Dynamic thermal integrated equivalent modeling method and device for regional heat supply network
Through directed graph division and optimization of regional heating networks, a hybrid semi-mechanical cluster segmented equivalence model is constructed, which solves the balance problem of high-precision dynamic thermal modeling and fast equivalence modeling in the existing technology, and improves heating Network management efficiency and accuracy.
Patent Information
- Application Number
- CN202411036436.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-31
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-07-31
AI Technical Summary
The prior art is difficult to achieve a balance between high-precision dynamic thermal modeling and fast equivalent modeling in regional heating networks, resulting in large amounts of calculation or insufficient measurement accuracy.
Through the division algorithm, the heating network is abstracted into a directed graph, and the preliminary cluster segmented equivalent model is obtained, and a hybrid semi-mechanical cluster segmented equivalent model is optimized and constructed, and the compensation is combined with the actual temperature data, and dynamic thermal integration of the regional thermal network is generated. Value model.
The combination of high-precision dynamic thermal modeling and fast equivalent modeling of the regional heating network is realized, which improves system design and management efficiency, increases flexibility and fault tolerance, and reduces modeling difficulty.
Smart Images

Figure CN119004966A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of heating systems, and in particular to a method and device for dynamic thermal integration equivalent modeling of a regional heating network. Background Art
[0002] The district heating network is an important part of the modern urban integrated energy network and has been widely used in residential life, industrial manufacturing and other scenarios in northern China. In order to fully share the heat transmission infrastructure, relevant regions usually gather to establish a district heating network. In order to ensure the safety, energy saving and low-carbon operation of the heating network, it is necessary to quickly and accurately model the dynamic thermal power of the heating network.
[0003] At present, there have been many studies on improving the penetration rate of new energy and energy utilization rate through the flexibility of digital modeling of heating systems. In some of these modeling studies, the internal dynamic thermal of the heating network is refined by using finite element methods represented by solving partial differential equations. However, this brings a huge amount of calculation to the operation of the system, because it is necessary to construct all the micro-element thermals of all the pipelines of the heating network according to the boundary conditions of the data to obtain the thermal state of the required nodes. In fact, in state estimation, most problems only need to obtain the estimated value of an internal node to express the measurement data of the actual sensor. In consideration of the calculation speed, in the problem of building the heating network model, another part of the scheme uses the lumped parameter equation representing the whole to perform equivalent modeling without considering the internal dynamics of the pipeline. This brings about the problem that when the network topology is too complex and the number of internal sensors is insufficient, the integrated equivalent model of multi-level connections is difficult to model and the measurement accuracy of the approximate processing scheme is insufficient. Therefore, how to comprehensively consider the local high-precision dynamic thermal modeling and the overall fast equivalent modeling to realize the integrated equivalent modeling of the dynamic thermal of the regional heating network is a technical problem that needs to be solved. Summary of the invention
[0004] In view of this, the purpose of this application is to provide a method and device for dynamic thermal integration equivalent modeling of a regional heating network, which effectively solves the problem of how to comprehensively consider local high-precision dynamic thermal modeling and overall rapid equivalent modeling to achieve dynamic thermal integration equivalent modeling of a regional heating network.
[0005] In a first aspect, an embodiment of the present application provides a method for dynamic thermal integration equivalent modeling of a district heating network, the method comprising:
[0006] Based on the partitioning algorithm, the directed graph abstracted from the heating network in the target area is partitioned to obtain a plurality of preliminary cluster segmented equivalent models containing unknown parameters; the edges of the directed graph are partitioned to obtain a plurality of intermediate nodes; the nodes of the directed graph are abstracted shunt nodes and pipeline nodes, and the edges of the directed graph are distances between shunt nodes and pipeline nodes;
[0007] Based on the characteristics of the heating network, unknown parameters in the preliminary cluster segmentation equivalent model are identified to obtain a preliminary cluster segmentation equivalent model without the unknown parameters, and the preliminary cluster segmentation equivalent model without the unknown parameters is optimized into a cluster segmentation equivalent model; the cluster segmentation equivalent model is used to output the predicted temperature data of the intermediate node; the parameters of the cluster segmentation equivalent model include the identification data delay;
[0008] According to the identification data delay parameter, the predicted temperature data output by the cluster segmentation equivalent model and the collected actual temperature data are selected as feature vectors, a compensation network for compensating the cluster segmentation equivalent model is constructed, and the compensation network and the cluster segmentation equivalent model are integrated to obtain a hybrid semi-mechanism cluster segmentation equivalent model;
[0009] Based on the hybrid semi-mechanism cluster segmented equivalent model, the hybrid state estimation value of the key node that meets the preset conditions in the intermediate node is calculated, and multiple hybrid semi-mechanism cluster segmented equivalent models are cascaded based on the hybrid state estimation value of the key node to obtain a dynamic thermal integration equivalent model of the regional heating network. The dynamic thermal integration equivalent model is used to control the thermal transfer of the heating network in the region.
[0010] In combination with the first aspect, the embodiment of the present application provides a first possible implementation of the first aspect, wherein a directed graph abstracted from a heating network in a region is partitioned based on a partitioning algorithm to obtain a plurality of preliminary cluster segmented equivalent models containing unknown parameters, including:
[0011] Designing node partitioning rules based on the connection relationship between nodes in the heating network, and inputting the node partitioning rules into the partitioning algorithm;
[0012] The partitioning algorithm divides the abstracted directed graph into segments based on the node partitioning rule to obtain a preliminary cluster segmentation equivalence model.
[0013] In combination with the first aspect, the embodiment of the present application provides a second possible implementation of the first aspect, wherein the partitioning algorithm partitions the abstracted directed graph based on the node partitioning rule to obtain a preliminary cluster segmentation equivalent model, including:
[0014] Obtain the pipeline input temperature and the ambient temperature of the intermediate nodes obtained by segmenting between the nodes of the directed graph;
[0015] The preliminary cluster segmented equivalent model processes the pipeline input temperature and the ambient temperature of the intermediate node to obtain the predicted temperature data of the intermediate node and the predicted temperature change at the next moment.
[0016] In combination with the first aspect, the embodiment of the present application provides a third possible implementation of the first aspect, wherein the step of identifying unknown parameters in the preliminary cluster segmented equivalent model based on the characteristics of the heating network to obtain a preliminary cluster segmented equivalent model that does not contain unknown parameters includes:
[0017] Input the number of actual temperature data and the number of intermediate nodes of the heating network obtained by statistics, the penalty value of the intermediate node when the preset penalty condition is met, and the deviation between the predicted temperature data and the actual temperature data of the terminal pipeline node output by the preliminary cluster segmented equivalent model into the model parameter identification network;
[0018] The model parameter identification network processes the number of actual temperature data of the heating network, the number of intermediate nodes, the penalty value of the intermediate nodes when the preset penalty conditions are met, and the deviation between the predicted temperature data and the actual temperature data of the terminal pipeline nodes output by the preliminary cluster segmented equivalent model, obtains the parameter values corresponding to the unknown parameters in the preliminary cluster segmented equivalent model, eliminates the unknown parameters, and obtains the preliminary cluster segmented equivalent model that does not contain unknown parameters.
[0019] In combination with the first aspect, the embodiment of the present application provides a fourth possible implementation of the first aspect, wherein the optimizing the preliminary cluster segmentation equivalent model that does not contain unknown parameters into the cluster segmentation equivalent model includes:
[0020] Fixing the parameter values output by the model parameter identification network in the preliminary cluster segmentation equivalent model, replacing the unknown parameters in the preliminary cluster segmentation equivalent model with the parameter values, so as to optimize the preliminary cluster segmentation equivalent model into the cluster segmentation equivalent model;
[0021] The cluster segmentation equivalence model is verified using a verification data set, and the cluster segmentation equivalence model is optimized again according to the verification result.
[0022] In combination with the first aspect, the embodiment of the present application provides a fifth possible implementation of the first aspect, wherein the constructing a compensation network for compensating the cluster segmentation equivalent model, and fusing the compensation network and the cluster segmentation equivalent model to obtain a hybrid semi-mechanism cluster segmentation equivalent model, includes:
[0023] The compensation network is operated and migrated based on the proportional factor to obtain a proportional compensation network after migration;
[0024] The proportion compensation network after the migration process and the predicted temperature data are fused to obtain a hybrid semi-mechanism cluster segmented equivalent model.
[0025] In combination with the first aspect, the embodiment of the present application provides a sixth possible implementation of the first aspect, wherein the hybrid state estimation value of the key node that meets the preset conditions in the intermediate node is calculated based on the hybrid semi-mechanism cluster segmented equivalent model, and multiple hybrid semi-mechanism cluster segmented equivalent models are cascaded based on the hybrid state estimation value of the key node to obtain a dynamic thermal integration equivalent model of the regional heating network, including:
[0026] Calling the dynamic heat transfer characteristics of the key nodes, the connection relationship of the key nodes, and the hybrid state estimation value of the key nodes calculated by the hybrid semi-mechanism cluster segmented equivalent model;
[0027] Based on the mixed state estimation values of the key nodes, the dynamic thermal transfer characteristics and the connection relationship of the key nodes, a dynamic thermal integration equivalent model of the regional heating network is established in the form of an equivalent model.
[0028] In a second aspect, an embodiment of the present application provides a device for dynamic thermal integration equivalent modeling of a regional heating network, the device comprising:
[0029] A partitioning module is used to partition the directed graph obtained by abstracting the heating network in the target area based on a partitioning algorithm to obtain a plurality of preliminary cluster segmentation equivalent models containing unknown parameters; the edges of the directed graph are partitioned to obtain a plurality of intermediate nodes; the nodes of the directed graph are abstracted shunt nodes and pipeline nodes, and the edges of the directed graph are distances between shunt nodes and pipeline nodes;
[0030] An identification module, used for identifying unknown parameters in the preliminary cluster segmentation equivalent model based on the characteristics of the heating network, obtaining a preliminary cluster segmentation equivalent model without the unknown parameters, and optimizing the preliminary cluster segmentation equivalent model without the unknown parameters into a cluster segmentation equivalent model; the cluster segmentation equivalent model is used to output the predicted temperature data of the intermediate node; the parameters of the cluster segmentation equivalent model include the identification data delay;
[0031] A fusion module, used for selecting the predicted temperature data output by the cluster segmentation equivalent model and the collected actual temperature data as feature vectors according to the identification data delay parameter, constructing a compensation network for compensating the cluster segmentation equivalent model, and fusing the compensation network and the cluster segmentation equivalent model to obtain a hybrid semi-mechanism cluster segmentation equivalent model;
[0032] A cascade module is used to calculate the hybrid state estimation value of the key node in the intermediate node from the hybrid semi-mechanism cluster segmented equivalent model, and to cascade multiple hybrid semi-mechanism cluster segmented equivalent models based on the hybrid state estimation value of the key node to obtain a dynamic thermal integration equivalent model of the regional heating network to control the thermal transfer of the heating network in the region.
[0033] In a third aspect, an embodiment of the present application provides an electronic device, comprising: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate through the bus, and when the machine-readable instructions are executed by the processor, any step of a method for dynamic thermal integration equivalent modeling of a regional heating network is performed.
[0034] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, executes the steps of a method for dynamic thermal integration equivalent modeling of a regional heating network as described in any one of claims 1 to 7.
[0035] The embodiment of the present application provides a method for dynamic thermal integration equivalent modeling of a regional heating network, the method comprising: dividing a directed graph abstracted from the heating network in the target area based on a partitioning algorithm to obtain a plurality of preliminary cluster segmented equivalent models containing unknown parameters; dividing the edges of the directed graph to obtain a plurality of intermediate nodes; the nodes of the directed graph are abstracted shunt nodes and pipeline nodes, and the edges of the directed graph are the distances between the shunt nodes and the pipeline nodes; identifying the unknown parameters in the preliminary cluster segmented equivalent model based on the characteristics of the heating network to obtain a preliminary cluster segmented equivalent model that does not contain unknown parameters, and optimizing the preliminary cluster segmented equivalent model that does not contain unknown parameters to a cluster segmented equivalent model; the cluster segmented equivalent model is used to output predicted temperature data of intermediate nodes; the parameters of the cluster segmented equivalent model include identification data delay; according to the identification data delay parameters, selecting the predicted temperature data output by the cluster segmented equivalent model and the actual temperature data collected as feature vectors to construct a method for compensating the cluster segmented equivalent model. A compensation network is prepared, and the compensation network and the cluster segmented equivalent model are integrated to obtain a hybrid semi-mechanism cluster segmented equivalent model; based on the hybrid semi-mechanism cluster segmented equivalent model, the hybrid state estimation value of the key node that meets the preset conditions in the intermediate node is calculated, and multiple hybrid semi-mechanism cluster segmented equivalent models are cascaded based on the hybrid state estimation value of the key node to obtain a dynamic thermal integration equivalent model of the regional heating network. The dynamic thermal integration equivalent model is used to control the thermal transfer of the heating network in the region, and realizes the dynamic thermal equivalent modeling of the regional heating network. It can provide system design, optimize and improve management efficiency, increase the flexibility and scalability of the regional heating network, and improve the fault tolerance and reliability of the heating network. It can also be used for arbitrary precision approximation and division of the actual characteristics of the actual regional heating system, reducing the difficulty of modeling, and can be combined with a neural network model of arbitrary complexity, thereby comprehensively considering the local high-precision dynamic thermal modeling and the overall fast equivalent modeling to achieve the effect of dynamic thermal integration equivalent modeling of the regional heating network. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.
[0037] Figure 1 A flow chart showing a first method for dynamic thermal integration equivalent modeling of a district heating network provided in an embodiment of the present application is shown;
[0038] Figure 2A schematic diagram of a district heating network directed graph provided by an embodiment of the present application is shown;
[0039] Figure 3 A diagram showing the result of dividing the directed graph nodes into clusters of the regional heating network provided by an embodiment of the present application is shown;
[0040] Figure 4 A schematic diagram showing the mixed state estimation result provided by an embodiment of the present application is shown;
[0041] Figure 5 A schematic diagram of the structure of a first district heating network dynamic thermal integration equivalent modeling device provided in an embodiment of the present application is shown;
[0042] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION
[0043] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the drawings in the present application only serve the purpose of explanation and description and are not used to limit the scope of protection of the present application. In addition, it should be understood that the schematic drawings are not drawn in real proportion. The flowchart used in this application shows the operations implemented according to some embodiments of the present application. It should be understood that the operations of the flowchart can be implemented out of sequence, and the steps without logical context can be reversed in order or implemented simultaneously. In addition, those skilled in the art can add one or more other operations to the flowchart under the guidance of the content of the present application, or remove one or more operations from the flowchart.
[0044] In addition, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.
[0045] It should be noted that the term "comprising" will be used in the embodiments of the present application to indicate the existence of the features declared thereafter, but does not exclude the addition of other features.
[0046] In the current state estimation of regional heating networks, most problems only require obtaining the estimated value of an internal node to express the actual sensor measurement data. However, due to the consideration of calculation speed, in the construction of the heating network model, another part of the scheme adopts the lumped parameter equation representing the whole to perform equivalent modeling without considering the internal dynamics of the pipeline. This brings about the problem that when the network topology is too complex and the number of internal sensors is insufficient, the integrated equivalent model of multi-level connections is difficult to model and the measurement accuracy of the approximate processing scheme is insufficient. Therefore, how to comprehensively consider the local high-precision dynamic thermal modeling and the overall fast equivalent modeling to realize the dynamic thermal integrated equivalent modeling of the regional heating network is a technical problem that needs to be solved urgently.
[0047] Based on this, an embodiment of the present application provides a method and device for dynamic thermal integration equivalent modeling of a regional heating network, which is described below through an embodiment.
[0048] Example 1
[0049] To facilitate understanding of this embodiment, a method for dynamic thermal integration equivalent modeling of a district heating network disclosed in an embodiment of this application is first introduced in detail. Figure 1 A flow chart of a method for dynamic thermal integration equivalent modeling of a regional heating network is shown. The present application provides a method for dynamic thermal integration equivalent modeling of a regional heating network, the method comprising:
[0050] S101, based on a partitioning algorithm, a directed graph abstracted from a heating network in a target area is partitioned to obtain a plurality of preliminary cluster segmented equivalent models containing unknown parameters; the edges of the directed graph are partitioned to obtain a plurality of intermediate nodes; the nodes of the directed graph are abstracted shunt nodes and pipeline nodes, and the edges of the directed graph are distances between shunt nodes and pipeline nodes;
[0051] S102, identifying unknown parameters in the preliminary cluster segmentation equivalent model based on the characteristics of the heating network, obtaining a preliminary cluster segmentation equivalent model without the unknown parameters, and optimizing the preliminary cluster segmentation equivalent model without the unknown parameters into a cluster segmentation equivalent model; the cluster segmentation equivalent model is used to output the predicted temperature data of the intermediate node; the parameters of the cluster segmentation equivalent model include the identification data delay;
[0052] S103, selecting the predicted temperature data output by the cluster segmentation equivalent model and the collected actual temperature data as feature vectors according to the identification data delay parameter, constructing a compensation network for compensating the cluster segmentation equivalent model, and fusing the compensation network and the cluster segmentation equivalent model to obtain a hybrid semi-mechanism cluster segmentation equivalent model;
[0053] S104. Based on the hybrid semi-mechanism cluster segmented equivalent model, the hybrid state estimation value of the key node that meets the preset conditions in the intermediate node is calculated, and multiple hybrid semi-mechanism cluster segmented equivalent models are cascaded based on the hybrid state estimation value of the key node to obtain a dynamic thermal integration equivalent model of the regional heating network. The dynamic thermal integration equivalent model is used to control the thermal transfer of the heating network in the region.
[0054] In step S101, the entire process of dynamic thermal equivalent modeling of the regional heating network is executed based on a computer or other intelligent device with data analysis and processing capabilities. The partitioning algorithm is preset in the computer or computer and can be called when the directed graph needs to be partitioned, thereby obtaining a plurality of preliminary cluster segmented equivalent models containing unknown parameters based on the partitioning algorithm; and the directed graph is obtained by abstracting the heating network in the target area, and the edges of the directed graph are divided to obtain a plurality of intermediate nodes; the directed graph is a form of graph theory, which strips off the multiple attributes of the edges in the traditional heating network graph theory expression in a unified form of heating network graph theory, and the nodes of the directed graph are abstracted shunt nodes and pipeline nodes, and the shunt nodes are the locations of various sensors arranged at the intersection of the heating pipelines, and the pipeline nodes are The directed graph is established on the edge of non-edge station nodes and non-pipeline intersection shunt nodes. The front end of the pipeline node is connected to a station node or shunt node that provides temperature input data, and the back end is also connected to a station node or shunt node that provides temperature output data. The edge of the directed graph is the distance between the shunt node and the pipeline node; the weight of the edge of the directed graph is obtained according to the distance between the station node, the shunt node, and the pipeline node. The terminal nodes in this article are heating boilers, combined heat and power stations, heat storage stations, and secondary heat network heat exchange stations. The distance between the station node and the shunt node to the nearest connected pipeline node is 0, and the pipeline node to the next node is the complete edge distance, that is, the complete distance is taken before and after this pipeline node and on the subsequent edges to avoid the problem of repeated values. The directed graph is represented by the adjacency matrix (e ij ) n×n The abstract directed graph is G(V,E,Φ), where V represents the set of nodes, E represents the set of edges of the directed graph, the edge is the distance between the diversion node and the pipeline node to which it flows, Φ represents the node connection relationship, and e represents the set of edges of the diversion node. ij Expressed by formula (1):
[0055]
[0056] Among them, w ijis the distance between the ith node and the jth node, G(V,E,Φ) is a directed graph, and it is stipulated that in the diagonal matrix, the edge weight between the pipeline nodes of the station node and the diversion node is 0. The directed graph is expressed in the form of an adjacency matrix, which can be uniformly calculated and processed on the computer, providing a reference scheme for the connection relationship and key node selection of the heat network integrated equivalent model, where the key node is the intermediate node where the intersection point is located in the intermediate nodes obtained by dividing two edges with only one intersection point.
[0057] Among them, the pipeline node also converts equipment such as expansion joints and three-way valves into equivalent pipeline lengths, expressed by formulas (2)-(4):
[0058]
[0059] λ=f(Re,ε) (3)
[0060]
[0061] Among them, S j represents the local resistivity; d represents the diameter of the heating pipe; the friction coefficient λ of the heating pipe depends on the flow state of the hot water in the heating pipe and the roughness of the pipe wall; Re represents the Reynolds number, which is a dimensionless number that characterizes the flow state of the fluid; ε represents the relative roughness of the heating pipe wall; v is the viscosity coefficient of the medium movement, and ΔL is the equivalent conversion of equipment such as expansion joints and three-way valves into pipe length, that is, the length of the added edge of the directed graph. Through the above unified expression, any heating network equipment can be equivalently converted into a pipe and expressed in the form of a pipe model node, which can effectively express its thermal and hydraulic transmission characteristics as a connecting node.
[0062] In the specific implementation process of step S101, there is an embodiment of: dividing the directed graph abstracted from the heating network in the region based on a partitioning algorithm to obtain a plurality of preliminary cluster segmented equivalent models containing unknown parameters, including:
[0063] S1011, designing a node partitioning rule based on the connection relationship between nodes in the heating network, and inputting the node partitioning rule into a partitioning algorithm;
[0064] S1012: The partitioning algorithm divides the abstracted directed graph into segments based on the node partitioning rule to obtain a preliminary cluster segmentation equivalence model.
[0065] In steps S1011-S1012, the partitioning algorithm pre-stored in the computer can be the Tarjan algorithm based on depth-first search. The Tarjan algorithm was originally used for tree-shaped search of strongly connected components in directed graphs. By modifying the component partitioning conditions of the Tarjan algorithm, it can be applied to any heating network structure based on graph search and partitioning, thereby obtaining multiple node clusters after downscaling and deconstructing the complex heating network. These clusters provide support for local equivalent modeling schemes, and at the same time provide node locations for connection of each cluster, providing a reference for the dynamic thermal integration equivalent modeling scheme of the regional heating network.
[0066] Combined with the on-site graph structure of the heating network, a node division rule based on the graph theory expression of the heating network is designed to obtain multiple node cluster components. According to the established directed graph adjacency matrix, the following division conditions are available: (1) The head and tail of the cluster must be terminal nodes, that is, the head and tail should be the heat source node and the heat exchange station node; (2) When one end of the search has been used, the cluster is divided end-to-end from the nearest diversion node; (3) When the distance between the two endpoints is short, a separate cluster should be divided; (4) When the diversion node is at the end of the pipeline, the diversion node and the terminal node are divided as a whole.
[0067] The above division conditions correspond to the following designs respectively: (1) The proposed segmented equivalent model only uses terminal nodes for equivalent modeling; (2) A highly controllable heat network integration model without repeated paths is established, which can be achieved by using shunt nodes to connect clusters instead of relying on approximate methods to use repeated path modeling; (3) When the distance between the two endpoints is short, it shows that the terminal measurement point data can effectively approximate the shunt node, which improves the quality of the estimated value of the shunt node; (4) A suitable processing scheme is provided for cluster division at the end of the pipeline.
[0068] In the specific implementation process of step S1012, there is an embodiment in which the partitioning algorithm partitions the abstracted directed graph based on the node partitioning rule to obtain a preliminary cluster segmentation equivalent model, including:
[0069] S10121. Obtain pipeline input temperature and ambient temperature of intermediate nodes obtained by segmenting between nodes of the directed graph;
[0070] S10122. The preliminary cluster segmented equivalent model processes the pipeline input temperature and the ambient temperature of the intermediate node to obtain the predicted temperature data of the intermediate node and the predicted temperature change at the next moment.
[0071] In steps S10121-S10122, the preliminary cluster segmentation equivalent model is expressed by formula (5):
[0072]
[0073] Among them, X SS (k) represents the state vector of each node in the pipeline at time k. In this application, the state vector refers to the temperature. Assuming that there are N-1 intermediate nodes in the node cluster, X SS (k) = [x SS1 x SS2 …x SSN ] T Represents the predicted temperature data of each section of the pipeline after segmentation, U(k)=[u 1 (k)T amb (k)] T It represents the temperature data of the divided pipeline corresponding to the preliminary cluster segmentation equivalent model, and the pipeline input temperature u of the intermediate node 1 (k) and ambient temperature T amb (k), the pipeline input temperature is estimated based on the temperature data measured by the terminal node using a sensor or the like, Y d (k) represents the output predicted temperature data, is the predicted temperature change at the next moment, A, B, and C are the system matrix, input matrix, and state transfer matrix, respectively, which are expressed by formulas (6)-(8):
[0074]
[0075]
[0076] Where a and b represent the parameters of the system matrix A and the input matrix B in the segmented model.
[0077] The heat exchange station, heat source, expansion joint, three-way valve and other heat network equipment in the terminal node are equivalently expressed as follows Figure 2 As shown in the figure, the complex heat network topology is expressed in a unified graph theory and converted into an adjacency matrix that can be processed by a computer as follows (e ij ) n×n :
[0078]
[0079] Among them, e ijRepresents the length relationship between intermediate node i and intermediate node j. Subsequently, the Tarjan graph clustering algorithm with improved component partitioning rules for regional heating networks is used to perform deep node clustering on the directed graph expressed by the regional heating network. The Tarjan algorithm is an algorithm that initially solves the linear time of strongly connected components of a directed graph. It is based on a depth-first search algorithm for the graph, and each strongly connected component is a subtree in the search tree. In the application of regional heating networks, this method adaptively modifies the Tarjan algorithm so that it can handle the special properties of the graph theory of regional heating networks. First, the single heat source heating network is expressed as an adjacency matrix (e ij ) n×n Then, by applying the modified Tarjan algorithm, the adjacency matrix is divided into multiple node clusters. Each cluster contains one or more key nodes for interconnecting clusters, and these key nodes share the same pipeline thermal transfer equivalent characteristics. In this way, the complex thermal network equivalent modeling problem can be simplified into multiple smaller sub-model problems, and each sub-problem only involves one node cluster. Figure 2 The adjacency matrix of the single heat source heating network shown is divided into Figure 3 There are a total of 20 node clusters shown. Among them, HS represents the heat source node, S represents the pipeline model visualization equivalent node proposed by this method, SM represents the key node used to connect each node cluster, HE represents the heat exchange station node containing measurement data, A represents the abbreviated point, which contains multiple divided node clusters, Cluster represents the node cluster result divided by the adjacency matrix and component construction rules constructed according to the field topology, and the dotted line represents the result of the hybrid state estimation of the key nodes provided by this method corresponding to the actual physical distribution on the district heating network structure diagram.
[0080] In step S102, unknown parameters in the preliminary cluster segmented equivalent model are identified based on the characteristics of the heating network. If the most important characteristic of the heating network is temperature, the unknown parameters in the preliminary cluster segmented equivalent model are identified based on temperature. The unknown parameters in the preliminary cluster segmented equivalent model are replaced with the parameter values of the identified unknown parameters to obtain a preliminary cluster segmented equivalent model that does not contain unknown parameters. The preliminary cluster segmented equivalent model that does not contain unknown parameters is optimized to a cluster segmented equivalent model. The cluster segmented equivalent model is used to output predicted temperature data of intermediate nodes, that is, the predicted temperature data is not the actual temperature of the heating network. There are multiple parameters in the cluster segmented equivalent model, and the parameter of identification data delay is selected for use.
[0081] In the specific implementation process of step S102, there is an embodiment in which the unknown parameters in the preliminary cluster segmentation equivalent model are identified based on the characteristics of the heating network to obtain the preliminary cluster segmentation equivalent model that does not contain the unknown parameters, including:
[0082] S10211, inputting the number of actual temperature data of the heating network and the number of intermediate nodes obtained by statistics, as well as the penalty value of the intermediate node when the preset penalty condition is met and the deviation between the predicted temperature data of the terminal pipeline node output by the preliminary cluster segmented equivalent model and the actual temperature data into the model parameter identification network;
[0083] S10212. The model parameter identification network processes the number of actual temperature data of the heating network, the number of intermediate nodes, the penalty value of the intermediate nodes when the preset penalty conditions are met, and the deviation between the predicted temperature data and the actual temperature data of the terminal pipeline nodes output by the preliminary cluster segmented equivalent model, obtains the parameter values corresponding to the unknown parameters in the preliminary cluster segmented equivalent model, eliminates the unknown parameters, and obtains the preliminary cluster segmented equivalent model that does not contain unknown parameters.
[0084] In steps S1021-S1022, the model parameter identification network pre-stored in the computer is represented by an objective function, and the objective function of the model parameter identification network is represented by formula (9):
[0085]
[0086] Wherein, obj is the parameter value corresponding to the unknown parameter, N represents the number of predicted temperature data of the preliminary cluster segmentation equivalent model, error represents the deviation between the predicted temperature data of the terminal pipeline node output by the preliminary cluster segmentation equivalent model and the actual temperature data, Xest represents the number of intermediate nodes, and penalty represents the penalty value of the intermediate node i for the path exceeding the boundary condition and the identification of the mismatched mechanism parameter, and the actual temperature data is collected by the district heating network shown using a temperature sensor or other methods, wherein:
[0087]
[0088] Among them, T represents the sampling time step, state k represents the recognition result of the predicted temperature data of the terminal pipeline node at time k output by the preliminary cluster segmentation equivalent model, state k est represents the recognition result of the actual temperature of the intermediate node k, H represents the boundary condition of the intermediate node i, and a iirepresents the value of row i and column i in the system matrix A identified by the model parameter identification network, α i represents the convective heat transfer coefficient of the pipeline segment at the intermediate node i, C w represents specific heat capacity, ui w represents the flow rate of the pipe section, L i Indicates the circumference of the pipe section, d i represents the length of the pipe section, S represents the number of intermediate nodes with constructed mechanism constraints, and the model parameter identification network here can be understood as: (1) the error term represents the difference between the predicted temperature data output by the preliminary cluster segmented equivalent model constructed by the model parameter value identified by the model parameter identification network and the actual temperature data, which is used to ensure that the output of the equivalent model is close to the true value; (2) the penalty term represents the temperature data of the intermediate nodes with internal mechanism constraints identified by the model parameter identification network. The temperature data should be within the data boundary, and the identification parameters should satisfy the mechanism constraints as much as possible.
[0089] For the optimization problem of a single unknown parameter proposed by the model parameter identification network, the simulated annealing algorithm is used to solve the optimal value of the unknown parameter, and the parameter identification of the cluster segmented equivalent model is realized from the perspective of global search.
[0090] In the specific implementation process of step S102, there is another embodiment: optimizing the preliminary cluster segmentation equivalent model that does not contain unknown parameters into a cluster segmentation equivalent model includes:
[0091] S10221, fixing the parameter values output by the model parameter identification network in the preliminary cluster segmentation equivalent model, replacing the unknown parameters in the preliminary cluster segmentation equivalent model with the parameter values, so as to optimize the preliminary cluster segmentation equivalent model into the cluster segmentation equivalent model;
[0092] S10222. Use a verification data set to verify the cluster segmentation equivalence model, and optimize the cluster segmentation equivalence model again according to the verification result.
[0093] In steps S10221-S10222, the cluster segmentation equivalent model obtained based on the optimization of the preliminary cluster segmentation equivalent model is expressed by formula (13):
[0094]
[0095] Where P = -u w / d-αL / Fc w ρ w ,u(k)=[T in w(k)T amb (k)],u wrepresents the instantaneous flow velocity of the pipeline in m / s, Tamb(t) is the ambient temperature in °C, α is the convection heat transfer coefficient with respect to the environment in J / (㎡·℃·s), L is the circumference of the pipeline in m, d represents the total length of the pipeline in m, and F is the cross-sectional area of the pipeline in m 2 , c w is the specific heat capacity of water, in J / ℃, ρ w is the water density in kg / m 3 , at this time, through the temperature x ss (k) According to the structure of the matrix-modified state transfer matrix C, the predicted temperature data y(k) can be transformed into the predicted temperature data for the key nodes; and the cluster segmented equivalence model is verified using a pre-stored verification data set, the verification data set is the actual temperature data collected historically from the regional heating network, if the deviation between the output predicted temperature data and the actual temperature data is within a preset range, the verification result is successful, otherwise it is a failure, and when the verification result is a failure, the cluster segmented equivalence model is optimized again to ensure the accuracy of the cluster segmented equivalence model.
[0096] In step S103, the cluster segmented equivalent model is established for each intermediate node, and in actual use, all models will generate delays during operation, and the cluster segmented equivalent model generates multiple parameters when identifying the input data, one of which is the identification data delay parameter, and also generates nonlinear errors. Then, the predicted temperature data output by the cluster segmented equivalent model and the collected actual temperature data are selected as feature vectors for the identification data delay parameter, and a compensation network is constructed to compensate for the cluster segmented equivalent model, and the compensation network and the cluster segmented equivalent model are integrated to obtain a hybrid semi-mechanism cluster segmented equivalent model. At this time, the predicted temperature data output by the hybrid semi-mechanism cluster segmented equivalent model is closer to the actual temperature data, which greatly improves the accuracy of the output predicted temperature data.
[0097] The nonlinear error generated by the cluster segmentation equivalent model is expressed by formula (14):
[0098] A(z -1 )y(k)=B(z-1)u(k)+δ(k) (14)
[0099] The intermediate node receives the input temperature of the heat or water output from other intermediate nodes. (k), the output temperature y of the intermediate node to other intermediate nodes real(k) and the autoregression of the error e(k) between the predicted temperature data and the actual temperature data output by the cluster segmented equivalent model are used as input vectors and output vectors to train the neural network model, which is expressed by formulas (15)-(16):
[0100] U com (k)=[u(k)u(k-2)…u(kn b )y(k-1)y(k-2)…y(kn a )e(k-1)e(k-2)…e(kn a )] (15)
[0101] Y com (k) = e(k) (16)
[0102] Where e is the error term, input temperature U com (k) and output temperature Y com (k) can be expanded into a multi-output form, wherein the identification data delay parameter includes the delay order n of the cluster segmented equivalent model output predicted temperature data a and the delay level n of the input temperature b According to the computing performance, the longer the delay of the characteristics of the heating network, the more data, the better the nonlinear compensation performance, and the slower the computing speed, a compensation network is constructed to compensate for the cluster segmented equivalent model, wherein the compensation network is a timing neural network.
[0103] In the specific implementation process of step S103, there is an embodiment in which: the construction of a compensation network for compensating the cluster segmentation equivalent model, and fusing the compensation network and the cluster segmentation equivalent model to obtain a hybrid semi-mechanism cluster segmentation equivalent model includes:
[0104] S1031, performing calculation and migration processing on the compensation network based on the proportional factor to obtain a proportional compensation network after migration processing;
[0105] S1032, fusing the migrated proportional compensation network and the predicted temperature data to obtain a hybrid semi-mechanism cluster segmented equivalent model.
[0106] In steps S1031-S1032, the predicted temperature data after migration processing is merged with the compensation network after proportional calculation to obtain the compensation item of the predicted temperature data of the intermediate node, which is expressed by formula (17):
[0107]
[0108] Among them, e cp,trans (k) is the compensation value e of the compensation network cp(k) The compensation value transmitted to the intermediate node at time k after compensating the heating equipment in the terminal node, d forward Represents the distance from the input node to the intermediate node, where the input node is, other nodes for input hydraulic or thermal, d total represents the total length of the segmented pipeline, and the hybrid semi-mechanism cluster segmented equivalent model is obtained by integrating the compensation network and the cluster segmented equivalent model, which is expressed by formula (18):
[0109]
[0110] where g(k) is the data-driven compensation term output by the neural network model based on the bias dynamics, g(k) = [e cp,trans (k), e cp (k)] By establishing the compensation network and the cluster segmented equivalent model, a hybrid semi-mechanism segmented equivalent model for the node cluster is established, and the hybrid state estimation value of the internal node under the cluster equivalent model based on the heat exchange station boundary node data included in the terminal node and the mechanism constraint is obtained, which provides a means for obtaining the hybrid state estimation value, i.e., the predicted temperature data, for the key nodes of the regional heating network when the internal measurement points are insufficient, thereby providing effective support for the rapid cascade equivalent modeling of the regional heating network.
[0111] By identifying the internal transmission delay under the condition of fixed flow rate, aligning the time axis to process the delay relationship between the intermediate node and the output node, and the migration compensation result of the proportional processing neural network output value, the mixed state estimation value of the estimated node in the node cluster is obtained. It should be noted that the state estimation result of the preliminary identification can be directly obtained from the segmented state quantity x of the model. ss (k) proposes that the state estimation of the hybrid semi-mechanism needs to be performed by aligning the segmented delays under fixed flow rate and transferring the output of the neural network to e-data. cp,trans (k) Compensation, performance of hybrid state estimation Figure 4 shown.
[0112] In step S104, based on the hybrid semi-mechanism cluster segmented equivalent model, the hybrid state estimation value of the key node in the intermediate node that meets the preset conditions is calculated, and the hybrid state estimation value is the predicted temperature data, that is, the key node is the intermediate node where the intersection point is located in the intermediate nodes obtained by dividing two edges with only one intersection point, that is, the hybrid semi-mechanism cluster segmented equivalent model is established for the key node including a section of pipeline and the diversion node, and then the predicted temperature data of this key node is used as the input of the hybrid semi-mechanism cluster segmented equivalent model of the next key node, and multiple hybrid semi-mechanism cluster segmented equivalent models are cascaded based on the hybrid state estimation value of the key node to obtain a dynamic thermal integration equivalent model of the regional heating network, and the dynamic thermal integration equivalent model is used to control the thermal transfer of the heating network in the region.
[0113] In the specific implementation process of step S04, there is an embodiment in which: the hybrid state estimation value of the key node that meets the preset conditions in the intermediate node is calculated based on the hybrid semi-mechanism cluster segmented equivalent model, and multiple hybrid semi-mechanism cluster segmented equivalent models are cascaded based on the hybrid state estimation value of the key node to obtain a dynamic thermal integration equivalent model of the regional heating network, including:
[0114] S1041, calling the dynamic heat transfer characteristics of the key nodes, the connection relationship of the key nodes, and the hybrid state estimation value of the key nodes calculated by the hybrid semi-mechanism cluster segmented equivalent model;
[0115] S1042. A dynamic thermal integration equivalent model of the regional heating network is established in the form of an equivalent model based on the estimated values of the mixed states of the key nodes, the dynamic thermal transfer characteristics and the connection relationship of the key nodes.
[0116] In steps S1041-S1042, a dynamic thermal integration equivalent model of the district heating network is established and expressed by formula (19):
[0117]
[0118] Among them, M represents the number of node clusters divided by the Tarjan algorithm with improved component division rules. The dynamic thermal integration equivalent model of the regional heating network established above includes the dynamic thermal transfer characteristics of the key nodes, the connection relationship of the key nodes, and the mixed state estimation value of the key nodes calculated by the hybrid semi-mechanism cluster segmented equivalent model, thereby further ensuring the accuracy of the established dynamic thermal integration equivalent model of the regional heating network, ensuring its adaptability to the regional heating network, and being able to accurately reflect the temperature of thermal or hydraulic transfer of the regional heating network.
[0119] Based on different heat network dynamic thermal equivalent modeling methods and this method, the terminal output results of the model are evaluated by the mean absolute error MAE, root mean square error RMSE and mean absolute percentage error MAPE, which are calculated by formulas (20)-(22):
[0120]
[0121] In the formula, Represents the output prediction value, y i is the output actual value, N is the number of predicted samples, and it can be concluded that the final output effect of the parameter identification model obtained by solving the optimization problem after compensation of different neural network models is shown in Table 1. It can be seen that the effect of this method is better than other methods.
[0122] Table 1 Performance of thermal network dynamic equivalent modeling using different network model compensations in a 7-day test
[0123]
[0124] Example 2
[0125] The present application also provides a dynamic thermal integration equivalent modeling device for a district heating network, such as Figure 5 The block diagram of a device for dynamic thermal integration equivalent modeling of a district heating network is shown. The functions implemented by the device for dynamic thermal integration equivalent modeling of a district heating network correspond to the steps of executing a method for dynamic thermal integration equivalent modeling of a district heating network on a terminal device. The device can be understood as a component of a server including a processor, and the device includes:
[0126] The partitioning module 501 is used to partition the directed graph obtained by abstracting the heating network in the target area based on the partitioning algorithm to obtain multiple preliminary cluster segmentation equivalent models containing unknown parameters; the edges of the directed graph are partitioned to obtain multiple intermediate nodes; the nodes of the directed graph are the abstracted shunt nodes and pipeline nodes, and the edges of the directed graph are the distances between the shunt nodes and the pipeline nodes;
[0127] The identification module 502 is used to identify unknown parameters in the preliminary cluster segmentation equivalent model based on the characteristics of the heating network, obtain a preliminary cluster segmentation equivalent model without unknown parameters, and optimize the preliminary cluster segmentation equivalent model without unknown parameters into a cluster segmentation equivalent model; the cluster segmentation equivalent model is used to output the predicted temperature data of the intermediate node; the parameters of the cluster segmentation equivalent model include the identification data delay;
[0128] A fusion module 503 is used to select the predicted temperature data output by the cluster segmentation equivalent model and the collected actual temperature data as feature vectors according to the identification data delay parameter, construct a compensation network for compensating the cluster segmentation equivalent model, and fuse the compensation network and the cluster segmentation equivalent model to obtain a hybrid semi-mechanism cluster segmentation equivalent model;
[0129] The cascade module 504 is used to calculate the hybrid state estimation value of the key node in the intermediate node from the hybrid semi-mechanism cluster segmented equivalent model, and cascade multiple hybrid semi-mechanism cluster segmented equivalent models based on the hybrid state estimation value of the key node to obtain a dynamic thermal integration equivalent model of the regional heating network to control the thermal transfer of the heating network in the region.
[0130] In a feasible implementation, the division modules include:
[0131] A design module, for designing node partitioning rules based on the connection relationship between nodes in the heating network, and inputting the node partitioning rules into the partitioning algorithm;
[0132] The abstract module is used for the partitioning algorithm to segment the abstracted directed graph based on the node partitioning rule to obtain a preliminary cluster segmentation equivalence model.
[0133] In a feasible implementation manner, the division module further includes:
[0134] An acquisition module, used to acquire the pipeline input temperature and the ambient temperature of the intermediate nodes obtained by segmenting between the nodes of the directed graph;
[0135] The prediction module is used for processing the pipeline input temperature and the ambient temperature of the intermediate node by the preliminary cluster segmentation equivalent model to obtain the predicted temperature data of the intermediate node and the predicted temperature change at the next moment.
[0136] In a feasible implementation manner, the identification module includes:
[0137] A statistical module, used for inputting the number of actual temperature data of the heating network and the number of intermediate nodes obtained by statistics, as well as the penalty value of the intermediate node when the preset penalty condition is met and the deviation between the predicted temperature data of the terminal pipeline node output by the preliminary cluster segmented equivalent model and the actual temperature data, into the model parameter identification network;
[0138] A processing module is used for the model parameter identification network to process the number of actual temperature data of the heating network, the number of intermediate nodes, the penalty value of the intermediate nodes when the preset penalty conditions are met, and the deviation between the predicted temperature data and the actual temperature data of the terminal pipeline nodes output by the preliminary cluster segmented equivalent model, obtain the parameter values corresponding to the unknown parameters in the preliminary cluster segmented equivalent model, eliminate the unknown parameters, and obtain the preliminary cluster segmented equivalent model that does not contain unknown parameters.
[0139] In a feasible implementation manner, the identification module further includes:
[0140] A fixing module, used for fixing the parameter values output by the model parameter identification network in the preliminary cluster segmentation equivalent model, replacing the unknown parameters in the preliminary cluster segmentation equivalent model with the parameter values, so as to optimize the preliminary cluster segmentation equivalent model into the cluster segmentation equivalent model;
[0141] The verification module is used to verify the cluster segmentation equivalence model using a verification data set, and to optimize the cluster segmentation equivalence model again according to the verification result.
[0142] In a feasible implementation, the fusion module includes:
[0143] A calculation module, used for calculating and migrating the compensation network based on the proportional factor to obtain a proportional compensation network after the migration process;
[0144] The migration module is used to fuse the migrated proportional compensation network and the predicted temperature data to obtain a hybrid semi-mechanism cluster segmented equivalent model.
[0145] In a feasible implementation manner, the cascade module includes:
[0146] A calling module, used for calling the dynamic heat transfer characteristics of the key nodes, the connection relationship of the key nodes and the hybrid state estimation value of the key nodes calculated by the hybrid semi-mechanism cluster segmented equivalent model;
[0147] A module is established for establishing a dynamic thermal integration equivalent model of a regional heating network in the form of an equivalent model based on the mixed state estimation value of the key nodes, the dynamic thermal transfer characteristics and the connection relationship of the key nodes.
[0148] Example 3
[0149] The present application embodiment provides an electronic device, such as Figure 6As shown, it includes: a processor 601, a memory 602 and a bus 603, the memory 602 stores machine-readable instructions executable by the processor 601, when the electronic device is running, the processor 601 communicates with the memory 602 through the bus 603, and when the machine-readable instructions are executed by the processor 601, the steps of any one of the methods for dynamic thermal integration equivalent modeling of a regional heating network are performed.
[0150] Example 4
[0151] A computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the steps of any one of the methods for dynamic thermal integration equivalent modeling of a regional heating network are executed.
[0152] Those skilled in the art can clearly understand that, for the convenience and simplicity of description, the specific working process of the system and device described above can refer to the corresponding process in the method embodiment, and will not be repeated in this application. In the several embodiments provided in this application, it should be understood that the disclosed system, device and method can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.
[0153] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0154] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
[0155] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a platform server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard drives, ROM, RAM, magnetic disks, or optical disks.
[0156] The above are only specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
Claims
1. A method for dynamic thermal integration equivalent modeling of a district heating network, characterized in that: The method comprises: Based on the partitioning algorithm, the directed graph obtained by abstracting the heating network in the target area is partitioned to obtain multiple preliminary cluster segmentation equivalent models containing unknown parameters; the edges of the directed graph are partitioned to obtain multiple intermediate nodes; the nodes of the directed graph are abstracted shunt nodes and pipeline nodes, and the edges of the directed graph are the distances between the shunt nodes and the pipeline nodes; Based on the characteristics of the heating network, unknown parameters in the preliminary cluster segmentation equivalent model are identified to obtain a preliminary cluster segmentation equivalent model without the unknown parameters, and the preliminary cluster segmentation equivalent model without the unknown parameters is optimized into a cluster segmentation equivalent model; the cluster segmentation equivalent model is used to output the predicted temperature data of the intermediate node; the parameters of the cluster segmentation equivalent model include the identification data delay; According to the identification data delay parameter, the predicted temperature data output by the cluster segmentation equivalent model and the collected actual temperature data are selected as feature vectors, a compensation network for compensating the cluster segmentation equivalent model is constructed, and the compensation network and the cluster segmentation equivalent model are integrated to obtain a hybrid semi-mechanism cluster segmentation equivalent model; Based on the hybrid semi-mechanism cluster segmented equivalent model, the hybrid state estimation value of the key node that meets the preset conditions in the intermediate node is calculated, and multiple hybrid semi-mechanism cluster segmented equivalent models are cascaded based on the hybrid state estimation value of the key node to obtain a dynamic thermal integration equivalent model of the regional heating network. The dynamic thermal integration equivalent model is used to control the thermal transfer of the heating network in the region.
2. The method according to claim 1, characterized in that Based on the partitioning algorithm, the directed graph abstracted from the heating network in the region is partitioned to obtain multiple preliminary cluster segmentation equivalent models containing unknown parameters, including: Designing node partitioning rules based on the connection relationship between nodes in the heating network, and inputting the node partitioning rules into the partitioning algorithm; The partitioning algorithm divides the abstracted directed graph into segments based on the node partitioning rule to obtain a preliminary cluster segmentation equivalence model.
3. The method according to claim 2, characterized in that The partitioning algorithm partitions the abstracted directed graph based on the node partitioning rule to obtain a preliminary cluster segmentation equivalent model, including: Obtain the pipeline input temperature and the ambient temperature of the intermediate nodes obtained by segmenting between the nodes of the directed graph; The preliminary cluster segmented equivalent model processes the pipeline input temperature and the ambient temperature of the intermediate node to obtain the predicted temperature data of the intermediate node and the predicted temperature change at the next moment.
4. The method according to claim 1, characterized in that The method of identifying unknown parameters in the preliminary cluster segmentation equivalent model based on the characteristics of the heating network to obtain a preliminary cluster segmentation equivalent model that does not contain unknown parameters includes: Input the number of actual temperature data and the number of intermediate nodes of the heating network obtained by statistics, the penalty value of the intermediate node when the preset penalty condition is met, and the deviation between the predicted temperature data and the actual temperature data of the terminal pipeline node output by the preliminary cluster segmentation equivalent model into the model parameter identification network; The model parameter identification network processes the number of actual temperature data of the heating network, the number of intermediate nodes, the penalty value of the intermediate nodes when the preset penalty conditions are met, and the deviation between the predicted temperature data and the actual temperature data of the terminal pipeline nodes output by the preliminary cluster segmented equivalent model, obtains the parameter values corresponding to the unknown parameters in the preliminary cluster segmented equivalent model, eliminates the unknown parameters, and obtains the preliminary cluster segmented equivalent model that does not contain unknown parameters.
5. The method according to claim 1, characterized in that The step of optimizing the preliminary cluster segmentation equivalent model that does not contain unknown parameters into a cluster segmentation equivalent model comprises: Fixing the parameter values output by the model parameter identification network in the preliminary cluster segmentation equivalent model, replacing the unknown parameters in the preliminary cluster segmentation equivalent model with the parameter values, so as to optimize the preliminary cluster segmentation equivalent model into the cluster segmentation equivalent model; The cluster segmentation equivalence model is verified using a verification data set, and the cluster segmentation equivalence model is optimized again according to the verification result.
6. The method according to claim 1, characterized in that The constructing of a compensation network for compensating the cluster segmentation equivalent model, and fusing the compensation network and the cluster segmentation equivalent model to obtain a hybrid semi-mechanism cluster segmentation equivalent model, includes: The compensation network is operated and migrated based on the proportional factor to obtain a proportional compensation network after migration; The proportion compensation network after the migration process and the predicted temperature data are fused to obtain a hybrid semi-mechanism cluster segmented equivalent model.
7. The method according to claim 1, characterized in that The hybrid state estimation value of the key node that meets the preset conditions in the intermediate node is calculated based on the hybrid semi-mechanism cluster segmented equivalent model, and multiple hybrid semi-mechanism cluster segmented equivalent models are cascaded based on the hybrid state estimation value of the key node to obtain a dynamic thermal integration equivalent model of the regional heating network, including: Calling the dynamic heat transfer characteristics of the key nodes, the connection relationship of the key nodes, and the hybrid state estimation value of the key nodes calculated by the hybrid semi-mechanism cluster segmented equivalent model; Based on the mixed state estimation values of the key nodes, the dynamic thermal transfer characteristics and the connection relationship of the key nodes, a dynamic thermal integration equivalent model of the regional heating network is established in the form of an equivalent model.
8. A dynamic thermal integration equivalent modeling device for a regional heating network, characterized in that: The device comprises: A partitioning module is used to partition the directed graph obtained by abstracting the heating network in the target area based on a partitioning algorithm to obtain a plurality of preliminary cluster segmentation equivalent models containing unknown parameters; the edges of the directed graph are partitioned to obtain a plurality of intermediate nodes; the nodes of the directed graph are abstracted shunt nodes and pipeline nodes, and the edges of the directed graph are distances between shunt nodes and pipeline nodes; An identification module, used for identifying unknown parameters in the preliminary cluster segmentation equivalent model based on the characteristics of the heating network, obtaining a preliminary cluster segmentation equivalent model without the unknown parameters, and optimizing the preliminary cluster segmentation equivalent model without the unknown parameters into a cluster segmentation equivalent model; the cluster segmentation equivalent model is used to output the predicted temperature data of the intermediate node; the parameters of the cluster segmentation equivalent model include the identification data delay; A fusion module, used for selecting the predicted temperature data output by the cluster segmentation equivalent model and the collected actual temperature data as feature vectors according to the identification data delay parameter, constructing a compensation network for compensating the cluster segmentation equivalent model, and fusing the compensation network and the cluster segmentation equivalent model to obtain a hybrid semi-mechanism cluster segmentation equivalent model; A cascade module is used to calculate the hybrid state estimation value of the key node in the intermediate node from the hybrid semi-mechanism cluster segmented equivalent model, and to cascade multiple hybrid semi-mechanism cluster segmented equivalent models based on the hybrid state estimation value of the key node to obtain a dynamic thermal integration equivalent model of the regional heating network to control the thermal transfer of the heating network in the region.
9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate via the bus, and when the machine-readable instructions are executed by the processor, the steps of a method for dynamic thermal integration equivalent modeling of a regional heating network as described in any one of claims 1 to 7 are performed.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which, when executed by a processor, executes the steps of a method for dynamic thermal integration equivalent modeling of a regional heating network as described in any one of claims 1 to 7.
Citation Information
Patent Citations
Equivalent inertia online evaluation method and system in power grid quasi-steady state scene
CN117498359A
Optimized scheduling method for hybrid energy central heating system based on air source heat pump
CN118066588A
Coordination and optimization method and system for comprehensive electric-thermal energy system, and device, medium and program
WO2023082697A1
