A municipal drainage pipeline fault intelligent identification and prediction system and method

Abstract

The application discloses a kind of municipal drainage pipeline fault intelligent identification and prediction system, including data integration module, internal monitoring module and early warning module;The data integration module is used to store pipeline historical fault data, and establishes Markov chain model to predict future pipeline failure;The internal monitoring module is controlled by the data integration module, for monitoring the internal failure condition of pipeline;The early warning module receives the data of the internal monitoring module, for analyzing the internal failure type and level of pipeline and determining whether to issue an early warning, then return the data to the data integration module.The application also discloses a kind of municipal drainage pipeline fault intelligent identification and prediction method.The beneficial effects of the present application: the pipeline failure event can be predicted, and the internal condition of the pipeline is regularly monitored according to the prediction result, the internal failure type and severity of the pipeline are intelligently identified, a large amount of manpower and material resources are saved, and the pipeline operation and maintenance cost is reduced.

Description

Technical Field

[0001] This invention relates to the field of municipal drainage technology, and in particular to an intelligent identification and prediction system and method for municipal drainage pipeline faults. Background Technology

[0002] With the rapid urbanization in my country, municipal pipe networks are becoming increasingly complex, and the load on underground drainage networks is rising, placing higher demands on the performance of drainage pipes. Simultaneously, most urban municipal drainage networks are buried underground in complex geological environments, making timely fault diagnosis difficult. Therefore, it is necessary to monitor and identify faults in drainage networks and promptly repair or replace problematic pipe sections to minimize labor costs, avoid environmental pollution, and reduce economic losses. Due to the vast and complex nature of urban underground drainage systems, relying on manual methods is time-consuming, labor-intensive, and costly, and it is difficult to obtain timely information on drainage pipe faults. Therefore, the need for automated methods to promptly diagnose and handle faults in drainage pipelines is urgent. Summary of the Invention

[0003] This invention discloses an intelligent identification and prediction system and method for municipal drainage pipeline faults, which can effectively solve the technical problems involved in the background art.

[0004] To achieve the above objectives, the technical solution of the present invention is as follows: A municipal drainage pipeline fault intelligent identification and prediction system includes a data integration module, an internal monitoring module, and an early warning module; wherein, The data integration module is used to store historical pipeline fault data, pipeline internal image information, and basic pipeline operating parameters, and to establish a Markov chain model to predict future pipeline faults. The internal monitoring module is controlled by the data integration module and is used to monitor internal pipeline faults. The early warning module receives data from the internal monitoring module, analyzes the type and level of internal pipeline faults, determines whether to issue an early warning, and then returns the data to the data integration module.

[0005] As a preferred improvement of the present invention, the data integration module includes a data storage unit and a fault prediction unit; The data storage unit is used to store historical pipeline fault data, internal pipeline image information, and basic pipeline operating parameters. The fault prediction unit is used to establish a Markov chain model to predict future pipeline faults and to control the internal monitoring module to perform internal monitoring periodically.

[0006] As a preferred improvement of the present invention, the internal monitoring module includes an image acquisition unit and a data transmission unit; The image acquisition unit is used to acquire image information inside the pipeline; The data transmission unit is used to transmit image information of the inside of the pipeline to the early warning module.

[0007] As a preferred improvement of the present invention, the early warning module includes an image processing unit and an early warning unit; The image processing unit is used to identify and analyze the image information inside the pipeline to obtain the type and level of internal faults. The early warning unit is used to determine whether to issue an alarm based on the type and level of the internal pipeline fault and in conjunction with a Markov chain model, and then transmits the pipeline fault data to the data integration module for storage.

[0008] As a preferred improvement of the present invention, the image acquisition unit acquires internal image information of the pipeline by means of a pipeline robot equipped with a 360° rotatable infrared camera, and the data transmission unit transmits the internal image information of the pipeline to the early warning unit via wireless transmission.

[0009] As a preferred improvement of the present invention, the image processing unit operates based on a convolutional neural network model, including: Pre-classify existing pipeline faults; The images acquired by the data integration module are preprocessed to extract fault information; Establish a pipeline image recognition model; An optimizer was used to optimize and train a large number of historical images until the pipeline image recognition model became stable. Save the image recognition model, identify the pipeline fault condition through the model and output the pipeline fault type and level to obtain the internal fault state of the pipeline.

[0010] As a preferred improvement of the present invention, the image recognition model is an AlexNet convolutional neural network structure model, the optimizer is an Adam optimizer, and the pipeline fault levels include 1 to 5 levels.

[0011] As a preferred improvement of the present invention, pipeline faults include structural defects and functional defects. The structural defects include rupture, leakage, severe deformation, and pipeline misalignment. The functional defects include sludge deposition, blockage accumulation, and pipeline scaling.

[0012] A method for intelligent identification and prediction of municipal drainage pipeline faults based on the aforementioned intelligent identification and prediction system includes the following steps: S1: Treat the failure of municipal drainage pipes as a Markov chain that is only related to time; S2: Define the fault status of municipal drainage pipes as follows: Status 1, Status 2, Status 3, Status 4, and Status 5. Status 1 represents a brand-new drainage pipe in a fault-free state, while Status 5 represents a drainage pipe with a serious fault that urgently needs repair. S3: Based on the Markov chain, calculate the transition probability from state i to state i+1, and then predict the probability that the pipeline will move to state 4 or state 5 as time progresses, where i takes the value 1-4. S4: Based on the data integration module, the model parameters are obtained through simulation calculation using the Monte Carlo Markov chain method, and then the transition strength and strength matrix are calculated to determine the transition probability. S5: Based on the determined transition probability, predict the probability that the pipeline will reach state 4 over time, and based on the prediction result, periodically control the internal monitoring module to collect images and transmit them to the early warning module for fault status identification. S6: After fault identification, the early warning module transmits the data to the data integration module, updates the historical fault data, and repeats steps S4-S6 to continuously update the transition probability of the model.

[0013] The beneficial effects of this invention are: it can predict pipeline failure events and monitor the internal condition of the pipeline regularly based on the prediction results, intelligently identify the type and severity of internal pipeline failures, replace the traditional manual inspection mode for internal pipeline failures, save a lot of manpower and material resources, and reduce pipeline operation and maintenance costs. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort, wherein: Figure 1 This is a system framework diagram of a municipal drainage pipeline fault intelligent identification and prediction system and method according to the present invention; Figure 2 This is a flowchart of the intelligent identification and prediction system and method for municipal drainage pipeline faults according to the present invention. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0016] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative positional relationship and movement of each component in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indication will also change accordingly.

[0017] Furthermore, in this invention, descriptions involving "first," "second," etc., are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0018] In this invention, unless otherwise explicitly specified and limited, the terms "connection," "fixed," etc., should be interpreted broadly. For example, "fixed" can mean a fixed connection, a detachable connection, or an integral part; it can mean a mechanical connection or an electrical connection; it can mean a direct connection or an indirect connection through an intermediate medium; it can mean the internal communication of two components or the interaction between two components, unless otherwise explicitly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0019] Furthermore, the technical solutions of the various embodiments of the present invention can be combined with each other, but only if they are feasible for those skilled in the art. If the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such combination of technical solutions does not exist and is not within the scope of protection claimed by the present invention.

[0020] Please see Figure 1 and Figure 2 As shown, this invention provides an intelligent identification and prediction system for municipal drainage pipeline faults, including a data integration module 1, an internal monitoring module 2, and an early warning module 3. The data integration module 1 includes a data storage unit 11 and a fault prediction unit 12. The data storage unit 11 stores historical pipeline fault data, internal pipeline image information, and basic pipeline operating parameters. The fault prediction unit 12 uses the information from the data storage unit to establish a Markov chain model, predicts future pipeline fault conditions, and controls the internal monitoring module 2 to perform periodic internal monitoring.

[0021] The internal monitoring module 2 includes an image acquisition unit 21 and a data transmission unit 22. The image acquisition unit 21 uses a pipeline robot equipped with a 360° rotatable infrared camera to acquire image information inside the pipeline. The data transmission unit 22 transmits the image information inside the pipeline to the early warning module 3 wirelessly.

[0022] The early warning module 3 includes an image processing unit 31 and an early warning unit 32. The image processing unit 31 is used to identify and analyze the internal image information of the pipeline to obtain the type and level of internal pipeline faults. The early warning unit 32 determines whether to issue an alarm based on the type and level of internal pipeline faults and in conjunction with a Markov chain model, and then transmits the pipeline fault data to the data integration module 1 for storage. Specifically, pipeline fault types include structural defects and functional defects. Structural defects include rupture, leakage, severe deformation, and pipeline misalignment, while functional defects include sludge deposition, blockage accumulation, and pipeline scaling.

[0023] Specifically, the image processing unit 31 uses convolutional neural network technology to identify and analyze the internal image information of the pipeline, including: pre-classifying existing pipeline faults; pre-processing the images collected by the data integration module 1 to extract fault information; establishing a pipeline image recognition model, which is an AlexNet convolutional neural network structure model, wherein the loss function adopts the YOLOv5 cross-entropy loss function; using the Adam optimizer, optimizing and training it with a large number of historical images until the pipeline image recognition model tends to be stable; saving the image recognition model, identifying the pipeline fault condition through the model and outputting the pipeline fault type and level to obtain the internal fault state of the pipeline, i.e., five levels from 1 to 5.

[0024] This invention also provides a method for intelligent identification and prediction of faults in municipal drainage pipelines, specifically including the following steps: S1: Treat municipal drainage pipe failures as a time-dependent Markov chain.

[0025] S2: Defines the fault states of municipal drainage pipes as State 1, State 2, State 3, State 4, and State 5. State 1 represents a brand-new, fault-free drainage pipe, while State 5 represents a severely faulty drainage pipe requiring immediate repair. Over time, municipal drainage pipes will age, and the fault state will progress from 1 to 5. When the early warning module identifies a drainage pipe fault state as 4 or 5, the pipe will be repaired. The fault state of the repaired drainage pipe will be restored accordingly, specifically determined by the convolutional neural network technology of the early warning module.

[0026] S3: Assume that during the degradation process of the drainage pipe, the transition of fault states is gradual, i.e., from state 1 to state 2, from state 2 to state 3, from state 3 to state 4, and from state 4 to state 5; there will be no state jumps like from state 1 to state 3 in a very short time. Therefore, the aging process of the pipe is modeled as a continuous-time Markov chain. Based on the Markov chain, the transition probability from state i to state i+1 is determined, and then the probability of the pipe progressing to state 4 or state 5 over time is predicted, where i takes values ​​from 1 to 4.

[0027] S4: Based on the data integration module, the model parameters are obtained through simulation calculation using the Monte Carlo Markov chain method, and then the transfer intensity is calculated. The intensity matrix Q is used to determine the transition probability P(s,t), which is expressed by equation (1): (1) Where t is the current time point, s is the time point when the fault state was last identified as i, and the intensity matrix Q is represented by equation (2): (2) in, , , , and These represent the transition strength from state i to state i+1, respectively. Equation (3) represents: (3) in, , , , , Here are the model parameters to be determined: L is the pipe length, D is the diameter, S is the slope, and A is the drainage ditch area. , , , It is a normalization constant.

[0028] S5: Based on the above calculation results, predict the probability of the pipeline reaching fault 4 as time t progresses, and according to the prediction results, periodically control the internal monitoring module to collect images and transmit them to the early warning module for fault status identification.

[0029] S6: After fault identification, the early warning module transmits the data to the data integration module, updates historical fault data, and repeats steps S4-S6 to continuously update the model's transition probability, thereby improving the accuracy of fault prediction. Additionally, it is worth noting that the transition intensity function is not uniquely determined.

[0030] Example 1 The image processing unit uses convolutional neural network image recognition technology to identify and analyze image information inside the pipe, including the following steps: S01: Pre-classify existing drainage pipe faults, which can be divided into two main categories: (1) Structural defects: cracks, leaks, severe deformation, pipe misalignment; (2) Functional defects: sludge deposition, blockage accumulation, and pipe scaling.

[0031] S02: Preprocess the images in the data integration module. A Python program is written to filter, smooth, augment, and unify the image format, completing the image preprocessing to generate sample data.

[0032] S03: Construct a pipeline fault image recognition model. An AlexNet convolutional neural network structure is used, consisting of 8 layers, including convolutional layers and fully connected layers. The first 5 layers are convolutional layers, and the last 3 are fully connected layers. The output of the last fully connected layer contains a 1000-unit softmax layer. The response normalization layer follows the 1st and 2nd convolutional layers, and the max-pooling layer follows the response normalization layer and the 5th convolutional layer. The ReLU activation function is applied after the outputs of all convolutional and fully connected layers.

[0033] S04: Training the image classification and recognition model involves adjusting the weights, biases, and loss functions of each layer until the model stabilizes. Consider that the softmax function (represented by Equation 4) outputs mutually exclusive classes. (4) Since a certain fault may belong to multiple categories, the loss function adopted is YOLOv5, and the cross-entropy loss function is used for each category of identification result, as expressed by equation (5): (5) Where N represents the total number of categories, yi is the true value (0 or 1) of the current class, yi is the probability of the current class after the activation function, xi is the predicted value of the current class, and Lclass is the classification loss.

[0034] S05: Use the AdamOptimizer to optimize parameters and train the model until it stabilizes. Specific steps include: (1) Input an image with a size of 227*227*3; (2) The first convolutional layer: 11*11*3 convolution, and using ReLU activation function and max pooling layer to obtain a 27*27*96 feature map; (3) The second convolutional layer: a 5*5*48 convolution, and using the ReLU activation function and max pooling layer to obtain a 13*13*128 feature map; (4) The 3rd-4th convolutional layer: a 3*3*192 convolution, and using the ReLU activation function to obtain a 13*13*192 pixel layer; (5) The fifth convolutional layer: a 3*3*128 convolution, and using the ReLU activation function, after pooling, a 6*6*256 pixel layer is obtained; (6) The first and second fully connected layers: 4096 neurons are fully connected and use the ReLU activation function; (7) The third fully connected layer: 1000 neurons are fully connected, and the output layer is softmax with the probability values ​​of 1000 classes; (8) Calculate the cross-entropy loss function.

[0035] S06: Output Results. There are 10 output categories: Structural defect fault states, categorized into 1, 2, 3, 4, and 5. Category 1 indicates a brand-new, fault-free drainage pipe, while category 5 indicates a severely faulty drainage pipe requiring immediate repair. Functional defect fault states, also categorized into 1, 2, 3, 4, and 5. Category 1 indicates a brand-new, fault-free drainage pipe, while category 5 indicates a severely faulty drainage pipe. When the warning module identifies a drainage pipe fault state of 4 or 5, the pipe will be repaired.

[0036] S07: Save the image classification and recognition model, and repeat steps S02 to S05 in conjunction with the internal monitoring module to identify the fault status of municipal drainage pipelines.

[0037] The beneficial effects of this invention are: it can predict pipeline failure events and monitor the internal condition of the pipeline regularly based on the prediction results, intelligently identify the type and severity of internal pipeline failures, replace the traditional manual inspection mode for internal pipeline failures, save a lot of manpower and material resources, and reduce pipeline operation and maintenance costs.

[0038] Although embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. They can be applied to various fields suitable for the present invention. Other modifications can be easily made by those skilled in the art. Therefore, without departing from the general concept defined by the claims and their equivalents, the present invention is not limited to the specific details and the illustrations shown and described herein.

Claims

1. A municipal drainage pipeline fault intelligent identification and prediction system, characterized in that: It includes a data integration module, an internal monitoring module, and an early warning module; among which, The data integration module includes a data storage unit and a fault prediction unit. The data storage unit stores historical pipeline fault data, internal pipeline image information, and basic pipeline operating parameters. The fault prediction unit establishes a Markov chain model to predict future pipeline faults and controls the internal monitoring module to perform periodic internal monitoring. In the Markov chain model established by the fault prediction unit, the pipeline fault state can only be determined from the current state. To the next adjacent state +1 step-by-step migration, where The value can be 1-4; The internal monitoring module is controlled by the data integration module. The internal monitoring module includes an image acquisition unit and a data transmission unit. The image acquisition unit is used to acquire image information inside the pipeline, and the data transmission unit is used to transmit the image information inside the pipeline to the early warning module. The early warning module receives data from the internal monitoring module. The early warning module includes an image processing unit and an early warning unit. The image processing unit is used to identify and analyze the image information inside the pipeline to obtain the type and level of the internal pipeline fault. The early warning unit is used to determine whether to issue an alarm based on the type and level of the internal pipeline fault and in conjunction with a Markov chain model. Then, the pipeline fault data is transmitted to the data integration module for storage.

2. The intelligent identification and prediction system for municipal drainage pipeline faults according to claim 1, characterized in that: The image acquisition unit uses a pipeline robot equipped with a 360° rotatable infrared camera to acquire image information inside the pipeline, and the data transmission unit transmits the image information inside the pipeline to the early warning unit wirelessly.

3. The intelligent identification and prediction system for municipal drainage pipeline faults according to claim 1, characterized in that: The image processing unit operates based on a convolutional neural network model, including: Pre-classify existing pipeline faults; The images acquired by the data integration module are preprocessed to extract fault information; Establish a pipeline image recognition model; An optimizer was used to optimize and train a large number of historical images until the pipeline image recognition model became stable. Save the image recognition model, identify the pipeline fault condition through the model and output the pipeline fault type and level to obtain the internal fault state of the pipeline.

4. The intelligent identification and prediction system for municipal drainage pipeline faults according to claim 3, characterized in that: The image recognition model is an AlexNet convolutional neural network structure model, the optimizer is the Adam optimizer, and the pipeline fault levels include 1 to 5 levels.

5. The intelligent identification and prediction system for municipal drainage pipeline faults according to claim 3, characterized in that: Pipeline failures include structural defects and functional defects. Structural defects include ruptures, leaks, severe deformation, and pipe misalignment. Functional defects include sludge deposition, blockage buildup, and pipe scaling.

6. A method for intelligent identification and prediction of municipal drainage pipeline faults based on the intelligent identification and prediction system for municipal drainage pipeline faults according to any one of claims 1 or 5, characterized in that, Includes the following steps: S1: Treat the failure of municipal drainage pipes as a Markov chain that is only related to time; S2: Define the fault status of municipal drainage pipes as follows: Status 1, Status 2, Status 3, Status 4, and Status 5. Status 1 represents a brand-new drainage pipe in a fault-free state, while Status 5 represents a drainage pipe with a serious fault that urgently needs repair. S3: Based on the Markov chain, calculate the state... to state The transition probability of +1 is used to predict the probability that the pipeline will move to state 4 or state 5 as time progresses. Values ​​range from 1 to 4; S4: Based on the data integration module, the model parameters are obtained through simulation calculation using the Monte Carlo Markov chain method, and then the transition strength and strength matrix are calculated. Finally, the transition probability is determined; the transition strength... It can be expressed by the following formula: ； in, , , , , For the model parameters to be determined, For the length of the pipe, For diameter, For slope, The area of ​​the drainage ditch , , , It is a normalization constant; S5: Based on the determined transition probability, predict the probability that the pipeline will reach state 4 over time, and based on the prediction result, periodically control the internal monitoring module to collect images and transmit them to the early warning module for fault status identification. S6: After fault identification, the early warning module transmits the data to the data integration module, updates the historical fault data, and repeats steps S4-S6 to continuously update the transition probability of the model.

Images