Remote surface temperature adjusting method and system based on boiler heat supply

Abstract

The invention discloses a surface temperature remote adjustment method and system based on boiler heat supply, and relates to the technical field of temperature adjustment. Similarity comparison is carried out on a thermal field structure description unit constructed at the current moment and a plurality of thermal field structure description units stored in a historical thermal state feature library; screening out a plurality of historical state frames matched with the current thermodynamic state, calculating a comprehensive similarity score between the current thermodynamic state and each candidate historical state frame, and selecting the historical state frame with the maximum comprehensive similarity score as an optimal matching reference state of the current thermodynamic state; and based on the historical heat supply parameters corresponding to the optimal matching reference state, the current actual environment change and the real-time thermal demand difference are combined, and a current heat supply adjustment strategy is deduced through an adaptive parameter compensation algorithm. According to the adjusting system, through multi-modal data fusion and intelligent analysis, the accuracy and self-adaptability of boiler heat supply adjustment are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of temperature regulation technology, and more specifically to a method and system for remote regulation of surface temperature based on boiler heating. Background Technology

[0002] In traditional boiler heating systems, maintaining stable and comfortable surface temperatures within a region typically relies on manual experience or fixed parameters for heating regulation, making it difficult to respond in real-time to complex and changing environmental conditions and variations in thermal distribution. However, with the advancement of refined urban heating management, there is an urgent need for a technological solution capable of accurately sensing real-time surface temperatures and their dynamic changes across the entire area, and intelligently regulating them in conjunction with boiler operating status. Given that existing systems lack the ability to comprehensively analyze multidimensional environmental variables and thermal field structures, and cannot effectively identify local thermal characteristics and achieve adaptive regulation, there is an urgent need for an intelligent method based on multimodal data fusion and historical state matching to achieve remote and precise regulation of surface temperature based on boiler heating. Summary of the Invention

[0003] The purpose of this invention is to provide a method and system for remote regulation of surface temperature based on boiler heating, so as to solve the problems in the prior art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a method for remotely regulating surface temperature based on boiler heating, the method comprising the following steps:

[0005] Acquire real-time surface temperature information covering the entire area, synchronously collect multi-dimensional environmental variables of the boiler, form multi-modal data fusion input, and aggregate the continuously collected multi-modal data into a thermal field state frame with time and space labels;

[0006] The system performs feature analysis and local structure mining on the overall thermal distribution, marks thermal field feature points, analyzes the temperature distribution pattern in the neighborhood of the thermal field feature points, determines the type of role the thermal field feature points play in the local thermal field, and constructs thermal field structure description units that characterize the local thermal topology.

[0007] The thermal field structure description unit constructed at the current moment is compared with multiple thermal field structure description units stored in the historical thermal state feature library. Multiple historical state frames that match the current thermal state are selected. The comprehensive similarity score between the current thermal state and each candidate historical state frame is calculated. The historical state frame with the highest comprehensive similarity score is selected as the best matching reference state for the current thermal state. Based on the historical heating parameters corresponding to the best matching reference state, combined with the current actual environmental changes and the difference in real-time heat demand, the current heating regulation strategy is derived through an adaptive parameter compensation algorithm.

[0008] In a preferred embodiment, filtering out multiple historical state frames that match the current thermal state includes the following steps:

[0009] By evaluating the degree of matching in temperature distribution topology, spatial location correspondence, and temperature difference pattern, multiple historical state frames that match the current thermal state are selected, i.e., all historical state frames with a matching degree greater than the matching threshold.

[0010] In a preferred embodiment, calculating a comprehensive similarity score between the current thermal state and each candidate historical state frame includes the following steps:

[0011] After obtaining a set of candidate historical state frames, the comprehensive similarity score between the current thermal state and each candidate historical state frame is calculated. The comprehensive similarity score takes into account the similarity of temperature field structure, the matching degree of heating parameters, and the consistency of external environment.

[0012] In a preferred embodiment, the heating regulation strategy includes boiler supply water temperature regulation operation suggestions to establish a thermodynamic state mapping relationship between the current state and historical stable states.

[0013] In a preferred embodiment, the overall thermal distribution is characterized and local structure is mined, and thermal field feature points are marked, including the following steps:

[0014] Based on the obtained thermal field state frame, focus on the temperature change gradient and spatial points exhibiting transitional characteristics on the heat transfer path, and mark them as thermal field feature points. These thermal field feature points are located at the heating source, radiator layout points, joints of door and window enclosure structures, or areas where heat loss is concentrated.

[0015] In a preferred embodiment, the role of a thermal field feature point in a local thermal field is determined by analyzing the temperature distribution pattern within its neighborhood, including the following steps:

[0016] For each identified thermal field feature point, the role it plays in the local thermal field is determined by analyzing the temperature distribution pattern in its neighborhood. Some thermal field feature points exhibit thermal concentration characteristics due to the convergence of surrounding temperatures, corresponding to the outlet of heating output equipment or near the main pipeline. Other thermal field feature points form thermal depression characteristics due to the diffusion of heat from them to the surrounding areas, corresponding to the outer boundary, weak heat dissipation area, or location with poor heat transmission.

[0017] In a preferred embodiment, constructing a thermodynamic field structure description unit that characterizes a local thermodynamic topology includes the following steps:

[0018] Based on the combination of thermal field feature points and their surrounding temperature relationships, a thermal field structure description unit is constructed to characterize the local thermal topology pattern.

[0019] Each thermal field structure description unit consists of several interrelated thermal field feature points. By analyzing the temperature difference, relative position, and heat flow direction relationship between the thermal field feature points, a local thermal state fingerprint is formed to describe the change pattern of thermal distribution in the region.

[0020] The generated thermal field structure description units are uniformly stored and managed through an index structure, forming a thermal state feature library covering multiple historical moments.

[0021] In a preferred embodiment, real-time surface temperature information covering the entire area is acquired, multi-dimensional environmental variables of the boiler are collected synchronously, and multi-modal data fusion input is formed. The continuously collected multi-modal data is aggregated into a thermal field state frame with a time-space label, including the following steps:

[0022] Collect boiler operating parameters, pipeline control valve opening status, and external meteorological data;

[0023] According to the set time period and spatial partitioning rules, the continuously collected multimodal data is aggregated into a thermal field state frame with spatiotemporal labels.

[0024] Each frame of data records the surface temperature value at each monitoring point, and also associates it with the physical area identifier, heating input, and environmental background at that time.

[0025] In a preferred embodiment, for heating scenarios with three-dimensional distribution or multi-story structures, temperature data is smoothed or layered and aggregated along the vertical direction to generate an average thermal field snapshot that reflects the overall thermal distribution trend.

[0026] This application also provides a remote surface temperature control system based on boiler heating, including:

[0027] Acquire real-time surface temperature information covering the entire area, and simultaneously collect multi-dimensional environmental variables of the boiler to form a multi-modal data fusion input;

[0028] Acquisition module: Aggregates continuously acquired multimodal data into thermal field state frames with spatiotemporal tags;

[0029] The tagging module performs feature analysis on the overall thermal distribution and mines local structures, marking feature points of the thermal field;

[0030] Description Unit Construction Module: By analyzing the temperature distribution pattern within the neighborhood of a thermal field feature point, the module determines the type of role that feature point plays in the local thermal field and constructs a thermal field structure description unit that characterizes the local thermal topology.

[0031] The filtering module compares the thermal field structure description unit constructed at the current moment with multiple thermal field structure description units stored in the historical thermal state feature library to filter out multiple historical state frames that match the current thermal state.

[0032] State matching module: Calculates the comprehensive similarity score between the current thermal state and each candidate historical state frame, and selects the historical state frame with the highest comprehensive similarity score as the best matching reference state for the current thermal state;

[0033] Strategy derivation module: Based on the historical heating parameters corresponding to the best matching reference state, and combined with the current actual environmental changes and real-time heat demand differences, the current heating regulation strategy is derived through an adaptive parameter compensation algorithm.

[0034] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0035] This invention, combining current environmental changes with real-time differences in heat demand, introduces an adaptive parameter compensation algorithm. This algorithm dynamically adjusts and derives the most suitable heating regulation strategy for the current situation, enabling a shift from experience-driven to data-driven and intelligent decision-making. This not only improves heating efficiency and energy utilization but also significantly enhances the system's adaptability and response speed to complex operating conditions, ultimately achieving the technical effects of improving the user's heating experience and reducing energy consumption and operating costs.

[0036] This invention acquires real-time surface temperature information for the entire region and simultaneously collects multi-dimensional environmental variables of boiler operation. It organically integrates data from different sources and assigns spatiotemporal labels to form a complete thermal field state frame, thereby achieving comprehensive and high-precision perception of the operating status of the heating system and the distribution of the thermal environment, providing a detailed data foundation for subsequent analysis.

[0037] This invention, through feature analysis of the overall distribution of the thermal field and local structure mining, can accurately mark key feature points in the thermal field, and further analyze the role of these feature points in the local thermal field. It constructs a structural description unit that can accurately characterize the local thermal topology pattern, enabling the system to deeply understand the internal structure and dynamic change law of thermal distribution and enhance the cognitive ability of complex thermal environments.

[0038] This invention efficiently compares the current thermal field structure description unit with multiple historical states in the historical state feature library, which can quickly filter out the historical reference state that best matches the current state, and select the optimal match based on a comprehensive similarity scoring mechanism. This effectively utilizes historical operating experience to provide a reliable reference for current heating regulation. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0040] Figure 1 This is a flowchart of the identification method of the present invention. Detailed Implementation

[0041] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0042] Example: This example provides a method for remotely regulating surface temperature based on boiler heating. Please refer to [link / reference]. Figure 1 As shown, the adjustment method includes the following steps:

[0043] Relying on various types of sensing devices deployed in the heating area, including temperature monitoring probes, infrared thermal imaging devices, and environmental temperature and humidity acquisition devices distributed on the ground or under the floor, the system simultaneously acquires real-time surface temperature information covering the entire area; it also simultaneously collects boiler operating parameters (such as water supply temperature, flow rate, and pressure), pipeline control valve opening status, and external meteorological data (such as ambient temperature, solar radiation intensity, and wind force level) and other multi-dimensional environmental variables to form a multi-modal data fusion input.

[0044] According to the set time period (e.g., every 5 minutes) and spatial zoning rules (e.g., divided by functional area, floor, or building plan grid), the continuously collected multimodal data is aggregated into thermal field state frames with spatiotemporal labels. Each frame of data not only records the surface temperature value at each monitoring point, but also associates it with the physical area identifier and the heating input and environmental background at that time.

[0045] For complex heating scenarios with three-dimensional distribution or multi-story structures, temperature data is further smoothed or layered along the vertical direction (e.g., floor height) to eliminate interference from single-point abrupt changes or sensor errors, generating an average thermal field snapshot that reflects the overall thermal distribution trend. This processing helps to weaken local noise and enhance the overall characteristics of regional thermal distribution.

[0046] Based on the obtained thermal field state frames, feature analysis and local structure mining are performed on the overall thermal distribution. Special attention is paid to spatial points with significant temperature gradients and obvious transitional characteristics along the heat transfer path, which are marked as thermal field feature points. These points are often located near heating sources, radiator placement points, joints in building envelopes such as doors and windows, or areas with concentrated heat loss; they are key nodes in the thermal field topology.

[0047] For each identified thermal field feature point, the role it plays in the local thermal field is determined by analyzing the temperature distribution pattern in its neighborhood. For example, some points exhibit thermal concentration characteristics because the surrounding temperature converges towards them, usually corresponding to equipment outlets or near main pipelines with strong heat output; while other points form thermal depression characteristics because heat diffuses from them to the surrounding areas, commonly found at outer boundaries, weak heat dissipation areas, or locations with poor heat transmission.

[0048] Based on the combination of the aforementioned thermal field feature points and their surrounding temperature relationships, a thermal field structure description unit capable of characterizing local thermal topological patterns is constructed. Each thermal field structure description unit typically consists of several (e.g., three) interconnected thermal field feature points. By analyzing the temperature difference, relative position, and heat flow direction relationship between the thermal field feature points, a distinctive local thermal state fingerprint is formed to describe the changing patterns of thermal distribution in that region.

[0049] All generated thermal field structure description units are uniformly stored and managed through efficient index structures (such as hash tables or KD trees), forming a thermal state feature library covering multiple historical moments.

[0050] The thermal field structure description unit constructed at the current moment is compared one by one with multiple thermal field structure description units stored in the historical thermal state feature library. By evaluating their matching degree in terms of temperature distribution topology, spatial correspondence, and temperature difference patterns, multiple historical state frames that match the current thermal state are selected (i.e., all historical state frames with a matching degree greater than the matching threshold, representing potentially referable past thermal distribution states). These historical state frames reflect stable heating states experienced under similar environmental and load conditions.

[0051] After obtaining a set of candidate historical state frames, a comprehensive similarity score is further calculated between the current thermal state and each candidate historical state frame. This score comprehensively considers multiple dimensions such as temperature field structure similarity, heating parameter matching degree (e.g., boiler output, valve opening), and external environment consistency (e.g., outdoor temperature, solar radiation intensity). The historical state frame with the highest comprehensive similarity score is selected as the best matching reference state for the current thermal state, and the heating operation parameters used in this state are extracted as the benchmark input for regulation and optimization.

[0052] Based on historical heating parameters corresponding to the optimal matching reference state, and combined with current environmental changes (such as sudden temperature drops and increased human activity) and real-time differences in heat demand, an adaptive parameter compensation algorithm is used to derive the most suitable heating regulation strategy, including operational suggestions such as fine-tuning of boiler supply water temperature. This regulation strategy can essentially be viewed as establishing a thermodynamic state mapping relationship between the current state and historical stable states, providing a direction for adjustment from the current abnormal or fluctuating state to a stable and comfortable state.

[0053] This embodiment also provides a remote surface temperature control system based on boiler heating, including:

[0054] Acquire real-time surface temperature information covering the entire area, and simultaneously collect multi-dimensional environmental variables of the boiler to form a multi-modal data fusion input;

[0055] Acquisition module: Aggregates continuously acquired multimodal data into thermal field state frames with spatiotemporal labels, and sends the thermal field state frames to the tagging module;

[0056] The tagging module performs feature analysis and local structure mining on the overall thermal distribution, tags thermal field feature points, and sends these feature points to the description unit construction module.

[0057] Description Unit Construction Module: By analyzing the temperature distribution pattern within the neighborhood of a thermal field feature point, the module determines the type of role that feature point plays in the local thermal field and constructs a thermal field structure description unit that characterizes the local thermal topology. The thermal field structure description unit is then sent to the filtering module.

[0058] The filtering module compares the thermal field structure description unit constructed at the current moment with multiple thermal field structure description units stored in the historical thermal state feature library to filter out multiple historical state frames that match the current thermal state. The historical state frames are then sent to the state matching module.

[0059] State matching module: Calculates the comprehensive similarity score between the current thermal state and each candidate historical state frame, selects the historical state frame with the largest comprehensive similarity score as the best matching reference state for the current thermal state, and sends the best matching reference state to the strategy derivation module;

[0060] Strategy derivation module: Based on the historical heating parameters corresponding to the best matching reference state, and combined with the current actual environmental changes and real-time heat demand differences, the current heating regulation strategy is derived through an adaptive parameter compensation algorithm.

[0061] The following is a detailed description of each step of the adjustment method in this application:

[0062] Relying on various types of sensing devices deployed in the heating area, including temperature monitoring probes, infrared thermal imaging devices, and environmental temperature and humidity acquisition devices distributed on the ground or under the floor, the system simultaneously acquires real-time surface temperature information covering the entire area; it also simultaneously collects boiler operating parameters (such as water supply temperature, flow rate, and pressure), pipeline control valve opening status, and external meteorological data (such as ambient temperature, solar radiation intensity, and wind force level) and other multi-dimensional environmental variables to form a multi-modal data fusion input.

[0063] According to the set time period (e.g., every 5 minutes) and spatial zoning rules (e.g., divided by functional area, floor, or building plan grid), the continuously collected multimodal data is aggregated into thermal field state frames with spatiotemporal labels. Each frame of data not only records the surface temperature value at each monitoring point, but also associates it with the physical area identifier and the heating input and environmental background at that time.

[0064] For complex heating scenarios with three-dimensional distribution or multi-story structures, temperature data is further smoothed or layered along the vertical direction (e.g., floor height) to eliminate interference from single-point abrupt changes or sensor errors, generating an average thermal field snapshot that reflects the overall thermal distribution trend. This processing helps to weaken local noise and enhance the overall characteristics of regional thermal distribution.

[0065] In one embodiment disclosed in this application, a high-precision temperature monitoring probe (such as an embedded digital temperature sensor with a sampling accuracy of ±0.1℃) deployed on the ground or under the floor directly collects continuous temperature data near the ground surface.

[0066] Infrared thermal imaging devices (such as fixed mid-wave infrared cameras with spatial resolution ≤1cm / pixel) acquire temperature distribution images of large areas of the earth's surface in a non-contact manner, supplementing the spatial coverage blind spots of traditional point monitoring.

[0067] The environmental temperature and humidity acquisition device (integrated temperature and humidity sensor, measurement range -20℃~60℃ / 0~100%RH, accuracy ±0.5℃ / ±2%RH) synchronously records the environmental temperature and humidity parameters of key outdoor and indoor locations.

[0068] Boiler-side operating parameters (including supply water temperature (unit: °C, collected by a platinum resistance sensor installed on the main pipeline) and circulating water flow rate (unit: m³) 3 The core input variables for heating are: flow rate ( / h, measured in real time by an electromagnetic flow meter), pressure (unit: MPa, monitored by a pressure transmitter), burner power (unit: kW, calculated based on gas flow rate and calorific value) and the opening status of pipeline control valves (unit: % opening, fed back by electric actuators).

[0069] External meteorological data (such as ambient temperature (from regional weather stations or miniature weather sensors), solar radiation intensity (unit: W / m²) 2 The data, including radiation sensor measurements, wind speed (converted to wind speed m / s, affecting heat dissipation of the building envelope), and wind force level (converted to wind speed m / s, affecting heat dissipation of the building envelope), are used as boundary condition inputs to form a multimodal data fusion input covering user-side thermal response, boiler operating status, and environmental disturbances.

[0070] In one embodiment disclosed in this application, continuously acquired multimodal data is spatiotemporally aligned and structurally encapsulated according to a preset time period (every 5 minutes, which can be adjusted to 1-10 minutes according to actual accuracy requirements) and spatial zoning rules (such as dividing the building into residential areas, public activity areas, and equipment rooms according to building function, or dividing it into 1F-A area, 2F-B area, and X-axis 3m × Y-axis 3m grid units according to floor / plane grid). This generates thermal field state frames with clear spatiotemporal labels. The core structure of each frame of data includes three parts:

[0071] ① Surface temperature values ​​at each monitoring point (stored in matrix form, with the dimension being spatial partition × monitoring sensor ID).

[0072] ② Associated physical area identifiers (such as 3F-living room grid (5,5) or boiler room outlet pipes);

[0073] ③ Current heating input parameters (water supply temperature, flow rate, valve opening, etc.) and environmental background parameters (outdoor temperature, solar radiation, etc.).

[0074] For example, the thermal field status frame of the living room area on the first floor of a residential building at 14:00 will record the real-time values ​​of five temperature probes in that area (such as 22.1℃, 21.9℃, 22.3℃, etc.), the corresponding grid coordinates ((1,1) to (1,5)), water supply temperature (45℃), flow rate (12m³ / h), outdoor temperature (18℃), and solar radiation intensity (300W / m²), ensuring data traceability and spatial correlation.

[0075] In one embodiment disclosed in this application, for complex heating scenarios with three-dimensional distribution (such as floors 1-10 of a high-rise residential building) or multi-story structures, the raw temperature data may suffer from noise problems such as single-point sensor failure (e.g., a probe momentarily drifts to 30°C, significantly deviating from the normal range of 20-25°C) and local heat source interference (e.g., abnormal local temperature rise caused by electrical equipment heating). Therefore, a vertical (floor height) data smoothing and layered aggregation processing logic is introduced:

[0076] Temperature data from monitoring points within the same vertical zone (such as all grid units on the 3rd floor or the 3rd to 5th floors covered by a certain vertical heating riser) are grouped according to preset stratification rules (such as dividing by natural floors or by each 3-meter layer in height intervals).

[0077] A local mean filtering algorithm based on a sliding window is adopted (processing logic: for the temperature data sequence within each layer, select N consecutive time points (such as the current time and one period before and after, for a total of 3 data points) or M spatially adjacent monitoring points (such as 4 adjacent probes in the same floor), remove outliers that exceed the threshold range (such as ±3σ, where σ is the standard deviation of the historical temperature of the layer), and calculate the arithmetic mean of the remaining valid data as the representative temperature of the layer), generating an average thermal field snapshot that can reflect the overall thermal distribution trend.

[0078] The core objective is to reduce the interference of local noise (such as abrupt changes caused by single-point probe failure) on the overall judgment, while preserving the true thermal gradient changes (such as the temperature of higher floors being slightly lower than that of lower floors due to faster heat dissipation).

[0079] For example, in a certain period, the temperature of floors 5-7 of a 10-story residential building briefly rose to 28℃ (normal range 22-24℃) due to direct sunlight on a probe on floor 5. By using sliding window mean filtering (selecting current and previous period data from 12 valid probes on floors 5-7, removing the 28℃ outlier, and calculating the mean of the remaining 11 data points), the representative temperature of floors 5-7 was corrected to 23.5℃ (close to the actual thermal level). The resulting vertically stratified average thermal field snapshot can accurately reflect the overall distribution trend of higher temperatures (24-25℃) in the lower floors (floors 1-3), moderate temperatures (23-24℃) in the middle floors (floors 4-6), and lower temperatures (22-23℃) in the upper floors (floors 7-10), providing reliable basic data for subsequent thermal field characteristic analysis and adjustment strategy generation.

[0080] Based on the obtained thermal field state frames, feature analysis and local structure mining are performed on the overall thermal distribution. Special attention is paid to spatial points with significant temperature gradients and obvious transitional characteristics along the heat transfer path, which are marked as thermal field feature points. These points are often located near heating sources, radiator placement points, joints in building envelopes such as doors and windows, or areas with concentrated heat loss; they are key nodes in the thermal field topology.

[0081] For each identified thermal field feature point, the role it plays in the local thermal field is determined by analyzing the temperature distribution pattern in its neighborhood. For example, some points exhibit thermal concentration characteristics because the surrounding temperature converges towards them, usually corresponding to equipment outlets or near main pipelines with strong heat output; while other points form thermal depression characteristics because heat diffuses from them to the surrounding areas, commonly found at outer boundaries, weak heat dissipation areas, or locations with poor heat transmission.

[0082] Based on the combination of the aforementioned thermal field feature points and their surrounding temperature relationships, a thermal field structure description unit capable of characterizing local thermal topological patterns is constructed. Each thermal field structure description unit typically consists of several (e.g., three) interconnected thermal field feature points. By analyzing the temperature difference, relative position, and heat flow direction relationship between the thermal field feature points, a distinctive local thermal state fingerprint is formed to describe the changing patterns of thermal distribution in that region.

[0083] All generated thermal field structure description units are uniformly stored and managed through efficient index structures (such as hash tables or KD trees), forming a thermal state feature library covering multiple historical moments.

[0084] In one embodiment disclosed in this application, based on the temperature values ​​(stored in matrix form) and spatial coordinates (determined by the sensor deployment location or grid division coordinates) of each monitoring point in the thermal field state frame, the temperature gradient vector of each point is calculated (processing logic: for each point, select the K nearest neighbor monitoring points in its spatial neighborhood (K value is dynamically set according to spatial resolution, such as 8 adjacent points in a 3×3 grid area), calculate the temperature difference vector between the current point and the neighboring points (ΔT=T_current-T_neighbor), and then synthesize the direction and magnitude of the fastest temperature change (gradient magnitude=||ΔT||, gradient direction=arctan(ΔT_y / ΔT_x))).

[0085] Combined with the physical characteristics of heat transfer (such as the heat conduction path usually extending along the building structure or pipeline direction), identify the spatial points where the temperature change gradient is significant (such as the gradient amplitude exceeding the global threshold τ, which is set to 2-3 times the average gradient of the same type of area according to historical data statistics) and located at the key nodes of the heat transfer path (such as the outlet of the main heating pipeline, the inlet and outlet of the radiator, the gaps of doors and windows, etc.), and mark them as heat field feature points. These feature points are usually distributed in three types of areas:

[0086] One is the heat supply energy input end (such as near the outlet pipeline of the boiler, at the outlet of the circulation pump), which is manifested as high-temperature aggregation and gradient divergence; the second is the turning point of the heat transfer path (such as the pipeline elbow, valve interface), where local temperature anomalies are caused by flow resistance; the third is the heat loss sensitive area (such as the joints of the peripheral enclosure structure, the inner side of the glass curtain wall), where gradient convergence is formed due to the higher heat dissipation intensity than the surrounding area.

[0087] In an implementation manner disclosed in the present application, for each identified heat field feature point, further analyze the temperature distribution pattern within its neighborhood range to determine the role type (i.e., functional attribute) of this point in the local heat field. Specifically:

[0088] Taking the feature point as the center, define a neighborhood range with a radius of r (r is set according to the spatial grid density, such as 0.5-1 meter, or covering 3×3 monitoring grid cells), extract the temperature values of all monitoring points within the neighborhood, and calculate the statistical characteristics of the temperature distribution in this neighborhood (including the mean T_mean, variance T_var, maximum temperature difference ΔT_max = T_max - T_min) and the heat flow direction indication parameter (by comparing the relative magnitude relationship between the temperature T_feature of the feature point and the neighborhood temperature:

[0089] If T_feature is significantly higher than the neighborhood mean (such as T_feature > T_mean + δ, δ is the neighborhood temperature difference threshold, usually taking 0.5-1 °C), it is determined that heat diffuses to the neighborhood (divergent role); if T_feature is significantly lower than the neighborhood mean (such as T_feature < T_mean - δ), it is determined that heat converges to the feature point (convergent role)).

[0090] Based on the combination of the above statistical characteristics, the role type of the feature point can be divided into four types of modes:

[0091] Thermal concentration node (divergent + high T_feature), commonly found at the outlet of heating equipment (such as the end of the main boiler pipeline), characterized by high self-temperature and decreasing surrounding temperature;

[0092] Thermal diffusion node (divergent + medium T_feature), corresponding to the surface of the radiator or the branch point of the heating riser, manifested as heat evenly radiating to the surroundings;

[0093] Heat accumulation nodes (convergence type + low T_feature) are often found at the joints of external walls and uninsulated pipe sections, where heat accumulates due to the low external temperature environment or heat dissipation defects.

[0094] Thermal transition nodes (high T_var but no obvious concentration / convergence trend) are usually located in areas with gentle thermal gradients or mixed heat transfer areas (such as the junction of walls of different materials).

[0095] For example, a monitoring point directly above the radiator outlet in the living room of a residential building has a neighborhood average temperature of 23℃, its own temperature of 42℃ (T_feature>T_mean+19℃), and a neighborhood temperature difference ΔT_max of 18℃. It is determined to be a heat concentration node, corresponding to a point of strong energy release from the heat supply output.

[0096] In one embodiment disclosed in this application, each structural description unit consists of 3 to 5 interconnected feature points (3 values, representing energy source point, transition node, and boundary point respectively). By analyzing the temperature difference relationship between these points (e.g., ΔT_1→2=T_point1-T_point2, reflecting the direction of heat flow), relative spatial position (e.g., Euclidean distance d_point1→2, combined with the building plan layout to judge the rationality of the heat transfer path), and consistency of heat flow direction (e.g., whether the temperature gradient direction of multiple adjacent points points to the same convergence point or away from the same source point), a unique local thermal state fingerprint is generated.

[0097] For example, in a localized area of ​​a radiator-wall-window, the structural description unit might include:

[0098] Feature point A (radiator outlet, heat concentration node, temperature 42℃), feature point B (midpoint between radiator and wall, heat diffusion node, temperature 28℃), and feature point C (wall joint near window, heat convergence node, temperature 19℃). The temperature difference relationship of this unit is ΔT_A→B=14℃ (heat is transferred from the radiator to the wall) and ΔT_B→C=9℃ (heat continues to diffuse towards the window). The relative positions show that the three points are linearly arranged along the heat flow direction (distance A→B is 0.3 meters, B→C is 0.5 meters), and the temperature gradient direction is consistent (both from high temperature to low temperature). Finally, a fingerprint pattern of high temperature source → medium temperature transition → low temperature convergence is formed to describe the thermal topology of strong radiation from the radiator, conduction through the wall, and heat loss near the window in this local area.

[0099] In one embodiment disclosed in this application, when the structural description unit at the current moment contains the fingerprint of ΔT_1→2≈14℃+ΔT_2→3≈9℃+role sequence [concentration→diffusion→convergence], a hash table can be used to quickly locate units in the historical state database that have the same or similar fingerprints (allowing a temperature difference tolerance of ±1℃ and interchangeability of role types), providing a data foundation for subsequent optimal matching reference state selection. The construction of this feature database not only realizes the long-term storage and reuse of the microstructure of the thermal field, but also lays a key data foundation for the generation of intelligent adjustment strategies based on historical experience.

[0100] The thermal field structure description unit constructed at the current moment is compared one by one with multiple thermal field structure description units stored in the historical thermal state feature library. By evaluating their matching degree in terms of temperature distribution topology, spatial correspondence, and temperature difference patterns, multiple historical state frames that match the current thermal state are selected (i.e., all historical state frames with a matching degree greater than the matching threshold, representing potentially referable past thermal distribution states). These historical state frames reflect stable heating states experienced under similar environmental and load conditions.

[0101] After obtaining a set of candidate historical state frames, a comprehensive similarity score is further calculated between the current thermal state and each candidate historical state frame. This score comprehensively considers multiple dimensions such as temperature field structure similarity, heating parameter matching degree (e.g., boiler output, valve opening), and external environment consistency (e.g., outdoor temperature, solar radiation intensity). The historical state frame with the highest comprehensive similarity score is selected as the best matching reference state for the current thermal state, and the heating operation parameters used in this state are extracted as the benchmark input for regulation and optimization.

[0102] Based on historical heating parameters corresponding to the optimal matching reference state, and combined with current environmental changes (such as sudden temperature drops and increased human activity) and real-time differences in heat demand, an adaptive parameter compensation algorithm is used to derive the most suitable heating regulation strategy, including operational suggestions such as fine-tuning of boiler supply water temperature. This regulation strategy can essentially be viewed as establishing a thermodynamic state mapping relationship between the current state and historical stable states, providing a direction for adjustment from the current abnormal or fluctuating state to a stable and comfortable state.

[0103] In one embodiment disclosed in this application, a similarity comparison of the thermal field structure description units is performed, and the specific logic is as follows:

[0104] The structural description unit generated at the current moment (including metadata such as feature point coordinates, role type, temperature difference relationship, and spatial position relationship) is compared one by one with all historical structural description units stored in the historical thermal state feature library (covering stable heating states under multiple time periods and environmental conditions). The degree of matching between the two is evaluated in key dimensions such as temperature distribution topology (e.g., whether the role type combination of feature points is consistent, such as the current sequence of central node → diffusion node → convergence node, and whether historical candidate units have the same or similar role sequence), spatial position correspondence (e.g., whether the relative spatial layout between feature points matches, such as the three feature points in the current unit are linearly arranged with a spacing of 0.3 m / 0.5 m / 0.8 m, and whether the spatial topology of the corresponding feature points in the historical unit has a similar geometric relationship), and temperature difference pattern (e.g. whether the temperature difference between adjacent feature points is of the same order of magnitude, such as ΔT_1→2=14℃±1℃ and ΔT_2→3=9℃±1℃ in the current unit, and whether the corresponding difference in the historical unit is within the allowable deviation range).

[0105] The comparison process uses preset matching thresholds (e.g., role type matching rate ≥90%, spatial topological similarity ≥85%, temperature difference deviation ≤10%) to filter and retain only historical state frames with matching degrees exceeding the thresholds (i.e., potentially referable past thermal distribution states). These frames typically correspond to heating states that operate stably and achieve comfort targets under similar environmental conditions (e.g., outdoor temperature, solar radiation intensity) and load demands (e.g., indoor occupant activity levels, building envelope heat loss). For example, if the current time is a sunny winter afternoon (outdoor temperature 5℃, solar radiation intensity 400W / m²), the comparison will be used to determine the heating status. 2 The characteristic point role sequence and temperature difference pattern of the living room radiator area are compared with the same period on a historical day (outdoor temperature 6℃, solar radiation 380W / m). 2 If the stable states of the historical frames are highly similar, then the historical frame will be included in the candidate set.

[0106] In one embodiment disclosed in this application, after obtaining a set of candidate historical state frames, a comprehensive similarity score is further calculated. This score is a weighted comprehensive result of multi-dimensional similarity indicators, specifically:

[0107] The temperature field structure similarity sub-score is based on parameters such as feature point role type matching degree (e.g., consistency rate between current and historical role sequences), spatial topological similarity (e.g., mean Euclidean distance deviation of relative positions of feature points), and temperature difference pattern consistency (e.g., mean relative error of differences between adjacent points). After normalization (e.g., mapping each parameter to the 0~1 interval), the parameters are weighted and summed (weights are configurable, for example, role matching degree accounts for 40%, spatial topology accounts for 30%, and temperature difference accounts for 30%).

[0108] The heating parameter matching degree sub-score compares the key operating parameters of the boiler (such as water supply temperature, circulation flow rate, and valve opening) in the current and historical status frames, calculates the relative deviation of the parameter values ​​(e.g., the deviation between the current water supply temperature of 45℃ and the historical value of 43℃ is (45-43) / 43≈4.7%, and the score is normalized), and combines the engineering importance weight of each parameter (e.g., water supply temperature weight 50%, valve opening weight 30%, flow rate weight 20%) to obtain the sub-score;

[0109] The external environment consistency sub-score assesses the similarity of external meteorological conditions (such as outdoor temperature, solar radiation intensity, and wind force) between the current and historical state frames. It is calculated by the ratio of the absolute difference of environmental parameters (such as the difference in outdoor temperature ΔT_outdoor = |current 5℃ - historical 6℃| = 1℃) to the environmental sensitivity threshold (such as the outdoor temperature sensitivity threshold 2℃) (1 / 2 = 50%, normalized score), combined with the comprehensive deviation assessment of parameters such as solar radiation intensity.

[0110] The overall similarity score is calculated by weighting the three sub-scores mentioned above according to preset weights (e.g., temperature field structure accounts for 50%, heating parameters account for 30%, and external environment accounts for 20%), and the historical state frame with the highest score is selected as the best matching reference state. If the current candidate set contains historical frame A (overall score 0.82), historical frame B (overall score 0.76), and historical frame C (overall score 0.85), then historical frame C will be selected as the best reference, and its corresponding heating parameters (e.g., water supply temperature 44℃, valve opening 75%, circulation flow rate 11m³) will be used. 3 / h) will be used as the baseline input for regulation optimization.

[0111] In one embodiment disclosed in this application, by analyzing the differences between the current actual environment, real-time thermal demand, and historical reference states, historical parameters are dynamically corrected to establish a thermal state mapping relationship from the current fluctuating state to the historical stable state (similar to rigid body transformation in spatial registration, i.e., adjusting key parameters to achieve state alignment while keeping the overall thermal distribution structure unchanged). The processing steps are as follows:

[0112] Identify key differences between the current and historical reference states, including environmental changes (such as a 3°C drop in current outdoor temperature or a 200 W / m² decrease in solar radiation intensity compared to historical levels). 2 ), real-time heat demand differences (e.g., increased activity in a room leads to a rise in actual heat demand, which is reflected in the monitoring point temperature being 1°C lower than the historical average for the same period);

[0113] Engineering impact weights based on differences (e.g., for every 1°C decrease in outdoor temperature, the water supply temperature needs to be increased by 0.8~1.2°C; for every 100W / m² decrease in solar radiation intensity) 2The valve opening needs to be increased by 5% to 8%; for every 1°C of temperature deviation caused by increased personnel activity, the water supply temperature needs to be adjusted by 0.5 to 1°C. Calculate the compensation for each heating parameter (for example: if the current outdoor temperature is 3°C lower than the historical average, the water supply temperature needs to be compensated by 3 × 1.0 = 3°C according to the weighted calculation; if solar radiation decreases by 150W / m²...). 2 The valve opening needs to be compensated by 150 / 100×6%=9%; if personnel activity causes the temperature at the monitoring point to be 1℃ lower, the water supply temperature needs to be compensated by an additional 0.8℃, for a total compensation of 3+0.8=3.8℃.

[0114] The compensation amount is superimposed on historical reference parameters (e.g., historical water supply temperature 44℃ + 3.8℃ ≈ 47.8℃, historical valve opening 75% + 9% ≈ 84%) to generate the current adjustment strategy (e.g., increase the boiler water supply temperature to 48℃ (rounded to an integer), and adjust the valve opening to 85%).

[0115] The essence of this strategy is to guide heating from abnormal or fluctuating states (such as localized overcooling or uneven temperature distribution) to a stable and comfortable state (such as uniform temperature across the entire area and meeting design requirements) by using historical stable state parameter benchmarks and quantitative compensation for current actual deviations, while avoiding energy waste caused by over-adjustment. For example, in the aforementioned residential living room area, if the current monitoring point temperature is 21℃ (historical average for the same period is 22℃), through comprehensive similarity matching and parameter compensation, the final output water supply temperature is adjusted to 48℃, and the valve opening is set at 85%. This can effectively improve the heat supply intensity of the area, allowing the temperature to rise back to the target range (22~24℃) within 15~30 minutes, achieving precise and efficient remote adjustment.

[0116] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0117] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for remote regulation of surface temperature based on boiler heating, characterized in that: The adjustment method includes the following steps: Acquire real-time surface temperature information covering the entire area, synchronously collect multi-dimensional environmental variables of the boiler, form multi-modal data fusion input, and aggregate the continuously collected multi-modal data into a thermal field state frame with time and space labels; The system performs feature analysis and local structure mining on the overall thermal distribution, marks thermal field feature points, analyzes the temperature distribution pattern in the neighborhood of the thermal field feature points, determines the type of role the thermal field feature points play in the local thermal field, and constructs thermal field structure description units that characterize the local thermal topology. The thermal field structure description unit constructed at the current moment is compared with multiple thermal field structure description units stored in the historical thermal state feature library. Multiple historical state frames that match the current thermal state are selected. The comprehensive similarity score between the current thermal state and each candidate historical state frame is calculated. The historical state frame with the highest comprehensive similarity score is selected as the best matching reference state for the current thermal state. Based on the historical heating parameters corresponding to the best matching reference state, combined with the current actual environmental changes and the difference in real-time heat demand, the current heating regulation strategy is derived through an adaptive parameter compensation algorithm.

2. The method for remote regulation of surface temperature based on boiler heating according to claim 1, characterized in that: Filtering out multiple historical state frames that match the current thermal state includes the following steps: By evaluating the degree of matching in temperature distribution topology, spatial location correspondence, and temperature difference pattern, multiple historical state frames that match the current thermal state are selected, i.e., all historical state frames with a matching degree greater than the matching threshold.

3. The method for remote regulation of surface temperature based on boiler heating according to claim 2, characterized in that: Calculate the comprehensive similarity score between the current thermal state and each candidate historical state frame, including the following steps: After obtaining a set of candidate historical state frames, the comprehensive similarity score between the current thermal state and each candidate historical state frame is calculated. The comprehensive similarity score takes into account the similarity of temperature field structure, the matching degree of heating parameters, and the consistency of external environment.

4. The method for remote regulation of surface temperature based on boiler heating according to claim 3, characterized in that: The heating regulation strategy includes boiler supply water temperature regulation operation suggestions, which are used to establish a thermodynamic state mapping relationship between the current state and historical stable states.

5. The method for remote regulation of surface temperature based on boiler heating according to claim 1, characterized in that: The process of performing feature analysis and local structure mining on the overall thermal distribution, and marking characteristic points of the thermal field, includes the following steps: Based on the obtained thermal field state frame, focus on the temperature change gradient and spatial points exhibiting transitional characteristics on the heat transfer path, and mark them as thermal field feature points. These thermal field feature points are located at the heating source, radiator layout points, joints of door and window enclosure structures, or areas where heat loss is concentrated.

6. The method for remote regulation of surface temperature based on boiler heating according to claim 2, characterized in that: By analyzing the temperature distribution pattern within the neighborhood of a characteristic point in a thermal field, the type of role that characteristic point plays in the local thermal field is determined, including the following steps: For each identified thermal field feature point, the role it plays in the local thermal field is determined by analyzing the temperature distribution pattern in its neighborhood. Some thermal field feature points exhibit thermal concentration characteristics due to the convergence of surrounding temperatures, corresponding to the outlet of heating output equipment or near the main pipeline. Other thermal field feature points form thermal depression characteristics due to the diffusion of heat from them to the surrounding areas, corresponding to the outer boundary, weak heat dissipation area, or location with poor heat transmission.

7. The method for remote regulation of surface temperature based on boiler heating according to claim 6, characterized in that: Constructing a thermodynamic field structure description unit that characterizes a local thermodynamic topology includes the following steps: Based on the combination of thermal field feature points and their surrounding temperature relationships, a thermal field structure description unit is constructed to characterize the local thermal topology pattern. Each thermal field structure description unit consists of several interrelated thermal field feature points. By analyzing the temperature difference, relative position, and heat flow direction relationship between the thermal field feature points, a local thermal state fingerprint is formed to describe the change pattern of thermal distribution in the region. The generated thermal field structure description units are uniformly stored and managed through an index structure, forming a thermal state feature library covering multiple historical moments.

8. The method for remote regulation of surface temperature based on boiler heating according to claim 1, characterized in that: To acquire real-time surface temperature information covering the entire area, simultaneously collect multi-dimensional environmental variables from the boiler, form a multi-modal data fusion input, and aggregate the continuously collected multi-modal data into a thermodynamic field state frame with a time-space label, the following steps are included: Collect boiler operating parameters, pipeline control valve opening status, and external meteorological data; According to the set time period and spatial partitioning rules, the continuously collected multimodal data is aggregated into a thermal field state frame with spatiotemporal labels. Each frame of data records the surface temperature value at each monitoring point, and also associates it with the physical area identifier, heating input, and environmental background at that time.

9. The method for remote regulation of surface temperature based on boiler heating according to claim 8, characterized in that: For heating scenarios with three-dimensional distribution or multi-story structures, temperature data is smoothed or aggregated in layers along the vertical direction to generate an average thermal field snapshot that reflects the overall thermal distribution trend.

10. A remote surface temperature regulation system based on boiler heating, used to implement the regulation method according to any one of claims 1-9, characterized in that: include: Acquire real-time surface temperature information covering the entire area, and simultaneously collect multi-dimensional environmental variables of the boiler to form a multi-modal data fusion input; Acquisition module: Aggregates continuously acquired multimodal data into thermal field state frames with spatiotemporal tags; The tagging module performs feature analysis on the overall thermal distribution and mines local structures, marking feature points of the thermal field; Description Unit Construction Module: By analyzing the temperature distribution pattern within the neighborhood of a thermal field feature point, the module determines the type of role that feature point plays in the local thermal field and constructs a thermal field structure description unit that characterizes the local thermal topology. The filtering module compares the thermal field structure description unit constructed at the current moment with multiple thermal field structure description units stored in the historical thermal state feature library to filter out multiple historical state frames that match the current thermal state. State matching module: Calculates the comprehensive similarity score between the current thermal state and each candidate historical state frame, and selects the historical state frame with the highest comprehensive similarity score as the best matching reference state for the current thermal state; Strategy derivation module: Based on the historical heating parameters corresponding to the best matching reference state, and combined with the current actual environmental changes and real-time heat demand differences, the current heating regulation strategy is derived through an adaptive parameter compensation algorithm.

Images