Nondestructive testing system and method for boiler pressure pipeline based on image processing

Abstract

The invention discloses a non-destructive testing system and method for a boiler pressure pipeline based on image processing, relates to the field of non-destructive testing, and solves the problem that an existing detection method for the boiler pressure pipeline is poor in effect. The sensor module is used for carrying out sensor detection on the boiler according to the boiler image processing data and respectively obtaining a pipeline pressure detection coefficient, a pipeline temperature detection coefficient and a pipeline vibration frequency detection coefficient to obtain boiler sensor data; and the nondestructive testing module is used for carrying out nondestructive testing on the boiler pipeline to be detected according to the boiler sensor data and the boiler image processing data, and issuing an abnormal early warning according to a detection result. The accuracy and comprehensiveness of nondestructive testing of the boiler pressure pipeline can be ensured.

Description

Technical Field

[0001] The invention belongs to the field of nondestructive testing and relates to image processing technology, in particular to a nondestructive testing system and method for a boiler pressure pipe based on image processing. Background Art

[0002] The existing non-destructive testing system for boiler pressure pipes has the following defects when testing boiler pressure pipes:

[0003] 1. The existing detection methods require professional operators to conduct regular inspections, and the single inspection time is long, which increases the inspection cost;

[0004] Existing detection methods rarely use advanced technologies such as artificial intelligence and machine learning, resulting in a lack of automation and intelligence in the detection process;

[0005] Therefore, we propose a nondestructive testing system and method for boiler pressure pipes based on image processing. Summary of the invention

[0006] In view of the deficiencies in the prior art, the object of the present invention is to provide a nondestructive testing system and method for boiler pressure pipes based on image processing. The present invention is based on obtaining multiple detection area images inside the boiler pipe to be inspected, analyzing the multiple detection area images by establishing a boiler detection image model, dividing the boiler to be inspected into a first detection result boiler and a second detection result boiler according to the analysis result, obtaining boiler image processing data, performing sensor detection on the second detection result boiler according to the boiler image processing data, respectively obtaining a pipeline pressure detection coefficient, a pipeline temperature detection coefficient and a pipeline vibration frequency detection coefficient, obtaining boiler sensor data, performing nondestructive testing on the boiler pipe to be inspected according to the boiler sensor data and the boiler image processing data, and issuing an abnormal warning according to the detection result.

[0007] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: The specific working process of each module of a non-destructive testing system for boiler pressure pipes based on image processing is as follows:

[0008] Image data module: used to obtain multiple images of the inspection area inside the pipeline of the boiler to be inspected, analyze the multiple inspection area images by establishing a boiler inspection image model, divide the boiler to be inspected into a first inspection result boiler and a second inspection result boiler according to the analysis results, and obtain boiler image processing data;

[0009] Sensor module: used for performing sensor detection on the second detection result boiler according to the boiler image processing data, respectively obtaining the pipeline pressure detection coefficient, the pipeline temperature detection coefficient and the pipeline vibration frequency detection coefficient, and obtaining the boiler sensor data;

[0010] Nondestructive testing module: used to perform nondestructive testing on the boiler pipeline to be inspected based on boiler sensor data and boiler image processing data, and issue abnormal warnings based on the test results.

[0011] Furthermore, the image data module acquires boiler image processing data as follows:

[0012] The interior of the boiler pipe to be inspected is divided into a number of image inspection areas of equal area, and an image is acquired for each image inspection area respectively to obtain a plurality of inspection area images;

[0013] Create boiler inspection image model;

[0014] Using the boiler detection image model to analyze multiple detection area images, to obtain boiler image processing data;

[0015] The details are as follows:

[0016] Use the boiler detection model to detect and identify each detection area image. If the boiler detection image model detects that the detection area image is abnormal, the corresponding detection area image is marked as a first type of boiler internal image. If the boiler detection image model does not detect that the detection area image is abnormal, the corresponding detection area image is marked as a second type of boiler internal image.

[0017] If there is a first type of boiler internal image in the multiple detection area images corresponding to the pipeline of the boiler to be detected, the boiler to be detected is marked as a first detection result boiler;

[0018] If the first type of boiler internal image does not exist in the multiple detection area images corresponding to the pipeline of the boiler to be detected, the boiler to be detected is marked as a second detection result boiler.

[0019] Furthermore, the image data module creates a boiler detection image model as follows:

[0020] The data crawler technology is used to crawl images using the internal images of boiler tubes as keywords to obtain multiple boiler tube sample images;

[0021] By means of manual recognition, the parts of the boiler tube sample images with pipeline abnormalities are marked, and the boiler tube sample images with pipeline abnormalities are marked as abnormal boiler tube sample images;

[0022] Dividing a plurality of boiler tube sample images into an image test set and an image training set according to a preset division ratio;

[0023] Create an image recognition model through an existing artificial intelligence platform, and use the image training set to train the image recognition model in sequence until the image recognition model completes the training of each image in the image training set;

[0024] The trained image recognition model is tested using the image test set, and the recognition success rate of the image recognition model is obtained. When the recognition accuracy is lower than the target recognition accuracy, the image training set is continued to be used to train the image recognition model until the recognition accuracy is greater than or equal to the target recognition accuracy. When the recognition accuracy is greater than the target recognition accuracy, the image recognition model training is completed, and a boiler detection image model is obtained.

[0025] Furthermore, the image data module acquires boiler sensor data as follows:

[0026] Acquire boiler image processing data, and acquire a first detection result boiler and a second detection result boiler respectively according to the boiler image processing data;

[0027] During the period of normal operation of the boiler in the second test result, a pipeline nondestructive testing cycle is marked;

[0028] Perform air pressure testing on the second test result boiler in the pipeline nondestructive testing cycle to obtain a pipeline pressure testing coefficient;

[0029] Perform air temperature detection on the second detection result boiler in the pipeline nondestructive detection cycle to obtain the pipeline temperature detection coefficient;

[0030] Perform pipeline vibration frequency detection on the second detection result boiler in the pipeline nondestructive detection cycle to obtain a pipeline vibration frequency detection coefficient;

[0031] The pipeline pressure detection coefficient, the pipeline temperature detection coefficient, and the pipeline vibration frequency detection coefficient are defined as boiler sensor data.

[0032] Furthermore, the sensor module acquires the pipeline pressure detection coefficient as follows:

[0033] During the pipeline nondestructive testing cycle, a number of pressure testing points are randomly selected, and the air pressure value corresponding to each pressure testing point of the second test result boiler is respectively measured to obtain multiple air pressure values;

[0034] Calculate the average of the multiple air pressure values ​​obtained to obtain the detected average air pressure value;

[0035] The variance of the multiple air pressure values ​​obtained is calculated to obtain the detected air pressure variance;

[0036] Compare the obtained multiple air pressure values, mark the air pressure value with the largest value as the first characteristic air pressure value, mark the air pressure value with the smallest value as the second characteristic air pressure value, calculate the difference between the first characteristic air pressure value and the second characteristic air pressure value, and take the absolute value of the obtained difference to obtain the air pressure detection peak-to-valley difference;

[0037] The pipeline pressure detection coefficient is obtained by calculating the detected average air pressure value, the detected air pressure variance and the air pressure detection peak-to-valley difference;

[0038] Calculate the pipeline pressure detection coefficient.

[0039] Furthermore, the sensor module acquires the pipeline temperature detection coefficient as follows:

[0040] During the pipeline nondestructive testing cycle, a number of temperature detection points are randomly selected, and the air temperature value corresponding to each temperature detection point of the second test result boiler is respectively obtained to obtain multiple air temperature values;

[0041] Calculate the average of the multiple air temperature values ​​obtained to obtain the detected average air temperature value;

[0042] The variance of the multiple air temperature values ​​obtained is calculated to obtain the detected air temperature variance;

[0043] Compare the values ​​of the multiple air temperature values ​​obtained, mark the air temperature value with the largest value as the first characteristic air temperature value, mark the air temperature value with the smallest value as the second characteristic air temperature value, calculate the difference between the first characteristic air temperature value and the second characteristic air temperature value, and take the absolute value of the obtained difference to obtain the air temperature detection peak-to-valley difference;

[0044] The pipeline temperature detection coefficient is obtained by calculating the detected average air temperature value, the detected air temperature variance and the air temperature detection peak-to-valley difference;

[0045] Calculate the pipeline temperature detection coefficient.

[0046] Furthermore, the sensor module acquires the pipeline vibration frequency detection coefficient as follows:

[0047] During the pipeline nondestructive testing period, a number of vibration frequency detection points are randomly selected, and the pipeline vibration frequency value corresponding to each vibration frequency detection point of the second test result boiler is respectively obtained to obtain a plurality of pipeline vibration frequency values;

[0048] Calculate the average of the multiple pipeline vibration frequency values ​​obtained to obtain the detected average pipeline vibration frequency value;

[0049] The obtained multiple pipeline vibration frequency values ​​are subjected to variance calculation to obtain the detection pipeline vibration frequency variance;

[0050] Compare the values ​​of the multiple pipeline vibration frequency values ​​obtained, mark the pipeline vibration frequency value with the largest value as the first characteristic pipeline vibration frequency value, mark the pipeline vibration frequency value with the smallest value as the second characteristic pipeline vibration frequency value, calculate the difference between the first characteristic pipeline vibration frequency value and the second characteristic pipeline vibration frequency value, and take the absolute value of the obtained difference to obtain the peak-to-valley difference of air vibration frequency detection;

[0051] The pipeline vibration frequency detection coefficient is obtained by calculating the detected average pipeline vibration frequency value, the detected pipeline vibration frequency variance and the detected vibration frequency detection peak-to-valley difference;

[0052] Calculate the pipeline vibration frequency detection coefficient.

[0053] Furthermore, the nondestructive detection module performs boiler abnormality warning, as follows:

[0054] Obtain boiler image processing data and boiler sensor data, and obtain a first detection result boiler and a second detection result boiler respectively according to the boiler image processing data;

[0055] Issue a warning of unqualified test for the boiler with the first test result;

[0056] Performing sensor data detection on the second detection result boiler according to the boiler sensor data;

[0057] The details are as follows:

[0058] According to the boiler sensor data, the pipeline pressure detection coefficient, the pipeline temperature detection coefficient and the pipeline vibration frequency detection coefficient are respectively obtained;

[0059] Obtain pipeline pressure detection benchmark coefficient, pipeline temperature detection benchmark coefficient and pipeline vibration frequency detection benchmark coefficient respectively;

[0060] The pipeline pressure detection coefficient, the pipeline temperature detection coefficient, the pipeline vibration frequency detection coefficient, the pipeline pressure detection reference coefficient, the pipeline temperature detection reference coefficient and the pipeline vibration frequency detection reference coefficient are calculated to obtain the boiler abnormality detection coefficient;

[0061] Calculate the boiler abnormality detection coefficient;

[0062] Obtain the boiler abnormality detection coefficient threshold, perform numerical comparison between the boiler abnormality detection coefficient and the boiler abnormality detection coefficient threshold, and issue an abnormality warning based on the numerical comparison result.

[0063] Furthermore, the nondestructive testing module issues an early warning to the boiler of the second testing result, as follows:

[0064] Obtain pipeline pressure detection coefficient threshold, pipeline temperature detection coefficient threshold and pipeline vibration frequency detection coefficient threshold respectively;

[0065] The pipeline pressure detection coefficient threshold, pipeline temperature detection coefficient threshold, pipeline vibration frequency detection coefficient threshold, pipeline pressure detection reference coefficient, pipeline temperature detection reference coefficient and pipeline vibration frequency detection reference coefficient are calculated to obtain a boiler abnormality detection coefficient threshold;

[0066] When the boiler abnormality detection coefficient is greater than or equal to the boiler abnormality detection coefficient threshold, an abnormality warning is issued for the boiler pipeline;

[0067] When the boiler abnormality detection coefficient is less than the boiler abnormality detection coefficient threshold, no warning is issued for the boiler pipeline.

[0068] A nondestructive testing method for a boiler pressure pipe based on image processing comprises the following steps:

[0069] Step S1: obtaining a plurality of inspection area images of the interior of the pipeline of the boiler to be inspected, analyzing the plurality of inspection area images by establishing a boiler inspection image model, dividing the boiler to be inspected into a first inspection result boiler and a second inspection result boiler according to the analysis result, and obtaining boiler image processing data;

[0070] Step S2: performing sensor detection on the boiler of the second detection result according to the boiler image processing data, respectively obtaining a pipeline pressure detection coefficient, a pipeline temperature detection coefficient, and a pipeline vibration frequency detection coefficient, and obtaining boiler sensor data;

[0071] Step S3: Perform nondestructive testing on the boiler pipeline to be tested based on the boiler sensor data and the boiler image processing data, and issue an abnormal warning based on the test results.

[0072] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0073] 1. The present invention analyzes multiple detection area images by establishing a boiler detection image model, and divides the boiler to be detected into a first detection result boiler and a second detection result boiler according to the analysis result, thereby realizing preliminary detection of the boiler pressure pipeline and ensuring the intelligence of the boiler pressure pipeline detection process;

[0074] 2. The present invention obtains the pipeline pressure detection coefficient, pipeline temperature detection coefficient and pipeline vibration frequency detection coefficient respectively to further detect the boiler of the second detection result, so as to conduct a more comprehensive evaluation and detection of the overall operating status of the boiler pressure pipeline. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0076] Figure 1 is the overall system block diagram of the present invention;

[0077] Figure 2 It is a diagram of the implementation steps of the present invention;

[0078] Figure 3 This is a sample image of an abnormal boiler tube in the present invention. DETAILED DESCRIPTION

[0079] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0080] Embodiment 1

[0081] See also Figure 1 , the present invention provides a technical solution: a nondestructive testing system for a boiler pressure pipe based on image processing, comprising an image data module, a sensor module, a nondestructive testing module and a server, wherein the image data module, the sensor module and the nondestructive testing module are respectively connected to the server, and the server controls the image data module, the sensor module and the nondestructive testing module respectively;

[0082] The image data module obtains a plurality of inspection area images of the interior of the pipeline of the boiler to be inspected, analyzes the plurality of inspection area images by establishing a boiler inspection image model, divides the boiler to be inspected into a first inspection result boiler and a second inspection result boiler according to the analysis result, and obtains boiler image processing data;

[0083] The interior of the boiler pipe to be inspected is divided into a number of image inspection areas of equal area, and an image is acquired for each image inspection area respectively to obtain a plurality of inspection area images;

[0084] Create boiler inspection image model;

[0085] The details are as follows:

[0086] The data crawler technology is used to crawl images using the internal images of boiler tubes as keywords to obtain multiple boiler tube sample images;

[0087] See also Figure 3, by means of manual recognition, the parts of the boiler tube sample images with pipeline anomalies are marked, and the boiler tube sample images with pipeline anomalies are marked as abnormal boiler tube sample images;

[0088] It should be noted here that:

[0089] The specific situations corresponding to the pipeline anomalies existing in the boiler tube sample images involved herein include but are not limited to corrosion, cracks and deformation;

[0090] Dividing a plurality of boiler tube sample images into an image test set and an image training set according to a preset division ratio;

[0091] It should be noted here that:

[0092] In the present application, the preset division ratio involved here is specifically 3:7, that is, the division ratio of the image test set and the image training set is 3:7;

[0093] Create an image recognition model through an existing artificial intelligence platform, and use the image training set to train the image recognition model in sequence until the image recognition model completes the training of each image in the image training set;

[0094] The trained image recognition model is tested by the image test set, and the recognition success rate of the image recognition model is obtained. When the recognition accuracy rate is less than the target recognition accuracy rate, the image recognition model is continuously trained using the image training set until the recognition accuracy rate is greater than or equal to the target recognition accuracy rate. When the recognition accuracy rate is greater than the target recognition accuracy rate, the image recognition model training is completed, and a boiler detection image model is obtained.

[0095] It should be noted here that:

[0096] In this application, the target recognition accuracy is defined as 95% here;

[0097] Using the boiler detection image model to analyze multiple detection area images, to obtain boiler image processing data;

[0098] The details are as follows:

[0099] Use the boiler detection model to detect and identify each detection area image. If the boiler detection image model detects that the detection area image is abnormal, the corresponding detection area image is marked as a first type of boiler internal image. If the boiler detection image model does not detect that the detection area image is abnormal, the corresponding detection area image is marked as a second type of boiler internal image.

[0100] If there is a first type of boiler internal image in the plurality of inspection area images corresponding to the pipeline of the boiler to be inspected, the boiler to be inspected is marked as a first inspection result boiler;

[0101] If the first type of boiler internal image does not exist in the multiple detection area images corresponding to the pipeline of the boiler to be detected, the boiler to be detected is marked as a second detection result boiler;

[0102] The image data module acquires the boiler image processing data and transmits it to the sensor module and the nondestructive testing module;

[0103] The sensor module performs sensor detection on the second detection result boiler according to the boiler image processing data, obtains the pipeline pressure detection coefficient, the pipeline temperature detection coefficient and the pipeline vibration frequency detection coefficient respectively, and obtains the boiler sensor data;

[0104] Acquire boiler image processing data, and acquire a first detection result boiler and a second detection result boiler respectively according to the boiler image processing data;

[0105] During the period of normal operation of the boiler in the second test result, a pipeline nondestructive testing cycle is marked;

[0106] It should be noted here that:

[0107] In the present application, the second test result shows that the boiler maintains the same working condition during the pipeline nondestructive testing cycle;

[0108] Perform air pressure testing on the second test result boiler in the pipeline nondestructive testing cycle to obtain a pipeline pressure testing coefficient;

[0109] The details are as follows:

[0110] During the pipeline nondestructive testing cycle, a number of pressure testing points are randomly selected, and the air pressure value corresponding to each pressure testing point of the second test result boiler is respectively measured to obtain multiple air pressure values;

[0111] Calculate the average of the multiple air pressure values ​​obtained to obtain the detected average air pressure value;

[0112] The variance of the multiple air pressure values ​​obtained is calculated to obtain the detected air pressure variance;

[0113] Compare the obtained multiple air pressure values, mark the air pressure value with the largest value as the first characteristic air pressure value, mark the air pressure value with the smallest value as the second characteristic air pressure value, calculate the difference between the first characteristic air pressure value and the second characteristic air pressure value, and take the absolute value of the obtained difference to obtain the air pressure detection peak-to-valley difference;

[0114] The pipeline pressure detection coefficient is obtained by calculating the detected average air pressure value, the detected air pressure variance and the air pressure detection peak-to-valley difference;

[0115] The pipeline pressure detection coefficient is calculated, and the specific formula is as follows:

[0116] Kyx=Ylp+Ylc+Yfg;

[0117] Among them, Kyx is the pipeline pressure detection coefficient, Y lp is the detected average air pressure value, Y lc is the detected air pressure variance, and Yfg is the peak-to-valley difference of air pressure detection;

[0118] Perform air temperature detection on the second detection result boiler in the pipeline nondestructive detection cycle to obtain the pipeline temperature detection coefficient;

[0119] The details are as follows:

[0120] During the pipeline nondestructive testing cycle, a number of temperature detection points are randomly selected, and the air temperature value corresponding to each temperature detection point of the second test result boiler is respectively obtained to obtain multiple air temperature values;

[0121] Calculate the average of the multiple air temperature values ​​obtained to obtain the detected average air temperature value;

[0122] The variance of the multiple air temperature values ​​obtained is calculated to obtain the detected air temperature variance;

[0123] Compare the values ​​of the multiple air temperature values ​​obtained, mark the air temperature value with the largest value as the first characteristic air temperature value, mark the air temperature value with the smallest value as the second characteristic air temperature value, calculate the difference between the first characteristic air temperature value and the second characteristic air temperature value, and take the absolute value of the obtained difference to obtain the air temperature detection peak-to-valley difference;

[0124] The pipeline temperature detection coefficient is obtained by calculating the detected average air temperature value, the detected air temperature variance and the air temperature detection peak-to-valley difference;

[0125] Calculate the pipeline temperature detection coefficient, the specific formula is as follows:

[0126] Wyx=Wlp+Wlc+Wfg;

[0127] Among them, Wyx is the pipeline temperature detection coefficient, Wl p is the detected average air temperature value, Wl c is the detected air temperature variance, and Wfg is the peak-to-valley difference of air temperature detection;

[0128] Perform pipeline vibration frequency detection on the second detection result boiler in the pipeline nondestructive detection cycle to obtain a pipeline vibration frequency detection coefficient;

[0129] The details are as follows:

[0130] During the pipeline nondestructive testing period, a number of vibration frequency detection points are randomly selected, and the pipeline vibration frequency value corresponding to each vibration frequency detection point of the second test result boiler is respectively obtained to obtain a plurality of pipeline vibration frequency values;

[0131] Calculate the average of the multiple pipeline vibration frequency values ​​obtained to obtain the detected average pipeline vibration frequency value;

[0132] The obtained multiple pipeline vibration frequency values ​​are subjected to variance calculation to obtain the detection pipeline vibration frequency variance;

[0133] Compare the values ​​of the multiple pipeline vibration frequency values ​​obtained, mark the pipeline vibration frequency value with the largest value as the first characteristic pipeline vibration frequency value, mark the pipeline vibration frequency value with the smallest value as the second characteristic pipeline vibration frequency value, calculate the difference between the first characteristic pipeline vibration frequency value and the second characteristic pipeline vibration frequency value, and take the absolute value of the obtained difference to obtain the peak-to-valley difference of air vibration frequency detection;

[0134] The pipeline vibration frequency detection coefficient is obtained by calculating the detected average pipeline vibration frequency value, the detected pipeline vibration frequency variance and the detected vibration frequency detection peak-to-valley difference;

[0135] The pipeline vibration frequency detection coefficient is calculated, and the specific formula is as follows:

[0136] Gyx=Glp+Glc+Gfg;

[0137] Among them, Gyx is the pipeline vibration frequency detection coefficient, G lp is the detected average pipeline vibration frequency value, G lc is the detected pipeline vibration frequency variance, and Gfg is the peak-to-valley difference of pipeline vibration frequency detection;

[0138] The pipeline pressure detection coefficient, pipeline temperature detection coefficient and pipeline vibration frequency detection coefficient are defined as boiler sensor data;

[0139] The sensor data module acquires the boiler sensor data and transmits it to the nondestructive testing module;

[0140] The non-destructive testing module performs non-destructive testing on the boiler pipeline to be tested based on the boiler sensor data and boiler image processing data, and issues abnormal warnings based on the test results;

[0141] Obtain boiler image processing data and boiler sensor data, and obtain a first detection result boiler and a second detection result boiler respectively according to the boiler image processing data;

[0142] Issue a warning of unqualified test for the boiler with the first test result;

[0143] Performing sensor data detection on the second detection result boiler according to the boiler sensor data;

[0144] The details are as follows:

[0145] According to the boiler sensor data, the pipeline pressure detection coefficient, the pipeline temperature detection coefficient and the pipeline vibration frequency detection coefficient are respectively obtained;

[0146] Obtain pipeline pressure detection benchmark coefficient, pipeline temperature detection benchmark coefficient and pipeline vibration frequency detection benchmark coefficient respectively;

[0147] It should be noted here that:

[0148] The pressure detection reference coefficient, pipeline temperature detection reference coefficient and pipeline vibration frequency detection reference coefficient involved here are respectively the optimal pressure detection coefficient, optimal temperature detection coefficient and optimal vibration frequency detection coefficient of the boiler under the current working condition;

[0149] The pipeline pressure detection coefficient, the pipeline temperature detection coefficient, the pipeline vibration frequency detection coefficient, the pipeline pressure detection reference coefficient, the pipeline temperature detection reference coefficient and the pipeline vibration frequency detection reference coefficient are calculated to obtain the boiler abnormality detection coefficient;

[0150] The boiler abnormality detection coefficient is calculated, and the specific formula is as follows:

[0151] Gyy=|Kyx-Kyxj|+|Wyx-Wyxj|+|Gyx-Gyxj|;

[0152] Among them, Gyy is the boiler abnormality detection coefficient, Kyx is the pipeline pressure detection coefficient, Wyx is the pipeline temperature detection coefficient, Gyx is the pipeline vibration frequency detection coefficient, Kyxj is the pipeline pressure detection coefficient, Wyxj is the pipeline temperature detection coefficient, and Gyxj is the pipeline vibration frequency detection coefficient;

[0153] Obtain a boiler abnormality detection coefficient threshold, compare the boiler abnormality detection coefficient with the boiler abnormality detection coefficient threshold, and issue an abnormality warning based on the numerical comparison result;

[0154] The details are as follows:

[0155] Obtain pipeline pressure detection coefficient threshold, pipeline temperature detection coefficient threshold and pipeline vibration frequency detection coefficient threshold respectively;

[0156] The pipeline pressure detection coefficient threshold, pipeline temperature detection coefficient threshold, pipeline vibration frequency detection coefficient threshold, pipeline pressure detection reference coefficient, pipeline temperature detection reference coefficient and pipeline vibration frequency detection reference coefficient are calculated to obtain a boiler abnormality detection coefficient threshold;

[0157] The boiler abnormality detection coefficient threshold is calculated, and the specific formula is as follows:

[0158] Gyyy=|Kyxy-Kyxj|+|Wyxy-Wyxj|+|Gyxy-Gyxj|;

[0159] Among them, Gyyy is the boiler abnormality detection coefficient threshold, Kyxy is the pipeline pressure detection coefficient threshold, Wyxy is the pipeline temperature detection coefficient threshold, Gyxy is the pipeline vibration frequency detection coefficient, Kyxj is the pipeline pressure detection coefficient, Wyxj is the pipeline temperature detection coefficient, and Gyxj is the pipeline vibration frequency detection coefficient;

[0160] When the boiler abnormality detection coefficient is greater than or equal to the boiler abnormality detection coefficient threshold, an abnormality warning is issued for the boiler pipeline;

[0161] When the boiler abnormality detection coefficient is less than the boiler abnormality detection coefficient threshold, no warning is issued for the boiler pipeline.

[0162] In this application, if corresponding calculation formulas appear, the above calculation formulas are all dimensionless and take their numerical calculations. The weight coefficients, proportional coefficients and other coefficients in the formulas are set to a result value obtained by quantifying each parameter. The size of the weight coefficient and the proportional coefficient can be determined as long as it does not affect the proportional relationship between the parameter and the result value.

[0163] Embodiment 2

[0164] See also Figure 2 Based on another concept of the same invention, a nondestructive testing method for boiler pressure pipes based on image processing is now proposed, comprising the following steps:

[0165] Step S1: obtaining a plurality of inspection area images of the interior of the pipeline of the boiler to be inspected, analyzing the plurality of inspection area images by establishing a boiler inspection image model, dividing the boiler to be inspected into a first inspection result boiler and a second inspection result boiler according to the analysis result, and obtaining boiler image processing data;

[0166] Step S11: Divide the interior of the boiler pipe to be inspected into a number of image inspection areas of equal area, and acquire an image of each image inspection area to obtain a plurality of inspection area images;

[0167] Step S12: creating a boiler detection image model;

[0168] The step S12 further includes the following specific steps:

[0169] Step S121: crawling images using the data crawler technology with the internal image of the boiler pipe as a keyword to obtain multiple boiler pipe sample images;

[0170] Step S122: marking the parts of the boiler tube sample images where there are pipeline anomalies by manual recognition, and marking the boiler tube sample images where there are pipeline anomalies as abnormal boiler tube sample images;

[0171] Step S123: dividing the plurality of boiler tube sample images into an image test set and an image training set according to a preset division ratio;

[0172] Step S124: creating an image recognition model through an existing artificial intelligence platform, and using the image training set to train the image recognition model in sequence until the image recognition model completes the training of each image in the image training set;

[0173] Step S125: testing the trained image recognition model through the image test set, and obtaining the recognition success rate of the image recognition model. When the recognition accuracy rate is less than the target recognition accuracy rate, continue to train the image recognition model using the image training set until the recognition accuracy rate is greater than or equal to the target recognition accuracy rate. When the recognition accuracy rate is greater than the target recognition accuracy rate, the image recognition model training is completed, and a boiler detection image model is obtained.

[0174] Step S13: using the boiler detection image model to analyze multiple detection area images to obtain boiler image processing data;

[0175] The step S13 further includes the following specific steps:

[0176] Step S131: Use the boiler detection model to detect and identify each detection area image. If the boiler detection image model detects that the detection area image is abnormal, the corresponding detection area image is marked as a first type of boiler internal image. If the boiler detection image model does not detect that the detection area image is abnormal, the corresponding detection area image is marked as a second type of boiler internal image.

[0177] Step S132: if there is a first type of boiler internal image in the plurality of inspection area images corresponding to the pipeline of the boiler to be inspected, the boiler to be inspected is marked as a first inspection result boiler;

[0178] Step S134: if the first type of boiler internal image does not exist in the plurality of detection area images corresponding to the pipeline of the boiler to be detected, the boiler to be detected is marked as a second detection result boiler;

[0179] Step S2: performing sensor detection on the boiler of the second detection result according to the boiler image processing data, respectively obtaining a pipeline pressure detection coefficient, a pipeline temperature detection coefficient, and a pipeline vibration frequency detection coefficient, and obtaining boiler sensor data;

[0180] Step S21: Obtain boiler image processing data, and obtain a first detection result boiler and a second detection result boiler respectively according to the boiler image processing data;

[0181] Step S22: marking a pipeline nondestructive testing cycle within the period of time when the second test result indicates that the boiler is operating normally;

[0182] Step S23: Perform air pressure detection on the second detection result boiler in the pipeline nondestructive detection cycle to obtain a pipeline pressure detection coefficient;

[0183] The step S23 further includes the following specific steps:

[0184] Step S231: randomly selecting a number of pressure detection points within the pipeline nondestructive testing cycle, and respectively performing a second test result on the air pressure value of the boiler corresponding to each pressure detection point to obtain a plurality of air pressure values;

[0185] Step S232: Calculate the average of the multiple air pressure values ​​obtained to obtain the detected average air pressure value;

[0186] Step S233: performing variance calculation on the obtained multiple air pressure values ​​to obtain the detected air pressure variance;

[0187] Step S234: compare the obtained multiple air pressure values, mark the air pressure value with the largest value as the first characteristic air pressure value, mark the air pressure value with the smallest value as the second characteristic air pressure value, calculate the difference between the first characteristic air pressure value and the second characteristic air pressure value, and take the absolute value of the obtained difference to obtain the air pressure detection peak-to-valley difference;

[0188] Step S235: Calculate the average air pressure value, the air pressure variance and the peak-to-valley difference of the air pressure detection to obtain the pipeline pressure detection coefficient;

[0189] The pipeline pressure detection coefficient is calculated, and the specific formula is as follows:

[0190] Kyx=Ylp+Ylc+Yfg;

[0191] Among them, Kyx is the pipeline pressure detection coefficient, Y lp is the detected average air pressure value, Y lc is the detected air pressure variance, and Yfg is the peak-to-valley difference of air pressure detection;

[0192] Step S24: performing air temperature detection on the second detection result boiler in the pipeline nondestructive detection cycle to obtain a pipeline temperature detection coefficient;

[0193] The step S24 further includes the following specific steps:

[0194] Step S241: randomly selecting a number of temperature detection points within the pipeline nondestructive testing cycle, and respectively performing a second test result on the air temperature value corresponding to each temperature detection point of the boiler to obtain a plurality of air temperature values;

[0195] Step S242: Calculate the average of the multiple air temperature values ​​obtained to obtain the detected average air temperature value;

[0196] Step S243: performing variance calculation on the obtained multiple air temperature values ​​to obtain the detected air temperature variance;

[0197] Step S244: compare the obtained multiple air temperature values, mark the air temperature value with the largest value as the first characteristic air temperature value, mark the air temperature value with the smallest value as the second characteristic air temperature value, calculate the difference between the first characteristic air temperature value and the second characteristic air temperature value, and take the absolute value of the obtained difference to obtain the air temperature detection peak-to-valley difference;

[0198] Step S245: Calculate the detected average air temperature value, the detected air temperature variance, and the air temperature detection peak-to-valley difference to obtain a pipeline temperature detection coefficient;

[0199] Calculate the pipeline temperature detection coefficient, the specific formula is as follows:

[0200] Wyx=Wlp+Wlc+Wfg;

[0201] Among them, Wyx is the pipeline temperature detection coefficient, Wl p is the detected average air temperature value, Wl c is the detected air temperature variance, and Wfg is the peak-to-valley difference of air temperature detection;

[0202] Step S25: performing pipeline vibration frequency detection on the second detection result boiler in the pipeline nondestructive detection cycle to obtain a pipeline vibration frequency detection coefficient;

[0203] The step S25 further includes the following specific steps:

[0204] Step S251: randomly selecting a number of vibration frequency detection points within the pipeline nondestructive testing cycle, and respectively calculating the pipeline vibration frequency value corresponding to each vibration frequency detection point of the second test result boiler to obtain a plurality of pipeline vibration frequency values;

[0205] Step S252: Calculate the average of the multiple pipeline vibration frequency values ​​obtained to obtain the detected average pipeline vibration frequency value;

[0206] Step S253: performing variance calculation on the obtained multiple pipeline vibration frequency values ​​to obtain the detection pipeline vibration frequency variance;

[0207] Step S254: compare the obtained multiple pipeline vibration frequency values, mark the pipeline vibration frequency value with the largest value as the first characteristic pipeline vibration frequency value, mark the pipeline vibration frequency value with the smallest value as the second characteristic pipeline vibration frequency value, calculate the difference between the first characteristic pipeline vibration frequency value and the second characteristic pipeline vibration frequency value, and take the absolute value of the obtained difference to obtain the peak-to-valley difference of air vibration frequency detection;

[0208] Step S255: Calculate the pipeline vibration frequency detection coefficient by detecting the average pipeline vibration frequency value, the pipeline vibration frequency variance, and the vibration frequency detection peak-to-valley difference;

[0209] The pipeline vibration frequency detection coefficient is calculated, and the specific formula is as follows:

[0210] Gyx=Glp+Glc+Gfg;

[0211] Among them, Gyx is the pipeline vibration frequency detection coefficient, G lp is the detected average pipeline vibration frequency value, G lc is the detected pipeline vibration frequency variance, and Gfg is the peak-to-valley difference of pipeline vibration frequency detection;

[0212] Step S26: defining the pipeline pressure detection coefficient, the pipeline temperature detection coefficient and the pipeline vibration frequency detection coefficient as boiler sensor data;

[0213] Step S3: Perform nondestructive testing on the boiler pipeline to be tested based on the boiler sensor data and the boiler image processing data, and issue an abnormal warning based on the test results;

[0214] Step S31: Obtain boiler image processing data and boiler sensor data, and obtain a first detection result boiler and a second detection result boiler respectively according to the boiler image processing data;

[0215] Step S32: issuing a warning of unqualified detection for the boiler with the first detection result;

[0216] Step S33: performing sensor data detection on the boiler of the second detection result according to the boiler sensor data;

[0217] The step S33 further includes the following specific steps:

[0218] Step S331: acquiring the pipeline pressure detection coefficient, pipeline temperature detection coefficient and pipeline vibration frequency detection coefficient respectively according to the boiler sensor data;

[0219] Step S332: respectively obtaining a pipeline pressure detection reference coefficient, a pipeline temperature detection reference coefficient, and a pipeline vibration frequency detection reference coefficient;

[0220] Step S333: obtaining a boiler abnormality detection coefficient by calculating the pipeline pressure detection coefficient, the pipeline temperature detection coefficient, the pipeline vibration frequency detection coefficient, the pipeline pressure detection reference coefficient, the pipeline temperature detection reference coefficient and the pipeline vibration frequency detection reference coefficient;

[0221] The boiler abnormality detection coefficient is calculated, and the specific formula is as follows:

[0222] Gyy=|Kyx-Kyxj|+|Wyx-Wyxj|+|Gyx-Gyxj|;

[0223] Among them, Gyy is the boiler abnormality detection coefficient, Kyx is the pipeline pressure detection coefficient, Wyx is the pipeline temperature detection coefficient, Gyx is the pipeline vibration frequency detection coefficient, Kyxj is the pipeline pressure detection coefficient, Wyxj is the pipeline temperature detection coefficient, and Gyxj is the pipeline vibration frequency detection coefficient;

[0224] Step S334: obtaining a boiler abnormality detection coefficient threshold, performing a numerical comparison between the boiler abnormality detection coefficient and the boiler abnormality detection coefficient threshold, and issuing an abnormality warning according to the numerical comparison result;

[0225] The step S334 further includes the following specific steps:

[0226] Step S3341: respectively obtaining a pipeline pressure detection coefficient threshold, a pipeline temperature detection coefficient threshold, and a pipeline vibration frequency detection coefficient threshold;

[0227] Step S3342: Calculate the pipeline pressure detection coefficient threshold, pipeline temperature detection coefficient threshold, pipeline vibration frequency detection coefficient threshold, pipeline pressure detection reference coefficient, pipeline temperature detection reference coefficient, and pipeline vibration frequency detection reference coefficient to obtain a boiler abnormality detection coefficient threshold;

[0228] Step S3343: When the boiler abnormality detection coefficient is greater than or equal to the boiler abnormality detection coefficient threshold, an abnormality warning is issued for the boiler pipeline;

[0229] Step S3344: When the boiler abnormality detection coefficient is less than the boiler abnormality detection coefficient threshold, no warning is issued for the boiler pipeline.

[0230] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well.

Claims

1. A nondestructive testing system for boiler pressure pipes based on image processing, comprising: Image data module: used to obtain multiple images of the inspection area inside the pipeline of the boiler to be inspected, analyze the multiple inspection area images by establishing a boiler inspection image model, divide the boiler to be inspected into a first inspection result boiler and a second inspection result boiler according to the analysis results, and obtain boiler image processing data; Sensor module: used for performing sensor detection on the second detection result boiler according to the boiler image processing data, respectively obtaining the pipeline pressure detection coefficient, the pipeline temperature detection coefficient and the pipeline vibration frequency detection coefficient, and obtaining the boiler sensor data; Nondestructive testing module: used to perform nondestructive testing on the boiler pipeline to be inspected based on the boiler sensor data and boiler image processing data, and issue abnormal warnings based on the test results.

2. The nondestructive testing system for boiler pressure pipes based on image processing according to claim 1 is characterized in that: The image data module acquires boiler image processing data, specifically as follows: The interior of the boiler pipe to be inspected is divided into a number of image inspection areas of equal area, and an image is acquired for each image inspection area respectively to obtain a plurality of inspection area images; Create boiler inspection image model; Using the boiler detection image model to analyze multiple detection area images, to obtain boiler image processing data; The details are as follows: Use the boiler detection model to detect and identify each detection area image. If the boiler detection image model detects that the detection area image is abnormal, the corresponding detection area image is marked as a first type of boiler internal image. If the boiler detection image model does not detect that the detection area image is abnormal, the corresponding detection area image is marked as a second type of boiler internal image. If there is a first type of boiler internal image in the plurality of inspection area images corresponding to the pipeline of the boiler to be inspected, the boiler to be inspected is marked as a first inspection result boiler; If the first type of boiler internal image does not exist in the multiple detection area images corresponding to the pipeline of the boiler to be detected, the boiler to be detected is marked as a second detection result boiler.

3. The nondestructive testing system for boiler pressure pipes based on image processing according to claim 2 is characterized in that: The image data module creates a boiler detection image model, as follows: The data crawler technology is used to crawl images using the internal images of boiler tubes as keywords to obtain multiple boiler tube sample images; By means of manual recognition, the parts of the boiler tube sample images where there are pipeline anomalies are marked, and the boiler tube sample images where there are pipeline anomalies are marked as abnormal boiler tube sample images; Dividing a plurality of boiler tube sample images into an image test set and an image training set according to a preset division ratio; Create an image recognition model through an existing artificial intelligence platform, and use the image training set to train the image recognition model in sequence until the image recognition model completes the training of each image in the image training set; The trained image recognition model is tested using the image test set, and the recognition success rate of the image recognition model is obtained. When the recognition accuracy is lower than the target recognition accuracy, the image training set is continued to be used to train the image recognition model until the recognition accuracy is greater than or equal to the target recognition accuracy. When the recognition accuracy is greater than the target recognition accuracy, the image recognition model training is completed, and a boiler detection image model is obtained.

4. The nondestructive testing system for boiler pressure pipes based on image processing according to claim 1 is characterized in that: The image data module acquires the boiler sensor data as follows: Acquire boiler image processing data, and acquire a first detection result boiler and a second detection result boiler respectively according to the boiler image processing data; During the period of normal operation of the boiler in the second test result, a pipeline nondestructive testing cycle is marked; Perform air pressure testing on the second test result boiler in the pipeline nondestructive testing cycle to obtain a pipeline pressure testing coefficient; Perform air temperature detection on the second detection result boiler in the pipeline nondestructive detection cycle to obtain the pipeline temperature detection coefficient; Perform pipeline vibration frequency detection on the second detection result boiler in the pipeline nondestructive detection cycle to obtain a pipeline vibration frequency detection coefficient; The pipeline pressure detection coefficient, the pipeline temperature detection coefficient, and the pipeline vibration frequency detection coefficient are defined as boiler sensor data.

5. The nondestructive testing system for boiler pressure pipes based on image processing according to claim 4 is characterized in that: The sensor module obtains the pipeline pressure detection coefficient as follows: During the pipeline nondestructive testing cycle, a number of pressure testing points are randomly selected, and the air pressure value corresponding to each pressure testing point of the second test result boiler is respectively measured to obtain multiple air pressure values; Calculate the average of the multiple air pressure values ​​obtained to obtain the detected average air pressure value; The variance of the multiple air pressure values ​​obtained is calculated to obtain the detected air pressure variance; Compare the obtained multiple air pressure values, mark the air pressure value with the largest value as the first characteristic air pressure value, mark the air pressure value with the smallest value as the second characteristic air pressure value, calculate the difference between the first characteristic air pressure value and the second characteristic air pressure value, and take the absolute value of the obtained difference to obtain the air pressure detection peak-to-valley difference; The pipeline pressure detection coefficient is obtained by calculating the detected average air pressure value, the detected air pressure variance and the air pressure detection peak-to-valley difference; Calculate the pipeline pressure detection coefficient.

6. The nondestructive testing system for boiler pressure pipes based on image processing according to claim 4 is characterized in that: The sensor module obtains the pipeline temperature detection coefficient as follows: During the pipeline nondestructive testing cycle, a number of temperature detection points are randomly selected, and the air temperature value corresponding to each temperature detection point of the second test result boiler is respectively obtained to obtain multiple air temperature values; Calculate the average of the multiple air temperature values ​​obtained to obtain the detected average air temperature value; The variance of the multiple air temperature values ​​obtained is calculated to obtain the detected air temperature variance; Compare the values ​​of the multiple air temperature values ​​obtained, mark the air temperature value with the largest value as the first characteristic air temperature value, mark the air temperature value with the smallest value as the second characteristic air temperature value, calculate the difference between the first characteristic air temperature value and the second characteristic air temperature value, and take the absolute value of the obtained difference to obtain the air temperature detection peak-to-valley difference; The pipeline temperature detection coefficient is obtained by calculating the detected average air temperature value, the detected air temperature variance and the air temperature detection peak-to-valley difference; Calculate the pipeline temperature detection coefficient.

7. The nondestructive testing system for boiler pressure pipes based on image processing according to claim 4 is characterized in that: The sensor module obtains the pipeline vibration frequency detection coefficient as follows: During the pipeline nondestructive testing period, a number of vibration frequency detection points are randomly selected, and the pipeline vibration frequency value corresponding to each vibration frequency detection point of the second test result boiler is respectively obtained to obtain a plurality of pipeline vibration frequency values; Calculate the average of the multiple pipeline vibration frequency values ​​obtained to obtain the detected average pipeline vibration frequency value; The obtained multiple pipeline vibration frequency values ​​are subjected to variance calculation to obtain the detection pipeline vibration frequency variance; Compare the values ​​of the multiple pipeline vibration frequency values ​​obtained, mark the pipeline vibration frequency value with the largest value as the first characteristic pipeline vibration frequency value, mark the pipeline vibration frequency value with the smallest value as the second characteristic pipeline vibration frequency value, calculate the difference between the first characteristic pipeline vibration frequency value and the second characteristic pipeline vibration frequency value, and take the absolute value of the obtained difference to obtain the peak-to-valley difference of air vibration frequency detection; The pipeline vibration frequency detection coefficient is obtained by calculating the detected average pipeline vibration frequency value, the detected pipeline vibration frequency variance and the detected vibration frequency detection peak-to-valley difference; Calculate the pipeline vibration frequency detection coefficient.

8. The nondestructive testing system for boiler pressure pipes based on image processing according to claim 1 is characterized in that: The nondestructive detection module performs boiler abnormality warning, as follows: Obtain boiler image processing data and boiler sensor data, and obtain a first detection result boiler and a second detection result boiler respectively according to the boiler image processing data; Issue a warning of unqualified test for the boiler with the first test result; Performing sensor data detection on the second detection result boiler according to the boiler sensor data; The details are as follows: According to the boiler sensor data, the pipeline pressure detection coefficient, the pipeline temperature detection coefficient and the pipeline vibration frequency detection coefficient are respectively obtained; Obtain pipeline pressure detection benchmark coefficient, pipeline temperature detection benchmark coefficient and pipeline vibration frequency detection benchmark coefficient respectively; The pipeline pressure detection coefficient, the pipeline temperature detection coefficient, the pipeline vibration frequency detection coefficient, the pipeline pressure detection reference coefficient, the pipeline temperature detection reference coefficient and the pipeline vibration frequency detection reference coefficient are calculated to obtain the boiler abnormality detection coefficient; Calculate the boiler abnormality detection coefficient; Obtain the boiler abnormality detection coefficient threshold, perform numerical comparison between the boiler abnormality detection coefficient and the boiler abnormality detection coefficient threshold, and issue an abnormality warning based on the numerical comparison result.

9. The nondestructive testing system for boiler pressure pipes based on image processing according to claim 8 is characterized in that: The nondestructive testing module issues an early warning to the boiler with the second test result, as follows: Obtain pipeline pressure detection coefficient threshold, pipeline temperature detection coefficient threshold and pipeline vibration frequency detection coefficient threshold respectively; The pipeline pressure detection coefficient threshold, pipeline temperature detection coefficient threshold, pipeline vibration frequency detection coefficient threshold, pipeline pressure detection reference coefficient, pipeline temperature detection reference coefficient and pipeline vibration frequency detection reference coefficient are calculated to obtain a boiler abnormality detection coefficient threshold; When the boiler abnormality detection coefficient is greater than or equal to the boiler abnormality detection coefficient threshold, an abnormality warning is issued for the boiler pipeline; When the boiler abnormality detection coefficient is less than the boiler abnormality detection coefficient threshold, no warning is issued for the boiler pipeline.

10. A non-destructive testing method for boiler pressure piping based on image processing, applicable to a non-destructive testing system for boiler pressure piping based on image processing according to any one of claims 1 to 9, characterized in that: The nondestructive testing method comprises the following steps: Step S1: obtaining a plurality of inspection area images of the interior of the pipeline of the boiler to be inspected, analyzing the plurality of inspection area images by establishing a boiler inspection image model, dividing the boiler to be inspected into a first inspection result boiler and a second inspection result boiler according to the analysis result, and obtaining boiler image processing data; Step S2: performing sensor detection on the boiler of the second detection result according to the boiler image processing data, respectively obtaining a pipeline pressure detection coefficient, a pipeline temperature detection coefficient, and a pipeline vibration frequency detection coefficient, and obtaining boiler sensor data; Step S3: Perform nondestructive testing on the boiler pipeline to be tested based on the boiler sensor data and the boiler image processing data, and issue an abnormal warning based on the test results.

Images