Air conditioner energy-saving control system based on environmental perception

Abstract

The invention relates to the technical field of energy-saving control systems, in particular to an air conditioner energy-saving control system based on environmental perception. According to the technical scheme, the system comprises an environment sensing module, a central control module, an execution mechanism and a strong and weak current integrated platform. Environmental factors are sensed in real time through the multi-parameter sensor network, and the optimization weight of each parameter is dynamically adjusted in combination with a dynamic optimization algorithm, so that the system can be matched with the operation power of the air conditioning system and the actual load requirement, the building energy consumption is reduced, and the energy cost is saved; meanwhile, an edge computing architecture is adopted to achieve rapid analysis and decision making of environmental parameters, the response speed of the system is increased, in addition, the continuity and stability of system operation are guaranteed through a fault self-healing mechanism, the adaptability of the system to different environments and use requirements is improved through a genetic algorithm, and the system reliability is improved. And the strong and weak current integrated platform reduces wiring complexity and system cost, can be integrated with an existing building automatic control system, and reduces reconstruction difficulty and cost.

Description

Technical Field

[0001] The present invention relates to the technical field of energy-saving control systems, and in particular to an air-conditioning energy-saving control system based on environment perception. Background Art

[0002] The air conditioning energy-saving control system mainly consists of three parts: sensors, controllers, and actuators. The sensors are responsible for real-time monitoring of indoor and outdoor environmental parameters such as temperature and humidity, and transmit this data to the controller. The controller, as the core of the system, calculates the optimal control strategy based on the received data, combined with the preset temperature setting value and energy-saving algorithm. The actuator adjusts the operating status of the air conditioning equipment according to the controller's instructions to achieve a balance between energy saving and comfort. With the development of intelligent buildings, air conditioning energy-saving control systems play a key role in improving building energy efficiency and improving indoor environmental quality. Traditional air conditioning energy-saving control systems mainly rely on fixed temperature settings or single environmental parameters for adjustment, which lacks intelligence. The generation of control instructions also depends on a single or a few parameters. It lacks the ability to collaboratively analyze multiple parameters, cannot cope with complex environmental changes, and has a delayed response. Therefore, we propose an air conditioning energy-saving control system based on environmental perception. Summary of the Invention

[0003] The purpose of the present invention is to propose an air conditioning energy-saving control system based on environmental perception to address the problems of insufficient intelligence, lack of multi-parameter collaborative analysis capabilities, and delayed response of traditional air conditioning energy-saving control systems in the background technology.

[0004] The technical solution of the present invention is an air-conditioning energy-saving control system based on environmental perception, which includes an environmental perception module, a central control module, an actuator and a strong and weak electricity integrated platform, wherein:

[0005] The environmental perception module is used to collect environmental data and pre-process the environmental data;

[0006] The central control module is used to analyze the environmental data processed by the environmental perception module, make control decisions and optimizations, and issue control instructions;

[0007] The actuator is used to adjust the operation of the air-conditioning equipment according to the control instructions given by the central control module;

[0008] The integrated power and weak electricity platform is used to ensure communication and power supply between the environmental perception module, central control module, and actuator.

[0009] Optionally, the environment perception module includes a multi-parameter sensor network and a data fusion unit, wherein:

[0010] The multi-parameter sensor network includes a sensor network integrating a temperature and humidity sensor, a CO2 concentration sensor, a PM2.5 detector, an infrared people counter and a light intensity sensor;

[0011] The data fusion unit performs a fusion operation on the sensor data monitored by the multi-parameter sensor network through an extended Kalman filter algorithm and extracts key characteristic parameters, wherein the system state vector x = [T, H] T , where T represents the indoor temperature, H represents the indoor humidity, and the state equation is a random walk model, which is expressed as follows:

[0012]

[0013] Among them, w T,k-1 and w H,k-1 are the process noise of temperature T and humidity H respectively;

[0014] The observation equation is:

[0015]

[0016] Among them, z T,k and z H,k are the measured values ​​of the temperature sensor and humidity sensor, respectively, v T,k and v H,k are the observation noises corresponding to the measurement values ​​of the temperature sensor and the humidity sensor respectively;

[0017] Dynamic weights are introduced when calculating the key characteristic parameters, where the number of people N is given a high weight w when calculating the population density D. N , the calculation formula is:

[0018]

[0019] Where S is the space area;

[0020] And according to the air quality index calculation formula and the air quality parameters measured by the environmental perception module, the comprehensive air quality index AQI is obtained. When calculating the AQI sub-item value corresponding to each pollutant, dynamic weights are assigned according to time and scene, among which, The calculation formula is:

[0021]

[0022] in, is the CO2 concentration, is the conversion function from CO2 concentration to AQI sub-item value, is a dynamic weight value. The final comprehensive air quality index AQI formula is expressed as:

[0023]

[0024] Among them, AOI PM2.5 It is the AQI sub-item value corresponding to the PM2.5 concentration value.

[0025] Optionally, the central control module includes a multi-objective optimization model based on a dynamic optimization algorithm, a fuzzy logic reasoning unit based on genetic algorithm optimization, and a priority strategy unit;

[0026] The multi-objective optimization model includes optimization objectives, which include energy-saving index, comfort and air quality index. The energy-saving objective function is E(f,α,…), where f is the frequency of the variable frequency compressor, α is the control variable such as the opening of the fresh air valve, and the comfort objective function is C(T i ,T s ,…), where T i is the actual indoor temperature, T s To set the comfortable temperature, the air quality objective function is A(AQI), where AQI is the air quality index. The comprehensive objective function formula is expressed as:

[0027] F=w E ×E+w C ×C+w A ×A

[0028] Among them, w E 、w C 、w A are the weights of energy saving index, comfort level and air quality index respectively, and w E +w C +w A =1;

[0029] The fuzzy logic reasoning unit includes a fuzzy rule base, which is used to map environmental parameters into control decisions, wherein the control decision includes the frequency of the variable frequency compressor, and the environmental parameters include indoor and outdoor temperature and humidity, human density and air quality index, wherein the fuzzified input variables indoor and outdoor temperature and humidity ΔT=T i -T s , personnel density D and air quality index AQI, and obtain the fuzzy output of the compressor frequency through fuzzy reasoning according to the preset fuzzy rules, and then obtain the frequency value through defuzzification operation.

[0030] Optionally, the fuzzy rule base is encoded by a genetic algorithm, each fuzzy rule is regarded as a gene, and multiple fuzzy rules form a chromosome. Assuming that there are n rules in the fuzzy rule base, the chromosome is represented by an ordered gene sequence Chr=[g1,g1,…,g n ], where gi represents the i-th fuzzy rule. The genetic algorithm iteratively optimizes the chromosome through selection, crossover, and mutation operations. The selection operation adopts the roulette wheel selection method. Suppose there are m chromosomes in the population and the fitness value of the j-th chromosome is F j , then the probability P that the jth chromosome is selected j The calculation formula is:

[0031]

[0032] The crossover operation sets the probability p c Gene exchange is performed on two randomly selected parent chromosomes to generate offspring chromosomes. The mutation operation is performed with a set probability p m Randomly change the genes on the chromosomes and introduce new genes.

[0033] Optionally, the priority strategy unit is used to evaluate the environmental anomaly level and perform real-time analysis through a neural network model embedded in the local controller, wherein the environmental anomaly level index R is assumed to be:

[0034]

[0035] Among them, C PM2.5 is the PM2.5 concentration, C lim is the national standard limit for PM2.5 concentration;

[0036] The neural network model takes environmental parameters as input and outputs the predicted value of environmental anomaly level after multi-layer convolution and pooling operations. The starting signal S of the fresh air system is:

[0037]

[0038] when When the system sets the improvement of air quality index as high priority, it starts the fresh air system.

[0039] Optionally, the actuator includes a variable frequency compressor and an intelligent air valve, wherein:

[0040] The variable frequency compressor uses a vector control algorithm to control the voltage and frequency of the motor. In a three-phase AC motor, the stator current is decomposed into the excitation current component i d and the torque current component i q , respectively control the magnetic flux and torque of the motor. According to the mathematical model of the motor, the stator voltage equation is:

[0041]

[0042] Among them, u d ,uq are the d-axis and q-axis components of the stator voltage, R s is the stator resistance, L s is the stator inductance, ω e is the motor electrical angular velocity;

[0043] The load demand is predicted when the sensor fails by using the fault self-recovery prediction model based on the long short-term memory network. The input is the historical operation data sequence X = [x1, x1, ... x n ], where x i Including the frequency, voltage, current, and cooling capacity information of the variable frequency compressor, the long short-term memory network processes the input sequence through the input gate, forget gate, output gate, and memory unit, and outputs the predicted load demand;

[0044] The intelligent air valve is based on the return air temperature T r The air valve opening α is controlled by a linkage control algorithm with the air quality index AQI, and the control quantity u of the air valve opening is calculated by the PID control algorithm. The control quantity u is expressed as follows:

[0045]

[0046] Where, e = T r-set -T r or e=AQI set -AQI, where T r-set is the return air temperature set point, AQI set is the air quality index setting value, K p is the proportionality coefficient, K i is the integral coefficient, K d is the differential coefficient, and α=f(u).

[0047] Optionally, the actuator further includes a cold and hot source group control subsystem based on a distributed collaborative control algorithm, the cold and hot source group control subsystem includes an ECS-7000MR controller, and the ECS-7000MR controller is used to distribute and coordinate the load among multiple units. Assume that the total system load demand is L, there are i units in total, and the load distribution coefficient of each unit is λ i (i=1,2,…,n), then the load of the i-th unit is:

[0048] L i =λ i L

[0049] in, Load distribution coefficient λ iDynamically adjust according to the performance parameters and current operating status of each unit. Each local controller uses edge computing to analyze the local unit operating data and environmental parameters in real time and calculate the load distribution coefficient adjustment value Δλ of the local unit. i , and the strong and weak electricity integrated platform communicates with other controllers.

[0050] Optionally, the integrated platform for strong and weak electricity includes an internal bus communication and fault processing unit and a LONWORKS bus communication mechanism based on edge computing optimization;

[0051] The internal bus communication and fault handling unit includes an internal bus for communication between modules. The data frame format is set to [Header, Data, CRC], where the Header contains the source address, destination address, and data type information, the Data is the actual transmitted data, and the CRC is a cyclic redundancy check code used for data verification.

[0052] Optionally, when a communication failure occurs at a node in the internal bus, the internal bus communication and fault processing unit automatically switches to a backup communication path. The main communication path is P1, the backup communication path is P2, and the node status monitoring variable is S i , i is the node number), S i =1 means the node is normal, S i =0 indicates a node failure. When a node failure is detected, the communication path switching control signal C is:

[0053]

[0054] Assume that the historical communication data is H = [h1,h1,…h m ], where h k Contains the data sent and received by the faulty node at different times, and uses the sliding average method to predict future data demand. The formula is:

[0055]

[0056] Where N is the sliding window size.

[0057] Optionally, when the LONWORKS bus communication mechanism based on edge computing optimization sends data, the neuron chip encapsulates the upper-layer application data into a data frame that complies with the LONWORKS protocol, adds the source address, destination address, and check code information, and after the neuron chip at the receiving end receives the data frame, it determines whether it is its own data based on the destination address. If so, it performs verification and decapsulation and passes the data to the upper-layer application;

[0058] Each controller in the system pre-processes and caches the data to be sent and received through edge computing capabilities. When sending data, the local controller performs a preliminary analysis of the environmental parameters, calculates key feature data, and transmits the key feature data through the LONWORKS bus. Assume that the amount of original environmental parameter data is D0, the amount of key feature data after edge computing pre-processing is D1, and the data compression ratio is When receiving data, edge computing is used to cache and pre-parse the data.

[0059] In summary, this application includes at least one of the following beneficial technical effects:

[0060] This invention uses a multi-parameter sensor network to perceive indoor and outdoor temperature and humidity, human density, and air quality environmental factors in real time. In combination with a dynamic optimization algorithm, it dynamically adjusts the optimization weights of various parameters according to different times and scenarios, enabling the system to accurately match the operating power of the air-conditioning system with the actual load demand, effectively reducing building energy consumption and saving energy costs. At the same time, it adopts an edge computing architecture to achieve rapid analysis and decision-making of environmental parameters, reduce data transmission delays, and use a priority strategy to dynamically adjust the control priority according to the level of environmental anomalies, thereby improving the system's response speed, reducing the time to generate control instructions, effectively improving indoor comfort, and providing users with a more comfortable indoor environment.

[0061] The fault self-healing mechanism in the present invention can automatically switch to a prediction mode based on historical data when a sensor fails, and adopts a long-short-term memory network model to predict the current environmental status based on historical environmental parameters and operating data, thereby ensuring the continuity and stability of system operation, avoiding system operation abnormalities caused by sensor failure, and improving system reliability and stability;

[0062] The present invention optimizes the fuzzy rule base through genetic algorithms, and continuously iterates to find the optimal fuzzy rule combination through selection, crossover, and mutation operations, so that the control strategy can better adapt to complex and changing environmental conditions, and improve the system's adaptability to different environments and usage requirements;

[0063] The integrated power and weak current platform in the present invention adopts the ECS-7000M series controller, which reduces the wiring complexity and system cost. At the same time, it uses the LONWORKS bus to realize peer-to-peer communication between controllers and adopts standardized communication protocols and interface design. It can be seamlessly integrated with the existing building automation system, reducing the transformation cost and implementation cycle, facilitating the upgrade and expansion of the system. For the air-conditioning system transformation project of existing buildings, it can better adapt to the original system architecture and reduce the difficulty and cost of transformation. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1The structural block diagram of the air-conditioning energy-saving control system based on environmental perception of the present invention is given. DETAILED DESCRIPTION

[0065] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0066] Example

[0067] The air conditioning energy-saving control system based on environmental perception proposed in this invention is as follows: Figure 1 As shown, it includes an environmental perception module, a central control module, an actuator and a strong and weak electricity integrated platform.

[0068] The environmental perception module is used to collect and pre-process environmental data, including a multi-parameter sensor network and a data fusion unit, where:

[0069] The multi-parameter sensor network includes a sensor network that integrates temperature and humidity sensors, CO2 concentration sensors, PM2.5 detectors, infrared people counters and light intensity sensors. For areas with dense population and relatively poor ventilation, a temperature and humidity sensor is deployed every 50 square meters, a CO2 concentration sensor is set up every 80 square meters, and a PM2.5 detector is deployed every 120 square meters. In other areas, a temperature and humidity sensor, a CO2 concentration sensor and a PM2.5 detector are deployed every 100 square meters. At least one infrared people counter is installed at the entrances and exits of each area, and a light intensity sensor is installed every 150 square meters in window areas and public lighting areas. All sensors are connected to the data fusion unit via wired or wireless means. Various sensors in the environmental perception module collect environmental data in real time with a cycle of 10 seconds and transmit the data to the data fusion unit.

[0070] The data fusion unit performs fusion operations on the sensor data monitored by the multi-parameter sensor network through the extended Kalman filter algorithm and extracts key characteristic parameters. In which, the system state vector x = [T, H] T , where T represents the indoor temperature, H represents the indoor humidity, and the state equation is a random walk model, which is expressed as follows:

[0071]

[0072] Among them, w T,k-1 and w H,k-1 are the process noise of temperature T and humidity H respectively;

[0073] The observation equation is:

[0074]

[0075] Among them, z T,k and z H,k are the measured values ​​of the temperature sensor and humidity sensor, respectively, v T,k and v H,k are the observation noise corresponding to the measurement values ​​of the temperature sensor and the humidity sensor, respectively. Dynamic weights are introduced when calculating key characteristic parameters. When calculating the population density D, a high weight w is given to the number of people N. N , the calculation formula is:

[0076]

[0077] Among them, S is the spatial area. At the same time, according to the air quality index calculation formula and the air quality parameters measured by the environmental perception module, the comprehensive air quality index AQI is obtained. When calculating the AQI sub-item value corresponding to each pollutant, dynamic weights are assigned according to time and scene. Among them, The calculation formula is:

[0078]

[0079] in, is the CO2 concentration, is the conversion function from CO2 concentration to AQI sub-item value, is a dynamic weight value. The final comprehensive air quality index AQI formula is expressed as:

[0080]

[0081] Among them, AOI PM2.5 It is the AQI sub-item value corresponding to the PM2.5 concentration value.

[0082] The present invention uses a multi-parameter sensor network to perceive indoor and outdoor temperature and humidity, population density, and air quality environmental factors in real time, and combines it with a dynamic optimization algorithm to dynamically adjust the optimization weights of various parameters according to different times and scenarios, so that the system can accurately match the operating power of the air-conditioning system with the actual load demand, effectively reducing building energy consumption and saving energy costs.

[0083] The central control module is used to analyze the environmental data processed by the environmental perception module and make control decisions and optimizations, and give control instructions. It includes a multi-objective optimization model based on a dynamic optimization algorithm, a fuzzy logic reasoning unit based on genetic algorithm optimization, and a priority strategy unit. The central control module uses a high-performance server as the hardware platform of the central control module and installs customized control software. The software integrates a multi-objective optimization model based on fuzzy logic and genetic algorithm, as well as a priority strategy algorithm.

[0084] The multi-objective optimization model includes optimization objectives, which include energy saving index, comfort and air quality index. The energy saving objective function is E(f,α,…), where f is the frequency of the variable frequency compressor, α is the control variable such as the opening of the fresh air valve, and the comfort objective function is C(T i ,T s ,…), where T i is the actual indoor temperature, T s To set the comfortable temperature, the air quality objective function is A(AQI), where AQI is the air quality index. The comprehensive objective function formula is expressed as:

[0085] F=w E ×E+w C ×C+w A ×A

[0086] Among them, w E 、w C 、w A are the weights of energy saving index, comfort level and air quality index respectively, and w E +w C +w A =1, the present invention adopts edge computing architecture to achieve rapid analysis and decision-making of environmental parameters, reducing data transmission delay.

[0087] The fuzzy logic reasoning unit includes a fuzzy rule base, which is used to map environmental parameters into control decisions, wherein the control decision includes the frequency of the variable frequency compressor, and the environmental parameters include indoor and outdoor temperature and humidity, occupancy density, and air quality index. The fuzzified input variables indoor and outdoor temperature and humidity ΔT = T i -T s , population density D and air quality index AQI, and obtain the fuzzy output of the compressor frequency through fuzzy reasoning according to the preset fuzzy rules, and then obtain the frequency value through defuzzification operation, and encode the fuzzy rule base through genetic algorithm. Each fuzzy rule is regarded as a gene, and multiple fuzzy rules form a chromosome. Assuming that there are n rules in the fuzzy rule base, the chromosome is represented by an ordered gene sequence Chr=[g1,g1,…,g n ], where g iRepresents the i-th fuzzy rule. The genetic algorithm iteratively optimizes the chromosome through selection, crossover, and mutation operations. The selection operation adopts the roulette wheel selection method. Suppose there are m chromosomes in the population and the fitness value of the j-th chromosome is F j , then the probability P that the jth chromosome is selected j The calculation formula is:

[0088]

[0089] Crossover operation to set the probability p c Gene exchange is performed on two randomly selected parent chromosomes to generate offspring chromosomes, and the mutation operation is performed with a set probability p m The genes of chromosomes are randomly changed, new genes are introduced, the fuzzy rule base is optimized through genetic algorithms, and the optimal fuzzy rule combination is continuously iterated through selection, crossover, and mutation operations, so that the control strategy can better adapt to complex and changeable environmental conditions, and improve the system's adaptability to different environments and usage requirements.

[0090] The priority strategy unit is used to evaluate the level of environmental anomaly and perform real-time analysis through the neural network model embedded in the local controller. The environmental anomaly level index R is assumed to be:

[0091]

[0092] Among them, C PM2.5 is the PM2.5 concentration, C lim The national standard limit for PM2.5 concentration is used. The neural network model takes environmental parameters as input and outputs the predicted value of environmental abnormality level after multi-layer convolution and pooling operations. The starting signal S of the fresh air system is:

[0093]

[0094] when When the air quality index is improved, the system sets the improvement as a high priority and starts the fresh air system. The local controller embedded with the neural network model is distributed between the equipment on each floor. It is responsible for data interaction with the environmental perception module and preliminary data processing and analysis. The local controller communicates with the central control server through a high-speed local area network, uploads the processed key data to the server, and receives the control instructions issued by the server at the same time. The present invention uses a priority strategy to dynamically adjust the control priority according to the level of environmental abnormality, which improves the response speed of the system, reduces the generation time of control instructions, effectively improves indoor comfort, and provides users with a more comfortable indoor environment.

[0095] The actuators are used to adjust the operation of the air conditioning equipment according to the control instructions given by the central control module. They include 50 variable frequency compressors and 300 intelligent air valves, including:

[0096] The variable frequency compressor uses a vector control algorithm to control the voltage and frequency of the motor. In a three-phase AC motor, the stator current is decomposed into the excitation current component i d and the torque current component i q , respectively control the magnetic flux and torque of the motor. According to the mathematical model of the motor, the stator voltage equation is:

[0097]

[0098] Among them, u d ,u q are the d-axis and q-axis components of the stator voltage, R s is the stator resistance, L s is the stator inductance, ω e is the motor electrical angular velocity, and the load demand is predicted when the sensor fails through the fault self-healing prediction model based on the long short-term memory network. Assume that the input is the historical operation data sequence X = [x1, x1, ... x n ], where x i Including variable frequency compressor frequency, voltage, current, and cooling capacity information, the long short-term memory network processes the input sequence through input gates, forget gates, output gates, and memory units, and outputs the predicted load demand.

[0099] The intelligent air valves are installed at each branch of the air conditioning duct. The intelligent air valves are based on the return air temperature T r The air valve opening α is controlled by the linkage control algorithm with the air quality index AQI, and the control quantity u of the air valve opening is calculated by the PID control algorithm. The control quantity u formula is expressed as:

[0100]

[0101] Where, e = T r-set -T r or e=AQI set -AQI, where T r-set is the return air temperature set point, AQI set is the air quality index setting value, K p is the proportionality coefficient, K i is the integral coefficient, K d is the differential coefficient, and α=f(u). When the AQI exceeds the standard or the return air temperature deviates greatly from the set value, the intelligent air valve increases its opening to enhance air circulation and heat exchange efficiency.

[0102] The actuator also includes a cold and hot source group control subsystem based on a distributed collaborative control algorithm. The cold and hot source group control subsystem includes an ECS-7000MR controller. The ECS-7000MR controller is used to distribute and coordinate the load among multiple units. Assume that the total system load demand is L, there are i units in total, and the load distribution coefficient of each unit is λ i (i=1,2,…,n), then the load of the i-th unit is:

[0103] L i =λ i L

[0104] in, Load distribution coefficient λ i Dynamically adjust according to the performance parameters and current operating status of each unit. Each local controller uses edge computing to analyze the local unit operating data and environmental parameters in real time and calculate the load distribution coefficient adjustment value Δλ of the local unit. i , and the strong and weak electricity integrated platform communicates with other controllers.

[0105] The integrated power and weak current platform is used to ensure communication and power supply between the environmental perception module, central control module, and actuators. It includes an internal bus communication and fault handling unit and a LONWORKS bus communication mechanism optimized based on edge computing. 80 ECS-7000 series controllers with integrated sensor interfaces, motor protection, and energy consumption metering functions are installed in the equipment rooms and distribution boxes on each floor.

[0106] The internal bus communication and fault handling unit includes an internal bus for communication between modules. The data frame format is [Header, Data, CRC], where Header contains source address, destination address, and data type information, Data is the actual transmitted data, and CRC is the cyclic redundancy check code used for data verification. When a communication failure occurs at a node on the internal bus, the internal bus communication and fault handling unit automatically switches to the backup communication path. The main communication path is P1, the backup communication path is P2, and the node status monitoring variable is S i , i is the node number), S i =1 means the node is normal, S i =0 indicates a node failure. When a node failure is detected, the communication path switching control signal C is:

[0107]

[0108] Assume that the historical communication data is H = [h1,h1,…h m ], where h k Contains the data sent and received by the faulty node at different times, and uses the sliding average method to predict future data demand. The formula is:

[0109]

[0110] Among them, N is the sliding window size. The fault self-healing mechanism can automatically switch to the prediction mode based on historical data when the sensor fails, and adopt the long-short-term memory network model to predict the current environmental status according to historical environmental parameters and operating data, thereby ensuring the continuity and stability of system operation, avoiding system operation abnormalities caused by sensor failure, and improving the reliability and stability of the system.

[0111] When sending data, the LONWORKS bus communication mechanism optimized for edge computing encapsulates the upper-layer application data into a data frame that complies with the LONWORKS protocol, adding the source address, destination address, and checksum information. After receiving the data frame, the receiving end neuron chip determines whether it is its own data based on the destination address. If so, it verifies and decapsulates the data and passes it to the upper-layer application.

[0112] Each controller in the system pre-processes and caches the data to be sent and received through edge computing capabilities. When sending data, the local controller performs a preliminary analysis of the environmental parameters, calculates key feature data, and transmits the key feature data through the LONWORKS bus. Assume that the amount of original environmental parameter data is D0, the amount of key feature data after edge computing pre-processing is D1, and the data compression ratio is When receiving data, edge computing is used to cache and pre-parse the data.

[0113] The integrated power and weak current platform in the present invention adopts the ECS-7000M series controller, which reduces the wiring complexity and system cost. At the same time, it uses the LONWORKS bus to realize peer-to-peer communication between controllers and adopts standardized communication protocols and interface design. It can be seamlessly integrated with the existing building automation system, reducing the transformation cost and implementation cycle, facilitating the upgrade and expansion of the system. For the air-conditioning system transformation project of existing buildings, it can better adapt to the original system architecture and reduce the difficulty and cost of transformation.

[0114] The above specific embodiments are merely several optional embodiments of the present invention. Based on the technical solutions of the present invention and the relevant inspirations of the above embodiments, those skilled in the art may make various alternative improvements and combinations to the above specific embodiments.

Claims

1. Air conditioning energy-saving control system based on environmental perception, characterized by: It includes an environmental perception module, a central control module, an actuator, and a strong and weak electricity integrated platform, including: The environmental perception module is used to collect environmental data and pre-process the environmental data; The central control module is used to analyze the environmental data processed by the environmental perception module, make control decisions and optimizations, and issue control instructions; The actuator is used to adjust the operation of the air-conditioning equipment according to the control instructions given by the central control module; The integrated power and weak electricity platform is used to ensure communication and power supply between the environmental perception module, central control module, and actuator.

2. The air conditioning energy-saving control system based on environmental perception according to claim 1 is characterized in that: The environment perception module includes a multi-parameter sensor network and a data fusion unit, wherein: The multi-parameter sensor network includes a sensor network integrating a temperature and humidity sensor, a CO2 concentration sensor, a PM2.5 detector, an infrared people counter and a light intensity sensor; The data fusion unit performs a fusion operation on the sensor data monitored by the multi-parameter sensor network through an extended Kalman filter algorithm and extracts key characteristic parameters, wherein the system state vector x = [T, H] T , where T represents the indoor temperature, H represents the indoor humidity, and the state equation is a random walk model, which is expressed as follows: Among them, w T,k-1 and w H,k-1 are the process noise of temperature T and humidity H respectively; The observation equation is: Among them, z T,k and z H,k are the measured values ​​of the temperature sensor and humidity sensor, respectively, v T,k and v H,k are the observation noises corresponding to the measurement values ​​of the temperature sensor and the humidity sensor respectively; Dynamic weights are introduced when calculating the key characteristic parameters, where the number of people N is given a high weight w when calculating the population density D. N , the calculation formula is: Where S is the space area; And according to the air quality index calculation formula and the air quality parameters measured by the environmental perception module, the comprehensive air quality index AQI is obtained. When calculating the AQI sub-item value corresponding to each pollutant, dynamic weights are assigned according to time and scene, among which, The calculation formula is: in, is the CO2 concentration, is the conversion function from CO2 concentration to AQI sub-item value, is a dynamic weight value, and the comprehensive air quality index AQI formula is expressed as: Among them, AOI PM2.5 It is the AQI sub-item value corresponding to the PM2.5 concentration value.

3. The air conditioning energy-saving control system based on environmental perception according to claim 1, characterized in that: The central control module includes a multi-objective optimization model based on a dynamic optimization algorithm, a fuzzy logic reasoning unit based on genetic algorithm optimization, and a priority strategy unit; The multi-objective optimization model includes optimization objectives, which include energy-saving index, comfort and air quality index. The energy-saving objective function is E(f,α,…), where f is the frequency of the variable frequency compressor, α is the control variable such as the opening of the fresh air valve, and the comfort objective function is C(T i ,T s ,…), where T i is the actual indoor temperature, T s To set the comfortable temperature, the air quality objective function is A(AQI), where AQI is the air quality index. The comprehensive objective function formula is expressed as: F=w E ×E+w C ×C+w A ×A Among them, w E 、w C 、w A are the weights of energy saving index, comfort level and air quality index respectively, and w E +w C +w A =1; The fuzzy logic reasoning unit includes a fuzzy rule base, which is used to map environmental parameters into control decisions, wherein the control decision includes the frequency of the variable frequency compressor, and the environmental parameters include indoor and outdoor temperature and humidity, human density and air quality index, wherein the fuzzified input variables indoor and outdoor temperature and humidity ΔT=T i -T s , personnel density D and air quality index AQI, and obtain the fuzzy output of the compressor frequency through fuzzy reasoning according to the preset fuzzy rules, and then obtain the frequency value through defuzzification operation.

4. The air conditioning energy-saving control system based on environmental perception according to claim 3 is characterized in that: The fuzzy rule base is encoded by genetic algorithm. Each fuzzy rule is regarded as a gene. Multiple fuzzy rules form a chromosome. Assuming that there are n rules in the fuzzy rule base, the chromosome is represented by an ordered gene sequence Chr=[g1,g1,…,g n ], where g i represents the i-th fuzzy rule. The genetic algorithm iteratively optimizes the chromosome through selection, crossover, and mutation operations. The selection operation adopts the roulette wheel selection method. Suppose there are m chromosomes in the population and the fitness value of the j-th chromosome is F j , then the probability P that the jth chromosome is selected j The calculation formula is: The crossover operation sets the probability p c Gene exchange is performed on two randomly selected parent chromosomes to generate offspring chromosomes. The mutation operation is performed with a set probability p m Randomly change the genes on the chromosomes and introduce new genes.

5. The air conditioning energy-saving control system based on environmental perception according to claim 3 is characterized in that: The priority strategy unit is used to evaluate the environmental anomaly level and perform real-time analysis through a neural network model embedded in the local controller, wherein the environmental anomaly level index R is assumed to be: Among them, C PM2.5 is the PM2.5 concentration, C lim is the national standard limit for PM2.5 concentration; The neural network model takes environmental parameters as input and outputs the predicted value of environmental anomaly level after multi-layer convolution and pooling operations. The starting signal S of the fresh air system is: when When the system sets the improvement of air quality index as high priority, it starts the fresh air system.

6. The air conditioning energy-saving control system based on environmental perception according to claim 1, characterized in that: The actuator includes a variable frequency compressor and an intelligent air valve, wherein: The variable frequency compressor uses a vector control algorithm to control the voltage and frequency of the motor. In a three-phase AC motor, the stator current is decomposed into the excitation current component i d and the torque current component i q , respectively control the magnetic flux and torque of the motor. According to the mathematical model of the motor, the stator voltage equation is: Among them, u d ,u q are the d-axis and q-axis components of the stator voltage, R s is the stator resistance, L s is the stator inductance, ω e is the motor electrical angular velocity; The load demand is predicted when the sensor fails by using the fault self-recovery prediction model based on the long short-term memory network. The input is the historical operation data sequence X = [x1, x1, ... x n ], where x i Including the frequency, voltage, current, and cooling capacity information of the variable frequency compressor, the long short-term memory network processes the input sequence through the input gate, forget gate, output gate, and memory unit, and outputs the predicted load demand; The intelligent air valve is based on the return air temperature T r The air valve opening α is controlled by a linkage control algorithm with the air quality index AQI, and the control quantity u of the air valve opening is calculated by the PID control algorithm. The control quantity u is expressed as follows: Where, e = T r-set -T r or e=AQI set -AQI, where T r-set is the return air temperature set point, AQI set is the air quality index setting value, K p is the proportionality coefficient, K i is the integral coefficient, K d is the differential coefficient, and α=f(u).

7. The air conditioning energy-saving control system based on environmental perception according to claim 6, characterized in that: The actuator also includes a cold and hot source group control subsystem based on a distributed collaborative control algorithm. The cold and hot source group control subsystem includes an ECS-7000MR controller. The ECS-7000MR controller is used to distribute and coordinate the load among multiple units. Assume that the total system load demand is L, there are i units in total, and the load distribution coefficient of each unit is λ i (i=1,2,…,n), then the load of the i-th unit is: L i =λ i L in, Load distribution coefficient λ i Dynamically adjust according to the performance parameters and current operating status of each unit. Each local controller uses edge computing to analyze the local unit operating data and environmental parameters in real time and calculate the load distribution coefficient adjustment value Δλ of the local unit. i , and the strong and weak electricity integrated platform communicates with other controllers.

8. The air conditioning energy-saving control system based on environmental perception according to claim 1, characterized in that: The integrated platform for strong and weak electricity includes an internal bus communication and fault handling unit and a LONWORKS bus communication mechanism based on edge computing optimization; The internal bus communication and fault handling unit includes an internal bus for communication between modules. The data frame format is set to [Header, Data, CRC], where the Header contains the source address, destination address, and data type information, the Data is the actual transmitted data, and the CRC is a cyclic redundancy check code used for data verification.

9. The air conditioning energy-saving control system based on environmental perception according to claim 8, characterized in that: When a communication failure occurs at a node in the internal bus, the internal bus communication and fault processing unit automatically switches to the backup communication path. The main communication path is P1, the backup communication path is P2, and the node status monitoring variable is S. i , i is the node number, S i =1 means the node is normal, S i =0 indicates a node failure. When a node failure is detected, the communication path switching control signal C is: Assume that the historical communication data is H = [h1,h1,…h m ], where h k Contains the data sent and received by the faulty node at different times, and uses the sliding average method to predict future data demand. The formula is: Where N is the sliding window size.

10. The air conditioning energy-saving control system based on environmental perception according to claim 8, characterized in that: When the LONWORKS bus communication mechanism based on edge computing optimization sends data, the neuron chip encapsulates the upper-layer application data into a data frame that complies with the LONWORKS protocol, adds the source address, destination address, and check code information, and after the neuron chip at the receiving end receives the data frame, it determines whether it is its own data based on the destination address. If so, it performs verification and decapsulation and passes the data to the upper-layer application; Each controller in the system pre-processes and caches the data to be sent and received through edge computing capabilities. When sending data, the local controller performs a preliminary analysis of the environmental parameters, calculates key feature data, and transmits the key feature data through the LONWORKS bus. Assume that the amount of original environmental parameter data is D0, the amount of key feature data after edge computing pre-processing is D1, and the data compression ratio is When receiving data, edge computing is used to cache and pre-parse the data.

Images