Building and municipal facility operation and maintenance management method based on multi-agent collaborative decision-making

Abstract

This invention discloses a method for the operation and maintenance management of buildings and municipal facilities based on multi-agent collaborative decision-making, relating to the field of building operation and maintenance management. This method combines a dynamic evaluation mechanism of the thermal inertia index Hk with a path collaborative identification mechanism of the spatial thermal coupling map G. For the first time, this invention establishes a complete closed-loop control path from microclimate parameter acquisition, thermal inertia calculation, response lag assessment, collaborative buffer path identification, valve collaborative adjustment, and secondary sensing feedback. This mechanism enables building air conditioning systems to achieve a fully automatic multi-agent adaptive adjustment process involving perception, decision-making, and execution. Compared to traditional air conditioning systems, it not only responds more promptly but also exhibits stronger system stability and energy-saving potential under complex thermal disturbances, providing an efficient and scalable solution for the intelligent environmental operation and maintenance of complex spaces such as domes.

Description

Technical Field

[0001] This invention relates to the field of building operation and maintenance management technology, specifically a method for the operation and maintenance management of buildings and municipal facilities based on multi-agent collaborative decision-making. Background Technology

[0002] With the continuous development of green building concepts and smart city infrastructure, large public buildings are showing a trend towards diversification and intelligence in their structural forms and operation and maintenance models. In particular, new architectural spaces such as exhibition halls and museums with large-span glass dome structures have higher requirements for the dynamic response and control of indoor microclimate environments. Due to problems such as strong sunlight, high heat load fluctuations, large vertical temperature differences, and drastic changes in personnel density, conventional centralized air conditioning control modes face challenges in terms of response speed and adjustment accuracy.

[0003] Current environmental control systems in large exhibition halls generally adopt a mode of air conditioning regulation based on unified parameters for the entire hall or prediction of host operating status. Essentially, this aims at overall average response and provides coarse-grained control of the building space. However, within glass dome buildings, local spaces such as the south-facing area with strong sunlight, the central negotiation area, areas with high pedestrian traffic, and elevated walkways exhibit significant differences in heat capacity, light intensity, ventilation, and usage frequency. This leads to a lag in response from conventional methods based on the average values ​​of sensors throughout the hall, resulting in high energy consumption, uneven temperature and humidity, and poor local comfort. Furthermore, the lack of effective spatial heat diffusion prediction mechanisms and regional regulation path identification based on physical thermal coupling relationships results in low energy distribution regulation efficiency and makes it difficult to implement regulation strategies.

[0004] The aforementioned shortcomings stem primarily from three factors: First, existing control strategies fail to detect differences in thermal inertia between blocks, ignoring the varying response speeds of different areas to the same regulatory behavior. Second, the system lacks a thermal diffusion path identification mechanism based on spatial coupling structures, hindering collaborative control based on adjacent block thermal buffers. Third, the lack of an intelligent agent mechanism that integrates the model and the building's physical scene leads to a broken feedback loop, disconnecting strategy execution from subsequent re-evaluation. These control mismatches directly result in abnormal effects such as frequent temperature fluctuations in areas, unstable humidity control in key exhibit areas, discomfort caused by direct cold air blowing, and ineffective amplification of air conditioning system energy consumption, severely impacting exhibition hall operational efficiency and visitor comfort. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method for the operation and maintenance management of building and municipal facilities based on multi-agent collaborative decision-making, which solves the problems mentioned in the background technology.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for operation and maintenance management of building and municipal facilities based on multi-agent collaborative decision-making, comprising the following steps: S1. Divide the interior of the dome exhibition hall into multiple microclimate blocks Z. Set up a microclimate block agent AgentZ in each microclimate block Z to collect environmental parameter information in real time. Upload the collected environmental parameter information to the central multi-agent collaborative server and perform preprocessing to obtain the thermal inertia standard dataset. S2. Construct a thermodynamic characteristic model based on the standard thermal inertia dataset, calculate the thermal inertia index Hk for each microclimate block Z, set the thermal inertia threshold Hth for preliminary comparative evaluation, and trigger a collaborative response mechanism based on the preliminary comparative evaluation results. S3. After triggering the collaborative response mechanism, based on the building BIM model and the regional thermal coupling relationship, identify the set N of adjacent blocks of the high inertia block, and calculate the valve collaborative adjustment value Fadj. S4. The coordinated adjustment value Fadj of the air valve is sent to the corresponding microclimate block Z intelligent agent through the building automation system interface. The microclimate block Z intelligent agent controls the air supply equipment in its microclimate block Z to perform environmental adjustment according to the received coordinated adjustment value Fadj of the air valve.

[0007] Preferably, S1 includes S11; S11. Divide the interior of the dome exhibition hall into multiple microclimate blocks Z. The microclimate blocks Z are divided into 9 independent microclimate blocks Z by physical space division based on building lighting model, functional type, indoor heat capacity structure characteristics, air conditioning vent server distribution boundary and CAD structure and CFD airflow simulation results. Each independent microclimate block Z is labeled as the kth microclimate block Zk. An edge processing terminal is deployed in each microclimate block Z as a microclimate block agent (AgentZ). The microclimate block agent (AgentZ) includes a sensing access unit, a data transmission unit, and an instruction response unit.

[0008] Preferably, S1 further includes S12; S12. Deploy environmental acquisition devices in each microclimate block Z to collect environmental parameter information in real time. Integrate the collected environmental parameter information into the corresponding microclimate block agent AgentZ through the sensing access unit. Then, label the environmental parameter information with the current k-th microclimate block Zk tag through the data transmission unit, perform local time serialization processing and unified structured encapsulation of data format to obtain the data encapsulation package, and then transmit the data encapsulation package to the central multi-agent collaborative server using the MQTT interface protocol. The environmental data acquisition equipment includes a temperature sensor, a humidity sensor, a wind speed sensor, a light sensor, and an infrared people sensor. The environmental parameter information includes the air temperature Tz (Zk) of the kth microclimate block Z, the relative humidity Sdz (Zk) of the kth microclimate block Z, the wind speed Vs (Zk) of the kth microclimate block Z, the solar radiation intensity Gs (Zk) of the kth microclimate block Z, and the population density Nocc (Zk) of the kth microclimate block Z.

[0009] Preferably, S1 further includes S13; S13. In the central multi-agent collaborative server, the data package is received in real time, and the environmental parameter information is extracted by decapsulation. The data stream is partitioned and classified based on the Z label of each microclimate block. The environmental parameter information of each microclimate block Z is preprocessed to obtain the thermal inertia standard dataset. The preprocessing includes feature extraction and dimensionless processing; The feature extraction is performed by extracting features from the environmental parameter information after data cleaning to obtain a set of key factor parameters for thermal inertia modeling; The set of key factor parameters includes the heat capacity factor C (Zk) of the kth microclimate block Z, the heat transfer coefficient factor U (Zk) of the kth microclimate block Z, the temperature fluctuation standard deviation BTz (Zk) of the kth microclimate block Z, and the solar disturbance response gradient δGs (Zk) / δTz (Zk) of the kth microclimate block Z. The dimensionless processing integrates the key factor parameter set with environmental parameter information, converts them into dimensionless standard values ​​using the extreme value normalization method, and summarizes them to obtain a thermal inertia standard dataset.

[0010] Preferably, S2 includes S21; S21. In the central multi-agent collaborative server, a thermodynamic feature model is constructed based on the thermal inertia standard dataset. The thermodynamic feature model is constructed for microclimate block Z through a built-in thermal response modeling algorithm to describe the block's response capability to external disturbances and air conditioning regulation behavior. Then, the real-time acquired thermal inertia standard dataset is input into the thermodynamic feature model to calculate and output the thermal inertia index Hk of each microclimate block Z. The thermal inertia index Hk is calculated and output using the following thermodynamic characteristic model; ; In the formula, Hk(Zk) represents the thermal inertia index of the k-th microclimate block Z. This represents the average air temperature of the k-th microclimate block Z. Let Gs(Zk) represent the rate of change in solar radiation intensity Gs(Zk) of the k-th microclimate block Z. denoted as the response rate of the air temperature Tz(Zk) in the k-th microclimate block Z; a1 and a2 represent the preset empirical coefficients of the disturbance response term correction and the stability correction term, respectively, with initial values ​​set to 0.6 and 0.4.

[0011] Preferably, S2 further includes S22; S22. In the central multi-agent collaborative server, based on the statistical results of the thermal inertia index Hk of the same microclimate block Z within 24 historical operating cycles, the critical value between the normal and abnormal states of thermal response performance is selected as the thermal inertia threshold Hth. Next, the thermal inertia index Hk of each microclimate block Z, acquired in real time, is compared and evaluated with the thermal inertia threshold Hth to determine the risk of thermal response lag in microclimate block Z. Based on the preliminary comparison and evaluation results, a collaborative response mechanism is triggered. The specific evaluation content is as follows: When the thermal inertia index Hk > thermal inertia threshold Hth, it is determined that the current microclimate block Z has a risk of thermal response lag, the current microclimate block is marked as a target block to be coordinated and regulated, and the coordinated response mechanism is triggered. When the thermal inertia index Hk ≤ thermal inertia threshold Hth, the current microclimate block Z is determined to be in an autonomous and stable adjustment state, and the local agent of the current microclimate block continues to execute the conventional prediction and adjustment strategy.

[0012] Preferably, S3 further includes S31; S31. After triggering the collaborative response mechanism, a spatial thermal coupling map G is constructed in the central multi-agent collaborative server to identify the paths that need to be coordinated and buffered. The spatial thermal coupling map G is constructed by taking each microclimate block Z as a map node, establishing a connection edge between any two k-th microclimate blocks Zk with structural thermal conduction relationship and the adjacent j-th microclimate block Zj, and determining the edge weight based on the actual building structure thermal simulation results and historical energy flow data, which serves as the structural coupling coefficient Y(Zk, Zj) between the k-th microclimate block Zk and the adjacent j-th microclimate block Zj. The central server takes all microclimate blocks Z with thermal inertia index Hk higher than thermal inertia threshold Hth as the main control area, then traverses the set of directly adjacent blocks Nk in the spatial thermal coupling map G of the main control area, and identifies adjacent blocks whose thermal inertia index Hk(Zj) of adjacent microclimate block j is lower than the thermal inertia index Hk(Zk) of adjacent microclimate block Zj as cooperative buffers.

[0013] Preferably, S3 further includes S32; S32. The central multi-agent collaborative server issues a valve strategy adjustment instruction to the microclimate block agent Agent Z corresponding to the microclimate block Z in the collaborative buffer. The valve strategy adjustment command extracts the set of directly adjacent blocks Nk and the structural coupling coefficient Y(Zk, Zj) between the kth microclimate block Zk and the adjacent jth microclimate block Zj. It then calculates and outputs the valve collaborative adjustment value Fadj by combining the thermal inertia index Hk(Zj) of the adjacent jth microclimate block Zj with the thermal inertia index Hk(Zk) of the kth microclimate block Zj. The valve collaborative adjustment value Fadj is then sent to the command response unit to control the valve actuator in the current collaborative buffer to perform a preset opening adjustment. The coordinated adjustment value Fadj of the air valve is calculated and output using the following algorithm formula; ; In the formula, Fbase(Zk) represents the basic opening degree of the damper in the k-th microclimate block Z, and its value is dimensionless; Fadj(Zk) represents the coordinated adjustment value of the damper in the k-th microclimate block Z; and Href represents the precise value of the reference thermal inertia index, which is dimensionless. This represents an empirical coefficient, with a value range of 0.1-0.3.

[0014] Preferably, S4 includes S41; S41. The step of sending the valve collaborative adjustment value Fadj to the corresponding microclimate block agent AgentZ includes: the central multi-agent collaborative server encapsulates the valve collaborative adjustment value Fadj into a data instruction conforming to the Modbus protocol through the communication interface of the building automation system (BAS), and sends the data instruction to the corresponding microclimate block agent AgentZ instruction response unit according to the device address identifier of the microclimate block Z. Then, through the air valve controller in the command response unit, the air valve controller adjusts the operating parameters of the air supply equipment corresponding to the microclimate block Z in real time after receiving the air valve coordinated adjustment value Fadj.

[0015] Preferably, S4 further includes S42; S42. After the air valve controller of microclimate block Z performs environmental adjustment on the air supply equipment of its microclimate block Z by executing the air valve collaborative adjustment value Fadj, it restarts the sensing access unit of the microclimate block intelligent agent Agent Z after a preset 60 minutes, collects the adjusted environmental parameter information, inputs it into the thermodynamic characteristic model, outputs the thermal inertia index Hk, iterates and evaluates the thermal inertia, and if the risk of thermal response lag is still present after three adjustments, it triggers a manual warning through the central multi-agent collaborative server, and at the same time, manual inspection adjusts the microclimate block Z with the current risk of thermal response lag.

[0016] This invention provides a method for the operation and maintenance management of building and municipal facilities based on multi-agent collaborative decision-making. It has the following beneficial effects: (1) This method sets up multiple microclimate block intelligent agents inside the dome exhibition hall and classifies and manages them and identifies risks based on the thermal inertia index Hk of each microclimate block. Compared with the existing centralized air conditioning control method based on static temperature setpoints, it can significantly improve the sensitivity and control accuracy of local environmental response. The thermal inertia index Hk, combined with heat capacity, heat transfer, disturbance sensitivity and stability indicators, constitutes a multi-dimensional perception model for the dynamic load of the building, enabling the system to accurately identify response lag risks, thereby making more forward-looking pre-control strategy deployments and improving the overall air conditioning energy efficiency and comfort guarantee level.

[0017] (2) This method constructs a spatial thermal coupling map G with microclimate blocks as nodes and integrates a static structural thermal resistance model with a dynamic temperature-coordinated behavior coupling index. For the first time, it realizes the graph calculation modeling of building heat conduction paths, enabling the system to accurately identify the buffer paths with the greatest heat diffusion potential and low-inertia adjacent blocks after triggering a coordinated response. By constructing the valve coordinated adjustment value Fadj through the structural coupling coefficient Y(Zk, Zj), the system achieves precise adjustment and control of the air supply equipment in the buffer block. Under the premise of not affecting the environmental stability of the main exhibition area, it effectively buffers the heat load of the main control area, reduces the response time delay and energy consumption redundancy of the overall system, and significantly improves the intelligence level of multi-block coupling control.

[0018] (3) This method combines the dynamic evaluation mechanism of the thermal inertia index Hk with the path collaborative identification mechanism of the spatial thermal coupling spectrum G. For the first time, this invention establishes a complete closed-loop control path from microclimate parameter acquisition, thermal inertia calculation, response lag evaluation, collaborative buffer path identification, valve collaborative adjustment, and secondary sensing feedback. This mechanism enables the building air conditioning system to achieve a fully automatic multi-agent adaptive adjustment process of perception-decision-execution. Compared with traditional air conditioning systems, it not only responds more promptly, but also has stronger system stability and energy-saving potential in the face of complex thermal disturbances, providing an efficient and scalable solution for the intelligent operation and maintenance of complex spaces such as domes. Attached Figure Description

[0019] Figure 1 This is a schematic diagram illustrating the steps of the building and municipal facility operation and maintenance management method based on multi-agent collaborative decision-making of the present invention; Figure 2 This is a schematic diagram of the connection between AgentZ, the intelligent agent for microclimate blocks. Figure 3 This is a schematic diagram of the Z-zone division of microclimate areas. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Example 1 This invention provides a method for the operation and maintenance management of building and municipal facilities based on multi-agent collaborative decision-making. Please refer to [link / reference]. Figure 1 This includes the following steps: S1. Divide the interior of the dome exhibition hall into multiple microclimate blocks Z. Set up a microclimate block agent AgentZ in each microclimate block Z to collect environmental parameter information in real time. Upload the collected environmental parameter information to the central multi-agent collaborative server and perform preprocessing to obtain the thermal inertia standard dataset. S2. Construct a thermodynamic characteristic model based on the standard thermal inertia dataset, calculate the thermal inertia index Hk for each microclimate block Z, set the thermal inertia threshold Hth for preliminary comparative evaluation, and trigger a collaborative response mechanism based on the preliminary comparative evaluation results. S3. After triggering the collaborative response mechanism, based on the building BIM model and the regional thermal coupling relationship, identify the set N of adjacent blocks of the high inertia block, and calculate the valve collaborative adjustment value Fadj. S4. The coordinated adjustment value Fadj of the air valve is sent to the corresponding microclimate block Z intelligent agent through the building automation system interface. The microclimate block Z intelligent agent controls the air supply equipment in its microclimate block Z to perform environmental adjustment according to the received coordinated adjustment value Fadj of the air valve.

[0022] In this embodiment, the method first divides the interior of the dome exhibition hall into multiple microclimate blocks Z with physical boundaries, and deploys an independent microclimate block agent (AgentZ) within each block, constructing a distributed sensing and edge processing network for the entire area. This setup enables real-time acquisition and preliminary local processing of multi-dimensional environmental parameters such as air temperature, humidity, wind speed, light intensity, and personnel density, forming a structured thermal inertia standard dataset which is then uploaded to the central multi-agent collaborative server, significantly improving the system's dynamic control over regional thermal behavior. Based on the uploaded dataset, the central collaborative server constructs a thermodynamic feature model and introduces the thermal inertia index Hk as a key evaluation indicator. The thermal inertia index Hk considers not only the heat capacity and heat transfer coefficient of the microclimate block but also integrates the external solar disturbance response gradient and temperature time series volatility, thereby comprehensively measuring the temperature response speed and stability of a block under external disturbances or air conditioning regulation. To ensure the accuracy of risk identification, a thermal inertia threshold Hth is further set, and based on this, the thermal response capability of each block is dynamically evaluated. Once the thermal inertia index Hk of a certain block exceeds the threshold Hth, it is determined that the block has a risk of thermal response lag, triggering a subsequent collaborative response mechanism. In this mechanism, a spatial thermal coupling map G is constructed by calling the building BIM model and historical energy consumption data. The weights of each edge in the map are generated based on a dual weighting of the coupling relationship between structural thermal resistance, shared boundary area, and block temperature sequence, forming a structural coupling coefficient Y(Zk,Zj), which is used to realistically reflect the heat diffusion potential between different blocks. The adjacent low-inertia buffer zones of the high-inertia main control zone are identified as heat diffusion target paths, and the damper collaborative adjustment value Fadj is calculated. This value is dynamically extrapolated based on the thermal inertia gradient and coupling strength to ensure that heat actively diffuses along a controllable path, improving overall control efficiency. Finally, the damper collaborative adjustment value Fadj is sent to the agent AgentZ of the corresponding microclimate block Z through the building automation system interface, which then controls the local air supply equipment to execute the corresponding environmental regulation strategy. To achieve complete control logic, a delayed sampling mechanism is set after adjustment to re-collect environmental parameters for thermal inertia iterative evaluation. If the thermal response status is not improved after three consecutive adjustments, a manual inspection and early warning is automatically triggered to ensure long-term stable operation.

[0023] Example 2 Please see Figure 1 , Figure 2 and Figure 3 Specifically: S1 includes S11; S11. Divide the interior of the dome exhibition hall into multiple microclimate blocks Z. The microclimate blocks Z are physically divided based on the building lighting model, the type of use function, the characteristics of the indoor heat capacity structure, the distribution boundary of the air conditioning vent server, and the CAD structure and CFD airflow simulation results, forming 9 independent microclimate blocks Z. Each independent microclimate block Z is labeled as the kth microclimate block Zk. Specifically: The first microclimate zone Z1 is located on the south side near the glass curtain wall, in a plant display area that receives direct sunlight and has a high latent heat of evaporation. The second microclimate zone, Z2, is located in the inner ring area on the north side, an exhibition area requiring constant temperature and humidity. The third microclimate zone, Z3, is located in the negotiation area directly below the dome, where there is high foot traffic and periodic heat load. The fourth microclimate block Z4 is located in the skywalk area where there is strong high-altitude cold radiation and unstable wind speed. Microclimate blocks Z5 to Z9 are located in other temporary exhibition areas or equipment areas, with different structural features, and are flexibly divided according to thermal coupling relationships. An edge processing terminal is deployed in each microclimate block Z as a microclimate block agent (AgentZ). The microclimate block agent (AgentZ) includes a sensing access unit, a data transmission unit, and an instruction response unit. The sensing access unit extracts environmental parameter information by connecting to sensors and building lighting models set in the temporal part of each microclimate block Z. The data transmission unit integrates the extracted environmental parameter information and uploads it to the central multi-agent collaborative server; The instruction response unit receives and executes policy instructions from the central multi-agent collaborative server.

[0024] S1 also includes S12; S12. Deploy environmental acquisition devices in each microclimate block Z to collect environmental parameter information in real time. Integrate the collected environmental parameter information into the corresponding microclimate block agent AgentZ through the sensing access unit. Then, label the environmental parameter information with the current k-th microclimate block Zk tag through the data transmission unit, perform local time serialization processing and unified structured encapsulation of data format to obtain the data encapsulation package, and then transmit the data encapsulation package to the central multi-agent collaborative server using the MQTT interface protocol. Environmental data acquisition equipment includes temperature sensors, humidity sensors, wind speed sensors, light sensors, and infrared people sensors; The environmental parameter information includes the air temperature Tz (Zk) of the k-th microclimate block Z, the relative humidity Sdz (Zk) of the k-th microclimate block Z, the wind speed Vs (Zk) of the k-th microclimate block Z, the solar radiation intensity Gs (Zk) of the k-th microclimate block Z, and the population density Nocc (Zk) of the k-th microclimate block Z.

[0025] S1 also includes S13; S13. In the central multi-agent collaborative server, the data package is received in real time, and the environmental parameter information is extracted by decapsulation. The data stream is partitioned and classified based on the Z label of each microclimate block. The environmental parameter information of each microclimate block Z is preprocessed to obtain the thermal inertia standard dataset. Preprocessing includes feature extraction and dimensionless processing; Feature extraction involves extracting features from environmental parameter information after data cleaning to obtain a set of key factor parameters for thermal inertia modeling. The key factor parameter set includes the heat capacity factor C (Zk) of the k-th microclimate block Z, the heat transfer coefficient factor U (Zk) of the k-th microclimate block Z, the temperature fluctuation standard deviation BTz (Zk) of the k-th microclimate block Z, and the solar disturbance response gradient δGs (Zk) / δTz (Zk) of the k-th microclimate block Z. Dimensionless processing integrates key factor parameter sets with environmental parameter information, converts them into dimensionless standard values ​​using extreme value normalization, and then summarizes them to obtain a thermal inertia standard dataset. The heat capacity factor C(Zk) of the k-th microclimate block Z is calculated by comparing the fluctuation of wind speed Vs(Zk) over a continuous time period with the response lag relationship of air temperature Tz(Zk) of the corresponding k-th microclimate block Z. A sliding window difference is used in the data layer, and a fitting algorithm is used to extract the response intensity between air supply changes and temperature lag, which is used to infer the heat capacity level. This calculation does not depend on building material parameters, but is achieved through a data-driven approach. The heat transfer coefficient factor U(Zk) of the kth microclimate block Z: During the nighttime period when there is no human disturbance (Nocc(Zk) is stable) and the light change is small, the heat transfer capacity is estimated by extracting the temperature drop process of the outdoor temperature and the air temperature Tz(Zk) of the kth microclimate block Z, combining the temperature change slope to construct a heat transfer attenuation model, and using the least squares method to fit the linear relationship between the external temperature difference and the internal temperature rate. The standard deviation of temperature fluctuation in the k-th microclimate block Z, BTz(Zk): Based directly on the time series of air temperature Tz(Zk) in the k-th microclimate block Z within a set time window (such as 30 minutes or 1 hour), the standard deviation statistical method is used to evaluate the volatility and characterize the environmental stability characteristics of the microclimate zone. Solar disturbance response gradient of the k-th microclimate block Z By constructing a joint curve of the rate of change of solar intensity Gs(Zk) and air temperature Tz(Zk) in the k-th microclimate block Z, the significant change segment of solar intensity is extracted, and the ratio of the rate of change of solar intensity to the rate of temperature rise in this segment is calculated. The correlation is fitted by linear regression to obtain the slope of the response of the block to solar disturbance.

[0026] In this embodiment, the method addresses the issue within the dome-shaped exhibition hall where, due to its large, open space, complex structure, and mixed functional distribution, drastically different thermal responses in different areas under the same air supply strategy often occur, making it difficult for traditional centralized control systems to effectively regulate local conditions. To address this, the dome-shaped exhibition hall is divided into nine microclimate blocks Z with differences in physical characteristics, thermal capacity, and usage. Precise boundary delineation is achieved based on CFD simulation results, lighting models, and building structural boundaries, effectively preventing the spread of thermal disturbances to non-target areas. The microclimate block intelligent agent AgentZ deployed within each block not only possesses distributed data sensing capabilities but can also execute policy responses through partitioned execution, breaking the lag pattern of traditional systems' "centralized control, passive edge control." After deploying environmental acquisition devices in each microclimate block Z and connecting them to the intelligent agent, high-frequency sensing and acquisition of multi-dimensional environmental parameters such as temperature, humidity, wind speed, light intensity, and personnel density can be achieved. Especially in plant display areas or negotiation areas with frequent personnel movement or severe sunlight interference, the combined operation of infrared people sensors and light sensors can capture instantaneous disturbance changes, avoiding misjudgments in the control strategy. By locally structuring and encapsulating the collected data and tagging it with blocks before transmitting it to the central collaborative server, the subsequent model processing can be ensured to have block independence and traceability, solving the pain point of control failure caused by data loss or chaotic labeling in traditional systems. Further, preprocessing the environmental parameter information uploaded by each block in the central multi-agent collaborative server is the core of realizing "thermal inertia-based" model control. Introducing four key factors—heat capacity factor C (Zk), heat transfer coefficient U (Zk), temperature fluctuation standard deviation BTz (Zk), and solar disturbance response slope—can break away from the low dynamic limitations of traditional "material parameter tables" and instead establish a dynamic response model through a data-driven approach, which is particularly suitable for scenarios with complex structures and frequent changes in operating status, such as exhibition halls. For example, the estimation of the heat capacity factor is based on the time difference between air supply fluctuations and temperature response, truly reflecting the heat absorption and slow release capacity within the block, effectively identifying high-inertia areas and predicting the risk of control delays; similarly, the heat transfer coefficient is automatically modeled through a nighttime undisturbed window, avoiding the influence of interference factors such as personnel activity on the judgment results, making the results closer to the physical reality.

[0027] Example 3 Please see Figure 1 Specifically: S2 includes S21; S21. In the central multi-agent collaborative server, a thermodynamic feature model is constructed based on the thermal inertia standard dataset. The thermodynamic feature model constructs a thermodynamic feature model for microclimate block Z through a built-in thermal response modeling algorithm, which is used to describe the block's response capability to external disturbances and air conditioning regulation behavior. Then, the real-time acquired thermal inertia standard dataset is input into the thermodynamic feature model to calculate and output the thermal inertia index Hk of each microclimate block Z. The thermal inertia index Hk is calculated and output using the following thermodynamic characteristic model; ; In the formula, Hk(Zk) represents the thermal inertia index of the k-th microclimate block Z. This represents the average air temperature of the k-th microclimate block Z. Let Gs(Zk) represent the rate of change in solar radiation intensity Gs(Zk) of the k-th microclimate block Z. denoted as the response rate of the air temperature Tz(Zk) in the k-th microclimate block Z; a1 and a2 represent the preset empirical coefficients for the disturbance response term correction and stability correction term, respectively, with initial values ​​set to 0.6 and 0.4, respectively, derived from the regulation sensitivity calibration during the system debugging phase; fine-tuning will be performed in subsequent operation based on the effectiveness of the strategy and energy consumption indicators to form an adaptive optimization mechanism for the model. In this model, the thermodynamic characteristic model is constructed from the theory of thermal balance and thermal response delay in building physics. In this thermodynamic characteristic model, in order to adapt to the complex system of multiple microclimate blocks Z and the case of no clear structural area A, the parameter normalization modeling idea is adopted. The heat transfer coefficient factor U and the heat capacity factor C constitute the "heat transfer rate ratio per unit heat capacity" as an indicator to measure the response speed. Furthermore, the disturbance sensitivity term and the stability fluctuation term are introduced to form an extended thermal inertia index model. in, The thermal inertia index (Hk) represents the basic lag index, which indicates the heat transfer capacity per unit time under unit heat capacity. The larger the value, the faster the heat enters and exits, and the faster the response. The smaller the thermal inertia index (Hk), the lower the thermal inertia; the smaller the value, the higher the thermal inertia. This indicates the correction for the disturbance response term, reflecting the sensitivity of solar disturbances to temperature. A larger value indicates a more significant impact from changes in sunlight, which can easily cause local temperature rises. This term is added to the basic response term as a multiplicative term to adjust the overall thermal inertia index Hk, reflecting external instability. The stability correction term represents the stability ratio of the temperature time series. The larger B is, the larger the denominator and the larger the thermal inertia index Hk, reflecting stronger fluctuations and greater difficulty in control. The more stable the data, the smaller the value, and the better the thermal inertia can be controlled. The contribution ratios of the control disturbance term and the fluctuation term are determined using a1 and a2. The core idea of ​​this formula is to comprehensively express thermal inertia from three dimensions: Structural heat capacity dimension (C / U): a basic judgment of heat transfer and heat storage capacity, representing the structure's "inherent" response speed to thermal changes. The larger U is and the smaller C is, the faster the response (the larger H is). External disturbance response dimension (δGs / δTz): represents the temperature rise caused by a unit light disturbance, reflecting the response magnitude of the block to external energy input; Internal stability dimension (temperature mean / standard deviation): This indicates the temperature fluctuation stability of the block. If the standard deviation is large, the block is less stable to disturbances and the pre-adjustment weight should be increased. The combined significance: This formula provides a "response baseline" through structural thermal parameters, captures "environmental change drivers" through disturbance gain, and identifies "control sensitivity" through fluctuation characteristics. The combination of these three constitutes a comprehensive thermal inertia judgment index.

[0028] S2 also includes S22; S22. In the central multi-agent collaborative server, based on the statistical results of the thermal inertia index Hk of the same microclimate block Z within 24 historical operating cycles, the critical value between the normal and abnormal states of thermal response performance is selected as the thermal inertia threshold Hth. Next, the thermal inertia index Hk of each microclimate block Z, acquired in real time, is compared and evaluated with the thermal inertia threshold Hth to determine the risk of thermal response lag in microclimate block Z. Based on the preliminary comparison and evaluation results, a collaborative response mechanism is triggered. The specific evaluation content is as follows: When the thermal inertia index Hk > thermal inertia threshold Hth, it is determined that the current microclimate block Z has a risk of thermal response lag, the current microclimate block is marked as a target block to be coordinated and regulated, and the coordinated response mechanism is triggered. When the thermal inertia index Hk ≤ thermal inertia threshold Hth, the current microclimate block Z is determined to be in an autonomous and stable adjustment state, and the local agent of the current microclimate block continues to execute the conventional prediction and adjustment strategy.

[0029] In this embodiment, due to significant differences in structural materials, solar disturbance, and human heat load among the various microclimate blocks Z during actual operation and maintenance, even under the same air conditioning control strategy, their thermal response performance exhibits obvious lag differences. Without proper differentiation, this can easily lead to uneven heating or energy waste. Therefore, in the central multi-agent collaborative server, a thermodynamic feature model is first constructed based on the previously extracted thermal inertia standard dataset. By introducing three independent parameter dimensions—structural heat capacity ratio (C / U), solar disturbance response gradient, and temperature fluctuation standard deviation—the building's thermal response characteristics are mapped to a unified thermal inertia index Hk. This modeling approach departs from the traditional static evaluation path that relies solely on material properties or area structure, instead employing dynamic behavior modeling to realistically reflect the response sensitivity and adjustment difficulty of microclimate blocks Z during operation. For example, if a block Zk experiences drastic temperature changes under strong solar radiation, the disturbance term δGs / δTz increases, and the thermal inertia index Hk also rises accordingly, allowing for early detection of its "instability tendency" without waiting for the environmental deviation to actually occur. To establish judgment criteria, the system further extracts the critical value between normal and runaway thermal response states using historical Hk statistical data from 24 operating cycles, setting it as the thermal inertia threshold Hth. By comparing the current thermal inertia index Hk with the thermal inertia threshold Hth in real time, the system can classify the state of each block. If the thermal inertia index Hk > the thermal inertia threshold Hth, it is marked as a block requiring coordinated regulation; if the thermal inertia index Hk ≤ the thermal inertia threshold Hth, local predictive regulation is maintained. This mechanism can complete the judgment and marking before the problem occurs, avoiding large-scale regulatory imbalances due to response lag. Overall, this implementation scheme combines the theoretical foundation of thermal science with data-driven dynamic adaptability, significantly improving the accuracy and energy efficiency matching of microclimate regulation, and is particularly suitable for complex architectural scenarios with high glass ratios and large ceiling heights, such as domed exhibition halls.

[0030] Example 4 Please see Figure 1 Specifically: S3 also includes S31; S31. After triggering the collaborative response mechanism, a spatial thermal coupling map G is constructed in the central multi-agent collaborative server to identify the paths that need to be coordinated and buffered. The spatial thermal coupling map G is constructed by taking each microclimate block Z as a map node, establishing a connection edge between any two k-th microclimate blocks Zk and adjacent j-th microclimate blocks Zj that have a structural thermal conduction relationship, and determining the edge weight based on the actual building structure thermal simulation results and historical energy flow data, which serves as the structural coupling coefficient Y(Zk, Zj) between the k-th microclimate block Zk and adjacent j-th microclimate block Zj. The central server takes all microclimate blocks Z with thermal inertia index Hk higher than thermal inertia threshold Hth as the main control area, then traverses the set of directly adjacent blocks Nk in the spatial thermal coupling map G of the main control area, and identifies adjacent blocks whose thermal inertia index Hk(Zj) of adjacent microclimate block Zj is lower than the thermal inertia index Hk(Zk) of adjacent microclimate block Zj as cooperative buffers. Based on the thermal parameters of the building structure, the contact area A(Zk,Zj) and boundary thermal resistance R(Zk,Zj) of the boundary shared by the k-th microclimate block Zk and the adjacent j-th microclimate block Zj are extracted using the Building Information Modeling (BIM) system, and a basic heat conduction model is established accordingly. Y(Zk,Zj)(1)=1 / R(Zk,Zj)·A(Zk,Zj); This method demonstrates the static heat conduction capacity between the two blocks, providing a basic evaluation basis for the physical properties of the building. Secondly, based on historical environmental operation data, the air temperature Tz(Zk) of the k-th microclimate block Z and the air temperature Tz(Zj) of the adjacent j-th microclimate block Z within a typical operating cycle are selected from the central server over a period of nearly 7 days. The behavioral coupling degree between the temperature fluctuations of the two blocks is obtained through the Pearson correlation coefficient, and a dynamic correlation coefficient is constructed. Y(Zk,Zj)(2)=max Δt Corr(Tz(Zk)(t),Tz(Zj)(t+Δt)); where Δt represents the response lag time between the thermal disturbances of the two blocks, which is used to identify whether there is a time delay in the coupling relationship. This analysis can reveal the effectiveness of the energy transfer path in the actual control process. max represents the maximum value function, and Corr represents the Pearson correlation coefficient. Where Δt represents the response lag time between the thermal disturbances of the two blocks, it is used to identify whether there is a time delay in the coupling relationship; this analysis can reveal the effectiveness of the energy transfer path in the actual control process. Finally, the central multi-agent collaborative server weights and fuses the coupling strengths from the two sources to form a comprehensive coupling coefficient: Y(Zk,Zj)=α·Y(Zk,Zj)(1)+(1-α)·Y(Zk,Zj)(2); Where α∈[0,1] is the empirical weighting coefficient, which is set according to the building type and historical energy consumption control effect; the structural coupling coefficient Y(Zk,Zj) is ultimately used as the edge weight in the graph G, and is used as the basic input parameter for collaborative path generation, buffer block identification and subsequent valve opening adjustment strategy calculation.

[0031] S3 also includes S32; S32. The central multi-agent collaborative server sends a valve strategy adjustment command to the microclimate block agent Agent Z corresponding to the collaborative buffer microclimate block Z to control the valve actuator of the buffer to perform a preset opening adjustment, thereby guiding heat diffusion between regions and realizing active thermal disturbance buffering and adjustment between multiple blocks. The damper strategy adjustment command extracts the structural coupling coefficient Y(Zk, Zj) between the directly adjacent block set Nk and the k-th microclimate block Zk and the adjacent j-th microclimate block Zj, and calculates and outputs the damper collaborative adjustment value Fadj by combining the thermal inertia index Hk(Zj) of the adjacent j-th microclimate block Zj and the thermal inertia index Hk(Zk) of the k-th microclimate block Z. Then, the damper collaborative adjustment value Fadj is sent to the command response unit to control the damper actuator of the current collaborative buffer to perform preset opening adjustment, which is used to dynamically adjust the regional air supply volume and guide the main control area to thermal diffusion to the low inertia area, thereby achieving the purpose of equalizing the temperature response rate. The damper coordinated adjustment value Fadj is calculated and output using the following algorithm formula; ; In the formula, Fbase(Zk) represents the basic opening degree of the damper in the k-th microclimate block Z, and its value is dimensionless; Fadj(Zk) represents the coordinated adjustment value of the damper in the k-th microclimate block Z; and Href represents the precise value of the reference thermal inertia index, which is dimensionless. This represents an empirical coefficient used to control the intensity of regulation, with a value range of 0.1-0.3. This formula is an improved integration of thermal diffusion control mechanism and aerodynamic distribution strategy. Its physical core comes from the following two principles: thermal diffusion gradient driving mechanism: referring to the idea of ​​the law of heat conduction, that is, heat naturally flows from the high thermal inertia region to the low thermal inertia region; fluid regulation control strategy: using the HVAC valve system to guide the direction of thermal diffusion by adjusting the air flow. This formula is derived innovatively based on the aforementioned fundamental principles as follows: Establish a thermal inertia gradient function: (Hk(Zk)-Hk(Zj) / Href) reflects the degree of thermal inertia difference between the main control region and the buffer zone. The larger the value, the stronger the diffusion potential. Introducing structural coupling weighting: The structural coupling coefficient Y(Zk, Zj) between the k-th microclimate block Zk and the adjacent j-th microclimate block Zj characterizes the heat transfer capacity between the two blocks, avoiding ineffective air supply to weak connection paths; empirical coefficient The design features an adjustable amplification factor to adapt to different building types or energy consumption strategy requirements. The final adjustment value Fadj of the proportional damper is output by multiplying the damper base opening Fbase(Zk) of the kth microclimate block Z by the adjustment function, ensuring compatibility with the original system's damper control mechanism.

[0032] In this embodiment, during actual air conditioning operation, some microclimate blocks Z exhibit high thermal inertia and delayed adjustment response due to poor structural insulation or severe local heat load. Relying solely on the self-regulation of this area would lead to a continuous increase in energy consumption. Therefore, after triggering the collaborative response mechanism, a spatial thermal coupling map G is constructed through a central multi-agent collaborative server. Boundary thermal parameters extracted from the building BIM model (such as contact area A and thermal resistance R) are combined with historical temperature behavior coupling analysis (Pearson correlation coefficient and response lag) to obtain the structural coupling coefficient Y(Zk,Zj) between the two blocks. This coefficient serves as a path identification basis, effectively filtering out low-inertia adjacent blocks with strong heat transfer channels as collaborative buffer zones, avoiding mis-issuing commands to weakly connected areas and failing to achieve actual adjustment effects. Furthermore, based on the thermal inertia difference gradient (Hk(Zk)-Hk(Zj)) and the structural coupling strength Y(Zk,Zj), combined with the current main control area valve base opening Fbase(Zk), the system calculates and outputs the valve collaborative adjustment value Fadj. This value, acting as an adjustment increment, is sent to Agent Z, an intelligent agent controlling the damper actuator in the buffer microclimate block Z, to achieve precise airflow regulation. This guides heat from the main control area to the buffer zone, achieving "indirect cooling" of the main control area. This approach overcomes the limitations of traditional air conditioning's "zone-based self-control" strategy, constructing a cross-block coupled control mechanism that significantly improves the system's overall response speed and energy consumption control level under multi-heat source disturbance conditions. Especially in scenarios such as dome-shaped exhibition halls with high-altitude layering, complex glass structures, and heterogeneous heat load distribution, this collaborative strategy can effectively buffer the heat accumulation effect, prevent concentrated overload of cooling load, and ensure regional comfort and system economy.

[0033] Example 5 Please see Figure 1 Specifically: S4 includes S41; S41. The step of sending the valve collaborative adjustment value Fadj to the corresponding microclimate block agent AgentZ includes: the central multi-agent collaborative server encapsulates the valve collaborative adjustment value Fadj into a data instruction conforming to the Modbus protocol through the communication interface of the building automation system (BAS), and sends the data instruction to the corresponding microclimate block agent AgentZ instruction response unit according to the device address identifier of the microclimate block Z. Then, through the air valve controller in the command response unit, after receiving the air valve coordinated adjustment value Fadj, the air valve controller adjusts the operating parameters of the air supply equipment corresponding to the microclimate block Z in real time.

[0034] S4 also includes S42; S42. After the air valve controller of microclimate block Z performs environmental adjustment on the air supply equipment of its microclimate block Z by executing the air valve collaborative adjustment value Fadj, it restarts the sensing access unit of the microclimate block intelligent agent Agent Z after a preset 60 minutes, collects the adjusted environmental parameter information, inputs it into the thermodynamic characteristic model, outputs the thermal inertia index Hk, iterates and evaluates the thermal inertia, and if the risk of thermal response lag is still present after three adjustments, it triggers a manual warning through the central multi-agent collaborative server, and at the same time, manual inspection adjusts the microclimate block Z with the current risk of thermal response lag.

[0035] In this embodiment, in actual multi-block air conditioning control, if the valve collaborative adjustment value Fadj is only calculated and issued but not adapted to the communication protocol of the field equipment or there is no continuous monitoring after control, it will lead to control failure or unrecognizable control lag. Therefore, this implementation method uses the Building Automation System (BAS) downlink to encapsulate the valve collaborative adjustment value Fadj with the Modbus protocol to ensure command format compatibility, and accurately maps the device address identifier based on the microclimate block Z, enabling the command to be quickly and directionally transmitted to the command response unit of the corresponding Agent Z, ensuring the timeliness and accuracy of control. In addition, to evaluate the control effect and avoid lag and ineffective control caused by "thermal inertia", this solution sets up an automatic wake-up of the sensing module 60 minutes after the valve is executed, re-collects environmental parameters, and re-inputs the thermodynamic characteristic model Hk calculation value. If three consecutive adjustments are still determined to be in a state of thermal response lag, it indicates that there may be structural problems or local load changes in the area, and the system will automatically trigger a manual warning to notify the operation and maintenance personnel to conduct inspections and interventions. This "intelligent-human joint mechanism" effectively improves the closed-loop nature of the system and reduces the risk of "long-term inefficient operation" in high-thermal inertia areas. Especially in scenarios with high requirements for temperature and humidity stability, such as exhibition halls and exhibition areas, it can significantly ensure environmental quality and operational efficiency.

[0036] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A building and municipal facility operation and maintenance management method based on multi-agent collaborative decision-making, characterized in that: Includes the following steps: S1. Divide the interior of the dome exhibition hall into multiple microclimate blocks Z. Set up a microclimate block agent AgentZ in each microclimate block Z to collect environmental parameter information in real time. Upload the collected environmental parameter information to the central multi-agent collaborative server and perform preprocessing to obtain the thermal inertia standard dataset. S2. Construct a thermodynamic characteristic model based on the standard thermal inertia dataset, calculate the thermal inertia index Hk for each microclimate block Z, set the thermal inertia threshold Hth for preliminary comparative evaluation, and trigger a collaborative response mechanism based on the preliminary comparative evaluation results. S2 includes S21; S21. In the central multi-agent collaborative server, a thermodynamic feature model is constructed based on the thermal inertia standard dataset. The thermodynamic feature model is constructed for microclimate block Z through a built-in thermal response modeling algorithm. Then, the real-time acquired thermal inertia standard dataset is input into the thermodynamic feature model to calculate and output the thermal inertia index Hk of each microclimate block Z. The thermal inertia index Hk is calculated and output using the following thermodynamic characteristic model; ； In the formula, Hk(Zk) represents the thermal inertia index of the k-th microclimate block Z. This represents the average air temperature of the k-th microclimate block Z. Let Gs(Zk) represent the rate of change in solar radiation intensity Gs(Zk) of the k-th microclimate block Z. denoted as the response rate of the air temperature Tz(Zk) in the k-th microclimate block Z; a1 and a2 represent the preset empirical coefficients for the disturbance response term correction and the stability correction term, respectively, with initial values ​​set to 0.6 and 0.4, respectively. S3. After triggering the collaborative response mechanism, based on the building BIM model and the regional thermal coupling relationship, identify the set N of adjacent blocks of the high inertia block, and calculate the valve collaborative adjustment value Fadj. S3 further includes S32; S32. The central multi-agent collaborative server issues a valve strategy adjustment instruction to the microclimate block agent Agent Z corresponding to the microclimate block Z in the collaborative buffer. The valve strategy adjustment command extracts the set of directly adjacent blocks Nk and the structural coupling coefficient Y(Zk, Zj) between the kth microclimate block Zk and the adjacent jth microclimate block Zj. It then calculates and outputs the valve collaborative adjustment value Fadj by combining the thermal inertia index Hk(Zj) of the adjacent jth microclimate block Zj with the thermal inertia index Hk(Zk) of the kth microclimate block Zj. The valve collaborative adjustment value Fadj is then sent to the command response unit to control the valve actuator in the current collaborative buffer to perform a preset opening adjustment. The coordinated adjustment value Fadj of the air valve is calculated and output using the following algorithm formula; ； In the formula, Fbase(Zk) represents the basic opening degree of the damper in the k-th microclimate block Z, and its value is dimensionless; Fadj(Zk) represents the coordinated adjustment value of the damper in the k-th microclimate block Z; and Href represents the precise value of the reference thermal inertia index, which is dimensionless. This represents an empirical coefficient, with a value range of 0.1-0.3; S4. The coordinated adjustment value Fadj of the air valve is sent to the corresponding microclimate block Z intelligent agent through the building automation system interface. The microclimate block Z intelligent agent controls the air supply equipment in its microclimate block Z to perform environmental adjustment according to the received coordinated adjustment value Fadj of the air valve.

2. The building and municipal facility operation and maintenance management method based on multi-agent collaborative decision making according to claim 1, characterized in that: S1 includes S11; S11. Divide the interior of the dome exhibition hall into multiple microclimate blocks Z. The microclimate blocks Z are divided into 9 independent microclimate blocks Z by physical space division based on building lighting model, functional type, indoor heat capacity structure characteristics, air conditioning vent server distribution boundary and CAD structure and CFD airflow simulation results. Each independent microclimate block Z is labeled as the kth microclimate block Zk. An edge processing terminal is deployed in each microclimate block Z as a microclimate block agent (AgentZ). The microclimate block agent (AgentZ) includes a sensing access unit, a data transmission unit, and an instruction response unit.

3. The building and municipal facility operation and maintenance management method based on multi-agent collaborative decision making according to claim 1, characterized in that: S1 further includes S12; S12. Deploy environmental acquisition devices in each microclimate block Z to collect environmental parameter information in real time. Integrate the collected environmental parameter information into the corresponding microclimate block agent AgentZ through the sensing access unit. Then, label the environmental parameter information with the current k-th microclimate block Zk tag through the data transmission unit, perform local time serialization processing and unified structured encapsulation of data format to obtain the data encapsulation package, and then transmit the data encapsulation package to the central multi-agent collaborative server using the MQTT interface protocol. The environmental data acquisition equipment includes a temperature sensor, a humidity sensor, a wind speed sensor, a light sensor, and an infrared people sensor. The environmental parameter information includes the air temperature Tz (Zk) of the kth microclimate block Z, the relative humidity Sdz (Zk) of the kth microclimate block Z, the wind speed Vs (Zk) of the kth microclimate block Z, the solar radiation intensity Gs (Zk) of the kth microclimate block Z, and the population density Nocc (Zk) of the kth microclimate block Z.

4. The building and municipal facility operation and maintenance management method based on multi-agent collaborative decision making according to claim 3, characterized in that: S1 also includes S13; S13. In the central multi-agent collaborative server, the data package is received in real time, and the environmental parameter information is extracted by decapsulation. The data stream is partitioned and classified based on the Z label of each microclimate block. The environmental parameter information of each microclimate block Z is preprocessed to obtain the thermal inertia standard dataset. The preprocessing includes feature extraction and dimensionless processing; The feature extraction is performed by extracting features from the environmental parameter information after data cleaning to obtain a set of key factor parameters for thermal inertia modeling; The set of key factor parameters includes the heat capacity factor C (Zk) of the kth microclimate block Z, the heat transfer coefficient factor U (Zk) of the kth microclimate block Z, the temperature fluctuation standard deviation BTz (Zk) of the kth microclimate block Z, and the solar disturbance response gradient δGs (Zk) / δTz (Zk) of the kth microclimate block Z. The dimensionless processing integrates the key factor parameter set with environmental parameter information, converts them into dimensionless standard values ​​using the extreme value normalization method, and summarizes them to obtain a thermal inertia standard dataset.

5. The method for operation and maintenance management of building and municipal facilities based on multi-agent collaborative decision-making as described in claim 1, characterized in that: S2 further includes S22; S22. In the central multi-agent collaborative server, based on the statistical results of the thermal inertia index Hk of the same microclimate block Z within 24 historical operating cycles, the critical value between the normal and abnormal states of thermal response performance is selected as the thermal inertia threshold Hth. Next, the thermal inertia index Hk of each microclimate block Z, acquired in real time, is compared and evaluated with the thermal inertia threshold Hth to determine the risk of thermal response lag in microclimate block Z. Based on the preliminary comparison and evaluation results, a collaborative response mechanism is triggered. The specific evaluation content is as follows: When the thermal inertia index Hk > thermal inertia threshold Hth, it is determined that the current microclimate block Z has a risk of thermal response lag, the current microclimate block is marked as a target block to be coordinated and regulated, and the coordinated response mechanism is triggered. When the thermal inertia index Hk ≤ thermal inertia threshold Hth, the current microclimate block Z is determined to be in an autonomous and stable adjustment state, and the local agent of the current microclimate block continues to execute the conventional prediction and adjustment strategy.

6. The method for operation and maintenance management of building and municipal facilities based on multi-agent collaborative decision-making as described in claim 5, characterized in that: S3 also includes S31; S31. After triggering the collaborative response mechanism, a spatial thermal coupling map G is constructed in the central multi-agent collaborative server to identify the paths that need to be coordinated and buffered. The spatial thermal coupling map G is constructed by taking each microclimate block Z as a map node, establishing a connection edge between any two k-th microclimate blocks Zk with structural thermal conduction relationship and the adjacent j-th microclimate block Zj, and determining the edge weight based on the actual building structure thermal simulation results and historical energy flow data, which serves as the structural coupling coefficient Y(Zk, Zj) between the k-th microclimate block Zk and the adjacent j-th microclimate block Zj. The central server takes all microclimate blocks Z with thermal inertia index Hk higher than thermal inertia threshold Hth as the main control area, then traverses the set of directly adjacent blocks Nk in the spatial thermal coupling map G of the main control area, and identifies adjacent blocks whose thermal inertia index Hk(Zj) of adjacent microclimate block j is lower than the thermal inertia index Hk(Zk) of adjacent microclimate block Zj as cooperative buffers.

7. The building and municipal facility operation and maintenance management method based on multi-agent collaborative decision-making according to claim 1, characterized in that: S4 includes S41; S41. The step of sending the valve collaborative adjustment value Fadj to the corresponding microclimate block agent AgentZ includes: the central multi-agent collaborative server encapsulates the valve collaborative adjustment value Fadj into a data instruction conforming to the Modbus protocol through the communication interface of the building automation system (BAS), and sends the data instruction to the corresponding microclimate block agent AgentZ instruction response unit according to the device address identifier of the microclimate block Z. Then, through the air valve controller in the command response unit, the air valve controller adjusts the operating parameters of the air supply equipment corresponding to the microclimate block Z in real time after receiving the air valve coordinated adjustment value Fadj.

8. The building and municipal facility operation and maintenance management method based on multi-agent collaborative decision-making according to claim 7, characterized in that: S4 also includes S42; S42. After the air valve controller of microclimate block Z performs environmental adjustment on the air supply equipment of its microclimate block Z by executing the air valve collaborative adjustment value Fadj, it restarts the sensing access unit of the microclimate block intelligent agent Agent Z after a preset 60 minutes, collects the adjusted environmental parameter information, inputs it into the thermodynamic characteristic model, outputs the thermal inertia index Hk, iterates and evaluates the thermal inertia, and if the risk of thermal response lag is still present after three adjustments, it triggers a manual warning through the central multi-agent collaborative server, and at the same time, manual inspection adjusts the microclimate block Z with the current risk of thermal response lag.

Images