A real-time detection system for urban gas pipe network

Abstract

The utility model relates to a kind of urban gas pipe network real-time detection system, it is characterized by including sensor layer, storage node layer, computing node layer and network control platform;Sensor layer, storage node layer, computing node layer and network control platform are connected by network, and computing node layer includes multiple computing nodes and controller.

Description

Technical Field

[0001] This invention relates to the field of detection, specifically to a real-time detection system for urban gas pipeline networks. Background Technology

[0002] The urban gas pipeline real-time monitoring system is an intelligent monitoring system composed of sensors, computing nodes, and an internet management platform. Utilizing internet technology, it monitors data such as pressure, flow rate, confined space conditions, and gas leaks, recording and analyzing the pipeline network's operational status in real time, ensuring the normal operation of the urban pipeline network. Furthermore, through the cooperation of sensors and computing nodes, most of the detection and calculation work can be performed remotely, eliminating the need for frequent data transfer between the network management platform and remote sensors, thus alleviating network resource constraints to some extent.

[0003] However, existing technologies still have the following problems: Since the computing node layer closest to the sensor undertakes the main computing tasks, it needs to read data from the sensor's associated memory and perform corresponding calculations. Therefore, when the network scale is large, the number of computing tasks that need to be executed in real time is very high, causing congestion in data transmission between the memory and computing nodes. On the one hand, multiple computing nodes simultaneously reading data from a single storage node results in low efficiency due to network congestion, wasting the computing resources of the computing nodes. On the other hand, frequent connections from computing nodes with weaker data caching capabilities or computing power to storage nodes can also easily cause excessive bandwidth diversion from the storage node's output, reducing the efficiency of the storage node's data output.

[0004] Therefore, there is a need to provide a real-time monitoring system for urban gas pipeline networks that can reasonably configure memory and computing node resources to maximize the data transmission and computing efficiency of the real-time monitoring system for urban gas pipeline networks. Summary of the Invention

[0005] The technical problem this invention aims to solve is the following shortcomings of existing technologies: Since the computing node layer closest to the sensor undertakes the main computing tasks, it needs to read data from the sensor's associated memory and perform corresponding calculations. Therefore, when the network scale is large, the number of computing tasks requiring real-time execution is very high, causing congestion in data transmission between the memory and computing nodes. On the one hand, multiple computing nodes simultaneously reading data from a single storage node results in low data reading efficiency due to network congestion, wasting the computing resources of the computing nodes. On the other hand, frequent connections by computing nodes with weaker data caching capabilities or computing power to the storage node can also easily cause excessive bandwidth diversion from the storage node's output, reducing the efficiency of the storage node's data output.

[0006] The technical solution adopted by this invention to solve its technical problem is:

[0007] A real-time monitoring system for urban gas pipeline networks includes a sensor layer, a storage node layer, a computing node layer, and a network control platform. The sensor layer, storage node layer, computing node layer, and network control platform are connected through a network. The computing node layer includes multiple computing nodes and a controller.

[0008] Specifically, multiple computing nodes are connected to the storage node layer and the controller to read data from the storage nodes according to the controller's instructions, perform calculations, and finally feed the calculation results back to the controller.

[0009] Specifically, the sensor layer includes a variety of measurement sensors.

[0010] Specifically, the measurement sensors include: flow measurement instruments, temperature measurement instruments, and pressure measurement instruments.

[0011] Specifically, the storage node layer includes multiple storage nodes for storing real-time data detected by multiple sensors in the sensor layer; the storage node layer is also connected to the computing node layer.

[0012] Specifically, the controller connects multiple computing nodes and multiple storage nodes in the storage node layer, and is used to distribute computing tasks to the computing nodes according to the instructions of the network control platform.

[0013] Specifically, the controller is also used to detect the network status and computing capabilities of storage nodes and compute nodes.

[0014] Specifically, the management of computing nodes is achieved based on the network status and computing capability status of the aforementioned storage and computing nodes.

[0015] Specifically, the network control platform connects to the controller of the computing node layer, and issues computing tasks to the computing node layer according to the overall control needs.

[0016] A monitoring method based on the real-time monitoring system of the urban gas pipeline network, which manages the computing nodes through a controller.

[0017] The beneficial effect of the real-time monitoring system for urban gas pipeline networks provided in this application is that the reasonable configuration of memory and computing node resources maximizes the data transmission and computing efficiency of the real-time monitoring system for urban gas pipeline networks. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the structure of the real-time monitoring system for urban gas pipeline networks provided in this application. Detailed Implementation

[0019] The present invention will now be described in more detail with reference to the accompanying drawings, which illustrate preferred embodiments of the invention. It should be understood that those skilled in the art can modify the invention described herein while still achieving the beneficial effects of the invention. Therefore, the following description should be understood as being of broad knowledge to those skilled in the art and is not intended to limit the invention.

[0020] For clarity, not all features of the actual embodiments are described. In the following description, well-known functions and structures are not detailed in detail, as they would confuse the invention with unnecessary detail. It should be understood that in the development of any actual embodiment, numerous implementation details must be made to achieve the developer's specific objectives.

[0021] To make the objectives and features of the present invention more apparent and understandable, the specific embodiments of the present invention will be further described below with reference to the accompanying drawings. It should be noted that the drawings are all in a very simplified form and use non-precise ratios, intended only to facilitate and clearly illustrate the objectives of the embodiments of the present invention.

[0022] The real-time monitoring system for urban gas pipeline networks provided in this application includes a sensor layer, a storage node layer, a computing node layer, and a network control platform. The sensor layer, storage node layer, computing node layer, and network control platform are connected via a network.

[0023] The sensor layer includes various measurement sensors, such as flow meters, temperature meters, and pressure meters. The real-time data measured by the sensor layer is transmitted to the storage node layer. The storage node layer includes multiple storage nodes for storing the real-time data detected by the various sensors in the sensor layer. The storage node layer is also connected to the computing node layer.

[0024] The computing node layer includes a controller and multiple computing nodes. These computing nodes are connected to the storage node layer and the controller, and are used to read data from the storage nodes according to instructions from the controller, perform calculations, and finally feed the calculation results back to the controller.

[0025] The controller in the compute node layer is connected to the network control platform to receive compute tasks. The controller connects multiple compute nodes and multiple storage nodes in the storage node layer, distributing compute tasks to the compute nodes according to instructions from the network control platform. The controller also detects the network status and computing capabilities of the storage and compute nodes, and determines the combined state of multiple compute nodes during task execution based on these statuses.

[0026] The network status includes one or more of the following network parameters, which include: congestion control window value for data transmission of at least one of multiple storage nodes, reception window value for data reception of at least one of multiple computing nodes, congestion control window value for data transmission, data caching speed, data transmission efficiency between storage nodes and computing nodes or a pair of computing nodes, maximum bandwidth, computing power of each computing node, etc.

[0027] The aforementioned combined states include: centralized combined state, distributed combined state, or hybrid combined state. In the centralized combined state configuration, after the storage node transmits the data required for computation to a central computing node, the central computing node caches the data and, based on the divisibility of the overall computation task, divides the overall computation task into multiple sub-computation tasks. Based on these sub-computation tasks, the data is further divided into at least one data block, and different data blocks are sent to other computing nodes connected to the central computing node. Upon receiving the data blocks, the other computing nodes execute the corresponding sub-computations. The central computing node may execute some sub-computations or not. In the distributed combined state configuration, based on the divisibility of the overall computation task, the overall computation task is divided into multiple sub-computation tasks. Based on these sub-computation tasks, the data is divided into at least one data block. The storage node sends different data blocks to computing nodes that execute the corresponding sub-computation tasks. Upon receiving the data blocks, the multiple computing nodes execute their respective sub-computation tasks. The hybrid combinatorial state configuration is a combination of centralized and distributed combinatorial states. That is, for a whole computational task, based on the divisibility of the whole computational task, the whole computational task is divided into multiple sub-computational tasks. Some sub-tasks are computed through centralized combinatorial states, while other sub-tasks are computed through distributed combinatorial states.

[0028] The network control platform is connected to the controller of the computing node layer, and it issues computing tasks to the computing node layer according to the overall control needs.

[0029] Based on the aforementioned real-time monitoring system for urban gas pipeline networks, this application also provides a method for real-time monitoring of urban gas pipeline networks. Specifically, it includes the following steps:

[0030] The first step is for the sensor layer to collect real-time data and store it in the corresponding computing node of the memory layer.

[0031] Specifically, the aforementioned real-time data includes flow rate, pressure, temperature, design parameters, optical parameters, etc. The real-time data collected by the sensor layer is transmitted and stored in real time on the corresponding computing nodes in the memory layer.

[0032] Preferably, data from related sensors can be stored in the same storage node. For example, sensor data located on the same branch, detection data of different attributes at the same end, or sensor data downstream of the same gas station can be stored in the same storage node.

[0033] The second step involves the network control platform initiating at least one computing task based on specific triggering conditions or computing rules, and then distributing the computing task to the controller at the computing node layer.

[0034] Specifically, the triggering condition refers to the computation that the network control platform needs to initiate when the aforementioned triggering conditions are met. For example, when the real-time monitoring value of a specific parameter or the intermediate value obtained through real-time detection calculation exceeds a preset threshold or meets other preset conditions, the network control platform initiates a calculation for a certain security performance and sends the computer program for the above calculation task to the controller of the computing node layer. Or, more preferably, it calls the computer program pre-stored inside the controller of the computing node layer by sending an instruction, thereby further reducing the total amount of data transmission.

[0035] The third step is for the computing node layer controller to perform a logical minimum partitioning of the received computing task.

[0036] Specifically, after receiving the computation program of the computation task, the computing node layer controller divides the received computation program into the logically smallest subtasks according to the divisibility of the program logic. The logically smallest subtask refers to the sub-computation task that cannot be further divided in the program logic, i.e., T = {T1, ..., Tj}, j ≥ 1.

[0037] Furthermore, the computing node layer controller queries the amount of data required to run the computing program of each of the logically smallest subtasks, D = {D1, ..., Dj}, j ≥ 1, based on the computing program of the logically smallest subtask mentioned above.

[0038] The fourth step involves the compute node layer controller determining the combined state of multiple compute nodes during the execution of a compute task based on the compute task and its corresponding required data, as well as the network status and compute capability status of the storage nodes and compute nodes mentioned above.

[0039] The fourth step specifically includes the following steps:

[0040] Step 4.1, using the formula D = Min(D) EX1 +……+D EXi The controller divides the computation task program into at least one execution task block T. EX1 , ..., T EXi The execution task block T EX1 , ..., TEXi The corresponding data are: D EX1 , ..., D EXi , where D EX1 ={D+……+D a1}, D EX2 ={D a1+1 +……+D a1+a2},……,D EXi =={D a(i-1)+1 +……+D a1+a2+……+ai}, 1≤i≤j, a1+a2+……+a i =j, a1, a2, ..., a i Each represents at least one execution task block T EX1 , ..., T EXi The number of the smallest subtasks contained in each.

[0041] The data corresponding to the smallest subtasks obtained through logical partitioning often contains some overlap. Therefore, before partitioning the tasks, the smallest subtasks are reorganized by setting the minimum total data volume as a condition, thereby obtaining a partitioning method that minimizes the amount of data to be transmitted for the execution task blocks. Preferably, due to the correlation of the program, only the number of clusters of the smallest subtasks is adjusted, without significantly changing their order.

[0042] Step 4.2, determine the primary compute node, specifically including:

[0043] 1) Establish configuration equations, that is, establish the target configuration with the shortest total data throughput time and the shortest congestion window detection time.

[0044]

[0045]

[0046] Where m represents the node number, m = 1 to i, T t T represents the total transmission time. m-delay D represents exm The delay time T between the storage node storing the data and sending it. cm-delay D represents the delay time between each computing node receiving information and sending it back to the storage node. EXm Wm represents the data size of the execution task block for this compute node. exm The available bandwidth from the storage node storing the data to the compute node, rwnd m B represents the size of the sliding window for receiving information by the computing node. m T represents the cache speed of the compute node. cwnd Indicates the congestion window detection time. ssthresh i represents the historical average value of the congestion window of this computing node in bytes, and ssthresh i represents the initial value of the slow start threshold.

[0047] 2) Substitute the above data from all computing nodes into the configuration equation and solve the configuration equation using a genetic algorithm. The genetic algorithm specifically includes: randomly generating a population as the initial solution; finding a suitable encoding scheme to encode individuals in the population, such as floating-point encoding or binary encoding; using the function value of the multimodal function as the fitness of the individual, calculating the fitness of each individual in the population; selecting parent and parent individuals to participate in reproduction based on their fitness, with the principle that individuals with higher fitness are more likely to be selected; performing genetic operations on the selected parent and parent individuals, i.e., copying the genes of the parent and parent, and using crossover, mutation, and other operators to generate offspring; Step 6: determining, based on certain criteria, whether to continue executing the algorithm or find the individual with the highest fitness among all offspring as the solution and return to end the program, using the solved computing node as the master computing node, and using the execution task block corresponding to the calculated master computing node as the execution task block corresponding to that master node.

[0048] Step 4.3 Selecting subordinate compute nodes specifically includes:

[0049] 1) Determine the minimum number of subtasks x contained in the execution task block corresponding to each master node. If x = 1, determine that the master node is in distributed composite state, that is, the master computing node does not need to set up slave computing nodes. If x is greater than 1, proceed to step 2).

[0050] 2) Judgment Is it less than Where i represents the number of the computing node, T i-delay T represents the delay time for the computing node to send information. ci-delay D represents the delay time for the computing node to receive information. EXi W represents the data size of the task block executed by this computing node. i This represents the average available bandwidth from the master compute node to other compute nodes. rwndi B represents the size of the sliding window for the information sent by the computing node. i This represents the cache speed of the compute node, where k is a constant, taking the value of an integer between 2 and 5, and P... i The parameter represents the theoretical computing power of the computing node. If not, it is determined that the master node uses a distributed composite state, that is, the master computing node does not need to set up slave computing nodes. If not, it is determined that the master node uses a centralized composite state and step 3 is executed.

[0051] 3) Determine the minimum number of subtasks x in the execution task block allocated on the master computing node A. Based on the minimum number of subtasks x, determine the number of subordinate nodes and select from other available nodes (nodes that are not currently working or have been partitioned into tasks). The x largest computational nodes are designated as slave nodes of the master computational node A. Where P... i The parameters representing the theoretical computing capabilities of the candidate computing nodes are: Wi, which is the available bandwidth between the candidate computing nodes and the master computing node; and Bi, which is the cache speed of the candidate computing nodes. In other words, node A and x slave nodes form a centralized combination to execute computing tasks.

[0052] Fifth, the controller sends the task blocks to the corresponding master computing nodes according to the combination method of the divided node execution tasks, and sends the combination method to the master computing nodes. The master computing nodes then perform further calculations and / or allocate data according to the combination method.

[0053] The beneficial effect of the real-time monitoring system for urban gas pipeline networks provided in this application is that the reasonable configuration of memory and computing node resources maximizes the data transmission and computing efficiency of the real-time monitoring system for urban gas pipeline networks.

[0054] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for real-time detection of urban gas pipeline networks, characterized in that: The method includes the following steps: The first step is to collect real-time data and store it in the corresponding computing nodes of the memory layer; The second step is to start at least one computing task and send the computing task to the controller of the computing node layer. The third step is to perform a logical minimum partitioning of the received computing task. The fourth step involves determining the combination state of multiple computing nodes during the execution of the computing task based on the computing task and its corresponding data requirements, as well as the network status and computing capability status of the storage and computing nodes. These combination states include: centralized combination state, distributed combination state, or hybrid combination state. In the centralized combination state, the storage node transmits the data required for computing to a central computing node. This central computing node then caches the data, divides the overall computing task into multiple sub-computing tasks, and further divides the data into at least one data block based on these sub-computing tasks. Different data blocks are then sent to other computing nodes connected to this central computing node, and the other computing nodes execute the corresponding sub-computations upon receiving the data blocks. In the distributed combination state, the overall computing task is divided into multiple sub-computing tasks, and the data is divided into at least one data block based on these sub-computing tasks. The storage node sends different data blocks to the computing nodes that execute the sub-computing tasks corresponding to the data blocks. Multiple computing nodes then execute their respective sub-computing tasks upon receiving the data blocks. The hybrid combination state is a combination of centralized and distributed combination states. Specifically, the fourth step includes the following steps: Step 4.1, using the formula D = Min(D) EX1 +……+D EXi The computation task program is divided into at least one logically partitioned minimum subtask; wherein the controller divides the computation task program into at least one execution task block T. EX1 , ..., T EXi Execute task block T EX1 , ..., T EXi The corresponding data are: D EX1 , ..., D EXi ; Step 4.2, determine the primary compute node, specifically including: 1) Establish the target configuration equations that minimize the total data throughput time and the congestion window detection time; 2) Input the above data of all computing nodes into the configuration equation and solve the configuration equation using a genetic algorithm; take the solved computing node as the master computing node, and take the execution task block corresponding to the master computing node as the execution task block corresponding to the master computing node. Step 4.3: Select subordinate computing nodes and determine the combined state of multiple computing nodes; Fifth, the controller sends the task blocks to the corresponding master computing nodes based on the combined state of the tasks executed by the nodes, and sends the combined state to the master computing nodes. The master computing nodes then perform further calculations and / or allocate data according to the combination method.

2. The real-time detection method for urban gas pipeline networks according to claim 1, characterized in that: The real-time data mentioned above includes flow rate, pressure, temperature, design parameters, and optical parameters.

3. The real-time detection method for urban gas pipeline networks according to claim 1, characterized in that: In the third step, after receiving the computation program of the computation task, the computation node layer controller divides the received computation program of the computation task into the logically smallest subtasks according to the divisibility of the program logic. The logically smallest subtask refers to the sub-computation task that cannot be further divided in the program logic, that is, T={T1,……,Tj}, j≥1.

4. The real-time detection method for urban gas pipeline networks according to claim 3, characterized in that: The computing node layer controller queries the amount of data required to run the computing program of each of the logically smallest subtasks, D = {D1, ..., Dj}, j ≥ 1, based on the computing program of the logically smallest subtask mentioned above.

5. The real-time detection method for urban gas pipeline networks according to claim 1, characterized in that: Step 4.1 also includes reorganizing the smallest subtasks by setting the minimum total data volume as the condition.

6. The method for real-time detection of urban gas pipeline networks according to claim 1, characterized in that: Step 4.1 also includes adjusting only the number of clusters of the smallest subtasks.

7. The method for real-time detection of urban gas pipeline networks according to claim 1, characterized in that: In step 4.2, establishing the configuration equations includes: establishing the target configuration that minimizes the total data throughput time and the congestion window detection time.

8. The real-time detection method for urban gas pipeline networks according to claim 1, characterized in that: In step 4.2, the genetic algorithm specifically includes: randomly generating a population as the initial solution.

9. The method for real-time detection of urban gas pipeline networks according to claim 1, characterized in that: In step 4.2, the genetic algorithm specifically includes: finding the individual with the highest fitness among all offspring as the solution and returning it to end the program.

10. A real-time detection system for urban gas pipeline networks, comprising a sensor layer, a storage node layer, a computing node layer, and a network control platform; the sensor layer, storage node layer, computing node layer, and network control platform are connected via a network, the computing node layer includes multiple computing nodes and a controller, and the real-time detection system for urban gas pipeline networks is used to execute the real-time detection method for urban gas pipeline networks according to any one of claims 1-9.

Images