage tank provided by the embodiment of the application has the following technical effects:
1. acquiring a first magnetic leakage signal and a first ultrasonic signal which are acquired information of a newly thrown oil storage tank wall and comprise acquisition position information; acquiring a second magnetic leakage signal and a second ultrasonic signal, wherein the second magnetic leakage signal and the second ultrasonic signal are acquired information of the wall of the oil tank to be detected, and the acquired position information in the first magnetic leakage signal and the second magnetic leakage signal and the first ultrasonic signal and the second ultrasonic signal have correspondence; preprocessing all acquired magnetic flux leakage signals and ultrasonic signals; dividing the preprocessed magnetic leakage signal and ultrasonic signal according to a preset step length to obtain N training samples and N test samples, wherein N, N are positive integers; training a neural network model through the N training samples to obtain a multi-type signal cyclic convolution neural network; inputting the n test samples into the multi-type signal cyclic convolution neural network to obtain abnormal sample information; and determining the corrosion position of the tank wall of the oil storage tank according to the abnormal sample information. The method achieves the technical effects of intelligently training the multi-type signal cyclic convolution neural network based on an unsupervised training mode, thereby detecting the corrosion condition of the tank wall, improving the information utilization rate and simultaneously improving the corrosion detection accuracy.
2. Through utilizing magnetic leakage signal and ultrasonic signal, establish based on unsupervised study the polymorphic type signal circulation convolution neural network to improve information reuse rate, and then reach the technological effect of reinforcing oil storage tank corrosion detection degree of accuracy.
Example two
Based on the method for detecting corrosion of the tank wall of the oil storage tank in the foregoing embodiment, the invention also provides a system for detecting corrosion of the tank wall of the oil storage tank, referring to fig. 5, where the system includes:
the first obtaining unit 11 is configured to obtain a first magnetic leakage signal and a first ultrasonic signal, where the first magnetic leakage signal and the first ultrasonic signal are information collected by newly throwing into a tank wall of an oil storage tank and both include collection position information;
the second obtaining unit 12 is configured to obtain a second magnetic flux leakage signal and a second ultrasonic signal, where the second magnetic flux leakage signal and the second ultrasonic signal are information collected by a tank wall of the oil storage tank to be detected, and the collected position information in the first magnetic flux leakage signal and the second magnetic flux leakage signal and the collected position information in the first ultrasonic signal and the second ultrasonic signal have a correspondence;
the first execution unit 13 is configured to perform preprocessing on all acquired magnetic leakage signals and ultrasonic signals;
a third obtaining unit 14, where the third obtaining unit 14 is configured to segment the magnetic leakage signal and the ultrasonic signal after the preprocessing according to a preset step size to obtain N training samples and N test samples, where N, N are positive integers;
a fourth obtaining unit 15, where the fourth obtaining unit 15 is configured to train a neural network model through the N training samples to obtain a multi-type signal cyclic convolution neural network;
a fifth obtaining unit 16, where the fifth obtaining unit 16 is configured to input the n test samples into the multi-type signal cyclic convolution neural network, so as to obtain abnormal sample information;
a first determination unit 17, wherein the first determination unit 17 is used for determining the corrosion position of the oil storage tank wall according to the abnormal sample information.
Further, the system further comprises:
a sixth obtaining unit for obtaining oil tank size information;
the seventh obtaining unit is used for calculating and obtaining signal acquisition parameters according to the size information of the oil storage tank, wherein the signal acquisition parameters comprise an acquisition starting position, an instrument moving direction and moving step length information;
an eighth obtaining unit, configured to obtain a first mark position according to the acquisition start position;
and the ninth obtaining unit is used for acquiring and obtaining a magnetic leakage signal and an ultrasonic signal according to the moving direction and the moving step length information of the instrument based on the first mark position.
Further, the system further comprises:
the first setting unit, the first setting unit is used for all leak magnetic signals, ultrasonic signal that the collection obtained carry out the preliminary treatment, include: baseline correction, anomaly detection, loss recovery.
Further, the system further comprises:
a tenth obtaining unit, configured to obtain a first leakage magnetic signal set and a first ultrasonic signal set according to the first leakage magnetic signal and the first ultrasonic signal;
an eleventh obtaining unit, configured to obtain the N training samples according to the first leakage magnetic signal set and the first ultrasonic signal set;
a twelfth obtaining unit, configured to obtain a second leakage magnetic signal set and a second ultrasonic signal set according to the second leakage magnetic signal and the second ultrasonic signal;
a thirteenth obtaining unit, configured to obtain the n test samples according to the second leakage magnetic signal set and the second ultrasonic signal set.
Further, the system further comprises:
a fourteenth obtaining unit configured to perform step a: respectively inputting the first leakage magnetic signal sets in the N training samples into a first convolution feature extraction network to obtain first convolution feature information;
a fifteenth obtaining unit configured to perform step b: respectively inputting the first ultrasonic signal sets in the N training samples into a second convolution feature extraction network to obtain second convolution feature information;
a sixteenth obtaining unit configured to perform step c: taking the first convolution feature information as a label of a second convolution feature extraction network, and performing ultrasonic extraction network parameter optimization on the second convolution feature extraction network by using a random gradient descent method through a minimum loss function to obtain a second updated convolution feature extraction network;
a seventeenth obtaining unit configured to perform step d: obtaining second updated convolution characteristic information output by a second updated convolution characteristic extraction network;
an eighteenth obtaining unit configured to perform step e: taking the second updated convolution characteristic information as a label of the first convolution characteristic extraction network, and optimizing the magnetic flux leakage extraction network parameters of the first convolution characteristic extraction network by using a random gradient descent through a minimized loss function to obtain a first updated convolution characteristic extraction network;
a nineteenth obtaining unit configured to perform step f: and e, repeating the steps c to e until the ultrasonic extraction network parameters and the magnetic flux leakage extraction network parameters reach preset requirements, and obtaining the multi-type signal cyclic convolution neural network.
Further, the system further comprises:
a twentieth obtaining unit, configured to obtain a first training sample matrix according to the N training samples;
a twenty-first obtaining unit, configured to perform convolution processing on the convolution layer of the first convolution feature extraction network to obtain a convolution output matrix;
a second execution unit to activate the convolution output matrix with an activation function;
and a twenty-second obtaining unit, configured to input the activated convolution matrix into a pooling layer of the first convolution feature extraction network, perform down-sampling, and output the result through a full connection layer to obtain the first convolution feature information.
Further, the system further comprises:
a twenty-third obtaining unit, configured to input the second leakage magnetic signal set and the second ultrasonic signal set in the n test samples into the multi-type signal training convolutional neural network, respectively, and obtain a leakage magnetic signal feature vector and an ultrasonic signal feature vector, respectively;
a twenty-fourth obtaining unit, configured to obtain, according to the leakage magnetic signal feature vector and the ultrasonic signal feature vector, a formula E | | | FM-FU||22Calculating to obtain a characteristic error, wherein an FM magnetic flux leakage signal characteristic vector, a Fu ultrasonic signal characteristic vector and E are characteristic errors;
a twenty-fifth obtaining unit, configured to obtain a preset abnormal threshold, where the preset abnormal threshold is a maximum error value in all training samples;
a second determining unit, configured to determine the feature error as the abnormal sample information when the feature error is not less than the preset abnormal threshold.
Further, the system further comprises:
a third execution unit to execute a formula
Figure BDA0003389264690000221
Performing baseline correction on all magnetic leakage signals and ultrasonic signals, wherein P is the number of channels of the acquired signals, k is the number of sampling counting points, and oijOriginal value, o, for the ith count point position for the jth channelij' is the correction value of the j channel at the ith counting point position, and s is the median of all channels;
a twenty-fifth obtaining unit, configured to perform anomaly detection on all magnetic flux leakage signals and ultrasonic signals based on a 3 σ criterion, obtain an anomaly point, and remove the anomaly point;
and the fourth execution unit is used for performing missing point recovery on the magnetic leakage signal and the ultrasonic signal after the abnormal points are removed.
The embodiments in the present specification are described in a progressive manner, and each embodiment focuses on differences from other embodiments, and the method for detecting corrosion of a tank wall of an oil storage tank in the first embodiment of fig. 1 and the specific example are also applicable to a system for detecting corrosion of a tank wall of an oil storage tank in the present embodiment. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Exemplary electronic device
The electronic apparatus of the embodiment of the present application is described below with reference to fig. 6.
Fig. 6 illustrates a schematic structural diagram of an electronic device according to an embodiment of the present application.
Based on the inventive concept of the method for detecting corrosion of a wall of an oil storage tank in the foregoing embodiments, the present invention further provides a system for detecting corrosion of a wall of an oil storage tank, on which a computer program is stored, which, when being executed by a processor, implements the steps of any one of the methods for detecting corrosion of a wall of an oil storage tank described above.
Where in fig. 6 a bus architecture (represented by bus 300), bus 300 may include any number of interconnected buses and bridges, bus 300 linking together various circuits including one or more processors, represented by processor 302, and memory, represented by memory 304. The bus 300 may also link together various other circuits such as peripherals, voltage regulators, power management circuits, and the like, which are well known in the art, and therefore, will not be described any further herein. A bus interface 305 provides an interface between the bus 300 and the receiver 301 and transmitter 303. The receiver 301 and the transmitter 303 may be the same element, i.e., a transceiver, providing a means for communicating with various other apparatus over a transmission medium.
The processor 302 is responsible for managing the bus 300 and general processing, and the memory 304 may be used for storing data used by the processor 302 in performing operations.
The application provides a corrosion detection method for a tank wall of an oil storage tank, which is applied to a corrosion detection system for the tank wall of the oil storage tank, wherein the method comprises the following steps: acquiring a first magnetic leakage signal and a first ultrasonic signal which are acquired information of a newly thrown oil storage tank wall and comprise acquisition position information; acquiring a second magnetic leakage signal and a second ultrasonic signal, wherein the second magnetic leakage signal and the second ultrasonic signal are acquired information of the tank wall of the oil storage tank to be detected, and the acquired position information in the first magnetic leakage signal and the second magnetic leakage signal and the first ultrasonic signal and the second ultrasonic signal have correspondences; preprocessing all acquired magnetic flux leakage signals and ultrasonic signals; dividing the preprocessed magnetic leakage signal and ultrasonic signal according to a preset step length to obtain N training samples and N test samples, wherein N, N are positive integers; training a neural network model through the N training samples to obtain a multi-type signal cyclic convolution neural network; inputting the n test samples into the multi-type signal cyclic convolution neural network to obtain abnormal sample information; and determining the corrosion position of the tank wall of the oil storage tank according to the abnormal sample information. The method solves the technical problems that in the prior art, the number of corrosion samples of the tank wall is small, the types are various, and online detection is difficult. The method achieves the technical effects of intelligently training the multi-type signal cyclic convolution neural network based on an unsupervised training mode, thereby detecting the corrosion condition of the tank wall, improving the information utilization rate and simultaneously improving the corrosion detection accuracy.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, apparatus, or computer program product. Accordingly, the present application may take the form of an entirely software embodiment, an entirely hardware embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application is in the form of a computer program product that may be embodied on one or more computer-usable storage media having computer-usable program code embodied therewith. And such computer-usable storage media include, but are not limited to: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk Memory, a Compact Disc Read-Only Memory (CD-ROM), and an optical Memory.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a system for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction system which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks. While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all such alterations and modifications as fall within the scope of the invention.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.

Claims (10)

1. A method for detecting corrosion of a tank wall of a storage tank, wherein the method is applied to a system for detecting corrosion of a tank wall of a storage tank, and the method comprises the following steps:
acquiring a first magnetic leakage signal and a first ultrasonic signal, wherein the first magnetic leakage signal and the first ultrasonic signal are information acquired by newly throwing into the tank wall of the oil storage tank and comprise acquisition position information;
acquiring a second magnetic leakage signal and a second ultrasonic signal, wherein the second magnetic leakage signal and the second ultrasonic signal are acquired information of the tank wall of the oil storage tank to be detected, and the acquired position information in the first magnetic leakage signal and the second magnetic leakage signal and the first ultrasonic signal and the second ultrasonic signal have correspondences;
preprocessing all acquired magnetic flux leakage signals and ultrasonic signals;
dividing the preprocessed magnetic leakage signal and ultrasonic signal according to a preset step length to obtain N training samples and N test samples, wherein N, N are positive integers;
training a neural network model through the N training samples to obtain a multi-type signal cyclic convolution neural network;
inputting the n test samples into the multi-type signal cyclic convolution neural network to obtain abnormal sample information;
and determining the corrosion position of the tank wall of the oil storage tank according to the abnormal sample information.
2. The method of claim 1, wherein obtaining the first leakage magnetic signal and the first ultrasonic signal is preceded by:
obtaining the size information of the oil storage tank;
calculating to obtain signal acquisition parameters according to the size information of the oil storage tank, wherein the signal acquisition parameters comprise an acquisition starting position, an instrument moving direction and moving step length information;
obtaining a first mark position according to the acquisition starting position;
and acquiring and obtaining a magnetic leakage signal and an ultrasonic signal according to the moving direction and the moving step length information of the instrument based on the first mark position.
3. The method of claim 1, wherein the preprocessing of all the acquired leakage magnetic signals and ultrasonic signals comprises: baseline correction, anomaly detection, loss recovery.
4. The method of claim 1, wherein the segmenting the preprocessed leakage magnetic signal and ultrasonic signal according to a preset step size to obtain N training samples and N test samples comprises:
obtaining a first magnetic leakage signal set and a first ultrasonic signal set according to the first magnetic leakage signal and the first ultrasonic signal;
obtaining the N training samples according to the first leakage magnetic signal set and the first ultrasonic signal set;
obtaining a second magnetic leakage signal set and a second ultrasonic signal set according to the second magnetic leakage signal and the second ultrasonic signal;
and obtaining the n test samples according to the second leakage magnetic signal set and the second ultrasonic signal set.
5. The method of claim 4, wherein said training a neural network model through said N training samples to obtain a multi-type signal cyclic convolution neural network comprises:
step a: respectively inputting the first leakage magnetic signal sets in the N training samples into a first convolution feature extraction network to obtain first convolution feature information;
step b: respectively inputting the first ultrasonic signal sets in the N training samples into a second convolution feature extraction network to obtain second convolution feature information;
step c: taking the first convolution feature information as a label of a second convolution feature extraction network, and performing ultrasonic extraction network parameter optimization on the second convolution feature extraction network by using a random gradient descent method through a minimum loss function to obtain a second updated convolution feature extraction network;
step d: obtaining second updated convolution characteristic information output by a second updated convolution characteristic extraction network;
step e: taking the second updated convolution characteristic information as a label of the first convolution characteristic extraction network, and optimizing the magnetic flux leakage extraction network parameters of the first convolution characteristic extraction network by using a random gradient descent through a minimized loss function to obtain a first updated convolution characteristic extraction network;
step f: and e, repeating the steps c to e until the ultrasonic extraction network parameters and the magnetic flux leakage extraction network parameters reach preset requirements, and obtaining the multi-type signal cyclic convolution neural network.
6. The method of claim 5, wherein the inputting the first leakage magnetic signal sets of the N training samples into a first convolution feature extraction network respectively to obtain first convolution feature information comprises:
obtaining a first training sample matrix according to the N training samples;
performing convolution processing on the convolution layer of the first convolution feature extraction network to obtain a convolution output matrix;
activating the convolution output matrix through an activation function;
and inputting the activated convolution matrix into a pooling layer of the first convolution feature extraction network for down-sampling, and outputting through a full connection layer to obtain the first convolution feature information.
7. The method of claim 4, wherein said inputting said n test samples into said multi-type signal cyclic convolutional neural network to obtain abnormal sample information comprises:
respectively inputting a second magnetic leakage signal set and a second ultrasonic signal set in the n test samples into the multi-type signal training convolutional neural network to respectively obtain a magnetic leakage signal characteristic vector and an ultrasonic signal characteristic vector;
according to the magnetic leakage signal characteristic vector and the ultrasonic signal characteristic vector, passing through a formula
Figure FDA0003389264680000041
Calculating to obtain a characteristic error, wherein an FM magnetic flux leakage signal characteristic vector, a Fu ultrasonic signal characteristic vector and E are characteristic errors;
obtaining a preset abnormal threshold value which is the maximum error value in all training samples;
and when the characteristic error is not less than the preset abnormal threshold value, determining the characteristic error as the abnormal sample information.
8. The method of claim 3, wherein the preprocessing of all the acquired leakage magnetic signals and ultrasonic signals comprises: baseline correction, anomaly detection, loss recovery, including:
according to the formula
Figure FDA0003389264680000042
Performing baseline correction on all magnetic leakage signals and ultrasonic signals, wherein P is the number of channels of the acquired signals, k is the number of sampling counting points, and oijOriginal value, o, for the ith count point position for the jth channelij' is the correction value of the j channel at the ith counting point position, and s is the median of all channels;
carrying out anomaly detection on all magnetic flux leakage signals and ultrasonic signals based on a 3 sigma criterion to obtain abnormal points, and removing the abnormal points;
and performing missing point recovery on the magnetic flux leakage signal and the ultrasonic signal after the abnormal points are removed.
9. A storage tank wall corrosion detection system, wherein the system comprises:
a first obtaining unit: the first obtaining unit is used for obtaining a first magnetic leakage signal and a first ultrasonic signal, wherein the first magnetic leakage signal and the first ultrasonic signal are information collected by newly throwing into the tank wall of the oil storage tank and comprise collected position information;
a second obtaining unit: the second obtaining unit is used for obtaining a second magnetic leakage signal and a second ultrasonic signal, the second magnetic leakage signal and the second ultrasonic signal are collected information of the tank wall of the oil storage tank to be detected, and the collected position information in the first magnetic leakage signal and the second magnetic leakage signal and the collected position information in the first ultrasonic signal and the second ultrasonic signal have correspondences;
a first execution unit: the first execution unit is used for preprocessing all acquired magnetic leakage signals and ultrasonic signals;
a third obtaining unit: the third obtaining unit is configured to segment the magnetic leakage signal and the ultrasonic signal after the preprocessing according to a preset step length to obtain N training samples and N test samples, where N, N are positive integers;
a fourth obtaining unit: the fourth obtaining unit is used for training the neural network model through the N training samples to obtain a multi-type signal cyclic convolution neural network;
a fifth obtaining unit: the fifth obtaining unit is configured to input the n test samples into the multi-type signal cyclic convolution neural network to obtain abnormal sample information;
a first determination unit: the first determining unit is used for determining the corrosion position of the tank wall of the oil storage tank according to the abnormal sample information.
10. A storage tank wall corrosion detection system comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor executes the program to perform the steps of the method of any one of claims 1 to 8.
CN202111462839.3A 2021-12-02 2021-12-02 Method and system for detecting corrosion of tank wall of oil storage tank Active CN114199992B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111462839.3A CN114199992B (en) 2021-12-02 2021-12-02 Method and system for detecting corrosion of tank wall of oil storage tank

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111462839.3A CN114199992B (en) 2021-12-02 2021-12-02 Method and system for detecting corrosion of tank wall of oil storage tank

Publications (2)

Publication Number Publication Date
CN114199992A true CN114199992A (en) 2022-03-18
CN114199992B CN114199992B (en) 2024-04-05

Family

ID=80650274

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111462839.3A Active CN114199992B (en) 2021-12-02 2021-12-02 Method and system for detecting corrosion of tank wall of oil storage tank

Country Status (1)

Country Link
CN (1) CN114199992B (en)

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104458895A (en) * 2014-12-08 2015-03-25 清华大学 Three-dimensional pipeline leakage flux imaging detection method and system
CN109783906A (en) * 2018-12-29 2019-05-21 东北大学 A kind of pipeline detection magnetic flux leakage data intelligent analysis system and method
CN110068609A (en) * 2019-05-31 2019-07-30 中国计量大学 A kind of compound bearing device inside fault detection system of the ultrasonic accurate measurement of leakage field Rough Inspection combination
CN110412120A (en) * 2019-06-26 2019-11-05 中国石油天然气股份有限公司 Pipeline crack detection method and device
US20200210826A1 (en) * 2018-12-29 2020-07-02 Northeastern University Intelligent analysis system using magnetic flux leakage data in pipeline inner inspection
CN111664365A (en) * 2020-06-07 2020-09-15 东北石油大学 Oil and gas pipeline leakage detection method based on improved VMD and 1DCNN
EP3722800A1 (en) * 2019-04-09 2020-10-14 Rosen Swiss AG Method for determining the geometry of a defect on the basis of nondestructive measurement method using direct inversion
CN111965246A (en) * 2020-08-11 2020-11-20 太原理工大学 Scraper machine fault detection method and detection system based on multi-information fusion
CN112668527A (en) * 2020-12-31 2021-04-16 华南理工大学 Ultrasonic guided wave semi-supervised imaging detection method
CN113051325A (en) * 2021-03-25 2021-06-29 陕西延长石油（集团）有限责任公司 Storage tank detection data display analysis method, system, medium, equipment and application

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104458895A (en) * 2014-12-08 2015-03-25 清华大学 Three-dimensional pipeline leakage flux imaging detection method and system
CN109783906A (en) * 2018-12-29 2019-05-21 东北大学 A kind of pipeline detection magnetic flux leakage data intelligent analysis system and method
US20200210826A1 (en) * 2018-12-29 2020-07-02 Northeastern University Intelligent analysis system using magnetic flux leakage data in pipeline inner inspection
EP3722800A1 (en) * 2019-04-09 2020-10-14 Rosen Swiss AG Method for determining the geometry of a defect on the basis of nondestructive measurement method using direct inversion
CN110068609A (en) * 2019-05-31 2019-07-30 中国计量大学 A kind of compound bearing device inside fault detection system of the ultrasonic accurate measurement of leakage field Rough Inspection combination
CN110412120A (en) * 2019-06-26 2019-11-05 中国石油天然气股份有限公司 Pipeline crack detection method and device
CN111664365A (en) * 2020-06-07 2020-09-15 东北石油大学 Oil and gas pipeline leakage detection method based on improved VMD and 1DCNN
CN111965246A (en) * 2020-08-11 2020-11-20 太原理工大学 Scraper machine fault detection method and detection system based on multi-information fusion
CN112668527A (en) * 2020-12-31 2021-04-16 华南理工大学 Ultrasonic guided wave semi-supervised imaging detection method
CN113051325A (en) * 2021-03-25 2021-06-29 陕西延长石油（集团）有限责任公司 Storage tank detection data display analysis method, system, medium, equipment and application

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
张增晓: "基于声发射检测的储罐底板腐蚀评估方法研究", 中国优秀硕士学位论文全文数据库 工程科技I辑, no. 2, 15 February 2021 (2021-02-15), pages 31 - 36 *
田景文 等: "人工神经网络算法研究及应用", 31 July 2006, 北京理工大学出版社, pages: 174 - 179 *

Also Published As

Publication number Publication date
CN114199992B (en) 2024-04-05

Similar Documents

Publication Publication Date Title
CN110942072A (en) Quality evaluation-based quality scoring and detecting model training and detecting method and device
CN111008959A (en) Grading ring defect detection method, device, medium and equipment based on aerial image
CN112926642A (en) Multi-source domain self-adaptive intelligent mechanical fault diagnosis method and system
CN109242165A (en) A kind of model training and prediction technique and device based on model training
CN113936210A (en) Anti-collision method for tower crane
CN116227586A (en) Meta learning fault diagnosis method and device based on depth residual error shrinkage prototype network
CN115793590A (en) Data processing method and platform suitable for system safety operation and maintenance
CN115171045A (en) YOLO-based power grid operation field violation identification method and terminal
CN110889338A (en) Unsupervised railway track bed foreign matter detection and sample construction method and unsupervised railway track bed foreign matter detection and sample construction device
CN112346974B (en) Depth feature embedding-based cross-mobile application program instant defect prediction method
CN116152674A (en) Dam unmanned aerial vehicle image crack intelligent recognition method based on improved U-Net model
CN117274212A (en) Bridge underwater structure crack detection method
CN113780287A (en) Optimal selection method and system for multi-depth learning model
CN115063337A (en) Intelligent maintenance decision-making method and device for buried pipeline
CN110320802B (en) Complex system signal time sequence identification method based on data visualization
CN115239733B (en) Crack detection method and apparatus, terminal device and storage medium
CN111126694A (en) Time series data prediction method, system, medium and device
CN114199992A (en) Oil storage tank wall corrosion detection method and system
CN116186609A (en) LSTM-based manned submersible sensor fault data enhancement method
CN114550460B (en) Rail transit anomaly detection method and device and storage medium
CN113420733B (en) Efficient distributed big data acquisition implementation method and system
CN113920311A (en) Remote sensing image segmentation method and system based on edge auxiliary information
Jia et al. Defect severity assessment model for sewer pipeline based on automated pipe calibration
Gao et al. Detection method of potholes on highway pavement based on yolov5
CN113205215A (en) Knowledge-based battlefield situation prediction method

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant