Novel heat supply network hydraulic simulation method and system based on large model and medium

Abstract

The invention provides a novel heat supply network hydraulic simulation method and system based on a large model and a medium. The method comprises the steps that a historical operation fusion feature vector training model is constructed, an optimized adjacent matrix is generated according to real-time GIS topological data, a roughness distribution matrix of the inner wall of the heat supply pipe network is obtained by processing through a preset pipe network roughness recognition model according to an image of the inner wall of the heat supply pipe network in combination with a real-time operation fusion feature vector, and the roughness distribution matrix of the inner wall of the heat supply pipe network is obtained. The method comprises the steps of obtaining an optimized adjacency matrix, generating hydraulic operation prediction data of a heat exchange station through processing of a preset heat supply hydraulic prediction model of the heat exchange station, generating heat supply network hydraulic distribution data through processing of a preset heat supply network hydraulic prediction model in combination with the obtained optimized adjacency matrix, and finally performing visual rendering processing to generate a heat supply network hydraulic simulation scene graph. Outputting to a heat supply network dispatching center for displaying; through data fusion, topological optimization, microcosmic characteristic extraction, mesoscopic prediction, macroscopic distribution and visual output processing, heat supply network hydraulic simulation based on the large model is realized.

Description

Technical Field

[0001] This application relates to the field of heating system technology, and more specifically, to a novel hydraulic simulation method, system, and medium for heating networks based on a large model. Background Technology

[0002] Existing hydraulic simulation technologies for heating networks mainly rely on fixed formulas (such as the Darcy-Weisbach equation), which are difficult to adapt to the dynamic changes in roughness caused by pipe network aging. Data-driven models often use single-modal data (such as flow rate data alone), ignoring the implicit characteristics of pipe network vibration and noise. These technologies have significant limitations when facing complex heating network structures and variable operating conditions, resulting in large deviations between simulation results and actual operation. Furthermore, they fail to fully utilize the data analysis and prediction capabilities of cutting-edge large-scale modeling technologies, making breakthroughs in model adaptability, simulation accuracy, and real-time performance difficult. Therefore, there is an urgent need to construct a novel hydraulic simulation method for heating networks that integrates multimodal data, supports dynamic topology, and possesses cross-scale simulation capabilities.

[0003] Effective technical solutions are urgently needed to address the above problems. Summary of the Invention

[0004] The purpose of this application is to provide a novel hydraulic simulation method, system, and medium for heating networks based on a large model. This method can achieve hydraulic simulation of heating networks based on a large model through data fusion, topology optimization, microscopic characteristic extraction, mesoscopic prediction, macroscopic distribution, and visualization output processing.

[0005] Firstly, this application provides a novel hydraulic simulation method for heating networks based on a large model, comprising the following steps: Acquire historical hydraulic operation record data, heating network system operation record data, and heating network environment and topology data for a preset time period, and perform data preprocessing and data fusion to obtain the historical operation fusion feature vector of the heating network; Acquire real-time GIS topology data and compare it with preset baseline GIS topology data; generate an optimized adjacency matrix based on the comparison results. The image of the inner wall of the heating pipeline network is acquired and processed by a preset pipeline roughness recognition model to obtain the roughness distribution matrix of the inner wall of the heating pipeline network. Based on the roughness distribution matrix, the hydraulic prediction data of the heat exchange station's hot water supply is generated by processing the preset heat exchange station hot water supply hydraulic prediction model to generate hydraulic operation prediction data of the heat exchange station within a preset time period. Based on the optimized adjacency matrix and hydraulic operation prediction data, the data is processed through a preset heating network hydraulic prediction model to generate heating network hydraulic distribution data. The hydraulic distribution data of the heating network is visualized and rendered to generate a hydraulic simulation scene of the heating network, which is then output to the heating network dispatch center for display.

[0006] Optionally, in the novel heating network hydraulic simulation method based on a large model described in this application, the step of acquiring historical heating network hydraulic operation record data, heating network system operation record data, and heating network environment and topology data for a preset historical time period, and performing data preprocessing and data fusion to obtain the historical operation fusion feature vector of the heating network includes: Acquire historical data on hydraulic operation of the heating network, operation of the heating network system, and environmental and topological data for a preset time period. The hydraulic operation data of the heating network includes temperature, flow rate, and pressure; The operating data of the heating network system includes vibration acceleration data and sound pressure level; The heating network environment and topology data include heating network pipeline surface status data and GIS topology data; The temperature, flow rate, and pressure are subjected to outlier removal and missing value linear interpolation to obtain optimized temperature, optimized flow rate, and optimized pressure. The vibration acceleration data is processed by short-time Fourier transform to obtain the vibration time-frequency matrix, and the Mel-frequency cepstral coefficients of the sound pressure level are extracted to generate the sound pressure Mel-frequency cepstral coefficients. The surface condition data and GIS topology data of the heating network pipeline are encoded to obtain surface condition feature vectors and topology feature vectors. The optimized temperature, optimized flow rate, optimized pressure, vibration time-frequency matrix, and sound pressure Mel-frequency cepstral coefficients are combined with the surface state feature vector and topological feature vector and processed through a preset multimodal data conversion model to obtain the historical operation fusion feature vector of the heating network.

[0007] Optionally, in the novel heating network hydraulic simulation method based on a large model described in this application, the step of acquiring real-time GIS topology data, comparing it with preset benchmark GIS topology data, and generating an optimized adjacency matrix based on the comparison results includes: Acquire real-time GIS topology data and construct a real-time heating topology map, including a real-time heating node set and a real-time heating edge set; A benchmark heating topology map is constructed based on the preset benchmark GIS topology data, including the benchmark heating node set and the benchmark heating edge set, and matrix transformation is performed to generate the benchmark adjacency matrix. The real-time heating node set and real-time heating edge set are processed with the reference heating node set and reference heating edge set using a preset Jaccard similarity algorithm to obtain a similarity value; The similarity value is compared with a preset similarity threshold. If the similarity value is greater than a preset similarity threshold, then the baseline adjacency matrix is ​​an optimized adjacency matrix; If the similarity value is less than or equal to a preset similarity threshold, the baseline adjacency matrix is ​​optimized based on the real-time heating node set and the real-time heating edge set to generate an optimized adjacency matrix.

[0008] Optionally, in the novel hydraulic simulation method for heating networks based on a large model described in this application, the step of acquiring an image of the inner wall of the heating network and processing it through a preset network roughness recognition model to obtain the roughness distribution matrix of the inner wall of the heating network includes: The initial pipeline roughness recognition model is trained by combining the historical operation fusion feature vector with the preset historical heating pipeline inner wall image and the marked historical roughness distribution matrix to obtain the trained preset pipeline roughness recognition model. The system acquires real-time hydraulic operation records of the heating network, real-time operation records of the heating network system, and real-time environmental and topological data of the heating network. It then performs data preprocessing and data fusion to obtain the real-time operation fusion feature vector of the heating network. The image of the inner wall of the heating pipeline network is acquired, and then processed by a preset pipeline roughness recognition model in combination with the real-time operation fusion feature vector to obtain the roughness distribution matrix of the inner wall of the heating pipeline network.

[0009] Optionally, in the novel heating network hydraulic simulation method based on a large model described in this application, the step of processing the roughness distribution matrix using a preset heat exchange station hot water supply hydraulic prediction model to generate hydraulic operation prediction data for the heat exchange station within a preset time period includes: The real-time hydraulic operation record data of the heating network is normalized to obtain hydraulic operation sequence characteristic data. The roughness distribution matrix is ​​subjected to feature extraction to obtain roughness space feature data; Based on the hydraulic operation sequence characteristic data and the roughness spatial characteristic data, the data are processed by a preset heat exchange station hot water supply hydraulic prediction model to generate hydraulic operation prediction data for the heat exchange station within a preset time period.

[0010] Optionally, in the novel heating network hydraulic simulation method based on a large model described in this application, the step of processing the optimized adjacency matrix and hydraulic operation prediction data through a preset heating network hydraulic prediction model to generate heating network hydraulic distribution data includes: The optimized adjacency matrix is ​​transformed to obtain the heating diagram structure data; The heating diagram structure data, hydraulic operation prediction data, and real-time operation record data are input into a preset heating network hydraulic prediction model for processing to generate heating network hydraulic distribution data. The hydraulic distribution data of the heating network includes pressure field distribution data, flow field distribution data, and temperature field distribution data.

[0011] Secondly, this application provides a novel hydraulic simulation system for a large-scale heating network. The system includes a memory and a processor. The memory contains a program for a novel hydraulic simulation method for a large-scale heating network. When executed by the processor, the program for this novel hydraulic simulation method for a large-scale heating network performs the following steps: Acquire historical hydraulic operation record data, heating network system operation record data, and heating network environment and topology data for a preset time period, and perform data preprocessing and data fusion to obtain the historical operation fusion feature vector of the heating network; Acquire real-time GIS topology data and compare it with preset baseline GIS topology data; generate an optimized adjacency matrix based on the comparison results. The image of the inner wall of the heating pipeline network is acquired and processed by a preset pipeline roughness recognition model to obtain the roughness distribution matrix of the inner wall of the heating pipeline network. Based on the roughness distribution matrix, the hydraulic prediction data of the heat exchange station's hot water supply is generated by processing the preset heat exchange station hot water supply hydraulic prediction model to generate hydraulic operation prediction data of the heat exchange station within a preset time period. Based on the optimized adjacency matrix and hydraulic operation prediction data, the data is processed through a preset heating network hydraulic prediction model to generate heating network hydraulic distribution data. The hydraulic distribution data of the heating network is visualized and rendered to generate a hydraulic simulation scene of the heating network, which is then output to the heating network dispatch center for display.

[0012] Optionally, in the novel heating network hydraulic simulation system based on a large model described in this application, the step of acquiring historical heating network hydraulic operation record data, heating network system operation record data, and heating network environment and topology data for a preset historical time period, and performing data preprocessing and data fusion to obtain the historical operation fusion feature vector of the heating network includes: Acquire historical data on hydraulic operation of the heating network, operation of the heating network system, and environmental and topological data for a preset time period. The hydraulic operation data of the heating network includes temperature, flow rate, and pressure; The operating data of the heating network system includes vibration acceleration data and sound pressure level; The heating network environment and topology data include heating network pipeline surface status data and GIS topology data; The temperature, flow rate, and pressure are subjected to outlier removal and missing value linear interpolation to obtain optimized temperature, optimized flow rate, and optimized pressure. The vibration acceleration data is processed by short-time Fourier transform to obtain the vibration time-frequency matrix, and the Mel-frequency cepstral coefficients of the sound pressure level are extracted to generate the sound pressure Mel-frequency cepstral coefficients. The surface condition data and GIS topology data of the heating network pipeline are encoded to obtain surface condition feature vectors and topology feature vectors. The optimized temperature, optimized flow rate, optimized pressure, vibration time-frequency matrix, and sound pressure Mel-frequency cepstral coefficients are combined with the surface state feature vector and topological feature vector and processed through a preset multimodal data conversion model to obtain the historical operation fusion feature vector of the heating network.

[0013] Optionally, in the novel heating network hydraulic simulation system based on a large model described in this application, the step of acquiring real-time GIS topology data, comparing it with preset benchmark GIS topology data, and generating an optimized adjacency matrix based on the comparison results includes: Acquire real-time GIS topology data and construct a real-time heating topology map, including a real-time heating node set and a real-time heating edge set; A benchmark heating topology map is constructed based on the preset benchmark GIS topology data, including the benchmark heating node set and the benchmark heating edge set, and matrix transformation is performed to generate the benchmark adjacency matrix. The real-time heating node set and real-time heating edge set are processed with the reference heating node set and reference heating edge set using a preset Jaccard similarity algorithm to obtain a similarity value; If the similarity value is greater than a preset similarity threshold, then the baseline adjacency matrix is ​​an optimized adjacency matrix; If the similarity value is less than or equal to a preset similarity threshold, the baseline adjacency matrix is ​​optimized based on the real-time heating node set and the real-time heating edge set to generate an optimized adjacency matrix.

[0014] Thirdly, this application also provides a computer-readable storage medium storing a program for a novel hydraulic simulation method for a large-scale heating network, wherein when the program is executed by a processor, it implements the steps of the novel hydraulic simulation method for a large-scale heating network as described in any of the preceding claims.

[0015] As can be seen from the above, the novel heating network hydraulic simulation method, system and medium based on a large model provided in this application realizes the heating network hydraulic simulation based on a large model through data fusion, topology optimization, micro-characteristic extraction, meso-level prediction, macro-level distribution and visualization output processing.

[0016] Other features and advantages of this application will be set forth in the following description and will be apparent in part from the description or may be learned by practicing embodiments of this application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating a novel hydraulic simulation method for heating networks based on a large model, provided for embodiments of this application; Figure 2 A flowchart illustrating the process of obtaining the historical operation fusion feature vector of a heating network using a novel large-model-based hydraulic simulation method for heating networks, as provided in this application embodiment; Figure 3 A flowchart illustrating the generation of an optimized adjacency matrix for a novel hydraulic simulation method for heating networks based on a large model, provided for embodiments of this application; Figure 4 This is a flowchart illustrating the process of obtaining the roughness distribution matrix of the inner wall of a heating network using a novel hydraulic simulation method based on a large model, as provided in this application embodiment. Detailed Implementation

[0019] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0020] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0021] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating a novel hydraulic simulation method for heating networks based on a large model, as described in some embodiments of this application. This novel hydraulic simulation method for heating networks based on a large model is used in terminal devices, such as computers and mobile terminals. The novel hydraulic simulation method for heating networks based on a large model includes the following steps: S11. Obtain historical hydraulic operation record data, heating network system operation record data, and heating network environment and topology data for a preset time period, and perform data preprocessing and data fusion to obtain the historical operation fusion feature vector of the heating network. S12. Obtain real-time GIS topology data and compare it with preset benchmark GIS topology data, and generate an optimized adjacency matrix based on the comparison results. S13. Obtain an image of the inner wall of the heating pipe network and process it through a preset pipe network roughness recognition model to obtain the roughness distribution matrix of the inner wall of the heating pipe network. S14. Based on the roughness distribution matrix, the hydraulic prediction model for the hot water supply of the heat exchange station is used to process the hydraulic operation prediction data of the heat exchange station within a preset time period. S15. Based on the optimized adjacency matrix and hydraulic operation prediction data, the data is processed through a preset heating network hydraulic prediction model to generate heating network hydraulic distribution data. S16. Visualize and render the hydraulic distribution data of the heating network to generate a hydraulic simulation scene diagram of the heating network, and output it to the heating network dispatch center for display.

[0022] It should be noted that, firstly, based on the acquired historical multimodal data, a historical operational fusion feature vector is obtained. Then, combined with preset historical heating network inner wall images and marked historical roughness distribution matrices, an initial network roughness identification model is trained to obtain a trained preset network roughness identification model. This model is used to correct the hydraulic calculation friction coefficient from the heat exchange station outlet to the network end, making the meso-level flow and pressure predictions more consistent with the actual aging state of the pipelines and reducing simulation errors caused by roughness estimation deviations. Secondly, based on the comparison results between real-time GIS topology data and preset benchmark GIS topology data, an optimized adjacency matrix is ​​generated as the basis for constructing the entire network's hydraulic calculation topology. This ensures that the model can reflect real-time topology changes due to newly added pipelines and valve status changes, making the macro-level pressure, flow, and temperature field simulation results consistent with the actual network structure. Finally, the roughness distribution matrix extracted from the micro-features is combined with real-time hydraulic operation records of the heating network. The system processes the water supply hydraulic prediction model of the heat exchange station to generate hydraulic operation prediction data for a preset time period. This data is used to achieve high-precision prediction of the hydraulic operation curve of the heat exchange station. Then, the optimized adjacency matrix and the hydraulic operation prediction data obtained from the meso-level are processed through the preset heating network hydraulic prediction model to generate heating network hydraulic distribution data. This data is used to achieve cross-scale collaborative correction at the macro level (combining micro-roughness and meso-level output deviation). Finally, RGB color mapping is performed based on the pressure field distribution data. The flow rate value is dynamically mapped by the line thickness, forming a two-dimensional visualization of "color + line width" with the pressure field color. For example, a thick red line represents a high-pressure, high-flow pipe section. A semi-transparent halo effect is superimposed on the surface of the pipe network model. The temperature gradient is distinguished by the halo transparency, forming a spatial relationship with the pressure field (e.g., high-temperature pipe sections usually correspond to high-pressure areas). Finally, a heating network hydraulic simulation scene diagram is generated and output to the heating network dispatch center for display.

[0023] Please refer to Figure 2 , Figure 2 This is a flowchart illustrating the process of obtaining the historical operation fusion feature vector of a heating network using a novel large-model-based hydraulic simulation method in some embodiments of this application. According to embodiments of the present invention, the steps of acquiring historical heating network hydraulic operation record data, heating network system operation record data, and heating network environment and topology data over a preset historical time period, and performing data preprocessing and data fusion to obtain the historical operation fusion feature vector of the heating network include: S21. Obtain historical preset time period data of heating network hydraulic operation record, heating network system operation record data and heating network environment and topology data; S211, The hydraulic operation record data of the heating network includes temperature, flow rate and pressure; S212, The operation record data of the heating network system includes vibration acceleration data and sound pressure level; S213, The heating network environment and topology data include heating network pipeline surface status data and GIS topology data; S221. Perform outlier removal and missing value linear interpolation to fill in the temperature, flow rate and pressure to obtain optimized temperature, optimized flow rate and optimized pressure; S222. The vibration acceleration data is processed by short-time Fourier transform to obtain the vibration time-frequency matrix, and the Mel-frequency cepstral coefficients of the sound pressure level are extracted to generate the sound pressure Mel-frequency cepstral coefficients. S223. Encode the surface status data and GIS topology data of the heating network pipeline to obtain surface status feature vector and topology feature vector; S23. The optimized temperature, optimized flow rate, optimized pressure, vibration time-frequency matrix, and sound pressure Mel-frequency cepstral coefficients are combined with the surface state feature vector and topological feature vector and processed through a preset multimodal data conversion model to obtain the historical operation fusion feature vector of the heating network.

[0024] It should be noted that, in order to improve the accuracy of the heating network hydraulic simulation, this embodiment comprehensively considers the heating network hydraulic operation, heating network system operation, and heating network environment and topology data over a preset historical time period (such as the past 3 years). This includes both basic physical quantities and implicit characteristic quantities as well as environmental and topological quantities. The surface state of the heating network pipelines includes pipeline exposure, ground subsidence, and water accumulation. The surface state data of the heating network pipelines is represented by different values. GIS topology data refers to a digital information set representing the spatial structural relationships of the heating network, including heating nodes, heating edges, and the association rules between nodes and edges. Missing values ​​are filled based on a quadratic interpolation of the trend in adjacent time periods (e.g., for missing values ​​in a pressure drop interval, the slope of the preceding and following 5 data points is used for calculation). After data preprocessing, timestamps and preset node GPS coordinates are introduced to generate 512. The historical operation fusion feature vector contains 8-dimensional timestamp encoding, 16-dimensional spatial coordinate encoding, 128-dimensional temporal features extracted from optimized temperature, optimized flow rate, and optimized pressure, 128-dimensional latent features (energy spectrum features of 64-dimensional vibration time-frequency matrix + 64-dimensional sound pressure Mel-frequency cepstral coefficients), 96-dimensional environmental topological features, and 136-dimensional cross-correlation features. Among them, the preset multimodal data transformation model adopts the Transformer-XL model, which is based on a large number of historical samples of optimized temperature, optimized flow rate, optimized pressure, vibration time-frequency matrix, and sound pressure Mel-frequency cepstral coefficients, combined with surface state feature vectors and topological feature vectors, as well as the corresponding historical operation fusion feature vectors. It learns cross-modal and cross-temporal correlations through an attention mechanism and introduces a memory caching mechanism to handle long sequence dependencies for training.

[0025] Please refer to Figure 3 , Figure 3This is a flowchart illustrating the generation of an optimized adjacency matrix for a novel hydraulic simulation method for heating networks based on a large model, as described in some embodiments of this application. According to an embodiment of the present invention, the step of acquiring real-time GIS topology data, comparing it with preset benchmark GIS topology data, and generating an optimized adjacency matrix based on the comparison results includes: S31. Obtain real-time GIS topology data and construct a real-time heating topology map, including a real-time heating node set and a real-time heating edge set. S32. Construct a benchmark heating topology map based on the preset benchmark GIS topology data, including the benchmark heating node set and the benchmark heating edge set, and perform matrix transformation processing to generate a benchmark adjacency matrix. S33. The real-time heating node set and the real-time heating edge set are processed with the reference heating node set and the reference heating edge set using a preset Jaccard similarity algorithm to obtain a similarity value. S34. Compare the similarity value with a preset similarity threshold. S341. If the similarity value is greater than the preset similarity threshold, then the baseline adjacency matrix is ​​an optimized adjacency matrix. S342. If the similarity value is less than or equal to a preset similarity threshold, the baseline adjacency matrix is ​​optimized based on the real-time heating node set and the real-time heating edge set to generate an optimized adjacency matrix.

[0026] It should be noted that real-time GIS topology data is collected, parsed into real-time heating node sets (including node IDs and coordinates), and real-time heating edge sets (including pipe diameters and lengths), and a real-time heating topology map is constructed. Simultaneously, based on preset benchmark GIS data, a benchmark heating topology map containing benchmark node sets and benchmark edge sets is generated and converted into a benchmark adjacency matrix. The matrix elements are pipe diameters; an empty matrix is ​​0. The similarity between the real-time and benchmark node sets and edge sets is calculated separately. For example, if the benchmark heating edge set is 800, adding one edge changes the real-time heating edge set to 801. The intersection of the two sets (resulting in 800) is then divided by the union (resulting in 801). The edge set similarity value is approximately 0.9988 (800 / 801 ≈ 0.9988). Similarly, the node similarity value can be obtained. The edge set similarity value and the node similarity value are weighted and summed (where the corresponding weight values ​​are pre-constructed by those skilled in the art and can be dynamically adjusted according to the simulation situation) to obtain the similarity value. Finally, the topology change is identified by threshold comparison. If the similarity is greater than the preset similarity threshold, the baseline adjacency matrix is ​​directly used as the optimized adjacency matrix. Otherwise, the matrix is ​​updated according to the real-time heating node set and the real-time heating edge set. For example, if a new edge is added, the corresponding diameter is filled in, and if an edge is deleted, it is set to 0. Finally, the optimized adjacency matrix is ​​generated to achieve dynamic topology adaptation.

[0027] Please refer to Figure 4 , Figure 4This is a flowchart illustrating the acquisition of the roughness distribution matrix of the inner wall of a heating network using a novel large-model-based hydraulic simulation method for heating networks, as described in some embodiments of this application. According to embodiments of the present invention, the step of acquiring an image of the inner wall of the heating network and processing it using a preset network roughness recognition model to obtain the roughness distribution matrix of the inner wall of the heating network includes: S41. The initial pipeline roughness recognition model is trained based on the historical operation fusion feature vector, the preset historical heating pipeline inner wall image, and the marked historical roughness distribution matrix to obtain the trained preset pipeline roughness recognition model. S42. Acquire real-time hydraulic operation record data of the heating network, real-time operation record data of the heating network system, and real-time environmental and topological data of the heating network, and perform data preprocessing and data fusion to obtain the real-time operation fusion feature vector of the heating network; S43. Obtain an image of the inner wall of the heating pipeline network, and process it by combining the real-time operation fusion feature vector with a preset pipeline roughness recognition model to obtain the roughness distribution matrix of the inner wall of the heating pipeline network.

[0028] It should be noted that the historical operation fusion feature vector is combined with the preset historical heating network inner wall image and the marked historical roughness distribution matrix to train the initial model. This allows the model to learn the correlation between "operation features, image features and roughness", resulting in a trained network roughness recognition model. Real-time hydraulic operation data (real-time temperature, real-time flow and real-time pressure), real-time operation data of the heating network system (real-time vibration acceleration data and real-time sound pressure level), and real-time environmental and topological data are collected. After preprocessing and fusion, a real-time operation fusion feature vector is generated. Then, real-time heating network inner wall images are obtained and input into the trained model. Through correlation analysis, the roughness distribution matrix of the network inner wall is output to reflect the roughness values ​​at different locations and the influence of the micro-layer on the hydraulic simulation of the heating network.

[0029] According to an embodiment of the present invention, the step of processing the roughness distribution matrix using a preset heat exchange station hot water supply hydraulic prediction model to generate hydraulic operation prediction data for the heat exchange station within a preset time period includes: The real-time hydraulic operation record data of the heating network is normalized to obtain hydraulic operation sequence characteristic data. The roughness distribution matrix is ​​subjected to feature extraction to obtain roughness space feature data; Based on the hydraulic operation sequence characteristic data and the roughness spatial characteristic data, the data are processed by a preset heat exchange station hot water supply hydraulic prediction model to generate hydraulic operation prediction data for the heat exchange station within a preset time period.

[0030] It should be noted that the real-time hydraulic operation records of the heating network, including real-time temperature, real-time flow, and real-time pressure, are normalized to extract the hydraulic operation sequence feature data that changes over time. Spatial features are extracted from the roughness distribution matrix through convolution operations to obtain roughness spatial feature data reflecting the spatial distribution of the pipe inner wall roughness. These are then input into the preset heat exchange station hot water supply hydraulic prediction model. The model learns the correlation between the two to generate hydraulic operation prediction data for the heat exchange station within a preset time period, reflecting the hydraulic meso-level simulation of the heating network. The preset heat exchange station hot water supply hydraulic prediction model is obtained by acquiring a large amount of historical samples of hydraulic operation sequence feature data, roughness spatial feature data, and corresponding hydraulic operation prediction data, and is trained by using LSTM to capture temporal patterns and convolutional layers to extract spatial features.

[0031] According to an embodiment of the present invention, the step of processing the optimized adjacency matrix and hydraulic operation prediction data through a preset heating network hydraulic prediction model to generate heating network hydraulic distribution data includes: The optimized adjacency matrix is ​​transformed to obtain the heating diagram structure data; The heating diagram structure data, hydraulic operation prediction data, and real-time operation record data are input into a preset heating network hydraulic prediction model for processing to generate heating network hydraulic distribution data. The hydraulic distribution data of the heating network includes pressure field distribution data, flow field distribution data, and temperature field distribution data.

[0032] It should be noted that the optimized adjacency matrix is ​​converted into heating map structure data containing node coordinates, pipe diameter, and length attributes, constructing a digital representation of the heating network topology. The heating map structure data serves as the topological foundation, hydraulic operation prediction data serves as a meso-level reference, and real-time operation record data serves as the input for measured verification. A pre-set heating network hydraulic prediction model is established, as shown in the Transformer model. Through graph structure inference and multi-source data fusion, the output includes pressure field distribution data, flow field distribution data, and temperature field distribution data covering the entire network, such as 100m×100m grid precision. This achieves a global quantitative representation of the heating network hydraulic state based on a large model. The pre-set heating network hydraulic prediction model is obtained by acquiring a large amount of historical sample heating map structure data, hydraulic operation prediction data, real-time operation record data, and corresponding heating network hydraulic distribution data. Based on the graph Transformer architecture, topological relationships are constructed using graph structure data, and the hydraulic conduction laws are learned through multi-source input for training.

[0033] It is worth mentioning that, according to embodiments of the present invention, it further includes: The real-time hydraulic operation record data of the heating network is compared with the corresponding hydraulic operation prediction data to obtain the prediction deviation rate; If the prediction deviation rate is greater than the preset prediction deviation warning threshold, then the deviation duration that is greater than the preset prediction deviation warning threshold is counted. If the deviation duration exceeds a preset deviation duration threshold, it is determined to be a fault node; The real-time running fusion feature vector corresponding to the fault node is input into a preset fault type diagnosis model for processing to obtain fault type label data. The fault handling strategy is obtained by querying the preset fault handling database based on the fault type label data.

[0034] It should be noted that, in order to respond to simulation faults in real time, the prediction deviation rate is first calculated. The prediction deviation rate refers to the absolute value of the difference between the real-time hydraulic operation record data of the heating network and the corresponding hydraulic operation prediction data, and then the ratio of this difference to the hydraulic operation prediction data. The fault point is determined based on the prediction deviation. The real-time operation fusion feature vector of the fault point is then input into the preset fault type diagnosis model for processing to obtain fault type label data. Fault types, such as valve jamming and pipeline leakage, are represented by different numerical values. The corresponding fault handling strategy is then obtained based on the obtained fault type. For example, when valve jamming occurs, the flow coefficient of the associated pipe section is multiplied by a correction factor of 0.7 to 0.9. The preset fault type diagnosis model is obtained by training with real-time operation fusion feature vectors and corresponding fault type label data from a large number of historical samples. The preset fault handling database is pre-constructed by those skilled in the art based on simulation analysis and can be dynamically adjusted.

[0035] This invention also discloses a novel hydraulic simulation system for heating networks based on a large model, comprising a memory and a processor. The memory includes a program for a novel hydraulic simulation method for heating networks based on a large model. When the processor executes the program for the novel hydraulic simulation method for heating networks based on a large model, it performs the following steps: Acquire historical hydraulic operation record data, heating network system operation record data, and heating network environment and topology data for a preset time period, and perform data preprocessing and data fusion to obtain the historical operation fusion feature vector of the heating network; Acquire real-time GIS topology data and compare it with preset baseline GIS topology data; generate an optimized adjacency matrix based on the comparison results. The image of the inner wall of the heating pipeline network is acquired and processed by a preset pipeline roughness recognition model to obtain the roughness distribution matrix of the inner wall of the heating pipeline network. Based on the roughness distribution matrix, the hydraulic prediction data of the heat exchange station's hot water supply is generated by processing the preset heat exchange station hot water supply hydraulic prediction model to generate hydraulic operation prediction data of the heat exchange station within a preset time period. Based on the optimized adjacency matrix and hydraulic operation prediction data, the data is processed through a preset heating network hydraulic prediction model to generate heating network hydraulic distribution data. The hydraulic distribution data of the heating network is visualized and rendered to generate a hydraulic simulation scene of the heating network, which is then output to the heating network dispatch center for display.

[0036] It should be noted that, firstly, based on the acquired historical multimodal data, a historical operational fusion feature vector is obtained. Then, combined with preset historical heating network inner wall images and marked historical roughness distribution matrices, an initial network roughness identification model is trained to obtain a trained preset network roughness identification model. This model is used to correct the hydraulic calculation friction coefficient from the heat exchange station outlet to the network end, making the meso-level flow and pressure predictions more consistent with the actual aging state of the pipelines and reducing simulation errors caused by roughness estimation deviations. Secondly, based on the comparison results between real-time GIS topology data and preset benchmark GIS topology data, an optimized adjacency matrix is ​​generated as the basis for constructing the entire network's hydraulic calculation topology. This ensures that the model can reflect real-time topology changes due to newly added pipelines and valve status changes, making the macro-level pressure, flow, and temperature field simulation results consistent with the actual network structure. Finally, the roughness distribution matrix extracted from the micro-features is combined with real-time hydraulic operation records of the heating network. The system processes the water supply hydraulic prediction model of the heat exchange station to generate hydraulic operation prediction data for a preset time period. This data is used to achieve high-precision prediction of the hydraulic operation curve of the heat exchange station. Then, the optimized adjacency matrix and the hydraulic operation prediction data obtained from the meso-level are processed through the preset heating network hydraulic prediction model to generate heating network hydraulic distribution data. This data is used to achieve cross-scale collaborative correction at the macro level (combining micro-roughness and meso-level output deviation). Finally, RGB color mapping is performed based on the pressure field distribution data. The flow rate value is dynamically mapped by the line thickness, forming a two-dimensional visualization of "color + line width" with the pressure field color. For example, a thick red line represents a high-pressure, high-flow pipe section. A semi-transparent halo effect is superimposed on the surface of the pipe network model. The temperature gradient is distinguished by the halo transparency, forming a spatial relationship with the pressure field (e.g., high-temperature pipe sections usually correspond to high-pressure areas). Finally, a heating network hydraulic simulation scene diagram is generated and output to the heating network dispatch center for display.

[0037] According to an embodiment of the present invention, the step of acquiring historical hydraulic operation record data of the heating network, operation record data of the heating network system, and environmental and topological data of the heating network over a preset historical time period, and performing data preprocessing and data fusion to obtain a historical operation fusion feature vector of the heating network includes: Acquire historical data on hydraulic operation of the heating network, operation of the heating network system, and environmental and topological data for a preset time period. The hydraulic operation data of the heating network includes temperature, flow rate, and pressure; The operating data of the heating network system includes vibration acceleration data and sound pressure level; The heating network environment and topology data include heating network pipeline surface status data and GIS topology data; The temperature, flow rate, and pressure are subjected to outlier removal and missing value linear interpolation to obtain optimized temperature, optimized flow rate, and optimized pressure. The vibration acceleration data is processed by short-time Fourier transform to obtain the vibration time-frequency matrix, and the Mel-frequency cepstral coefficients of the sound pressure level are extracted to generate the sound pressure Mel-frequency cepstral coefficients. The surface condition data and GIS topology data of the heating network pipeline are encoded to obtain surface condition feature vectors and topology feature vectors. The optimized temperature, optimized flow rate, optimized pressure, vibration time-frequency matrix, and sound pressure Mel-frequency cepstral coefficients are combined with the surface state feature vector and topological feature vector and processed through a preset multimodal data conversion model to obtain the historical operation fusion feature vector of the heating network.

[0038] It should be noted that, in order to improve the accuracy of the heating network hydraulic simulation, this embodiment comprehensively considers the heating network hydraulic operation, heating network system operation, and heating network environment and topology data over a preset historical time period (such as the past 3 years). This includes both basic physical quantities and implicit characteristic quantities as well as environmental and topological quantities. The surface state of the heating network pipelines includes pipeline exposure, ground subsidence, and water accumulation. The surface state data of the heating network pipelines is represented by different values. GIS topology data refers to a digital information set representing the spatial structural relationships of the heating network, including heating nodes, heating edges, and the association rules between nodes and edges. Missing values ​​are filled based on a quadratic interpolation of the trend in adjacent time periods (e.g., for missing values ​​in a pressure drop interval, the slope of the preceding and following 5 data points is used for calculation). After data preprocessing, timestamps and preset node GPS coordinates are introduced to generate 512. The historical operation fusion feature vector contains 8-dimensional timestamp encoding, 16-dimensional spatial coordinate encoding, 128-dimensional temporal features extracted from optimized temperature, optimized flow rate, and optimized pressure, 128-dimensional latent features (energy spectrum features of 64-dimensional vibration time-frequency matrix + 64-dimensional sound pressure Mel-frequency cepstral coefficients), 96-dimensional environmental topological features, and 136-dimensional cross-correlation features. Among them, the preset multimodal data transformation model adopts the Transformer-XL model, which is based on a large number of historical samples of optimized temperature, optimized flow rate, optimized pressure, vibration time-frequency matrix, and sound pressure Mel-frequency cepstral coefficients, combined with surface state feature vectors and topological feature vectors, as well as the corresponding historical operation fusion feature vectors. It learns cross-modal and cross-temporal correlations through an attention mechanism and introduces a memory caching mechanism to handle long sequence dependencies for training.

[0039] According to an embodiment of the present invention, the step of acquiring real-time GIS topology data, comparing it with preset benchmark GIS topology data, and generating an optimized adjacency matrix based on the comparison result includes: Acquire real-time GIS topology data and construct a real-time heating topology map, including a real-time heating node set and a real-time heating edge set; A benchmark heating topology map is constructed based on the preset benchmark GIS topology data, including the benchmark heating node set and the benchmark heating edge set, and matrix transformation is performed to generate the benchmark adjacency matrix. The real-time heating node set and real-time heating edge set are processed with the reference heating node set and reference heating edge set using a preset Jaccard similarity algorithm to obtain a similarity value; The similarity value is compared with a preset similarity threshold. If the similarity value is greater than a preset similarity threshold, then the baseline adjacency matrix is ​​an optimized adjacency matrix; If the similarity value is less than or equal to a preset similarity threshold, the baseline adjacency matrix is ​​optimized based on the real-time heating node set and the real-time heating edge set to generate an optimized adjacency matrix.

[0040] It should be noted that real-time GIS topology data is collected, parsed into real-time heating node sets (including node IDs and coordinates), and real-time heating edge sets (including pipe diameters and lengths), and a real-time heating topology map is constructed. Simultaneously, based on preset benchmark GIS data, a benchmark heating topology map containing benchmark node sets and benchmark edge sets is generated and converted into a benchmark adjacency matrix. The matrix elements are pipe diameters; an empty matrix is ​​0. The similarity between the real-time and benchmark node sets and edge sets is calculated separately. For example, if the benchmark heating edge set is 800, adding one edge changes the real-time heating edge set to 801. The intersection of the two sets (resulting in 800) is then divided by the union (resulting in 801). The edge set similarity value is approximately 0.9988 (800 / 801 ≈ 0.9988). Similarly, the node similarity value can be obtained. The edge set similarity value and the node similarity value are weighted and summed (where the corresponding weight values ​​are pre-constructed by those skilled in the art and can be dynamically adjusted according to the simulation situation) to obtain the similarity value. Finally, the topology change is identified by threshold comparison. If the similarity is greater than the preset similarity threshold, the baseline adjacency matrix is ​​directly used as the optimized adjacency matrix. Otherwise, the matrix is ​​updated according to the real-time heating node set and the real-time heating edge set. For example, if a new edge is added, the corresponding diameter is filled in, and if an edge is deleted, it is set to 0. Finally, the optimized adjacency matrix is ​​generated to achieve dynamic topology adaptation.

[0041] According to an embodiment of the present invention, the step of acquiring an image of the inner wall of a heating pipeline network and processing it through a preset pipeline roughness recognition model to obtain a roughness distribution matrix of the inner wall of the heating pipeline network includes: The initial pipeline roughness recognition model is trained by combining the historical operation fusion feature vector with the preset historical heating pipeline inner wall image and the marked historical roughness distribution matrix to obtain the trained preset pipeline roughness recognition model. The system acquires real-time hydraulic operation records of the heating network, real-time operation records of the heating network system, and real-time environmental and topological data of the heating network. It then performs data preprocessing and data fusion to obtain the real-time operation fusion feature vector of the heating network. The image of the inner wall of the heating pipeline network is acquired, and then processed by a preset pipeline roughness recognition model in combination with the real-time operation fusion feature vector to obtain the roughness distribution matrix of the inner wall of the heating pipeline network.

[0042] It should be noted that the historical operation fusion feature vector is combined with the preset historical heating network inner wall image and the marked historical roughness distribution matrix to train the initial model. This allows the model to learn the correlation between "operation features, image features and roughness", resulting in a trained network roughness recognition model. Real-time hydraulic operation data (real-time temperature, real-time flow and real-time pressure), real-time operation data of the heating network system (real-time vibration acceleration data and real-time sound pressure level), and real-time environmental and topological data are collected. After preprocessing and fusion, a real-time operation fusion feature vector is generated. Then, real-time heating network inner wall images are obtained and input into the trained model. Through correlation analysis, the roughness distribution matrix of the network inner wall is output to reflect the roughness values ​​at different locations and the influence of the micro-layer on the hydraulic simulation of the heating network.

[0043] According to an embodiment of the present invention, the step of processing the roughness distribution matrix using a preset heat exchange station hot water supply hydraulic prediction model to generate hydraulic operation prediction data for the heat exchange station within a preset time period includes: The real-time hydraulic operation record data of the heating network is normalized to obtain hydraulic operation sequence characteristic data. The roughness distribution matrix is ​​subjected to feature extraction to obtain roughness space feature data; Based on the hydraulic operation sequence characteristic data and the roughness spatial characteristic data, the data are processed by a preset heat exchange station hot water supply hydraulic prediction model to generate hydraulic operation prediction data for the heat exchange station within a preset time period.

[0044] It should be noted that the real-time hydraulic operation records of the heating network, including real-time temperature, real-time flow, and real-time pressure, are normalized to extract the hydraulic operation sequence feature data that changes over time. Spatial features are extracted from the roughness distribution matrix through convolution operations to obtain roughness spatial feature data reflecting the spatial distribution of the pipe inner wall roughness. These are then input into the preset heat exchange station hot water supply hydraulic prediction model. The model learns the correlation between the two to generate hydraulic operation prediction data for the heat exchange station within a preset time period, reflecting the hydraulic meso-level simulation of the heating network. The preset heat exchange station hot water supply hydraulic prediction model is obtained by acquiring a large amount of historical samples of hydraulic operation sequence feature data, roughness spatial feature data, and corresponding hydraulic operation prediction data, and is trained by using LSTM to capture temporal patterns and convolutional layers to extract spatial features.

[0045] According to an embodiment of the present invention, the step of processing the optimized adjacency matrix and hydraulic operation prediction data through a preset heating network hydraulic prediction model to generate heating network hydraulic distribution data includes: The optimized adjacency matrix is ​​transformed to obtain the heating diagram structure data; The heating diagram structure data, hydraulic operation prediction data, and real-time operation record data are input into a preset heating network hydraulic prediction model for processing to generate heating network hydraulic distribution data. The hydraulic distribution data of the heating network includes pressure field distribution data, flow field distribution data, and temperature field distribution data.

[0046] It should be noted that the optimized adjacency matrix is ​​converted into heating map structure data containing node coordinates, pipe diameter, and length attributes, constructing a digital representation of the heating network topology. The heating map structure data serves as the topological foundation, hydraulic operation prediction data serves as a meso-level reference, and real-time operation record data serves as the input for measured verification. A pre-set heating network hydraulic prediction model is established, as shown in the Transformer model. Through graph structure inference and multi-source data fusion, the output includes pressure field distribution data, flow field distribution data, and temperature field distribution data covering the entire network, such as 100m×100m grid precision. This achieves a global quantitative representation of the heating network hydraulic state based on a large model. The pre-set heating network hydraulic prediction model is obtained by acquiring a large amount of historical sample heating map structure data, hydraulic operation prediction data, real-time operation record data, and corresponding heating network hydraulic distribution data. Based on the graph Transformer architecture, topological relationships are constructed using graph structure data, and the hydraulic conduction laws are learned through multi-source input for training.

[0047] It is worth mentioning that, according to embodiments of the present invention, it further includes: The real-time hydraulic operation record data of the heating network is compared with the corresponding hydraulic operation prediction data to obtain the prediction deviation rate; If the prediction deviation rate is greater than the preset prediction deviation warning threshold, then the deviation duration that is greater than the preset prediction deviation warning threshold is counted. If the deviation duration exceeds a preset deviation duration threshold, it is determined to be a fault node; The real-time running fusion feature vector corresponding to the fault node is input into a preset fault type diagnosis model for processing to obtain fault type label data. The fault handling strategy is obtained by querying the preset fault handling database based on the fault type label data.

[0048] It should be noted that, in order to respond to simulation faults in real time, the prediction deviation rate is first calculated. The prediction deviation rate refers to the absolute value of the difference between the real-time hydraulic operation record data of the heating network and the corresponding hydraulic operation prediction data, and then the ratio of this difference to the hydraulic operation prediction data. The fault point is determined based on the prediction deviation. The real-time operation fusion feature vector of the fault point is then input into the preset fault type diagnosis model for processing to obtain fault type label data. Fault types, such as valve jamming and pipeline leakage, are represented by different numerical values. The corresponding fault handling strategy is then obtained based on the obtained fault type. For example, when valve jamming occurs, the flow coefficient of the associated pipe section is multiplied by a correction factor of 0.7 to 0.9. The preset fault type diagnosis model is obtained by training with real-time operation fusion feature vectors and corresponding fault type label data from a large number of historical samples. The preset fault handling database is pre-constructed by those skilled in the art based on simulation analysis and can be dynamically adjusted.

[0049] A third aspect of the present invention provides a readable storage medium storing a program for a novel hydraulic simulation method for a large-scale heating network, wherein when the program is executed by a processor, it implements the steps of the novel hydraulic simulation method for a large-scale heating network as described in any of the preceding claims.

[0050] This invention discloses a novel hydraulic simulation method, system, and medium for heating networks based on a large model. Through data fusion, topology optimization, microscopic characteristic extraction, mesoscopic prediction, macroscopic distribution, and visualization output processing, it realizes hydraulic simulation of heating networks based on a large model.

[0051] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0052] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0053] In addition, in the various embodiments of the present invention, each functional unit can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0054] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0055] Alternatively, if the integrated units of this invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a readable storage medium. Based on this understanding, the technical solutions of the embodiments of this invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

Claims

1. A large model-based new heat network hydraulic simulation method, characterized in that, Includes the following steps: Acquire historical hydraulic operation record data, heating network system operation record data, and heating network environment and topology data for a preset time period, and perform data preprocessing and data fusion to obtain the historical operation fusion feature vector of the heating network; Acquire real-time GIS topology data and compare it with preset baseline GIS topology data; generate an optimized adjacency matrix based on the comparison results. The image of the inner wall of the heating pipeline network is acquired and processed by a preset pipeline roughness recognition model to obtain the roughness distribution matrix of the inner wall of the heating pipeline network. Based on the roughness distribution matrix, the hydraulic prediction data of the heat exchange station's hot water supply is generated by processing the preset heat exchange station hot water supply hydraulic prediction model to generate hydraulic operation prediction data of the heat exchange station within a preset time period. Based on the optimized adjacency matrix and hydraulic operation prediction data, the data is processed through a preset heating network hydraulic prediction model to generate heating network hydraulic distribution data. The hydraulic distribution data of the heating network is visualized and rendered to generate a hydraulic simulation scene of the heating network, which is then output to the heating network dispatch center for display.

2. The large model-based novel heat network hydraulic simulation method according to claim 1, characterized in that, The process of acquiring historical hydraulic operation record data, heating network system operation record data, and heating network environment and topology data for a preset time period, and performing data preprocessing and data fusion to obtain the historical operation fusion feature vector of the heating network includes: Acquire historical data on hydraulic operation of the heating network, operation of the heating network system, and environmental and topological data for a preset time period. The hydraulic operation data of the heating network includes temperature, flow rate, and pressure; The operating data of the heating network system includes vibration acceleration data and sound pressure level; The heating network environment and topology data include heating network pipeline surface status data and GIS topology data; The temperature, flow rate, and pressure are subjected to outlier removal and missing value linear interpolation to obtain optimized temperature, optimized flow rate, and optimized pressure. The vibration acceleration data is processed by short-time Fourier transform to obtain the vibration time-frequency matrix, and the Mel-frequency cepstral coefficients of the sound pressure level are extracted to generate the sound pressure Mel-frequency cepstral coefficients. The surface condition data and GIS topology data of the heating network pipeline are encoded to obtain surface condition feature vectors and topology feature vectors. The optimized temperature, optimized flow rate, optimized pressure, vibration time-frequency matrix, and sound pressure Mel-frequency cepstral coefficients are combined with the surface state feature vector and topological feature vector and processed through a preset multimodal data conversion model to obtain the historical operation fusion feature vector of the heating network.

3. The large model-based novel heat network hydraulic simulation method according to claim 2, characterized in that, The process of acquiring real-time GIS topology data, comparing it with preset benchmark GIS topology data, and generating an optimized adjacency matrix based on the comparison results includes: Acquire real-time GIS topology data and construct a real-time heating topology map, including a real-time heating node set and a real-time heating edge set; A benchmark heating topology map is constructed based on the preset benchmark GIS topology data, including the benchmark heating node set and the benchmark heating edge set, and matrix transformation is performed to generate the benchmark adjacency matrix. The real-time heating node set and real-time heating edge set are processed with the reference heating node set and reference heating edge set using a preset Jaccard similarity algorithm to obtain a similarity value; The similarity value is compared with a preset similarity threshold. If the similarity value is greater than a preset similarity threshold, then the baseline adjacency matrix is ​​an optimized adjacency matrix; If the similarity value is less than or equal to a preset similarity threshold, the baseline adjacency matrix is ​​optimized based on the real-time heating node set and the real-time heating edge set to generate an optimized adjacency matrix.

4. The large model-based novel heat network hydraulic simulation method according to claim 3, characterized in that, The process of acquiring an image of the inner wall of the heating pipeline network and processing it using a preset pipeline roughness recognition model to obtain a roughness distribution matrix of the inner wall of the heating pipeline network includes: The initial pipeline roughness recognition model is trained by combining the historical operation fusion feature vector with the preset historical heating pipeline inner wall image and the marked historical roughness distribution matrix to obtain the trained preset pipeline roughness recognition model. The system acquires real-time hydraulic operation records of the heating network, real-time operation records of the heating network system, and real-time environmental and topological data of the heating network. It then performs data preprocessing and data fusion to obtain the real-time operation fusion feature vector of the heating network. The image of the inner wall of the heating pipeline network is acquired, and then processed by a preset pipeline roughness recognition model in combination with the real-time operation fusion feature vector to obtain the roughness distribution matrix of the inner wall of the heating pipeline network.

5. The large model-based novel heat network hydraulic simulation method according to claim 4, characterized in that, The process of processing the roughness distribution matrix using a preset heat exchange station hot water supply hydraulic prediction model to generate hydraulic operation prediction data for the heat exchange station within a preset time period includes: The real-time hydraulic operation record data of the heating network is normalized to obtain hydraulic operation sequence characteristic data. The roughness distribution matrix is ​​subjected to feature extraction to obtain roughness space feature data; Based on the hydraulic operation sequence characteristic data and the roughness spatial characteristic data, the data are processed by a preset heat exchange station hot water supply hydraulic prediction model to generate hydraulic operation prediction data for the heat exchange station within a preset time period.

6. The large model-based novel heat network hydraulic simulation method according to claim 5, characterized in that, The process of generating heating network hydraulic distribution data by processing the optimized adjacency matrix and hydraulic operation prediction data through a preset heating network hydraulic prediction model includes: The optimized adjacency matrix is ​​transformed to obtain the heating diagram structure data; The heating diagram structure data, hydraulic operation prediction data, and real-time operation record data are input into a preset heating network hydraulic prediction model for processing to generate heating network hydraulic distribution data. The hydraulic distribution data of the heating network includes pressure field distribution data, flow field distribution data, and temperature field distribution data.

7. A novel hydraulic simulation system for heating networks based on a large model, characterized in that, The system includes a memory and a processor. The memory contains a program for a novel hydraulic simulation method for heating networks based on a large model. When the processor executes the program for this novel hydraulic simulation method for heating networks based on a large model, it performs the following steps: Acquire historical hydraulic operation record data, heating network system operation record data, and heating network environment and topology data for a preset time period, and perform data preprocessing and data fusion to obtain the historical operation fusion feature vector of the heating network; Acquire real-time GIS topology data and compare it with preset baseline GIS topology data; generate an optimized adjacency matrix based on the comparison results. The image of the inner wall of the heating pipeline network is acquired and processed by a preset pipeline roughness recognition model to obtain the roughness distribution matrix of the inner wall of the heating pipeline network. Based on the roughness distribution matrix, the hydraulic prediction data of the heat exchange station's hot water supply is generated by processing the preset heat exchange station hot water supply hydraulic prediction model to generate hydraulic operation prediction data of the heat exchange station within a preset time period. Based on the optimized adjacency matrix and hydraulic operation prediction data, the data is processed through a preset heating network hydraulic prediction model to generate heating network hydraulic distribution data. The hydraulic distribution data of the heating network is visualized and rendered to generate a hydraulic simulation scene of the heating network, which is then output to the heating network dispatch center for display.

8. The novel heating network hydraulic simulation system based on a large model according to claim 7, characterized in that, The process of acquiring historical hydraulic operation record data, heating network system operation record data, and heating network environment and topology data for a preset time period, and performing data preprocessing and data fusion to obtain the historical operation fusion feature vector of the heating network includes: Acquire historical data on hydraulic operation of the heating network, operation of the heating network system, and environmental and topological data for a preset time period. The hydraulic operation data of the heating network includes temperature, flow rate, and pressure; The operating data of the heating network system includes vibration acceleration data and sound pressure level; The heating network environment and topology data include heating network pipeline surface status data and GIS topology data; The temperature, flow rate, and pressure are subjected to outlier removal and missing value linear interpolation to obtain optimized temperature, optimized flow rate, and optimized pressure. The vibration acceleration data is processed by short-time Fourier transform to obtain the vibration time-frequency matrix, and the Mel-frequency cepstral coefficients of the sound pressure level are extracted to generate the sound pressure Mel-frequency cepstral coefficients. The surface condition data and GIS topology data of the heating network pipeline are encoded to obtain surface condition feature vectors and topology feature vectors. The optimized temperature, optimized flow rate, optimized pressure, vibration time-frequency matrix, and sound pressure Mel-frequency cepstral coefficients are combined with the surface state feature vector and topological feature vector and processed through a preset multimodal data conversion model to obtain the historical operation fusion feature vector of the heating network.

9. The novel heating network hydraulic simulation system based on a large model according to claim 8, characterized in that, The process of acquiring real-time GIS topology data, comparing it with preset benchmark GIS topology data, and generating an optimized adjacency matrix based on the comparison results includes: Acquire real-time GIS topology data and construct a real-time heating topology map, including a real-time heating node set and a real-time heating edge set; A benchmark heating topology map is constructed based on the preset benchmark GIS topology data, including the benchmark heating node set and the benchmark heating edge set, and matrix transformation is performed to generate the benchmark adjacency matrix. The real-time heating node set and real-time heating edge set are processed with the reference heating node set and reference heating edge set using a preset Jaccard similarity algorithm to obtain a similarity value; If the similarity value is greater than a preset similarity threshold, then the baseline adjacency matrix is ​​an optimized adjacency matrix; If the similarity value is less than or equal to a preset similarity threshold, the baseline adjacency matrix is ​​optimized based on the real-time heating node set and the real-time heating edge set to generate an optimized adjacency matrix.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a novel hydraulic simulation method program for a large-scale heating network. When the program is executed by a processor, it implements the steps of a novel hydraulic simulation method for a large-scale heating network as described in any one of claims 1 to 6.

Images