Automatic hydraulic balance adjusting system and method for central heating pipe network

Abstract

The invention relates to the technical field of hydraulic balance adjustment, and particularly discloses a system and a method for automatically adjusting hydraulic balance of a central heating pipe network, which are characterized in that a time queue of user room temperature acquired by a temperature sensor is acquired, and the difference between each time point and a target room temperature is calculated to obtain a time queue of room temperature deviation values; meanwhile, a pipe network pressure distribution topological graph is obtained, feature extraction is carried out through a data analysis and processing technology based on artificial intelligence, sequence coding is carried out on a time queue of room temperature deviation values, and therefore the valve opening value is automatically recommended according to key clue nonlinear significant interactive representation between the two; and a response adjusting instruction of the electric adjusting valve is generated by comparing the real-time value of the valve opening degree with the real-time value of the valve opening degree. By accurately calculating the room temperature deviation value and combining the pipe network pressure distribution characteristics, the valve opening degree of each area can be more accurately adjusted, it is ensured that heat distribution is more reasonable, and the situation of overheating or insufficient heating is reduced.

Description

Technical Field

[0001] The present application relates to the technical field of hydraulic balance regulation, and more specifically, to a system and method for automatic hydraulic balance regulation of a centralized heating network. Background Art

[0002] As an efficient heating method widely used in modern cities and large buildings, the centralized heating system relies on central heat sources such as boilers and thermal power plants to generate heat, and transports hot water or steam to various users through a complex pipe network. This system can not only achieve large-scale heat distribution, but also effectively reduce overall energy consumption and operating costs, with significant economic and environmental benefits.

[0003] However, the regulation method of traditional centralized heating system is relatively extensive and difficult to adjust in real time according to actual needs. Due to the complexity of the pipe network and the ever-changing user needs, hydraulic imbalance is very likely to occur in different areas. The uneven distribution of pipe network pressure causes overheating in some user areas and insufficient heating in other areas. This not only causes serious energy waste, but also greatly reduces user satisfaction. A more scientific and accurate regulation plan is urgently needed.

[0004] Therefore, an automatic adjustment scheme for the hydraulic balance of a central heating network is desired. Summary of the invention

[0005] The present application provides a system and method for automatic adjustment of the hydraulic balance of a centralized heating network. By accurately calculating the room temperature offset value and combining the pressure distribution characteristics of the network, the system can more accurately adjust the valve opening of each area, ensure more reasonable heat distribution, and reduce overheating or insufficient heating.

[0006] According to one aspect of the present application, a method for automatic adjustment of hydraulic balance of a centralized heating network is provided, comprising: obtaining a time queue of user room temperatures collected by a temperature sensor; calculating the difference between the user room temperature and the target room temperature at each time point in the time queue of the user room temperature to obtain a time queue of room temperature offset values; obtaining a network pressure distribution topology map; performing cross-modal interaction on the network pressure distribution topology map and the time queue of the room temperature offset values ​​to obtain a room temperature offset-network pressure cross-modal significant interaction feature, comprising: performing feature extraction on the network pressure distribution topology map and the time queue of the room temperature offset values ​​to obtain a network pressure distribution feature and a room temperature offset timing correlation feature; performing cross-modal interaction on the room temperature offset value; performing cross-modal interaction on the network pressure distribution topology map and the time queue of the room temperature offset values ​​to obtain a room temperature offset-network pressure cross-modal significant interaction feature; performing cross-modal interaction on the room temperature offset value; performing cross-modal interaction on the network pressure distribution topology map and the time queue of the room temperature offset value; performing cross-modal interaction on ... The time series correlation characteristics of the room temperature offset and the pipeline pressure distribution characteristics are subjected to nonlinear cross-domain response of the key clue of room temperature offset-pipeline pressure to obtain the significant cross-modal interaction characteristics of the room temperature offset-pipeline pressure; based on the significant cross-modal interaction characteristics of the room temperature offset-pipeline pressure, the response adjustment instructions of the electric control valve are generated; wherein, the pipeline pressure distribution topology diagram and the time queue of the room temperature offset value are respectively subjected to feature extraction to obtain the pipeline pressure distribution characteristics and the time series correlation characteristics of the room temperature offset, including: extracting the pipeline pressure distribution characteristics of the pipeline pressure distribution topology diagram to obtain the pipeline pressure distribution characteristics; and performing sequence encoding on the time queue of the room temperature offset value to obtain the time series correlation characteristics of the room temperature offset.

[0007] According to another aspect of the present application, a centralized heating network hydraulic balance automatic adjustment system is provided, comprising: a user room temperature acquisition module, used to acquire a time queue of user room temperatures collected by a temperature sensor; a room temperature offset value calculation module, used to calculate the difference between the user room temperature and the target room temperature at each time point in the user room temperature time queue to obtain a time queue of room temperature offset values; a network pressure distribution topology acquisition module, used to acquire a network pressure distribution topology; a cross-modal interaction processing module, used to perform cross-modal interaction on the network pressure distribution topology and the room temperature offset value time queue to obtain a room temperature offset-network pressure cross-modal significant interaction feature. Characteristic, wherein the cross-modal interaction processing module includes: a feature extraction unit, used to perform feature extraction on the pipeline pressure distribution topology diagram and the time queue of the room temperature offset value respectively to obtain the pipeline pressure distribution feature and the room temperature offset timing correlation feature; a cross-modal significant interaction feature extraction unit, used to perform room temperature offset-pipeline pressure key clue nonlinear cross-domain response on the room temperature offset timing correlation feature and the pipeline pressure distribution feature to obtain the room temperature offset-pipeline pressure cross-modal significant interaction feature; a response adjustment instruction generation module, used to generate a response adjustment instruction for the electric control valve based on the room temperature offset-pipeline pressure cross-modal significant interaction feature.

[0008] The present application provides a centralized heating network hydraulic balance automatic adjustment system and method, which obtains the time queue of user room temperature collected by a temperature sensor, and calculates the difference between each time point and the target room temperature to obtain the time queue of the room temperature offset value, and simultaneously obtains the network pressure distribution topology map, and extracts the features of the network pressure distribution topology map by using artificial intelligence-based data analysis and processing technology, and performs sequence encoding on the time queue of the room temperature offset value, so as to automatically recommend the valve opening value according to the key clue nonlinear significant interaction between the room temperature offset time series correlation feature and the network pressure distribution feature, and generate the response adjustment instruction of the electric control valve by comparing with the real-time value of the valve opening. The present application accurately calculates the room temperature offset value and combines the network pressure distribution feature, so that the system can more accurately adjust the valve opening of each area, ensure more reasonable heat distribution, and reduce overheating or insufficient heating. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings of the embodiments of the present application are briefly introduced below. Obviously, the drawings described below only relate to some embodiments of the present application, and are not intended to limit the present application.

[0010] Figure 1 This is a schematic flow chart of a method for automatic adjustment of the hydraulic balance of a centralized heating network according to an embodiment of the present application.

[0011] Figure 2 This is a data flow diagram of the method for automatic adjustment of the hydraulic balance of a centralized heating network according to an embodiment of the present application.

[0012] Figure 3 This is a schematic flow chart of step S4 in the method for automatic adjustment of hydraulic balance of a centralized heating network according to an embodiment of the present application.

[0013] Figure 4 It is a schematic flow chart of step S41 in the method for automatic adjustment of hydraulic balance of a centralized heating network according to an embodiment of the present application.

[0014] Figure 5 This is a schematic flowchart of step S42 in the method for automatic adjustment of hydraulic balance of a centralized heating network according to an embodiment of the present application.

[0015] Figure 6 This is a schematic flowchart of step S423 in the method for automatic adjustment of hydraulic balance of a centralized heating network according to an embodiment of the present application.

[0016] Figure 7 This is a schematic flowchart of step S424 in the method for automatic adjustment of hydraulic balance of a centralized heating network according to an embodiment of the present application.

[0017] Figure 8It is a schematic block diagram of the automatic adjustment system for hydraulic balance of a centralized heating network according to an embodiment of the present application. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work also fall within the scope of protection of the present application.

[0019] In response to the above technical problems, the technical concept of the present application is to obtain the time queue of the user's room temperature collected by the temperature sensor, and calculate the difference between each time point and the target room temperature to obtain the time queue of the room temperature offset value, and at the same time obtain the pipe network pressure distribution topology map, and extract the features of the pipe network pressure distribution topology map by using artificial intelligence-based data analysis and processing technology, and sequence encode the time queue of the room temperature offset value, so as to automatically recommend the valve opening value according to the key clue nonlinear significant interaction between the room temperature offset time series correlation characteristics and the pipe network pressure distribution characteristics, and generate the response adjustment instruction of the electric control valve by comparing with the real-time value of the valve opening. By accurately calculating the room temperature offset value and combining the pipe network pressure distribution characteristics, the system of the present application can more accurately adjust the valve opening of each area, ensure more reasonable heat distribution, and reduce overheating or insufficient heating.

[0020] Specifically, in the technical solution of the present application, if Figure 1 and Figure 2 As shown, the method for automatic adjustment of hydraulic balance of a centralized heating network includes: S1, obtaining a time queue of user room temperatures collected by a temperature sensor; S2, calculating the difference between the user room temperature and the target room temperature at each time point in the time queue of the user room temperature to obtain a time queue of room temperature offset values; S3, obtaining a network pressure distribution topology map; S4, performing cross-modal interaction on the network pressure distribution topology map and the time queue of room temperature offset values ​​to obtain a room temperature offset-network pressure cross-modal significant interaction feature; S5, generating a response adjustment instruction for the electric control valve based on the room temperature offset-network pressure cross-modal significant interaction feature.

[0021] Specifically, in step S1, the time queue of the user's room temperature collected by the temperature sensor is obtained. It should be understood that the purpose of obtaining the time queue of the user's room temperature collected by the temperature sensor is to accurately understand the actual indoor temperature changes of each user and compare and analyze them with the target room temperature, so as to provide data support for adjusting the valve opening. In this way, it can ensure that the heat distribution of each area is more reasonable, reduce overheating or insufficient heating, and improve the overall heating efficiency and user satisfaction.

[0022] In a specific embodiment, temperature sensors are installed in the user's room to monitor the indoor temperature changes in real time. These sensors record the current room temperature value at a certain time interval (such as every minute, every hour, etc.) and transmit this data to the central control system. In this way, a time series data set containing room temperature information at multiple time points is formed, that is, the time queue of the user's room temperature. This time queue not only reflects the actual room temperature at each time point, but also implies the trend and pattern of temperature changes over time, which is crucial to understanding the operating status of the heating system.

[0023] Specifically, in step S2, the difference between the user room temperature and the target room temperature at each time point in the time queue of the user room temperature is calculated to obtain a time queue of the room temperature offset value. It should be understood that the time queue of the user room temperature only reflects the changes in the actual room temperature, but cannot directly reflect the gap with the ideal heating state (i.e., the target room temperature). Based on this, in order to be able to more intuitively understand the operating effect of the heating system in the heating system network, the present application quantifies the heating effect by calculating the difference between the user room temperature and the target room temperature at each time point in the time queue of the user room temperature, and expresses the degree of deviation between the actual heating and the expected heating in a specific numerical form to obtain a time queue of the room temperature offset value. That is, the room temperature offset value, as the key data for measuring the gap between the actual room temperature and the target room temperature, provides a direct basis for subsequent adjustment of the heating system.

[0024] Specifically, in step S3, a network pressure distribution topology diagram is obtained. It should be understood that by understanding the pressure distribution in the network, the pressure conditions of each node and their mutual influence can be more accurately determined, thereby providing a basis for accurately adjusting the valve opening, ensuring that heat can be reasonably distributed, and avoiding overheating or insufficient heating in certain areas. Specifically, the network pressure distribution topology diagram is a graphical representation of the heating network structure and its internal pressure distribution. It not only shows the positional relationship between each pipe and node, but also contains the pressure value information at each location. This detailed view helps to identify potential problem points in the network, such as local pressure anomalies and uneven pressure distribution. If these problems are not handled in a timely manner, they may lead to low heating efficiency or equipment damage.

[0025] In a specific embodiment, in order to obtain the pipe network pressure distribution topology, a series of pressure sensors need to be deployed at key nodes of the network. These sensors monitor the pressure changes at the location in real time and transmit the data to the central control system. In addition, professional software tools are needed to process and analyze the collected data to generate an intuitive pipe network pressure distribution map. This process may involve complex algorithms and techniques, such as two-dimensional convolution operations, which are used to extract meaningful pressure features from the original data, and then build a comprehensive and detailed pipe network pressure distribution model. The pipe network pressure distribution topology obtained in this way can help operators and automation systems better understand the current operating status of the heating system, discover potential problems and take corresponding measures to solve them. For example, when it is found that the pressure in a certain area is too high or too low, the normal pressure level can be restored by adjusting the valve opening of the relevant area, thereby ensuring the stable operation and efficient operation of the entire heating network. At the same time, combined with the time queue of the user's room temperature, the adjustment strategy can be further optimized to achieve more accurate heat distribution and control. This can not only improve the user's comfort experience, but also effectively save energy and reduce operating costs.

[0026] Specifically, in step S4, cross-modal interaction is performed on the pipe network pressure distribution topology diagram and the time queue of the room temperature offset value to obtain a significant cross-modal interaction feature of room temperature offset-pipe network pressure.

[0027] In one embodiment, Figure 3 As shown, cross-modal interaction is performed on the pipeline network pressure distribution topology diagram and the time queue of the room temperature offset value to obtain the room temperature offset-pipeline network pressure cross-modal significant interaction characteristics, including: S41, feature extraction is performed on the pipeline network pressure distribution topology diagram and the time queue of the room temperature offset value respectively to obtain the pipeline network pressure distribution characteristics and the room temperature offset time series correlation characteristics; S42, room temperature offset-pipeline network pressure key clue nonlinear cross-domain response is performed on the room temperature offset time series correlation characteristics and the pipeline pressure distribution characteristics to obtain the room temperature offset-pipeline network pressure cross-modal significant interaction characteristics.

[0028] In one embodiment, Figure 4 As shown, feature extraction is performed on the pipeline network pressure distribution topology diagram and the time queue of the room temperature offset value respectively to obtain the pipeline network pressure distribution feature and the room temperature offset timing correlation feature, including: S411, extracting the pipeline network pressure distribution feature of the pipeline network pressure distribution topology diagram to obtain the pipeline network pressure distribution feature; S412, serially encoding the time queue of the room temperature offset value to obtain the room temperature offset timing correlation feature.

[0029] In one embodiment, in step S411, the pipeline network pressure distribution topology map is subjected to pipeline network pressure distribution feature extraction to obtain the pipeline network pressure distribution feature, including: extracting the pipeline network pressure distribution feature based on two-dimensional convolution on the pipeline network pressure distribution topology map to obtain the pipeline network pressure distribution feature map as the pipeline network pressure distribution feature. It should be understood that considering that the pipeline network pressure distribution topology map is essentially a data representation of a two-dimensional structure, it contains the pressure information of each node in the pipeline network and the connection relationship between the nodes. The two-dimensional convolution operation is very suitable for processing such data with a two-dimensional structure. It can effectively identify these complex patterns without changing the two-dimensional structure of the data. Based on this, in the technical solution of the present application, the pipeline network pressure distribution topology map is subjected to pipeline network pressure distribution feature extraction based on two-dimensional convolution to slide the convolution kernel on the image, and feature extraction is performed on the local area, so as to effectively capture the local pressure features and structural features in the pipeline network pressure distribution topology map, and obtain the pipeline network pressure distribution feature map. In this way, through the two-dimensional convolution operation, key pressure distribution features, such as the local change trend of pressure, the area with large pressure gradient, the area with relatively stable pressure, etc., can be extracted from the pipeline network pressure distribution topology map. These characteristics are very important for in-depth understanding of the hydraulic conditions of the pipeline network. They can help discover possible problems in the pipeline network, such as abnormal local pressure, uneven pressure distribution, etc., thereby providing a basis for taking corresponding regulatory measures.

[0030] In one embodiment, in step S412, the time queue of the room temperature offset value is sequence-encoded to obtain the room temperature offset timing association feature, including: using a sequence encoder based on a one-dimensional causal convolution model to encode the time queue of the room temperature offset value to obtain an implicit encoding vector of the room temperature offset timing association feature as the room temperature offset timing association feature. It should be understood that considering that the time queue of the room temperature offset value is a typical time series data, it contains the deviation information of the room temperature relative to the target room temperature at different time points. Time series data has the characteristic of front-to-back dependence, that is, the data at the current time point may be affected by the data at the past time point. In order to effectively process this temporal sequence and association information, the present application obtains an implicit encoding vector of the room temperature offset timing association feature by sequence-encoding the time queue of the room temperature offset value. In particular, in a specific example of the present application, a sequence encoder based on a one-dimensional causal convolution model is used to encode the time queue of the room temperature offset value to obtain an implicit encoding vector of the room temperature offset timing association feature. In particular, the time queue of room temperature offset values ​​is a one-dimensional data sequence arranged in chronological order, with obvious precedence and causal relationship, that is, the room temperature offset at the current time point depends on the past time point, but not on the future time point. The one-dimensional causal convolution model can just meet the processing requirements of this causal relationship. It only considers the data of the past and current time points during convolution calculation, and does not use the data of the future time points. This is consistent with the actual situation of the time queue of room temperature offset values ​​and can accurately capture the causal information in the time series. In addition, the one-dimensional causal convolution model can slide on the time series through convolution kernels of different sizes to extract local features of the time queue of room temperature offset values. It can effectively capture the short-term change characteristics in the time series, such as sudden fluctuations in room temperature offset values, small-scale trend changes, etc., so as to more carefully characterize the characteristics of the time series of room temperature offset values.

[0031] Specifically, in step S42, the room temperature offset time series correlation feature and the pipe network pressure distribution feature are subjected to nonlinear cross-domain response of the key clue of room temperature offset-pipe network pressure to obtain the significant cross-modal interaction feature of the room temperature offset-pipe network pressure. It should be understood that considering that the room temperature offset time series correlation feature comes from the analysis of the time queue of the room temperature offset value, it reflects the change of room temperature over time; while the pipe network pressure distribution feature is obtained after processing the pipe network pressure topology map, reflecting the pressure state of the pipe network. These two types of data belong to different fields, and there are complex nonlinear relationships and hidden associations between them. That is, changes in pipe network pressure will affect the flow and distribution of hot water, and then affect the room temperature of the user; and changes in room temperature may also feedback and affect the demand for pipe network pressure regulation. However, traditional linear analysis methods are difficult to effectively handle such complex relationships between cross-domain data. Based on this, the present application obtains the significant cross-modal interaction feature of the room temperature offset-pipe network pressure by performing nonlinear cross-domain response of the key clue of room temperature offset-pipe network pressure on the room temperature offset time series correlation feature and the pipe network pressure distribution feature. In this way, the information of both room temperature offset and pipe network pressure can be integrated to highlight the significant interaction between the two. For example, it may be found that there is a close connection between a specific pipe network pressure distribution pattern and a certain trend of room temperature deviation. These interactive characteristics are very important for a deep understanding of the operating laws of the heating system and provide stronger support for the precise regulation of the heating system.

[0032] In one embodiment, Figure 5 As shown, the room temperature offset time series association feature and the pipe network pressure distribution feature are subjected to nonlinear cross-domain response of the key clue of room temperature offset-pipe network pressure to obtain the cross-modal significant interaction feature of the room temperature offset-pipe network pressure, including: S421, core clue point convolution encoding of the implicit coding vector of the room temperature offset time series association feature to obtain the room temperature offset time series core clue coding vector; S422, extracting the pipe network pressure distribution core clue coding vector from the pipe network pressure distribution feature map; S423, constructing a room temperature offset-pipe network pressure core clue weaving template matrix between the room temperature offset time series core clue coding vector and the pipe network pressure distribution core clue coding vector; S424, using the room temperature offset-pipe network pressure core clue weaving template matrix as a priori information constraint matrix, cross-modal template constraints are performed on the implicit coding vector of the room temperature offset time series association feature and the pipe network pressure distribution feature map to obtain the room temperature offset-pipe network pressure cross-modal significant interaction coding vector as the room temperature offset-pipe network pressure cross-modal significant interaction feature.

[0033] Specifically, in step S421, by performing core clue point convolution coding on this implicit coding vector, a core clue coding vector of the room temperature offset timing can be obtained. This process enables the system to quickly locate the key time period that affects the user's comfort, so as to make targeted adjustments. For example, when a sudden drop in temperature at night is detected, the system can increase the heating intensity in advance to prevent the user from feeling cold. At the same time, point convolution coding strengthens the semantically significant modal specific features to ensure that the system can respond accurately even under complex and changing environmental conditions. Further, this coding method helps the system better understand the deep-seated reasons behind temperature changes. For example, in some cases, a sudden change in external temperature may cause the room temperature in multiple areas to fluctuate simultaneously. Through point convolution coding, the system can identify which areas are most likely to be affected and pre-adjust in advance to ensure that all users can enjoy a comfortable indoor temperature. In addition, point convolution coding can effectively filter out redundant information and focus on those truly meaningful change points. For example, ignore those time periods with small temperature fluctuations and focus on those time periods with significant changes. This greatly improves the system's computing efficiency and response speed, allowing the system to process only the most critical temperature change data, thereby generating effective adjustment instructions more quickly and reducing unnecessary waste of resources. Ultimately, through this series of operations, the system is able to generate a highly concentrated and representative room temperature offset timing core clue encoding vector. This process not only helps the system quickly identify the key factors that affect user comfort, but also improves the accuracy and effectiveness of the overall adjustment strategy. Specifically, the process is expressed as: ;in, is the implicit encoding vector of the room temperature offset timing correlation feature, is the point convolutional coding, is the bias vector, is the activation function, is the room temperature offset timing core cue encoding vector.

[0034] Specifically, in step S422, in the automatic adjustment system of the hydraulic balance of the centralized heating network, it is a crucial step to extract the core clue encoding vector of the network pressure distribution from the network pressure distribution feature map. In this process, the system first obtains the network pressure distribution topology map, which shows the pressure conditions of each node and pipeline in detail. Since the network pressure distribution map usually carries stronger spatial dimension information, the processing method is different from the time series data of the room temperature offset value. Specifically, the system uses the regional attention mechanism to ensure the co-retention of spatial information and semantic features. In this way, the pressure data of the significant area is reweighted, giving priority to those areas that have a greater impact on the overall system. Next, the system further enhances the expression ability of structured information by performing a convolution operation on the network pressure distribution feature map. Based on the local receptive field of the convolution operation, the system can enrich the structured information expression ability of the modality and capture the situation of local pressure fluctuations. The process of extracting the core clue encoding vector from the network pressure distribution feature map helps the system better understand the pressure distribution pattern and its dynamic changes inside the network. For example, in some cases, changes in external temperature may cause the network pressure in multiple areas to fluctuate simultaneously. Through the convolution operation, the system can identify which areas are most likely to be affected and make pre-adjustments in advance to ensure that all users can enjoy a comfortable indoor temperature. Furthermore, this encoding method can also effectively filter out redundant information and focus on those truly meaningful change points. Specifically, the process is expressed as: ;in, is the characteristic diagram of the pipe network pressure distribution, is the meshing operation, and They are the first, second, and third in the set of pipeline network pressure distribution characteristic diagrams. and The local characteristic diagram of the pressure distribution of the pipe network, is a convolutional code with a convolution kernel of 3×3. yes function, It is the first in the set of local enhanced characteristic maps of pipeline network pressure distribution. The local enhanced characteristic map of the pressure distribution in the pipeline network, is a global pooling operation, It is the first in the set of local enhanced characteristic vectors of the pressure distribution in the pipeline network. The local enhanced eigenvector of the network pressure distribution, and Take the maximum and minimum values ​​respectively. and is the modulation parameter, and They are The mean and variance of is the set of local significant coefficients of the pipe network pressure distribution. The local significant coefficient of the pressure distribution in the pipe network is Indicated by natural constant The exponential function value with base , is the number of vectors in the set of local enhanced characteristic vectors of the pipe network pressure distribution, It is the core clue encoding vector of the pipeline network pressure distribution.

[0035] In one embodiment, Figure 6 As shown, in step S423, a room temperature offset-pipe network pressure core clue weaving template matrix is ​​constructed between the room temperature offset timing core clue coding vector and the pipe network pressure distribution core clue coding vector, including: S4231, using a feature mapping function to associate the room temperature offset timing core clue coding vector and the pipe network pressure distribution core clue coding vector to obtain a room temperature offset-pipe network pressure core association matrix; S4232, performing different-driven semantic alignment on the room temperature offset-pipe network pressure core association matrix to obtain the room temperature offset-pipe network pressure core clue weaving template matrix.

[0036] Specifically, the process of step S4231 is expressed by the formula: ;in, is the transpose operation, and is a feature mapping function, such as a linear mapping or a nonlinear kernel function, is the core correlation matrix of room temperature offset and network pressure.

[0037] In one embodiment, in step S4232, the room temperature offset-pipeline network pressure core association matrix is ​​subjected to different-driven semantic alignment to obtain the room temperature offset-pipeline network pressure core clue weaving template matrix, including: using a semantic alignment module based on a residual mechanism and feature mapping to perform semantic spatial alignment on the room temperature offset timing core clue encoding vector and the pipe network pressure distribution core clue encoding vector to obtain a room temperature offset-pipeline network pressure alignment vector and a pipe network pressure-room temperature offset alignment vector; specifically, the process is expressed by a formula: ;in, It is subtracted by position point. and They are the room temperature offset-pipeline network pressure alignment vector and the pipeline network pressure-room temperature offset alignment vector.

[0038] The room temperature offset-pipeline network pressure explicit modeling correlation matrix between the room temperature offset-pipeline network pressure alignment vector and the pipeline network pressure-room temperature offset alignment vector is calculated; specifically, the process is expressed by the formula: ;in, represents a natural constant, To calculate the square of the vector's two-norm, is the logarithmic function value with base 2, is matrix multiplication, is the room temperature offset-network pressure explicit modeling correlation matrix.

[0039] Based on the room temperature offset-pipeline network pressure explicit modeling correlation matrix, the room temperature offset-pipeline network pressure core correlation matrix is ​​compensated for modal explicit information to obtain the room temperature offset-pipeline network pressure core clue weaving template matrix. Specifically, the process is expressed by the formula: ;in, It is the room temperature offset-tube network pressure core clue weave template matrix.

[0040] It should be understood that the process of constructing the room temperature offset-pipeline network pressure core clue weaving template matrix between the room temperature offset time series core clue encoding vector and the pipe network pressure distribution core clue encoding vector provides a solid foundation for the subsequent precise adjustment of the electric control valve by deeply exploring the intrinsic connection between the two. First, the room temperature offset time series core clue encoding vector and the pipe network pressure distribution core clue encoding vector are extracted from a large amount of data through a series of complex algorithms. The former reflects the trend of the user's indoor temperature over time and its deviation from the target temperature, while the latter reveals the pressure distribution characteristics in the entire heating network. Next, the two encoding vectors are associated to construct the room temperature offset-pipeline network pressure core clue weaving template matrix. In this process, not only the changes in individual variables are paid attention to, but also the intricate interactive relationship between them is paid more attention. For example, the room temperature of a building is generally lower than the set value, and the corresponding pipe network pressure also shows abnormal fluctuations. By weaving the template matrix, the close connection between room temperature and pressure can be captured, and the possible causal relationship between the two can be further analyzed. In actual operation, the template matrix plays the role of a bridge, enabling simultaneous consideration of the two dimensions of room temperature and pressure. In addition, the process of constructing the woven template matrix also involves semantic alignment technology to ensure that data from different modalities can be effectively compared and analyzed within the same framework. This means that whether it is a time series describing room temperature changes or a spatial distribution map reflecting pipe network pressure, it can be quantified and compared under the same standard. This greatly improves the level of intelligence, enabling it to make accurate judgments in complex and changing environments. In this way, it is possible not only to identify which areas are most in need of priority treatment, but also to dynamically adjust strategies based on current conditions.

[0041] Preferably, for the room temperature offset temporal core cue encoding vector for different modality feature core cue representation and the core clue encoding vector of the pipe network pressure distribution , through its residual mechanism based on different modal mapping, a preliminary structured semantic alignment of universal explicit associations is performed in the shared representation space to obtain the room temperature offset-pipeline network pressure alignment vector Align vector with network pressure-room temperature offset Then, based on the global semantic spatialization representation of the graph structure under explicit modeling of mapping preference, the residual mechanism is further used to capture the "missing" information in the alignment process under different mapping modalities, and adaptive compensation is performed through explicit association to further capture the fine-grained significant difference information between modalities in the initial alignment process. In this way, the contradiction between the alignment complexity of the intrinsic driving mapping of cross-modal differences and the intuitiveness of the explicit association of the graph structure can be optimized, and the illusion of global semantic constraints between modalities under the condition of misalignment of fine-grained relationships between local salient regions can be reduced.

[0042] In one embodiment, Figure 7 As shown, in step S424, the room temperature offset-pipeline network pressure core clue weaving template matrix is ​​used as the prior information constraint matrix, and the room temperature offset time series correlation feature implicit coding vector and the pipe network pressure distribution feature map are subjected to cross-modal template constraints to obtain the room temperature offset-pipeline network pressure cross-modal significant interaction coding vector, including: S4241, the pipe network pressure distribution feature map is subjected to feature decoupling along the channel dimension to obtain a set of local feature matrices of the pipe network pressure distribution; specifically, the process is expressed by the formula: ;in, is the feature decoupling operation along the channel dimension, and is the first, second, and third in the set of local characteristic matrices of the network pressure distribution. and The local characteristic matrix of the pressure distribution in the pipe network.

[0043] S4242, using the room temperature offset time series associated feature implicit coding vector as a query vector, using each pipe network pressure distribution local feature matrix in the set of the pipe network pressure distribution local feature matrix as a key matrix, and using the room temperature offset-pipe network pressure core clue weaving template matrix as a priori information constraint matrix, heterogeneous transformation cross-modal constraint coding is performed on them to obtain a set of room temperature offset-pipe network pressure cross-modal fine-grained interactive coding vectors; specifically, the process is expressed by the formula: ;in, yes The scale is the matrix width multiplied by the matrix height, is the normalization function, is the first in the set of cross-modal fine-grained interaction encoding vectors of room temperature offset and pipe network pressure. A fine-grained cross-modal interaction encoding vector of room temperature offset and pipe network pressure.

[0044] S4243, calculating the position mean vector of the set of the room temperature offset-pipeline network pressure cross-modal fine-grained interaction coding vectors to obtain the room temperature offset-pipeline network pressure cross-modal significant interaction coding vector. Specifically, the process is expressed by the formula: ;in, is the number of vectors in the set of cross-modal fine-grained interaction encoding vectors of room temperature offset and pipe network pressure, is the room temperature offset-pipeline network pressure cross-modal significant interaction encoding vector.

[0045] It should be understood that first, the pipeline pressure distribution feature map is obtained, and the features along the channel dimension are decoupled to obtain a set of local feature matrices. These local feature matrices represent the pressure distribution in different areas of the pipeline network, so that local pressure fluctuations in complex pipeline networks can be captured more finely. Next, the implicit encoding vector of the room temperature offset time series correlation feature is used as the query vector, each local feature matrix in the set of local feature matrices of the pipeline pressure distribution is used as the key matrix, and the room temperature offset-pipeline pressure core clue weaving template matrix is ​​used as the prior information constraint matrix to perform heterogeneous conversion cross-modal constraint encoding. This process obtains responses from the local feature matrix through the attention mechanism to ensure that each local feature matrix can be effectively matched with the room temperature time series data. Specifically, when a sudden drop in room temperature in a certain area is detected, it uses the implicit encoding vector of the room temperature offset time series correlation feature as the query vector to find the local feature matrix related to this from the set of local feature matrices of the pipeline pressure distribution. Through the attention mechanism, it is possible to identify which local feature matrices are most critical to the current problem and give them higher weights. At the same time, the room temperature offset-pipeline network pressure core clue weaving template matrix is ​​used as a prior information constraint matrix to ensure that the interaction between these local feature matrices conforms to the overall dynamic characteristics. In this way, not only can the local pressure fluctuations be accurately captured, but also the intrinsic connection between these fluctuations and room temperature changes can be understood. Subsequently, a set of room temperature offset-pipeline network pressure cross-modal fine-grained interaction encoding vectors is generated. Each fine-grained interaction encoding vector reflects the combined representation of a local feature matrix and the room temperature time series data. Finally, the positional mean vector of these fine-grained interaction encoding vector sets is calculated to obtain the final room temperature offset-pipeline network pressure cross-modal significant interaction encoding vector. This process simplifies complex interaction results by aggregating information from multiple local areas while ensuring that key information is not lost.

[0046] Specifically, in step S5, based on the significant cross-modal interaction characteristics of the room temperature offset and the pipe network pressure, a response adjustment instruction of the electric control valve is generated. In one embodiment, based on the significant cross-modal interaction characteristics of the room temperature offset and the pipe network pressure, a response adjustment instruction of the electric control valve is generated, including: inputting the room temperature offset and the pipe network pressure cross-modal significant interaction coding vector into a decoder-based valve opening strategy generation engine to obtain a valve opening recommended value; based on the valve opening recommended value and the valve opening real-time value, generating the response adjustment instruction of the electric control valve.

[0047] That is, the room temperature offset-pipeline network pressure cross-modal significant interaction coding features obtained by nonlinear interaction of key clues using the implicit coding vector of the room temperature offset time series correlation feature and the pipe network pressure distribution characteristic diagram are decoded to automatically recommend the valve opening value. It should be understood that in a centralized heating system, the adjustment of the valve opening directly affects the hydraulic conditions of the pipe network, and then affects the pipe network pressure distribution and the final user room temperature. Therefore, using the coding vector containing these two aspects of key information as input can provide a comprehensive and accurate basis for the adjustment of the valve opening based on the inherent connection between various factors in the heating system. The decoder can understand and parse the complex interactive relationship between the room temperature and the pipe network pressure contained in the coding vector, and generate appropriate valve opening recommendation values ​​based on these relationships and pre-set rules and models.

[0048] Then, based on the recommended valve opening value and the real-time valve opening value, a response adjustment instruction of the electric control valve is generated. That is, the recommended valve opening value is an ideal opening value obtained based on a comprehensive analysis of multiple factors such as room temperature offset and pipe network pressure, reflecting the valve state required to achieve the best heating effect. The real-time valve opening value reflects the actual working state of the valve. By comparing the two, the gap between the current state of the valve and the ideal state can be clearly judged, thereby clarifying the direction of adjustment (whether to increase or decrease the opening) and the degree of adjustment, providing a data basis for generating accurate adjustment instructions.

[0049] In a specific embodiment, taking a specific area as an example, by analyzing the time series data of the room temperature of users in the area and the characteristics of the pressure distribution of the pipe network and decoding, it is concluded that the ideal recommended value of the valve opening in the current situation is 60%, while the actual opening of the valve (i.e., the real-time value) is 50%. In this context, based on the difference between the recommended value of the valve opening and the real-time value of the valve opening, a corresponding response adjustment instruction can be generated. The key here is to compare the two values ​​and determine how to adjust the state of the valve to reach or approach the recommended value. Since the goal is to adjust the valve opening from 50% to 60%, it is necessary to send an instruction to increase the valve opening to the electric control valve. In this case, the response adjustment instruction of the electric control valve may be a string of specifically encoded data information, which tells the valve controller the specific operation to be performed. For example, under the digital communication protocol, this may be a command code containing "increase the opening by 10 percentage points". After this command is transmitted to the electric control valve, the valve will open at a predetermined speed and manner until its opening reaches the recommended 60%.

[0050] In summary, according to the embodiment of the present application, the automatic adjustment system and method for the hydraulic balance of the centralized heating network are explained, which obtains the time queue of the user room temperature collected by the temperature sensor, and calculates the difference between each time point and the target room temperature to obtain the time queue of the room temperature offset value, and simultaneously obtains the network pressure distribution topology map, and extracts the features of the network pressure distribution topology map by using data analysis and processing technology based on artificial intelligence, and performs sequence encoding on the time queue of the room temperature offset value, so as to automatically recommend the valve opening value according to the nonlinear significant interaction representation of the key clues between the time series correlation characteristics of the room temperature offset and the network pressure distribution characteristics, and generate the response adjustment instruction of the electric control valve by comparing with the real-time value of the valve opening. By accurately calculating the room temperature offset value and combining the network pressure distribution characteristics, the system of the present application can more accurately adjust the valve opening of each area, ensure more reasonable heat distribution, and reduce overheating or insufficient heating.

[0051] Figure 8 Schematic diagram of the automatic hydraulic balance adjustment system of the central heating network of the present application. Figure 8As shown, the centralized heating network hydraulic balance automatic regulation system includes: a user room temperature acquisition module 10, which is used to obtain a time queue of user room temperatures collected by a temperature sensor; a room temperature offset value calculation module 20, which is used to calculate the difference between the user room temperature and the target room temperature at each time point in the user room temperature time queue to obtain a time queue of room temperature offset values; a network pressure distribution topology map acquisition module 30, which is used to obtain a network pressure distribution topology map; a cross-modal interaction processing module 40, which is used to perform cross-modal interaction on the network pressure distribution topology map and the time queue of the room temperature offset value to obtain a room temperature offset-network pressure cross-modal significant interaction feature; a response adjustment instruction generation module 50, which is used to generate a response adjustment instruction for the electric control valve based on the room temperature offset-network pressure cross-modal significant interaction feature.

[0052] In one embodiment, the cross-modal interaction processing module includes: a feature extraction unit, which is used to perform feature extraction on the pipeline pressure distribution topology map and the time queue of the room temperature offset value respectively to obtain the pipeline pressure distribution characteristics and the room temperature offset timing correlation characteristics; a cross-modal significant interaction feature extraction unit, which is used to perform room temperature offset-pipeline pressure key clue nonlinear cross-domain response on the room temperature offset timing correlation characteristics and the pipeline pressure distribution characteristics to obtain the room temperature offset-pipeline pressure cross-modal significant interaction characteristics.

[0053] Here, those skilled in the art can understand that the specific operations of each module in the above-mentioned centralized heating network hydraulic balance automatic adjustment system have been referred to above. Figures 1 to 7 The method for automatic adjustment of the hydraulic balance of a centralized heating network has been described in detail, and therefore, its repeated description will be omitted.

[0054] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0055] It should be understood that the specific examples in this article are only intended to help those skilled in the art better understand the embodiments of the present application, rather than to limit the scope of the embodiments of the present application.

[0056] It should also be understood that in the various embodiments of the present application, the size of the serial number of each process does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0057] It should also be understood that the various implementation methods described in this specification can be implemented individually or in combination, and the embodiments of the present application are not limited to this.

[0058] Unless otherwise stated, all technical and scientific terms used in the embodiments of the present application are the same as the meanings generally understood by those skilled in the art of the technical field of the present application. The terms used in the present application are only for the purpose of describing specific embodiments and are not intended to limit the scope of the present application. The term "and / or" used in the present application includes any and all combinations of one or more related listed items. The singular forms of "a kind of", "above" and "the" used in the embodiments of the present application and the appended claims are also intended to include majority forms, unless the context clearly indicates other meanings. In addition, the terms "first", "second" etc. are only used for description purposes and cannot be understood as indicating or suggesting relative importance.

[0059] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units 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 interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0060] The units described as separate components may or may not be physically separated, and the components shown as units 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.

[0061] 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.

[0062] 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 computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0063] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present 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 automatically adjusting the hydraulic balance of a central heating network, characterized in that: The method comprises: obtaining a time queue of user room temperature collected by a temperature sensor; calculating the difference between the user room temperature and the target room temperature at each time point in the time queue of the user room temperature to obtain a time queue of room temperature offset values; obtaining a pipe network pressure distribution topology map; performing cross-modal interaction on the pipe network pressure distribution topology map and the time queue of the room temperature offset values ​​to obtain a room temperature offset-pipe network pressure cross-modal significant interaction feature, comprising: performing feature extraction on the pipe network pressure distribution topology map and the time queue of the room temperature offset values ​​to obtain a pipe network pressure distribution feature and a room temperature offset time series correlation feature; performing cross-modal interaction on the room temperature offset time series correlation feature and the pipe network pressure distribution topology map; performing cross-modal interaction on the pipe network pressure distribution topology map and the time queue of the room temperature offset values ​​to obtain a pipe network pressure distribution feature and a room temperature offset time series correlation feature; performing cross-modal interaction on the pipe network pressure distribution topology map and the time queue of the room temperature offset values; performing cross-modal interaction on the room temperature offset time series correlation feature and the pipe network pressure distribution topology map ... The characteristic is to perform nonlinear cross-domain response of key clues of room temperature offset-pipeline pressure to obtain the significant cross-modal interaction characteristics of the room temperature offset-pipeline pressure; based on the significant cross-modal interaction characteristics of the room temperature offset-pipeline pressure, generate response adjustment instructions for the electric control valve; wherein, feature extraction is performed on the pipeline network pressure distribution topology diagram and the time queue of the room temperature offset value respectively to obtain the pipeline network pressure distribution characteristics and the room temperature offset timing correlation characteristics, including: extracting the pipeline network pressure distribution characteristics of the pipeline network pressure distribution topology diagram to obtain the pipeline network pressure distribution characteristics; and performing sequence encoding on the time queue of the room temperature offset value to obtain the room temperature offset timing correlation characteristics.

2. The method for automatic adjustment of hydraulic balance of a centralized heating network according to claim 1, characterized in that: Extracting the pipeline network pressure distribution characteristics from the pipeline network pressure distribution topology map to obtain the pipeline network pressure distribution characteristics, including: extracting the pipeline network pressure distribution characteristics based on two-dimensional convolution from the pipeline network pressure distribution topology map to obtain a pipeline network pressure distribution characteristic map as the pipeline network pressure distribution characteristics.

3. The method for automatic adjustment of hydraulic balance of a central heating network according to claim 2, characterized in that: The time queue of the room temperature offset value is sequentially encoded to obtain the room temperature offset timing association feature, including: using a sequence encoder based on a one-dimensional causal convolution model to encode the time queue of the room temperature offset value to obtain an implicit encoding vector of the room temperature offset timing association feature as the room temperature offset timing association feature.

4. The method for automatic adjustment of hydraulic balance of a centralized heating network according to claim 3, characterized in that: The room temperature offset time series association feature and the pipe network pressure distribution feature are subjected to nonlinear cross-domain response of the key clue of room temperature offset-pipe network pressure to obtain the cross-modal significant interaction feature of the room temperature offset-pipe network pressure, including: core clue point convolution encoding of the implicit coding vector of the room temperature offset time series association feature to obtain the room temperature offset time series core clue coding vector; extracting the pipe network pressure distribution core clue coding vector from the pipe network pressure distribution feature map; constructing a room temperature offset-pipe network pressure core clue weaving template matrix between the room temperature offset time series core clue coding vector and the pipe network pressure distribution core clue coding vector; using the room temperature offset-pipe network pressure core clue weaving template matrix as a priori information constraint matrix, cross-modal template constraints are performed on the implicit coding vector of the room temperature offset time series association feature and the pipe network pressure distribution feature map to obtain the room temperature offset-pipe network pressure cross-modal significant interaction coding vector as the room temperature offset-pipe network pressure cross-modal significant interaction feature.

5. The method for automatic adjustment of hydraulic balance of a central heating network according to claim 4, characterized in that: Constructing a room temperature offset-pipe network pressure core clue weaving template matrix between the room temperature offset timing core clue coding vector and the pipe network pressure distribution core clue coding vector, including: using a feature mapping function to associate the room temperature offset timing core clue coding vector with the pipe network pressure distribution core clue coding vector to obtain a room temperature offset-pipe network pressure core association matrix; performing different-driven semantic alignment on the room temperature offset-pipe network pressure core association matrix to obtain the room temperature offset-pipe network pressure core clue weaving template matrix.

6. The method for automatic adjustment of hydraulic balance of a central heating network according to claim 5, characterized in that: The room temperature offset-pipeline network pressure core association matrix is ​​subjected to different-driven semantic alignment to obtain the room temperature offset-pipeline network pressure core clue weaving template matrix, including: using a semantic alignment module based on residual mechanism and feature mapping to perform semantic spatial alignment on the room temperature offset timing core clue encoding vector and the pipe network pressure distribution core clue encoding vector to obtain a room temperature offset-pipeline network pressure alignment vector and a pipe network pressure-room temperature offset alignment vector; calculating a room temperature offset-pipeline network pressure explicit modeling correlation matrix between the room temperature offset-pipeline network pressure alignment vector and the pipe network pressure-room temperature offset alignment vector; based on the room temperature offset-pipeline network pressure explicit modeling correlation matrix, the room temperature offset-pipeline network pressure core association matrix is ​​subjected to inter-modal explicit information compensation to obtain the room temperature offset-pipeline network pressure core clue weaving template matrix.

7. The method for automatic adjustment of hydraulic balance of a central heating network according to claim 6, characterized in that: The room temperature offset-pipeline network pressure core clue weaving template matrix is ​​used as a prior information constraint matrix, and the room temperature offset time series associated feature implicit coding vector and the pipe network pressure distribution feature map are cross-modal template constrained to obtain the room temperature offset-pipe network pressure cross-modal significant interaction coding vector, including: performing feature decoupling along the channel dimension of the pipe network pressure distribution feature map to obtain a set of pipe network pressure distribution local feature matrices; using the room temperature offset time series associated feature implicit coding vector as a query vector, each pipe network pressure distribution local feature matrix in the set of pipe network pressure distribution local feature matrices as a key matrix and the room temperature offset-pipe network pressure core clue weaving template matrix as a prior information constraint matrix, performing heterogeneous transformation cross-modal constraint coding on them to obtain a set of room temperature offset-pipe network pressure cross-modal fine-grained interaction coding vectors; calculating the positional mean vector of the set of room temperature offset-pipe network pressure cross-modal fine-grained interaction coding vectors to obtain the room temperature offset-pipe network pressure cross-modal significant interaction coding vector.

8. The method for automatic adjustment of hydraulic balance of a central heating network according to claim 7, characterized in that: Based on the room temperature offset-pipeline network pressure cross-modal significant interaction characteristics, a response adjustment instruction of the electric control valve is generated, including: inputting the room temperature offset-pipeline network pressure cross-modal significant interaction encoding vector into a decoder-based valve opening strategy generation engine to obtain a valve opening recommended value; based on the valve opening recommended value and the valve opening real-time value, a response adjustment instruction of the electric control valve is generated.

9. A centralized heating network hydraulic balance automatic adjustment system, used to implement the centralized heating network hydraulic balance automatic adjustment method according to any one of claims 1 to 8, characterized in that: include: A user room temperature acquisition module is used to acquire a time queue of user room temperature collected by a temperature sensor; a room temperature offset value calculation module is used to calculate the difference between the user room temperature and the target room temperature at each time point in the user room temperature time queue to obtain a time queue of room temperature offset values; a pipe network pressure distribution topology acquisition module is used to acquire a pipe network pressure distribution topology; a cross-modal interaction processing module is used to perform cross-modal interaction on the pipe network pressure distribution topology and the time queue of the room temperature offset value to obtain a room temperature offset-pipe network pressure cross-modal significant interaction feature, wherein the cross-modal interaction processing module includes: a feature extraction unit is used to perform feature extraction on the pipe network pressure distribution topology and the time queue of the room temperature offset value respectively to obtain a pipe network pressure distribution feature and a room temperature offset time series correlation feature; a cross-modal significant interaction feature extraction unit is used to perform a nonlinear cross-domain response of the room temperature offset-pipe network pressure key clue on the room temperature offset time series correlation feature and the pipe network pressure distribution feature to obtain the room temperature offset-pipe network pressure cross-modal significant interaction feature; a response adjustment instruction generation module is used to generate a response adjustment instruction for the electric control valve based on the room temperature offset-pipe network pressure cross-modal significant interaction feature.

Images