Composite fiber sucker rod detection method

Abstract

The invention discloses a method for detecting a composite fiber sucker rod. The method comprises the following steps: S1, establishing a microscopic image detection model of the composite fiber sucker rod; s2, preparing a detection sample, and determining an observation surface; s3, shooting the observation surface of the detection sample to obtain a target microscopic image; s4, the target microscopic image obtained in the step S3 is input into the microscopic image detection model, established in the step S1, of the composite fiber sucker rod, a detection result is obtained through analysis, the composite fiber sucker rod to be detected can be rapidly detected through the model, the quality problem of the composite fiber sucker rod can be determined in a classified mode, the detection accuracy is high, and the detection efficiency is high.

Description

Technical Field

[0001] The invention relates to the field of petrochemical detection, in particular to a composite fiber sucker rod detection method. Background Art

[0002] Composite sucker rod is a new type of structural material that appeared in the mid-1960s and is one of the most advanced high-performance composite materials. It is a composite continuous sucker rod reinforced by high-modulus, high-strength carbon fiber and large deformation, good impact resistance glass fiber. The carbon fiber is in the core of the rod body, mainly bearing the longitudinal tensile load, and the glass fiber is on the outside, mainly providing lateral impact protection. Composite sucker rods completely overcome the shortcomings of traditional steel sucker rods such as heavy weight, easy corrosion, easy fatigue fracture, joint breakage and piston effect, and provide good technical support for the increasingly complex oil reservoir development. The composite sucker rod can not only improve oil production efficiency and reduce the accident rate, but also greatly reduce oil production costs. The composite fiber continuous sucker rod widely adopts pultrusion molding process in the production process. Due to the particularity of its manufacturing process and constituent materials, the sucker rods produced under reasonable process may still have quality problems.

[0003] The commonly used non-destructive testing methods for the production and application of this type of composite materials are currently five conventional technologies: X-ray testing, ultrasonic testing, magnetic particle testing, penetration testing and eddy current testing. Ultrasonic testing, digital image testing and pulse infrared thermal imaging testing have been widely used in the field of industrial production, but the testing methods of composite products in the field of petroleum and petrochemicals are still in the exploratory stage, especially for the fiber-resin interface relationship in the pultrusion process, aging, fatigue and other failure mechanisms and life prediction technologies caused by long-term use. There are few studies, and the relevant detection system is imperfect. Therefore, in recent years, the detection method based on cross-sectional scanning electron microscope microscopic images has attracted the attention of researchers and has gradually been applied. However, the mainstream image analysis method has the disadvantages of low accuracy, imperfect functions and low degree of automation. How to realize the specification index detection of composite fiber sucker rods in the detection and evaluation of oilfield enterprises and improve the detection efficiency and accuracy is a technical problem to be solved by technicians in this field. Summary of the invention

[0004] In view of the above problems, the present invention provides a composite fiber sucker rod detection method, which has the advantages of high detection efficiency and reliable detection results.

[0005] The technical solution of the present invention is:

[0006] A composite fiber sucker rod detection method comprises the following steps:

[0007] S1. Establishing a microscopic image detection model for composite fiber sucker rods;

[0008] S2, prepare the test sample and determine the observation surface;

[0009] S3, photographing the observation surface of the test sample to obtain a target microscopic image;

[0010] S4, inputting the target microscopic image obtained in step S3 into the composite fiber sucker rod microscopic image detection model established in step S1, and analyzing and obtaining the detection result;

[0011] Step S1, the method for establishing a composite fiber sucker rod microscopic image detection model is as follows:

[0012] 4) Prepare model training samples and obtain target microscopic images;

[0013] 5) Sample processing to obtain amplified microscopic images;

[0014] 6) Build a detection model using the target microscopic image and the amplified image.

[0015] In step S2, the test sample preparation method is as follows: cut a cross-section of the sucker rod with a length of 1-2 mm, and grind, polish, clean and dry the slice in sequence.

[0016] After the drying, a gold coating needs to be sprayed on the sample surface.

[0017] In step S3, the sample is photographed using a scanning electron microscope.

[0018] The training samples in step 1) include microscopic images of qualified composite sucker rods and microscopic images of unqualified composite fiber sucker rods. The microscopic images are subjected to contrast enhancement processing to obtain target microscopic images.

[0019] In step 2), the method for preparing the amplified microscopic image is as follows:

[0020] a. Mark the target microscopic image with glass fibers to obtain a mask image;

[0021] b. synthesizing the mask image and the target microscopic image to obtain a synthesized image;

[0022] c. Determine the energy value in the composite image, and determine the amplified microscopic image based on the energy value.

[0023] The beneficial effects of the present invention are:

[0024] By constructing a composite fiber sucker rod microscopic image detection model through microscopic images, the composite fiber sucker rod to be detected can be quickly detected through the model, and the quality problems of the composite fiber sucker rod can be classified and determined, with high detection accuracy and efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a schematic diagram of a flow chart of a composite fiber sucker rod detection method according to an embodiment of the present invention;

[0026] Figure 2 It is a schematic diagram of the selection of shooting points for a cross-section sample of a composite fiber sucker rod detection method according to an embodiment of the present invention;

[0027] Figure 3 is a scanned image 1 of a composite fiber sucker rod detection method according to an embodiment of the present invention;

[0028] Figure 4 is a scanned image 2 of a composite fiber sucker rod detection method according to an embodiment of the present invention;

[0029] Figure 5 is an enhanced image corresponding to the scanned image 1 of a composite fiber sucker rod detection method according to an embodiment of the present invention;

[0030] Figure 6 is an enhanced image corresponding to the scanned image 2 of a composite fiber sucker rod detection method according to an embodiment of the present invention;

[0031] Figure 7 It is a training model diagram corresponding to the scanned image 1 of a composite fiber sucker rod detection method according to an embodiment of the present invention;

[0032] Figure 8 It is a training model diagram corresponding to the scanned image 2 of a composite fiber sucker rod detection method described in an embodiment of the present invention. DETAILED DESCRIPTION

[0033] The embodiments of the present invention will be further described below in conjunction with the accompanying drawings.

[0034] Example:

[0035] like Figure 1 As shown, a composite fiber sucker rod detection method comprises the following steps:

[0036] S1. Establishing a microscopic image detection model for composite fiber sucker rods;

[0037] Training sample preparation: prepare N qualified composite fiber sucker rod microscopic images and N unqualified composite fiber sucker rod microscopic images respectively.

[0038] Preprocessing of microscopic image training samples: The contrast enhancement process is performed by using the Clahe (Contrast-Limited Adaptive Histogram Equalization) algorithm to obtain a contrast-enhanced target microscopic image.

[0039] The target microscopic image is annotated with a glass fiber target using Labelme manual annotation software to obtain a mask image.

[0040] The mask image and the target microscopic image are jointly input into a preset cyclic generative adversarial network to obtain a synthetic image.

[0041] Determine the energy value of the corresponding synthesized image according to the gradient between pixels in the image;

[0042] The energy value of the corresponding image is determined by the following formula:

[0043]

[0044] In the formula, x and y represent the coordinates of pixel I in the x-axis and y-axis directions respectively. is the symbol of partial derivative, Represent the gradient of pixel I in the x-axis and y-axis directions respectively.

[0045] All the synthetic images are sorted in descending order of energy value, and the synthetic images with energy values ​​ranked first by a preset percentage are determined as amplified microscopic images;

[0046] The target microscopic image and the amplified microscopic image are input into the Mask-RCNN network to complete the model training and obtain the intelligent detection model.

[0047] S2, prepare the test sample and determine the observation surface;

[0048] Cut the cross-section of the sucker rod with a length range of 1 to 2 mm, and prevent the rod from cracking during the cutting process. Prepare the sample by grinding it on a polishing machine with coarse to fine water-abrasive sandpaper under running water, and then polish it with polishing fabric and artificial diamond paste after grinding. After polishing, clean the sample surface with running water, wipe the rod dry with clean fabric or dry it naturally, and finally spray gold coating on the surface of the material.

[0049] Glass-carbon composite fiber sucker rod is a sucker rod made of carbon fiber or glass fiber as reinforcement material, epoxy resin as matrix, and produced by pultrusion molding process. The center of its cross section is the carbon fiber layer, and the center outer layer is glass fiber, which contains multiple bundles of unidirectionally arranged carbon fiber, glass fiber and resin mixed coating, in addition, it is also mixed with pores that may be generated during the preparation process.

[0050] like Figure 2 As shown, the number of fibers and the pores of the resin matrix at different points in the sample collection area are selected for identification and statistics:

[0051] The outer layer area (taking a rod with a diameter of 22 mm as an example: selecting the area with a diameter of 20 mm to 22 mm) has 8 fixed scanning points; the inner layer area (taking a rod with a diameter of 22 mm as an example: selecting the area with a diameter of 4 mm to 20 mm) has 24 random scanning points. The above 32 points are all selected with a 200μm field of view, and the fiber number and resin matrix pores are identified and calculated with a 100μm field of view. Some fibers that cannot be directly identified are supplemented with a 50μm field of view for identification and calculation.

[0052] The white, nearly circular part in the glass fiber layer scanning image is the cross section of the glass fiber, with a diameter of approximately 15 μm to 20 μm. The flocculent part between the glass fibers is the resin material.

[0053] S3, photographing the observation surface of the test sample to obtain a target microscopic image;

[0054] The Thermo Fisher Scientific Prisma E environmental scanning electron microscope was used. Experimental conditions: low vacuum probe, pressure 70pa, voltage 20KV. After the rod sample was prepared, the sample was placed on the scanning electron microscope stage and the brightness of the observation surface was adjusted. The microscopic images of different positions of the rod were scanned respectively. The outer area (taking the rod with a diameter of 22mm as an example: the area with a diameter of 20mm to 22mm was selected) had 8 fixed scanning points; the inner area (taking the rod with a diameter of 22mm as an example: the area with a diameter of 4mm to 20mm was selected) had 24 random scanning points. Select the field of view at a lower magnification, then adjust the focus under the custom high-power microscope, and shoot and save the image after the image is clear.

[0055] The microscopic image is contrast enhanced by using the Clahé algorithm (Contrast-Limited Adaptive Histogram Equalization) to obtain a contrast-enhanced target microscopic image.

[0056] S4, inputting the target microscopic image obtained in step S3 into the composite fiber sucker rod microscopic image detection model established in step S1, and analyzing and obtaining the detection result;

[0057] The microscopic image to be inspected is first processed by clahe contrast enhancement and then input into the intelligent detection model. The model will output categories, which means that the composite fiber sucker rods are divided into qualified groups and defective groups, so as to more accurately judge the status of the composite fiber sucker rods.

[0058] The failure of composite fiber sucker rods mainly includes broken wire peeling, local splitting, continuous splitting and debonding of fiber / resin interface layer. For reinforcing materials, the number of fibers, fiber fineness, volume content and uniformity are key specifications that determine mechanical properties, and can be used as parameters to determine whether their structural design and process control meet the standards. When the sucker rod is subjected to bending load, matrix cracking and fiber breakage first appear on the lower surface of the outer glass fiber, and the failure sequence is: cracking of the outer glass fiber matrix ~ breakage of the outer glass fiber ~ debonding of the bonding layer ~ failure of the inner carbon fiber. In addition, the resin hydrolysis and plasticization caused by water molecules entering the rod body will lead to debonding of the fiber / resin interface layer. During the use of the rod, the alternating fatigue load it bears will aggravate the generation and spread of fatigue cracks inside the rod. Therefore, the number and size of pores in the resin matrix during the production of the rod are important factors affecting the aging and service life of the rod.

[0059] like Figure 3-Figure 8 As shown, after image processing, the resin matrix pore content of 32 sampling points in different areas and outer layers, the fiber density and fiber arrangement of different partitions in each sampling point are comprehensively analyzed; combined with the comprehensive analysis of the intelligent detection data of unqualified rod samples, it is determined whether the rod has defects.

[0060] The above-mentioned embodiments only express the specific implementation of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A composite fiber sucker rod detection method, characterized in that: The steps include: S1. Establishing a microscopic image detection model for composite fiber sucker rods; S2, prepare the test sample and determine the observation surface; S3, photographing the observation surface of the test sample to obtain a target microscopic image; S4, inputting the target microscopic image obtained in step S3 into the composite fiber sucker rod microscopic image detection model established in step S1, and analyzing and obtaining the detection result; Step S1, the method for establishing a composite fiber sucker rod microscopic image detection model is as follows: 1) Prepare model training samples and obtain target microscopic images; 2) Sample processing to obtain amplified microscopic images; 3) Build a detection model using the target microscopic image and the amplified image.

2. A composite fiber sucker rod detection method according to claim 1, characterized in that: In step S2, the test sample preparation method is as follows: cut a cross-section of the sucker rod with a length of 1-2 mm, and grind, polish, clean and dry the slice in sequence.

3. A composite fiber sucker rod detection method according to claim 2, characterized in that: After the drying, a gold coating needs to be sprayed on the sample surface.

4. A composite fiber sucker rod detection method according to claim 1, characterized in that: In step S3, the sample is photographed using a scanning electron microscope.

5. A composite fiber sucker rod detection method according to claim 1, characterized in that: The training samples in step 1) include microscopic images of qualified composite sucker rods and microscopic images of unqualified composite fiber sucker rods. The microscopic images are subjected to contrast enhancement processing to obtain target microscopic images.

6. A composite fiber sucker rod detection method according to claim 1, characterized in that: In step 2), the method for preparing the amplified microscopic image is as follows: a. Mark the target microscopic image with glass fibers to obtain a mask image; b. synthesizing the mask image and the target microscopic image to obtain a synthesized image; c. Determine the energy value in the composite image, and determine the amplified microscopic image based on the energy value.

Images