Heat supply network model optimization method based on heat supply pipeline network topology transformation

Abstract

The invention relates to the technical field of heat supply network intelligent modeling, and discloses a heat supply network model optimization method based on heat supply pipeline network topological transformation. The method comprises the following steps: receiving an original operation signal from a monitoring platform, and performing credibility screening to obtain a heat supply basic data set; the method comprises the following steps: depicting an actual connection state of a heat supply pipe network in a topological graph form, constructing a basic topological graph, and identifying a pipe network skeleton structure with master-slave hierarchical features from the basic topological graph; based on this, a heat supply pipe network hydraulic operation simulation model is established; during operation, continuously comparing a model prediction value with a measured value, and triggering a pipe network structure re-identification instruction when the deviation continuously exceeds a threshold value; the instruction is executed, the basic topological graph is dynamically updated, the core channel and the standby channel are re-divided, and then the hydraulic simulation model is corrected through the updated pipe network framework structure. According to the invention, the simulation model can automatically identify and adapt to the change of the actual structure of the pipe network, the dynamic self-correction of the model is realized, and the accuracy and reliability of the model in long-term operation are improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent modeling technology for heating pipe networks, specifically a method for optimizing heating network models based on topology transformation of heating pipe networks. Background Technology

[0002] Existing heating network simulation models typically rely on design drawings or one-time surveys to construct a static network topology. These models fix the connections between components such as heating stations and valves, and perform hydraulic and thermal calculations based on this. After the model is put into operation, even if actual observation data deviates significantly from model predictions, conventional optimization methods are limited to adjusting parameters such as pipe resistance coefficients and heat source output, while treating the network's physical connections and functional hierarchy as unchanged.

[0003] This static modeling approach has inherent limitations. In actual operation, heating networks can experience changes in effective delivery paths and functional hierarchy due to valve status switching, pipeline modifications, and changes in user access. Static models cannot detect these actual topological changes, leading to a gradual mismatch between the model and the physical system, and increasing simulation prediction bias. When biases occur, current technologies lack a mechanism to automatically determine whether the bias stems from changes in the underlying network structure, and are unable to dynamically reconstruct the model's basic topology and functional understanding during operation. This makes it difficult for simulation models to accurately serve scheduling decisions over the long term.

[0004] A method is needed to enable heating network models to have self-diagnostic and structural update capabilities. The key lies in how to define the criteria for model inaccuracies and automatically trigger reconstruction, and how to re-identify the actual connection status and core hydraulic paths of the network based on actual operating data during reconstruction, thereby breaking the limitations of static models and achieving synchronization between the model and the dynamic evolution of the physical network. Summary of the Invention

[0005] The purpose of this invention is to provide a heating network model optimization method based on heating pipeline network topology transformation, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, this invention provides a heating network model optimization method based on heating pipeline network topology transformation, the method comprising: The system receives a set of raw operating signals from the heating system monitoring platform. The set of raw operating signals includes time-series data from heat sources, heating stations, pipeline valves, and pressure sensors. The original set of operating signals is subjected to credibility screening to select stable and reliable heating operation signals and remove abnormal signals with missing data or physical contradictions to obtain the basic heating dataset. The actual connection status of the heating network is depicted in the form of a topology diagram, in which heating stations and key valves are graph nodes and connecting pipelines are graph edges, thus constructing the basic topology diagram of the heating network; The core transmission channels with high influence and the secondary backup branch channels are identified in the basic topology map, and a pipeline skeleton structure with master-slave hierarchical characteristics is generated. Based on the heating basic dataset and the pipeline skeleton structure, a hydraulic operation simulation model of the heating pipeline network is established. During the operation of the heating system, the predicted values ​​of the hydraulic operation simulation model of the heating network are continuously compared with the observed values ​​of the heating basic dataset. When the deviation continues to exceed the threshold, a network structure re-identification command is triggered. The pipeline structure re-identification command is executed to dynamically update the basic topology map and re-divide the core delivery channel and the backup branch channel; The updated pipeline framework structure is used to correct the hydraulic operation simulation model of the heating pipeline network, resulting in an updated hydraulic operation simulation model of the heating pipeline network.

[0007] Preferably, the credibility screening includes: Check the continuity of each signal in the original set of operating signals within a preset time window, and mark the signal segments that have interruptions; Verify the physical and logical relationships between signals, including verifying the rationality of upstream and downstream pressure signals in the same pipe section, the rationality of inlet and outlet temperatures of the heating station, and the correspondence between valve opening and flow rate; Signals that do not meet the continuity requirements or physical logic relationships, along with their associated data, are isolated from the heating basic dataset to form a set of abnormal signals to be verified. The signal combinations with temporal and spatial correlations are extracted from the set of abnormal signals to be verified, providing potential abnormal structural features for subsequent topological transformations.

[0008] Preferably, the basic topology diagram for constructing the heating pipe network includes: Based on the heating network design drawings and equipment ledger, establish an initial logical topology diagram that includes all known heating stations, valves and pipe section connections; The equipment identifiers and location information in the heating basic dataset are mapped and verified one by one with the nodes and edges in the initial logical topology diagram. Nodes and edges that have actually been abandoned in the drawing are removed, and edges that are not recorded in the drawing but actually have data interaction are added. Based on the results of verification and supplementation, a basic topology diagram reflecting the actual physical connection status of the heating network is generated.

[0009] Preferably, the generation of the pipeline skeleton structure with master-slave hierarchical characteristics includes: In the basic topology diagram, the cumulative heat load and average flow rate carried by each edge within a preset historical period are calculated; Based on the cumulative heat load and average flow rate, all edges in the basic topology graph are sorted in descending order, and a specific number of edges at the top of the sort are selected as a candidate set of high-influence edges. Analyze the connection relationships of edges in the candidate set of high-influence edges, and identify the combination of mutually connected edges with the same transmission direction as candidate transmission channels; Assess the irreplaceability of each candidate transport channel in the heat medium transport and distribution process, identify the candidate transport channels with irreplaceability higher than a specified value as core transport channels, and classify the remaining candidate transport channels or those not selected as backup branch channels. The core transport channel and the backup branch channel together constitute the pipeline network skeleton structure, wherein the core transport channel constitutes the main level and the backup branch channel constitutes the secondary level.

[0010] Preferably, the establishment of the hydraulic operation simulation model of the heating network includes: Based on the aforementioned pipeline network skeleton structure, a set of basic hydraulic equations describing the flow distribution and pressure distribution in the heating pipeline network is established. Multiple sets of operating data under historical stable operating conditions are extracted from the heating basic dataset and used as input and output samples of the basic hydraulic equations. Using the input and output samples, the pipe network resistance characteristic coefficients in the basic hydraulic equations are identified and calibrated. The calibrated set of basic hydraulic equations, which includes the determination of resistance characteristic coefficients, is established as the hydraulic operation simulation model for the heating network.

[0011] Preferably, the triggering pipeline structure re-identification instruction includes: Real-time acquisition of current observations from the heating system's basic dataset; Input the key state variables from the current observations into the hydraulic operation simulation model of the heating network to obtain the corresponding model prediction values; Calculate the pointwise deviations between the model predictions and the current observations on key state quantities, and perform a moving average of the pointwise deviations within a time window; When the deviation value after moving average continues to exceed the preset deviation threshold, and the covered pipeline area exceeds the preset range threshold, the hydraulic operation simulation model of the heating pipeline is determined to be inaccurate, and a pipeline structure re-identification instruction is generated. The pipeline network structure re-identification instruction includes the identifier of the pipeline network area with excessive deviation, a list of key status quantities, and the duration of the deviation.

[0012] Preferably, the dynamic updating of the basic topology graph includes: Receive the pipeline structure re-identification command and lock the pipeline area with excessive deviation contained in the command; Retrieve the set of abnormal signals to be verified generated in the pipeline network area in the near future; Analyze the spatiotemporal distribution patterns of the signal combinations in the abnormal signal cluster to be verified, and identify signal features that may indicate changes in the physical connection of the pipeline network; Based on the signal characteristics, within the corresponding pipeline area in the basic topology map, the connection relationship between nodes and edges is modified or added / removed to form an update proposal. Based on the latest data from the heating infrastructure dataset, the rationality of the update proposal is verified. Once confirmed, the infrastructure topology map is dynamically updated to generate a revised infrastructure topology map.

[0013] Preferably, the re-division of the core transport channel and the backup branch channel includes: Based on the revised base topology graph, the current operating metrics for each edge in the graph are recalculated; Based on the recalculated current operating indicators, the sorting of edges, identification of candidate transport channels, and evaluation of irreplaceability are performed again. Based on the latest assessment results, the core delivery channels and backup branch channels were redefined in the revised basic topology map, generating an updated pipeline network skeleton structure.

[0014] Preferably, the modification of the hydraulic operation simulation model of the heating network includes: The updated pipeline skeleton structure is compared with the original pipeline skeleton structure to identify the changed edges and nodes. For the changed edge, adjust its corresponding resistance characteristic coefficient or add or remove the corresponding hydraulic equation in the basic hydraulic equation set of the heating network hydraulic operation simulation model. For specific key nodes that have changed, add or modify their boundary condition constraint equations in the hydraulic operation simulation model of the heating network. Using the latest heating data set for the current period, the parameters of the adjusted heating network hydraulic operation simulation model are recalibrated to obtain the updated heating network hydraulic operation simulation model.

[0015] Preferably, the method further includes a verification step for the update process: After obtaining the updated hydraulic operation simulation model of the heating network, new observation data of the heating system operation are collected within an independent verification time window. The new operational observation data is input into the updated hydraulic operation simulation model of the heating network, and the prediction bias of the model on the new data is calculated. If the prediction deviation is lower than the preset verification threshold, the model update is confirmed to be effective. If the prediction deviation is still higher than the preset verification threshold, a new round of pipeline structure re-identification instructions will be triggered to start the iterative optimization process until the prediction deviation meets the requirements.

[0016] Compared with the prior art, the beneficial effects of the present invention are: During operation, the predicted values ​​from the hydraulic simulation model are continuously compared with real-time observations from the monitoring platform, and a threshold condition is set for the deviation to continuously exceed. This effectively distinguishes between short-term deviations caused by random disturbances and systematic, persistent inaccuracies caused by actual changes in the pipeline network structure. Only the latter triggers a re-identification command, avoiding frequent and unnecessary model reconstruction due to data noise, ensuring the necessity and accuracy of model updates, and achieving intelligent judgment of the root causes of model failure.

[0017] When the re-identification command is triggered, the method does not perform parameter correction on the original fixed topology. Instead, it dynamically updates the basic topology map depicting the pipeline network connections. Based on the updated topology and actual operational data, it re-divides the core channels and secondary backup branches in the pipeline network that actually undertake the main transportation tasks. This allows the model to automatically detect and adapt to changes in the actual hydraulic paths and functional levels caused by valve operations, pipeline modifications, etc. By correcting the model's most fundamental topological framework, the simulation model maintains consistency with the current real structure of the physical system, improving the model's long-term applicability and simulation accuracy under scenarios of dynamic changes in the pipeline network structure. Attached Figure Description

[0018] Figure 1 This is a schematic diagram illustrating the working principle of the heating network model optimization method based on heating pipeline network topology transformation as described in this invention. Figure 2 A flowchart for credibility screening; Figure 3 A flowchart for generating the pipeline network skeleton structure; Figure 4 A bar chart comparing the flow rates at key edges of the heating network; Figure 5 This is a radar chart of a predictive model for a thermal system. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see Figure 1This invention provides a method for optimizing a heating network model based on the topology transformation of a heating pipeline network. The method includes: receiving a set of raw operating signals from a heating system monitoring platform, which contains time-series data from heat sources, heating stations, pipeline valves, and pressure sensors; performing reliability screening on the raw operating signal set to select stable and reliable heating operating signals and eliminating abnormal signals with missing data or physical inconsistencies to obtain a basic heating dataset; depicting the actual connection state of the heating pipeline network in the form of a topology graph, using heating stations and key valves as graph nodes and connecting pipelines as graph edges to construct a basic topology graph of the heating pipeline network; identifying high-influence core transmission channels and secondary backup branch channels in the basic topology graph to generate a pipeline network skeleton structure with master-slave hierarchical characteristics; and establishing a hydraulic operation simulation model of the heating pipeline network based on the basic heating dataset and the pipeline skeleton structure. During the operation of the heating system, continuously comparing the predicted values ​​of the hydraulic operation simulation model with the observed values ​​of the basic heating dataset, and triggering a pipeline structure re-identification command when the deviation continuously exceeds a threshold. The pipeline structure re-identification command is executed to dynamically update the basic topology map and re-divide the core delivery channels and backup branch channels. Using the updated pipeline skeleton structure, the hydraulic operation simulation model of the heating pipeline network is corrected to obtain the updated hydraulic operation simulation model of the heating pipeline network.

[0021] In one embodiment of the present invention, see [reference] Figure 2 The credibility screening process includes checking the continuity of each signal in the original operating signal set within a preset time window and marking signal segments with interruptions. It verifies the physical and logical relationships between signals, including verifying the rationality of upstream and downstream pressure signals within the same pipe segment, the rationality of inlet and outlet temperatures at heating stations, and the correspondence between valve opening and flow rate. Signals that do not meet continuity requirements or physical and logical relationships, along with their associated data, are isolated from the heating network's basic dataset, forming a set of abnormal signals to be verified. Signal combinations with temporal and spatial correlations are extracted from this set to provide potential abnormal structural features for subsequent topology transformation. The construction of the basic topology map of the heating network involves establishing an initial logical topology map containing all known heating stations, valves, and pipe segment connections based on the heating network design drawings and equipment ledger. Equipment identifiers and location information from the heating network's basic dataset are mapped and verified one-to-one with nodes and edges in the initial logical topology map. Nodes and edges that are actually obsolete in the drawings are removed, and edges that are not recorded in the drawings but actually have data interaction are added. Based on the verification and addition results, a basic topology map reflecting the true physical connection status of the heating network is generated.

[0022] In the specific implementation, for a heating network area containing multiple heating stations and valve nodes, reliability screening and construction of the basic topology map of the heating network were carried out. The original operating signal set included the time series data of the heat source outlet pressure sensor PS-101, the inlet pressure sensor PIT-05 and outlet temperature sensor TIT-052 of the heating station HS-05, the opening signal AO-12 of pipeline valve V-12, and the associated flow meter FIQ-07. The preset time window was set to a continuous 24 hours. Item-by-item inspection revealed that the signal value of flow meter FIQ-07 remained at zero from 2:00 AM to 3:00 AM on the same day, while the opening signal AO-12 of upstream pressure sensor PIT-05 and downstream valve V-12 both showed normal fluctuations during the same period. This signal segment of the hour was marked as a signal segment with an interruption. During the verification of physical logic relationships, it was found that the opening signal AO-12 of pipeline valve V-12 showed 80% at 10:00 AM on the same day, but the instantaneous flow value of its associated flow meter FIQ-07 was lower than 30% of the historical average flow rate at the same opening. This correspondence was deemed unreasonable. Signal segments with interruptions and signals that did not meet physical logic relationships, along with their associated data, were isolated from the heating system's basic dataset, forming a set of abnormal signals to be verified. From this set of abnormal signals, the abnormal patterns jointly presented by valve V-12, flow meter FIQ-07, and pressure signals from adjacent pipe sections within a specific time period of the day were extracted. This combination of signals with temporal and spatial correlation was recorded as a potential abnormal structural feature.

[0023] In some embodiments, the basic topology map of the heating network is constructed based on the heating network design drawings and equipment ledgers for the region. An initial logical topology map is established, which includes heating stations HS-05 and HS-06, valves V-12 and V-13, and the pipe segments connecting these nodes. The identifiers and spatial coordinate information of each device in the heating basic dataset are mapped and verified with the nodes and edges in the initial logical topology map. Verification reveals that heating station HS-08, recorded in the design drawings, has no pressure, temperature, or flow signals in any historical or current data. Therefore, the node HS-08 and its connecting edges are determined to be actually abandoned nodes and edges in the initial logical topology map and are removed. Meanwhile, the data mapping process revealed that the pressure signals PIT-05 and PIT-07 of heating stations HS-05 and HS-07 exhibited a high degree of synchronicity and strong correlation in their dynamic changes. However, the initial logical topology diagram did not record a direct pipe connection between the two stations. Therefore, an edge connecting heating stations HS-05 and HS-07, which was not recorded in the drawings but actually involved data interaction, was added. The basic topology diagram was generated based on the verified and supplemented results, reflecting the true state of the physical connections of the heating network in this area. The corrected connection relationships provided an accurate structural basis for subsequent analysis.

[0024] Optionally, a flow estimation model can be introduced to assist in verifying the correspondence between valve opening and flow rate. The specific implementation of the flow estimation model in verifying this correspondence is as follows: First, the valve flow coefficient is determined based on the characteristic curve or historical operating data provided by the valve manufacturer. Then, the flow area ratio is obtained by mapping the real-time collected valve opening signal value through a preset proportional function. Simultaneously, the pressure difference data measured by pressure sensors installed upstream and downstream of the valve is acquired. Based on these parameters, an estimated volumetric flow rate is calculated, and this estimated value is continuously compared with the actual observed value of the flow meter. When the deviation exceeds the threshold range set according to historical normal operating conditions, and external factors such as sensor malfunction or communication interruption have been checked and ruled out, it is determined that there is a physical and logical contradiction in the correspondence between valve opening and flow rate. The relevant signals are then included in the set of abnormal signals to be verified for further analysis. The flow estimation model is established based on the valve characteristic curve and the upstream and downstream pressure difference, and its expression is: ;

[0025] in: This represents the estimated volumetric flow rate. This represents the valve flow coefficient. This represents the valve opening signal value. This represents the proportional function between valve opening and flow area. This represents the measured pressure difference across the valve. It can be understood that, in practical implementation, the observed flow rate value of the flow meter FIQ-07... Calculated with the flow estimation model If the deviation under the same operating conditions continues to exceed the preset range, and the possibility of instrument failure has been ruled out, the correspondence between the valve opening signal AO-12 and the observed flow value of the flow meter FIQ-07 is determined to be a physical and logical contradiction, and the relevant signals are classified into the abnormal signal set to be checked.

[0026] In one embodiment of the present invention, see [reference] Figure 3The process of generating a pipeline network skeleton structure with master-slave hierarchical characteristics includes: calculating the cumulative heat load and average flow rate of each edge within a preset historical period in the basic topology graph; sorting all edges in the basic topology graph in descending order based on the cumulative heat load and average flow rate, and selecting a specific number of edges at the top of the sort as a candidate set of high-influence edges; analyzing the connectivity relationships of edges in the candidate set of high-influence edges, and identifying interconnected edges with the same transport direction as candidate transport channels; evaluating the irreplaceability of each candidate transport channel in the heat medium distribution process, and identifying candidate transport channels with irreplaceability higher than a specified value as core transport channels, while other candidate transport channels or unselected edges are classified as backup branch channels. The core transport channels and backup branch channels together constitute the pipeline network skeleton structure, with the core transport channels forming the master level and the backup branch channels forming the slave level. The establishment of a hydraulic operation simulation model for the heating pipeline network includes establishing a set of basic hydraulic equations describing the flow distribution and pressure distribution in the heating pipeline network based on the pipeline network skeleton structure. Multiple sets of operational data under historical stable operating conditions are extracted from the heating system's basic dataset and used as input and output samples for the fundamental hydraulic equations. Using these input and output samples, the network resistance characteristic coefficients in the fundamental hydraulic equations are identified and calibrated. The calibrated fundamental hydraulic equations, containing the determined resistance characteristic coefficients, are then established as the hydraulic operation simulation model for the heating network.

[0027] In practical implementation, the generation of a pipeline network skeleton structure with master-slave hierarchical characteristics and the establishment of a hydraulic operation simulation model for the heating pipeline network are based on the existing basic topology map of the heating pipeline network. The preset historical period is set to the previous complete heating season. In practical implementation, for each edge in the basic topology map, the cumulative heat load and average flow rate it carries within the preset historical period are calculated. The cumulative heat load is obtained by integrating the product of the heat flow rate associated with that edge and time. Based on the two indicators of cumulative heat load and average flow rate, all edges in the basic topology map are sorted in descending order, and the edges with the highest ranking are selected to form a candidate set of high-influence edges. For example, the edges with the highest combined cumulative heat load and average flow rate are selected. The connection relationships of the edges in the candidate set of high-influence edges are analyzed, and the combination of edges that are interconnected and have the same transport direction is identified as a candidate transport channel. For example, a series of continuous edges from the main outlet pipe of the heat source plant to the inlet of regional heat exchange station A are identified as a candidate transport channel.

[0028] In some embodiments, the irreplaceability of each candidate transport channel in the heat medium distribution process is evaluated. The evaluation process considers the additional transport capacity and system pressure drop changes that the backup path needs to meet when the candidate transport channel is hypothetically shut down. Candidate transport channels with irreplaceability higher than a specified value are identified as core transport channels, for example, a specified value of 0.8. Channels with an irreplaceability higher than this value are considered core transport channels, and other candidate transport channels or edges not selected into the high-influence candidate set are classified as backup branch channels. Core transport channels and backup branch channels together constitute the pipeline network skeleton structure, with core transport channels forming the main level and backup branch channels forming the subordinate level. A hydraulic operation simulation model of the heating pipeline network is established based on the pipeline network skeleton structure. A set of basic hydraulic equations describing the flow distribution and pressure distribution in the heating pipeline network is established. The set of basic hydraulic equations includes nodal flow continuity equations and loop pressure balance equations.

[0029] Optionally, when calculating the cumulative heat load, an integral formula is used for quantification. The expression of the integral formula is as follows: ;

[0030] in: Represents the time interval The cumulative heat load within, The density of the heat transfer medium Specific heat capacity, representing the heat transfer medium Represents the volumetric flow rate that changes over time. This represents the temperature difference between the supply and return water as it changes over time. Multiple sets of operational data under historical stable operating conditions are extracted from the heating system's basic dataset and used as input and output samples for the fundamental hydraulic equations. The input samples include the outlet pressure of each heat source and the valve opening of each heating station; the output samples include the observed pressure at each node and the flow rate of each pipe segment. It can be understood that the network resistance characteristic coefficients in the fundamental hydraulic equations are identified and calibrated using these input and output samples. These coefficients include the pipe friction coefficient and the local resistance coefficient. The calibration process is completed by solving an optimization problem that minimizes the deviation between the model's calculated values ​​and the observed values. The calibrated fundamental hydraulic equations, containing the determined resistance characteristic coefficients, are then established as the hydraulic operation simulation model of the heating network.

[0031] In one embodiment of the present invention, triggering the pipeline structure re-identification instruction includes acquiring the current observation values ​​of the heating basic dataset in real time. Key state variables from the current observation values ​​are input into the heating pipeline hydraulic operation simulation model to obtain the corresponding model prediction values. The point-by-point deviation between the model prediction values ​​and the current observation values ​​on the key state variables is calculated, and a moving average is performed on the point-by-point deviation within a time window. When the deviation value after moving average continuously exceeds a preset deviation threshold, and the covered pipeline area exceeds a preset range threshold, the heating pipeline hydraulic operation simulation model is determined to be inaccurate, and a pipeline structure re-identification instruction is generated. The pipeline structure re-identification instruction includes the identifier of the pipeline area with excessive deviation, a list of key state variables, and the duration of the deviation.

[0032] In practice, the process of triggering the pipeline structure re-identification command involves continuously comparing the predicted output of the heating pipeline hydraulic operation simulation model with the actual observation data. The current observation values ​​of the heating basic dataset are acquired in real time. These values ​​include the real-time reading of the inlet pressure sensor PIT-03 (HS-03) at heating station (0.85 MPa), the real-time reading of the inlet pressure sensor PIT-07 (HS-07) at heating station (0.62 MPa), and the real-time reading of the main pipe pressure sensor PM-12 (1.05 MPa). Key state variables from the current observation values ​​are input into the heating pipeline hydraulic operation simulation model. These key state variables include the setpoints for the outlet pressure of each heat source and the opening settings of the main valves, resulting in corresponding model prediction values. These prediction values ​​include predicted pressures for PIT-03, PIT-07, and PM-12, which are 0.82 MPa, 0.60 MPa, and 1.02 MPa, respectively.

[0033] The model calculates the point-by-point deviations between predicted values ​​and current observations on key state quantities: 0.03 MPa for PIT-03, 0.02 MPa for PIT-07, and 0.03 MPa for PM-12. A moving average is calculated over a 30-minute time window, with the moving average calculated every minute, generating a series of continuous moving average deviation values. If the moving average deviation consistently exceeds a preset deviation threshold (e.g., a pressure deviation threshold of 0.025 MPa) and covers a network area exceeding a preset range threshold (e.g., an area involving three or more adjacent monitoring points), the heating network hydraulic operation simulation model is deemed inaccurate. Generate a pipeline structure re-identification instruction. The pipeline structure re-identification instruction includes the pipeline area identifier that exceeds the deviation limit, a list of key status quantities, and the duration of the deviation. The pipeline area identifier is "North Main Pipeline". The list of key status quantities includes PIT-03, PIT-07, and PM-12. The duration of the deviation is 45 minutes.

[0034] In some embodiments, the point-by-point deviation is calculated using the absolute difference form. An optional method for calculating the moving average is to take the arithmetic mean of the deviation values ​​at all sampling points within a time window; the moving average deviation value... Calculated using the following formula: ;

[0035] in: This represents the moving average deviation value calculated at time t, where N represents the number of sampling points included in the time window. Represents the moment The observed values, Represents the moment The model prediction value, where t represents the current time index. It can be understood that, in practical implementation, when calculating the results for sensor PM-12... The pressure was greater than 0.025 MPa for 45 consecutive minutes, and the pressure of adjacent sensors PIT-03 and PIT-07 was also greater during the same period. It is also greater than 0.025 MPa, meeting the conditions of continuous deviation exceeding the limit and coverage area exceeding the range. The logic for triggering the pipeline structure re-identification command is activated, and the generated pipeline structure re-identification command indicates that the system has identified existing pipeline structure changes or model mismatches.

[0036] In one embodiment of the present invention, dynamically updating the basic topology map includes receiving a pipeline structure re-identification command and locking the pipeline area containing deviations exceeding limits. A set of anomaly signals to be verified recently generated for that pipeline area is retrieved. The spatiotemporal distribution pattern of signal combinations in the set of anomaly signals to be verified is analyzed to identify signal features indicating changes in the physical connections of the pipeline. Based on the signal features, the connection relationships between nodes and edges are modified or added / removed within the corresponding pipeline area in the basic topology map to form an update proposal. The rationality of the update proposal is verified based on the latest time period data in the heating basic dataset. After confirmation, the basic topology map is dynamically updated to generate a revised basic topology map. Re-dividing the core delivery channel and backup branch channels includes recalculating the current operating indicators of each edge in the revised basic topology map. Based on the recalculated current operating indicators, the edges are sorted again, candidate delivery channels are identified, and their irreplaceability is evaluated. Based on the latest evaluation results, the core delivery channel and backup branch channel are re-determined in the revised basic topology map to generate an updated pipeline skeleton structure.

[0037] In practice, the dynamic updating and re-division of the core transmission channels and backup branch channels of the basic topology map begins with receiving the pipeline structure re-identification command. The pipeline structure re-identification command indicates that the pipeline area with excessive deviation is the "Northern Main Network and Branch Connection Point". After locking the pipeline area of ​​the "Northern Main Network and Branch Connection Point", the set of abnormal signals to be verified generated in this pipeline area in the near future is retrieved. The set of abnormal signals to be verified includes a combination of signals with spatiotemporal correlation, such as abnormal fluctuations in the opening signal of valve V-08, the reading of the branch flowmeter FIQ-04 connected to the heating station HS-04 returning to zero, and the abnormal increase in the pressure value of the pressure sensor PM-05 of the adjacent main pipeline. By analyzing the spatiotemporal distribution pattern of the signal combinations in the set of abnormal signals to be verified, the signal characteristics indicating changes in the physical connection of the pipeline are identified. The signal characteristics are that when the flow of a branch pipeline returns to zero, the pressure of its upstream main pipeline increases abnormally, and the flow of another adjacent branch pipeline increases synchronously. This pattern suggests that a bypass pipeline has been activated or an existing branch pipeline has been closed. Based on signal characteristics, within the pipeline area corresponding to the "Northern Main Network and Branch Connection Point" in the basic topology map, the connection relationships between nodes and edges are modified to form an update proposal. The update proposal adds a new connection edge E-new between the main pipeline node N-05 and the heating station HS-04, and marks the original connection edge E-old as "pending confirmation." The rationality of the update proposal is verified using the latest time period data in the heating basic dataset. The latest time period data shows that the pressure difference and flow rate relationship of the virtual pipe segment corresponding to the newly added connection edge E-new conforms to hydraulic characteristics. After confirmation, the basic topology map is dynamically updated to generate a revised basic topology map.

[0038] In some embodiments, the reclassification of core transport channels and backup branch channels is performed based on the revised base topology. Based on the revised base topology, the current operating indicators for each edge in the graph are recalculated, including the average flow rate and average load rate over the past 24 hours. Based on the recalculated current operating indicators, the edges are re-sorted, candidate transport channels are identified, and their irreplaceability is assessed. Based on the latest assessment results, the core transport channels and backup branch channels are redefined in the revised base topology. An updated pipeline network skeleton structure is generated. Refer to Table 1 for a comparison of edge operating indicators before and after the update.

[0039] Table 1: Changes in operational metrics and hierarchical levels before and after the critical edge update

[0040] Optionally, a comprehensive metric can be used when recalculating the current operating metrics of the edges. To simultaneously reflect the stability of flow rate and load, the relationship is as follows: ;

[0041] in: The current comprehensive indicators representing the edge's operation. This represents the average flow of the edge during the calculation period. This represents the standard deviation of the flow of the edge within the calculation period. This represents the average load factor of the edge during the calculation period. This is understandable. Edges with higher values ​​indicate larger and less volatile transport volumes and higher load levels, thus receiving higher priority in the sorting process. This is based on the recalculated current operating parameters. The edges are sorted in descending order, and the sorting result directly affects the identification of candidate transport channels and subsequent irreplaceability assessment. During the irreplaceability assessment, candidate transport channels are simulated for closure, and the minimum additional pump power increment required by the system to meet the same heat load demand is calculated. Increment The larger the value, the higher the irreplaceability of the candidate transmission channel, and the more certain it is to be the new core transmission channel. After evaluation, the newly added edge E-new, due to taking over the traffic of the original edge E-old, and... The high value indicates that it has been identified as a core transport channel in the updated pipeline skeleton structure, while the original E-old has been removed from the pipeline skeleton structure due to its zero flow.

[0042] See Figure 4 This is a bar chart comparing the flow rates of key edges in a heating network, clearly showing the average flow rate changes of different edge identifiers before and after the update. This chart corresponds to the flow distribution adjustment after the heating network topology update, reflecting the optimization logic of the network's backbone structure. The new edge E-new replaces the original edge E-old, achieving a smooth flow transfer; the flow rate of the core channel E-main decreases slightly, indicating that the new edge shares some of the load; the flow rate of the standby channel E-alt increases, reflecting the coordinated adjustment of the network's standby branches. This type of chart is typically used to verify the effectiveness of heating network topology optimization, helping maintenance personnel intuitively judge the rationality of the network's flow distribution, and is one of the core visualization tools for assessing the operational status of thermal systems.

[0043] In one embodiment of the present invention, modifying the hydraulic operation simulation model of the heating network includes comparing the updated network skeleton structure with the original network skeleton structure to identify the changed edges and nodes. For the changed edges, the corresponding resistance characteristic coefficients are adjusted or corresponding hydraulic equations are added or removed from the basic hydraulic equation set of the heating network hydraulic operation simulation model. For the changed specific key nodes, their boundary condition constraint equations are added or modified in the heating network hydraulic operation simulation model. Using the latest heating basic dataset, the adjusted heating network hydraulic operation simulation model is recalibrated to obtain the updated heating network hydraulic operation simulation model. The verification step of the update process includes, after obtaining the updated heating network hydraulic operation simulation model, collecting new heating system operation observation data within an independent verification time window. The new operation observation data is input into the updated heating network hydraulic operation simulation model, and the prediction deviation of the model on the new data is calculated. If the prediction deviation is lower than a preset verification threshold, the model update is confirmed to be effective. If the prediction deviation is still higher than the preset verification threshold, a new round of pipeline structure re-identification instructions will be triggered to start the iterative optimization process until the prediction deviation meets the requirements.

[0044] In practical implementation, the modification of the hydraulic operation simulation model of the heating network involves comparing the updated network skeleton structure with the original network skeleton structure to identify the changed edges and nodes. The updated network skeleton structure includes the newly added core transmission channel edge E-new, while the original network skeleton structure includes the deactivated edge E-old. The identified changed edges include the newly added edge E-new and the removed edge E-old. The changed nodes involve nodes N-05 and HS-04 connected to the newly added edge E-new. For the changed edges, the corresponding resistance characteristic coefficients are adjusted or corresponding hydraulic equations are added or removed from the basic hydraulic equation set of the heating network hydraulic operation simulation model. For the newly added edge E-new, an equation describing its flow rate and pressure difference relationship is added to the basic hydraulic equation set. For the removed edge E-old, its corresponding equation is commented out or deleted from the basic hydraulic equation set. For specific key nodes that have undergone changes, their boundary condition constraint equations are added or modified in the hydraulic operation simulation model of the heating network. Node N-05, having changed from connecting a single path to connecting a branch path, has its flow balance equation modified. Using the latest heating data set, the adjusted hydraulic operation simulation model of the heating network is recalibrated to obtain an updated model. Parameter recalibration primarily involves re-identifying the resistance characteristic coefficients of the newly added edge E-new and the resistance characteristic coefficients of the affected adjacent pipe sections.

[0045] In some embodiments, the parameter recalibration process is implemented by solving an optimization problem. The optimization objective is to minimize the overall deviation between the model predictions and the centrally observed values ​​in the latest heating baseline dataset. An optional deviation metric is to calculate the sum of squared residuals of the predicted pressure and flow rates at each monitoring point. The parameter recalibration process utilizes an objective function. To guide, objective function The expression is: ; in: The objective function value that needs to be minimized during the parameter recalibration process is represented by... This represents the number of monitoring points used for calibration. Representing the Pressure observation values ​​at each monitoring point The updated simulation model of the hydraulic operation of the heating network represents the first Calculated pressure values ​​at each monitoring point Representing the Flow observation values ​​at each monitoring point The updated simulation model of the hydraulic operation of the heating network represents the first Calculated flow rate at each monitoring point Representing the first The weighting coefficients for the pressure and flow residuals at each monitoring point. The adjusted hydraulic operation simulation model of the heating network adjusts its internal set of resistance characteristic coefficients. To make the objective function The value reaches its minimum.

[0046] The verification steps of the update process are performed after obtaining the updated hydraulic simulation model of the heating network. After obtaining the updated hydraulic simulation model, new operational observation data of the heating system are collected within an independent verification time window. This independent verification time window is a continuous 8 hours following the period used for model correction and parameter recalibration. The new operational observation data is input into the updated hydraulic simulation model of the heating network, and the prediction deviation of the model on the new data is calculated. The prediction deviation is the root mean square error (RMSE) calculated for key state variables. If the prediction deviation is lower than a preset verification threshold, for example, the RMS error of pressure prediction is lower than 0.015 MPa and the RMS error of flow prediction is lower than 2% of the rated flow, then the model update is confirmed to be effective. If the prediction deviation is still higher than the preset verification threshold, a new round of network structure re-identification is triggered, initiating the iterative optimization process. The iterative optimization process restarts from identifying potential structural changes until the prediction deviation meets the requirements in the new round of verification. In some embodiments, when the initial verification of the prediction bias does not meet the requirements, the analysis time range of the set of anomalous signals to be checked is expanded, and it is checked whether there are other uncaptured related signal changes, thereby generating more accurate update proposals in subsequent iterations.

[0047] See Figure 5 This is a radar chart of a thermal system prediction model, used for multi-dimensional performance evaluation. The shape of the radar chart indicates that this model performs best in the pressure prediction accuracy dimension; the small differences in scores across dimensions suggest that the model performs well in terms of prediction accuracy, robustness, and response performance, with no obvious weaknesses; the overall score is in the range of 0.90 to 0.94, classifying it as a high-performing thermal system prediction model. Radar charts are a common tool for multi-dimensional performance evaluation. In heating systems, these charts are mainly used to compare the overall performance of different prediction models; identify the model's strengths and areas for improvement; and visually demonstrate the model's reliability to maintenance personnel, providing a basis for system decision-making.

[0048] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0049] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A heating network model optimization method based on heating pipeline network topology transformation, characterized in that, The method includes: The system receives a set of raw operating signals from the heating system monitoring platform. The set of raw operating signals includes time-series data from heat sources, heating stations, pipeline valves, and pressure sensors. The original set of operating signals is subjected to credibility screening to select stable and reliable heating operation signals and remove abnormal signals with missing data or physical contradictions to obtain the basic heating dataset. The actual connection status of the heating network is depicted in the form of a topology diagram, in which heating stations and key valves are graph nodes and connecting pipelines are graph edges, thus constructing the basic topology diagram of the heating network; The core transmission channels with high influence and the secondary backup branch channels are identified in the basic topology map, and a pipeline skeleton structure with master-slave hierarchical characteristics is generated. Based on the heating basic dataset and the pipeline skeleton structure, a hydraulic operation simulation model of the heating pipeline network is established. During the operation of the heating system, the predicted values ​​of the hydraulic operation simulation model of the heating network are continuously compared with the observed values ​​of the heating basic dataset. When the deviation continues to exceed the threshold, a network structure re-identification command is triggered. The pipeline structure re-identification command is executed to dynamically update the basic topology map and re-divide the core delivery channel and the backup branch channel; The updated pipeline framework structure is used to correct the hydraulic operation simulation model of the heating pipeline network, resulting in an updated hydraulic operation simulation model of the heating pipeline network.

2. The heating network model optimization method based on heating pipeline network topology transformation according to claim 1, characterized in that, The credibility screening includes: Check the continuity of each signal in the original set of operating signals within a preset time window, and mark the signal segments that have interruptions; Verify the physical and logical relationships between signals, including verifying the rationality of upstream and downstream pressure signals in the same pipe section, the rationality of inlet and outlet temperatures of the heating station, and the correspondence between valve opening and flow rate; Signals that do not meet the continuity requirements or physical logic relationships, along with their associated data, are isolated from the heating basic dataset to form a set of abnormal signals to be verified. The signal combinations with temporal and spatial correlations are extracted from the set of abnormal signals to be verified, providing potential abnormal structural features for subsequent topological transformations.

3. The heating network model optimization method based on heating pipeline network topology transformation according to claim 2, characterized in that, The basic topology diagram for constructing the heating pipeline network includes: Based on the heating network design drawings and equipment ledger, establish an initial logical topology diagram that includes all known heating stations, valves and pipe section connections; The equipment identifiers and location information in the heating basic dataset are mapped and verified one by one with the nodes and edges in the initial logical topology diagram. Nodes and edges that have actually been abandoned in the drawing are removed, and edges that are not recorded in the drawing but actually have data interaction are added. Based on the results of verification and supplementation, a basic topology diagram reflecting the actual physical connection status of the heating network is generated.

4. The heating network model optimization method based on heating pipeline network topology transformation according to claim 1, characterized in that, The generation of the pipeline skeleton structure with master-slave hierarchical characteristics includes: In the basic topology diagram, the cumulative heat load and average flow rate carried by each edge within a preset historical period are calculated; Based on the cumulative heat load and average flow rate, all edges in the basic topology graph are sorted in descending order, and a specific number of edges at the top of the sort are selected as a candidate set of high-influence edges. Analyze the connection relationships of edges in the candidate set of high-influence edges, and identify the combination of mutually connected edges with the same transmission direction as candidate transmission channels; Assess the irreplaceability of each candidate transport channel in the heat medium transport and distribution process, identify the candidate transport channels with irreplaceability higher than a specified value as core transport channels, and classify the remaining candidate transport channels or those not selected as backup branch channels. The core transport channel and the backup branch channel together constitute the pipeline network skeleton structure, wherein the core transport channel constitutes the main level and the backup branch channel constitutes the secondary level.

5. The heating network model optimization method based on heating pipeline network topology transformation according to claim 4, characterized in that, The establishment of the hydraulic operation simulation model for the heating network includes: Based on the aforementioned pipeline network skeleton structure, a set of basic hydraulic equations describing the flow distribution and pressure distribution in the heating pipeline network is established. Multiple sets of operating data under historical stable operating conditions are extracted from the heating basic dataset and used as input and output samples of the basic hydraulic equations. Using the input and output samples, the pipe network resistance characteristic coefficients in the basic hydraulic equations are identified and calibrated. The calibrated set of basic hydraulic equations, which includes the determination of resistance characteristic coefficients, is established as the hydraulic operation simulation model for the heating network.

6. The heating network model optimization method based on heating pipeline network topology transformation according to claim 5, characterized in that, The triggering command for re-identifying the pipeline structure includes: Real-time acquisition of current observations from the heating system's basic dataset; Input the key state variables from the current observations into the hydraulic operation simulation model of the heating network to obtain the corresponding model prediction values; Calculate the pointwise deviations between the model predictions and the current observations on key state quantities, and perform a moving average of the pointwise deviations within a time window; When the deviation value after moving average continues to exceed the preset deviation threshold, and the covered pipeline area exceeds the preset range threshold, the hydraulic operation simulation model of the heating pipeline is determined to be inaccurate, and a pipeline structure re-identification instruction is generated. The pipeline network structure re-identification instruction includes the identifier of the pipeline network area with excessive deviation, a list of key status quantities, and the duration of the deviation.

7. The heating network model optimization method based on heating pipeline network topology transformation according to claim 6, characterized in that, The dynamic updating of the basic topology graph includes: Receive the pipeline structure re-identification command and lock the pipeline area with excessive deviation contained in the command; Retrieve the set of abnormal signals to be verified generated in the pipeline network area in the near future; Analyze the spatiotemporal distribution patterns of the signal combinations in the abnormal signal cluster to be verified, and identify signal features that may indicate changes in the physical connection of the pipeline network; Based on the signal characteristics, within the corresponding pipeline area in the basic topology map, the connection relationship between nodes and edges is modified or added / removed to form an update proposal. Based on the latest data from the heating infrastructure dataset, the rationality of the update proposal is verified. Once confirmed, the infrastructure topology map is dynamically updated to generate a revised infrastructure topology map.

8. The heating network model optimization method based on heating pipeline network topology transformation according to claim 7, characterized in that, The re-division of the core transport channel and the backup branch channel includes: Based on the revised base topology graph, the current operating metrics for each edge in the graph are recalculated; Based on the recalculated current operating indicators, the sorting of edges, identification of candidate transport channels, and evaluation of irreplaceability are performed again. Based on the latest assessment results, the core delivery channels and backup branch channels were redefined in the revised basic topology map, generating an updated pipeline network skeleton structure.

9. The heating network model optimization method based on heating pipeline network topology transformation according to claim 8, characterized in that, The correction of the hydraulic operation simulation model of the heating network includes: The updated pipeline skeleton structure is compared with the original pipeline skeleton structure to identify the changed edges and nodes. For the changed edge, adjust its corresponding resistance characteristic coefficient or add or remove the corresponding hydraulic equation in the basic hydraulic equation set of the heating network hydraulic operation simulation model. For specific key nodes that have changed, add or modify their boundary condition constraint equations in the hydraulic operation simulation model of the heating network. Using the latest heating data set for the current period, the parameters of the adjusted heating network hydraulic operation simulation model are recalibrated to obtain the updated heating network hydraulic operation simulation model.

10. The heating network model optimization method based on heating pipeline network topology transformation according to claim 9, characterized in that, The method also includes a verification step for the update process: After obtaining the updated hydraulic operation simulation model of the heating network, new observation data of the heating system operation are collected within an independent verification time window. The new operational observation data is input into the updated hydraulic operation simulation model of the heating network, and the prediction bias of the model on the new data is calculated. If the prediction deviation is lower than the preset verification threshold, the model update is confirmed to be effective. If the prediction deviation is still higher than the preset verification threshold, a new round of pipeline structure re-identification instructions will be triggered to start the iterative optimization process until the prediction deviation meets the requirements.

Images