Building heating ventilation system control method, device, medium and system

Abstract

This application provides a control method, device, medium, and system for a building HVAC system, belonging to the field of building HVAC control technology. The method collects multi-source heterogeneous environmental data inside and outside the building through a distributed sensor network, preprocesses the multi-source heterogeneous environmental data to obtain initial multi-dimensional feature vectors for each HVAC control area within the building; determines the building's static attributes and temporal variation characteristics for each HVAC control area, and based on these characteristics, corrects the initial multi-dimensional feature vectors for each HVAC control area to obtain target multi-dimensional feature vectors for each HVAC control area; for any given HVAC control area, based on the target multi-dimensional feature vector, determines the parameter control value for the HVAC control area, and controls the operation of HVAC equipment within the HVAC control area according to the parameter control value. This method reduces the frequency of adverse operating conditions and lowers system energy consumption.

Description

Technical Field

[0001] This application belongs to the field of building HVAC control technology, and in particular relates to a control method, equipment, medium and system for a building HVAC system. Background Technology

[0002] In the field of modern intelligent building technology, building heating, ventilation and air conditioning systems are the core infrastructure for ensuring the quality of the indoor environment of buildings and meeting the needs of people living and working. They are widely used in various commercial and civil buildings.

[0003] Most existing building HVAC systems use threshold control, rule control, or simple weighted control based on environmental parameters such as temperature and humidity to regulate the operation of air conditioning, fresh air, and heating equipment.

[0004] However, the existing control methods of building HVAC systems do not make sufficient comprehensive use of multi-source environmental information from indoor and outdoor environments, which limits the system's judgment of the actual environment, resulting in poor operating conditions such as delayed adjustment or frequent start-stop. Summary of the Invention

[0005] This application provides a control method, device, medium, and system for a building HVAC system, which solves the technical problem that the existing control methods of building HVAC systems do not make comprehensive use of multi-source indoor and outdoor environmental information, which limits the system's judgment of the actual environment and leads to adverse operating conditions such as adjustment lag or frequent start-stop.

[0006] In a first aspect, embodiments of this application provide a control method for a building heating, ventilation, and air conditioning system, including:

[0007] Multi-source heterogeneous environmental data inside and outside the building are collected through a distributed sensor network, and the multi-source heterogeneous environmental data are preprocessed to obtain the initial multi-dimensional feature vectors of each HVAC control area in the building.

[0008] Determine the static attributes and temporal variation characteristics of buildings in each HVAC control zone;

[0009] Based on the static attributes and temporal variation characteristics of each HVAC control area, the initial multidimensional feature vectors of each HVAC control area are corrected to obtain the target multidimensional feature vectors corresponding to each HVAC control area; the temporal variation characteristics are used to reflect the periodic patterns of different HVAC control areas in the building over time.

[0010] For any HVAC control zone, based on the target multidimensional feature vector corresponding to the HVAC control zone, the parameter control value of the HVAC control zone is determined, and the operation of the HVAC equipment in the HVAC control zone is controlled according to the parameter control value;

[0011] Based on the target multidimensional feature vector corresponding to the HVAC control area, the parameter control values ​​of the HVAC control area are determined, including:

[0012] For the control parameters of multiple HVAC equipment in the HVAC control area, based on the physical influence relationship between the control parameters and each feature dimension in the target multidimensional feature vector, feature dimensions that have a physical influence relationship with the control parameters are selected from the target multidimensional feature vector as the target feature dimension set corresponding to the control parameters.

[0013] Determine the weights of each target feature dimension in the target feature dimension set corresponding to the parameter to be controlled;

[0014] For any parameter to be controlled, the weighted sum of the weights of the target feature dimensions corresponding to the parameter to be controlled is used to obtain the target control factor corresponding to the parameter to be controlled.

[0015] Based on the target control factor, the benchmark control factor of the parameter to be controlled and its corresponding benchmark parameter value, determine the parameter control value of the parameter to be controlled;

[0016] Among them, the target feature dimension sets corresponding to different parameters to be controlled are different, and the weight polarity of the same feature dimension in different parameters to be controlled is determined according to the physical influence relationship between the corresponding feature dimension and the parameter to be controlled.

[0017] In one possible implementation, the multi-source heterogeneous environmental data is preprocessed to obtain initial multidimensional feature vectors for each HVAC control zone within the building, including:

[0018] Based on historical multi-source heterogeneous environmental data of the HVAC system, the static attributes and temporal change characteristics of buildings in each HVAC control area, the baseline feature thresholds for different feature dimensions in each HVAC control area are determined.

[0019] Based on the baseline feature thresholds of different feature dimensions, the multi-source heterogeneous environmental data are normalized to obtain the initial multi-dimensional feature vectors of each HVAC control area in the building.

[0020] In one possible implementation, based on the building's static attributes and temporal variation characteristics of each HVAC control area, the initial multidimensional feature vector of each HVAC control area is modified to obtain the target multidimensional feature vector corresponding to each HVAC control area, including:

[0021] For any HVAC control zone, the static deviation of the initial multidimensional feature vector of each HVAC control zone is corrected based on the static attributes of the building in the HVAC control zone.

[0022] Based on the temporal variation characteristics of the HVAC control area, the initial multidimensional feature vectors of each HVAC control area are modified in a second step to obtain the target multidimensional feature vectors corresponding to each HVAC control area.

[0023] In one possible implementation, based on the target multidimensional feature vector corresponding to the HVAC control area, the weights of the target feature dimensions corresponding to each parameter to be controlled are determined, including:

[0024] Determine the degree of single-dimensional deviation in different dimensions of the multidimensional feature vector of the target corresponding to the HVAC control area;

[0025] Based on the target multidimensional feature vectors corresponding to each HVAC control zone and the degree of deviation of different dimensions of the target multidimensional feature vectors in the single dimension, with the global energy consumption optimization as the optimization objective, under the preset constraints, the decision model generates the weights of the target feature dimensions corresponding to each parameter to be controlled.

[0026] The preset constraints include at least one of the following:

[0027] Operating boundary constraints for HVAC equipment;

[0028] Interlocking constraints between heating and cooling equipment;

[0029] Zone load limit constraints.

[0030] In one possible implementation, after determining the building static attributes and temporal variation characteristics of each HVAC control zone, the method further includes:

[0031] Based on multi-source heterogeneous environmental data, determine the rate of environmental change for each meteorological dimension outside the building;

[0032] For any meteorological dimension, when the rate of environmental change in the meteorological dimension exceeds the preset rate of change, the corresponding compensation coefficient for the meteorological dimension is determined.

[0033] After correcting the initial multidimensional feature vectors of each HVAC control area based on the building's static attributes and temporal variation characteristics, and obtaining the target multidimensional feature vectors corresponding to each HVAC control area, the following steps are taken:

[0034] Based on the compensation coefficients corresponding to multiple meteorological dimensions, the target multidimensional feature vectors corresponding to each HVAC control area are corrected and updated to obtain new target multidimensional feature vectors corresponding to each HVAC control area.

[0035] In one possible implementation, the multi-source heterogeneous environmental data includes real-time HVAC loads, which characterize the real-time supply and demand status of heating and cooling loads in the HVAC controlled area; the method further includes:

[0036] For any HVAC control zone, the standard HVAC load of the building is determined through a multi-field coupled load model based on multi-source heterogeneous environmental data.

[0037] Determine the load deviation between the real-time HVAC load and the standard HVAC load, and determine the deviation correction factor corresponding to the load deviation; the deviation correction factor is used to correct the measurement error of the real-time HVAC load.

[0038] Based on the deviation correction coefficient and real-time HVAC load, the corrected HVAC load of the HVAC control area is determined; the corrected HVAC load is the load data after error correction by the multi-field coupled load model.

[0039] Among them, the multi-field coupled load model is a mechanism model constructed by coupling multiple types of indoor heat transfer physical fields, which is used to generate the theoretical HVAC load of the HVAC control area under the environment of multi-source heterogeneous environmental data.

[0040] After preprocessing multi-source heterogeneous environmental data to obtain initial multi-dimensional feature vectors for each HVAC control zone within the building, the method also includes:

[0041] Based on the corrected HVAC load, the initial multidimensional feature vector is updated to obtain the updated initial multidimensional feature vector.

[0042] Secondly, embodiments of this application provide a control device for a building heating, ventilation, and air conditioning system, comprising:

[0043] The acquisition module is used to collect multi-source heterogeneous environmental data inside and outside buildings through a distributed sensor network.

[0044] The processing module is used to preprocess multi-source heterogeneous environmental data to obtain the initial multi-dimensional feature vectors of each HVAC control area in the building.

[0045] The processing module is also used to determine the static attributes and temporal variation characteristics of buildings in each HVAC control zone.

[0046] The processing module is also used to correct the initial multidimensional feature vectors of each HVAC control area based on the building's static attributes and temporal variation characteristics, so as to obtain the target multidimensional feature vectors corresponding to each HVAC control area; the temporal variation characteristics are used to reflect the periodic patterns of different HVAC control areas in the building over time.

[0047] The processing module is also used to determine the parameter control value of any HVAC control area based on the target multidimensional feature vector corresponding to the HVAC control area, and control the operation of the HVAC equipment in the HVAC control area according to the parameter control value.

[0048] The processing module is also used to select the feature dimensions that have a physical influence relationship with the parameter to be controlled from the target multidimensional feature vector based on the physical influence relationship between the parameter to be controlled and each feature dimension in the target multidimensional feature vector, and use them as the target feature dimension set corresponding to the parameter to be controlled.

[0049] The processing module is also used to determine the weights of each target feature dimension in the target feature dimension set corresponding to the parameter to be adjusted.

[0050] The processing module is also used to perform a weighted summation of the weights of the target feature dimensions corresponding to any parameter to be controlled, so as to obtain the target control factor corresponding to the parameter to be controlled.

[0051] The processing module is also used to determine the parameter control value of the parameter to be controlled based on the target control factor, the benchmark control factor of the parameter to be controlled and its corresponding benchmark parameter value.

[0052] Among them, the target feature dimension sets corresponding to different parameters to be controlled are different, and the weight polarity of the same feature dimension in different parameters to be controlled is determined according to the physical influence relationship between the corresponding feature dimension and the parameter to be controlled.

[0053] In one possible implementation, the processing module is further configured to determine the baseline feature thresholds for different feature dimensions in each HVAC control area based on historical multi-source heterogeneous environmental data of the HVAC system, building static attributes and temporal change characteristics of each HVAC control area.

[0054] The processing module is also used to normalize multi-source heterogeneous environmental data based on benchmark feature thresholds of different feature dimensions to obtain the initial multi-dimensional feature vectors of each HVAC control area in the building.

[0055] In one possible implementation, the processing module is further configured to, for any HVAC control zone, correct the static deviation of the initial multidimensional feature vector of each HVAC control zone based on the building static attributes of the HVAC control zone.

[0056] The processing module is also used to perform secondary correction processing on the initial multidimensional feature vectors of each HVAC control area based on the temporal variation characteristics of the HVAC control area, so as to obtain the target multidimensional feature vectors corresponding to each HVAC control area.

[0057] In one possible implementation, the processing module is further configured to determine the weights of the target feature dimensions corresponding to each parameter to be controlled based on the target multidimensional feature vector corresponding to the HVAC control area.

[0058] The processing module is also used to perform a weighted summation of the weights of the target feature dimensions corresponding to any parameter to be regulated, so as to obtain the target regulation factor corresponding to the parameter to be regulated.

[0059] The processing module is also used to determine the parameter control value of the parameter to be controlled based on the target control factor, the benchmark control factor of the parameter to be controlled and its corresponding benchmark parameter value.

[0060] In one possible implementation, the processing module is further configured to determine the degree of single-dimensional deviation in different dimensions of the target multidimensional feature vector corresponding to the HVAC control area.

[0061] The processing module is also used to generate the weights of the target feature dimensions corresponding to each parameter to be controlled through a decision model, based on the target multidimensional feature vectors corresponding to each HVAC control zone and the degree of deviation of the single dimension in different dimensions of the target multidimensional feature vectors, with the global energy consumption optimization as the optimization objective, under preset constraints.

[0062] The preset constraints include at least one of the following:

[0063] Operating boundary constraints for HVAC equipment;

[0064] Interlocking constraints between heating and cooling equipment;

[0065] Zone load limit constraints.

[0066] In one possible implementation, the processing module is also used to determine the rate of environmental change for each meteorological dimension outside the building based on multi-source heterogeneous environmental data.

[0067] The processing module is also used to determine the compensation coefficient for any meteorological dimension when the rate of environmental change in the meteorological dimension exceeds the preset rate of change.

[0068] The processing module is also used to correct and update the target multidimensional feature vectors corresponding to each HVAC control area based on the compensation coefficients corresponding to multiple meteorological dimensions, so as to obtain new target multidimensional feature vectors corresponding to each HVAC control area.

[0069] In one possible implementation, the processing module is further configured to determine the standard HVAC load of a building for any given HVAC control zone based on multi-source heterogeneous environmental data and through a multi-field coupled load model.

[0070] The processing module is also used to determine the load deviation between the real-time HVAC load and the standard HVAC load, and to determine the deviation correction factor corresponding to the load deviation; the deviation correction factor is used to correct the measurement error of the real-time HVAC load.

[0071] The processing module is also used to determine the corrected HVAC load of the HVAC control area based on the deviation correction coefficient and the real-time HVAC load; the corrected HVAC load is the load data after error correction by the multi-field coupled load model;

[0072] Among them, the multi-field coupled load model is a mechanism model constructed by coupling multiple types of indoor heat transfer physical fields, which is used to generate the theoretical HVAC load of the HVAC control area under the environment of multi-source heterogeneous environmental data.

[0073] The processing module is also used to update the initial multidimensional feature vector based on the corrected HVAC load, so as to obtain the updated initial multidimensional feature vector.

[0074] Thirdly, embodiments of this application provide a building HVAC system, including: a controller and a distributed sensor network, wherein the distributed sensor network is disposed in multiple HVAC control areas within the building and outside the building; the controller is used to perform the first aspect and / or various possible implementations of the first aspect.

[0075] Fourthly, embodiments of this application provide an electronic device, including: a memory and a processor;

[0076] The memory stores the instructions that the computer executes;

[0077] The processor executes computer execution instructions stored in memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.

[0078] Fifthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.

[0079] In a sixth aspect, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.

[0080] The control method for building HVAC systems provided in this application collects multi-source heterogeneous environmental data inside and outside the building through a distributed sensor network, forming initial multi-dimensional feature vectors for each HVAC control area. This improves the completeness of the building HVAC system's perception of the building's environmental status and the sufficiency of information utilization. Then, by combining the building's static attributes and temporal change characteristics of each HVAC control area, the initial multi-dimensional feature vectors are corrected to obtain target multi-dimensional feature vectors. Based on these target multi-dimensional feature vectors, the parameters of each HVAC control area are determined, and the corresponding HVAC equipment is controlled. This method improves the globality, adaptability, and coordination of HVAC system control, taking into account energy consumption control, environmental comfort, and equipment operational stability. It effectively reduces the frequency of adverse operating conditions and improves the user experience. Attached Figure Description

[0081] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0082] Figure 1 This is a flowchart illustrating the control method for the building HVAC system provided in this application. Figure 1 ;

[0083] Figure 2 This is a flowchart illustrating the control method for the building HVAC system provided in this application. Figure 2 ;

[0084] Figure 3 This is a structural schematic diagram of the control device for the building HVAC system provided in this application;

[0085] Figure 4 This is a schematic diagram of the structure of the electronic device provided in this application.

[0086] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0087] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0088] The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented, for example, in orders other than those illustrated or described herein.

[0089] In this application, the terms "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0090] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with relevant laws, regulations and standards, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0091] First, let me explain the terms used in this application:

[0092] Rule Model: A static reasoning model built on preset industry logic, business constraints, threshold judgments and fixed judgment logic, relying on manually defined judgment conditions, priorities and execution processes to complete the result output.

[0093] Weighted model: Differentiated weight coefficients are assigned to multi-dimensional input features. The feature values ​​and corresponding weights are weighted and summed to integrate the influence of various indicators and obtain the final output result. The weights can be fixed or dynamically adjusted.

[0094] Lookup table model: The scenario conditions, input parameters and corresponding standard results are mapped and stored in advance to form a data table. During actual operation, the preset values ​​in the table are matched directly with the real-time input index to complete the fast value retrieval response.

[0095] Regression model: Based on massive historical sample data, machine learning is used to train and fit the continuous numerical mapping relationship between input and output, automatically mining the inherent correlation patterns in the data, and realizing accurate prediction, numerical fitting and trend inference of continuous quantities.

[0096] Decision model: It is a complete logical operation system used for data analysis, logical judgment, numerical calculation, and finally automatically outputting the optimal control scheme and operating parameters.

[0097] Linear programming model: A mathematical decision-making model that solves for the optimal extremum of a linear objective function under a set of linear equality and inequality constraints. It is often used in optimal allocation decision-making scenarios such as resource allocation, quota scheduling, and load allocation.

[0098] Quadratic programming model: A convex optimization decision-making model with a quadratic objective function and linear constraints. It can balance the optimization of benefits and losses, deviations and costs, and is suitable for precise optimal decision-making with fluctuating deviations.

[0099] Heuristic optimization model: Relying on empirical strategies and intelligent optimization algorithms, iteratively searches to find near-optimal feasible solutions that meet actual engineering needs within limited computing power and time constraints, and is suitable for complex nonlinear and multi-constraint scenarios.

[0100] In the field of modern intelligent building technology, building heating, ventilation and air conditioning (HVAC) systems are the core infrastructure for ensuring the quality of the indoor environment of buildings and meeting the needs of people living and working. They are widely used in building environments with multiple areas, multiple populations, and significant load changes, such as office buildings, commercial complexes, hospitals, and schools. Their core task is to coordinate and regulate HVAC equipment such as air conditioning, fresh air, and heating to maintain indoor temperature, humidity, and air quality at a suitable level, while also taking into account energy consumption control and operational stability.

[0101] In the above-mentioned usage scenarios, building HVAC systems typically cover multiple HVAC control areas within a building. These areas vary significantly in terms of function, population density, usage time, and the degree of influence from external weather conditions. To achieve environmental perception and equipment control, the system generally relies on sensors, controllers, and actuators installed inside and outside the building to form a basic operating architecture.

[0102] Most existing building HVAC systems use threshold control, rule control, or simple weighted control based on environmental parameters such as temperature and humidity to regulate the operation of air conditioning, fresh air, and heating equipment.

[0103] For example, when a certain detection value exceeds a set threshold, the device is triggered to start or stop, or parameters such as air conditioning supply, fresh air supply, heating and water supply are switched according to preset rules. Some systems also generate control results by assigning fixed weights to a small number of parameters such as temperature and humidity.

[0104] However, the existing control methods of building HVAC systems do not make sufficient comprehensive use of multi-source environmental information from indoor and outdoor environments, which limits the system's judgment of the actual environment. As a result, adverse operating conditions such as response lag, local overcooling and overheating, air quality deterioration, and frequent equipment start-ups and shutdowns are likely to occur.

[0105] In addition, existing solutions often fail to make full use of the inherent attributes and time-varying patterns of buildings, such as differences in building orientation, building envelope, and periodic load changes between weekdays and non-weekdays. This makes it difficult for control strategies to adapt to complex dynamic environments and to achieve a balance between comfort, energy efficiency, and operational coordination.

[0106] To address the aforementioned problems, this application provides a control method for a building HVAC system. This method collects multi-source heterogeneous environmental data from inside and outside the building using a distributed sensor network, and preprocesses the collected data to obtain initial multi-dimensional feature vectors for each HVAC control area. It then determines the building's static attributes and temporal variation characteristics for each HVAC control area, and corrects the initial multi-dimensional feature vectors based on these characteristics to obtain the target multi-dimensional feature vector for the corresponding area. For any given HVAC control area, it determines parameter control values ​​based on the corresponding target multi-dimensional feature vector and controls the operation of HVAC equipment according to these parameter control values. This method relies on an operational architecture consisting of a distributed sensor network, controllers, and HVAC actuators, incorporating multi-source environmental information, inherent regional attributes, and temporal variation patterns into a unified analysis process. This transforms the adjustment basis of HVAC equipment from single-parameter judgment to control decisions oriented towards the comprehensive state of the area, thereby solving problems such as lag in adjustment, insufficient adaptability, and uncoordinated regional control in existing systems, effectively reducing the frequency of adverse operating conditions.

[0107] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0108] Figure 1 A flowchart illustrating the control method for a building HVAC system provided in this application embodiment. Figure 1 The executing entity in this embodiment can be, for example, the controller of a building HVAC system. Figure 1 As shown, the control method for a building HVAC system provided in this embodiment includes:

[0109] S101: Collect multi-source heterogeneous environmental data inside and outside the building through a distributed sensor network, and preprocess the multi-source heterogeneous environmental data to obtain the initial multi-dimensional feature vectors of each HVAC control area in the building.

[0110] In this embodiment, the distributed sensor network is used to collect environmental data inside and outside the building. Its function is to provide a real-time and continuous data source for subsequent feature vector construction. For example, the distributed sensor network can consist of multiple sensor nodes deployed in office areas, conference rooms, corridors, wards, waiting areas, equipment rooms, building facades, and rooftop meteorological monitoring points. Each sensor node collects temperature, humidity, and other data. Multi-source heterogeneous environmental data, such as concentration, particulate matter concentration, personnel density, door and window opening status, outdoor temperature, outdoor humidity, wind speed, wind direction, solar irradiance, and equipment operating conditions, are uploaded to the control side via wired bus, industrial Ethernet, wireless LAN, low-power wireless communication link, or building automation communication protocol.

[0111] Understandably, multi-source heterogeneous environmental data is used as raw input information to characterize the state of the building's internal and external environment. This multi-source heterogeneous environmental data can be, for example, continuous numerical data, or state data such as switch quantities, alarm status quantities, and operating levels.

[0112] The initial multidimensional feature vector is used to characterize the relative offset of the initial environmental state of each HVAC control area relative to the dynamically determined comfort space environment. Essentially, it is the result of unified quantification and standardization of multiple environmental variables in the same HVAC control area within the same sampling period, so as to ensure the temporal synchronization and consistent operating conditions of various environmental data, and provide the basic input for subsequent correction processing and control parameter generation.

[0113] The building HVAC system collects multi-source heterogeneous environmental data inside and outside the building through a distributed sensor network, and performs outlier removal, missing value imputation, noise filtering, regional fusion, and normalization on the multi-source heterogeneous environmental data. After completing the preprocessing, the building HVAC system maps data with different dimensions and different value ranges to a standardized data space to form an initial multidimensional feature vector, which is the initial multidimensional feature vector of each HVAC control area in the building.

[0114] In one possible implementation, the building HVAC system can divide the building into HVAC control zones according to floors, orientation, functional zoning, and the coverage of terminal equipment. Each HVAC control zone can correspond to a room, an open office area, or a joint control unit consisting of multiple adjacent spaces.

[0115] Within each control cycle of a building HVAC system, the distributed sensor network performs data acquisition at a preset sampling frequency. For example, environmental parameters are sampled at a cycle of 30 seconds to 5 minutes, equipment operating parameters are sampled at a cycle of 5 seconds to 1 minute, and outdoor meteorological parameters are sampled at a cycle of 1 minute to 10 minutes. After receiving the data uploaded by each node, the controller first completes the aggregation and alignment according to the area identifier, timestamp, and data type to eliminate the time deviation caused by different sampling cycles.

[0116] In one possible implementation, for data that did not arrive at the same time, the controller can use nearest neighbor time alignment, linear interpolation, or sliding window averaging to form a regional dataset with a unified time, thereby ensuring that subsequent operations are based on the same analysis time point.

[0117] In one possible implementation, the preprocessing of multi-source heterogeneous environmental data mainly involves outlier removal, missing value imputation, noise filtering, region-based fusion, and normalization.

[0118] Outlier removal can be performed simultaneously based on physical boundary constraints and statistical criteria. For example, outlier removal can be performed when the indoor temperature exceeds 0°C to 50°C or the relative humidity exceeds 0% to 100%. When the concentration shows obvious pulse jumps and the deviation from the adjacent time exceeds the preset threshold, the corresponding data is marked as abnormal; missing value filling can be achieved by forward hold, compensation by the average of historical similar time periods, or collaborative estimation by adjacent sensors; noise filtering can suppress short-term jitter through moving average, median filtering, or low-pass filtering.

[0119] When multiple similar sensors are installed in the same HVAC control area, weights can be set according to the reliability of the installation location, the health status of the sensor, and the spatial representativeness, and the similar data can be weighted and averaged to obtain the regional characteristic value.

[0120] For example, temperature sensors near the air supply vent and temperature sensors near the air return vent in the conference room can be assigned different weights to reduce the impact of local airflow disturbances on the overall area assessment.

[0121] In one possible implementation, the dimensions of the initial multidimensional feature vector may include regional temperature deviation, regional humidity deviation, and regional... Concentration levels, regional population density index, outdoor meteorological influencing factors, equipment operating load index, etc.

[0122] For example, if X represents the original characteristics of a certain region, and Xmin and Xmax represent the lower limit and upper limit of the characteristic, respectively, then the standardized value x′ = (X - Xmin) / (Xmax - Xmin) can be obtained by linear normalization. For parameters with a comfort target center value, such as temperature and humidity, it can be further converted into relative target deviation to more directly reflect the intensity of regulation demand.

[0123] Understandably, by transforming raw sensing data that is scattered in origin, complex in type, and inconsistent in time scale into a computable unified regional state expression, it is possible to avoid local judgment biases caused by relying solely on a single temperature or humidity parameter. This allows each HVAC control zone to correspond to an initial multidimensional feature vector with a unified dimension, clear meaning, and direct participation in subsequent correction calculations at the current moment. This provides a data foundation for subsequent comprehensive consideration of the building's inherent attributes and time-varying patterns, thereby improving the globality and response accuracy of control decisions.

[0124] S102: Determine the static attributes and temporal variation characteristics of buildings in each HVAC control zone.

[0125] S103: Based on the static building attributes and temporal change characteristics of each HVAC control area, the initial multidimensional feature vectors of each HVAC control area are corrected to obtain the target multidimensional feature vectors corresponding to each HVAC control area.

[0126] Among them, the static attributes of the building are the inherent attribute data of different HVAC control areas in the building; the temporal variation characteristics are used to reflect the periodic patterns of the changes of different HVAC control areas in the building over time.

[0127] Understandably, building static attributes are technical information used to describe the fixed physical or structural characteristics of a building. Their role is to provide a basis for long-term or gradual correction of feature vectors. These building static attributes may include, for example, building orientation, window area ratio, heat transfer coefficient of building envelope, floor location, distribution of nearby heat source equipment, functional zoning type, space volume, floor height, upper limit of personnel capacity, and HVAC terminal type.

[0128] The role of time-series change characteristics is to provide a time dimension basis for dynamic correction; specifically, the time-series change characteristics can be the result of periodic analysis of historical operating data, or they can be generated by combining building schedule management system, access control system, reservation system or business scheduling information, reflecting the differences between working days and non-working days, peak and trough occupancy during the day, seasonal changes in hot and cold loads, and regular usage characteristics of specific functional areas.

[0129] The target multidimensional feature vector is a modified result that integrates the static attributes and temporal change characteristics of the building. It is used as a direct input for subsequent control decisions and is closer to the actual load state and control requirements of the area than the initial multidimensional feature vector.

[0130] The controller establishes an attribute database and a historical time-series feature database for the smart building where the HVAC system is deployed, thereby obtaining the corresponding static attributes and time-series change characteristics of the building.

[0131] For the static attributes of the building, the fixed attributes of each HVAC control area are encoded as static correction factors. In addition, the time-related data in the multi-source heterogeneous data are matched with the temporal change characteristics of different HVAC control areas to obtain dynamic correction factors. The initial multidimensional feature vectors of each HVAC control area are corrected by using the static correction factors and dynamic correction factors to obtain the target multidimensional feature vectors corresponding to each HVAC control area.

[0132] In one possible implementation, the regional attribute database can be generated from architectural design drawings, energy consumption modeling parameters, equipment ledgers, on-site survey results, and system debugging records, and written into the system deployment phase; the historical time series feature database can be formed by statistical learning of data from each region over the past few weeks, months, or seasons.

[0133] For example, if the static attributes of a building include "south-facing area, top floor area, glass curtain wall area, and conference room", then the south-facing area may have a higher solar heat gain coefficient in the summer afternoon, the top floor area may have a higher external temperature coupling coefficient, the glass curtain wall area may have a larger building envelope fluctuation influence coefficient, and the conference room and waiting area may have a higher instantaneous crowd gathering sensitivity coefficient. In this way, the static correction factor corresponding to different HVAC control areas can be determined.

[0134] In one possible implementation, the static correction factor can be obtained through empirical setting, historical calibration, or simulation verification, and mapped to the corresponding dimension of the initial multidimensional feature vector to perform the first layer of correction for temperature, humidity, air quality, and load trends.

[0135] For example, when both areas detect an increase in room temperature, the actual control implications are different for the western-facing area and the inner office area. The former indicates a continuous temperature rise caused by short-term solar heat gain, while the latter indicates that the temperature rise is more related to heat dissipation from people and equipment. Therefore, the controller will assign different static deviation correction coefficients to the two areas to make the correction results consistent with the actual thermal behavior of the area.

[0136] In one possible implementation, after the initial multidimensional feature vector is corrected for static attributes, the controller further performs a secondary correction based on the temporal variation features.

[0137] Among these, temporal variation characteristics can be extracted through time series decomposition, periodic clustering, sliding window statistics, or rule-based schedule mapping; for example, for office areas, the personnel density from 9:00 AM to 11:00 AM and from 2:00 PM to 5:00 PM on weekdays can be obtained based on historical data. Concentration is usually higher than at other times; for conference rooms, the increase in occupancy during a certain period can be inferred from meeting reservation records and historical usage habits; for hospital waiting areas, peak patterns that change over time can be formed based on outpatient schedules; for school classrooms, the switching characteristics between classes and breaks can be formed based on the timetable.

[0138] The controller matches the current time, date type, and seasonal information with the temporal change characteristics of the corresponding region to obtain the corresponding dynamic correction factor, and then uses this dynamic correction factor to perform a secondary correction on the feature vector that has already undergone static correction.

[0139] For example, the target multidimensional feature vector can be calculated using the following formula 1:

[0140]

[0141] in, Represents the target's multidimensional feature vector. Represents the initial multidimensional feature vector. This represents the correction amount after mapping the static attributes of the building. This represents the correction amount after mapping the time-series change features. , , These represent the weighting coefficients for the original state, static attributes, and time patterns, respectively. The symbol ⊙ indicates multiplication by dimension or weighted operation by dimension.

[0142] It is understandable that Formula 1 integrates the current state, the inherent attributes of the region, and the short-term future trend into the same expression system, so that the target multidimensional feature vector can simultaneously reflect the immediate state and potential changing trends.

[0143] In one possible implementation, the controller can also iteratively update the static correction factor and the dynamic correction factor based on historical control effects.

[0144] For example, when a region experiences temperature control lag, frequent overshooting, or slow air quality recovery for several consecutive days, the building HVAC system can retrospectively analyze the target multidimensional feature vector and the actual feedback error for the corresponding period, and adaptively adjust the relevant correction coefficients to improve the fit of subsequent prediction and control.

[0145] Understandably, by first compensating for static deviations by combining static attributes such as building orientation, building envelope, and functional zoning, and then making dynamic trend corrections by combining temporal change characteristics such as personnel work and rest, workday cycle, and seasonal fluctuations, the target multidimensional feature vector no longer remains on the surface state of the current sensor sampling value. It can more accurately reflect the comprehensive load and change direction of different HVAC control areas, thereby reducing the control lag of HVAC equipment, reducing the risk of mismatch between HVAC control areas due to attribute differences, and providing a unified and reliable decision-making basis for subsequent equipment coordinated control.

[0146] S104: For any HVAC control zone, based on the target multidimensional feature vector corresponding to the HVAC control zone, determine the parameter control value of the HVAC control zone, and control the operation of the HVAC equipment in the HVAC control zone according to the parameter control value.

[0147] Understandably, parameter control values ​​are the actual adjustment amounts used to control the operation of HVAC equipment. Their function is to convert the target multi-dimensional feature vector into control commands at the equipment execution level. These parameter control values ​​can be, for example, air conditioning supply air temperature setpoint, supply air volume setpoint, fresh air valve opening, water valve opening, fan speed, cold and heat source supply water temperature correction, underfloor heating manifold branch opening, return air ratio, or dehumidification operation intensity, etc.

[0148] Heating, ventilation, and air conditioning (HVAC) equipment is a collection of devices that perform environmental control, which may include combined air conditioning units, fresh air handling units, fan coil units, variable air volume terminals, hot and cold water valves, electric air valves, circulating water pumps, heat exchangers, boilers, chillers, and actuators and controllers that work with them.

[0149] The target multidimensional feature vector, as an input that has integrated the region's instantaneous state, static attributes, and temporal patterns, can reflect the actual adjustment needs of the current HVAC control area in terms of comfort, air quality, and energy consumption by the magnitude and interrelationships of its various dimensions.

[0150] In this embodiment of the application, the building HVAC system sets a set of parameters to be controlled for each HVAC control area and establishes a mapping relationship between the target multidimensional feature vector and the control values ​​of each parameter; wherein, the mapping relationship can be realized by a rule model, a weighted model, a lookup table model or a trained regression model.

[0151] Based on the mapping relationship between the target multidimensional feature vector and the parameter control values, the parameter control values ​​corresponding to different target multidimensional feature vectors are determined for different HVAC control areas acquired in the current transaction.

[0152] Based on the mapping relationship between the target multidimensional feature vector and the parameter control values, the parameter control values ​​corresponding to different target multidimensional feature vectors are determined for different HVAC control areas acquired in the current transaction. After obtaining the parameter control values, the controller sends control commands to the corresponding execution devices according to the device layer communication protocol and receives execution feedback.

[0153] For example, if the target area is equipped with fan coil units and fresh air terminals, the controller can simultaneously issue fan speed settings, fresh air valve opening settings, and water valve opening settings; if the area is connected to a centralized air conditioning system, the controller can also upload the area-level control values ​​to the building group control platform, which will coordinate the unit supply air temperature, total fresh air volume, and cold and heat source output distribution to avoid conflicts between local control and global control.

[0154] In one possible implementation, after the control command is issued, the system continues to read sensor feedback, compare the environmental changes before and after execution with the expected deviation, and when it is found that the equipment does not act according to the command, the regional state is not improved enough, or crosstalk occurs between adjacent regions, the parameter correction or command retransmission mechanism in the next control cycle can be triggered, thereby forming a closed-loop regulation process.

[0155] Understandably, by converting the target multidimensional feature vector into executable parameter control values ​​and directly driving the coordinated operation of HVAC equipment such as air conditioning, fresh air, and heating, a complete closed-loop link from regional comprehensive status identification to equipment execution control is realized. This enables the control results to simultaneously take into account temperature and humidity comfort, air quality, and equipment synergy, reducing the problems of frequent start-stop, cooling and heating offsetting, and mutual restraint between fresh air and temperature control under traditional threshold control, thereby improving system operation stability and energy-saving effect.

[0156] The control method for a building HVAC system provided in this embodiment collects multi-source heterogeneous environmental data inside and outside the building through a distributed sensor network, and preprocesses the multi-source heterogeneous environmental data to obtain initial multi-dimensional feature vectors for each HVAC control area within the building; determines the static attributes and temporal variation characteristics of each HVAC control area, and corrects the initial multi-dimensional feature vectors of each HVAC control area based on the static attributes and temporal variation characteristics of each HVAC control area to obtain the target multi-dimensional feature vectors corresponding to each HVAC control area; for any HVAC control area, determines the parameter control value of the HVAC control area based on the target multi-dimensional feature vector corresponding to the HVAC control area, and controls the operation of the HVAC equipment in the HVAC control area according to the parameter control value. This method incorporates multi-source heterogeneous environmental data, building static attributes, and temporal variation characteristics into a unified analysis framework. This transforms the control basis of HVAC equipment from a single threshold or a few fixed rules into a dynamic decision-making process oriented towards the overall regional state. As a result, it can more accurately identify the actual load characteristics and evolution trends of different HVAC control areas, reduce control response lag, lower the probability of local overcooling, overheating, and air quality degradation, suppress frequent equipment start-ups and shutdowns, and the phenomenon of mutual cancellation of heat and cold. This improves the comprehensive balance between comfort, energy efficiency, and operational coordination of HVAC systems in multi-regional building environments.

[0157] Figure 2 A flowchart illustrating the control method for a building HVAC system provided in this application embodiment. Figure 2 .like Figure 2 As shown, in this embodiment... Figure 1 Based on the embodiments, the control method of the building HVAC system is described in detail. The control method of the building HVAC system shown in this embodiment includes:

[0158] S201: Collect multi-source heterogeneous environmental data inside and outside buildings through a distributed sensor network.

[0159] Step S201 is similar to step S101 above, and will not be repeated here.

[0160] S202: Based on historical multi-source heterogeneous environmental data of the HVAC system, the static attributes of buildings in each HVAC control area and the temporal change characteristics, determine the baseline feature thresholds for different feature dimensions in each HVAC control area.

[0161] S203: Based on the baseline feature thresholds of different feature dimensions, normalize the multi-source heterogeneous environmental data to obtain the initial multi-dimensional feature vectors of each HVAC control area in the building.

[0162] Understandably, the baseline feature threshold is used to characterize the reference range of each feature dimension and serve as a unified benchmark for subsequent normalization processing to eliminate the dimensional differences between different data dimensions such as temperature, humidity, carbon dioxide concentration, and personnel density.

[0163] Historical multi-source heterogeneous environmental data includes sampling values ​​of the building's internal and external environment, equipment operating status, and regional load-related records recorded during the historical operation of the building's HVAC system. The building's static attributes include area, orientation, thermal parameters of the building envelope, and functional uses. The temporal variation characteristics are used to reflect the periodic load differences on weekdays, non-working days, and different time periods.

[0164] Normalization is a processing method that maps data from various feature dimensions to a uniform scale. The resulting initial multidimensional feature vector can be composed of multiple dimensions such as temperature, humidity, air quality, and operating status, so as to facilitate subsequent feature fusion and control calculations.

[0165] Historical multi-source heterogeneous environmental data is cleaned to remove outlier and missing values. A regionalized statistical model is established by combining the static attributes of each HVAC control area. Then, time-series variation features are extracted based on the load distribution patterns of different time periods. For each feature dimension, the statistical quantile, fluctuation range, and comfort constraint boundary in the target area are calculated and determined as the corresponding benchmark feature threshold.

[0166] After determining the baseline feature thresholds, linear or interval mapping is performed on the currently collected multi-source heterogeneous environmental data according to the baseline feature thresholds corresponding to each dimension, so that data of different dimensions are uniformly mapped to the preset interval, thereby generating the initial multi-dimensional feature vectors of each HVAC control area.

[0167] In this embodiment, the server, edge computing gateway, or building controller that implements the normalization process can all use existing general-purpose computing hardware. In practical applications, other models of this component can also be selected, and this embodiment does not impose any special restrictions on this.

[0168] For example, for the temperature dimension, the lower and upper limits of the comfort zone can be used as threshold references; for the humidity and air quality dimensions, the safe threshold range can be determined based on historical operating data and regional usage.

[0169] In this embodiment, a specific comfort range (e.g., 20°C to 26°C) is established for each HVAC zone, replacing the single threshold trigger point in traditional environmental control; by mapping physical measurement values ​​to relative position values ​​within the range [0,1] (e.g., 25°C is mapped to 0.83), the control signal is transformed from a discrete "on / off" state to a continuous control adjustment quantity; at the same time, the obtained relative position values ​​are double-corrected using static correction factors and time-series dynamic correction factors.

[0170] Because the control basis changes from "a point" to "a range", and the output changes from "0 or 1" to "any value between 0 and 1", the temperature fluctuations near the critical point no longer trigger a hard switch between "on" and "off". The equipment always operates in a continuous adjustment mode rather than a start-stop mode, thereby reducing frequent start-stops from the root.

[0171] Understandably, normalizing the initial multidimensional feature vector enables environmental data from different areas of a building to be expressed under a unified benchmark, avoiding feature distortion caused by different units or excessive regional differences, and providing stable input for subsequent parameter control based on the target multidimensional feature vector. Since the determination of the benchmark feature threshold takes into account historical operating data, inherent building attributes, and time variation patterns, the normalization result can more accurately reflect the current real state of each area, thereby improving the adaptability of HVAC control to complex dynamic environments and reducing energy consumption fluctuations caused by local over-adjustment and frequent equipment start-ups and shutdowns.

[0172] S204: Determine the static attributes and temporal variation characteristics of buildings in each HVAC control zone.

[0173] Step S204 is similar to step S102 above, and will not be described again here.

[0174] S205: For any HVAC control zone, based on the building static attributes of the HVAC control zone, the static deviation of the initial multidimensional feature vector of each HVAC control zone is corrected.

[0175] S206: Based on the temporal variation characteristics of the HVAC control area, the initial multidimensional feature vectors of each HVAC control area are modified in a second step to obtain the target multidimensional feature vectors corresponding to each HVAC control area.

[0176] Understandably, building static attributes are used to characterize the inherent conditions of a building that remain unchanged or change very little over a long period of time. These can include building orientation, floor location, thermal resistance of the building envelope, window area, shading conditions, and regional functional attributes. Static deviations are used to describe the long-term deviation of environmental characteristics caused by the above-mentioned inherent conditions. For example, the temperature deviation caused by sunlight in south-facing areas, or the temperature deviation caused by strong heat transfer in areas near exterior walls. Temporal variation characteristics are used to characterize the periodic patterns of changes in the same HVAC control area over time, reflecting the differences in environmental load between weekdays and non-weekdays, daytime and nighttime, and peak and off-peak periods of personnel entry and exit.

[0177] In this embodiment, when correcting the initial multidimensional feature vector using the static attributes and temporal variation characteristics of the building, a static attribute mapping relationship for each HVAC control area can be established based on historical operating data. The building orientation, envelope parameters, and area function type are then mapped to static correction coefficients (i.e., static correction factors). These correction coefficients are then superimposed on dimensions such as temperature, humidity, carbon dioxide concentration, or fresh air demand in the initial multidimensional feature vector to form a corrected initial multidimensional feature vector. Next, combined with the load fluctuation patterns, personnel arrival and departure patterns, and outdoor weather time-period variation patterns obtained by time-slice statistics (i.e., corresponding dynamic correction factors), the corrected initial multidimensional feature vector is dynamically compensated again, enabling the target multidimensional feature vector to simultaneously reflect static structural differences and temporal variation trends.

[0178] In one possible implementation, static correction and temporal correction can be calculated by the controller according to preset weights, or they can be completed by a parameter model trained from historical samples. In practical applications, the model can also be selected in other forms, and this application does not impose any special restrictions on this.

[0179] Understandably, by superimposing the static attributes of the building with dynamic corrections brought about by temporal changes, the target multidimensional feature vector is made closer to the actual load state of the area, thereby improving the accuracy of subsequent parameter control value calculations. Since the orientation, building envelope, and usage rhythm of different areas are all included in the unified correction process, the control system can more accurately distinguish the differences between areas, reduce the overcooling, overheating, and air quality fluctuations caused by single threshold judgments, and reduce energy consumption losses and operational conflicts caused by frequent equipment start-ups and shutdowns.

[0180] S207: For the controllable parameters corresponding to multiple HVAC equipment in the HVAC control area, based on the physical influence relationship between the controllable parameters and each feature dimension in the target multidimensional feature vector, select the feature dimensions that have a physical influence relationship with the controllable parameters from the target multidimensional feature vector, and use them as the target feature dimension set corresponding to the controllable parameters.

[0181] S208: Determine the weights of each target feature dimension in the target feature dimension set corresponding to the parameter to be controlled.

[0182] S209: For any parameter to be controlled, sum the weights of the target feature dimensions corresponding to the parameter to be controlled to obtain the target control factor corresponding to the parameter to be controlled.

[0183] S210: Determine the parameter control value of the parameter to be controlled based on the target control factor, the benchmark control factor of the parameter to be controlled, and its corresponding benchmark parameter value.

[0184] S211: Control the operation of HVAC equipment in the HVAC control zone according to the parameter adjustment values.

[0185] Among them, the parameters to be controlled are used to receive the calculation results of the target multidimensional feature vector and convert them into actual control commands; there are multiple parameters that need to be controlled for the HVAC equipment (such as air conditioning units) in the HVAC control area, such as the supply air temperature, fresh air volume, return air valve opening degree or water supply temperature of the corresponding HVAC equipment (such as air conditioner).

[0186] The weights of the target feature dimensions are used to characterize the differences in the degree of influence of each feature dimension on the control parameters, and can map information such as regional heat load, population density, outdoor weather and equipment operating conditions into calculable control correlation quantities.

[0187] The target regulation factor is used to summarize the weights of multidimensional features and form a single regulation coefficient, serving as an intermediate bridge for calculating the parameter regulation value.

[0188] The baseline control factor is used to represent the reference value of the default adjustment intensity under standard conditions, and the baseline parameter value is used to represent the default value under the standard operating conditions of the equipment. Together, they form the basis for the calculation of the final control quantity.

[0189] The physical influence relationship is determined based on the physical heat and mass transfer mechanism of the building HVAC system, reflecting the objective influence relationship between each characteristic dimension and the parameter to be controlled; among which, the physical influence relationship includes direct influence relationship and indirect influence relationship.

[0190] Specifically, firstly, a direct influence relationship refers to the feature dimension directly affecting the parameter to be controlled; for example, the physical influence relationship corresponding to the fresh air volume parameter is: fresh air volume demand and... Concentration deviation is positively correlated (if) Increased concentration necessitates increased fresh air volume; it is positively correlated with population density index (increased population density requires increased fresh air volume); and negatively correlated with outdoor air enthalpy (increased outdoor air enthalpy necessitates reducing the introduction of high-temperature, high-humidity fresh air to avoid increasing the cooling and dehumidification burden). Therefore, the target feature dimension set corresponding to this fresh air volume parameter includes: Concentration deviation, population density index, and outdoor air enthalpy.

[0191] The physical influence relationships corresponding to the supply air temperature parameter are as follows: the supply air temperature demand is positively correlated with the indoor temperature deviation (if the indoor temperature in the corresponding HVAC control area increases, the supply air temperature needs to be reduced to increase cooling), positively correlated with the indoor humidity deviation (if the humidity in the corresponding HVAC control area increases, the supply air temperature needs to be reduced to increase dehumidification), positively correlated with the thermal impact correction of the building envelope (if the heat transfer of the building envelope in the static attributes of the corresponding HVAC control area is enhanced, the supply air temperature needs to be reduced), and positively correlated with the outdoor meteorological load index (if the outdoor load increases, the supply air temperature needs to be reduced). Therefore, the target feature dimension set corresponding to this supply air temperature parameter includes: indoor temperature deviation, humidity deviation, thermal impact correction of the building envelope, and outdoor meteorological load index.

[0192] The physical influence relationship of the return air valve opening parameter is as follows: the required return air valve opening is negatively correlated with the indoor temperature deviation (if the temperature rises, the return air valve opening needs to be reduced to increase the proportion of fresh air), and... Concentration deviation is negatively correlated (if) If the concentration increases, the return air valve opening needs to be reduced to increase the proportion of fresh air. Therefore, the target feature dimension set corresponding to the return air valve opening parameter includes: indoor temperature deviation, Concentration deviation.

[0193] Secondly, indirect influence refers to the fact that the feature dimension indirectly affects the corresponding parameter to be controlled by influencing other dimensions in the target multidimensional feature vector.

[0194] For example, outdoor temperature indirectly affects the supply air temperature parameter by influencing the thermal effect correction of the building envelope (if the outdoor temperature rises, the heat transfer of the corresponding building envelope increases, bringing additional heat load to the corresponding HVAC control area, resulting in an increase in the thermal effect correction of the building envelope, requiring more cooling, and at this time, the supply air temperature control factor increases); outdoor humidity indirectly affects the fresh air volume parameter by influencing the outdoor air enthalpy (if the outdoor humidity rises, the corresponding outdoor air enthalpy increases, thereby reducing the fresh air volume control factor).

[0195] It is understandable that both direct and indirect influences are determined based on the physical heat and mass transfer mechanisms of HVAC systems, and together they constitute the physical influence relationship between the parameters to be controlled and the characteristic dimensions.

[0196] In this embodiment of the application, when adjusting the parameters of multiple HVAC devices in the HVAC control area, for any HVAC device, there are multiple parameters to be determined and adjusted, and different parameters to be adjusted are independently matched with exclusive associated feature dimensions and weight systems.

[0197] In other words, different parameters to be controlled correspond to different control objectives, and each is independently matched with its corresponding associated feature dimension; the target feature dimension sets corresponding to different parameters to be controlled are different from each other, and the weight polarity of the same feature dimension in different parameters to be controlled is determined according to the physical influence relationship between the feature dimension and the parameter to be controlled.

[0198] Because the weight of the same feature dimension in the set of target feature dimensions corresponding to different parameters to be controlled is determined according to the physical influence relationship between the feature dimension and the parameter to be controlled, the same feature dimension corresponds to different control directions in the calculation of different parameters to be controlled.

[0199] For example, the indoor temperature deviation value in the target feature dimension set has a positive impact on the supply air temperature control factor (if the temperature rises, the supply air temperature needs to be reduced to increase cooling), but may have a negative impact on the fresh air volume control factor (if the temperature rises, the introduction of high-temperature fresh air should be reduced to avoid increasing the cooling burden), reflecting the differentiated response of different parameters to the same feature dimension.

[0200] After determining the target multidimensional feature vector for each HVAC control zone, the physical influence relationship between any parameter to be controlled and each feature dimension in the target multidimensional feature vector is used to select the feature dimension that is strongly correlated with the current parameter to be controlled, and this feature dimension is determined as the target feature dimension corresponding to the parameter to be controlled. Then, the corresponding preset weight is assigned, and the target control factor is obtained by weighted summation. Finally, the conversion is completed by combining the corresponding benchmark quantity (i.e., benchmark parameter value), and the final usable parameter control value is output. This parameter control value is then sent to the corresponding execution device (HVAC equipment) to control the adjustment of working parameters such as the opening degree of the corresponding HVAC equipment's air valve, supply air temperature, fan frequency, or valve opening.

[0201] Understandably, the target multidimensional feature vector is first converted into a weight set corresponding to each parameter to be controlled, and then the target control factor is formed by weighted summation of the differentiated feature dimensions. Finally, the target control factor is generated by combining the benchmark control factor and the benchmark parameter value. Since the parameter control value directly reflects the deviation between the regional comprehensive state and the equipment benchmark state, it can keep the control command consistent with the actual load of the region and reduce the lag and misadjustment caused by single threshold control.

[0202] In this embodiment, by collecting data from each HVAC control zone within the building and generating target multidimensional feature vectors for each zone, the state of each zone is uniformly expressed in the same high-dimensional feature space. This allows the building's HVAC system to assess the usage requirements of each HVAC control zone from a global perspective, avoiding cross-zone energy cancellation due to cooling in zone A and heating in zone B. Simultaneously, since each zone and parameter makes coordinated decisions based on the same data rather than independent decisions, different parameters to be controlled (such as fresh air volume and supply air temperature) independently calculate control factors based on their own exclusive target feature dimension sets. When the concentration increases, the fresh air volume control factor responds independently, while the supply air temperature control factor remains unaffected. This avoids the mutual restraint between fresh air demand and temperature control demand within the area, resulting in the cancellation of hot and cold air, and eliminates energy waste caused by a lack of global coordination.

[0203] By combining the target control factor with the benchmark control factor, the equipment adjustment range can be made more in line with the actual needs of the region, thereby reducing the problems of frequent start-stop and local over-adjustment, and helping to balance comfort, stability and energy consumption control.

[0204] In one possible implementation, the multiple HVAC parameters include at least two of the following: supply air temperature, fresh air volume, return air valve opening, and water supply temperature.

[0205] In one possible implementation, the weights of each feature dimension in the target feature dimension set corresponding to each parameter to be controlled can be configured according to the importance (or degree of influence) of different feature dimensions on the corresponding parameter to be controlled, so that different parameters can form differentiated responses to the same multidimensional feature vector.

[0206] In one possible implementation, the weights of the target feature dimension set corresponding to each parameter to be controlled can be configured according to the importance (or degree of influence) of different feature dimensions on the corresponding parameter to be controlled, so that different parameters can form differentiated responses to the same multidimensional feature vector.

[0207] For example, regarding fresh air volume control, Concentration deviation, population density index, and outdoor air enthalpy can be used as primary input dimensions; for supply air temperature control, indoor temperature deviation, humidity deviation, building envelope thermal impact correction, and outdoor meteorological load index can be used as primary input dimensions; if using Indicates the first Normalized values ​​of each feature dimension. This indicates the weight of that dimension, and the target control factor. It can be represented as Subsequently, parameter adjustment values It can be represented as ,in, Indicates the reference parameter value. Indicates the benchmark control factor. This represents the adjustment sensitivity coefficient. Through the above calculation process, the multidimensional environmental state can be compressed into an executable adjustment intensity, and the output control quantity is ensured to fall within the executable range of the device by the reference value constraint, thus avoiding control overshoot or ineffective adjustment.

[0208] In another possible implementation, the single-dimensional deviation values ​​in the target multidimensional feature vector can be classified first, for example, into temperature control deviation, humidity control deviation, air quality deviation and equipment operation risk. Then, the comprehensive deviation value of each category can be calculated, and the total comprehensive deviation value can be obtained according to the preset priority and weight allocation ratio. Finally, the total comprehensive deviation value can be converted into parameter control value through the preset mapping relationship.

[0209] It is understandable that the above method of determining parameter control values ​​is applicable to multi-objective control scenarios that need to balance comfort, energy consumption, and equipment safety constraints; for example, in situations where the outdoor temperature and humidity are high but the indoor temperature and humidity are high. When the concentration is high, the system does not simply increase the fresh air volume, but rather combines the supply air dehumidification capacity, return air ratio and cooling capacity to generate comprehensive control commands in order to reduce the increase in energy consumption and dehumidification burden caused by simply introducing a large amount of high enthalpy fresh air.

[0210] In one possible implementation, the weights of the target feature dimensions corresponding to each parameter to be controlled can be determined by determining the degree of single-dimensional deviation of different dimensions in the target multidimensional feature vector corresponding to the HVAC control area; based on the target multidimensional feature vector corresponding to each HVAC control area and the degree of single-dimensional deviation of different dimensions in the target multidimensional feature vector, with global energy consumption optimization as the optimization objective, under preset constraints, the weights of the target feature dimensions corresponding to each parameter to be controlled are generated through a decision model.

[0211] The preset constraints include at least one of the following: operating boundary constraints for HVAC equipment; interlocking constraints for heating and cooling equipment; and upper limit constraints for zoned loads.

[0212] Understandably, the degree of deviation in a single dimension is used to characterize the magnitude of deviation of a single dimension from the target state. It is usually obtained by normalizing the difference between the current value and the corresponding target value, comfort zone, or benchmark threshold, and is used to reflect the strength of the contribution of that dimension to the control demand. The decision model can be constructed as a constrained optimization solution unit. Its input includes the multi-dimensional feature vector of each region's target, the degree of deviation in each dimension, and the equipment operating status information. It outputs the weight allocation results of each parameter to be controlled on different target feature dimensions, so that the weight allocation is consistent with the global energy consumption optimization target. The decision model can be, for example, a linear programming, quadratic programming, or heuristic optimization model.

[0213] The target multidimensional feature vector of each HVAC control zone is decomposed into dimensions, and the deviation of each dimension from the corresponding target value is calculated. The deviation is then used as a feature importance factor and input into the decision model. Through the decision model, the control parameters such as supply air temperature, fan speed, valve opening, and fresh air volume are jointly solved. During the solution process, the upper and lower limits of the parameters are limited according to the operating boundary constraints of HVAC equipment, the interlock constraints of cooling and heating equipment are used to avoid simultaneous strong operation of cooling and heating in the same area, and the upper limit constraints of the zone load limit restrict the amount of cooling and heating that can be allocated in a single area from not exceeding the set threshold. This ensures that the decision model outputs weighted results that conform to equipment safety, system coordination, and zone load-bearing capacity.

[0214] Understandably, this weighting result is used to determine subsequent parameter adjustment values, giving higher adjustment weights to dimensions with larger deviations and lower weights to dimensions with smaller deviations. This guides control resources to prioritize target dimensions that have a more significant impact on comfort and energy consumption. Since the weighting generation process considers the differences in characteristics across multiple regions and the global energy consumption target, the system can avoid energy waste caused by excessive adjustment in local areas and reduce the risk of heating / cooling conflicts and overload. At the same time, it transforms HVAC control from a single threshold judgment to a joint decision based on multi-dimensional characteristic deviations and constraint optimization, improving adaptability to complex load change scenarios, enhancing the consistency and stability of multi-region adjustment, and reducing overall system energy consumption while meeting comfort requirements.

[0215] For example, the operating boundary constraints of HVAC equipment refer to the inherent safe operating limits and working ranges of various HVAC equipment, which are the physical hardware boundaries of the equipment; the interlocking constraints of cooling and heating equipment are the linkage constraint logic set for the same HVAC control area, which prevents the cooling equipment and heating equipment from being in working state at the same time; the zoning load limit constraint conditions refer to the preset maximum allowable cooling and heating load supply limit for each HVAC control area, which is determined by the zoning area, the building envelope, and the design rated load.

[0216] In one possible implementation, after determining the static building attributes and temporal variation characteristics of each HVAC control area, the environmental change rate of each meteorological dimension outside the building can be determined based on multi-source heterogeneous environmental data. For any meteorological dimension, if the environmental change rate of the meteorological dimension exceeds the preset change rate, the corresponding compensation coefficient of the meteorological dimension is determined. After correcting the initial multidimensional feature vector of each HVAC control area based on the static building attributes and temporal variation characteristics of each HVAC control area to obtain the target multidimensional feature vector corresponding to each HVAC control area, the target multidimensional feature vector corresponding to each HVAC control area is corrected and updated based on the compensation coefficients corresponding to multiple meteorological dimensions to obtain the new target multidimensional feature vector corresponding to each HVAC control area.

[0217] Among them, the rate of environmental change is an indicator used to characterize how fast various meteorological dimensions outside the building change. For example, the rate of environmental change can be calculated from multi-source heterogeneous environmental data such as outdoor temperature, humidity, wind speed, and solar radiation intensity, and is used to identify sudden meteorological changes and trigger compensation mechanisms.

[0218] The compensation coefficient is an adjustment amount used to correct sudden weather changes. It can be correlated with the magnitude and duration of changes in meteorological dimensions and the degree of regional influence, so as to enhance the system's ability to respond to rapid changes in the external environment.

[0219] The new target multidimensional feature vector is used to characterize the final feature vector after meteorological compensation correction. In other words, the new target multidimensional feature vector refers to the comprehensive state of each HVAC control area that is more in line with real-time control requirements under sudden environmental changes.

[0220] Temperature, humidity, wind speed, wind direction, and irradiance data are collected by meteorological sensors installed on building facades, roofs, or surrounding areas. After time synchronization and outlier removal of multi-source heterogeneous environmental data, the change in each meteorological dimension per unit time is calculated according to a preset time window to obtain the rate of environmental change. When the rate of change of a certain meteorological dimension exceeds a preset threshold, the controller can generate a corresponding compensation coefficient based on the direction and magnitude of change of that meteorological dimension and the current seasonal operating conditions. The compensation coefficients of multiple meteorological dimensions are mapped to the target multidimensional feature vectors corresponding to each HVAC control area, and the dimensions such as temperature, humidity, air age, and heat load deviation are corrected and updated to obtain new target multidimensional feature vectors, which are used as inputs for subsequent parameter control value calculations.

[0221] For example, by using a distributed sensor network to detect the environmental conditions outside the building in real time, the compensation coefficient related to heating is increased when the outdoor temperature drops suddenly for a short period of time, and the compensation coefficient related to fresh air exchange and air supply is increased when the wind speed increases suddenly and is accompanied by an increase in outdoor cooling load.

[0222] When a sudden change in the external meteorological dimension is detected, a compensation coefficient is introduced in a timely manner to re-correct the target multidimensional feature vector, making the target features closer to the actual changes in heat and humidity load. This avoids the lagging judgment caused by relying solely on the static attributes and periodic characteristics of the building. Furthermore, by incorporating rapid outdoor fluctuation factors into the feature update process, the response speed of the building's HVAC system to sudden weather events is improved, reducing over-adjustment or under-adjustment of HVAC equipment, and improving regional comfort and operational stability.

[0223] In one possible implementation, the multi-source heterogeneous environmental data includes real-time HVAC loads, which are used to characterize the real-time supply and demand status of heating and cooling loads in the HVAC control area.

[0224] For any HVAC control area, based on multi-source heterogeneous environmental data, the standard HVAC load of the building is determined through a multi-field coupled load model; the load deviation between the real-time HVAC load and the standard HVAC load is determined, and the deviation correction coefficient corresponding to the load deviation is determined; based on the deviation correction coefficient and the real-time HVAC load, the corrected HVAC load of the HVAC control area is determined; after preprocessing the multi-source heterogeneous environmental data to obtain the initial multi-dimensional feature vector of each HVAC control area in the building, the initial multi-dimensional feature vector is updated based on the corrected HVAC load to obtain the updated initial multi-dimensional feature vector.

[0225] Among them, the deviation correction coefficient is used to correct the measurement error of real-time HVAC load; the multi-field coupled load model is a mechanism model constructed by coupling multiple types of indoor heat transfer physical fields, used to generate the theoretical HVAC load of the HVAC control area under the environment of multi-source heterogeneous environmental data; the corrected HVAC load is the load data after error correction by the multi-field coupled load model.

[0226] In this embodiment of the application, the real-time HVAC load is obtained by fusing multi-source data such as regional return air temperature, supply air volume, supply and return water temperature difference, and personnel density, and is used to characterize the current balance between heating and cooling demand and supply in the region.

[0227] The standard HVAC load is calculated by a multi-field coupled load model. This model couples physical processes such as heat transfer in the building envelope, indoor air convection heat transfer, and heat dissipation from the human body. It takes building static properties, outdoor meteorological data, and regional operation data as inputs and outputs the standard load corresponding to the theoretical heat balance.

[0228] Load deviation is used to reflect the difference between real-time sampled values ​​and theoretical values. The deviation correction coefficient can be obtained by the deviation ratio, historical error statistics or time-varying correction function, and is used to perform multiplicative or equivalent correction on real-time HVAC load, thereby forming the corrected HVAC load.

[0229] The modified HVAC load can be used as a write-back value to participate in the feature vector update, so that the load-related dimensions in the original multidimensional features are replaced with calibrated values, thereby reducing the impact of sensor drift and local disturbances on subsequent control.

[0230] For example, a multi-field coupled load model can be composed of a heat conduction sub-model, an air flow sub-model, and a human thermal comfort sub-model coupled together. The model parameters are jointly calibrated by the floor orientation, thermal resistance of the building envelope, window opening status, and equipment operating conditions. Before entering the model, multi-source heterogeneous environmental data undergoes time synchronization, outlier removal, and dimension normalization to form an initial feature vector for the region. Then, the load dimension in the feature vector is updated by combining the corrected HVAC load, so that the updated vector can more accurately reflect the actual operating status of the region.

[0231] Understandably, standard HVAC loads can provide a theoretical reference for real-time HVAC loads, deviation correction coefficients can compensate for measurement errors, and corrected HVAC loads can serve as more precise control inputs, thereby improving the accuracy and stability of building HVAC regulation. After updating the initial multidimensional feature vector based on the corrected HVAC load, the ability of subsequent feature analysis to characterize the actual load state is enhanced, which helps to reduce misjudgments and control lags, improve regional comfort, and reduce ineffective energy consumption.

[0232] In this embodiment, by mapping physical measurement values ​​to relative position values ​​within a continuous interval for parameter smoothing control, the system oscillations caused by hard switching between "on" and "off" in traditional threshold control are eliminated. By using time-series change characteristics (personnel arrival and departure patterns, differences between weekdays and non-weekdays, etc.) to enable the building HVAC system to predict load change trends and adjust in advance, the oscillation process of "over-limit / over-adjustment" in traditional PID control is avoided. Through differentiated response, different parameters are calculated independently. When a certain dimension changes, only the control parameters related to it respond, while other parameters are unaffected. This avoids the chain oscillations of "one parameter fluctuation triggering all parameters to move together" in traditional control, thus making the system more stable.

[0233] Meanwhile, by controlling the continuous relative position value based on the region's exclusive comfort zone, the equipment no longer undergoes repeated switching between "stop / start," eliminating the impact energy consumption of 3 to 7 times the rated current during startup. Through the unified characteristic space of multiple regions and the optimal weight allocation of global energy consumption, the control commands of each region are coordinated at the global level, avoiding energy waste caused by the offsetting of hot and cold. Differentiated response further avoids the additional energy consumption caused by the mutual restraint between fresh air and temperature control within the region, thereby comprehensively reducing the system's energy consumption and improving the system's energy efficiency.

[0234] The control method for a building HVAC system provided in this embodiment collects multi-source heterogeneous environmental data inside and outside the building through a distributed sensor network; based on the historical multi-source heterogeneous environmental data of the HVAC system, the static attributes of the building in each HVAC control area, and the temporal variation characteristics, it determines the baseline feature thresholds for different feature dimensions in each HVAC control area; based on the baseline feature thresholds for different feature dimensions, it normalizes the multi-source heterogeneous environmental data to obtain the initial multi-dimensional feature vectors for each HVAC control area within the building; it determines the static attributes and temporal variation characteristics of each HVAC control area, and for any HVAC control area, based on the static attributes of the building in the HVAC control area, it corrects the static deviation of the initial multi-dimensional feature vectors of each HVAC control area. The method involves: based on the temporal variation characteristics of the HVAC control area, performing a secondary correction process on the initial multidimensional feature vectors of each HVAC control area to obtain the target multidimensional feature vectors corresponding to each HVAC control area; for any HVAC control area, determining the weights of the target feature dimensions corresponding to each parameter to be controlled based on the target multidimensional feature vectors corresponding to the HVAC control area; for any parameter to be controlled, weighted summing of the weights of the target feature dimensions corresponding to the parameter to be controlled to obtain the target control factor corresponding to the parameter to be controlled; and determining the parameter control value of the parameter to be controlled based on the target control factor, the benchmark control factor of the parameter to be controlled, and the corresponding benchmark parameter value, and controlling the operation of the HVAC equipment in the HVAC control area according to the parameter control value. This method relies on an operational architecture consisting of a distributed sensor network, controllers, and HVAC actuators. It incorporates multi-source environmental information, regional inherent attributes, and temporal variation patterns into a unified analysis process. This transforms the adjustment of HVAC equipment from a single-parameter judgment to a control decision based on the comprehensive state of the region. As a result, it solves the problems of existing systems such as adjustment lag, insufficient adaptability, and regional control incoordination. It improves the globality, adaptability, and coordination of HVAC regulation, reduces response lag and frequent start-stop, improves indoor comfort and air quality, reduces energy consumption, and effectively reduces the frequency of adverse operating conditions.

[0235] Figure 3 A schematic diagram of the control device for the building HVAC system provided in this application. Figure 3 As shown, this application provides a control device for a building HVAC system. The control device 300 for the building HVAC system includes:

[0236] The acquisition module 301 is used to collect multi-source heterogeneous environmental data inside and outside the building through a distributed sensor network.

[0237] The processing module 302 is used to preprocess multi-source heterogeneous environmental data to obtain the initial multi-dimensional feature vectors of each HVAC control area in the building.

[0238] The processing module 302 is also used to determine the static attributes and temporal change characteristics of buildings in each HVAC control zone.

[0239] The processing module 302 is also used to correct the initial multidimensional feature vector of each HVAC control area based on the building's static attributes and temporal change characteristics, so as to obtain the target multidimensional feature vector corresponding to each HVAC control area; the temporal change characteristics are used to reflect the periodic pattern of the changes of different HVAC control areas in the building over time.

[0240] The processing module 302 is also used to determine the parameter control value of any HVAC control area based on the target multidimensional feature vector corresponding to the HVAC control area, and control the operation of the HVAC equipment in the HVAC control area according to the parameter control value.

[0241] The processing module 302 is also used to select, based on the physical influence relationship between the parameters to be controlled and each feature dimension in the target multidimensional feature vector, the feature dimensions that have a physical influence relationship with the parameters to be controlled for the multiple HVAC equipment in the HVAC control area, and use them as the target feature dimension set corresponding to the parameters to be controlled.

[0242] The processing module 302 is also used to determine the weight of each target feature dimension in the target feature dimension set corresponding to the parameter to be adjusted.

[0243] The processing module 302 is also used to perform a weighted summation of the weights of the target feature dimensions corresponding to any parameter to be controlled, so as to obtain the target control factor corresponding to the parameter to be controlled.

[0244] The processing module 302 is also used to determine the parameter control value of the parameter to be controlled based on the target control factor, the benchmark control factor of the parameter to be controlled and its corresponding benchmark parameter value.

[0245] Among them, the target feature dimension sets corresponding to different parameters to be controlled are different, and the weight polarity of the same feature dimension in different parameters to be controlled is determined according to the physical influence relationship between the corresponding feature dimension and the parameter to be controlled.

[0246] In one possible implementation, the processing module 302 is further configured to determine the baseline feature thresholds for different feature dimensions in each HVAC control area based on historical multi-source heterogeneous environmental data of the HVAC system, the static building attributes and temporal change characteristics of each HVAC control area.

[0247] The processing module 302 is also used to normalize multi-source heterogeneous environmental data based on the benchmark feature thresholds of different feature dimensions to obtain the initial multi-dimensional feature vectors of each HVAC control area in the building.

[0248] In one possible implementation, the processing module 302 is further configured to, for any HVAC control zone, perform static deviation correction processing on the initial multidimensional feature vector of each HVAC control zone based on the building static attributes of the HVAC control zone.

[0249] The processing module 302 is also used to perform secondary correction processing on the initial multidimensional feature vectors of each HVAC control area based on the temporal change characteristics of the HVAC control area, so as to obtain the target multidimensional feature vectors corresponding to each HVAC control area.

[0250] In one possible implementation, the processing module 302 is further configured to determine the degree of single-dimensional deviation in different dimensions of the target multidimensional feature vector corresponding to the HVAC control area.

[0251] The processing module 302 is also used to generate the weights of the target feature dimensions of each parameter to be controlled through a decision model, based on the target multidimensional feature vectors corresponding to each HVAC control area and the degree of deviation of the single dimension of different dimensions in the target multidimensional feature vectors, with the global energy consumption optimization as the optimization objective, under preset constraints.

[0252] The preset constraints include at least one of the following:

[0253] Operating boundary constraints for HVAC equipment;

[0254] Interlocking constraints between heating and cooling equipment;

[0255] Zone load limit constraints.

[0256] In one possible implementation, the processing module 302 is further configured to determine the rate of environmental change for each meteorological dimension outside the building based on multi-source heterogeneous environmental data.

[0257] The processing module 302 is also used to determine the compensation coefficient corresponding to any meteorological dimension when the rate of environmental change in the meteorological dimension exceeds the preset rate of change.

[0258] The processing module 302 is also used to correct and update the target multidimensional feature vector corresponding to each HVAC control area based on the compensation coefficients corresponding to multiple meteorological dimensions, so as to obtain the new target multidimensional feature vector corresponding to each HVAC control area.

[0259] In one possible implementation, the processing module 302 is further configured to determine the standard HVAC load of a building for any HVAC control zone based on multi-source heterogeneous environmental data and through a multi-field coupled load model.

[0260] The processing module 302 is also used to determine the load deviation between the real-time HVAC load and the standard HVAC load, and to determine the deviation correction coefficient corresponding to the load deviation; the deviation correction coefficient is used to correct the measurement error of the real-time HVAC load.

[0261] The processing module 302 is also used to determine the corrected HVAC load of the HVAC control area based on the deviation correction coefficient and the real-time HVAC load; the corrected HVAC load is the load data after error correction by the multi-field coupled load model;

[0262] Among them, the multi-field coupled load model is a mechanism model constructed by coupling multiple types of indoor heat transfer physical fields, which is used to generate the theoretical HVAC load of the HVAC control area under the environment of multi-source heterogeneous environmental data.

[0263] The processing module 302 is also used to update the initial multidimensional feature vector based on the corrected HVAC load to obtain the updated initial multidimensional feature vector.

[0264] The control device for the building HVAC system provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0265] Figure 4 A schematic diagram of the structure of the electronic device provided in this application. Figure 4 As shown, the electronic device 400 provided in this embodiment includes at least one processor 401 and a memory 402. Optionally, the device 400 further includes a communication component 403. The processor 401, memory 402, and communication component 403 are connected via a bus 404.

[0266] In a specific implementation, at least one processor 401 executes computer execution instructions stored in memory 402, causing at least one processor 401 to perform the above-described method.

[0267] The specific implementation process of processor 401 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0268] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0269] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0270] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0271] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0272] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0273] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0274] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0275] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0276] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0277] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0278] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0279] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0280] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A control method for a building heating, ventilation, and air conditioning system, characterized in that, include: Multi-source heterogeneous environmental data inside and outside the building are collected through a distributed sensor network, and the multi-source heterogeneous environmental data are preprocessed to obtain the initial multi-dimensional feature vectors of each HVAC control area in the building. Determine the static building attributes and temporal variation characteristics of each HVAC control zone; Based on the building's static attributes and temporal variation characteristics of each HVAC control area, the initial multidimensional feature vectors of each HVAC control area are modified to obtain the target multidimensional feature vectors corresponding to each HVAC control area; the temporal variation characteristics are used to reflect the periodic patterns of different HVAC control areas in the building over time. For any HVAC control zone, based on the target multidimensional feature vector corresponding to the HVAC control zone, the parameter control value of the HVAC control zone is determined, and the operation of the HVAC equipment in the HVAC control zone is controlled according to the parameter control value; The step of determining the parameter control values ​​of the HVAC control area based on the target multidimensional feature vector corresponding to the HVAC control area includes: For the control parameters corresponding to multiple HVAC devices in the HVAC control area, based on the physical influence relationship between the control parameters and each feature dimension in the target multidimensional feature vector, feature dimensions that have a physical influence relationship with the control parameters are selected from the target multidimensional feature vector and used as the target feature dimension set corresponding to the control parameters. Determine the weight of each target feature dimension in the target feature dimension set corresponding to the parameter to be adjusted; For any of the parameters to be controlled, the weights of the target feature dimensions corresponding to the parameter to be controlled are summed to obtain the target control factor corresponding to the parameter to be controlled; Based on the target control factor, the benchmark control factor of the parameter to be controlled and its corresponding benchmark parameter value, the parameter control value of the parameter to be controlled is determined; Among them, the target feature dimension sets corresponding to different parameters to be controlled are different, and the weight polarity of the same feature dimension in different parameters to be controlled is determined according to the physical influence relationship between the corresponding feature dimension and the parameter to be controlled.

2. The method according to claim 1, characterized in that, The preprocessing of the multi-source heterogeneous environmental data to obtain the initial multidimensional feature vectors of each HVAC control area within the building includes: Based on the historical multi-source heterogeneous environmental data of the HVAC system, the static building attributes and temporal change characteristics of each HVAC control area, the baseline feature thresholds for different feature dimensions in each HVAC control area are determined. Based on the benchmark feature thresholds of the different feature dimensions, the multi-source heterogeneous environmental data is normalized to obtain the initial multi-dimensional feature vectors of each HVAC control area within the building.

3. The method according to claim 2, characterized in that, The process involves refining the initial multidimensional feature vectors of each HVAC control area based on the building's static attributes and temporal variation characteristics to obtain the target multidimensional feature vectors corresponding to each HVAC control area, including: For any of the HVAC control zones, the static deviation of the initial multidimensional feature vectors of each HVAC control zone is corrected based on the building static attributes of the HVAC control zone. Based on the temporal variation characteristics of the HVAC control area, the initial multidimensional feature vectors of each HVAC control area are modified in a second step to obtain the target multidimensional feature vectors corresponding to each HVAC control area.

4. The method according to claim 1, characterized in that, The step of determining the weights of the target feature dimensions corresponding to each parameter to be controlled based on the target multidimensional feature vector corresponding to the HVAC control area includes: Determine the degree of single-dimensional deviation in different dimensions of the target multidimensional feature vector corresponding to the HVAC control area; Based on the target multidimensional feature vectors corresponding to each HVAC control zone and the degree of deviation of the single dimension of the target multidimensional feature vectors, with global energy consumption optimization as the optimization objective, under preset constraints, the decision model generates the weights of the target feature dimensions corresponding to each parameter to be controlled. The preset constraints include at least one of the following: Operating boundary constraints for HVAC equipment; Interlocking constraints between heating and cooling equipment; Zone load limit constraints.

5. The method according to claim 1, characterized in that, After determining the building static attributes and temporal variation characteristics of each HVAC control zone, the method further includes: Based on the aforementioned multi-source heterogeneous environmental data, the rate of environmental change for each meteorological dimension outside the building is determined; For any meteorological dimension, if the rate of environmental change in the meteorological dimension exceeds a preset rate of change, a compensation coefficient corresponding to the meteorological dimension is determined. After correcting the initial multidimensional feature vectors of each HVAC control area based on the building's static attributes and temporal variation characteristics to obtain the target multidimensional feature vectors corresponding to each HVAC control area, the process includes: Based on the compensation coefficients corresponding to the multiple meteorological dimensions, the target multidimensional feature vectors corresponding to each HVAC control area are corrected and updated to obtain new target multidimensional feature vectors corresponding to each HVAC control area.

6. The method according to claim 1, characterized in that, The multi-source heterogeneous environmental data includes real-time HVAC loads, which are used to characterize the real-time supply and demand status of heating and cooling loads in the HVAC control area; the method further includes: For any of the aforementioned HVAC control zones, based on the multi-source heterogeneous environmental data, the standard HVAC load of the building is determined through a multi-field coupled load model; The load deviation between the real-time HVAC load and the standard HVAC load is determined, and a deviation correction factor corresponding to the load deviation is determined; the deviation correction factor is used to correct the measurement error of the real-time HVAC load. Based on the deviation correction coefficient and the real-time HVAC load, the corrected HVAC load for the HVAC control area is determined; the corrected HVAC load is the load data after error correction by the multi-field coupled load model. The multi-field coupled load model is a mechanism model constructed by coupling multiple types of indoor heat transfer physical fields, which is used to generate the theoretical HVAC load of the HVAC control area under the environment of multi-source heterogeneous environmental data. After preprocessing the multi-source heterogeneous environmental data to obtain the initial multidimensional feature vectors of each HVAC control area within the building, the method further includes: Based on the corrected HVAC load, the initial multidimensional feature vector is updated to obtain the updated initial multidimensional feature vector.

7. A building heating, ventilation, and air conditioning system, characterized in that, include: A controller and a distributed sensor network, wherein the distributed sensor network is located in multiple HVAC control areas within the building and outside the building; The controller is used to perform the method according to any one of claims 1-6.

8. An electronic device, characterized in that, include: Memory, processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-6.

Images