Intelligent pipeline flushing control method and system based on pressure change

Abstract

The invention provides an intelligent pipeline flushing control method and system based on pressure change, a target pipeline is divided into a plurality of sub-pipelines according to length, pipe diameter change or an easy-to-block risk area, and pressure sensors are arranged at the two ends of each sub-pipeline so as to collect pressure data in the pipeline in real time; a spray head with an electromagnetic valve is arranged at one end of each section of sub-pipeline and is opened or closed according to an instruction, and the water spraying pressure and flow are adjusted; a graph neural network model based on deep learning is adopted, the pressure data of all the sections of pipelines serve as input, and the pressure transmission relation and the blockage influence rule between all the sections of pipelines are learned through training; pressure data collected by the pressure sensor are received in real time, the trained graph neural network model is called for analysis, a nozzle flushing control instruction is generated and sent to the corresponding nozzle, and specifically, the problem of pipeline blockage can be accurately solved through segmented monitoring and intelligent control.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of intelligent pipeline flushing control, and particularly relates to a pipeline intelligent flushing control method and system based on pressure change. BACKGROUND

[0002] In various industrial production, water treatment and environmental protection fields, the application of medium adding devices such as activated carbon is extremely wide.

[0003] In the long-term operation process of the existing medium adding device, the inside of the pipeline is prone to residual drug crystallization, impurity precipitation and the like, which causes pipeline blockage, poor flow, affects the accuracy and stability of medium addition, and further reduces production efficiency and increases maintenance cost.

[0004] The traditional pipeline flushing method is mostly manual flushing, which is tedious to operate and the flushing effect is difficult to guarantee, and cannot meet the needs of automatic production. At the same time, some automatic flushing devices have complex structure and high cost, and do not have wide applicability. SUMMARY

[0005] Therefore, the embodiment of the application provides a pipeline intelligent flushing control method and system based on pressure change, which aims to realize the function of automatically flushing the pipeline by monitoring the pressure of the pipeline and intelligently controlling the electric valve, solve the problems of easy blockage and inconvenient flushing of the pipeline of the existing medium adding device, and improve the operation efficiency and stability of the medium adding device.

[0006] The first aspect of the embodiment of the application provides a pipeline intelligent flushing control method based on pressure change, which comprises the following steps: The target pipeline is divided into a plurality of sub-pipelines according to length, pipe diameter change or risk area of easy blockage, wherein a pressure sensor is arranged at both ends of each sub-pipeline to collect pressure data in the pipeline in real time, and a nozzle with an electromagnetic valve is arranged at one end of each sub-pipeline to be opened or closed according to an instruction and to adjust water pressure and flow; A graph neural network model based on deep learning is adopted to input the pressure data of each pipeline segment, learn the pressure transmission relationship and blockage influence law between each pipeline segment through training, and generate a nozzle flushing control instruction. The pressure data collected by the pressure sensor is received in real time, the trained graph neural network model is called for analysis, a nozzle flushing control instruction is generated, and the nozzle flushing control instruction is sent to the corresponding nozzle.

[0007] Further, the step of adopting the graph neural network model based on deep learning to input the pressure data of each pipeline segment, learning the pressure transmission relationship and blockage influence law between each pipeline segment through training comprises the following steps: In the normal operation stage of the pipeline, the pressure sensor data of each section of the pipeline is continuously collected, and the pressure change under different working conditions is recorded as a normal state sample; Different degrees of blockage are set in different sub-pipelines, and the pressure data of each section of the pipeline when blockage occurs is collected to construct an abnormal state sample; The collected data is preprocessed, including data cleaning and normalization, to obtain preprocessed data; According to the graph neural network, each section of the sub-pipeline is regarded as a node in the graph, and the connection relationship between the pipelines is regarded as an edge. The mutual influence of the pressure changes of each section of the pipeline is learned through information transmission between nodes. To predict the blockage degree of each section of the pipeline, the mean square error is used as the loss function, and the stochastic gradient descent algorithm is used for model training to obtain the trained graph neural network model.

[0008] Further, the step of real-time receiving pressure data collected by the pressure sensor, calling the trained graph neural network model for analysis, generating a nozzle flushing control instruction, and sending it to the corresponding nozzle includes: Real-time receive pressure data collected by the pressure sensor, and input the trained graph neural network model, input the prediction result; According to the prediction result, determine the abnormal position and the influence range, and according to the abnormal position and the influence range, combine the pressure change of the adjacent pipeline to determine the required flushing pressure and flushing time of the nozzle; According to the flushing pressure and the flushing time, control the corresponding nozzle to flush the pipeline, wherein during the flushing process, the pressure change is continuously monitored, if the pressure returns to normal, stop flushing; if the pressure does not return to normal, increase the flushing intensity or start the adjacent pipeline nozzle to assist flushing.

[0009] Further, in the step of real-time receiving pressure data collected by the pressure sensor, calling the trained graph neural network model for analysis, generating a nozzle flushing control instruction, and sending it to the corresponding nozzle, the fuzzy logic model is combined with the graph neural network model. The graph neural network model is used to predict the pipeline blockage state, and the fuzzy logic is used to generate a flushing control instruction based on the prediction result and the real-time pressure change.

[0010] Further, in the fuzzy logic model, the pressure change rate and pressure difference of each section of the pipeline are used as the input of the fuzzy logic model; the nozzle flushing pressure and flushing time are used as the output of the fuzzy logic model; based on actual experience and expert knowledge, the fuzzy rules of the fuzzy logic model are formulated; according to the fuzzy rules, the input variables are subjected to fuzzy reasoning to obtain fuzzy output results, and then the fuzzy output results are converted into specific flushing pressure and flushing time through defuzzification to guide the nozzle to perform flushing operation.

[0011] Further, the graph neural network model adopts a graph convolution network as a basic architecture, wherein 3-5 layers of graph convolution network layers are arranged, each layer of graph convolution network layer updates its own features by aggregating information of adjacent nodes; a ReLU activation function is used after each layer of convolution network layer; the number of output layer neurons is the same as the number of pipeline sections, and each neuron corresponds to a predicted value of the blockage degree of a pipeline section.

[0012] Further, in the step of taking prediction of the blockage degree of each pipeline section as a target, adopting a mean square error as a loss function, and using a stochastic gradient descent algorithm to train the model to obtain the trained graph neural network model, the loss function formula is: ; Wherein, n is the number of samples, is an actual blockage degree label, is a predicted blockage degree of the model.

[0013] The second aspect of the embodiment of the present application provides a pipeline intelligent flushing control system based on pressure change, which is used to realize the pipeline intelligent flushing control method based on pressure change provided by the first aspect of the present application. The system comprises: A division module is configured to divide a target pipeline into a plurality of sub-pipeline sections according to length, pipe diameter change or a risk area prone to blockage, wherein pressure sensors are arranged at both ends of each sub-pipeline section to collect pressure data in the pipeline in real time, and a spray head with an electromagnetic valve is arranged at one end of each sub-pipeline section to open or close according to an instruction and adjust water pressure and flow rate. A training module is configured to use a graph neural network model based on deep learning to input pressure data of each pipeline section, and learn pressure transmission relationship and blockage influence law between each pipeline section through training. An analysis module is configured to receive pressure data collected by the pressure sensor in real time, call the trained graph neural network model for analysis, generate a spray head flushing control instruction, and send the instruction to the corresponding spray head.

[0014] The third aspect of the embodiment of the present application provides a computer readable storage medium, comprising: The readable storage medium stores one or more programs, which are executed by a processor to realize the pipeline intelligent flushing control method based on pressure change as described in the first aspect.

[0015] The fourth aspect of the embodiment of the present application provides an electronic device, comprising a memory and a processor, wherein: The memory is used to store a computer program. The processor is used to realize the pipeline intelligent flushing control method based on pressure change as described in the first aspect when executing the computer program stored on the memory.

[0016] The pipeline intelligent flushing control method and system based on pressure change provided in the embodiments of the present application, by dividing the target pipeline into several sub-pipelines according to length, pipe diameter change or easy-to-block risk area, wherein pressure sensors are arranged at both ends of each sub-pipeline to collect pressure data in the pipeline in real time, a nozzle with an electromagnetic valve is arranged at one end of each sub-pipeline to be opened or closed according to instructions and to adjust water pressure and flow; a graph neural network model based on deep learning is adopted to input pressure data of each pipeline segment, and to learn pressure transmission relationship and blockage influence law between each pipeline segment through training; real-time pressure data collected by the pressure sensor is received, the trained graph neural network model is called for analysis, nozzle flushing control instructions are generated, and the instructions are sent to the corresponding nozzle, specifically, the segmented monitoring and intelligent control can accurately respond to the pipeline blockage problem. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 Structure schematic diagram of adding device to medium; Figure 2 Implementation flowchart of a pipeline intelligent flushing control method based on pressure change provided for the first embodiment of the present application; Figure 3 Structure block diagram of a pipeline intelligent flushing control system based on pressure change provided for the second embodiment of the present application; Figure 4 Structure block diagram of an electronic device provided for the third embodiment of the present application.

[0018] Reference signs: 1, medium supply tank; 2, medium pump; 3, pressure sensor; 4, electric valve; 5, DI water flushing branch pipe; 6, main pipeline; 7, equipment needing medium; 8, controller; 9, signal line. DETAILED DESCRIPTION

[0019] In order to facilitate the understanding of the present application, the present application will be described more fully below with reference to the related drawings. The drawings show several embodiments of the present application. However, the present application can be realized in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present application more thorough and comprehensive.

[0020] It should be understood that when an element as a layer, region or plate is referred to as being "on" another element, it can be directly on the other element or intervening elements can also be present. In addition, it should be understood that when an element is referred to as being "connected" to another element, it can be directly connected to the other element or intervening elements can also be present. As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the description of the application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items.

[0022] For example, the basic medium adding device has only a short pipeline, and specifically, the medium adding device mainly comprises a medium supply tank, a pressure sensor, an electric valve, a controller, a medium pump, a DI water flushing branch, and a main pipeline. The medium supply tank is used to store a medium preparation liquid, and is connected to a device requiring medium addition through a pipeline. The pressure sensor is installed at one end of the main pipeline close to the outlet of the medium pump, and monitors the pressure change in the pipeline in real time, and converts the pressure signal into an electric signal and transmits the electric signal to the controller. It should be noted that the flushing branch is connected to the main pipeline close to the outlet of the medium pump through a tee joint, and the opening and closing state of the flushing branch is controlled by the controller. One end of the flushing branch is in communication with the main pipeline, and the other end is connected to flushing DI water, and the DI water is deionized water.

[0023] More specifically, the controller is internally preset with a normal medium adding pressure range and a flushing trigger pressure threshold value. In the normal medium adding state, when the pressure value detected by the pressure sensor is within the normal pressure range, the electric valve on the flushing branch is closed. When the pressure value detected by the pressure sensor is higher than the flushing trigger pressure threshold value, it indicates that the pipeline may be blocked, and the controller issues an instruction to open the electric valve on the flushing branch, and the DI water enters the main pipeline through the flushing branch to flush the main pipeline to remove impurities and blockages in the pipeline. After cleaning, when the pressure value detected by the pressure sensor is within the normal pressure range, the electric valve on the flushing branch is closed.

[0024] Please refer to Figure 1The structure schematic view of the medium adding device is shown in the figure, and the connection relationship of each component is that the medium supply tank 1 is stably installed at a suitable position to ensure that it can be safely stored, the medium supply tank 1 is connected to the inlet end of the medium pump 2 through the main pipeline 6, the equipment 7 needing to add medium is connected to the outlet end of the medium pump 2 through the main pipeline 6, the pressure sensor 3 is installed on the main pipeline 6 close to the outlet end of the medium pump 2, the electric valve 4 is installed on the DI water flushing branch 5, the DI water flushing branch 5 is connected between the outlet end of the medium pump 2 and the pressure sensor 3 through a tee joint, and the pressure sensor 3 and the electric valve 4 are connected to the controller 8 through the signal line 9.

[0025] It can be understood that the above only gives a simple pipeline flushing control scheme based on pressure change, but for longer and different specification pipelines, the pressures of different parts of the pipeline and the pressure when the pipeline is not smooth are different, and only by setting one pressure sensor and determining the pressure value collected by the pressure sensor, the blockage in the pipeline cannot be timely removed to ensure the smoothness of the pipeline. Therefore, a pipeline intelligent flushing control method based on pressure change is needed, and the specific implementation mode is shown as follows.

[0026] Embodiment one The embodiment one of the application provides a pipeline intelligent flushing control method based on pressure change, please refer to Figure 2 The embodiment one of the application provides a pipeline intelligent flushing control method based on pressure change, please refer to

[0027] Step S01, the target pipeline is divided into several sub-pipelines according to length, pipe diameter change or easy blockage risk area.

[0028] Among them, the two ends of each sub-pipeline are provided with pressure sensors to collect pressure data in the pipeline in real time, and one end of each sub-pipeline is provided with a spray head with an electromagnetic valve to open or close according to instructions and adjust water pressure and flow. It should be noted that the target pipeline can be divided into several sub-pipelines according to length, pipe diameter change or easy blockage risk area by manual method.

[0029] Step S02, a graph neural network model based on deep learning is used to input the pressure data of each section of the pipeline, and the pressure transmission relationship and blockage influence law between each section of the pipeline are learned through training.

[0030] Specifically, in the normal operation stage of the pipeline, the pressure sensor data of each section of the pipeline is continuously collected, and the pressure change under different working conditions is recorded as a normal state sample; The pipeline blockage experiment can be simulated by human, different degrees of blockage are set in different sub-pipelines, the pressure data of each section of the pipeline when the blockage occurs is collected, and an abnormal state sample is constructed; The collected data is preprocessed, including data cleaning and normalization, to obtain preprocessed data; According to the graph neural network, each section of the sub-pipeline is regarded as a node in the graph, and the connection relationship between the pipelines is regarded as an edge. The mutual influence of the pressure changes of each section of the pipeline is learned through information transmission between nodes. In other words, each section of the sub-pipeline is abstracted as a node in the graph, and the node features are real-time pressure data collected by the pressure sensor of the section of the pipeline, historical pressure change trend and other information. The physical connection relationship between the pipelines is defined as an edge, and the weight of the edge can be set according to factors such as the pipe diameter and length of the pipeline, reflecting the transmission efficiency of pressure between different pipelines. To predict the blockage degree of each section of the pipeline, the mean square error is used as the loss function, and the stochastic gradient descent algorithm is used for model training to obtain the trained graph neural network model. It should be noted that the graph neural network model uses a graph convolution network as the basic architecture. In this architecture, 3-5 layers of graph convolution network layers are set. Each layer of the graph convolution network layer updates its own features by aggregating the information of adjacent nodes. For example, the first layer of the graph convolution network layer receives the original node pressure data and learns the local pressure change pattern through convolution operation. The subsequent layers further extract high-level features and capture the complex correlations between different pipelines. The number of neurons in each layer can be adjusted according to the data dimension and computing resources. In the embodiment of the present application, the number of neurons starts from 128 and is optimized through experiments. A ReLU activation function is used after each convolution network layer, and the formula is , which introduces a nonlinear factor and enhances the model's ability to express complex pressure change relationships. The number of neurons in the output layer is the same as the number of pipeline sections, and each neuron corresponds to a blockage degree prediction value of a section of the pipeline. The output range is [0, 1], and the closer the value is to 1, the higher the blockage degree.

[0031] More specifically, in the training process, the mean square error is used as the loss function, and the loss function formula is: ; where n is the number of samples, is the actual blockage degree label, is the blockage degree predicted by the model, and MSE can effectively measure the average error between the model prediction value and the true value. Optimizing this function can improve the prediction accuracy of the model.

[0032] Further, the stochastic gradient descent (SGD) algorithm and its variants (such as the Adam optimizer) are selected for model training. In the embodiment of the present application, the Adam optimizer is used, and the initial learning rate is set to 0.001. The learning rate is dynamically adjusted according to the loss change in the training process to prevent the model from falling into a local optimal solution. At the same time, to avoid overfitting, an L2 regularization term can be added in the training process to constrain the size of the model parameters.

[0033] Further, the data set is divided into training set, validation set and test set, the ratio is 7:2:1. In the training process, the batch training method is adopted, the batch size is set to 32-64, and the loss is calculated and the model parameters are updated every iteration from the training set. After completing an epoch (traversing the training set once), the model performance is evaluated on the validation set, and the model parameters are adjusted or the training is terminated in advance (early stopping method) to prevent overfitting.

[0034] In step S03, real-time pressure data collected by the pressure sensor is received, a trained graph neural network model is called for analysis, a nozzle flushing control instruction is generated, and the nozzle is sent.

[0035] It should be noted that during the flushing process, the medium supply is not performed, and the medium supply pipeline is cut off. After the medium in the pipeline is emptied, the flushing is performed, wherein the waste liquid generated by the flushing is discharged through the drain port of the controllable switch reserved at the tail of the pipeline.

[0036] In the embodiment of the application, real-time pressure data collected by the pressure sensor is received and input into a trained graph neural network model, and a prediction result is input, wherein the pressure sensor collects pressure data of each section of the pipeline at a fixed frequency of 1 time per second. The frequency can be adjusted according to the complexity of the pipeline working condition and the response speed requirement. For example, for a pipeline prone to blockage or high flow rate, the frequency can be increased to 5 times per second. In addition, a preliminary threshold of pressure change can be set, such as an increase of 1.5 times in the average value of the pressure of a section of the pipeline compared with the average value during normal operation, triggering a preliminary warning. After the preliminary warning, the real-time pressure data of the section of the pipeline and the adjacent pipeline are input into the trained neural network model, and the model outputs a blockage degree prediction value. If the prediction value exceeds 0.7 (which can be adjusted according to the actual situation), it is determined that the section of the pipeline is blocked; According to the prediction result, the abnormal position and the influence range are determined. Specifically, according to the blockage degree prediction value of each section of the pipeline output by the neural network model, the core position of the blockage is determined in combination with the pressure sensor data. If the prediction value of a section of the pipeline is the highest and exceeds the threshold value, it is positioned as the blockage point. Further, the pressure change of the adjacent pipeline in the upstream and downstream directions is analyzed with the blockage point as the center. When the pressure change rate of the adjacent pipeline is lower than the preset value, the expansion is stopped, and the pipeline in the range is determined as the affected area; Subsequently, according to the abnormal position and the influence range, the flushing pressure and the flushing time required by the nozzle are determined in combination with the pressure change of the adjacent pipeline. It should be noted that the flushing pressure is calculated based on the blockage degree, the length of the pipeline and the pressure of the adjacent pipeline according to the formula P = P0 * k * S, wherein P0 is a reference pressure, k is a coefficient, and S is a degree of blockage; compensation is performed according to a pressure difference between an upstream and a downstream, and the pressure is increased by 0.2 MPa for each 1 MPa of the upstream pressure that is higher than a normal value, and the pressure is decreased by 0.1 MPa for each 1 MPa of the downstream pressure that is lower than the normal value; a formula is P = P0 * k * S, wherein P0 is a reference pressure, k is a coefficient, and S is a degree of blockage; and the flushing time is determined according to the degree of blockage and the pipeline volume. wherein β is a time coefficient, and V is a pipeline volume; and the flushing time is dynamically adjusted according to a pressure recovery condition, and the flushing time is extended in proportion if a pressure drop speed is lower than an expected value. The corresponding nozzle is controlled to flush the pipeline according to the flushing pressure and the flushing time, and the pressure change is continuously monitored during the flushing process; if the pressure returns to normal, the flushing is stopped; if the pressure does not return to normal, the flushing intensity is increased or the adjacent nozzle is started to assist in flushing.

[0037] In some other embodiments of the present application, in a pipeline blockage scenario, the pressure change and the degree of blockage are not a simple linear relationship, and there are various interference factors (such as fluid flow fluctuation and environmental temperature change). In order to generate more reasonable flushing control instructions, specifically, a fuzzy logic model and a graph neural network model are combined, the graph neural network model is used to predict the pipeline blockage state, and the fuzzy logic is used to generate the flushing control instructions based on the prediction result and the real-time pressure change.

[0038] It should be noted that in the fuzzy logic model, the pressure change rate and the pressure difference of each pipeline segment are taken as inputs of the fuzzy logic model. For example, the pressure change rate can be divided into "low change rate", "medium change rate", and "high change rate"; the pressure difference can be divided into "small difference", "medium difference", and "large difference"; the nozzle flushing pressure and the flushing time are taken as outputs of the fuzzy logic model. For example, the flushing pressure can be divided into "low pressure", "medium pressure", and "high pressure"; the flushing time can be divided into "short time", "medium time", and "long time"; the fuzzy rules of the fuzzy logic model are formulated based on actual experience and expert knowledge. In the embodiments of the present application, if the pressure change rate is "high" and the pressure difference is "large", the flushing pressure is "high pressure" and the flushing time is "long time"; if the pressure change rate is "low" and the pressure difference is "small", the flushing pressure is "low pressure" and the flushing time is "short time". A complete rule base is formed through a large number of actual cases and optimization; the fuzzy output result is obtained through fuzzy reasoning on the input variables according to the fuzzy rules; and the fuzzy output result is converted into specific flushing pressure and flushing time through defuzzification to guide the nozzle to perform the flushing operation. The defuzzification can be realized by the gravity method.

[0039] Finally, the flushing effect is evaluated, and the pressure recovery rate and the pressure fluctuation range are calculated respectively. The calculation formula of the pressure recovery rate is If the pressure recovery rate exceeds 90%, it is determined that the flushing effect is good; the pressure fluctuation range refers to the pressure fluctuation within 10 minutes after flushing, and if the pressure fluctuation range is within ±5% of the normal pressure, it is considered stable.

[0040] If the flushing effect does not meet the expectation, the system automatically analyzes the reason, such as incomplete removal of the blockage or improper setting of the flushing parameters, and adjusts the flushing strategy again, including changing the flushing mode, adjusting the pressure and time, etc., and starts the secondary flushing.

[0041] In summary, the embodiment of the present application proposes a pipeline intelligent flushing control method based on pressure change. The target pipeline is divided into several sub-pipelines according to length, pipe diameter change or risk area prone to blockage. Pressure sensors are installed at both ends of each sub-pipeline to collect real-time pressure data in the pipeline. A nozzle with an electromagnetic valve is provided at one end of each sub-pipeline to open or close according to instructions and adjust the water pressure and flow rate. A graph neural network model based on deep learning is used to input the pressure data of each pipeline segment, and the pressure transmission relationship and blockage influence law between each pipeline segment are learned through training. Real-time pressure data collected by the pressure sensor is analyzed by calling the trained graph neural network model to generate nozzle flushing control instructions and send them to the corresponding nozzle. Specifically, segmented monitoring and intelligent control can accurately respond to pipeline blockage problems.

[0042] Embodiment two The embodiment two of the present application provides a pipeline intelligent flushing control system 200 based on pressure change. Please refer to Figure 3 , which is a structural diagram of a pipeline intelligent flushing control system based on pressure change. The pipeline intelligent flushing control system 200 based on pressure change comprises: A division module 21 is used to divide the target pipeline into several sub-pipelines according to length, pipe diameter change or risk area prone to blockage. Pressure sensors are installed at both ends of each sub-pipeline to collect real-time pressure data in the pipeline. A nozzle with an electromagnetic valve is provided at one end of each sub-pipeline to open or close according to instructions and adjust the water pressure and flow rate. A training module 22 is used to input the pressure data of each pipeline segment into a graph neural network model based on deep learning to learn the pressure transmission relationship and blockage influence law between each pipeline segment through training. The graph neural network model uses a graph convolution network as a basic architecture. A 3-5 layer graph convolution network layer is set, and each layer of the graph convolution network layer updates its own features by aggregating the information of adjacent nodes. A ReLU activation function is used after each convolution network layer. The number of output layer neurons is the same as the number of pipeline segments, and each neuron corresponds to a blockage degree prediction value of a pipeline segment. The analysis module 23 is configured to receive pressure data collected by the pressure sensor in real time, call the trained graph neural network model to perform analysis, generate a nozzle flushing control instruction, and send the nozzle flushing control instruction to the corresponding nozzle. The graph neural network model is used to predict the pipe blockage state, and the fuzzy logic is used to generate the flushing control instruction based on the prediction result and real-time pressure change. In the fuzzy logic model, the pressure change rate and pressure difference of each pipe section are taken as the input of the fuzzy logic model, the nozzle flushing pressure and flushing time are taken as the output of the fuzzy logic model, the fuzzy rules of the fuzzy logic model are formulated based on actual experience and expert knowledge, the input variables are subjected to fuzzy reasoning according to the fuzzy rules, the fuzzy output result is obtained, and the fuzzy output result is converted into specific flushing pressure and flushing time through defuzzification to guide the nozzle to perform the flushing operation.

[0043] Further, in some other embodiments of the present application, the training module 22 comprises: A recording unit is configured to continuously collect pipe section pressure sensor data in the normal operation stage of the pipe, record pressure changes under different working conditions, and take the normal state samples. A collection unit is configured to set different degrees of blockage in different sub-pipes, collect pressure data of each pipe section when blockage occurs, and construct abnormal state samples. A preprocessing unit is configured to preprocess the collected data, including data cleaning and normalization processing, to obtain preprocessed data. A definition unit is configured to regard each sub-pipe as a node in a graph and the connection relationship between the pipes as an edge according to the graph neural network, and learn the mutual influence of the pressure changes of each pipe section through information transmission between the nodes. A training unit is configured to take the prediction of the blockage degree of each pipe section as the target, use the mean square error as the loss function, use the stochastic gradient descent algorithm to train the model, obtain the trained graph neural network model, and the loss function formula is: ; Wherein, n is the number of samples, is the actual blockage degree label, is the blockage degree predicted by the model.

[0044] Further, in some other embodiments of the present application, the analysis module 23 comprises: A receiving unit is configured to receive pressure data collected by the pressure sensor in real time, input the trained graph neural network model, and input the prediction result. The determining unit is configured to determine an abnormal position and an influence range according to the prediction result, and determine a required flushing pressure and a flushing time of the spray head according to the abnormal position and the influence range and a pressure change of a neighboring pipeline. The control unit is configured to control the corresponding spray head to flush the pipeline according to the flushing pressure and the flushing time, wherein during the flushing process, the pressure change is continuously monitored, if the pressure returns to normal, the flushing is stopped, and if the pressure does not return to normal, the flushing intensity is increased or the neighboring pipeline spray head is started to assist in flushing.

[0045] Embodiment three Embodiment three of the present application proposes an electronic device, please refer to Figure 4 , which is a structural block diagram of an electronic device, comprising a memory 20, a processor 10, and a computer program 30 stored in the memory and executable on the processor, wherein the processor 10 implements the above-mentioned pipeline intelligent flushing control method based on pressure change when executing the computer program 30.

[0046] In some embodiments, the processor 10 can be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip, which is used to run program codes or process data stored in the memory 20, such as executing access restriction programs.

[0047] In some embodiments, the memory 20 can be an internal storage unit of the electronic device, such as a hard disk of the electronic device. In other embodiments, the memory 20 can also be an external storage device of the electronic device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device. Further, the memory 20 can include both the internal storage unit and the external storage device of the electronic device. The memory 20 can be used not only to store application software and various data of the electronic device, but also to temporarily store data that has been output or will be output.

[0048] The present application also proposes a computer readable storage medium having a computer program stored thereon, wherein the program is executed by a processor to implement the above-mentioned pipeline intelligent flushing control method based on pressure change.

[0049] Those skilled in the art can appreciate that the logic and / or steps represented in the flow diagrams, or otherwise described herein, can be embodied in any computer-readable medium for use by an instruction execution system, apparatus, or device, such as a computer-based system, processor- containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions, or in conjunction with which the instructions can be executed. In the context of this specification, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. The computer-readable medium can be, for example but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, device, or propagation medium.

[0050] More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection (electronic) having one or more wires, a portable computer diskette (magnetic), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can also be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, for example via an optical scanner, then compiled, interpreted, or otherwise processed, and stored in a computer memory in a form that can be later executed by a computer. In some embodiments, the computer-readable medium can be a transmission medium that can contain, store, or carry the program for use by or in connection with an instruction execution system, apparatus, or device.

[0051] It should be understood that aspects of the application can be implemented in hardware, software, firmware, or combinations thereof. In the embodiments described above, various steps or methods can be implemented, for example, by software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, and in another embodiment, any of the following technologies, or combinations thereof, can be used: a discrete logic circuit having logic gates for implementing logic functions upon data signals, an application specific integrated circuit having appropriate combinational logic gates, a programmable gate array (PGA), a field programmable gate array (FPGA), and / or the like.

[0052] In the description of the specification, the description of the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" and the like means that the specific feature, structure, material, or characteristic being described is included in at least one embodiment or example of the application. The illustrative description of the above terms in the specification does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0053] The above embodiments only express several implementation manners of the present application, the description is more specific and detailed, but it cannot be understood as the limitation of the patent scope of the present application. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, several modifications and improvements can be made, which belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

Claims

1. A method for intelligent flushing control of a pipeline based on pressure variation, the method comprising: determining a pressure variation in the pipeline; and flushing the pipeline based on the determined pressure variation. The method comprises: Divide the target pipeline into several sub-pipelines according to length, pipe diameter change or risk area of blockage, wherein both ends of each sub-pipeline are provided with pressure sensors to collect pressure data in the pipeline in real time, and one end of each sub-pipeline is provided with a nozzle with an electromagnetic valve to open or close according to instructions and adjust water jet pressure and flow; Using a graph neural network model based on deep learning, input the pressure data of each section of the pipeline, and learn the pressure transmission relationship and blockage influence law between each section of the pipeline through training; Real-time receive pressure data collected by pressure sensors, call the trained graph neural network model for analysis, generate nozzle flushing control instructions, and send them to the corresponding nozzle.

2. The method of claim 1, wherein, The step of using a graph neural network model based on deep learning, inputting the pressure data of each section of the pipeline, and learning the pressure transmission relationship and blockage influence law between each section of the pipeline through training comprises: During normal operation of the pipeline, continuously collect pressure sensor data of each section of the pipeline, record pressure changes under different working conditions, and use them as normal state samples; Set different degrees of blockage in different sub-pipelines, collect pressure data of each section of the pipeline when blockage occurs, and construct abnormal state samples; Preprocess the collected data, including data cleaning and normalization, to obtain preprocessed data; According to the graph neural network, each sub-pipeline is regarded as a node in the graph, and the connection relationship between the pipelines is regarded as an edge, and the mutual influence of the pressure changes of each section of the pipeline is learned through information transmission between nodes; To predict the blockage degree of each section of the pipeline, use mean square error as the loss function, and use the stochastic gradient descent algorithm to train the model to obtain the trained graph neural network model.

3. The method of claim 2, wherein, The step of real-time receiving pressure data collected by pressure sensors, calling the trained graph neural network model for analysis, generating nozzle flushing control instructions, and sending them to the corresponding nozzle comprises: Real-time receive pressure data collected by pressure sensors, and input the trained graph neural network model to obtain the prediction result; According to the prediction result, determine the abnormal position and the influence range, and according to the abnormal position and the influence range, combine the pressure change of the adjacent pipeline to determine the required flushing pressure and flushing time of the nozzle; According to the flushing pressure and the flushing time, control the corresponding nozzle to flush the pipeline, wherein during the flushing process, the pressure change is continuously monitored, if the pressure returns to normal, the flushing is stopped, if the pressure does not return to normal, the flushing intensity is increased or the adjacent pipeline nozzle is started to assist flushing.

4. The method of claim 3, wherein, In the step of real-time receiving pressure data collected by pressure sensors, calling the trained graph neural network model for analysis, generating nozzle flushing control instructions, and sending them to the corresponding nozzle, a fuzzy logic model is combined with the graph neural network model, the graph neural network model is used to predict the pipeline blockage state, and the fuzzy logic is used to generate flushing control instructions based on the prediction result and real-time pressure change.

5. The method of claim 4, wherein, In the fuzzy logic model, the rate of change of the pressure of each section of the pipeline and the pressure difference value are taken as the input of the fuzzy logic model; the nozzle flushing pressure and the flushing time are taken as the output of the fuzzy logic model; the fuzzy rules of the fuzzy logic model are formulated based on actual experience and expert knowledge; the fuzzy output result is obtained through fuzzy reasoning on the input variables according to the fuzzy rules, and the fuzzy output result is converted into specific flushing pressure and flushing time through defuzzification to guide the nozzle to perform the flushing operation.

6. The method of claim 2, wherein, The graph neural network model adopts a graph convolution network as a basic architecture, wherein 3-5 layers of graph convolution network layers are arranged, each layer of graph convolution network layer updates its own features by aggregating the information of adjacent nodes; a ReLU activation function is used after each convolution network layer; the number of output layer neurons is the same as the number of pipeline sections, and each neuron corresponds to a predicted value of the blocking degree of a section of pipeline.

7. The pressure change based intelligent flushing control method for pipes as claimed in claim 2 wherein, In the step of taking the prediction of the blocking degree of each section of pipeline as the target, using mean square error as the loss function, and using the stochastic gradient descent algorithm to train the model to obtain the trained graph neural network model, the loss function formula is: ； where n is the number of samples, is the actual congestion level label, is the model predicted congestion level.

8. A pressure change based intelligent flushing control system for a pipe system, characterized in that, The system is used to implement the intelligent pipeline flushing control method based on pressure change according to any one of claims 1-7. The division module is configured to divide the target pipeline into a plurality of sub-pipelines according to length, pipe diameter change or risk area prone to blocking, wherein pressure sensors are arranged at both ends of each sub-pipeline to collect pressure data in the pipeline in real time, and a nozzle with an electromagnetic valve is arranged at one end of each sub-pipeline to open or close according to an instruction and adjust water pressure and flow rate. The training module is configured to use a graph neural network model based on deep learning to input the pressure data of each section of pipeline, and learn the pressure transmission relationship and blocking influence law between each section of pipeline through training. The analysis module is configured to receive the pressure data collected by the pressure sensor in real time, call the trained graph neural network model for analysis, generate a nozzle flushing control instruction, and send the nozzle flushing control instruction to the corresponding nozzle.

9. A computer-readable storage medium, characterized in that, The readable storage medium stores one or more programs, which are executed by the processor to implement the intelligent pipeline flushing control method based on pressure change according to any one of claims 1-7. The electronic device includes a memory and a processor.

10. An electronic device, comprising: The memory is used to store computer programs. The processor is used to execute the computer programs stored in the memory to implement the intelligent pipeline flushing control method based on pressure change according to any one of claims 1-7. ​

Images