Pipeline data management method based on GIS platform

Abstract

The invention relates to a pipeline data management method based on a GIS platform, and belongs to the technical field of modeling assistance. Pipeline data and station yard image data are collected, the pipeline data are screened, recognized and corrected to obtain standard pipeline data, the station yard image data are processed and modeled, a three-dimensional scene model is published through a GIS platform, and the standard pipeline data are hooked to the three-dimensional scene model. A modernized means is provided for management of pipeline data, the data can be managed more conveniently and quickly, and the safety emergency management efficiency of the pipeline is effectively improved. Meanwhile, in consideration of the possibility that the data is maliciously tampered, the block chain is used for storing the data.

Description

Technical Field

[0001] The invention relates to the field of modeling assistance technology, and in particular to a pipeline data management method based on a GIS platform. Background Art

[0002] Since pipeline facilities are distributed in a wide area and have complex underground structures with strong geographical characteristics, in daily work, it is necessary not only to retrieve conventional attributes such as text pipe diameter and construction time, but also to obtain geographic spatial information data about pipelines. If conventional archive management mode is adopted, it will seriously affect the timeliness and accuracy of emergency accident handling and safety maintenance positioning, and seriously affect the efficiency of emergency work; and conventional data storage methods are easily tampered with, affecting data reliability. GIS provides a modern means for studying and analyzing various pipeline resource information. Using this technology, we can analyze various forms of information pipeline resource environment data in different temporal states more conveniently, quickly and accurately, and generate new data we need. The functions of acquiring, storing, processing and analyzing spatial data by geographic information technology can effectively improve the efficiency of pipeline safety emergency management; blockchain storage method can prevent data from being maliciously tampered with; therefore, it is particularly important to invent a pipeline data management method based on GIS platform.

[0003] After searching, the Chinese invention patent "A method, device and system for underground pipeline data management" (application publication number CN107818535A) discloses a method, device and system for underground pipeline data management. Although the invention verifies the accuracy of pipeline data information through a data rollback mechanism, significantly improving the integrity and accuracy of pipeline data information uploaded by pipeline ownership units, the accuracy of data anomaly verification is low, and there is a lack of pipeline spatial geographic scenes. It cannot provide virtual scenes of underground pipelines and high-consequence areas for safety hazard management in a timely manner, which is inconvenient for staff to locate and investigate hazards. In addition, the data processing efficiency of existing pipeline data management methods is low, and the data is easily tampered with maliciously, affecting the accuracy of data analysis and prediction during production operation. To this end, we propose a pipeline data management method based on a GIS platform. Summary of the invention

[0004] The purpose of the present invention is to provide a pipeline data management method based on a GIS platform to solve the problem of low pipeline data processing efficiency in the prior art.

[0005] In order to solve the above technical problems, the present invention provides a pipeline data management method based on a GIS platform, the method comprising the following steps:

[0006] 1) Obtain pipeline data and station image data;

[0007] 2) Inputting the pipeline data into the trained machine learning model to classify the pipeline data, identify normal pipeline data and abnormal pipeline data, and correct the abnormal pipeline data, thereby obtaining standard pipeline data including the corrected pipeline data and normal pipeline data;

[0008] 3) Process the image data to obtain high-definition images, use the high-definition images to complete three-dimensional scene modeling, publish the three-dimensional scene model through the three-dimensional engine of the GIS platform, and attach the standard pipeline data to the three-dimensional scene model.

[0009] The beneficial effects of the above technical solution are: identifying and correcting pipeline data, effectively improving the accuracy of pipeline data, establishing a three-dimensional scene model, linking standard pipeline data with the three-dimensional scene model, giving pipeline data spatial geographic characteristics, and making pipeline management more convenient.

[0010] Furthermore, the pipeline data in step 1) includes: pipeline code, weld code, and weld coordinates.

[0011] The beneficial effect of the above technical solution is: selecting a variety of pipeline data facilitates the subsequent management of the pipeline data.

[0012] Furthermore, the abnormal types of the abnormal pipeline data classified in step 2) include pipeline coding format abnormality, weld coding format abnormality and weld coordinate abnormality.

[0013] The beneficial effect of the above technical solution is that the abnormal types of pipeline data are classified to facilitate the subsequent correction of the data.

[0014] Furthermore, the machine learning model in step 2) is a convolutional neural network.

[0015] The beneficial effect of the above technical solution is that the selection of convolutional neural network to classify pipeline data is more accurate.

[0016] Furthermore, in step 1), the specific method for acquiring the station image data is to determine the flight altitude and camera angle of the station drone image shooting, plan the drone flight route and control point positions, and obtain the station image pictures by drone shooting.

[0017] The beneficial effect of the above technical solution is that it is more convenient and faster to use drones to collect station image data.

[0018] Furthermore, the image data processing in step 2) is to decompose the collected image data frame by frame, and then perform block processing according to the display ratio of each frame of image data, and analyze and extract the high-frequency components in the data of the blocked image through Fourier transform, and then smooth the high-frequency noise at the edge of the image in the data through Gaussian filtering to obtain a high-definition image.

[0019] The beneficial effects of the above technical solution are: further processing of image data to improve data accuracy to facilitate the next step of three-dimensional scene modeling.

[0020] Furthermore, in the process of attaching the standard pipeline data to the three-dimensional scene model in step 3), it is also necessary to combine the pipeline drawing materials and pipeline survey data, and then complete the matching with the three-dimensional scene model through the pipeline unique code to complete the three-dimensional scene model loading and attribute attachment.

[0021] The beneficial effect of the above technical solution is: combining the standard pipeline data with the three-dimensional scene model so that the standard pipeline data has spatial geographic characteristics for easy management.

[0022] Furthermore, after the standard pipeline data is connected, the data needs to be stored using blockchain, and the storage steps include:

[0023] (1) Preprocess the 3D scene model data and the corresponding pipeline data into a unified format, and process the data into blocks. When joining the blockchain network, each node in the blockchain network generates a local public-private key pair as its own identification code in the network. When the node waits for the local role to become a candidate node, it broadcasts the leadership application information to other nodes in the network;

[0024] (2) After the application is approved, the candidate node becomes the leader node and the other nodes become follower nodes. The leader node then broadcasts the block record information. After receiving the information, the follower node broadcasts the received information to other follower nodes and records the number of repetitions. It uses the information with the most repetitions to generate a block header and sends a verification application to the leader node.

[0025] (3) After verification, the leader node sends an add command and enters a dormant period. During the dormant period, it cannot apply to become a leader node again until the end of the dormant period. After the follower node receives the confirmation message, it adds the newly generated groups of blocks to the blockchain and returns the candidate identity.

[0026] The beneficial effect of the above technical solution is: using blockchain to store data to prevent the data from being maliciously tampered with. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a flow chart of a pipeline data management method based on a GIS platform of the present invention;

[0028] Figure 2 It is a station image picture taken by a drone in an embodiment of the present invention;

[0029] Figure 3 is a high-definition image in an embodiment of the present invention;

[0030] Figure 4 is a three-dimensional scene model of a station in an embodiment of the present invention;

[0031] Figure 5 It is a three-dimensional station model published by using the three-dimensional model publishing technology of the GIS platform in the embodiment of the present invention;

[0032] Figure 6 It is a three-dimensional model for completing data mounting in the embodiment of the present invention. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is further described in detail below with reference to the accompanying drawings and embodiments.

[0034] Implementation example of pipeline data management method based on GIS platform:

[0035] The present invention proposes an effective design evaluation method to solve the problem of low pipeline data processing efficiency. The flowchart of the method is as follows: Figure 1 As shown, the specific implementation steps of this method are as follows:

[0036] 1. Collect pipeline data through the data collection APP.

[0037] The pipeline data includes: pipeline code, weld code, weld coordinates; use drones to shoot station images, determine the shooting altitude and 45-degree angle of the camera, plan the drone flight route and the positions of five control points, and use drones to complete the shooting to obtain station image pictures.

[0038] 2. Input the pipeline data into the trained machine learning model to classify the pipeline data, identify normal pipeline data and abnormal pipeline data, and correct the abnormal pipeline data to obtain standard pipeline data including corrected pipeline data and normal pipeline data; the abnormal types of abnormal pipeline data include pipeline coding format abnormality, weld coding format abnormality and weld coordinate abnormality, and the classification of pipeline data is based on GBT 51296-2018 Petrochemical Engineering Digital Delivery Standard.

[0039] In this embodiment, a convolutional neural network is selected as the machine learning model. As other implementation methods, recursive neural networks and genetic algorithms can also be selected.

[0040] The specific training method of the machine learning model is to divide the data into a training set and a test set, and use a clustering algorithm to import the training set samples one by one into the convolutional neural network to obtain the initial model for abnormal pipeline data detection. Based on the training set and the test set, the weights of the initial model are continuously updated to obtain the optimal model with the minimum root mean square error. In this way, abnormal pipeline data is obtained, and the abnormal pipeline data is corrected to obtain the corrected pipeline data. The specific calculation formula of the local error δ of the neuron in the convolutional neural network is:

[0041]

[0042] Among them, among them, represents the local error of the t-th position of the q-th neuron in the l-th hidden layer, represents the output of the t-th position of the q-th neuron in the l-th hidden layer, e represents the e-th neuron, represents the sum of the output of the l+1th hidden layer and the offset of the l-1th hidden layer within the range of the eth neuron, Represents the offset between the input data of the e-th neuron and the q-th neuron;.

[0043] When the local error δ exceeds the set threshold, the weight w of the adjustment model is trained and updated according to the learning rule equation, and the preset learning rate η (l) And the step length lists all possible data results;

[0044] The specific calculation formula of the learning rule equation is:

[0045]

[0046] In the formula, represents the weight of the kth neuron in the lth hidden layer, η (l) is the learning rate parameter corresponding to each network layer. Generally, the learning rate is set to an equal value. represents the local error of the kth neuron in the lth hidden layer, Represents the output result of the p-th neuron at the t-th position in the l-1-th output layer.

[0047] 3. Process the image data. Decompose the collected image data frame by frame, and then process it in blocks according to the display ratio of each frame of the image data. Use Fourier transform to analyze and extract the high-frequency components of the divided images, and then use Gaussian filtering to smooth the high-frequency noise at the edge of the image in the data to obtain high-definition images, and use high-definition images to complete 3D scene modeling.

[0048] Specific steps for 3D scene modeling.

[0049] Calculate the average grayscale value of each high-definition image data, then compare the grayscale value of each group of pixels in each image data after segmentation with the calculated average value, and use all pixels with grayscale values ​​greater than the average value as the segmentation target, and all pixels with grayscale values ​​less than the average value as the background of the segmented image. Then, use high-definition image automatic modeling technology to build a three-dimensional scene model of the station based on the segmentation results. Publish the three-dimensional scene model through the three-dimensional engine of the GIS platform, and attach the standard pipeline data to the three-dimensional scene model.

[0050] Standard pipeline data hookup.

[0051] The specific operation of standard pipeline data attachment is that in the process of attaching standard pipeline data to the three-dimensional scene model, it is also necessary to combine pipeline drawing materials and pipeline survey data, and then complete the matching with the three-dimensional scene model through the unique pipeline code to complete the three-dimensional scene model loading and attribute attachment.

[0052] 4. Use blockchain technology to store pipeline data. The specific steps of blockchain storage are as follows:

[0053] 1) Preprocess each model data and corresponding pipeline data into a unified format, and process each data into blocks. When entering the network, each node in the blockchain network generates a local public and private key pair as its own identification code in the network. When a node waits for the local role to become a candidate node, it broadcasts the leadership application information to other nodes in the network;

[0054] 2) After the application is approved, the candidate node becomes the leader node, and the other nodes become follower nodes. The leader node then broadcasts the block record information. After receiving the information, the follower node broadcasts the received information to other follower nodes and records the number of repetitions. It uses the information with the most repetitions to generate a block header and sends a verification application to the leader node;

[0055] 3) After verification, the leader node sends an add command and enters a dormant period. During the dormant period, it cannot apply to become a leader node again until the end of the dormant period. After the follower node receives the confirmation message, it adds each newly generated group of blocks to the blockchain and returns the candidate identity.

[0056] Combine the following Figure 1 , the technical solution of the pipeline data management method based on GIS platform of the present invention is further explained, and the specific steps are as follows:

[0057] 1. Data collection: Collect pipeline data through the data collection APP and use drones to take station images.

[0058] (1) Intelligently collect pipeline data through the data collection APP.

[0059] The pipeline data mainly includes but is not limited to: pipeline code, weld code, weld X coordinate, weld Y coordinate, weld Z coordinate, and welding method.

[0060] (2) Through on-site survey, determine the shooting altitude and 45-degree camera angle of the station image, plan the flight route of the drone and the positions of the five control points, and complete the shooting to obtain the station image, such as Figure 2 shown.

[0061] 2. Data processing.

[0062] (1) Pipeline data processing.

[0063] Using convolutional neural networks and feature classification methods, standard data and abnormal data are obtained, marked, and equally divided into training sets and test sets.

[0064] 1. According to GBT 51296-2018 petrochemical engineering digital delivery standard, the collected pipeline data is classified to obtain standard data and abnormal data that meet the model training requirements;

[0065] Standard data is shown in Table 1:

[0066] Table 1

[0067]

[0068] The abnormal data is shown in Table 2:

[0069] Table 2

[0070]

[0071] 2. Label and classify standard data and abnormal data.

[0072] The pipeline coding field format exception flag is shown in Table 3:

[0073] Table 3

[0074]

[0075] The pipeline code field format is normally marked as shown in Table 4:

[0076] Table 4

[0077]

[0078] The abnormal markings of weld X coordinates are shown in Table 5:

[0079] Table 5

[0080]

[0081] The normal marking of the weld X coordinate is shown in Table 6:

[0082] Table 6

[0083]

[0084] The abnormal mark of weld code field format is shown in Table 7:

[0085] Table 7

[0086]

[0087] The normal format of the weld code field is shown in Table 8:

[0088] Table 8

[0089]

[0090] 3. Divide the labeled standard data and abnormal data equally into training sets and test sets. The training set is used to train the model, and the test set is used to evaluate the performance of the model.

[0091] The training set is shown in Table 9:

[0092] Table 9

[0093]

[0094] The test set is shown in Table 10:

[0095] Table 10

[0096]

[0097] (2) Image data processing.

[0098] The collected image information is decomposed frame by frame, and then divided into blocks according to the display ratio of each group of image data. The high-frequency components of the divided images are analyzed and extracted through Fourier transform, and the high-frequency noise points on the edge of the image in the data are smoothed through Gaussian filtering to obtain high-definition images such as Figure 3 shown.

[0099] 3. Model training.

[0100] (1) Using the clustering algorithm, the training set samples are imported into the convolutional neural network one by one to obtain the initial model for pipeline data anomaly detection.

[0101] (2) Initialize the network connection weight w to 0.5 within the specified interval 0-1 of the initial model. Train the model through a set of training sets, then test the model through the test set, record the actual output of the model, and preset the expected output of the initial model. Compare the expected network output with the actual network output as shown in Table 11.

[0102] Table 11

[0103]

[0104] (3) Calculate the local error δ of the neuron.

[0105] The specific calculation formula of local error is as follows:

[0106]

[0107] Among them, among them, represents the local error of the t-th position of the q-th neuron in the l-th hidden layer, represents the output of the t-th position of the q-th neuron in the l-th hidden layer, e represents the e-th neuron, represents the sum of the output of the l+1th hidden layer and the offset of the l-1th hidden layer within the range of the eth neuron, Represents the offset between the input data of the e-th neuron and the q-th neuron;

[0108] When the local error δ exceeds the threshold set by the staff, the weight w of the adjustment model is trained and updated according to the learning rule equation, and the preset learning rate η (l) And the step length lists all possible data results;

[0109] The learning rule equation is as follows:

[0110]

[0111] In the formula, represents the weight of the kth neuron in the lth hidden layer, η (l) is the learning rate parameter corresponding to each network layer. Generally, the learning rate is set to an equal value. represents the local error of the kth neuron in the lth hidden layer, Represents the output result of the p-th neuron at the t-th position in the l-1-th output layer.

[0112] (4) Complete the above training and testing of the model for all training sets and test sets, and calculate the root mean square error of the detection results; by continuously adjusting the weight w to minimize the root mean square error value, the optimal model for pipeline data outlier detection is obtained.

[0113] Abnormal feedback is shown in Table 12:

[0114] Table 12

[0115]

[0116] 4. Correction of abnormal pipeline data.

[0117] According to the abnormal feedback results output by the pipeline data abnormality verification model, each group of abnormal data is accurately fed back to the staff, and the abnormal data is corrected to obtain standard pipeline data.

[0118] The standard pipeline data is shown in Table 13:

[0119] Table 13

[0120]

[0121] 5. Three-dimensional modeling.

[0122] The average grayscale value of each high-definition image data is calculated respectively, and then the grayscale value of each group of pixels in each image data after segmentation is compared with the calculated average value, and all pixels with grayscale values ​​greater than the average value constitute the segmentation target, and all pixels with grayscale values ​​less than the average value constitute the background of the segmented image. Then, according to the segmentation results, the high-definition image automatic modeling technology is used to construct a three-dimensional scene model of the station. Figure 4 shown.

[0123] 6. Release the 3D model and complete data connection.

[0124] (1) Using GIS platform 3D model publishing technology to complete the 3D station model publishing Figure 5 shown.

[0125] (2) Based on the collected pipeline data, pipeline drawings, and pipeline survey data, the pipeline unique code is used to match the pipeline with the 3D model, load the 3D model, and attach attributes. Figure 6 shown.

[0126] 7. The specific steps of data storage.

[0127] (1) Preprocess each model data and corresponding pipeline data into a unified format, and process each data into qualified blocks. When joining the network, each node in the blockchain network generates a local public-private key pair as its own identification code in the network. When a node waits for its local role to become a candidate node, it broadcasts the leadership application information to other nodes in the network.

[0128] (2) After the application is approved, the candidate node becomes the leader node and the other nodes become follower nodes. The leader node then broadcasts the block record information. After receiving the information, the follower node broadcasts the received information to other follower nodes and records the number of repetitions. It uses the information with the most repetitions to generate a block header and sends a verification application to the leader node.

[0129] (3) After verification, the leader node sends an add command and enters a dormant period. During the dormant period, it cannot apply to become a leader node again until the end of the dormant period. After the follower node receives the confirmation message, it adds the newly generated groups of blocks to the blockchain and returns the candidate identity.

[0130] Specific implementation methods are given above, but the present invention is not limited to the described implementation methods. The basic idea of ​​the present invention lies in the above basic scheme. For ordinary technicians in this field, it does not take creative work to design various deformed models, formulas, and parameters according to the teachings of the present invention. Changes, modifications, substitutions, and variations of the implementation methods without departing from the principles and spirit of the present invention still fall within the scope of protection of the present invention.

Claims

1. A pipeline data management method based on GIS platform, characterized in that: The following steps are involved: 1) Obtain pipeline data and station image data; 2) Inputting the pipeline data into the trained machine learning model to classify the pipeline data, identify normal pipeline data and abnormal pipeline data, and correct the abnormal pipeline data, thereby obtaining standard pipeline data including the corrected pipeline data and normal pipeline data; 3) Process the image data to obtain high-definition images, use the high-definition images to complete three-dimensional scene modeling, publish the three-dimensional scene model through the three-dimensional engine of the GIS platform, and attach the standard pipeline data to the three-dimensional scene model.

2. The pipeline data management method based on GIS platform according to claim 1 is characterized in that: The pipeline data in step 1) includes: pipeline code, weld code, and weld coordinates.

3. The pipeline data management method based on GIS platform according to claim 2 is characterized in that: The abnormal types of the abnormal pipeline data classified in step 2) include pipeline coding format abnormality, weld coding format abnormality and weld coordinate abnormality.

4. The pipeline data management method based on GIS platform according to claim 1 is characterized in that: Step 2) The machine learning model is a convolutional neural network.

5. The pipeline data management method based on GIS platform according to claim 1 is characterized in that: Step 1) The specific method for acquiring station image data is to determine the flight altitude and camera angle of the station drone image shooting, plan the drone flight route and control point position, and use the drone to obtain the station image picture.

6. The pipeline data management method based on GIS platform according to claim 1 is characterized in that: In step 2), the image data is processed by decomposing the collected image data frame by frame, and then dividing the image data into blocks according to the display ratio of each frame. The high-frequency components in the divided image are analyzed and extracted through Fourier transform, and the high-frequency noise at the edge of the image in the data is smoothed through Gaussian filtering to obtain a high-definition image.

7. The pipeline data management method based on GIS platform according to claim 1 is characterized in that: In step 3), when attaching the standard pipeline data to the three-dimensional scene model, it is also necessary to combine the pipeline drawing materials and pipeline survey data, and then complete the matching with the three-dimensional scene model through the pipeline unique code to complete the three-dimensional scene model loading and attribute attachment.

8. The pipeline data management method based on GIS platform according to any one of claims 1 to 7, characterized in that: After completing the standard pipeline data connection, the data needs to be stored using blockchain, and the storage steps include: (1) Preprocess the 3D scene model data and the corresponding pipeline data into a unified format, and process the data into blocks. When joining the blockchain network, each node in the blockchain network generates a local public-private key pair as its own identification code in the network. When the node waits for the local role to become a candidate node, it broadcasts the leadership application information to other nodes in the network; (2) After the application is approved, the candidate node becomes the leader node and the other nodes become follower nodes. The leader node then broadcasts the block record information. After receiving the information, the follower node broadcasts the received information to other follower nodes and records the number of repetitions. It uses the information with the most repetitions to generate a block header and sends a verification application to the leader node. (3) After verification, the leader node sends an add command and enters a dormant period. During the dormant period, it cannot apply to become a leader node again until the end of the dormant period. After the follower node receives the confirmation message, it adds the newly generated groups of blocks to the blockchain and returns the candidate identity.

Images