Intelligent building equipment control method and system

Abstract

The application discloses a kind of intelligent building equipment control method and system, it is related to intelligent building equipment control technical field, including using Kruskal algorithm fusion the multimodal association topology of these two groups of data, construct minimum spanning tree, using the reciprocal of edge weight Set as pheromone value, using maximum flow algorithm-ant colony algorithm selects optimal control path;Using Takens embedding method converts phase space data, randomly selects a pair of adjacent phase space data from time series as initial point, using Lyapunov index analysis building equipment state, construct hidden Markov model, using Lyapunov index dynamic correction transition probability matrix and emission probability matrix, obtain the predicted state of building equipment.By constructing minimum spanning tree, and combining maximum flow algorithm-ant colony algorithm selects optimal control path, improve resource utilization and control accuracy, based on the combination of hidden Markov model and Lyapunov index, enhance the comprehensiveness and adaptability of equipment control.

Description

Technical Field

[0001] This invention relates to the field of intelligent building equipment control technology, and in particular to a smart building equipment control method and system. Background Technology

[0002] With the continuous advancement of smart city construction, building automation technology plays a vital role in improving the energy efficiency, comfort, and safety of buildings. Building automation systems typically include the control and management of equipment such as air conditioning, elevators, and lighting. The efficient operation of these devices can significantly improve the energy efficiency and user experience of buildings. In these systems, smart sensors are widely used to monitor environmental changes and the status of building equipment in real time, providing data support for optimizing control strategies. With the development of Internet of Things (IoT) technology, the interconnection of building equipment has become the foundation for realizing intelligent control. By collecting environmental and equipment data through sensors and combining them with advanced computing algorithms, intelligent scheduling and optimized control of building equipment can be achieved, ensuring that the comfort and energy efficiency of buildings are maximized.

[0003] Existing building equipment control systems still face many challenges in practical applications. Although smart sensors can provide multi-dimensional data, existing systems have relatively simple modeling of complex relationships between devices, making it difficult to select the optimal control path between devices. When predicting the status of building equipment, existing building management systems often rely on traditional rules and experience, resulting in a slow response to environmental changes and failing to meet the increasingly complex equipment management needs. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a smart building equipment control method and system, which solves the problem that although smart sensors can provide multi-dimensional data, existing systems have relatively simple modeling of complex relationships between devices, making it difficult to achieve optimal control path selection between devices. Existing building management systems often rely on traditional rules and experience when predicting the status of building equipment, resulting in a slow response to environmental changes and failing to meet the increasingly complex equipment management needs.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a smart building equipment control method, comprising,

[0008] The system uses smart sensors to collect environmental and building equipment data. The two sets of data are preprocessed, and the building equipment is defined as network nodes. The connection between the building equipment is defined as an edge. The Kruskal algorithm is used to fuse the multimodal association topology of the two sets of data to construct a minimum spanning tree. The inverse of the edge weight is used as the pheromone value. The maximum flow algorithm-ant colony algorithm is used to select the optimal control path.

[0009] The Takens embedding method is used to transform phase space data. A pair of adjacent phase space data is randomly selected from the time series as the initial point. The Lyapunov exponent is used to analyze the status of building equipment, and a hidden Markov model is constructed. The Lyapunov exponent is used to dynamically correct the transition probability matrix and the emission probability matrix to obtain the predicted status of building equipment.

[0010] A comprehensive objective function is constructed using a linear weighting method, solved using a particle swarm optimization algorithm to obtain the optimal individual, converted into scheduling instructions using a dynamic scheduling algorithm and executed, and monitored and adjusted using a PID control algorithm.

[0011] Store the two sets of data generated from collection and analysis.

[0012] As a preferred embodiment of the intelligent building equipment control method of the present invention, the method includes: collecting environmental and building equipment data using intelligent sensors, defining building equipment as network nodes, defining connections between building equipment as edges, fusing the multimodal association topology of these two sets of data using the Kruskal algorithm, constructing a minimum spanning tree, setting the pheromone value using the reciprocal of the edge weight, and selecting the optimal control path using the maximum flow algorithm-ant colony algorithm, including:

[0013] The intelligent sensors include sensors for monitoring light, temperature and humidity, vibration, bandwidth, and delay.

[0014] The environmental and building equipment data includes light intensity, temperature and humidity, vibration, bandwidth, and connection latency data;

[0015] Use IoT protocols to connect sensors to the central control system;

[0016] The building equipment includes air conditioning, elevators, lighting, and smart sensors;

[0017] Building equipment is defined as network nodes, and the connections between building equipment are defined as edges. The weighting coefficients of bandwidth and connection delay data are calculated separately using the analytic hierarchy process (AHP). The bandwidth and connection delay data are combined using the weighted summation method to calculate the connection weights between building equipment, which are defined as edge weights.

[0018] Sort the edge weights in ascending order, select the edge with the smallest weight as the first connection, then select the next edge with a smaller weight as the second connection. Use the disjoint-set data structure algorithm to progressively select edges connecting building equipment from the sorted edge weights until all building equipment is connected into a connected graph, and stop the selection to generate the minimum spanning tree.

[0019] The pheromone value is set using the reciprocal of the edge weight, the path selection probability is calculated using the ant colony algorithm, the edge weight is set as the path quality, the path is selected using heuristic information, the pheromone evaporation coefficient is set using the rule of thumb, and the pheromone increment is calculated using the pheromone increment formula.

[0020] After the ant selects a path, the pheromone concentration of the path is updated using a local pheromone update formula.

[0021] After each ant completes its path selection and pheromone update, it proceeds to the next selection, until all ants have completed their path selection and the update process stops.

[0022] After all ants have completed their path selection, the quality of the optimal path is calculated using a weighted summation method, the optimal pheromone increment is calculated using the pheromone increment formula, and the pheromone concentration of the optimal path is updated using the global pheromone update formula.

[0023] The ant colony algorithm is used to select the optimal control path based on the pheromone concentration and quality of the optimal path.

[0024] As a preferred embodiment of the intelligent building equipment control method of the present invention, the preprocessing of the two sets of collected data includes:

[0025] The two sets of collected data were aligned in time using a timestamp alignment method, outliers were identified and removed using the quartile method, missing values ​​were filled using the mean imputation method, and the two sets of imputed data were normalized.

[0026] As a preferred embodiment of the intelligent building equipment control method of the present invention, the following steps are included: using Takens embedding to transform phase space data, randomly selecting a pair of adjacent phase space data from the time series as initial points, using the Lyapunov exponent to analyze the building equipment state, constructing a hidden Markov model, and using the Lyapunov exponent to dynamically correct the transition probability matrix and emission probability matrix to obtain the predicted state of the building equipment, including:

[0027] Collect and preprocess historical environmental data to generate a training set;

[0028] Constructing a Hidden Markov Model and training the Hidden Markov Model using a training set includes using threshold discretization to discretize the preprocessed historical environmental data into hidden states, wherein the hidden states include normal, fault, and degradation.

[0029] The number of transitions from one hidden state to another in the preprocessed historical environmental data is statistically analyzed. The transition probability between each pair of states is calculated using maximum likelihood. The transition probabilities are then filled into the transition probability matrix and normalized to obtain the initial transition probability matrix.

[0030] The number of times the preprocessed historical environmental data appears in the hidden state is counted. The emission probability of each hidden state is calculated using maximum likelihood. The emission probabilities are filled into the emission probability matrix and normalized to obtain the initial emission probability matrix.

[0031] The preprocessed environmental data is sorted in chronological order to generate a time series.

[0032] The time series data is converted into phase space data by using mutual information and FNN methods to set the delay time and embedding dimension, and the Takens embedding method is used to convert the time series into phase space data.

[0033] Randomly select a pair of adjacent phase space data from the time series as the initial point;

[0034] Calculate the Euclidean distance between the initial points using the Euclidean distance formula;

[0035] A second time series is generated using a small perturbation method on the initial point, and the Euclidean distance between the first time series and the second time series is calculated using the Euclidean distance formula.

[0036] The Lyapunov exponent for time t is calculated using the tangent vector method.

[0037] If the Lyapunov exponent at time t is greater than 0, it indicates that the building equipment is unstable. The control factors of the initial transition probability matrix and the initial emission probability matrix are set by cross-validation, and the adjustment range of the transition probability and emission probability is calculated by dynamic adjustment method.

[0038] The transition probabilities in the initial transition probability matrix and the emission probabilities in the initial emission probability matrix are dynamically corrected using the proportional adjustment method, resulting in dynamically corrected transition probability matrices and emission probability matrices.

[0039] The transition probability matrix and emission probability matrix after dynamic correction are optimized using maximum a posteriori estimation to obtain the final transition probability matrix and emission probability matrix.

[0040] If the Lyapunov exponent of time t is less than or equal to 0, it indicates that the building equipment is normal, and the initial transition probability matrix and the initial emission probability matrix are used.

[0041] The maximum probability of building equipment being in a hidden state at time t is obtained using the Viterbi algorithm.

[0042] Real-time environmental data is input into a trained Hidden Markov Model to obtain the predicted state of building equipment.

[0043] As a preferred embodiment of the intelligent building equipment control method of the present invention, the step of constructing a comprehensive objective function using a linear weighted method and solving it using a particle swarm optimization algorithm to obtain the optimal individual includes:

[0044] The predicted status of building equipment is normalized.

[0045] Calculate the mean values ​​of the preprocessed temperature and humidity data respectively, and set them as the ideal temperature and humidity data;

[0046] The temperature and humidity deviation method is used to set the comfort objective function;

[0047] Use the mean absolute error to define the loss function;

[0048] The weight coefficients of the comfort objective function and the loss function are calculated separately using a data-driven method, and a comprehensive objective function is constructed using a linear weighting method.

[0049] An initial population is randomly generated, with each individual representing a combination of prediction error and comfort requirement.

[0050] Individuals are evaluated by substituting them into a comprehensive objective function, and their fitness values ​​are calculated.

[0051] Individuals are randomly selected for comparison using a tournament selection method, and individuals with higher fitness are selected as parents to participate in the crossover operation.

[0052] New offspring are produced by exchanging the genes of parent individuals at a single point of crossover.

[0053] After crossover, individuals are randomly selected and mutated using single-point mutation to generate new individuals;

[0054] The maximum number of iterations is set using a halting criterion. When the maximum number of iterations is reached, iteration stops, and the optimal individual is output, including the optimized prediction error and comfort requirements.

[0055] As a preferred embodiment of the intelligent building equipment control method of the present invention, the step of converting the data into scheduling instructions using a dynamic scheduling algorithm and executing them, and using a PID control algorithm for monitoring and adjustment, includes:

[0056] The optimal individual is converted into a scheduling instruction using a dynamic scheduling algorithm, which is then transmitted to the building equipment via an API interface. The building equipment receives the scheduling instruction and executes it.

[0057] Collect feedback data and calculate the difference between the feedback data and the optimal individual;

[0058] A judgment threshold is set using statistical methods. Differences are compared with the judgment threshold. Differences greater than or equal to the judgment threshold are adjusted using a PID control algorithm until the difference is less than the judgment threshold. The adjustment stops then, and the feedback data is monitored.

[0059] As a preferred embodiment of the intelligent building equipment control method of the present invention, the two sets of data collected and analyzed include:

[0060] The two sets of collected data and the optimal individuals generated from the analysis are stored in a central database, and secure access measures are set up. The central database backs up the stored data to the cloud and performs integrity checks on the stored data and backup data regularly. After the checks are completed, integrity check records are generated and stored synchronously in the central database.

[0061] Secondly, the present invention provides a smart building equipment control system, comprising,

[0062] The topology optimization module is used to collect environmental and building equipment data using smart sensors. It preprocesses the two sets of data, defines building equipment as network nodes, defines the connections between building equipment as edges, uses the Kruskal algorithm to fuse the multimodal association topology of the two sets of data, constructs a minimum spanning tree, uses the reciprocal of the edge weight as the pheromone value, and uses the maximum flow algorithm-ant colony algorithm to select the optimal control path.

[0063] The state prediction module is used to transform phase space data using the Takens embedding method. It randomly selects a pair of adjacent phase space data from the time series as the initial point, uses the Lyapunov exponent to analyze the state of building equipment, constructs a hidden Markov model, and uses the Lyapunov exponent to dynamically correct the transition probability matrix and emission probability matrix to obtain the predicted state of the building equipment.

[0064] The scheduling and monitoring module is used to construct a comprehensive objective function using a linear weighting method, solve it using a particle swarm optimization algorithm to obtain the optimal individual, convert it into scheduling instructions using a dynamic scheduling algorithm and execute them, and monitor and adjust it using a PID control algorithm.

[0065] The storage module is used to store the two sets of data generated during collection and analysis.

[0066] Thirdly, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the intelligent building equipment control method as described in the first aspect of the present invention.

[0067] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the intelligent building equipment control method as described in the first aspect of the present invention.

[0068] The beneficial effects of this invention are as follows: This invention collects environmental and building equipment data using intelligent sensors, preprocesses the two sets of data, defines building equipment as network nodes, defines the connections between building equipment as edges, uses the Kruskal algorithm to fuse the multimodal association topology of these two sets of data, constructs a minimum spanning tree, uses the reciprocal of the edge weight as the pheromone value, and uses the maximum flow algorithm-ant colony algorithm to select the optimal control path; it uses the Takens embedding method to transform phase space data, randomly selects a pair of adjacent phase space data from the time series as initial points, uses the Lyapunov exponent to analyze the state of building equipment, constructs a hidden Markov model, uses the Lyapunov exponent to dynamically correct the transition probability matrix and emission probability matrix, and obtains the predicted state of the building equipment; it uses a linear weighted method to construct a comprehensive objective function, uses a particle swarm optimization algorithm to solve it, obtains the optimal individual, uses a dynamic scheduling algorithm to convert it into scheduling instructions for execution, and uses a PID control algorithm for monitoring and adjustment; this improves response speed, resource utilization, and control accuracy, avoids the impact of equipment failure on building operation, and enhances the comprehensiveness and adaptability of equipment control. Attached Figure Description

[0069] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0070] Figure 1 This is a flowchart of the intelligent building equipment control method in Example 1.

[0071] Figure 2 This is a structural diagram of the intelligent building equipment control system in Example 1. Detailed Implementation

[0072] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0073] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0074] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0075] Example 1, referring to Figure 1 and Figure 2 This is the first embodiment of the present invention, which provides a smart building equipment control method, including the following steps:

[0076] S1. Use smart sensors to collect environmental and building equipment data, preprocess the two sets of data, define building equipment as network nodes, define the connection between building equipment as edges, use Kruskal algorithm to fuse the multimodal association topology of the two sets of data, construct the minimum spanning tree, use the reciprocal of the edge weight as the pheromone value, and use the maximum flow algorithm-ant colony algorithm to select the optimal control path.

[0077] Specifically, intelligent sensors are used to collect environmental and building equipment data. Building equipment is defined as network nodes, and connections between building equipment are defined as edges. The Kruskal algorithm is used to fuse the multimodal association topology of these two sets of data, constructing a minimum spanning tree. The reciprocal of the edge weight is used as the pheromone value, and the maximum flow algorithm-ant colony algorithm is used to select the optimal control path, including:

[0078] The intelligent sensors include sensors for monitoring light, temperature and humidity, vibration, bandwidth, and delay.

[0079] The environmental and building equipment data includes light intensity, temperature and humidity, vibration, bandwidth, and connection latency data;

[0080] Use IoT protocols to connect sensors to the central control system;

[0081] The building equipment includes air conditioning, elevators, lighting, and smart sensors;

[0082] Building equipment is defined as network nodes, and the connections between building equipment are defined as edges. The weighting coefficients of bandwidth and connection delay data are calculated separately using the analytic hierarchy process (AHP). The bandwidth and connection delay data are combined using the weighted summation method to calculate the connection weights between building equipment, which are defined as edge weights.

[0083] Sort the edge weights in ascending order, select the edge with the smallest weight as the first connection, then select the next edge with a smaller weight as the second connection. Use the disjoint-set data structure algorithm to progressively select edges connecting building equipment from the sorted edge weights until all building equipment is connected into a connected graph, and stop the selection to generate the minimum spanning tree.

[0084] The pheromone value is set using the reciprocal of the edge weight, the path selection probability is calculated using the ant colony algorithm, the edge weight is set as the path quality, the path is selected using heuristic information, the pheromone evaporation coefficient is set using the rule of thumb, and the pheromone increment is calculated using the pheromone increment formula.

[0085] After the ant chooses a path, the pheromone concentration along that path is updated using the local pheromone update formula, which is:

[0086] ,

[0087] in Let be the pheromone concentration on the ij-th path at time t+1. The pheromone evaporation coefficient. Let be the pheromone concentration on the ij-th path at time t. Let be the pheromone increment on the ij-th path;

[0088] After each ant completes its path selection and pheromone update, it proceeds to the next selection, until all ants have completed their path selection and the update process stops.

[0089] After all ants have completed their path selection, the quality of the optimal path is calculated using a weighted summation method, the optimal pheromone increment is calculated using the pheromone increment formula, and the pheromone concentration of the optimal path is updated using the global pheromone update formula, which is:

[0090] ,

[0091] in Let be the pheromone concentration on the ij-th path at time t+1. The pheromone evaporation coefficient. Let be the pheromone concentration on the ij-th path at time t. The optimal pheromone increment on the ij-th path;

[0092] The ant colony algorithm is used to select the optimal control path based on the pheromone concentration and quality of the optimal path.

[0093] By optimizing control paths and reducing ineffective device connections, resources can be effectively saved, energy consumption reduced, and control paths and device scheduling dynamically adjusted to ensure the system responds promptly to environmental and device status changes. The combination of minimum spanning tree and ant colony algorithm ensures that building equipment can work together efficiently and avoid system bottlenecks. Combined with intelligent sensors and optimization algorithms, building equipment control becomes more intelligent and automated, reducing manual intervention and improving the system's adaptability.

[0094] By collecting real-time environmental parameters (such as temperature, humidity, and vibration) and equipment status data (such as bandwidth and connection latency) within the building, the system can dynamically monitor and control building equipment. This helps to promptly detect equipment anomalies or environmental changes, enabling timely adjustments, avoiding resource waste, and improving building energy efficiency and comfort. The Kruskal algorithm optimizes the connection structure between building equipment by generating a minimum spanning tree, ensuring that data transmission between devices is not interfered with by redundant paths, thus improving system efficiency. The generated minimum spanning tree reduces bandwidth consumption and latency in the network, improving the smoothness of inter-device collaboration and forming the basis for optimized control. Combining the maximum flow algorithm and ant colony algorithm, the system can not only evaluate and select the optimal control path but also dynamically adapt to changes in different network states. The method ensures unobstructed information flow, while the ant colony algorithm optimizes path selection based on real-time data, enabling efficient transmission of control commands. This allows for more precise coordination and scheduling between devices, improving the response speed and accuracy of building control. The analytic hierarchy process (AHP) calculates bandwidth and connection delay weighting coefficients, ensuring that the importance of each factor in path selection is reasonably considered, and also improving the overall effect of network optimization. This ensures efficient and stable communication between devices. Pheromone evaporation ensures that the system gradually eliminates inefficient paths, while incremental updates enhance the pheromone concentration of high-quality paths, enabling the system to find the optimal control path in complex environments. The effective operation of this mechanism can improve the accuracy and efficiency of building equipment control, especially in complex systems with a large number of devices, ensuring the accurate execution of real-time control decisions.

[0095] Furthermore, the two sets of collected data were preprocessed, including:

[0096] The two sets of collected data were aligned in time using a timestamp alignment method, outliers were identified and removed using the quartile method, missing values ​​were filled using the mean imputation method, and the two sets of imputed data were normalized.

[0097] Timestamp alignment effectively eliminates errors caused by time differences, ensuring data is compared within the same time window, improving the temporal consistency of data, and avoiding misleading analysis results due to time differences. Especially when dealing with time-related analysis tasks, it provides a more reliable and consistent data foundation. Removing outliers improves data stability and accuracy, ensuring the reliability of subsequent analysis results. Especially in applications requiring high-precision calculations, it can avoid incorrect judgments or misleading conclusions caused by outliers. Mean imputation can prevent large errors caused by missing data. Through normalization, each feature of the data can be compared on a uniform scale, thereby avoiding the impact of certain features being too large or too small on the analysis results.

[0098] S2. Use Takens embedding to transform phase space data, randomly select a pair of adjacent phase space data from the time series as the initial point, use the Lyapunov exponent to analyze the status of building equipment, construct a hidden Markov model, use the Lyapunov exponent to dynamically correct the transition probability matrix and emission probability matrix, and obtain the predicted status of building equipment.

[0099] Specifically, the Takens embedding method is used to transform the phase space data. A pair of adjacent phase space data points are randomly selected from the time series as initial points. The Lyapunov exponent is used to analyze the building equipment status, constructing a hidden Markov model. The Lyapunov exponent is used to dynamically correct the transition probability matrix and emission probability matrix to obtain the predicted status of the building equipment, including:

[0100] Collect and preprocess historical environmental data to generate a training set;

[0101] Constructing a Hidden Markov Model and training the Hidden Markov Model using a training set includes using threshold discretization to discretize the preprocessed historical environmental data into hidden states, wherein the hidden states include normal, fault, and degradation.

[0102] The number of transitions from one hidden state to another in the preprocessed historical environmental data is statistically analyzed. The transition probability between each pair of states is calculated using maximum likelihood. The transition probabilities are then filled into the transition probability matrix and normalized to obtain the initial transition probability matrix.

[0103] The number of times the preprocessed historical environmental data appears in the hidden state is counted. The emission probability of each hidden state is calculated using maximum likelihood. The emission probabilities are filled into the emission probability matrix and normalized to obtain the initial emission probability matrix.

[0104] The preprocessed environmental data is sorted in chronological order to generate a time series.

[0105] The time series data is converted into phase space data by using mutual information and FNN methods to set the delay time and embedding dimension, and the Takens embedding method is used to convert the time series into phase space data.

[0106] Randomly select a pair of adjacent phase space data from the time series as the initial point;

[0107] Calculate the Euclidean distance between the initial points using the Euclidean distance formula;

[0108] A second time series is generated using a small perturbation method on the initial point, and the Euclidean distance between the first time series and the second time series is calculated using the Euclidean distance formula.

[0109] The Lyapunov exponent for time t is calculated using the tangent vector method, with the following formula:

[0110] ,

[0111] in Let be the Lyapunov exponent at time t. Let be the Euclidean distance between the initial points. Let be the Euclidean distance between the time series at time t and the second trajectory;

[0112] If the Lyapunov exponent at time t is greater than 0, it indicates that the building equipment is unstable. Cross-validation is used to set control factors for the initial transition probability matrix and the initial emission probability matrix, respectively. A dynamic adjustment method is used to calculate the adjustment range for the transition probability and emission probability, respectively. The formula is as follows:

[0113] ,

[0114] ,

[0115] in The adjustment range for the transition probability. The adjustment range for the launch probability. To emerge from the hidden state to hidden state The transition probability represents the normalized transition probability of each element in the initial transition probability matrix. In the hidden state The emission probability of the historical environmental data x after preprocessing, the elements in the original emission probability matrix, and the normalized emission probability;

[0116] The transition probabilities in the initial transition probability matrix are dynamically corrected using a proportional adjustment method, and the emission probabilities in the initial emission probability matrix are also dynamically corrected using the following formula:

[0117] ,

[0118] ,

[0119] The dynamically corrected transition probability matrix and emission probability matrix are obtained;

[0120] The transition probability matrix and emission probability matrix after dynamic correction are optimized using maximum a posteriori estimation to obtain the final transition probability matrix and emission probability matrix.

[0121] If the Lyapunov exponent of time t is less than or equal to 0, it indicates that the building equipment is normal, and the initial transition probability matrix and the initial emission probability matrix are used.

[0122] The maximum probability of building equipment being in a hidden state at time t is obtained using the Viterbi algorithm.

[0123] Real-time environmental data is input into a trained Hidden Markov Model to obtain the predicted state of building equipment.

[0124] By monitoring the Lyapunov exponent in real time and using hidden Markov models for state prediction, the system can provide early warnings before equipment failures or degradation occur, greatly improving the stability and reliability of equipment operation. It enables dynamic tracking and optimized scheduling of multiple equipment states, reduces unplanned equipment downtime, and improves the energy efficiency management level of buildings. By introducing Takens embedding and Lyapunov exponent analysis, the system can make more accurate predictions of equipment operating states, improve the intelligence level of building equipment management, and dynamically adjust the transition probability and emission probability matrices so that the system can maintain high prediction accuracy and control efficiency in constantly changing environments and equipment states.

[0125] By converting time-series data into phase-space data, Takens embedding reveals the underlying dynamic behavior of a system, fully capturing the nonlinear characteristics of building equipment. It extracts more state information from complex equipment operation data. By calculating the Lyapunov exponent of equipment in real time, the system can determine whether the equipment is in a stable operating state, improving the stability monitoring capability of building equipment, reducing equipment failure rates, and enhancing equipment reliability. Through model training using historical environmental data, the system can effectively learn the transition patterns of equipment states and their corresponding probability distributions, thereby predicting future equipment states. By modeling latent states and their transition patterns, more accurate equipment state prediction can be achieved, providing a more comprehensive solution for building equipment management. The management system provides intelligent decision support. When the Lyapunov index indicates that the equipment state is unstable, the system adjusts the transition probability matrix and emission probability matrix through cross-validation and dynamic adjustment methods. The dynamic adjustment mechanism enables the model to adapt to changes in equipment state, improving the model's response speed and accuracy to changes in equipment state. The Viterbi algorithm can accurately predict the state changes of building equipment by solving for the most likely hidden state sequence under a given observation sequence. This makes equipment state prediction not only limited to static analysis, but also takes into account the impact of historical state transitions, realizing dynamic prediction of future equipment state, improving the intelligence level of the building equipment management system, timely identifying equipment failures or degradation, and avoiding equipment damage or energy efficiency decline.

[0126] S3. Construct a comprehensive objective function using the linear weighting method, solve it using the particle swarm optimization algorithm to obtain the optimal individual, convert it into scheduling instructions using the dynamic scheduling algorithm and execute them, and use the PID control algorithm for monitoring and adjustment.

[0127] Specifically, a comprehensive objective function is constructed using a linear weighting method, and solved using a particle swarm optimization algorithm to obtain the optimal individual, including:

[0128] The predicted status of building equipment is normalized.

[0129] Calculate the mean values ​​of the preprocessed temperature and humidity data respectively, and set them as the ideal temperature and humidity data;

[0130] The comfort objective function is set using the temperature and humidity deviation method, and the formula is as follows:

[0131] ,

[0132] in Let N be the objective function for comfort, and N be the number of temperature and humidity sensors. and These are the l-th temperature and humidity data. and These are ideal temperature and humidity data;

[0133] The loss function is defined using the mean absolute error, and the formula is:

[0134] ,

[0135] in Let n be the loss function, and n be the number of building equipment. and These are the predicted and actual states of the z-th building equipment, respectively.

[0136] The weight coefficients of the comfort objective function and the loss function are calculated separately using a data-driven method, and a comprehensive objective function is constructed using a linear weighting method.

[0137] An initial population is randomly generated, with each individual representing a combination of prediction error and comfort requirement.

[0138] Individuals are evaluated by substituting them into a comprehensive objective function, and their fitness values ​​are calculated.

[0139] Individuals are randomly selected for comparison using a tournament selection method, and individuals with higher fitness are selected as parents to participate in the crossover operation.

[0140] New offspring are produced by exchanging the genes of parent individuals at a single point of crossover.

[0141] After crossover, individuals are randomly selected and mutated using single-point mutation to generate new individuals;

[0142] The maximum number of iterations is set using a halting criterion. When the maximum number of iterations is reached, iteration stops, and the optimal individual is output, including the optimized prediction error and comfort requirements.

[0143] By minimizing prediction errors and optimizing comfort goals, the system can achieve more precise equipment scheduling, improve building comfort and energy efficiency. The combination of data-driven methods and particle swarm optimization algorithms makes the optimization process more intelligent and automated, improving the adaptability of the building equipment management system. Through the global search capability of particle swarm optimization, the optimal balance between energy efficiency and comfort is ensured, improving the overall efficiency of the building management system.

[0144] In building equipment optimization, normalization helps eliminate scale differences between different data, enabling optimization algorithms to better handle multiple factors such as temperature, humidity, and equipment status, thereby improving algorithm stability and solution performance. By calculating the mean of temperature and humidity and setting ideal values, the system can clearly define the comfort target within the building. By minimizing the loss function, it ensures more precise equipment scheduling and control, thereby reducing equipment malfunctions and energy consumption, and improving system efficiency. Using a data-driven approach to dynamically adjust weights ensures that the system finds the optimal balance between comfort and energy efficiency, thus improving the reliability and adaptability of the optimization effect. In building equipment control optimization, particle swarm optimization can efficiently search for the optimal control scheme, handle complex multi-objective optimization problems, and find the global optimum in a short time, helping to improve equipment scheduling efficiency and reduce the waste of computing resources. Tournament selection can quickly select individuals with high fitness, while crossover and mutation operations can increase the diversity of the solution space and avoid getting trapped in local optima. The final output optimal individual represents the optimized balance between building equipment prediction error and comfort requirements, providing accurate scheduling decisions for building equipment control and optimizing energy efficiency and comfort.

[0145] Furthermore, a dynamic scheduling algorithm is used to convert the data into scheduling instructions for execution, and a PID control algorithm is used for monitoring and adjustment, including:

[0146] The optimal individual is converted into a scheduling instruction using a dynamic scheduling algorithm, which is then transmitted to the building equipment via an API interface. The building equipment receives the scheduling instruction and executes it.

[0147] Collect feedback data and calculate the difference between the feedback data and the optimal individual;

[0148] A judgment threshold is set using statistical methods. Differences are compared with the judgment threshold. Differences greater than or equal to the judgment threshold are adjusted using a PID control algorithm until the difference is less than the judgment threshold. The adjustment stops then, and the feedback data is monitored.

[0149] The combination of dynamic scheduling and PID control enables building equipment to respond to environmental and load changes in real time, ensuring that the equipment is always in optimal condition. By adjusting the equipment operation based on the difference between feedback data and target state, energy efficiency is maximized while improving user comfort. The combination of PID control algorithm and dynamic scheduling algorithm allows the system to flexibly adjust under different environmental and operating conditions, avoiding manual intervention and over-adjustment, and improving the system's adaptability. Through the use of API interface, the scheduling and control of building equipment becomes more intelligent and automated, improving the efficiency and sustainability of building management.

[0150] The dynamic scheduling algorithm automatically calculates and generates the optimal scheduling scheme based on real-time data and equipment status, improving the flexibility and response speed of building equipment scheduling. It can adapt to changes in demand under different time periods and load conditions. The API interface not only simplifies the data transmission process between systems but also ensures real-time information updates, greatly improving the system's automation and intelligence level. By collecting feedback data in real time and comparing it with the best individual, the system can understand the operating effect of the equipment in a timely manner, assess whether further adjustments are needed, accurately detect deviations in equipment operation, and promptly identify situations where equipment performance is substandard or environmental conditions are not ideal, providing data support for subsequent adjustments. Through statistical analysis, the system can avoid over-adjustment or ineffective adjustment, ensuring the accuracy and stability of the PID control process. PID control, through the coordination of three control strategies (proportional, integral, and derivative), enables the equipment to quickly and smoothly adjust to the ideal state. Through real-time monitoring and adjustment, the system can cope with various changes in the external environment or internal equipment status, making building equipment management more intelligent and automated.

[0151] S4. Store the two sets of data generated from collection and analysis;

[0152] Specifically, the data collected and analyzed are stored in two sets, including:

[0153] The two sets of collected data and the optimal individuals generated from the analysis are stored in a central database, and secure access measures are set up. The central database backs up the stored data to the cloud and performs integrity checks on the stored data and backup data regularly. After the checks are completed, integrity check records are generated and stored synchronously in the central database.

[0154] Centralized management makes data processing and analysis more efficient, better supporting subsequent data queries, processing, and mining. The introduction of secure access control measures helps prevent data leakage, tampering, or unauthorized access, thereby improving data confidentiality and protecting user privacy. The benefits of cloud backup include: improved data recoverability, ensuring that data is not lost in case of accidents; at the same time, cloud backup also allows data to be shared and accessed between different locations, enhancing system flexibility and accessibility, increasing system transparency, facilitating managers and auditors to check and confirm the integrity and security of data, and ensuring the controllability and compliance of the data processing process.

[0155] This embodiment also provides a smart building equipment control system, including:

[0156] The topology optimization module is used to collect environmental and building equipment data using smart sensors. It preprocesses the two sets of data, defines building equipment as network nodes, defines the connections between building equipment as edges, uses the Kruskal algorithm to fuse the multimodal association topology of the two sets of data, constructs a minimum spanning tree, uses the reciprocal of the edge weight as the pheromone value, and uses the maximum flow algorithm-ant colony algorithm to select the optimal control path.

[0157] The state prediction module is used to transform phase space data using the Takens embedding method. It randomly selects a pair of adjacent phase space data from the time series as the initial point, uses the Lyapunov exponent to analyze the state of building equipment, constructs a hidden Markov model, and uses the Lyapunov exponent to dynamically correct the transition probability matrix and emission probability matrix to obtain the predicted state of the building equipment.

[0158] The scheduling and monitoring module is used to construct a comprehensive objective function using a linear weighting method, solve it using a particle swarm optimization algorithm to obtain the optimal individual, convert it into scheduling instructions using a dynamic scheduling algorithm and execute them, and monitor and adjust it using a PID control algorithm.

[0159] The storage module is used to store the two sets of data generated during collection and analysis.

[0160] This embodiment also provides a computer device applicable to the intelligent building equipment control method, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the intelligent building equipment control method proposed in the above embodiment.

[0161] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0162] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the intelligent building equipment control method proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0163] In summary, this invention collects environmental and building equipment data using intelligent sensors, preprocesses the two sets of data, defines building equipment as network nodes, defines the connections between building equipment as edges, uses the Kruskal algorithm to fuse the multimodal association topology of these two sets of data, constructs a minimum spanning tree, uses the reciprocal of the edge weight as the pheromone value, and uses the maximum flow algorithm-ant colony algorithm to select the optimal control path; it uses the Takens embedding method to transform phase space data, randomly selects a pair of adjacent phase space data from the time series as initial points, uses the Lyapunov exponent to analyze the state of building equipment, constructs a hidden Markov model, and uses the Lyapunov exponent to dynamically correct the transition probability matrix and emission probability matrix to obtain the predicted state of building equipment; it uses a linear weighted method to construct a comprehensive objective function, uses the particle swarm optimization algorithm to solve it to obtain the optimal individual, uses a dynamic scheduling algorithm to convert it into scheduling instructions for execution, and uses a PID control algorithm for monitoring and adjustment; this improves response speed, resource utilization, and control accuracy, avoids the impact of equipment failure on building operation, and enhances the comprehensiveness and adaptability of equipment control.

[0164] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for controlling intelligent building equipment, characterized in that: include, The system uses smart sensors to collect environmental and building equipment data. The two sets of data are preprocessed, and the building equipment is defined as network nodes. The connection between the building equipment is defined as an edge. The Kruskal algorithm is used to fuse the multimodal association topology of the two sets of data to construct a minimum spanning tree. The inverse of the edge weight is used as the pheromone value. The maximum flow algorithm-ant colony algorithm is used to select the optimal control path. The Takens embedding method is used to transform phase space data. A pair of adjacent phase space data is randomly selected from the time series as the initial point. The Lyapunov exponent is used to analyze the status of building equipment, and a hidden Markov model is constructed. The Lyapunov exponent is used to dynamically correct the transition probability matrix and the emission probability matrix to obtain the predicted status of building equipment. A comprehensive objective function is constructed using a linear weighting method, solved using a particle swarm optimization algorithm to obtain the optimal individual, converted into scheduling instructions using a dynamic scheduling algorithm and executed, and monitored and adjusted using a PID control algorithm. Store the two sets of data generated from collection and analysis; The weighting coefficients for bandwidth and connection delay data are calculated separately using the analytic hierarchy process (AHP). The weighted summation method is then used to combine the bandwidth and connection delay data to calculate the connection weights between building equipment, which are defined as edge weights. Sort the edge weights in ascending order, select the edge with the smallest weight as the first connection, then select the next edge with a smaller weight as the second connection. Use the disjoint-set data structure algorithm to progressively select edges connecting building equipment from the sorted edge weights until all building equipment is connected into a connected graph, and stop the selection to generate the minimum spanning tree. The ant colony algorithm is used to select the optimal control path based on the pheromone concentration and quality of the optimal path. The time series data is converted into phase space data by using mutual information and FNN methods to set the delay time and embedding dimension, and the Takens embedding method is used to convert the time series into phase space data. Calculate the Euclidean distance between the initial points using the Euclidean distance formula; A second time series is generated using a small perturbation method on the initial point, and the Euclidean distance between the first time series and the second time series is calculated using the Euclidean distance formula. The Lyapunov exponent for time t is calculated using the tangent vector method. If the Lyapunov exponent at time t is greater than 0, it indicates that the building equipment is unstable. The control factors of the initial transition probability matrix and the initial emission probability matrix are set by cross-validation, and the adjustment range of the transition probability and emission probability is calculated by dynamic adjustment method. If the Lyapunov exponent of time t is less than or equal to 0, it indicates that the building equipment is normal, and the initial transition probability matrix and the initial emission probability matrix are used. The maximum probability of building equipment being in a hidden state at time t is obtained using the Viterbi algorithm. The transition probabilities in the initial transition probability matrix and the emission probabilities in the initial emission probability matrix are dynamically corrected using the proportional adjustment method, resulting in dynamically corrected transition probability matrices and emission probability matrices. The transition probability matrix and emission probability matrix are optimized using maximum a posteriori estimation to obtain the final transition probability matrix and emission probability matrix.

2. The intelligent building equipment control method as described in claim 1, characterized in that: The process involves using smart sensors to collect environmental and building equipment data, defining building equipment as network nodes, defining connections between building equipment as edges, fusing the multimodal association topology of these two sets of data using the Kruskal algorithm, constructing a minimum spanning tree, using the reciprocal of the edge weight as the pheromone value, and using the maximum flow algorithm-ant colony algorithm to select the optimal control path, including: The intelligent sensors include sensors for monitoring light, temperature and humidity, vibration, bandwidth, and delay. The environmental and building equipment data includes light intensity, temperature and humidity, vibration, bandwidth, and connection latency data; Use IoT protocols to connect sensors to the central control system; The building equipment includes air conditioning, elevators, lighting, and smart sensors; The ant colony algorithm is used to calculate the path selection probability, the edge weight is set as the path quality, heuristic information is used to select the path, the empirical rule is used to set the pheromone evaporation coefficient, and the pheromone increment formula is used to calculate the pheromone increment. After the ant selects a path, the pheromone concentration of the path is updated using a local pheromone update formula. After each ant completes its path selection and pheromone update, it proceeds to the next selection, until all ants have completed their path selection and the update process stops. After all ants have completed their path selection, the optimal path quality is calculated using a weighted summation method, the optimal pheromone increment is calculated using the pheromone increment formula, and the pheromone concentration of the optimal path is updated using the global pheromone update formula.

3. The intelligent building equipment control method as described in claim 2, characterized in that: The process involves using Takens embedding to transform phase space data, randomly selecting a pair of adjacent phase space data from the time series as initial points, analyzing building equipment status using the Lyapunov exponent, constructing a hidden Markov model, and dynamically correcting the transition probability matrix and emission probability matrix using the Lyapunov exponent to obtain the predicted status of the building equipment, including: Collect and preprocess historical environmental data to generate a training set; Constructing a Hidden Markov Model and training the Hidden Markov Model using a training set includes using threshold discretization to discretize the preprocessed historical environmental data into hidden states, wherein the hidden states include normal, fault, and degradation. The number of transitions from one hidden state to another in the preprocessed historical environmental data is statistically analyzed. The transition probability between each pair of states is calculated using maximum likelihood. The transition probabilities are then filled into the transition probability matrix and normalized to obtain the initial transition probability matrix. The number of times the preprocessed historical environmental data appears in the hidden state is counted. The emission probability of each hidden state is calculated using maximum likelihood. The emission probabilities are filled into the emission probability matrix and normalized to obtain the initial emission probability matrix. The preprocessed environmental data is sorted in chronological order to generate a time series. Real-time environmental data is input into a trained Hidden Markov Model to obtain the predicted state of building equipment.

4. The intelligent building equipment control method as described in claim 3, characterized in that: The process of constructing a comprehensive objective function using a linear weighting method and solving it using a particle swarm optimization algorithm to obtain the optimal individual includes: The predicted status of building equipment is normalized. Calculate the mean values ​​of the preprocessed temperature and humidity data respectively, and set them as the ideal temperature and humidity data; The temperature and humidity deviation method is used to set the comfort objective function; Use the mean absolute error to define the loss function; The weight coefficients of the comfort objective function and the loss function are calculated separately using a data-driven method, and a comprehensive objective function is constructed using a linear weighting method. An initial population is randomly generated, with each individual representing a combination of prediction error and comfort requirement. Individuals are evaluated by substituting them into a comprehensive objective function, and their fitness values ​​are calculated. Individuals are randomly selected for comparison using a tournament selection method, and individuals with higher fitness are selected as parents to participate in the crossover operation. New offspring are produced by exchanging the genes of parent individuals at a single point of crossover. After crossover, individuals are randomly selected and mutated using single-point mutation to generate new individuals; The maximum number of iterations is set using a halting criterion. When the maximum number of iterations is reached, iteration stops, and the optimal individual is output, including the optimized prediction error and comfort requirements.

5. The intelligent building equipment control method as described in claim 4, characterized in that: The process of converting dynamic scheduling algorithms into scheduling instructions and executing them, and using PID control algorithms for monitoring and adjustment, includes: The optimal individual is converted into a scheduling instruction using a dynamic scheduling algorithm, which is then transmitted to the building equipment via an API interface. The building equipment receives the scheduling instruction and executes it. Collect feedback data and calculate the difference between the feedback data and the optimal individual; A judgment threshold is set using statistical methods. Differences are compared with the judgment threshold. Differences greater than or equal to the judgment threshold are adjusted using a PID control algorithm until the difference is less than the judgment threshold. The adjustment stops then, and the feedback data is monitored.

6. The intelligent building equipment control method as described in claim 5, characterized in that: The preprocessing of the two sets of collected data includes: The two sets of collected data were aligned in time using a timestamp alignment method, outliers were identified and removed using the quartile method, missing values ​​were filled using the mean imputation method, and the two sets of imputed data were normalized.

7. The intelligent building equipment control method as described in claim 6, characterized in that: The two sets of data collected and analyzed include: The two sets of collected data and the optimal individuals generated from the analysis are stored in a central database, and secure access measures are set up. The central database backs up the stored data to the cloud and performs integrity checks on the stored data and backup data regularly. After the checks are completed, integrity check records are generated and stored synchronously in the central database.

8. A smart building equipment control system, based on the smart building equipment control method according to any one of claims 1 to 7, characterized in that: include, The topology optimization module is used to collect environmental and building equipment data using smart sensors. It preprocesses the two sets of data, defines building equipment as network nodes, defines the connections between building equipment as edges, uses the Kruskal algorithm to fuse the multimodal association topology of the two sets of data, constructs a minimum spanning tree, uses the reciprocal of the edge weight as the pheromone value, and uses the maximum flow algorithm-ant colony algorithm to select the optimal control path. The state prediction module is used to transform phase space data using the Takens embedding method. It randomly selects a pair of adjacent phase space data from the time series as the initial point, uses the Lyapunov exponent to analyze the state of building equipment, constructs a hidden Markov model, and uses the Lyapunov exponent to dynamically correct the transition probability matrix and emission probability matrix to obtain the predicted state of the building equipment. The scheduling and monitoring module is used to construct a comprehensive objective function using a linear weighting method, solve it using a particle swarm optimization algorithm to obtain the optimal individual, convert it into scheduling instructions using a dynamic scheduling algorithm and execute them, and monitor and adjust it using a PID control algorithm. The storage module is used to store the two sets of data generated during collection and analysis.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the intelligent building equipment control method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the intelligent building equipment control method according to any one of claims 1 to 7.

Images