Pipeline monitoring method and system based on monitoring model

Abstract

The invention provides a pipeline monitoring method and system based on a monitoring model. The method comprises the steps that pipeline parameters at the current moment and an initial pipeline section image are acquired; inputting the pipeline parameters at the current moment and the initial pipeline section image into a pre-trained monitoring model; outputting a pipeline section image at the current moment through the monitoring model; wherein the monitoring model is obtained by inputting the pipeline data and the pipeline section image in a historical database into a preset monitoring model for training. The problems that existing pipeline monitoring precision is low, and the real situation in a pipeline is difficult to reflect are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of pipeline monitoring, and in particular to a pipeline monitoring method and system based on a monitoring model. Background Art

[0002] With the improvement of urbanization level and the development of infrastructure, pipeline transportation, as a mode of transportation with large transportation volume, economy, safety and small land occupation, is widely used in the transmission of liquids and gases, connecting various cities. Pipeline leakage will cause immeasurable losses to the surrounding ecological environment and economic development. Therefore, real-time monitoring of pipelines and real-time control of the entire life cycle of pipelines are of great significance to the healthy, safe and orderly development of cities.

[0003] Ultrasonic flow monitoring technology uses ultrasonic technology to monitor pipeline leakage. When the material or components of the pipeline deform and crack under operating conditions, strain energy will be released in the form of elastic waves in the pipeline. The receiving device performs non-destructive testing on the material by receiving the ultrasonic waves emitted by the transmitter. The ultrasonic sensor is outside the pipeline and can be monitored without disassembling the pipeline. However, the disadvantage of monitoring outside the pipeline is that it is greatly affected by the external environment of the pipeline and the monitoring accuracy is low. The principle of fiber optic sensing monitoring technology is that under the irradiation of a laser light source of a certain wavelength, a series of reflected light waves will be formed in the doped area of ​​the optical fiber. Since the wavelength of the reflected light wave is proportional to the refractive index of the doped area, the environment of the optical fiber can be monitored by measuring the wavelength change of the reflected light wave. This method outputs in-pipe monitoring, so it has very high accuracy, but the fiber optic sensor is greatly affected by the environment, and it cannot directly measure internal parameters of the pipeline such as flow.

[0004] The current pipeline monitoring methods use single sensors such as optical fiber sensors and pressure sensors to sense data for analysis, which cannot well describe the complex conditions of the actual pipeline environment. Long-term operation will lead to increasing distortion of pipeline monitoring. We need an algorithm that can simulate pipeline conditions in real time, so as to more accurately understand the real-time information of the pipeline, detect emergencies faster, and reduce property losses caused by untimely detection of pipeline leakage problems. Summary of the invention

[0005] The present invention provides a pipeline monitoring method and system based on a monitoring model, which are used to solve the problem that the existing pipeline monitoring has low accuracy and is difficult to reflect the real situation inside the pipeline.

[0006] The present invention provides a pipeline monitoring method based on a diffusion model, comprising: Obtain the pipeline parameters and initial pipeline cross-section image at the current moment; Inputting the pipeline parameters at the current moment and the initial pipeline cross-section image into the pre-trained monitoring model; Outputting the pipeline cross-section image at the current moment through the monitoring model; The monitoring model is obtained by inputting pipeline data and pipeline cross-section images in a historical database into a preset monitoring model for training.

[0007] According to a pipeline monitoring method based on a monitoring model provided by the present invention,

[0008] The monitoring model is obtained by inputting pipeline data and pipeline cross-section images in the historical database into a preset monitoring model for training, specifically including: Encode the pipeline parameters at the current moment in the historical database to obtain a coding vector, and encode the pipeline cross-section image at the previous moment to obtain an image vector; Concatenating the encoding vector and the image vector to form an input vector of a preset diffusion model; Inputting the input vector into the diffusion model to generate prediction noise of the pipeline state at the current moment, performing denoising based on the prediction noise through a decoder to generate a pipeline cross-sectional image at the current moment; The generated pipeline cross-section image at the current moment is compared with the pipeline cross-section image in the historical database, the diffusion model is iteratively adjusted, and finally the monitoring model is generated.

[0009] According to a pipeline monitoring method based on a monitoring model provided by the present invention, before inputting the input vector into the diffusion model, the method comprises: A set of initial solutions of the diffusion model are generated by randomly selecting pipeline parameters from the historical database through a preset differential evolution algorithm; The initial solution is cross-operated with pipeline parameters randomly selected from a historical database, and it is continuously iterated until preset conditions are met to generate initial parameters of the diffusion model.

[0010] According to a pipeline monitoring method based on a monitoring model provided by the present invention, the input vector is input into the diffusion model to generate a prediction noise of the pipeline state at the current moment, and denoising is performed through a decoder based on the prediction noise to generate a pipeline cross-sectional image at the current moment, specifically comprising: Inputting the input vector into the diffusion model, accelerating the convergence of the diffusion model through initial parameters, and generating prediction noise of the pipeline state at the current moment; The prediction noise is removed through the decoder of the diffusion model to generate the pipeline interface image at the current moment.

[0011] According to a pipeline monitoring method based on a monitoring model provided by the present invention, the input vector is input into the diffusion model, the diffusion model is accelerated to converge by initial parameters, and the prediction noise of the pipeline state at the current moment is generated, which specifically includes: inputting the input vector into the diffusion model, the diffusion model extracting an image vector from the input vector; Based on the initial parameters, multiple convolutional layers and linear layers of the diffusion model are used to accelerate convergence, extract the noise of the pipeline state at the current moment from the image vector, and output the predicted noise; The diffusion model gradually adds noise to the original pipe cross-section image until it becomes a pure noise image. The diffusion model restores the original image from the pure noise image through multiple convolutional layers and linear layers to complete the learning of the diffusion model's noise prediction function.

[0012] According to a pipeline monitoring method based on a monitoring model provided by the present invention, the step of obtaining pipeline parameters at the current moment includes: The preset optical fiber sensors, ultrasonic sensors, membrane oxygen sensors and nanocrystalline pH sensors are used to obtain the current pipeline parameters, including wall thickness, pipe diameter, elastic modulus, temperature, pressure, ambient pH value, dissolved oxygen content, flow rate and corrosion rate.

[0013] The present invention also provides a pipeline monitoring system based on a monitoring model, the system comprising: A data acquisition module is used to obtain the pipeline parameters and the initial pipeline cross-section image at the current moment; A data input module, used to input the pipeline parameters at the current moment and the initial pipeline cross-section image into the pre-trained monitoring model; An image output module, used to output the pipeline cross-section image at the current moment through the monitoring model; The monitoring model is obtained by inputting pipeline data and pipeline cross-section images in a historical database into a preset monitoring model for training.

[0014] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the pipeline monitoring method based on the monitoring model described above is implemented.

[0015] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the pipeline monitoring method based on the monitoring model as described above is implemented.

[0016] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the pipeline monitoring method based on any one of the monitoring models described above is implemented.

[0017] The present invention provides a pipeline monitoring method and system based on a monitoring model. A detection model for the pipeline is obtained by training based on a diffusion model. Only the pipeline parameters at the current moment obtained by the sensor and the initial pipeline cross-sectional image need to be input to output the pipeline cross-sectional image at the current moment, thereby realizing real-time monitoring of the pipeline, improving monitoring accuracy, and reflecting the real situation inside the pipeline. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 It is a flow chart of a pipeline monitoring method based on a monitoring model provided by the present invention.

[0020] Figure 2 This is an architecture diagram of a pipeline monitoring system based on a monitoring model provided by the present invention.

[0021] Figure 3 It is a schematic diagram of module connection of a pipeline monitoring system based on a monitoring model provided by the present invention.

[0022] Figure 4 It is a structural schematic diagram of the electronic device provided by the present invention.

[0023] Reference numerals: 110: data acquisition module; 120: data input module; 130: image output module; 410: processor; 420: communication interface; 430: memory; 440: communication bus. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are 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.

[0025] Combine the following Figure 1A pipeline monitoring method based on a monitoring model of the present invention is described, comprising: step 100, obtaining pipeline parameters at the current moment and an initial pipeline cross-section image.

[0026] In the present invention, optical fiber sensors, ultrasonic sensors, membrane oxygen sensors and nanocrystalline pH sensors are used to monitor the real-time data of the pipeline, and the cross-sectional image of the pipeline and the parameters of the surrounding environment are obtained: wall thickness (mm), pipe diameter (mm), elastic modulus (MPa), temperature (℃), pressure (MPa), environmental pH value, dissolved oxygen content (mg / L), flow rate (m³ / s), corrosion rate (μA / cm -2 ). Using multiple sensors to collect multiple pipeline parameters can help to more accurately reflect the pipeline cross-section image.

[0027] Step 200: Input the pipeline parameters at the current moment and the initial pipeline cross-section image into the pre-trained monitoring model.

[0028] In the present invention, after obtaining the pipeline data at the current moment, the initial pipeline cross-sectional image is obtained through the image acquisition device. After the medium flows in the pipeline, the image acquisition device no longer has the ability to acquire the pipeline cross-sectional image. The pipeline data at the current moment and the initial pipeline cross-sectional image are respectively encoded and spliced ​​together through the CLIP sensor encoder to form a comprehensive input vector; the spliced ​​comprehensive input vector is received through the Embedding Layer and embedded into a high-dimensional space to form an initial embedding vector, and the complex features and dependencies in the input vector are captured through Transformer Blocks to generate a rich feature representation; the intermediate output is normalized through the Norm Layer to keep the data distribution stable; the predicted noise is received through the DAE decoder, and the pipeline cross-sectional image at the current moment is generated and output through the denoising process.

[0029] Step 300: Outputting the pipeline cross-section image at the current moment through the monitoring model.

[0030] The monitoring model is obtained by inputting pipeline data and pipeline cross-section images in a historical database into a diffusion model for training.

[0031] In the present invention, the training process of the monitoring model is: Encode the pipeline parameters at the current moment in the historical database to obtain a coding vector, and encode the pipeline cross-section image at the previous moment to obtain an image vector; Concatenating the encoding vector and the image vector to form an input vector of a preset diffusion model; Inputting the input vector into the diffusion model to generate prediction noise of the pipeline state at the current moment, performing denoising based on the prediction noise through a decoder to generate a pipeline cross-sectional image at the current moment; The generated pipeline cross-section image at the current moment is compared with the pipeline cross-section image in the historical database, the diffusion model is iteratively adjusted, and finally the monitoring model is generated.

[0032] Before inputting the input vector into the diffusion model, the method includes: A set of initial solutions of the diffusion model are generated by randomly selecting pipeline parameters from the historical database through a preset differential evolution algorithm; The initial solution is cross-operated with pipeline parameters randomly selected from a historical database, and it is continuously iterated until preset conditions are met to generate initial parameters of the diffusion model.

[0033] In the present invention, the initial parameters refer to the initial values ​​of the model parameters set before training the model, which play an important role in accelerating the convergence speed and avoiding local optimum in the model training process.

[0034] Initial parameters usually include weights (the weights of the connections between neurons in each layer of the neural network) and biases (the bias term of each neuron in the neural network), which are information in the network. The weights and bias terms are used to accelerate the convergence speed of the diffusion model and discard the local optimal value.

[0035] Differential Evolution (DE) is also an optimization algorithm based on modern intelligence theory. It guides the direction of optimization search through the group intelligence generated by mutual cooperation and competition between individuals in the group. The basic idea of ​​the algorithm is: starting from a randomly generated initial population, a new individual is generated by summing the vector difference between any two individuals in the population with the third individual, and then the new individual is compared with the corresponding individual in the contemporary population. If the fitness of the new individual is better than that of the current individual, the new individual is used to replace the old individual in the generation, otherwise the old individual is still preserved. Through continuous evolution, excellent individuals are retained, inferior individuals are eliminated, and the search is guided to approach the optimal solution.

[0036] The differential optimization algorithm uses floating-point vectors for encoding to generate population individuals. In the DE algorithm optimization process, first, two individuals are selected from the parent individuals to generate differential vectors by vector difference; second, another individual is selected and summed with the differential vector to generate experimental individuals; then, the parent individuals and the corresponding experimental individuals are cross-operated to generate new offspring individuals; finally, a selection operation is performed between the parent individuals and the offspring individuals, and the individuals that meet the requirements are saved in the next generation population.

[0037] The main control parameters of the differential optimization algorithm include: population size (NP), scaling factor (F) and crossover probability (CR).

[0038] NP mainly reflects the amount of population information in the algorithm. The larger the NP value, the richer the population information contained. However, the consequence is that the amount of calculation increases, which is not conducive to solving the problem.

[0039] On the contrary, limiting the diversity of the population will not be conducive to the algorithm finding the global optimal solution and may even cause the search to stagnate.

[0040] CR mainly reflects the amount of information exchanged between offspring, parents, and intermediate variants during the crossover process. The larger the CR value, the greater the degree of information exchange. On the contrary, if the CR value is too small, the diversity of the population will decrease rapidly, which is not conducive to global optimization.

[0041] Compared with CR, F has a greater impact on algorithm performance. F mainly affects the algorithm's global optimization ability. The smaller F is, the better the algorithm's local search ability is. The larger F is, the more the algorithm can jump out of the local minimum point, but the convergence speed will be slower. In addition, F also affects the diversity of the population. The differential evolution algorithm is used to improve the training speed of the diffusion model.

[0042] The input vector is input into the diffusion model to generate prediction noise of the pipeline state at the current moment, and denoising is performed through a decoder based on the prediction noise to generate a pipeline cross-sectional image at the current moment, specifically including: Inputting the input vector into the diffusion model, accelerating the convergence of the diffusion model through initial parameters, and generating prediction noise of the pipeline state at the current moment; The prediction noise is removed through the decoder of the diffusion model to generate the pipeline interface image at the current moment.

[0043] The input vector is input into the diffusion model, and the diffusion model is accelerated to converge by using initial parameters to generate prediction noise of the pipeline state at the current moment, specifically including: inputting the input vector into the diffusion model, the diffusion model extracting an image vector from the input vector; Based on the initial parameters, multiple convolutional layers and linear layers of the diffusion model are used to accelerate convergence, extract the noise of the pipeline state at the current moment from the image vector, and output the predicted noise; The diffusion model gradually adds noise to the original pipe cross-section image until it becomes a pure noise image. The diffusion model restores the original image from the pure noise image through multiple convolutional layers and linear layers to complete the learning of the diffusion model's noise prediction function.

[0044] In the present invention, in the process of generating images by diffusion model, predicting noise is a critical operation, which affects the quality of the network generated images. It plays an important role in generating high-quality images, modeling complex distributions, and enhancing the robustness of the model.

[0045] First, noise is added to the original pipeline cross-section image. This process is usually completed through a process of gradually increasing noise, that is, noise is gradually added from the original pipeline cross-section image until it becomes a pure noise image.

[0046] Next, the diffusion model is trained to enable it to recover the original pipeline cross-section image from the noisy image, realizing the process of diffusion model learning denoising, that is, how to extract the information of the original pipeline cross-section image from the image containing noise.

[0047] Finally, in the generation stage, starting from random noise, the noise is gradually removed through the trained diffusion model with denoising ability, and finally a clear image is generated.

[0048] Predicting noise means that in the final generation stage, the model predicts the noise that should be removed in the current image at each step. During the operation of the diffusion model, the network generates a noise prediction based on the input parameters. The noise prediction is used to guide the model on how to process the current image in the subsequent denoising process, and finally generate a high-quality pipeline cross-section image through the DAE decoder.

[0049] Specifically, the complete process of diffusion model training is: The input includes the pipeline cross-section image at the previous moment and the pipeline parameters at the current moment; The CLIP sensor encoder encodes the pipeline cross-section image at the previous moment to obtain an image vector; encodes the pipeline parameters at the current moment to obtain a parameter vector; these two vectors are concatenated together to form the input vector of the diffusion model; the model generates prediction noise for the pipeline state at the current moment through multiple Transformer blocks, convolutional layers, and linear layers, and the prediction noise is denoised by the DAE decoder to generate the pipeline image at the current moment; Use image quality assessment metrics (such as FID, SSIM) to evaluate the quality of generated images; The training process involves multiple iterations, each of which updates the model parameters to gradually improve the quality of the generated images. Eventually, the monitoring model learns the mapping relationship between the pipeline cross-section images and the environmental parameters.

[0050] According to a pipeline monitoring method based on a monitoring model provided by the present invention, a detection model for the pipeline is obtained by training based on a diffusion model. Only the pipeline parameters at the current moment and the pipeline cross-sectional image at the previous moment obtained by the sensor need to be input, and the pipeline cross-sectional image at the current time can be output, thereby realizing real-time monitoring of the pipeline, improving monitoring accuracy, and reflecting the real situation inside the pipeline.

[0051] refer to Figure 2 and Figure 3 The present invention also discloses a pipeline monitoring system based on a monitoring model, the system comprising: The data acquisition module 110 is used to acquire the pipeline parameters and the initial pipeline cross-section image at the current moment; A data input module 120, used to input the pipeline parameters at the current moment and the initial pipeline cross-section image into the pre-trained monitoring model; An image output module 130, configured to output a pipeline cross-section image at the current moment through the monitoring model; The monitoring model is obtained by inputting pipeline data and pipeline cross-section images in a historical database into a preset monitoring model for training.

[0052] The monitoring model is obtained by inputting pipeline data and pipeline cross-section images in the historical database into a preset monitoring model for training, specifically including: Encode the pipeline parameters at the current moment in the historical database to obtain a coding vector, and encode the pipeline cross-section image at the previous moment to obtain an image vector; Concatenating the encoding vector and the image vector to form an input vector of a preset diffusion model; Inputting the input vector into the diffusion model to generate prediction noise of the pipeline state at the current moment, performing denoising based on the prediction noise through a decoder to generate a pipeline cross-sectional image at the current moment; The generated pipeline cross-section image at the current moment is compared with the pipeline cross-section image in the historical database, the diffusion model is iteratively adjusted, and finally the monitoring model is generated.

[0053] Before inputting the input vector into the diffusion model, the method includes: A set of initial solutions of the diffusion model are generated by randomly selecting pipeline parameters from the historical database through a preset differential evolution algorithm; The initial solution is cross-operated with pipeline parameters randomly selected from a historical database, and it is continuously iterated until preset conditions are met to generate initial parameters of the diffusion model.

[0054] Inputting the input vector into the diffusion model to generate prediction noise of the pipeline state at the current moment, performing denoising based on the prediction noise through a decoder to generate a pipeline cross-sectional image at the current moment, specifically includes: Inputting the input vector into the diffusion model, accelerating the convergence of the diffusion model through initial parameters, and generating prediction noise of the pipeline state at the current moment; The prediction noise is removed through the decoder of the diffusion model to generate the pipeline interface image at the current moment.

[0055] The input vector is input into the diffusion model, and the diffusion model is accelerated to converge by using initial parameters to generate prediction noise of the pipeline state at the current moment, specifically including: inputting the input vector into the diffusion model, the diffusion model extracting an image vector from the input vector; Based on the initial parameters, multiple convolutional layers and linear layers of the diffusion model are used to accelerate convergence, extract the noise of the pipeline state at the current moment from the image vector, and output the predicted noise; The diffusion model gradually adds noise to the original pipe cross-section image until it becomes a pure noise image. The diffusion model restores the original image from the pure noise image through multiple convolutional layers and linear layers to complete the learning of the diffusion model's noise prediction function.

[0056] The step of obtaining the pipeline parameters at the current moment includes: The preset optical fiber sensors, ultrasonic sensors, membrane oxygen sensors and nanocrystalline pH sensors are used to obtain the current pipeline parameters, including wall thickness, pipe diameter, elastic modulus, temperature, pressure, ambient pH value, dissolved oxygen content, flow rate and corrosion rate.

[0057] A pipeline monitoring system based on a monitoring model provided by the present invention obtains a detection model for the pipeline by training based on a diffusion model. It only needs to input the pipeline parameters at the current moment and the pipeline cross-sectional image at the previous moment obtained by the sensor to output the pipeline cross-sectional image at the current time, thereby realizing real-time monitoring of the pipeline, improving monitoring accuracy, and reflecting the real situation inside the pipeline.

[0058] Figure 4 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 4As shown, the electronic device may include: a processor 410, a communication interface 420, a memory 430 and a communication bus 440, wherein the processor 410, the communication interface 420 and the memory 430 communicate with each other through the communication bus 440. The processor 410 may call the logic instructions in the memory 430 to execute a pipeline monitoring method based on a monitoring model, the method comprising: obtaining pipeline parameters and an initial pipeline cross-sectional image at the current moment; inputting the pipeline parameters and the initial pipeline cross-sectional image at the current moment into a pre-trained monitoring model; outputting the pipeline cross-sectional image at the current moment through the monitoring model; wherein the monitoring model is obtained by training based on the pipeline data and the pipeline cross-sectional image in the historical database input into a preset monitoring model.

[0059] In addition, the logic instructions in the above-mentioned memory 430 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0060] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute a pipeline monitoring method based on a monitoring model provided by the above methods, the method including: obtaining pipeline parameters and an initial pipeline cross-sectional image at the current moment; inputting the pipeline parameters and the initial pipeline cross-sectional image at the current moment into a pre-trained monitoring model; and outputting the pipeline cross-sectional image at the current moment through the monitoring model; wherein the monitoring model is obtained by training based on pipeline data and pipeline cross-sectional images in a historical database input into a preset monitoring model.

[0061] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute a pipeline monitoring method based on a monitoring model provided by the above methods, the method comprising: obtaining pipeline parameters and an initial pipeline cross-sectional image at a current moment; inputting the pipeline parameters and the initial pipeline cross-sectional image at a current moment into a pre-trained monitoring model; and outputting the pipeline cross-sectional image at a current moment through the monitoring model; wherein the monitoring model is trained based on pipeline data and pipeline cross-sectional images in a historical database input into a preset monitoring model.

[0062] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0063] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0064] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A pipeline monitoring method based on a monitoring model, characterized in that: include: Obtain the pipeline parameters and initial pipeline cross-section image at the current moment; Inputting the pipeline parameters at the current moment and the initial pipeline cross-section image into the pre-trained monitoring model; Outputting the pipeline cross-section image at the current moment through the monitoring model; The monitoring model is obtained by inputting pipeline data and pipeline cross-section images in a historical database into a preset monitoring model for training.

2. The pipeline monitoring method based on the monitoring model according to claim 1 is characterized in that: The monitoring model is obtained by inputting pipeline data and pipeline cross-section images in the historical database into a preset monitoring model for training, specifically including: Encode the pipeline parameters at the current moment in the historical database to obtain a coding vector, and encode the pipeline cross-section image at the previous moment to obtain an image vector; Concatenating the encoding vector and the image vector to form an input vector of a preset diffusion model; Inputting the input vector into the diffusion model to generate prediction noise of the pipeline state at the current moment, performing denoising based on the prediction noise through a decoder to generate a pipeline cross-sectional image at the current moment; The generated pipeline cross-section image at the current moment is compared with the pipeline cross-section image in the historical database, the diffusion model is iteratively adjusted, and finally the monitoring model is generated.

3. The pipeline monitoring method based on the monitoring model according to claim 2 is characterized in that: Before inputting the input vector into the diffusion model, the method includes: A set of initial solutions of the diffusion model are generated by randomly selecting pipeline parameters from the historical database through a preset differential evolution algorithm; The initial solution is cross-operated with pipeline parameters randomly selected from a historical database, and it is continuously iterated until preset conditions are met to generate initial parameters of the diffusion model.

4. The pipeline monitoring method based on the monitoring model according to claim 2 is characterized in that: The step of inputting the input vector into the diffusion model, generating prediction noise of the pipeline state at the current moment, and performing denoising based on the prediction noise through a decoder to generate a pipeline cross-sectional image at the current moment specifically includes: Inputting the input vector into the diffusion model, accelerating the convergence of the diffusion model through initial parameters, and generating prediction noise of the pipeline state at the current moment; The prediction noise is removed through the decoder of the diffusion model to generate the pipeline interface image at the current moment.

5. The pipeline monitoring method based on the monitoring model according to claim 4 is characterized in that: The step of inputting the input vector into the diffusion model, accelerating the convergence of the diffusion model through initial parameters, and generating prediction noise of the pipeline state at the current moment specifically includes: inputting the input vector into the diffusion model, the diffusion model extracting an image vector from the input vector; Based on the initial parameters, multiple convolutional layers and linear layers of the diffusion model are used to accelerate convergence, extract the noise of the pipeline state at the current moment from the image vector, and output the predicted noise; The diffusion model gradually adds noise to the original pipe cross-section image until it becomes a pure noise image. The diffusion model restores the original image from the pure noise image through multiple convolutional layers and linear layers to complete the learning of the diffusion model's noise prediction function.

6. The pipeline monitoring method based on the monitoring model according to claim 1 is characterized in that: The step of obtaining the pipeline parameters at the current moment includes: The preset optical fiber sensors, ultrasonic sensors, membrane oxygen sensors and nanocrystalline pH sensors are used to obtain the current pipeline parameters, including wall thickness, pipe diameter, elastic modulus, temperature, pressure, ambient pH value, dissolved oxygen content, flow rate and corrosion rate.

7. A pipeline monitoring system based on a monitoring model, characterized in that: The system comprises: A data acquisition module is used to obtain the pipeline parameters and the initial pipeline cross-section image at the current moment; A data input module, used to input the pipeline parameters at the current moment and the initial pipeline cross-section image into the pre-trained monitoring model; An image output module, used to output a pipeline cross-section image at the current moment through the monitoring model; The monitoring model is obtained by inputting pipeline data and pipeline cross-section images in a historical database into a preset monitoring model for training.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the pipeline monitoring method based on the monitoring model as claimed in any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the pipeline monitoring method based on the monitoring model as claimed in any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the pipeline monitoring method based on the monitoring model as claimed in any one of claims 1 to 6 is implemented.

Images