Central air conditioner safety monitoring method and system based on BIM

Abstract

The invention provides a BIM-based central air conditioner safety monitoring method and system. The method comprises the following steps: generating a BIM model of a building; according to the internal structural layout of the building in the BIM model and the arrangement of the central air conditioning equipment, carrying out initial safety area division on the building, carrying out dynamic adjustment based on a risk index of a time change regular term, and updating the risk index into the BIM model in real time; fault features are extracted from the real-time environment data and the equipment state data, and whether a fault exists or not is preliminarily judged; constructing a fault traceability model according to a fault positioning result, and determining a fault propagation path; a digital twinborn model is constructed for risk prediction, and a preventive maintenance plan is automatically generated; and designing an adaptive optimization algorithm based on reinforcement learning to maximize the operation efficiency of the system and minimize the risk level, and performing closed-loop feedback optimization. The invention provides a comprehensive, intelligent and efficient central air-conditioning system management scheme, and has wide application prospect and practical value.

Description

Technical Field

[0001] The present invention belongs to the technical field of central air-conditioning safety monitoring, and in particular relates to a central air-conditioning safety monitoring method and system based on BIM. Background Art

[0002] In modern buildings, central air conditioning systems, as a highly efficient air conditioning equipment, have been widely used in commercial, public and industrial buildings. With the expansion of building scale and the complexity of building functions, the design, installation, operation and maintenance of central air conditioning systems have become increasingly complex. In order to ensure the efficient and safe operation of central air conditioning systems, building information modeling (BIM) technology has been gradually introduced into the management of central air conditioning systems in recent years. BIM technology can provide a three-dimensional digital model of a building, including the structure, space, equipment and pipeline information of the building, which provides strong support for the visual management and intelligent monitoring of the central air conditioning system. However, the combination of existing BIM technology and central air conditioning systems still has certain limitations and shortcomings.

[0003] First, although the BIM model can manage the detailed space and equipment information of the building, its application in the safety monitoring and fault diagnosis of the central air-conditioning system is not deep enough. Traditional central air-conditioning monitoring systems mainly rely on independent sensor networks. These systems can only monitor local operating data, such as temperature, humidity and air quality, but it is difficult to fully understand the overall operating status of the system and its interaction with the building environment. This limitation makes it difficult for the system to quickly locate the source of the fault and provide effective solutions when facing complex problems. It often needs to rely on experienced technicians for manual troubleshooting and repair. In addition, most of the existing central air-conditioning monitoring systems lack dynamic correlation analysis of the overall environment of the building and the operation of the equipment, and cannot adjust the operating parameters of the air-conditioning system in real time according to changes in the internal and external environment of the building, thereby reducing the operating efficiency and energy-saving effect of the system.

[0004] Secondly, the current maintenance of central air-conditioning systems is mostly post-maintenance, that is, handling after a fault occurs, and lacks forward-looking preventive maintenance methods. This not only increases the maintenance cost of the system, but also easily causes unplanned downtime of the system, affecting the environmental comfort and user experience in the building. Although some emerging fault prediction technologies have begun to be applied to central air-conditioning systems, such as fault prediction models based on big data and machine learning, these technologies often ignore the spatial layout of the building and the equipment installation environment, resulting in limited accuracy and practicality of the prediction results.

[0005] In addition, traditional maintenance methods lack intelligent support for maintenance work, and maintenance personnel usually rely on paper or simple digital manuals for operations, making it difficult to efficiently locate faults and perform repair operations in complex building environments. Summary of the invention

[0006] The purpose of the present invention is to design a central air-conditioning safety monitoring method and system based on BIM. By deeply integrating the BIM model with real-time monitoring data, fault diagnosis technology, and maintenance guidance system, comprehensive intelligent management of the central air-conditioning system can be achieved, which can not only improve the operating efficiency and reliability of the system, but also significantly reduce maintenance costs and enhance the comfort and safety of the building environment.

[0007] In order to achieve the above object, a first aspect of the present invention provides a central air conditioning safety monitoring method based on BIM, the method comprising the following steps:

[0008] S1. Use BIM software to generate a BIM model of the building. The three-dimensional coordinates of the BIM model are marked as P (x, y, z), representing all spatial locations inside the building. At the same time, the central air-conditioning system is monitored in real time, and the real-time data from the sensor network is integrated into the BIM model. Each type of data corresponds to a specific dimension, and is dynamically mapped with the equipment information in space. An adaptive filtering model is designed to pre-process the data transmitted into the BIM model in real time. Finally, the pre-processed data is mapped back to the equipment location and attributes in the BIM model to form a multi-level data update mechanism.

[0009] S2. According to the structural layout of the building and the arrangement of the central air-conditioning equipment in the BIM model, the building is initially divided into safety areas, and the safety areas are monitored in real time. The real-time environmental data is mapped to each safety area, and the multi-level risk index of each safety area is calculated. At the same time, the boundary of each safety area is dynamically adjusted according to the risk index based on the time-varying regularization term calculated in real time, and updated to the BIM model in real time; wherein the time-varying regularization term is expressed as follows:

[0010]

[0011] Among them, E ik (t) represents the environment state vector E i The kth component of (t), β k represents the risk weight, λ represents the regularization coefficient, which is used to adjust the impact of time changes on the risk index, m represents the total number of features, j represents the jth feature, k represents the kth feature, and t represents the current time; the time change regularization term is used to capture the impact of dynamic changes in environmental parameters on safety according to the central air-conditioning system. It means to find partial derivatives;

[0012] S3. Extract fault features from real-time environmental data and equipment status data, and make a preliminary judgment on whether there is a fault. If it is preliminarily judged that there is a fault, it is marked as a potential fault device, and then the fault location and cause analysis are performed on the potential fault device. The system automatically generates maintenance suggestions based on the results of the cause analysis and sends the maintenance suggestions to the maintenance personnel. At the same time, according to the severity and impact range of different faults, the system will adjust the equipment status in the BIM model in real time and push the fault information to the maintenance personnel; wherein, the preliminary judgment on whether there is a fault is expressed as follows:

[0013] D i (t)=∥F i (t)-F ref ∥

[0014] Among them, F ref Represents the normal operation reference characteristic vector of the equipment; F i (t) represents the fault feature vector; if the deviation D i (t) exceeds a certain preset threshold D th , then mark the device as a potential fault device;

[0015] S4. Build a fault tracing model based on the fault location results, determine the fault propagation path, and analyze the root cause of the fault based on the fault propagation path. The system automatically generates targeted repair guidance plans based on the root cause.

[0016] S5. Build a digital twin model based on the BIM model, build a risk prediction model based on historical data and real-time data to predict the risk of the digital twin model, and then automatically generate a preventive maintenance plan based on the results of the risk prediction to ensure that maintenance and adjustments are carried out before the equipment fails. At the same time, the actual operation data and maintenance results are recorded in real time and fed back to the digital twin model and risk prediction model to form an intelligent feedback mechanism;

[0017] S6. Design an adaptive optimization algorithm based on reinforcement learning to maximize the operating efficiency of the system and minimize the risk level. At the same time, introduce a dynamic adjustment mechanism to construct a multi-objective optimization function, so that the multi-objective optimization function can be adjusted in real time according to specific needs. Then run the adaptive optimization algorithm to compare the actual operating results of the system with the predicted results, and feed back the comparison deviation to the adaptive optimization algorithm for closed-loop feedback optimization.

[0018] Preferably, the method further comprises designing an intelligent analysis module to generate a detailed analysis report based on the real-time data and optimization results of the system; the analysis report includes system performance, energy consumption analysis, risk assessment and optimization suggestions, and the formula is as follows:

[0019] R report (t) = Aanalyze (D twin (t),O opt (t),F multi (t))

[0020] Among them, A analyze Represents the analytical function, combined with the digital twin model D twin (t), optimization result O opt (t) and multi-objective function F multi (t) Conduct comprehensive analysis; report (t) indicates analysis report.

[0021] Preferably, the adaptive filtering model is expressed as follows:

[0022]

[0023] Among them, D clean (t) represents the processed data, D(t) represents the filtered data vector; D real (t) represents the original real-time data vector; R k (D(t)) represents the kth regularization term, which is used to suppress different types of noise; w k represents the weight of the regularization term, indicating the influence of each type of noise in the current environment; λ represents the regularization coefficient, which controls the intensity of filtering;

[0024] The preprocessed data is mapped back to the equipment location and attributes in the BIM model to form a multi-level data update mechanism, as shown below:

[0025] M(t)=M(t-1)+F(D clean (t))

[0026] Where M(t) represents the BIM model dataset at time t; M(t-1) represents the model dataset at the previous time; F(D clean (t)) represents the data mapping function, which maps the cleaned data to the corresponding device and spatial location.

[0027] Preferably, the S2 specifically includes:

[0028] S201, divide the building into initial safety areas according to the internal structural layout of the building and the arrangement of the central air conditioning equipment, where each area R i The initial boundary R i (0) is determined based on the impact range of the equipment and the physical partitioning of the building. The impact range of the equipment is estimated by a thermodynamic model and is expressed as follows:

[0029] R i(0) = {(x,y,z)|P(x,y,z)∈P dev +δ i}

[0030] Among them, P(x,y,z) represents a spatial point, P dev represents the device position vector, δ i Indicates the influence range offset of device i, which is used to expand or reduce the influence area;

[0031] S202, define the environmental parameters as a multi-level vector E i (t), which includes parameters at different levels and then mapped to each security region R i Then according to E i (t) Calculate the multi-level risk index I for each area i (t), expressed as follows:

[0032]

[0033] Among them, E ik (t) represents the environment state vector E i The kth component of (t), β k represents the risk weight, λ represents the regularization coefficient, which is used to adjust the impact of time changes on the risk index, n represents the total number of indicators, m represents the total number of features, j represents the jth feature, k represents the kth feature, t represents the current time, It means to find partial derivatives;

[0034] S203, each area R i The border The risk index I will be calculated in real time. i (t) Perform dynamic adjustments;

[0035] S204, the system recalculates the environmental state E of all safety areas every time step Δt i (t+Δt) and risk index I i (t+Δt), based on the new risk indicators, the system will dynamically adjust the regional boundaries And update it to the BIM model in real time.

[0036] Preferably, the S203 includes:

[0037] Each area R i The border The risk index I will be calculated in real time. i (t) Dynamic adjustment is performed if the risk index of a certain area exceeds the preset threshold I th , the system will automatically adjust the boundaries of the region to increase or decrease the area of ​​influence, as shown below:

[0038]

[0039] in, represents the gradient of the risk indicator, γ represents the boundary adjustment step size, η represents the boundary adjustment sensitivity coefficient, sign(I i (t)-I th ) is used to indicate whether the risk indicator exceeds the threshold.

[0040] Preferably, fault location and cause analysis are performed based on the results of the preliminary judgment, specifically including:

[0041] According to the spatial position P of the equipment in the BIM model dev and connectivity parameter C i , determine the path along which the fault may propagate and identify the associated device j;

[0042] Based on the data correlation between associated devices, calculate the failure impact G of device i i (t), expressed as follows:

[0043]

[0044] in, represents the set of devices associated with device i, ρ ij represents the correlation coefficient between devices i and j, indicating the possibility of fault propagation; F j (t) represents the fault characteristics of device j;

[0045] Combined fault impact G i (t) and fault characteristics F i (t), analyze the most likely cause of the fault through optimization solution C i (t), the cause analysis formula is as follows:

[0046]

[0047] Among them, λ is the regularization coefficient used to avoid overfitting, and C represents the total number of connectivity parameters.

[0048] Preferably, the S4 comprises:

[0049] S401. Construct a fault tracing model, determine the root cause of the fault through the fault tracing model, and analyze the propagation path of the fault in the entire system, as shown below:

[0050]

[0051] Among them, C i,j represents the connection strength between devices i and j, d i,jrepresents the physical distance between the two devices, τ represents the influence attenuation constant; F i (t) represents the fault characteristics, G i (t) represents the fault influence; P i,j (t) represents the possibility of fault propagation path;

[0052] S402. Analyze the root cause of the fault according to the fault propagation path, as shown below:

[0053]

[0054] Among them, R i (t) represents the fault cause vector, α is the regularization coefficient of the fault propagation path, and ∥R∥1 is the sparsity constraint term of the root cause vector, which ensures that the most concise explanation of the fault cause is found;

[0055] S403, according to the fault cause vector R i (t), the system will automatically generate targeted repair guidance plans;

[0056] S404. After the repair is completed, the system will compare the actual operation data with the preset repair plan to evaluate the effect of the repair. If the repair is incomplete or a new fault occurs, the system will automatically update the fault detection and tracing model.

[0057] Preferably, the S5 specifically includes:

[0058] S501. Build the initial digital twin model D twin (t=0), expressed as follows:

[0059]

[0060] Among them, M BlM (0) represents the initial BIM model, S i (0) represents the initial state information of device i, H i (0) represents the historical maintenance record of equipment i; at the same time, the initial digital twin model D twin (t=0) to perform dynamic calibration and obtain the dynamic digital twin model D twin (t), expressed as follows:

[0061] D twin (t) = D twin (t-1)+κ·(S real (t)-S pred (t))

[0062] Where κ is the calibration coefficient, S real (t) represents the device status collected in real time, S pred(t) represents the equipment status predicted based on the digital twin model;

[0063] S502. Combine historical data and real-time data to build a risk prediction model R that introduces time decay and seasonal adjustment life (t), expressed as follows:

[0064]

[0065] Among them, W i represents the weight of historical data, H i (t) represents the historical data of device i, C i (t) represents the state vector of the digital twin model, τ represents the time decay constant, and T represents the seasonal period;

[0066] S503. Automatically generate a preventive maintenance plan M based on the results of the risk prediction model. pm (t), and execute the optimization scheduling algorithm o opt (t), adjust the operating parameters of the system, expressed as follows:

[0067]

[0068] Among them, λ represents the regularization parameter, O prev Indicates the previous scheduling status;

[0069] S504. During the system operation and maintenance process, the actual operation data and maintenance results will be recorded in real time and fed back to the digital twin model and risk prediction model, forming an intelligent feedback mechanism to optimize the system.

[0070] Preferably, the adaptive optimization algorithm based on reinforcement learning is opt , which is expressed as follows:

[0071]

[0072] Where E(t) represents the operating efficiency, R(t) represents the risk level, O represents the set of system parameters to be optimized, and γ represents the coefficient for balancing the operating efficiency and the risk level;

[0073] A dynamic adjustment mechanism is introduced to design a multi-objective optimization function, so that the optimization objectives can be adjusted in real time according to specific needs. The multi-objective optimization function F multi (t) is expressed as follows:

[0074]

[0075] Among them, F multi (t) represents the multi-objective optimization function, G k (t) represents the kth objective function, βk (t) represents the weight coefficient adjusted in real time.

[0076] In a second aspect of the present invention, a central air conditioning safety monitoring system based on BIM is provided, the system comprising:

[0077] BIM model building module is used to generate the BIM model of the building using BIM software. The three-dimensional coordinates of the BIM model are marked as P (x, y, z), representing all the spatial positions inside the building. At the same time, it monitors the central air-conditioning system in real time and integrates the real-time data from the sensor network into the BIM model. Each type of data corresponds to a specific dimension and is dynamically mapped with the equipment information in space. An adaptive filtering model is designed to pre-process the data transmitted into the BIM model in real time. Finally, the pre-processed data is mapped back to the equipment position and attributes in the BIM model to form a multi-level data update mechanism;

[0078] The risk monitoring module is used to divide the building into initial safety zones according to the structural layout of the building and the arrangement of the central air-conditioning equipment in the BIM model, and to monitor the safety zones in real time, map the real-time environmental data to each safety zone, and calculate the multi-level risk index of each safety zone. At the same time, the boundary of each safety zone is dynamically adjusted according to the risk index based on the time-varying regularization term calculated in real time, and updated to the BIM model in real time; wherein the time-varying regularization term is expressed as follows:

[0079]

[0080] Among them, E ik (t) represents the environment state vector E i The kth component of (t), β k represents the risk weight, λ represents the regularization coefficient, which is used to adjust the impact of time changes on the risk index, m represents the total number of features, j represents the jth feature, k represents the kth feature, and t represents the current time; the time change regularization term is used to capture the impact of dynamic changes in environmental parameters on safety according to the central air-conditioning system. It means to find partial derivatives;

[0081] The fault analysis module is used to extract fault features from real-time environmental data and equipment status data, and preliminarily determine whether there is a fault. If it is preliminarily determined that there is a fault, it is marked as a potential fault device, and then the potential fault device is fault located and cause analyzed. The system automatically generates maintenance suggestions based on the results of the cause analysis and sends the maintenance suggestions to the maintenance personnel. At the same time, according to the severity and impact range of different faults, the system will adjust the equipment status in the BIM model in real time and push the fault information to the maintenance personnel; wherein, the preliminary judgment of whether there is a fault is expressed as follows:

[0082] D i (t)=∥F i (t)-F ref ∥

[0083] Among them, F ref Represents the normal operation reference characteristic vector of the equipment; F i (t) represents the fault feature vector; if the deviation D i (t) exceeds a certain preset threshold D th , then mark the device as a potential fault device;

[0084] The fault location module is used to build a fault tracing model based on the results of fault location, determine the fault propagation path, and analyze the root cause of the fault based on the fault propagation path. The system automatically generates targeted repair guidance plans based on the root cause;

[0085] The fault prediction module is used to build a digital twin model based on the BIM model, build a risk prediction model based on historical data and real-time data to predict the risk of the digital twin model, and then automatically generate a preventive maintenance plan based on the results of the risk prediction to ensure that maintenance and adjustments are carried out before the equipment fails. At the same time, the actual operation data and maintenance results are recorded in real time and fed back to the digital twin model and risk prediction model to form an intelligent feedback mechanism;

[0086] The system optimization module is used to design an adaptive optimization algorithm based on reinforcement learning to maximize the system's operating efficiency and minimize the risk level. At the same time, a dynamic adjustment mechanism is introduced to construct a multi-objective optimization function, so that the multi-objective optimization function can be adjusted in real time according to specific needs. The adaptive optimization algorithm is then run to compare the actual operating results of the system with the predicted results, and the comparison deviation is fed back to the adaptive optimization algorithm for closed-loop feedback optimization.

[0087] The beneficial technical effects of the present invention are at least as follows:

[0088] (1) The present invention deeply integrates the BIM model with the real-time operating data of the central air-conditioning system, and uses the building space information and environmental parameters in the BIM model to automatically divide and dynamically adjust the safety zone in the building. Based on real-time monitoring data and combined with intelligent algorithms, the system can continuously evaluate the operating status of the air-conditioning system, predict possible safety risks in real time, such as air quality deterioration, equipment overload or refrigerant leakage, and provide visual risk warnings. This innovation solves the problem that the existing system cannot fully monitor and adjust the safety zone in real time, and significantly improves the safety and responsiveness of the central air-conditioning system.

[0089] (2) The present invention introduces a fault diagnosis and tracing mechanism based on artificial intelligence. When an abnormality is detected in the central air-conditioning system, the system can accurately locate the fault location through the BIM model and intuitively display the source of the fault in the three-dimensional model. Based on historical data and real-time data, the AI ​​algorithm can analyze the cause and impact range of the fault, and automatically generate a detailed fault analysis report and repair suggestions. This innovation overcomes the problem of traditional systems relying on manual experience and inaccurate positioning in fault diagnosis, greatly improving the efficiency and accuracy of fault handling.

[0090] (3) The present invention combines BIM with digital twin technology to achieve full life cycle management of the central air-conditioning system. By constructing a digital twin model of the central air-conditioning system and deeply integrating it with the BIM model, the system can dynamically monitor and manage the entire life cycle of the central air-conditioning design, installation, operation and maintenance. The system can identify and predict possible risks in advance, and formulate preventive maintenance plans to reduce the unplanned downtime of the system. In addition, the intelligent repair guidance provided by the system, combined with AR / VR technology, helps maintenance personnel perform maintenance operations efficiently. This innovation solves the problem of the lack of forward-looking maintenance and intelligent support in existing systems, and improves the reliability and maintenance efficiency of central air-conditioning systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0091] The present invention is further described using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0092] Figure 1 The present invention is a flow chart of the central air-conditioning safety monitoring method based on BIM.

[0093] Figure 2 This is a framework diagram of the central air-conditioning safety monitoring system based on BIM of the present invention. DETAILED DESCRIPTION

[0094] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0095] In one or more embodiments, Figure 1 As shown, a central air conditioning safety monitoring method based on BIM is disclosed, comprising the following steps S1 to S6:

[0096] S1. Use BIM software to generate a BIM model of the building. The three-dimensional coordinates of the BIM model are marked as P (x, y, z), representing all spatial locations inside the building. At the same time, real-time monitoring of the central air-conditioning system is performed, and real-time data from the sensor network is integrated into the BIM model. Each type of data corresponds to a specific dimension, which is dynamically mapped with the equipment information in space. An adaptive filtering model is designed to pre-process the data transmitted into the BIM model in real time. Finally, the pre-processed data is mapped back to the equipment location and attributes in the BIM model to form a multi-level data update mechanism.

[0097] Specifically, first, use BIM software to generate a three-dimensional model of the building. The three-dimensional coordinates of the model are marked as P (x, y, z), representing all the spatial positions inside the building. In order to accurately capture the layout of the central air-conditioning equipment, the location of each device is represented by a point with multi-level attributes, defined as:

[0098] P dev ={(x i ,y i ,z i ),F i ,G i ,C i}

[0099] Among them, (x i ,y i ,z i ) is the spatial coordinate of the i-th device; F i It is the function set of the equipment, including the functional parameters of the equipment (such as cooling capacity, air volume, etc.); G i is the geometric characteristics of the equipment (such as size, shape, material, etc.); C i It is the connectivity parameter of the device, which describes the connection relationship between the device and other devices or spaces (such as pipeline connection, power supply path, etc.).

[0100] Based on the traditional BIM model, the present invention adds the function set and connectivity parameters of the equipment to facilitate the subsequent dynamic integration and analysis with real-time data. In particular, the connectivity parameter C i The introduction of can help to quickly analyze the mutual influence between equipment when a fault occurs, which is particularly critical in complex central air-conditioning systems.

[0101] Furthermore, in order to realize real-time monitoring of the central air-conditioning system, it is necessary to integrate the real-time data from the sensor network into the BIM model. The real-time data includes temperature T(t), humidity H(t), air quality index A(t), fan speed S(t), and air quality index A(t). f (t) and compressor current I c(t), etc. Each type of data corresponds to a specific dimension and is dynamically mapped with the device information in space. The mapping formula is:

[0102]

[0103] Where: D i (t) is the comprehensive real-time data vector of device i at time t; R j (t) is the raw data from the jth sensor; α j It is a weight coefficient related to the device function and is automatically adjusted by the model training process to ensure the accuracy of the mapping.

[0104] Traditional real-time data integration is usually directly superimposed on the BIM model, but in this solution, the present invention introduces a multi-dimensional dynamic mapping method to make the data integration of each device more flexible and accurate. In particular, by adjusting the weight coefficient α j ,The present invention can dynamically adapt to the changes of building usage conditions, which is of great significance for achieving real-time monitoring of the system.

[0105] Furthermore, an adaptive filtering model is designed:

[0106] Real-time data usually contains noise and outliers, so it must be preprocessed before integration into the BIM model. This paper designs an adaptive filtering model to eliminate noise and correct data in real time. The filtering model is based on an innovative weighted regularization term, defined as:

[0107]

[0108] Where D(t) is the filtered data vector; D real (t) is the original real-time data vector; R k (D(t)) is the kth regularization term, which is used to suppress different types of noise; w k is the weight of the regularization term, which indicates the influence of each type of noise in the current environment; λ is the regularization coefficient, which controls the intensity of filtering.

[0109] By adding a weighted regularization term, the filter can automatically adapt to environmental changes and effectively suppress various noises (such as random noise and periodic interference), ensuring that the cleaned data accurately reflects the true state of the system. The introduction of this adaptive filtering model is particularly suitable for processing complex environmental data in central air-conditioning systems, ensuring the accuracy of subsequent monitoring and analysis.

[0110] Furthermore, after preprocessing, the cleaned data D clean (t) will be mapped back to the equipment location and attributes in the BIM model, forming a multi-level data update mechanism. The mapping process is as follows:

[0111] M(t)=M(t-1)+F(D clean (t))

[0112] Among them, M(t) is the BIM model dataset at time t; M(t-1) is the model dataset at the previous time; F(D clean (t)) is a data mapping function that maps the cleaned data to the corresponding device and spatial location.

[0113] Through multi-level data mapping, the model can dynamically reflect the current state of the building and the operation of the central air-conditioning system. In particular, the mapping function F(D clean (t)) can be calculated based on the device connectivity parameter C i Adaptive adjustments are made so that the model can still maintain efficient and accurate dynamic update capabilities when facing complex architectural environments.

[0114] S2. According to the structural layout of the building and the arrangement of central air-conditioning equipment in the BIM model, the building is initially divided into safety areas, and the safety areas are monitored in real time. The real-time environmental data is mapped to each safety area, and the multi-level risk index of each safety area is calculated. At the same time, the boundary of each safety area is dynamically adjusted according to the risk index based on the time-varying regularization term calculated in real time, and updated to the BIM model in real time.

[0115] Specifically, in the management of the central air-conditioning system, in order to ensure the environmental safety and comfort of each area inside the building, each area must be dynamically divided and monitored in real time based on the BIM model and real-time data. This step directly relies on the BIM model and real-time data integration results constructed in step S1, deeply combining the spatial information of the building with the real-time environmental data to dynamically adjust and monitor the safety area.

[0116] First, based on the BIM model, the building is initially divided into safety zones according to the internal structural layout and the arrangement of central air-conditioning equipment. i The initial boundary R i (0) is based on the impact range of the equipment and the physical partitioning of the building. The impact range of the equipment can be estimated by thermodynamic model, the formula is:

[0117] R i (0) = {(x,y,z)|P(x,y,z)∈P dev +δ i}

[0118] Among them, P(x,y,z) is a spatial point, P dev is the device position vector, δ iThe influence range offset of device i, which is used to expand or reduce the influence area.

[0119] Based on the traditional regional division, the impact range offset δ is introduced i To accommodate the thermodynamic impact of different equipment, this allows the initial division to more accurately reflect the actual impact range of the equipment, especially in buildings with complex air flows.

[0120] Furthermore, based on the real-time data in step 1, the environmental parameters (such as temperature T(t), humidity H(t), air quality index A(t), wind speed V(t)) are mapped to each safety area R i In order to more accurately reflect the environmental status in the region, the present invention defines the environmental status of each region as a multi-level vector E i (t), which includes parameters at different levels, such as local temperature gradients Humidity change rate Etc. The formula is as follows:

[0121]

[0122] Among them, E i (t) represents the region R i The state vector of the multi-level environment at time t.

[0123] Then, according to E i (t) Calculate the multi-level risk index I for each area i (t), this indicator introduces a regularization term to balance the influence of different environmental parameters. The formula is:

[0124]

[0125] Among them, E ik (t) is the environment state vector E i The kth component of (t), β k is the risk weight, and λ is the regularization coefficient, which is used to adjust the impact of time changes on risk indicators.

[0126] Compared with the traditional risk assessment method, this formula introduces a time-varying regularization term It is used to capture the impact of dynamic changes in environmental parameters on safety. This design is particularly suitable for scenarios where the environmental status in central air-conditioning systems changes rapidly, and can more accurately reflect the potential safety risks in the area.

[0127] Furthermore, each region R i The border The risk index I will be calculated in real time. i(t) Dynamic adjustment. If the risk index of a certain area exceeds the preset threshold I th , the system will automatically adjust the boundaries of the area to increase or decrease the area's influence. The specific adjustment formula is as follows:

[0128]

[0129] in, is the gradient of the risk indicator, γ is the boundary adjustment step size, η is the boundary adjustment sensitivity coefficient, sign(I i (t)-I th ) is used to indicate whether the risk indicator exceeds the threshold.

[0130] Furthermore, the boundary adjustment sensitivity coefficient η is introduced. This formula can dynamically adjust the amplitude of boundary changes according to the degree of risk index exceeding the standard, so that the system can more flexibly respond to changes in the complex environment inside the building. This design can provide more precise regional boundary control when the risk index is close to or far exceeds the threshold.

[0131] Furthermore, in order to ensure the safety status of each area inside the building, the system will periodically (every time step Δt) recalculate the environmental status E of all safety areas i (t+Δt) and risk index I i (t+Δt). Based on the new risk indicators, the system will dynamically adjust the regional boundaries The BIM model will be updated in real time. The updated BIM model will be used to guide further adjustments to the central air conditioning system.

[0132] By combining the calculation of risk indicators and boundary adjustment mechanisms with the dynamic update of the BIM model, the system can achieve continuous monitoring and safety management of various areas in the building. This circulation mechanism ensures that the central air-conditioning system can always maintain efficient and safe operation under different environmental conditions.

[0133] S3. Extract fault features from real-time environmental data and equipment status data, and make a preliminary judgment on whether there is a fault. If it is preliminarily judged that there is a fault, it will be marked as a potential fault device. Then, the fault location and cause analysis will be performed on the potential fault device. The system will automatically generate maintenance suggestions based on the results of the cause analysis and send the maintenance suggestions to the maintenance personnel. At the same time, according to the severity and impact range of different faults, the system will adjust the equipment status in the BIM model in real time and push the fault information to the maintenance personnel.

[0134] In the central air conditioning system, timely and accurate detection and diagnosis of faults are the key to ensuring the efficient and stable operation of the system. In the first two steps, the present invention has achieved comprehensive monitoring of the building environment and air conditioning system through BIM model construction and dynamic safety zone division. Next, in the fault detection and diagnosis step, the present invention will use this basic information to detect faults in real time through intelligent algorithms and accurately diagnose their causes and locations. The following are detailed solution steps.

[0135] First, based on real-time environmental data E i (t) and equipment status data S i (t) (such as fan speed, compressor current, etc.), extract fault features. For each device i, define a feature vector F i (t), which consists of multiple fault-sensitive indicators, including but not limited to temperature anomaly ΔT i (t), pressure fluctuation ΔP i (t) and current anomaly ΔI i (t), etc. Specifically expressed as:

[0136] F i (t) = [ΔT i (t),ΔP i (t),ΔI i (t),…]

[0137] Where, ΔT i (t) = T i (t)-T ref , T ref is the reference temperature of device i; ΔP i (t) = P i (t)-P ref , P ref is the reference pressure; ΔI i (t)Similarly defined.

[0138] By calculating the fault feature vector F i The deviation of (t) can be used to preliminarily determine whether there is a fault. The deviation calculation formula is:

[0139] D i (t)=∥F i (t)-F ref ∥

[0140] Among them, F ref is the reference characteristic vector for the normal operation of the equipment. If D i (t) exceeds a certain preset threshold D th , the device is marked as a potential fault device.

[0141] By defining specific fault feature vectors and deviation calculations, this solution can identify potential faults at an early stage and avoid omissions or false alarms caused by simple threshold judgments. The composition of the feature vector fully considers the various common fault modes in central air-conditioning systems, making detection more comprehensive and accurate.

[0142] Furthermore, for the detected potential faulty equipment, further fault location and cause analysis are performed. First, according to the spatial position P of the equipment in the BIM model, dev and connectivity parameter C i , determine the path along which the fault may propagate and identify the associated device j. Then, based on the data correlation between the associated devices, calculate the fault impact G of device i i (t):

[0143]

[0144] in, is the set of devices associated with device i, ρ ij is the correlation coefficient between devices i and j, indicating the possibility of fault propagation.

[0145] Combined with influence G i (t) and fault characteristics F i (t), analyze the most likely cause of the fault through optimization solution C i (t). The cause analysis formula is:

[0146]

[0147] Among them, λ is the regularization coefficient used to avoid overfitting, and C represents the total number of connectivity parameters.

[0148] According to the fault impact G i The introduction of (t) enables the scheme to consider the propagation effect of faults between devices, especially in complex central air-conditioning systems. This correlation analysis helps to locate the source of faults more accurately. The regularization term introduced in the cause analysis formula can also effectively improve the robustness of fault diagnosis and prevent the diagnosis model from overfitting to noisy data.

[0149] Furthermore, after completing the fault diagnosis, the system will i (t) Automatically generate maintenance suggestions, including but not limited to adjusting operating parameters, shutting down faulty equipment, starting backup systems, etc. Depending on the severity and impact of different faults, the system will also adjust the equipment status in the BIM model in real time and push fault information to maintenance personnel.

[0150] At the same time, the fault information F i (t) and the diagnosis result C i(t) will be recorded in the system database and used to update and train the fault detection model to improve the accuracy of future detection and diagnosis. The feedback update mechanism formula is:

[0151] W new =W old +η·ΔW

[0152] Among them, W is the weight parameter of the fault detection model, η is the learning rate, ΔW is the weight update amount, and the model is adjusted based on the latest fault diagnosis results.

[0153] Through intelligent fault decision-making, the system can automatically generate targeted maintenance plans, reducing the time and error of human judgment. At the same time, the feedback update mechanism based on fault diagnosis results ensures the continuous optimization and improvement of the detection model, enabling the system to maintain efficient fault identification capabilities in a constantly changing environment.

[0154] This intelligent fault detection and diagnosis solution achieves accurate detection and diagnosis of faults in central air-conditioning systems through a complete process from fault feature extraction, location and cause analysis to intelligent decision-making and feedback update. Each step is interconnected to ensure that the system can quickly respond to and handle various fault conditions in complex building environments, providing key fault management support for the entire patented solution.

[0155] S4. Build a fault tracing model based on the results of fault location, determine the fault propagation path, and analyze the root cause of the fault based on the fault propagation path. The system automatically generates targeted repair guidance plans based on the root cause.

[0156] In step S3, the present invention has detected and preliminarily located the fault point. Now, the present invention needs to determine the root cause of the fault through the fault tracing model and analyze its propagation path in the entire system. Assume that the fault feature is F i (t), the fault influence is G i (t), then the probability of the fault propagation path P o,j (t) can be expressed as:

[0157]

[0158] Among them, C i,j represents the connection strength between devices i and j, d i,j is the physical distance between the two devices, and τ is the influence attenuation constant.

[0159] By calculating the cumulative propagation probability P of all possible paths i,j(t), the system can determine the most likely fault propagation path and identify other potentially affected devices. By introducing the connection strength and physical distance between devices, the fault tracing model can more accurately evaluate the fault propagation path in the central air-conditioning system. This method is particularly suitable for complex large-scale air-conditioning systems and can effectively identify cascading faults in the system.

[0160] Furthermore, after determining the fault propagation path, the root cause analysis of the fault is further performed. Assume that the fault cause vector is R i (t), which is related to the fault propagation path P i,j (t) and fault characteristics F i (t) can be expressed as:

[0161]

[0162] Among them, α is the regularization coefficient of the fault propagation path, and ∥R∥1 is the sparsity constraint term of the root cause vector, which ensures that the most concise explanation of the fault cause is found.

[0163] By solving this optimization problem, the system can extract the main cause of the failure and provide a direct basis for subsequent repair work. The model introduces sparsity constraints, which can effectively extract the most fundamental cause behind complex fault phenomena, thereby avoiding overly complex explanations and unnecessary repair operations.

[0164] Furthermore, according to the root cause of the fault R i (t), the system will automatically generate a targeted repair guidance plan. The repair plan includes necessary operation steps, required tools, safety precautions, etc., and will provide specific operation locations and path planning in combination with the spatial information in the BIM model. Let the repair guidance vector be H i (t), its generation process is:

[0165] H i (t) = M(R i (t))+G(P i,j (t))

[0166] Among them, M is the repair operation mapping function, and G is the path planning function, which is calculated based on the equipment location and the spatial layout in the BIM model.

[0167] The system can also use AR (augmented reality) devices to overlay repair instructions on the real view, helping maintenance personnel complete repair tasks more intuitively.

[0168] By combining the cause of the fault with the BIM model, the intelligent repair guidance solution can provide maintenance personnel with high-precision and efficient operation suggestions. At the same time, the use of AR equipment makes the repair process more intuitive and convenient, reducing the complexity of operation and the error rate.

[0169] Furthermore, after the repair is completed, the system will compare the actual operation data with the preset repair plan to evaluate the repair effect. If the repair is incomplete or a new fault occurs, the system will automatically update the fault detection and tracing model for further analysis and adjustment. Let the repair error vector be E r (t), its update process is:

[0170] W new =W old +β·E r (t)

[0171] Among them, β is the learning rate of repair feedback, W is the weight parameter of the system model, and the updated model will be used for future fault detection and diagnosis. Through real-time feedback and model update mechanism, the system can continuously improve and optimize its own detection and repair capabilities after each fault handling, ensuring the long-term stable operation of the system.

[0172] This fault tracing and intelligent repair guidance solution achieves accurate tracing and efficient repair of faults in central air-conditioning systems through a complete process from fault propagation path analysis, root cause extraction to intelligent repair guidance generation and feedback update. Each step is interconnected to ensure that the system can respond and handle faults quickly in complex environments, providing key intelligent repair support for the entire patented solution.

[0173] S5. Build a digital twin model based on the BIM model, build a risk prediction model based on historical data and real-time data to predict the risks of the digital twin model, and then automatically generate a preventive maintenance plan based on the results of the risk prediction to ensure that maintenance and adjustments are performed before the equipment fails. At the same time, the actual operation data and maintenance results are recorded in real time, and fed back to the digital twin model and risk prediction model to form an intelligent feedback mechanism.

[0174] Specifically, in the management of central air-conditioning systems, it is crucial to ensure the safety and stability of the system throughout its life cycle. Full life cycle management covers not only the design, installation, and commissioning of the system, but also operation, maintenance, upgrades, and final retirement. This step will be based on the previously constructed BIM model, digital twin model, and real-time data and fault diagnosis results of the previous steps, and through innovative methods, intelligent and safe management of the entire life cycle of the central air-conditioning system will be achieved.

[0175] In step S1, the present invention has built a basic system model using BIM model and real-time data. In order to achieve full life cycle management, it is first necessary to build a dynamic digital twin model D twin (t). This model not only includes the spatial and equipment information of the building, but also combines the operating status of the equipment, historical maintenance records, fault data, etc. The initial digital twin model is defined as:

[0176]

[0177] Among them, M BlM (0) is the initial BIM model, S i (0) is the initial state information of device i, H i (0) is the historical maintenance record of equipment i. This model will be continuously updated and optimized throughout the life cycle of the system.

[0178] The digital twin model is not only a mirror of the physical system, but also needs to be dynamically calibrated to reflect the operating status and changes of the actual system. The calibration formula is:

[0179] D twin (t) = D twin (t-1)+κ·(S real (t)-S pred (t))

[0180] Where κ is the calibration coefficient, S real (t) is the device status collected in real time, S pred (t) is the device state predicted based on the model. This calibration mechanism ensures that the digital twin model is always consistent with the physical system.

[0181] Through dynamic calibration, the digital twin model can reflect the actual status of the central air conditioning system in real time and can self-correct according to the actual operating data. This real-time synchronization and calibration capability is essential for long-term monitoring and management of the system, especially when the system is aging or the environmental conditions change.

[0182] Furthermore, based on the digital twin model, the present invention uses the historical data and real-time data of the system to predict risks throughout the entire life cycle. Risk prediction includes not only the current operating risks, but also the potential risks in long-term operation. To achieve this goal, the present invention introduces a risk prediction model R with time decay and seasonal adjustment. life (t), defined as:

[0183]

[0184] Among them, W i is the weight of historical data, H i(t) is the historical data of device i, C i (t) is the state vector of the digital twin model, τ is the time decay constant, and T is the seasonal period.

[0185] By introducing the time decay term and seasonal adjustments The risk prediction model can simultaneously consider the aging effect of equipment and the impact of environmental changes on system safety. This method is particularly suitable for central air-conditioning systems that run for a long time and can identify possible long-term risks in advance.

[0186] Furthermore, based on the risk prediction results, the system automatically generates a preventive maintenance plan M pm (t) to ensure effective maintenance and adjustment before equipment failure. The preventive maintenance plan not only considers the current risk level, but also combines the service life of the equipment and available resources. The specific formula is as follows:

[0187] M pm (t) = A pm (R life (t),L i (t),R avail )

[0188] Among them, A pm Generates the function for the maintenance plan, L i (t) is the service life vector of equipment i, R avail is the available resource vector.

[0189] At the same time, the system will execute the optimization scheduling algorithm opt (t) to adjust the system's operating parameters to maximize its operating efficiency and safety. The scheduling optimization formula is:

[0190]

[0191] Among them, λ is the regularization parameter, O prev For the previous scheduling status, ensure the continuity and stability of scheduling optimization.

[0192] This step generates a more reasonable and targeted preventive maintenance plan by combining equipment life and available resources. At the same time, the scheduling optimization algorithm takes into account operational efficiency and safety to ensure that the system can maintain efficient operation with minimal risk.

[0193] Furthermore, during the system operation and maintenance process, the actual operation data and maintenance results will be recorded in real time and fed back to the digital twin model and risk prediction model to form an intelligent feedback mechanism. The feedback process can be described as:

[0194] D twin,new(t+1)=D twin (t)+α·F b (t)

[0195] Among them, α is the feedback update coefficient, F b (t) is the feedback vector, which contains the error and deviation between the actual operation and the prediction.

[0196] The system continuously optimizes and adjusts the prediction model and maintenance strategy by analyzing the feedback data to adapt to the actual operation of the system. This process ensures that the system can continuously improve its safety and stability throughout its life cycle.

[0197] The intelligent feedback mechanism ensures that the system can self-optimize and evolve during operation. By adjusting the digital twin model and risk prediction model in real time, the system can promptly adapt to environmental changes and the impact of system aging, and always maintain the best state.

[0198] S6. Design an adaptive optimization algorithm based on reinforcement learning to maximize the operating efficiency of the system and minimize the risk level. At the same time, introduce a dynamic adjustment mechanism to construct a multi-objective optimization function, so that the multi-objective optimization function can be adjusted in real time according to specific needs. Then run the adaptive optimization algorithm to compare the actual operating results of the system with the predicted results, and feed back the comparison deviation to the adaptive optimization algorithm for closed-loop feedback optimization.

[0199] Specifically, intelligent optimization and continuous improvement are key steps to ensure that the central air-conditioning system continues to operate efficiently and safely throughout its life cycle. In the previous steps, the present invention has established a strong foundation through digital twin models, full life cycle management, real-time monitoring and fault diagnosis. This step will build on the previous achievements and further achieve continuous improvement and performance enhancement of the system through intelligent optimization algorithms and adaptive mechanisms. The following are detailed solution steps.

[0200] In order to maintain the optimal operating state of the system in a dynamically changing environment, the present invention designs an adaptive optimization algorithm based on reinforcement learning. opt (t). The algorithm can continuously learn and optimize the parameter settings of the central air-conditioning system in actual operation. The optimization goal is to maximize the system's operating efficiency E(t) and minimize the risk level R(t). The optimization formula is as follows:

[0201]

[0202] Among them, O is the set of system parameters to be optimized, and γ represents the coefficient for balancing operating efficiency and risk level.

[0203] The system continuously adjusts and updates the parameters O based on real-time feedback data, and through the reinforcement learning mechanism, the algorithm can gradually converge to the optimal solution. opt (t) will be directly applied to the operation control of the system.

[0204] By introducing a reinforcement learning mechanism, the optimization algorithm can continuously learn and improve itself in a changing environment, ensuring that the system can maintain efficient and safe operation at different times. This algorithm is particularly suitable for central air conditioning systems because it can adaptively adjust parameters to cope with seasonal changes, equipment aging and other factors.

[0205] Furthermore, in actual operation, the optimization target of the system may change with changes in external conditions. Therefore, in addition to the standard operating efficiency E(t) and risk level R(t), the present invention introduces a dynamic adjustment mechanism so that the optimization target can be adjusted in real time according to specific needs. The dynamic adjustment formula of the multi-objective function is:

[0206]

[0207] Among them, F multi (t) is the comprehensive objective function, G k (t) is the kth objective function (such as energy consumption, comfort, maintenance cost, etc.), β k (t) is the weight coefficient adjusted in real time.

[0208] Through the design of this multi-objective function, the system can dynamically adjust the priority of optimization objectives under different operating stages or environmental conditions. For example, during high-load operation in summer, the system can increase the weight of comfort, while in winter it may focus more on energy consumption optimization.

[0209] Furthermore, in order to ensure that the optimization effect can be maintained for a long time, the present invention introduces a closed-loop feedback mechanism to compare the actual operation results of the system with the predictions and feed back the deviations to the optimization algorithm. The feedback formula is as follows:

[0210] O opt,new (t+1)=0 opt (t)+α·(O real (t)-O pred (t))

[0211] Among them, α is the feedback adjustment coefficient, O real (t) is the actual operating parameter of the system, O pred (t) is the optimized parameter for prediction.

[0212] Through this feedback mechanism, the system can continue to learn and improve, gradually reducing the deviation between actual operation and prediction.opt,new (t+1) will be used for the next cycle of operation, forming a closed loop of continuous improvement.

[0213] Furthermore, in order to help managers understand the operating status and optimization effect of the system, the present invention designs an intelligent analysis module, which can generate a detailed analysis report based on the real-time data and optimization results of the system. The analysis report includes system performance, energy consumption analysis, risk assessment, and optimization suggestions, and the formula is expressed as:

[0214] R report (t) = A analyze (D twin (t),O opt (t),F multi (t))

[0215] Among them, A analyze To analyze the function, a comprehensive analysis is performed by combining the digital twin model, optimization results and multi-objective functions.

[0216] The reports will be presented in a graphical form and provide specific optimization suggestions to help managers make decisions. These reports are not only used for daily management, but also can be used as a reference for future system improvements and upgrades.

[0217] In one or more embodiments, Figure 2 As shown, a BIM-based central air conditioning safety monitoring system is disclosed, and the system includes:

[0218] The BIM model construction module 101 is used to generate a BIM model of a building using BIM software. The three-dimensional coordinates of the BIM model are marked as P (x, y, z), representing all spatial positions inside the building. At the same time, the central air-conditioning system is monitored in real time, and real-time data from the sensor network is integrated into the BIM model. Each type of data corresponds to a specific dimension, and is dynamically mapped with the equipment information in space. An adaptive filtering model is designed to pre-process the data transmitted into the BIM model in real time. Finally, the pre-processed data is mapped back to the equipment position and attributes in the BIM model to form a multi-level data update mechanism;

[0219] The risk detection module 102 is used to divide the building into initial safety zones according to the structural layout of the building and the arrangement of the central air-conditioning equipment in the BIM model, and to monitor the safety zones in real time, map the real-time environmental data to each safety zone, and calculate the multi-level risk index of each safety zone. At the same time, the boundary of each safety zone is dynamically adjusted according to the risk index based on the time-varying regularization term calculated in real time, and updated to the BIM model in real time; wherein the time-varying regularization term is expressed as follows:

[0220]

[0221] Among them, E ik (t) represents the environment state vector E i The kth component of (t), β k represents the risk weight, λ represents the regularization coefficient, which is used to adjust the impact of time changes on the risk index, m represents the total number of features, j represents the jth feature, k represents the kth feature, t represents the current time, represents partial derivative; the time-varying regularization term is used to capture the impact of dynamic changes of environmental parameters on safety according to the central air-conditioning system;

[0222] The fault analysis module 103 is used to extract fault features from real-time environmental data and equipment status data, and preliminarily determine whether there is a fault. If it is preliminarily determined that there is a fault, it is marked as a potential fault device, and then the potential fault device is fault located and cause analyzed. The system automatically generates maintenance suggestions based on the results of the cause analysis and sends the maintenance suggestions to the maintenance personnel. At the same time, according to the severity and impact range of different faults, the system will adjust the equipment status in the BIM model in real time and push the fault information to the maintenance personnel; wherein, the preliminary determination of whether there is a fault is expressed as follows:

[0223] D i (t)=∥F i (t)-F ref ∥

[0224] Among them, F ref Represents the normal operation reference characteristic vector of the equipment; F i (t) represents the fault feature vector; if the deviation D l (t) exceeds a certain preset threshold D th , then mark the device as a potential fault device;

[0225] The fault location module 104 is used to build a fault tracing model based on the result of fault location, determine the fault propagation path, and analyze the root cause of the fault based on the fault propagation path. The system automatically generates a targeted repair guidance plan based on the root cause;

[0226] The fault prediction module 105 is used to build a digital twin model based on the BIM model, build a risk prediction model based on historical data and real-time data to predict the risk of the digital twin model, and then automatically generate a preventive maintenance plan based on the results of the risk prediction to ensure that maintenance and adjustment are performed before the equipment fails. At the same time, the actual operation data and maintenance results are recorded in real time and fed back to the digital twin model and the risk prediction model to form an intelligent feedback mechanism;

[0227] The system optimization module 106 is used to design an adaptive optimization algorithm based on reinforcement learning to maximize the operating efficiency of the system and minimize the risk level. At the same time, a dynamic adjustment mechanism is introduced to construct a multi-objective optimization function, so that the multi-objective optimization function can be adjusted in real time according to specific needs. The actual operating results of the system are run by the adaptive optimization algorithm to compare with the predicted results, and the comparison deviation is fed back to the adaptive optimization algorithm for closed-loop feedback optimization.

[0228] It is worth noting that the specific workflow of the BIM-based central air-conditioning safety monitoring system provided in the embodiment of the present invention is the same as the workflow of the BIM-based central air-conditioning safety monitoring method described in the above embodiment, and will not be repeated here.

[0229] Compared with the prior art, the BIM-based central air-conditioning safety monitoring system provided in the embodiment of the present invention uses BIM software to generate a BIM model of a building. The three-dimensional coordinates of the BIM model are marked as P (x, y, z), representing all spatial positions inside the building. At the same time, the central air-conditioning system is monitored in real time, and real-time data from the sensor network is integrated into the BIM model. Each type of data corresponds to a specific dimension, and is dynamically mapped with the equipment information in space. An adaptive filtering model is designed to pre-process the data transmitted into the BIM model in real time. Finally, the pre-processed data is mapped back to the equipment position and attributes in the BIM model, forming a multi-level data update mechanism. According to the structural layout of the building and the arrangement of central air-conditioning equipment in the BIM model, the building is initially divided into safety areas, and the safety areas are monitored in real time. The real-time environmental data is mapped to each safety area, and the multi-level risk index of each safety area is calculated. At the same time, the boundary of each safety area is dynamically adjusted according to the risk index based on the time-varying regularization term calculated in real time, and updated to the BIM model in real time; the time-varying regularization term is used to capture the impact of dynamic changes in environmental parameters on safety according to the central air-conditioning system; fault features are extracted from real-time environmental data and equipment status data, and a preliminary judgment is made as to whether a fault exists. If a fault is preliminarily judged to exist, it is marked as a potential fault. In the faulty equipment, the system then locates the fault and analyzes the cause of the potential faulty equipment. The system automatically generates maintenance suggestions based on the results of the cause analysis and sends the maintenance suggestions to the maintenance personnel. At the same time, according to the severity and impact range of different faults, the system will adjust the equipment status in the BIM model in real time and push the fault information to the maintenance personnel; build a fault tracing model based on the results of fault location, determine the fault propagation path, and analyze the root cause of the fault based on the fault propagation path. The system automatically generates targeted repair guidance plans based on the root cause; build a digital twin model based on the BIM model, and build a risk prediction model based on historical data and real-time data to conduct risk assessment on the digital twin model. Risk prediction is then performed, and then a preventive maintenance plan is automatically generated based on the results of risk prediction to ensure that maintenance and adjustments are performed before equipment failures occur. At the same time, the actual operating data and maintenance results are recorded in real time and fed back to the digital twin model and risk prediction model to form an intelligent feedback mechanism. An adaptive optimization algorithm based on reinforcement learning is designed to maximize the operating efficiency of the system and minimize the risk level. At the same time, a dynamic adjustment mechanism is introduced to construct a multi-objective optimization function, so that the multi-objective optimization function can be adjusted in real time according to specific needs. The actual operating results of the system are then run by the adaptive optimization algorithm to compare with the predicted results, and the comparison deviation is fed back to the adaptive optimization algorithm for closed-loop feedback optimization.

[0230] The embodiment of the present invention also provides a BIM-based central air conditioning safety monitoring device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the steps in the above-mentioned BIM-based central air conditioning safety monitoring method embodiment are implemented, for example Figure 1 or, the processor implements the functions of each module in the above-mentioned system embodiments when executing the computer program.

[0231] Exemplarily, the computer program may be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of completing specific functions, which are used to describe the execution process of the computer program in the BIM-based central air conditioning safety monitoring device.

[0232] The BIM-based central air-conditioning safety monitoring device can be a computing device such as a desktop computer, a notebook, a PDA, and a cloud server. The BIM-based central air-conditioning safety monitoring device can include, but is not limited to, a processor and a memory. Those skilled in the art can understand that the BIM-based central air-conditioning safety monitoring device can also include input and output devices, network access devices, buses, etc.

[0233] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASAC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the BIM-based central air-conditioning safety monitoring device, and uses various interfaces and lines to connect the various parts of the entire BIM-based central air-conditioning safety monitoring device.

[0234] The memory can be used to store the computer program and / or module, and the processor realizes various functions of the BIM-based central air-conditioning safety monitoring device by running or executing the computer program and / or module stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function, etc.; the data storage area can store data created according to the operation of the air-conditioning controller, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (SmartMedaaCard, SMC), a secure digital (SecureDagatal, SD) card, a flash card (FlashCard), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0235] Wherein, if the module integrated in the central air-conditioning safety monitoring device based on BIM is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0236] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0237] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A central air conditioning safety monitoring method based on BIM, characterized in that: The method comprises the following steps: S1. Use BIM software to generate a BIM model of the building. The three-dimensional coordinates of the BIM model are marked as P (x, y, z), representing all spatial locations inside the building. At the same time, the central air-conditioning system is monitored in real time, and the real-time data from the sensor network is integrated into the BIM model. Each type of data corresponds to a specific dimension, and is dynamically mapped with the equipment information in space. An adaptive filtering model is designed to pre-process the data transmitted into the BIM model in real time. Finally, the pre-processed data is mapped back to the equipment location and attributes in the BIM model to form a multi-level data update mechanism. S2. According to the structural layout of the building and the arrangement of the central air-conditioning equipment in the BIM model, the building is initially divided into safety areas, and the safety areas are monitored in real time. The real-time environmental data is mapped to each safety area, and the multi-level risk index of each safety area is calculated. At the same time, the boundary of each safety area is dynamically adjusted according to the risk index based on the time-varying regularization term calculated in real time, and updated to the BIM model in real time; wherein the time-varying regularization term is expressed as follows: Among them, E ik (t) represents the environment state vector E i The kth component of (t), β k represents the risk weight, λ represents the regularization coefficient, which is used to adjust the impact of time changes on the risk index, m represents the total number of features, j represents the jth feature, k represents the kth feature, and t represents the current time; the time change regularization term is used to capture the impact of dynamic changes in environmental parameters on safety according to the central air-conditioning system. It means to find partial derivatives; S3. Extract fault features from real-time environmental data and equipment status data, and make a preliminary judgment on whether there is a fault. If it is preliminarily judged that there is a fault, it is marked as a potential fault device, and then the fault location and cause analysis are performed on the potential fault device. The system automatically generates maintenance suggestions based on the results of the cause analysis and sends the maintenance suggestions to the maintenance personnel. At the same time, according to the severity and impact range of different faults, the system will adjust the equipment status in the BIM model in real time and push the fault information to the maintenance personnel; wherein, the preliminary judgment on whether there is a fault is expressed as follows: D i (t)=||F i (t)-F ref || Among them, F ref Represents the normal operation reference characteristic vector of the equipment; F i (t) represents the fault feature vector; if the deviation D i (t) exceeds a certain preset threshold D th , then mark the device as a potential fault device; S4. Build a fault tracing model based on the fault location results, determine the fault propagation path, and analyze the root cause of the fault based on the fault propagation path. The system automatically generates targeted repair guidance plans based on the root cause. S5. Build a digital twin model based on the BIM model, build a risk prediction model based on historical data and real-time data to predict the risk of the digital twin model, and then automatically generate a preventive maintenance plan based on the results of the risk prediction to ensure that maintenance and adjustments are carried out before the equipment fails. At the same time, the actual operation data and maintenance results are recorded in real time and fed back to the digital twin model and risk prediction model to form an intelligent feedback mechanism; S6. Design an adaptive optimization algorithm based on reinforcement learning to maximize the operating efficiency of the system and minimize the risk level. At the same time, introduce a dynamic adjustment mechanism to construct a multi-objective optimization function, so that the multi-objective optimization function can be adjusted in real time according to specific needs. Then run the adaptive optimization algorithm to compare the actual operating results of the system with the predicted results, and feed back the comparison deviation to the adaptive optimization algorithm for closed-loop feedback optimization.

2. The central air conditioning safety monitoring method based on BIM according to claim 1 is characterized in that: The method also includes designing an intelligent analysis module to generate a detailed analysis report based on the real-time data and optimization results of the system; the analysis report includes system performance, energy consumption analysis, risk assessment and optimization suggestions, and the formula is as follows: R report (t)=A analyze (D twin (t),O opt (t),F multi (t)) Among them, A analyze Represents the analytical function, combined with the digital twin model D twin (t), optimization result O opt (t) and multi-objective function F multi (t) Conduct comprehensive analysis; report (t) indicates analysis report.

3. The central air conditioning safety monitoring method based on BIM according to claim 1 is characterized in that: The adaptive filtering model is expressed as follows: Among them, D clean (t) represents the processed data, D(t) represents the filtered data vector; D real (t) represents the original real-time data vector; R k (D(t)) represents the kth regularization term, which is used to suppress different types of noise; W k represents the weight of the regularization term, indicating the influence of each type of noise in the current environment; λ represents the regularization coefficient, which controls the intensity of filtering; The pre-processed data is mapped back to the equipment location and attributes in the BIM model to form a multi-level data update mechanism, as shown below: M(t)=M(t-1)+F(D clean (t)) Where M(t) represents the BIM model dataset at time t; M(t-1) represents the model dataset at the previous time; F(D clean (t)) represents the data mapping function, which maps the cleaned data to the corresponding device and spatial location.

4. The central air conditioning safety monitoring method based on BIM according to claim 1 is characterized in that: The S2 specifically includes: S201, divide the building into initial safety areas according to the internal structural layout of the building and the arrangement of the central air conditioning equipment, where each area R i The initial boundary R i (0) is determined based on the impact range of the equipment and the physical partitioning of the building. The impact range of the equipment is estimated by a thermodynamic model and is expressed as follows: R i (0)={(x,y,z)|P(x,y,z)∈P dev +δ i } Among them, P(x, y, z) represents a spatial point, P dev represents the device position vector, δ i Indicates the influence range offset of device i, which is used to expand or reduce the influence area; S202, define the environmental parameters as a multi-level vector E i (t), which includes parameters at different levels and then mapped to each security region R i Then according to E i (t) Calculate the multi-level risk index I for each area i (t), expressed as follows: Among them, E ik (t) represents the environment state vector E i The kth component of (t), β k represents the risk weight, λ represents the regularization coefficient, which is used to adjust the impact of time changes on the risk index, n represents the total number of indicators, m represents the total number of features, j represents the jth feature, k represents the kth feature, and t represents the current time; S203, each area R i The border The risk index I will be calculated in real time. i (t) Perform dynamic adjustments; S204, the system recalculates the environmental state E of all safety areas every time step △t i (t+△t) and risk indicator I i (t+△t), according to the new risk indicators, the system will dynamically adjust the regional boundaries And update it to the BIM model in real time.

5. The central air conditioning safety monitoring method based on BIM according to claim 1 is characterized in that: The S203 includes: Each area R i The border The risk index I will be calculated in real time. i (t) Dynamic adjustment is performed if the risk index of a certain area exceeds the preset threshold I th , the system will automatically adjust the boundaries of the region to increase or decrease the area of ​​influence, as shown below: in, represents the gradient of the risk indicator, γ represents the boundary adjustment step size, η represents the boundary adjustment sensitivity coefficient, sign(I i (t)-I th ) is used to indicate whether the risk indicator exceeds the threshold.

6. The central air conditioning safety monitoring method based on BIM according to claim 1 is characterized in that: Based on the preliminary judgment results, fault location and cause analysis are performed, including: According to the spatial position P of the equipment in the BIM model dev and connectivity parameter C i , determine the path along which the fault may propagate and identify the associated device j; Based on the data correlation between associated devices, calculate the failure impact G of device i i (t), expressed as follows: in, represents the set of devices associated with device i, ρij represents the correlation coefficient between devices i and j, and represents the possibility of fault propagation; F j (t) represents the fault characteristics of device j; Combined fault impact G i (t) and fault characteristics F i (t), analyze the most likely cause of the fault through optimization solution C i (t), the cause analysis formula is as follows: Among them, λ is the regularization coefficient used to avoid overfitting, and C represents the total number of connectivity parameters.

7. The central air conditioning safety monitoring method based on BIM according to claim 1 is characterized in that: The S4 comprises: S401. Construct a fault tracing model, determine the root cause of the fault through the fault tracing model, and analyze the propagation path of the fault in the entire system, as shown below: Among them, C i，j represents the connection strength between devices i and j, d i，j represents the physical distance between the two devices, τ represents the influence attenuation constant; F i (t) represents the fault characteristics, G i (t) represents the fault influence; P i，j (t) represents the possibility of fault propagation path; S402. Analyze the root cause of the fault according to the fault propagation path, as shown below: Among them, R i (t) represents the fault cause vector, α is the regularization coefficient of the fault propagation path, and ||R||1 is the sparsity constraint term of the root cause vector, which ensures that the most concise explanation of the fault cause is found; S403, according to the fault cause vector R i (t), the system will automatically generate a repair guidance plan; S404. After the repair is completed, the system will compare the actual operation data with the preset repair plan to evaluate the effect of the repair. If the repair is incomplete or a new fault occurs, the system will automatically update the fault detection and tracing model.

8. The central air conditioning safety monitoring method based on BIM according to claim 1 is characterized in that: The S5 specifically includes: S501. Build the initial digital twin model D twin (t=0), expressed as follows: Among them, M BIM (0) represents the initial BIM model, S i (0) represents the initial state information of device i, H i (0) represents the historical maintenance record of equipment i; at the same time, the initial digital twin model D twin (t=0) to perform dynamic calibration and obtain the dynamic digital twin model D twin (t), expressed as follows: D twin (t)=D twin (t-1)+K·(S real (t)-S pred (t)) Where κ is the calibration coefficient, S real (t) represents the device status collected in real time, S pred (t) represents the equipment status predicted based on the digital twin model; S502. Combine historical data and real-time data to build a risk prediction model R that introduces time decay and seasonal adjustment life (t), expressed as follows: Among them, Wi represents the weight of historical data, H i (t) represents the historical data of device i, C i (t) represents the state vector of the digital twin model, τ represents the time decay constant, and T represents the seasonal period; S503: Automatically generate a preventive maintenance plan Mpm(t) based on the results of the risk prediction model and execute an optimization scheduling algorithm. opt (t), adjust the operating parameters of the system, expressed as follows: Among them, λ represents the regularization parameter, O prev Indicates the previous scheduling status; S504. During the system operation and maintenance process, the actual operation data and maintenance results will be recorded in real time and fed back to the digital twin model and risk prediction model, forming an intelligent feedback mechanism to optimize the system.

9. The central air conditioning safety monitoring method based on BIM according to claim 1 is characterized in that: The adaptive optimization algorithm based on reinforcement learning opt , which is expressed as follows: Where E(t) represents the operating efficiency, R(t) represents the risk level, O represents the set of system parameters to be optimized, and γ represents the coefficient for balancing the operating efficiency and the risk level; A dynamic adjustment mechanism is introduced to design a multi-objective optimization function, so that the optimization objectives can be adjusted in real time according to specific needs. The multi-objective optimization function F multi (t) is expressed as follows: Among them, F multi (t) represents the multi-objective optimization function, G k (t) represents the kth objective function, β k (t) represents the weight coefficient adjusted in real time.

10. The BIM-based central air conditioning safety monitoring system is characterized by: The system comprises: BIM model construction module is used to generate the BIM model of the building using BIM software. The three-dimensional coordinates of the BIM model are marked as P (x, y, z), representing all the spatial positions inside the building. At the same time, it monitors the central air-conditioning system in real time and integrates the real-time data from the sensor network into the BIM model. Each type of data corresponds to a specific dimension and is dynamically mapped with the equipment information in space. An adaptive filtering model is designed to pre-process the data transmitted into the BIM model in real time. Finally, the pre-processed data is mapped back to the equipment position and attributes in the BIM model to form a multi-level data update mechanism; The risk detection module is used to divide the building into initial safety zones according to the structural layout of the building and the arrangement of the central air-conditioning equipment in the BIM model, and to monitor the safety zones in real time, map the real-time environmental data to each safety zone, and calculate the multi-level risk index of each safety zone. At the same time, the boundary of each safety zone is dynamically adjusted according to the risk index based on the time-varying regularization term calculated in real time, and updated to the BIM model in real time; wherein the time-varying regularization term is expressed as follows: Among them, E ik (t) represents the environment state vector E i The kth component of (t), β k represents the risk weight, λ represents the regularization coefficient, which is used to adjust the impact of time changes on the risk index, m represents the total number of features, j represents the jth feature, k represents the kth feature, and t represents the current time; the time change regularization term is used to capture the impact of dynamic changes in environmental parameters on safety according to the central air-conditioning system. It means to find partial derivatives; The fault analysis module is used to extract fault features from real-time environmental data and equipment status data, and preliminarily determine whether there is a fault. If it is preliminarily determined that there is a fault, it is marked as a potential fault device, and then the potential fault device is fault located and cause analyzed. The system automatically generates maintenance suggestions based on the results of the cause analysis and sends the maintenance suggestions to the maintenance personnel. At the same time, according to the severity and impact range of different faults, the system will adjust the equipment status in the BIM model in real time and push the fault information to the maintenance personnel; wherein, the preliminary judgment of whether there is a fault is expressed as follows: D i (t)=||F i (t)-F ref || Among them, F ref Represents the normal operation reference characteristic vector of the equipment; F i (t) represents the fault feature vector; if the deviation D i (t) exceeds a certain preset threshold D th , then mark the device as a potential fault device; The fault location module is used to build a fault tracing model based on the results of fault location, determine the fault propagation path, and analyze the root cause of the fault based on the fault propagation path. The system automatically generates targeted repair guidance plans based on the root cause; The fault prediction module is used to build a digital twin model based on the BIM model, build a risk prediction model based on historical data and real-time data to predict the risk of the digital twin model, and then automatically generate a preventive maintenance plan based on the results of the risk prediction to ensure that maintenance and adjustments are carried out before the equipment fails. At the same time, the actual operation data and maintenance results are recorded in real time and fed back to the digital twin model and risk prediction model to form an intelligent feedback mechanism; The system optimization module is used to design an adaptive optimization algorithm based on reinforcement learning to maximize the system's operating efficiency and minimize the risk level. At the same time, a dynamic adjustment mechanism is introduced to construct a multi-objective optimization function, so that the multi-objective optimization function can be adjusted in real time according to specific needs. The adaptive optimization algorithm is then run to compare the actual operating results of the system with the predicted results, and the comparison deviation is fed back to the adaptive optimization algorithm for closed-loop feedback optimization.

Images