An intelligent analysis and diagnosis system for gas metering instrument caliber matching

Abstract

The present invention provides an intelligent analysis and diagnosis system for caliber matching of gas metering instruments, relating to the field of data processing technology. The system comprises: an extraction module for selecting, based on pipeline pressure gradient data, flow phase data, and user gas consumption time-series fluctuation data, a main pipeline intersection point P1, a user access point P2, and a pressure regulating station output point P3 as three dynamic topology reference points; a dynamic topology unit constructed based on these three points, which is then divided into regions and the pressure-flow coupling eigenvalues ​​of the subregions are extracted; a correction module for weightedly correcting the user gas demand mutation probability eigenvalue using the pressure-flow coupling eigenvalue to generate a corrected mutation probability eigenvalue; and a processing module for performing tensor concatenation of the corrected mutation probability eigenvalue with the pipeline equivalent inner diameter dynamic eigenvalue, which is then input into a fully connected diagnostic network to generate a caliber matching deviation index. This invention can improve the intelligent operation efficiency of gas supply.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to an intelligent analysis and diagnosis system for caliber matching of gas metering instruments. Background Art

[0002] A large commercial complex encompasses a variety of gas-using scenarios, including shopping malls, hotels, and restaurants. Gas demand is complex and fluctuates significantly. During initial construction, the caliber of gas meters was configured based on conventional estimates and experience. As operations progressed, the number of restaurants within the mall increased, and their operating hours varied. This led to insufficient gas supply during peak hours, forcing some businesses' gas equipment to operate at full capacity, impacting normal operations.

[0003] During off-peak hours, oversized metering instruments reduce metering accuracy, leading to inaccurate gas metering and frequent disputes between gas companies and merchants over gas volume settlements. Furthermore, long-term inappropriate gas supply has led to frequent malfunctions of some gas equipment due to unstable pressure, increasing maintenance costs. This highlights the shortcomings of some traditional gas metering caliber matching methods when dealing with complex gas usage scenarios. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide an intelligent analysis and diagnosis system for caliber matching of gas metering instruments, which can improve the intelligent operation efficiency of the gas supply system.

[0005] In order to solve the above technical problems, the technical solutions of the present invention are as follows:

[0006] In a first aspect, an intelligent analysis and diagnosis system for gas metering instrument caliber matching includes:

[0007] The acquisition module is used to collect pipeline pressure gradient data, flow phase data, and user gas consumption time series fluctuation data in real time through pressure sensors deployed at the three-dimensional spatial coordinate positions of the gas pipeline network trunk nodes, user access points, and pressure regulating stations;

[0008] An extraction module is configured to select the main pipeline intersection point P1, the user access point P2, and the pressure regulating station output point P3 as three dynamic topology reference points based on the pipeline pressure gradient data, flow phase data, and user gas consumption time series fluctuation data; construct a dynamic topology unit based on the three points, divide the unit into regions based on the Delaunay subdivision algorithm, and extract the pressure-flow coupling characteristic values ​​of the subregions;

[0009] a correction module, configured to perform weighted correction on the user gas demand mutation probability characteristic vector using the pressure-flow coupling characteristic value to generate a corrected mutation probability characteristic vector;

[0010] The processing module is used to perform tensor splicing on the corrected mutation probability feature vector and the dynamic feature vector of the pipeline equivalent inner diameter, and input the result into a fully connected diagnostic network to generate a caliber matching deviation index; when the deviation exceeds a set threshold, the linkage timing prediction network outputs a safe redundant caliber replacement plan.

[0011] In a second aspect, a method for intelligent analysis and diagnosis of gas metering instrument caliber matching is provided, the method comprising:

[0012] Step 1: Use pressure sensors deployed at the three-dimensional coordinate locations of the gas pipeline network trunk nodes, user access points, and pressure regulating stations to collect pipeline pressure gradient data, flow phase data, and user gas consumption time series fluctuation data in real time;

[0013] Step 2: Based on pipeline pressure gradient data, flow phase data, and user gas consumption time series fluctuation data, the main pipeline intersection point P1, user access point P2, and pressure regulating station output point P3 are selected as three vertices, and a dynamic triangle is formed based on the three vertices. The dynamic triangle is meshed using the Delaunay triangulation algorithm, and the pressure-flow coupling eigenvalue of the mesh is extracted.

[0014] Step 3: Perform weighted correction on the user gas demand mutation probability characteristic vector using the pressure-flow coupling characteristic value to obtain a corrected mutation probability characteristic vector;

[0015] In step 4, the corrected mutation probability feature vector is tensor-concatenated with the pipeline equivalent inner diameter dynamic feature vector and input into the fully connected diagnostic network to generate a caliber matching deviation index. When the deviation exceeds the threshold, the linked timing prediction network outputs a safe and redundant caliber replacement plan.

[0016] According to a third aspect, a computing device includes:

[0017] one or more processors;

[0018] The storage device is used to store one or more computer programs. When the one or more computer programs are executed by the one or more processors, the one or more processors implement the system.

[0019] In a fourth aspect, a computer-readable storage medium stores a computer program, which implements the system when executed by a processor.

[0020] The above solution of the present invention includes at least the following beneficial effects.

[0021] The acquisition module deploys pressure sensors at the main nodes of the gas pipeline network, user access points and pressure regulating stations to obtain pipeline pressure gradient data, flow phase data and user gas consumption timing fluctuation data in real time. Compared with the traditional method that relies on manual or static monitoring, the system can fully capture dynamic information during the operation of the pipeline network and avoid caliber matching deviations caused by data missing or lag.

[0022] Based on the collected multi-dimensional data, the extraction module innovatively selects three dynamic topological reference points to construct a dynamic topological unit, and uses the Delaunay decomposition algorithm to divide the area and extract the pressure-flow coupling characteristic values ​​of the sub-area. This method can effectively integrate the spatial topological structure of the pipeline network with the fluid mechanics characteristics, and convert the complex pipeline network operation status into quantifiable characteristic parameters, greatly improving the ability to analyze the operation laws of the pipeline network.

[0023] The correction module uses the pressure-flow coupling characteristic value to perform weighted correction on the characteristic vector of the probability of sudden changes in user gas demand, fully considering the impact of the pipeline network operation status on the user's gas consumption behavior. This dynamic correction mechanism can accurately capture sudden changes in user gas demand. Compared with the traditional fixed parameter prediction model, it can better fit the actual gas consumption scenario and avoid caliber matching errors caused by demand estimation deviations.

[0024] The processing module performs tensor splicing on the corrected mutation probability feature vector and the dynamic feature vector of the pipeline equivalent inner diameter, generates a caliber matching deviation index through a fully connected diagnostic network, and links the timing prediction network to output a safe and redundant caliber replacement plan when the deviation exceeds the threshold. This process realizes fully automated intelligent diagnosis from data input to decision output, which not only greatly improves the efficiency of caliber matching diagnosis, but also provides a reasonable caliber replacement plan based on scientific algorithms, effectively reducing pressure loss during gas transportation, improving metering accuracy, and reducing equipment wear and energy waste. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a schematic diagram of an intelligent analysis and diagnosis system for gas metering instrument caliber matching provided by an embodiment of the present invention.

[0026] Figure 2 The present invention provides a flowchart of an intelligent analysis and diagnosis method for gas metering instrument caliber matching. DETAILED DESCRIPTION

[0027] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0028] like Figure 1 As shown, an embodiment of the present invention provides an intelligent analysis and diagnosis system for gas metering instrument caliber matching, comprising:

[0029] The acquisition module is used to collect pipeline pressure gradient data, flow phase data, and user gas consumption time series fluctuation data in real time through pressure sensors deployed at the three-dimensional spatial coordinate positions of the gas pipeline network trunk nodes, user access points, and pressure regulating stations;

[0030] An extraction module is configured to select the main pipeline intersection point P1, the user access point P2, and the pressure regulating station output point P3 as three dynamic topology reference points based on the pipeline pressure gradient data, flow phase data, and user gas consumption time series fluctuation data; construct a dynamic topology unit based on the three points, divide the unit into regions based on the Delaunay subdivision algorithm, and extract the pressure-flow coupling characteristic values ​​of the subregions;

[0031] a correction module, configured to perform weighted correction on the user gas demand mutation probability characteristic vector using the pressure-flow coupling characteristic value to generate a corrected mutation probability characteristic vector;

[0032] The processing module is used to perform tensor splicing on the corrected mutation probability feature vector and the dynamic feature vector of the pipeline equivalent inner diameter, and input the result into a fully connected diagnostic network to generate a caliber matching deviation index; when the deviation exceeds a set threshold, the linkage timing prediction network outputs a safe redundant caliber replacement plan.

[0033] In an embodiment of the present invention, the acquisition module can obtain pipeline pressure gradient, flow phase and user gas consumption time series fluctuation data in real time by deploying pressure sensors at key locations of the gas pipeline network; compared with traditional manual monitoring or intermittent data acquisition methods, this system can realize all-round and dynamic monitoring of the pipeline network operation status, effectively avoiding caliber matching errors caused by data lag or loss.

[0034] The extraction module constructs topological units based on dynamic topological reference points and divides regions using the Delaunay decomposition algorithm to accurately extract pressure-flow coupling eigenvalues. This fully considers the dynamic relationship between the spatial topology of the pipeline network and fluid mechanics, enabling in-depth exploration of pipeline network operation patterns and transforming complex pipeline network operating conditions into quantifiable and analyzable characteristic parameters. The correction module uses pressure-flow coupling eigenvalues ​​to perform weighted corrections on the characteristic vectors of the probability of sudden changes in user gas demand, fully considering the impact of the pipeline network's operating status on user gas consumption behavior. This mechanism can dynamically adjust gas demand forecasts in real time, accurately capturing sudden changes in user gas demand. Compared to traditional fixed-parameter prediction models, it can better fit actual gas consumption scenarios and effectively avoid caliber matching errors caused by demand estimation deviations. The processing module uses tensor splicing and a fully connected diagnostic network to automatically generate caliber matching deviation indicators, and when the deviation exceeds the threshold, it links the timing prediction network to output a safe and redundant caliber replacement plan. This process is completely based on intelligent algorithms and models, and realizes full process automation from data processing to decision output, which improves diagnostic efficiency, effectively reduces gas transmission pressure loss, improves metering accuracy, reduces equipment wear and energy waste, and ensures the safe and stable operation of the gas supply system.

[0035] In a preferred embodiment of the present invention, pressure sensors are deployed at the three-dimensional coordinate positions of the gas pipeline network trunk nodes, user access points, and pressure regulating stations to collect pipeline pressure gradient data, flow phase data, and user gas consumption time series fluctuation data in real time, including:

[0036] Establish the pipe network topology mapping relationship based on the three-dimensional spatial coordinates of each sensor;

[0037] Synchronously collect real-time pressure values ​​along the pipeline between trunk nodes, calculate the pressure difference and associated spatial distance between adjacent nodes, and generate pressure gradient data;

[0038] Monitor the flow sensor output at the user access point, extract the phase offset and amplitude characteristics of the flow waveform within the cycle, and generate flow phase data;

[0039] The instantaneous fluctuation signal of gas consumption downstream of the pressure regulating station is recorded, and the fluctuation frequency and the absolute difference between the peak and the trough in unit time are calculated to form the time series fluctuation data of gas consumption of users.

[0040] In the embodiment of the present invention, the above steps, when applied specifically, can be implemented by the following steps, for example:

[0041] The three-dimensional spatial coordinates of pressure sensors deployed at the main nodes, user access points, and pressure regulating stations of the gas pipeline network are obtained. Each sensor has corresponding X, Y, and Z coordinate values, which reflect the sensor's spatial position in the actual pipeline network. Then, based on the actual layout and connection relationships of the gas pipeline network, sensors with connected relationships are associated. For example, if two main nodes are connected by a pipeline, the sensors corresponding to these two nodes are spatially marked as connected. For user access points and pressure regulating stations, their connection relationships with the main node sensors are also determined based on their actual connection locations to the pipeline network. Ultimately, a topological mapping relationship of the entire gas pipeline network is constructed, which is like drawing a pipeline network map with detailed connection information for each node.

[0042] By establishing a pipeline network topology mapping relationship, the spatial structure and node connection status of the gas pipeline network can be clearly presented. When calculating and analyzing data such as pressure and flow, the topological relationship can be combined to more accurately determine the data transmission path and mutual influence, thereby improving the accuracy and reliability of data processing and helping to fully understand the overall operation status of the pipeline network.

[0043] The pressure sensors distributed at the trunk nodes simultaneously collect data to obtain the real-time pressure value of each trunk node at the same moment. Next, for every two adjacent trunk nodes, the pressure difference between the two adjacent nodes is calculated by subtracting the pressure of the node with the smaller pressure value from the pressure of the node with the larger pressure value. The actual distance between the two adjacent nodes in three-dimensional space is then measured, and the pressure difference is associated with the corresponding spatial distance. For example, if node A and node B are adjacent, the pressure difference between them is 5 units, and the spatial distance is 10 meters, then this combination of pressure difference and spatial distance is used as a set of data. This process is repeated to calculate the pressure difference and spatial distance between all adjacent trunk nodes. By integrating this data, pressure gradient data reflecting the pressure change trend of the pipeline network is generated.

[0044] The pressure gradient data of the present invention can intuitively reflect the pressure changes of gas when it flows in the pipeline. By analyzing the pressure gradient data, it is possible to promptly detect whether there are abnormal pressure areas in the pipeline. For example, a sudden drop in pressure may mean pipeline leakage or blockage. This helps gas companies quickly locate fault points and take corresponding maintenance measures to ensure the safety and stability of gas transportation.

[0045] Continuously monitor the flow data output by the flow sensor at the user access point. Over time, this data will form a continuously changing flow waveform. Select a fixed time period and observe the changes in the flow waveform within this period. For the phase offset, compare it with the standard flow waveform under normal conditions to see whether the current flow waveform has shifted to the left or right on the time axis, and measure the duration of the shift. For the amplitude characteristic, find the maximum and minimum values ​​of the flow waveform within this period and calculate the difference between them. This difference is the range of change of the flow amplitude. Record the two sets of data, the phase offset and the amplitude characteristic. Repeat the above operation multiple times for each user access point to obtain data within multiple periods. Aggregate this data to generate the flow phase data.

[0046] Flow phase data accurately reflects the changing patterns and characteristics of gas usage. Phase offsets can reveal changes in gas usage timing, such as whether peak demand arrives earlier or later. Amplitude characteristics reflect the fluctuations in gas usage. By analyzing this data, gas companies can better understand user gas usage habits, rationally allocate gas resources, and improve the efficiency and stability of gas supply. This data also helps predict future gas demand.

[0047] Gas consumption data downstream of the pressure regulating station is recorded in real time. This data will fluctuate instantaneously over time, forming a series of fluctuation signals. A fixed time unit, such as one minute, is selected. The number of gas consumption fluctuations within this minute is counted. This number is the fluctuation frequency per unit time. Next, the maximum (peak) and minimum (trough) gas consumption values ​​within this time unit are identified, and the absolute difference between them is calculated. For example, if the maximum gas consumption within one minute is 20 cubic meters and the minimum is 10 cubic meters, the peak-to-trough absolute difference is 10 cubic meters. By recording the fluctuation frequency and peak-to-trough absolute difference within each time unit, and continuously counting these values ​​over multiple time units, and integrating these data, we can form user gas consumption time series fluctuation data, which clearly demonstrates the fluctuations in user gas consumption over time.

[0048] The proposed user gas usage time-series fluctuation data can intuitively display the instantaneous changes in gas usage by downstream users of the pressure regulating station. The fluctuation frequency reflects the frequency of gas usage changes, while the absolute difference between peaks and troughs reflects the severity of the changes in gas usage. By analyzing this data, gas companies can promptly identify abnormal fluctuations in user gas usage and take proactive countermeasures to ensure the stability and safety of gas supply. This also helps optimize the operating parameters of the pressure regulating station and improve gas delivery efficiency.

[0049] In a preferred embodiment of the present invention, based on the pipeline pressure gradient data, flow phase data and user gas consumption time series fluctuation data, the main pipeline intersection point P1, the user access point P2, and the pressure regulating station output point P3 are selected as three dynamic topology reference points, including:

[0050] According to the pipe network topology mapping relationship, the main pipe intersection point with the largest pressure change rate in the pressure gradient data is selected as P1, and its three-dimensional space coordinates are used as the initial reference coordinates;

[0051] Using the initial reference coordinates of the main pipeline intersection point P1 as a reference, combined with the phase offset and amplitude characteristics in the flow phase data, the spatial influence weight of the user access point P2 is calculated. The original three-dimensional coordinates of the user access point P2 are offset compensated based on the spatial influence weight to obtain the weighted real-time monitoring coordinates of the user access point P2.

[0052] Based on the weighted real-time monitoring coordinates of the user access point P2, the original coordinates of the pressure regulating station output point P3 are compensated for time series displacement using the fluctuation frequency and peak-to-trough difference in the user's gas consumption time series fluctuation data to obtain the real-time control coordinates of the pressure regulating station output point P3.

[0053] The initial reference coordinates of the main pipeline intersection point P1, the real-time monitoring coordinates of the user access point P2, and the real-time control coordinates of the pressure regulating station output point P3 are spatially synchronized and converted to a two-dimensional plane through dimensionality reduction mapping; the three points P1, P2, and P3 in the two-dimensional plane are connected to form a dynamic topological unit.

[0054] In the embodiment of the present invention, the above steps, when applied specifically, can be implemented by the following steps, for example:

[0055] Based on the established pipe network topology mapping, all main pipe intersections are identified. Next, the pressure gradient data for each main pipe intersection is examined, focusing on how pressure varies with spatial distance, or the pressure change rate. For example, the relationship between the pressure difference and distance between adjacent nodes at different intersections is compared. A large pressure change rate indicates significant pressure variation over a short distance. The main pipe intersection with the largest pressure change rate is identified as P1, and the 3D spatial coordinates of this point are recorded as the initial reference coordinates for subsequent operations, effectively selecting a key reference "anchor point" in the pipe network.

[0056] This method selects the main pipeline intersection point, where the rate of pressure change is the highest, as P1 because this point is often located where pressure fluctuations are most severe within the pipeline network, potentially representing a critical node or potential problem area for gas transmission. Using this point as the initial reference coordinate provides a highly representative reference point for subsequent analysis, helping to quickly locate abnormal pressure fluctuations within the pipeline network.

[0057] Using the three-dimensional coordinates of P1 as a reference, the phase offset and amplitude characteristics of the flow phase data for each user access point P2 are analyzed. The phase offset reflects the difference between the user's gas usage time and normal conditions, while the amplitude characteristics reflect the fluctuation in gas usage. Taking these two factors into consideration, for example, user access points with early gas usage and large fluctuations are likely to have a greater impact on the pipeline network and are therefore assigned a higher spatial impact weight; conversely, those with less impact are given a lower weight. After determining the spatial impact weight of each P2 point, its original three-dimensional coordinates are adjusted based on this weight. If the weight is large, the coordinates are spatially offset by a relatively large amount according to a specific rule; if the weight is small, the offset is small. Ultimately, the weighted real-time monitoring coordinates of user access point P2 are obtained, which better reflect the impact of this point on the actual operation of the pipeline network.

[0058] For example, the spatial influence weight determination process and value range of user access point P2 are as follows:

[0059] Setting evaluation indicators and grading:

[0060] Phase offset assessment categorizes the phase offset of a user's gas usage into five levels. Based on the typical gas usage time, an offset within ±15 minutes is considered "minor"; 15-30 minutes is "minor"; 30-60 minutes is "moderate"; 60-120 minutes is "significant"; and over 120 minutes is "extreme" (extreme). The longer the offset, the greater the difference in a user's gas usage time from the typical, potentially leading to a more significant impact on pipeline scheduling and pressure balance.

[0061] Amplitude characteristic assessment also categorizes gas usage into five levels. A gas usage fluctuation amplitude (the difference between peak and trough) within 5% of rated gas usage is considered "minimal fluctuation"; 5%-15% is considered "minor fluctuation"; 15%-30% is considered "moderate fluctuation"; 30%-50% is considered "large fluctuation"; and over 50% is considered "extreme fluctuation." Larger amplitudes indicate more dramatic changes in user gas usage, leading to greater impact on the pipeline network.

[0062] Quantitative scoring and weight calculation:

[0063] Each level is assigned a corresponding score: "Minimal offset or fluctuation" corresponds to 1 point, "Small offset / fluctuation" corresponds to 2 points, "Medium offset / fluctuation" corresponds to 3 points, "Large offset / fluctuation" corresponds to 4 points, and "Extreme offset or fluctuation" corresponds to 5 points. For each user access point P2, the corresponding level scores for its phase offset and amplitude characteristics are obtained, and the two scores are added together to obtain the comprehensive impact score for that point. For example, if the phase offset of user access point P2 is "Large offset" (4 points) and the amplitude characteristic is "Medium fluctuation" (3 points), its comprehensive impact score is 7 points.

[0064] Normalize the combined influence scores of all user access points P2 and map them to the range [0, 1] to obtain the spatial influence weight of each P2 point. Assuming the sum of the combined influence scores of all user access points P2 is S, and the combined influence score of a P2 point is s, then the spatial influence weight of that point, w, = s / S.

[0065] Weight value range:

[0066] The weight is close to 0. When the phase offset of the user access point P2 is "extremely small offset" and the amplitude characteristic is "extremely small fluctuation", its spatial influence weight is close to 0. This type of user has stable gas usage time and small fluctuation in gas usage, and the impact on the operation of the pipeline network can be ignored.

[0067] 0 < Weight ≤ 0.2. This weight is in this range if the phase offset and amplitude characteristics are in a "small offset," "small fluctuation," or lower combination. This indicates that user gas usage has little impact on the network, indicating normal and stable gas usage.

[0068] 0.2 < Weight ≤ 0.5. This weight is in this range when a combination of "medium shift" and "medium fluctuation" or lower occurs, or a combination of "large shift or fluctuation" and "very small or small shift or fluctuation" occurs. This indicates that user gas usage behavior has a certain impact on the pipeline network and requires appropriate attention in pipeline network analysis and scheduling.

[0069] 0.5<Weight≤0.8. If there is a combination of "large deviation" and "medium / large fluctuation", or a combination of "large deviation / fluctuation" and "small deviation or fluctuation", the weight is in this range, which means that the user's gas usage behavior has a greater impact on the pipeline network, which may cause local pressure fluctuations or changes in flow distribution in the pipeline network.

[0070] The weight is close to 1. When the phase offset at user access point P2 is "extremely large" and the amplitude characteristic is "extremely large," the weight is close to 1. This type of user gas usage has a significant impact on pipeline network operations, and targeted adjustments or enhanced monitoring of the pipeline network may be necessary.

[0071] By combining flow phase data to calculate spatial impact weights and compensate coordinates, this invention fully considers the impact of user gas usage behavior on the pipeline network. The real-time monitoring coordinates of the user access point P2 obtained in this way are no longer simply geographical coordinates, but incorporate the user's gas usage characteristics. This can more accurately reflect the actual role of the user access point in the pipeline network operation, helping gas companies to more accurately grasp the impact of user gas usage on the pipeline network and optimize gas distribution and scheduling strategies.

[0072] The frequency of fluctuations within a unit of time (e.g., an hour) is categorized into five levels: 5 or fewer is considered "very low frequency," 5-15 is "low frequency," 15-30 is "medium frequency," 30-50 is "high frequency," and over 50 is "very high frequency." The higher the frequency, the more frequent the user's gas consumption changes, and the more frequently the pressure regulating station needs to adjust its output.

[0073] Peak-to-valley difference grading is based on the rated output of the pressure regulating station and is divided into five levels based on the ratio of the peak-to-valley difference to the rated output. A difference ratio of less than 5% is considered "minimal fluctuation," 5%-15% is considered "small fluctuation," 15%-30% is considered "moderate fluctuation," 30%-50% is considered "large fluctuation," and over 50% is considered "extreme fluctuation." Larger differences indicate more dramatic changes in user gas consumption, and thus a greater impact on the pressure regulating station's output stability.

[0074] Comprehensive impact assessment:

[0075] An evaluation matrix was constructed by combining the five levels of fluctuation frequency with the five levels of peak-to-trough difference to form a 5×5 evaluation matrix. For example, a specific combination in the matrix corresponds to a "high frequency" fluctuation frequency and a "large fluctuation" peak-to-trough difference.

[0076] Determine the impact level. Based on different combinations, the degree to which the output status of the voltage regulating station is affected by users is divided into five levels. For example, combinations of "high frequency + large fluctuations" and "extremely high frequency + medium fluctuations" and above correspond to the "extreme impact" level; combinations of "high frequency + medium fluctuations" and "medium frequency + large fluctuations" correspond to the "significant impact" level, and so on, up to "extremely low frequency + minimal fluctuations" corresponding to the "minimal impact" level.

[0077] Compensation direction determination:

[0078] Forward compensation (time axis backward movement): When the user's gas consumption peak appears earlier or the fluctuation frequency increases, it means that the pressure regulating station needs to respond to these changes earlier. At this time, the original coordinates of P3 are compensated for positive displacement on the time axis, and the coordinate point is moved backward.

[0079] Negative compensation (time axis forward): If the user's gas consumption peak is delayed and the fluctuation frequency is reduced, it indicates that the response of the pressure regulating station can be appropriately delayed. At this time, the original coordinates of P3 are compensated for negative displacement on the time axis to move the coordinate point forward.

[0080] Compensation amplitude calculation:

[0081] Basic compensation unit setting, set a basic compensation time unit, such as 5 minutes. This unit represents the time displacement of the P3 coordinate at the minimum impact level (extremely small impact).

[0082] Gradual multiplier adjustment: Compensation multipliers are set for each of the five levels of comprehensive impact. "Minor impact" corresponds to 1x the base unit, "Minor impact" corresponds to 2x, "Medium impact" corresponds to 4x, "Large impact" corresponds to 8x, and "Extreme impact" corresponds to 16x. For example, if the comprehensive impact is "Extreme impact," the compensation range is 16 x 5 = 80 minutes.

[0083] The final compensation amount is determined based on the grading multiple and fine-tuned based on the specific value of the peak-to-trough difference. If the difference is in the upper half of the grade range, the upper limit of the compensation multiple for that grade is used; if it is in the lower half, the lower limit is used. For example, at the "Extreme Impact" grade, the compensation is 80 minutes if the difference is in the upper half, and 70 minutes if it is in the lower half.

[0084] Coordinate adjustment and real-time control of coordinate generation:

[0085] Time dimension displacement: The original coordinates of P3 are displaced in the time dimension according to the determined compensation direction and amplitude. For example, if the compensation direction is positive and the amplitude is 60 minutes, the time coordinate value of P3 will be increased by 60 minutes; if it is negative, the corresponding time value will be decreased.

[0086] The spatial coordinates are adjusted synchronously. While shifting in the time dimension, the impact of user gas consumption fluctuations on the spatial state of the pressure regulating station is taken into consideration. If the fluctuation causes a significant change in the output pressure or flow of the pressure regulating station, the coordinates of P3 are slightly adjusted in the spatial dimension. The adjustment amplitude is proportional to the degree of impact of the fluctuation.

[0087] Real-time control coordinates are generated. After adjustment in time and space dimensions, the real-time control coordinates of the pressure regulating station output point P3 are obtained. These coordinates reflect both the temporal correlation between the pressure regulating station output state and the user's gas consumption fluctuations, and the impact of the fluctuations on the spatial state of the pressure regulating station.

[0088] The present invention uses the time-series fluctuation data of user gas consumption to perform time-series displacement compensation on the coordinates of the output point P3 of the pressure regulating station, closely linking the dynamic changes of user gas consumption with the regulation and control of the pressure regulating station. This allows the real-time regulation coordinates of P3 to reflect the response of the pressure regulating station to the fluctuations in user gas consumption, helping gas companies to better understand the operating status of the pressure regulating station, ensure the stability and safety of gas supply, and achieve refined regulation and control of the pipeline network.

[0089] The dynamic topology unit construction process includes:

[0090] Spatial unified calibration:

[0091] To align coordinates, first determine a unified spatial origin, such as selecting the geometric center of the entire gas pipeline network coverage area as the origin. Then, for the initial reference coordinates of P1, the real-time monitoring coordinates of P2, and the real-time control coordinates of P3, calculate their offsets relative to this origin. For example, the original coordinates of P1 are (X1, Y1, Z1). With the new origin as a reference, calculate the distance it needs to be translated in the X, Y, and Z directions, and adjust its coordinates to the coordinate system based on the new origin. In the same way, translate the coordinates of P2 and P3 so that the coordinates of the three points are based on the same origin, achieving preliminary alignment.

[0092] Scale uniformity: Check that the three points have consistent scales along the X, Y, and Z axes. Since different areas of the actual pipe network may have different measurement units or precision, it is necessary to standardize the scale. For example, if the Z-axis coordinate of P1 is measured in meters, while the Z-axis coordinates of P2 and P3 are measured in centimeters, divide the Z-axis coordinate values ​​of P2 and P3 by 100 and convert them to meters. This ensures that the three points have the same scale along the three axes, completing the spatial calibration and placing them in the same reference system.

[0093] Dimensionality reduction mapping to a two-dimensional plane:

[0094] Determine the projection direction and select a suitable projection direction. A common projection direction is the direction perpendicular to the main laying plane of the pipeline network. For example, if the gas pipeline network is mainly laid on the underground plane, the vertical upward direction can be selected as the projection direction.

[0095] The projection operation projects the calibrated three-dimensional coordinates of points P1, P2, and P3 onto a two-dimensional plane along the selected projection direction. Imagine a beam of parallel light illuminating the three points from the projection direction. Their shadows on the two-dimensional plane are the projected coordinate positions. Calculate the corresponding X and Y coordinate values ​​of the three points on the two-dimensional plane respectively, discard the Z coordinate value, and obtain the new coordinates of the three points on the two-dimensional plane.

[0096] Coordinate adjustment and optimization: Check and adjust the projected 2D coordinates to ensure that the relative positions between points match the spatial relationships in the actual pipe network. If you find unreasonable overlap or offset in the positions of certain projected points, fine-tune the coordinate values ​​so that their layout on the 2D plane more accurately reflects the relative positions of the three points in the actual pipe network, facilitating subsequent analysis.

[0097] Constructing a dynamic topology unit:

[0098] To connect lines, use the line segment tool on a two-dimensional plane to connect points P1, P2, and P3 in sequence. Start from point P1 and draw a line segment to connect to point P2. Then draw a line segment from point P2 to connect to point P3. Finally, draw a line segment from point P3 back to point P1 to form a closed triangular area.

[0099] Information integration and annotation: the previously collected data related to the three points, such as the pressure change rate of point P1, the spatial influence weight of point P2, and the temporal displacement compensation basis of point P3, are marked on the corresponding points or triangular areas. Annotations can also be added to explain the relationship between these data and points, as well as their significance to the entire dynamic topology unit. This makes this triangular area not only a geometric figure, but also a visual analysis unit that integrates important information of key nodes in the pipeline network, and can intuitively present the spatial relationship and mutual influence between nodes.

[0100] This invention integrates and simplifies the complex spatial relationships and operational data of key nodes in the pipeline network by constructing a dynamic topology unit. This unit allows for intuitive analysis of the relationships between trunk pipeline intersections, user access points, and pressure regulating station output points, quickly understanding the operational characteristics of local areas within the pipeline network. This facilitates comprehensive and local analysis of the gas pipeline network, providing a clear analytical framework for subsequent operations such as regional segmentation based on the Delaunay decomposition algorithm and extraction of pressure-flow coupling eigenvalues, thereby improving the efficiency and accuracy of analysis of pipeline network operational status.

[0101] In a preferred embodiment of the present invention, a dynamic topological unit is constructed based on three points, the unit is divided into regions based on the Delaunay decomposition algorithm, and the pressure-flow coupling characteristic values ​​of the sub-regions are extracted, including:

[0102] A Delaunay triangulation algorithm is performed on the dynamic topological unit to generate a minimum angle partitioning structure containing at least three sub-regions, including: using the real-time two-dimensional plane coordinates of the three points P1, P2, and P3 in the dynamic topological unit as the triangulation input point set to construct an initial convex hull boundary; based on the real-time change amplitude of the user's gas consumption time series fluctuation data, when the fluctuation amplitude exceeds a set threshold, an auxiliary vertex Q1 is dynamically inserted at the midpoint of the line connecting P2 and P3; based on the phase offset of the flow phase data, an auxiliary vertex Q2 is inserted at the midpoint of the line connecting P1 and P2; Q1 and Q2 are added to the triangulation point set and the convex hull is updated to obtain an updated triangulation point set; traversing the updated triangulation point set, a diagonal exchange test is performed on a quadrilateral formed by any four adjacent points; if the minimum internal angle increases after the exchange, the topological connection relationship is updated; iterative optimization is performed until all quadrilaterals meet the Delaunay empty circle criterion to generate a final triangulation structure; each minimum topological unit in the final triangulation structure is defined as a sub-region to obtain a sub-region vertex coordinate set and an adjacency relationship table;

[0103] For each divided sub-region, the pipeline pressure gradient data associated with the vertex coordinates of the sub-region is extracted, and the amplitude characteristics of the flow phase data in the same spatial domain are simultaneously obtained;

[0104] Calculating the arithmetic mean of the product of the pressure gradient data and the flow phase amplitude characteristic, and outputting the basic spatial coupling factor of the sub-region;

[0105] Based on the basic spatial coupling factor, locate the gas consumption time series fluctuation data of the user access points within the sub-area, calculate the statistical variance value of the fluctuation data of the users within the continuous time window, perform linear normalization processing on the variance value in the interval [0, 1], and generate the time series correction coefficient;

[0106] Adding the basic spatial coupling factor and the timing correction coefficient to generate an intermediate coupling feature;

[0107] The intermediate coupling characteristic is multiplied by the spatial weight coefficient of the sub-area in the pipe network topology, and the pressure-flow coupling characteristic value is finally output.

[0108] In the embodiment of the present invention, the above steps, when applied specifically, can be implemented by the following steps, for example:

[0109] Based on the two-dimensional coordinates of points P1, P2, and P3, the relative position relationship of these three points is determined. Since there are only three points, they are either collinear or can form a triangle. In actual pipe network analysis, these three points are usually not collinear, so these three points are directly connected to form a triangle. This triangle is the minimum convex polygon that can enclose these three points, that is, the initial convex hull.

[0110] Mark the boundary direction, mark the three sides of the triangle in a clockwise or counterclockwise direction, and determine the connection order of each vertex.

[0111] Dynamically insert auxiliary vertices:

[0112] Fluctuation detection: monitor the fluctuation amplitude of user gas usage in real time and compare it with a pre-set threshold. For example, the threshold can be set at 1.5 times the average fluctuation amplitude. When the actual fluctuation amplitude exceeds this threshold, it indicates that the gas usage in the area is unstable.

[0113] Insert vertex Q1. When the fluctuation amplitude exceeds the threshold, calculate the coordinates of the midpoint of the line connecting P2 and P3. Specifically, take the average of the coordinates of P2 and P3 on the X and Y axes, and the resulting new coordinate point is Q1. Insert Q1 into the point set, which now contains four points: P1, P2, Q1, and P3.

[0114] Phase offset detection: Analyze the phase offset of the flow phase data and determine whether it has reached a certain level. For example, if the phase offset exceeds 20% of the normal fluctuation range, it is considered necessary to divide the area more finely.

[0115] Insert vertex Q2. If the phase offset reaches the set value, calculate the coordinates of the midpoint of the line connecting P1 and P2 in a similar way to calculating Q1. The resulting new coordinate point is Q2. Insert Q2 into the point set, forming a new subdivision point set consisting of P1, P2, P3, Q1, and Q2.

[0116] Update the convex hull boundary:

[0117] Re-determine the outer points. After adding Q1 and Q2, re-examine the positional relationship of these five points to find the outermost points. These points form the new convex hull boundary. Specifically, starting from any point, check each point in a direction (such as clockwise) to determine whether there are other points outside the line connecting the point and its adjacent points. If so, adjust the convex hull boundary until all outer points are determined.

[0118] Connect the new boundary by connecting the determined outer points in sequence to form a new convex polygon. This convex polygon is the minimum convex hull boundary containing the five points. For example, if the new outer points are in the order P1, Q2, P2, Q1, P3, then connect these five points in sequence to form a pentagonal convex hull boundary.

[0119] Diagonal swap detection and optimization:

[0120] Traverse the quadrilateral and find a quadrilateral consisting of any four adjacent points in the updated set of subdivision points. For example, there may be a quadrilateral consisting of P1, Q2, P2, Q1, or a quadrilateral consisting of Q2, P2, Q1, P3, etc.

[0121] To test the effect of diagonal swapping, consider the two diagonals of each quadrilateral separately. For example, for quadrilateral P1-Q2-P2-Q1, the two diagonals are P1-P2 and Q2-Q1. Calculate the minimum interior angle of the new triangle formed by swapping the diagonals and compare the minimum interior angles before and after the swap.

[0122] Update the topological connections. If swapping diagonals increases the minimum interior angle, the new connection method better meets the requirements of Delaunay triangulation. Replace the original diagonals with the new ones and update the topological connections of the quadrilaterals. For example, if swapping P1-P2 and Q2-Q1 increases the minimum interior angle, split the quadrilateral P1-Q2-P2-Q1 into two triangles: P1-Q2-Q1 and Q2-P2-Q1.

[0123] Iterative optimization repeats the diagonal swapping process described above, checking and optimizing all possible quadrilaterals until all quadrilaterals satisfy the Delaunay empty circle criterion, meaning that the circumcircle of each triangle contains no other points. At this point, the triangulation structure reaches its optimal state, generating the final triangulated structure.

[0124] Define sub-areas and adjacency relationships:

[0125] Divide the triangle into sub-regions, defining each triangle in the final triangulated structure as a sub-region and recording the coordinates of the three vertices of each sub-region. For example, one sub-region might consist of vertices P1, Q2, and Q1, while another sub-region might consist of vertices Q2, P2, and Q1, and so on.

[0126] Establish an adjacency table, check the shared edges between each sub-region and other sub-regions, and record the numbers of adjacent sub-regions. For example, if sub-region A and sub-region B share an edge, record A and B as adjacent in the adjacency table. In this way, a complete adjacency table is formed to describe the spatial connection relationship between each sub-region.

[0127] Calculation process for extracting sub-region pressure-flow coupling eigenvalues:

[0128] To find the pressure value, for each sub-region, check the coordinates of its three vertices in turn. Find the pressure values ​​corresponding to these three vertex coordinates in the pre-collected and stored original pipeline pressure gradient data. For example, if the coordinates of the three vertices of a sub-region are (X1, Y1), (X2, Y2), and (X3, Y3), search the pressure gradient data for the pressure values ​​recorded at these three coordinates and extract them one by one.

[0129] Flow Phase Amplitude Acquisition: Within the same spatial range, i.e., the pipe network area covered by the sub-region, flow phase data is found. From this data, the amplitude characteristic of the flow waveform is extracted, which is the difference between the maximum and minimum values ​​of the flow waveform. For example, if the maximum flow waveform in a sub-region is 100 cubic meters per hour and the minimum is 20 cubic meters per hour, the flow phase amplitude characteristic for this region is 80 cubic meters per hour.

[0130] Calculate the basic spatial coupling factor:

[0131] Data multiplication involves multiplying the extracted pressure gradient data within the subregion with the flow phase amplitude characteristics according to their corresponding relationships. For example, if the pressure values ​​corresponding to the three vertices of the subregion are P1, P2, and P3, and the flow phase amplitude characteristic is A, then P1×A, P2×A, and P3×A are calculated, respectively, to obtain three sets of product data.

[0132] To find the arithmetic mean, add the three sets of product data and divide them by the number of data (3 here) to calculate their arithmetic mean. This average value is the basic spatial coupling factor of the sub-region, which reflects the degree of correlation between pressure and flow in the sub-region in the spatial dimension.

[0133] Generate timing correction coefficients:

[0134] Locate the fluctuation data. Based on the scope of the sub-area, find the user access points under its jurisdiction. From the gas usage time series fluctuation data of these user access points, filter out the fluctuation data belonging to the sub-area. For example, if sub-area A covers three user access points, extract all the gas usage time series fluctuation data of these three user access points.

[0135] Calculate the variance of the extracted fluctuation data within a continuous time window. This variance measures the degree of dispersion of the fluctuation data. A larger variance indicates more drastic fluctuations in user gas usage. For example, by calculating gas usage data over a period of time, the variance of the fluctuation data for users in this subregion is 25.

[0136] Linear normalization applies a linear transformation to the calculated variance, mapping it to the interval [0, 1]. Specifically, the variance is scaled according to its relative size within the variance of all subregions. For example, if the maximum variance across all subregions is 100 and the minimum is 0, and the variance of the current subregion is 25, the resulting value after calculation and transformation falls within the interval [0, 1]. This value is the timing correction coefficient.

[0137] Create an intermediate coupling feature:

[0138] Add the data and add the basic spatial coupling factor calculated previously to the timing correction coefficient to obtain a new value, which is the intermediate coupling characteristic value. In this way, the pressure-flow correlation characteristics in the spatial dimension and the user gas consumption fluctuation characteristics in the time dimension are preliminarily integrated.

[0139] Compute the final eigenvalue:

[0140] Assign a spatial weight coefficient to each sub-region based on its importance within the overall network topology. This importance can be determined by factors such as the sub-region's proximity to the main pipeline and whether it contains key user access points. For example, a sub-region near a pressure regulating station with a high user density might be assigned a higher weight coefficient of 0.8, while a sub-region at the end of the network with fewer users might be assigned a lower weight coefficient of 0.2.

[0141] Calculate the final value and multiply the intermediate coupling eigenvalue by the spatial weight coefficient assigned to the sub-region. The result is the final pressure-flow coupling eigenvalue. This eigenvalue comprehensively considers the spatial characteristics of the sub-region, the time characteristics of the user's gas consumption, and the importance of the sub-region in the pipeline network, and can fully reflect the coupling relationship between pressure and flow in the sub-region.

[0142] By combining pipeline pressure gradient data, flow phase amplitude characteristics, and user gas consumption time series fluctuation data, a comprehensive assessment of the pipeline network operation status is conducted from both spatial and temporal dimensions, changing the limitations of previous single-dimensional analysis and more comprehensively reflecting the impact of gas flow characteristics and user gas consumption behavior on the pipeline network; the basic spatial coupling factor reflects the spatial correlation between pressure and flow in the sub-region, and the time series correction coefficient reflects the temporal fluctuation characteristics of user gas consumption. The intermediate coupling characteristics generated by the combination of the two can accurately characterize the unique operating characteristics of each sub-region. Combined with the spatial weight coefficient, it highlights the importance of key areas in the pipeline network, making the analysis results more in line with the actual pipeline network operation situation and helping to discover weak links and potential problem areas in the pipeline network.

[0143] By dynamically inserting auxiliary vertices at the midpoints of the lines connecting P2 and P3, and P1 and P2, the triangulation structure can be adaptively adjusted according to the user's gas consumption timing fluctuations and flow phase offsets, more precisely capturing the sensitive areas of pressure and flow changes in the pipeline network, and improving the analysis accuracy; the triangulated structure generated by the Delaunay triangulation algorithm has good geometric properties (maximizing the minimum internal angle), avoiding the appearance of narrow and long triangles, making the sub-region division more uniform and reasonable, and facilitating subsequent feature extraction and analysis; by combining pressure gradient data, flow phase amplitude characteristics, and gas consumption timing fluctuation data, the pipeline network operation status is comprehensively evaluated from both spatial and temporal dimensions, more comprehensively reflecting the impact of gas flow characteristics and user gas consumption behavior on the pipeline network; by assigning spatial weight coefficients to sub-regions, the influence of sub-regions where key nodes in the pipeline network (such as near pressure regulating stations and user-dense areas) are located can be highlighted.

[0144] In a preferred embodiment of the present invention, a weighted correction is performed on the user gas demand mutation probability characteristic vector using the pressure-flow coupling characteristic value to generate a corrected mutation probability characteristic vector, including:

[0145] The pressure-flow coupling eigenvalues ​​of each sub-region are input into the matrix converter, sorted by the spatial weight coefficients of the sub-regions in the pipe network topology, and scaled by the length of the pipes under each sub-region as a normalization factor to generate a diagonal correction coefficient matrix with the same dimension as the mutation probability eigenvector.

[0146] According to the preset user gas demand mutation probability eigenvector, the correction coefficient matrix and the mutation probability eigenvector are subjected to a Hadamard product operation to obtain a weighted eigenvector;

[0147] The weighted feature vector is normalized using the L2 norm to obtain a modified mutation probability feature vector.

[0148] In the embodiment of the present invention, the above steps, when applied specifically, can be implemented by the following steps, for example:

[0149] The pressure-flow coupling eigenvalues ​​calculated for each subregion are sorted according to their spatial weight coefficients in the network topology. Subregions with higher spatial weight coefficients have higher eigenvalues ​​in the sorting. For example, if the spatial weight coefficients for subregion A are 0.8, subregion B is 0.6, and subregion C is 0.4, the eigenvalues ​​are ranked A, B, and C.

[0150] Normalization scaling uses the length of the pipeline in each sub-region as a normalization factor to scale the sorted eigenvalues. Specifically, the eigenvalue of each sub-region is divided by the length of the pipeline in that sub-region to obtain the scaled eigenvalue. For example, if the eigenvalue of sub-region A is 10 and the pipeline length is 5 kilometers, the scaled value is 10 ÷ 5 = 2.

[0151] Generate a diagonal matrix, arrange the scaled eigenvalues ​​on the diagonal of the matrix in sequence, and fill the other positions with 0 to form a diagonal matrix. The dimension of this matrix is ​​the same as the eigenvector of the probability of sudden change in user gas demand.

[0152] Hadamard product operations, specifically including:

[0153] Prepare the original vector and obtain the preset user gas demand mutation probability feature vector. The vector contains multiple dimensions, each dimension representing the probability of a different type of gas demand mutation.

[0154] Perform element-by-element multiplication to perform a Hadamard product between the diagonal correction coefficient matrix and the mutation probability eigenvector, multiplying the corresponding elements. For example, the first diagonal element of the diagonal matrix is ​​multiplied by the first element of the eigenvector, the second diagonal element is multiplied by the second element, and so on. This results in a new vector, the weighted eigenvector, in which each element is corrected by the pressure-flow coupling eigenvalue.

[0155] L2 norm normalization:

[0156] Calculate the vector modulus and the L2 norm of the weighted eigenvector, which is the square root of the sum of the squares of the vector elements. Divide each element of the weighted eigenvector by its L2 norm to obtain the corrected mutation probability eigenvector. After normalization, the vector modulus is 1, which facilitates subsequent analysis and comparison while maintaining the relative proportional relationship between the elements of the vector.

[0157] In this embodiment of the present invention, by incorporating the pressure-flow coupling eigenvalue into the correction of the mutation probability eigenvector, the impact of the actual operating status of the pipeline network on user gas demand is fully considered. Changes in pipeline pressure and flow are often precursors to changes in user demand. This data-driven correction method can more accurately capture the sudden changes in user gas demand, improving the accuracy of the prediction model. The application of spatial weight coefficients makes the pressure-flow characteristics of key areas in the pipeline network (such as those near pressure regulating stations and areas with high user density) more influential on the sudden change probability. This helps highlight the potential risks in these areas, enabling gas companies to conduct more targeted monitoring and regulation, and proactively prevent potential large fluctuations in gas demand. Normalization eliminates the dimensional effects caused by differences in pipeline lengths across sub-areas, making the eigenvalues ​​of each sub-area comparable. Furthermore, L2 norm normalization ensures that the corrected eigenvectors are within a reasonable range, facilitating subsequent integration and analysis with other models or indicators, thereby improving the overall stability and reliability of the system.

[0158] In a preferred embodiment of the present invention, the modified mutation probability feature vector is tensor-concatenated with the pipeline equivalent inner diameter dynamic feature vector and input into a fully connected diagnostic network to generate a caliber matching deviation index, including:

[0159] According to the corrected mutation probability feature vector and the real-time updated pipeline equivalent inner diameter dynamic feature vector, it is expanded into a three-dimensional tensor along the feature dimension, where the first dimension is the time step, the second dimension is the spatial partition, and the third dimension is the feature channel;

[0160] The three-dimensional tensor is input into a three-layer fully connected diagnostic network, where:

[0161] In the first layer, the linear transformation layer reduces the input feature dimension to 1 / 2 of the original dimension and passes the ReLU activation function;

[0162] In the second layer, the linear transformation layer is further reduced to 1 / 4 of the original dimension and activated by the Sigmoid function;

[0163] The third layer, the single neuron linear layer outputs the original value of the deviation;

[0164] The original value output by the third layer is dynamically threshold calibrated, and the Z-score normalized offset of the current value relative to the benchmark distribution is calculated by fitting the benchmark distribution curve with the preset historical normal operating condition data.

[0165] In the embodiment of the present invention, the above steps, when applied specifically, can be implemented by the following steps, for example:

[0166] The corrected mutation probability feature vector and the real-time updated pipeline equivalent inner diameter dynamic feature vector are obtained. The former reflects the possibility of sudden changes in user gas demand, while the latter reflects the changes in pipeline inner diameter with the operation status of the pipeline network.

[0167] These two feature vectors are expanded into a three-dimensional tensor along the feature dimension. First, the time step is determined as the first dimension. For example, a time step of one hour is used to record data at different times. The second dimension is the spatial partition, corresponding to the previously divided pipe network sub-areas, each of which has corresponding data records. The third dimension is the feature channel, which stores the mutation probability feature and the dynamic feature of the equivalent inner diameter of the pipe. Data filling, in the order of time step, spatial partition, and feature channel, sequentially fills the corresponding data into each position of the three-dimensional tensor to form a complete three-dimensional data structure.

[0168] Fully connected diagnostic network operations:

[0169] The first layer of operations feeds the constructed three-dimensional tensor into the first layer of the fully connected diagnostic network. This layer is a linear transformation layer, which processes the input feature dimensions, reducing them to half their original dimensions. After this processing, the tensor is activated using the ReLU (Reduced Linear Unit) activation function, which converts all negative eigenvalues ​​to zero while retaining positive eigenvalues. This function selects more valuable features and enhances the network's nonlinear representation capabilities.

[0170] The second layer of operations: Data processed by the first layer enters the second layer, also a linear transformation layer. This layer further reduces the feature dimension to 1 / 4 of the original dimension, performing deeper data compression and feature extraction. It then passes through the Sigmoid activation function, which maps the feature values ​​to a range between 0 and 1, converting the features into probabilistic form.

[0171] In the third layer of operation, the data output by the second layer enters the third single-neuron linear layer, which performs final processing on the input data and outputs an original value of the deviation. This value reflects the preliminary evaluation result of the caliber matching of the gas metering equipment under the current pipeline network status.

[0172] Dynamic Threshold Calibration:

[0173] Filter out all records marked as "normal operating conditions" from historical data, exclude data in abnormal conditions such as pipeline leakage, equipment failure, and extreme gas consumption peaks, clean the filtered data, remove missing values ​​and obvious outliers, and ensure data integrity and reliability.

[0174] Statistical feature calculation, calculate the statistics of caliber matching related indicators in normal operating condition data, including mean (reflecting the average level under normal conditions), standard deviation (reflecting the degree of dispersion of data), quantiles (such as 25% quantile and 75% quantile, used to characterize the range of data distribution), etc.

[0175] Normal distribution fitting calculation process:

[0176] Extract raw data sets (such as historical deviation values) from historical normal operating condition data to match relevant indicators, ensuring that the data only contains records of stable network operation and no abnormal events. Sort the data sets by value from smallest to largest to facilitate subsequent observation of data distribution trends. For example, arrange 1,000 deviation data points under normal operating conditions from smallest to largest to observe their central tendency and dispersion range.

[0177] Draw a histogram. Divide the sorted data into several intervals (e.g., with an interval of 0.1), count the frequency of the data within each interval, and draw a histogram. The histogram can be used to initially determine whether the data exhibits a bell-shaped distribution with high values ​​in the middle and low values ​​on both sides—in other words, whether it conforms to the visual characteristics of a normal distribution.

[0178] Calculate the key parameters of the normal distribution:

[0179] To calculate the mean, μ, add up all the data points and divide by the total number of data points to get the mean. The mean represents the center of a normal distribution. For example, if the mean of 1000 data points is 1.2, the average deviation index under normal operating conditions is 1.2.

[0180] To calculate the standard deviation σ, first calculate the square of the difference between each data point and the mean, add up all the square values ​​and divide by the total amount of data to get the variance; then square the variance to get the standard deviation. The standard deviation reflects the degree of dispersion of the data. For example, a standard deviation of 0.3 indicates that under normal operating conditions, the deviation index usually fluctuates within the range of ±0.3 of the mean.

[0181] Generate a benchmark distribution curve:

[0182] Determine the shape of the curve. Based on the calculated mean and standard deviation, draw a bell-shaped curve with the mean as the center and the standard deviation as the width. The highest point of the curve corresponds to the mean position, and the curve gradually decreases on both sides. Theoretically, it extends to positive and negative infinity, but in practice, about 95% of the data falls within the range of the mean ± 2 times the standard deviation, and about 99.7% of the data falls within the range of the mean ± 3 times the standard deviation.

[0183] Mark key locations on the curve, such as the mean μ, mean ± 1 times the standard deviation (μ ± σ), and mean ± 2 times the standard deviation (μ ± 2σ), to clearly define the distribution range of data under normal operating conditions. For example, if the mean is 1.2 and the standard deviation is 0.3, the curve contains approximately 68% of the data between 0.9 (μ - σ) and 1.5 (μ + σ), and approximately 95% of the data between 0.6 (μ - 2σ) and 1.8 (μ + 2σ).

[0184] This method superimposes the probability density function curve of a normal distribution on a histogram to observe whether the curve generally matches the shape of the histogram. For example, if the peak of the histogram coincides with the peak position of the curve, and the frequency decay trend on both sides is consistent with the curve, the fit is good. Through visual observation or simple statistical methods (such as calculating whether the actual proportion of data within the ranges of μ±σ and μ±2σ is close to the theoretical proportions of 68% and 95%), the fitted normal distribution is verified to reasonably describe the data distribution characteristics. If the actual proportion is close to the theoretical proportion, the fit is considered effective. By fitting the normal distribution, the indicator distribution under normal operating conditions is converted into two key parameters: mean and standard deviation, clearly defining the numerical range of "normal" conditions (e.g., μ±2σ). Operators can quickly determine whether the current data falls within the normal range. For example, if the deviation index at a certain moment is 2.0, but μ = 1.2 and σ = 0.3, 2.0 > μ + 2σ × 1.8, which can be immediately identified as an abnormality.

[0185] Unified abnormality judgment standards, the standardized characteristics of normal distribution (such as Z-score) make it possible to use "multiple standard deviations from the mean" to measure the degree of deviation of indicators with different dimensions and value ranges. For example, whether it is the pressure difference indicator (unit kPa) or the flow fluctuation indicator (unit m 3 / h), can be converted into comparable abnormality levels through Z-score to avoid confusion in judgment caused by differences in indicator units.

[0186] The baseline distribution curve is not fixed; its mean and standard deviation are periodically recalculated based on newly collected normal operating data, updating the curve shape. This allows the system to adapt to long-term changes in network operation (such as user structure adjustments and equipment aging) and maintain the timeliness of the anomaly judgment criteria. For example, increased gas consumption during the winter heating season may cause the mean and standard deviation of the deviation index to shift upward overall. Refitting can avoid misinterpreting normal fluctuations as anomalies.

[0187] Based on the "3σ principle" of normal distribution (meaning 99.7% of data falls within μ ± 3σ), situations where the absolute value of the Z-score exceeds 3 are defined as "extremely high risk," automatically triggering an emergency alert. This statistically based alert logic eliminates the need for human intervention, reduces subjective judgment errors, and ensures a rapid response to extreme abnormalities.

[0188] Visual verification: compare the fitted distribution curve with the histogram or probability density plot of the actual data to check whether the fitting effect meets expectations and ensure that the curve can accurately reflect the distribution law of the indicators under normal working conditions.

[0189] Calculate the normalized offset (Z-score):

[0190] Locate the original value. Substitute the original deviation value output by the fully connected network into the mathematical expression of the benchmark distribution curve to determine its relative position within the benchmark distribution. For example, if the benchmark distribution is normal, and the original value is to the right of the mean, then the current state is deviating from the normal level in a positive direction.

[0191] To calculate the degree of deviation, calculate the difference between the original value x and the mean (x - μ) based on the mean μ and standard deviation σ of the benchmark distribution. Then, divide this difference by the standard deviation to obtain the standardized deviation (Z-score). This value represents the number of standard deviations the original value is away from the mean. For example, Z = 1 means the original value is 1 standard deviation above the mean, and Z = -1 means it is 1 standard deviation below the mean.

[0192] Mapping to a deviation indicator: Map the standardized offset to a specific deviation indicator based on business needs. For example, setting the absolute value range of the Z-score to correspond to the deviation level: |Z| ≤ 1 for "normal", 1 < |Z| ≤ 2 for "mild deviation", and |Z| > 2 for "severe deviation", making it easier for operators to intuitively understand the current pipeline network status.

[0193] The present invention realizes data fusion of time, space and different feature dimensions by constructing the corrected mutation probability feature vector and the dynamic feature vector of the pipeline equivalent inner diameter into a three-dimensional tensor. This fusion method can fully capture the dynamic information of changes in user demand and pipeline status during the operation of the pipeline network; the fully connected diagnostic network gradually extracts features and reduces the dimension of the input data through multi-layer linear transformation and activation function processing. It not only effectively reduces the data dimension and computational complexity, but also retains key features, highlights the factors that have an important impact on caliber matching, and improves the accuracy and efficiency of the analysis; dynamic threshold calibration is performed based on the benchmark distribution curve fitted by historical normal operating data, and the evaluation standard of the deviation index can be adaptively adjusted according to the actual operation of the pipeline network. Compared with the fixed threshold method, this method is more adaptable to changes in the operation status of the pipeline network, accurately identifies abnormal situations, avoids misjudgment and missed judgment, and improves the reliability of the caliber matching deviation index.

[0194] like Figure 2 As shown, an embodiment of the present invention further provides an intelligent analysis and diagnosis method for gas metering instrument caliber matching, comprising:

[0195] Step 1: Use pressure sensors deployed at the three-dimensional coordinate locations of the gas pipeline network trunk nodes, user access points, and pressure regulating stations to collect pipeline pressure gradient data, flow phase data, and user gas consumption time series fluctuation data in real time;

[0196] Step 2: Based on pipeline pressure gradient data, flow phase data, and user gas consumption time series fluctuation data, the main pipeline intersection point P1, user access point P2, and pressure regulating station output point P3 are selected as three vertices, and a dynamic triangle is formed based on the three vertices. The dynamic triangle is meshed using the Delaunay triangulation algorithm, and the pressure-flow coupling eigenvalue of the mesh is extracted.

[0197] Step 3: Perform weighted correction on the user gas demand mutation probability characteristic vector using the pressure-flow coupling characteristic value to obtain a corrected mutation probability characteristic vector;

[0198] In step 4, the corrected mutation probability feature vector is tensor-concatenated with the pipeline equivalent inner diameter dynamic feature vector and input into the fully connected diagnostic network to generate a caliber matching deviation index. When the deviation exceeds the threshold, the linked timing prediction network outputs a safe and redundant caliber replacement plan.

[0199] An embodiment of the present invention further provides a computing device comprising: a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, performs the above-described method. All implementations in the above-described method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0200] The embodiment of the present invention further provides a computer-readable storage medium storing instructions, which, when executed on a computer, causes the computer to execute the above-described method. All implementations in the above-described method embodiment are applicable to this embodiment and can achieve the same technical effects.

[0201] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. An intelligent analysis and diagnosis system for gas metering instrument caliber matching, characterized in that: include: The acquisition module is used to collect pipeline pressure gradient data, flow phase data, and user gas consumption time series fluctuation data in real time through pressure sensors deployed at the three-dimensional spatial coordinate positions of the gas pipeline network trunk nodes, user access points, and pressure regulating stations; The extraction module is used to select the main pipeline intersection point P1, the user access point P2, and the pressure regulating station output point P3 as three dynamic topology reference points based on the pipeline pressure gradient data, the flow phase data, and the user gas consumption time series fluctuation data; construct a dynamic topology unit based on the three points, divide the unit into regions based on the Delaunay subdivision algorithm, and extract the pressure-flow coupling characteristic value of the subregion, including: executing the Delaunay subdivision algorithm on the dynamic topology unit to generate a minimum angle division structure containing at least three subregions, including: using the real-time two-dimensional plane coordinates of the three points P1, P2, and P3 in the dynamic topology unit as the subdivision input point set, and constructing an initial convex Envelope boundary, based on the real-time change amplitude of the user's gas consumption time series fluctuation data, when the fluctuation amplitude exceeds the set threshold, an auxiliary vertex Q1 is dynamically inserted at the midpoint of the line connecting P2 and P3. According to the phase offset of the flow phase data, an auxiliary vertex Q2 is inserted at the midpoint of the line connecting P1 and P2. Q1 and Q2 are added to the segmentation point set and the convex hull is updated to obtain the updated segmentation point set. The updated segmentation point set is traversed and a diagonal exchange test is performed on the quadrilateral formed by any four adjacent points. If the minimum internal angle increases after the exchange, the topological connection relationship is updated. The iterative optimization is carried out until all quadrilaterals meet the Delaunay empty circle criterion to generate the final triangulated structure. Each minimum topological unit in the final triangulated structure is defined as a sub-region to obtain a sub-region vertex coordinate set and an adjacency table; for each divided sub-region, the pipeline pressure gradient data associated with the vertex coordinates of the sub-region is extracted, and the flow phase data amplitude characteristics in the same spatial domain are simultaneously obtained; the arithmetic mean of the product of the pressure gradient data and the flow phase amplitude characteristics is calculated, and the basic spatial coupling factor of the sub-region is output; based on the basic spatial coupling factor, the gas consumption time series fluctuation data of the user access point under the jurisdiction of the sub-region is located, the statistical variance value of the user fluctuation data under the jurisdiction in the continuous time window is calculated, and the variance value is linearly normalized in the interval [0, 1] to generate a time series correction coefficient; the basic spatial coupling factor and the time series correction coefficient are added to generate an intermediate coupling feature; the intermediate coupling feature is multiplied by the spatial weight coefficient of the sub-region in the pipeline network topology, and finally the pressure-flow coupling feature value is output; The correction module is used to perform weighted correction on the user gas demand mutation probability characteristic vector using the pressure-flow coupling characteristic value to generate a corrected mutation probability characteristic vector, including: The pressure-flow coupling eigenvalues ​​of each sub-region are input into the matrix converter, sorted by the spatial weight coefficients of the sub-regions in the pipe network topology, and scaled by the length of the pipes under each sub-region as a normalization factor to generate a diagonal correction coefficient matrix with the same dimension as the mutation probability eigenvector. According to the preset user gas demand mutation probability eigenvector, the correction coefficient matrix and the mutation probability eigenvector are subjected to a Hadamard product operation to obtain a weighted eigenvector; Perform L2 norm normalization on the weighted feature vector to obtain the corrected mutation probability feature vector; The processing module is used to perform tensor splicing on the modified mutation probability feature vector and the pipeline equivalent inner diameter dynamic feature vector, and input the result into a fully connected diagnostic network to generate a caliber matching deviation index, including: According to the corrected mutation probability feature vector and the real-time updated dynamic feature vector of the pipeline equivalent inner diameter, it is expanded into a three-dimensional tensor along the feature dimension, where the first dimension is the time step, the second dimension is the spatial partition, and the third dimension is the feature channel. The three-dimensional tensor is input into a three-layer fully connected diagnostic network, where: in the first layer, the linear transformation layer reduces the input feature dimension to 1 / 2 of the original dimension and passes the ReLU activation function; in the second layer, the linear transformation layer further reduces it to 1 / 4 of the original dimension and passes the Sigmoid activation function; in the third layer, the single neuron linear layer outputs the original value of the deviation, and the original value output by the third layer is dynamically threshold calibrated. The Z-score standardized offset of the current value relative to the benchmark distribution is calculated through the benchmark distribution curve fitted with the preset historical normal operating data; when the deviation exceeds the set threshold, the linkage time series prediction network outputs a safe redundant caliber replacement plan.

2. The intelligent analysis and diagnosis system for gas metering instrument caliber matching according to claim 1 is characterized in that: By deploying pressure sensors at the three-dimensional coordinate locations of the gas pipeline network trunk nodes, user access points, and pressure regulating stations, pipeline pressure gradient data, flow phase data, and user gas consumption time series fluctuation data are collected in real time, including: Establish the pipe network topology mapping relationship based on the three-dimensional spatial coordinates of each sensor; Synchronously collect real-time pressure values ​​along the pipeline between trunk nodes, calculate the pressure difference and associated spatial distance between adjacent nodes, and generate pressure gradient data; Monitor the flow sensor output at the user access point, extract the phase offset and amplitude characteristics of the flow waveform within the cycle, and generate flow phase data; The instantaneous fluctuation signal of gas consumption downstream of the pressure regulating station is recorded, and the fluctuation frequency and the absolute difference between the peak and the trough in unit time are calculated to form the time series fluctuation data of gas consumption of users.

3. The intelligent analysis and diagnosis system for gas metering instrument caliber matching according to claim 2 is characterized in that: Based on the pipeline pressure gradient data, flow phase data and user gas consumption time series fluctuation data, the main pipeline intersection point P1, user access point P2 and pressure regulating station output point P3 are selected as three dynamic topology reference points, including: According to the pipe network topology mapping relationship, the main pipe intersection point with the largest pressure change rate in the pressure gradient data is selected as P1, and its three-dimensional space coordinates are used as the initial reference coordinates; Using the initial reference coordinates of the main pipeline intersection point P1 as a reference, combined with the phase offset and amplitude characteristics in the flow phase data, the spatial influence weight of the user access point P2 is calculated. The original three-dimensional coordinates of the user access point P2 are offset compensated based on the spatial influence weight to obtain the weighted real-time monitoring coordinates of the user access point P2. Based on the weighted real-time monitoring coordinates of the user access point P2, the original coordinates of the pressure regulating station output point P3 are compensated for time series displacement using the fluctuation frequency and peak-to-trough difference in the user's gas consumption time series fluctuation data to obtain the real-time control coordinates of the pressure regulating station output point P3. The initial reference coordinates of the main pipeline intersection point P1, the real-time monitoring coordinates of the user access point P2, and the real-time control coordinates of the pressure regulating station output point P3 are spatially synchronized and converted to a two-dimensional plane through dimensionality reduction mapping; the three points P1, P2, and P3 in the two-dimensional plane are connected to form a dynamic topological unit.

4. An intelligent analysis and diagnosis method for gas metering instrument caliber matching, characterized in that: The method is applied to the system according to any one of claims 1 to 3, and the method comprises: Step 1: Use pressure sensors deployed at the three-dimensional coordinate locations of the gas pipeline network trunk nodes, user access points, and pressure regulating stations to collect pipeline pressure gradient data, flow phase data, and user gas consumption time series fluctuation data in real time; Step 2: Based on pipeline pressure gradient data, flow phase data, and user gas consumption time series fluctuation data, the main pipeline intersection point P1, user access point P2, and pressure regulating station output point P3 are selected as three vertices, and a dynamic triangle is formed based on the three vertices. The dynamic triangle is meshed using the Delaunay triangulation algorithm, and the pressure-flow coupling eigenvalue of the mesh is extracted. Step 3: Perform weighted correction on the user gas demand mutation probability characteristic vector using the pressure-flow coupling characteristic value to obtain a corrected mutation probability characteristic vector; In step 4, the corrected mutation probability feature vector is tensor-concatenated with the pipeline equivalent inner diameter dynamic feature vector and input into the fully connected diagnostic network to generate a caliber matching deviation index. When the deviation exceeds the threshold, the linked timing prediction network outputs a safe and redundant caliber replacement plan.

5. A computing device, characterized in that include: one or more processors; A storage device for storing one or more computer programs, wherein when the one or more computer programs are executed by the one or more processors, the one or more processors implement the method according to claim 4.

6. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, so that when the computer program is executed by a processor, the method according to claim 4 is implemented.

Images