Control method and device for multi-connected air conditioning unit

Abstract

The invention relates to the technical field of air conditioner control optimization, in particular to a control method and device for a multi-connected air conditioning unit, and the control method comprises the steps that an air conditioning area is divided to obtain a plurality of air conditioning sub-areas, and an air conditioning area map is constructed based on the air conditioning sub-areas; and performing video monitoring on the plurality of air conditioning sub-regions based on an infrared camera to obtain infrared videos of the plurality of air conditioning sub-regions, and performing image interception on the infrared videos according to a preset sampling period to obtain an infrared video frame sequence. The air conditioning area is reasonably divided, the whole space is divided into the multiple air conditioning sub-areas, the air conditioning system can be controlled more flexibly, the division can be adjusted according to the actual use condition, and the requirements of different numbers of users can be better met.

Description

Technical Field

[0001] The present invention relates to the technical field of air conditioning control optimization, and in particular to a control method and device for a multi-connected air conditioning unit. Background Art

[0002] Multi-split air conditioning unit is an efficient and flexible air conditioning system widely used in large buildings and commercial places. It connects multiple internal fan coils through an external condensing unit to achieve independent temperature control of different areas. This system can flexibly adjust the temperature of each indoor unit according to actual needs, providing higher comfort and energy efficiency.

[0003] Traditional methods usually use fixed control strategies and cannot be adjusted according to real-time environmental changes or fluctuations in user demand, which may lead to over-cooling or over-heating in some areas. Due to the lack of intelligent load forecasting and dynamic adjustment functions, traditional methods tend to maintain high energy consumption levels, resulting in unnecessary waste of resources, especially during periods of low utilization. Traditional methods usually rely on manual monitoring and manual adjustment, with slow response speeds, and are unable to respond to personnel flow and environmental changes in a timely manner, thus affecting comfort. Traditional methods often do not have real-time data collection and analysis capabilities, making it difficult to obtain usage and energy consumption data for each area, which makes optimization decisions difficult. Summary of the invention

[0004] The technical problem to be solved by the present invention is to overcome the disadvantages of the above-mentioned prior art and provide a control method and device for a multi-split air-conditioning unit.

[0005] The technical solution adopted to solve the above technical problems is: a control method for a multi-connected air conditioning unit, comprising:

[0006] Dividing the air conditioning area into multiple air conditioning sub-areas, and constructing an air conditioning area map based on the multiple air conditioning sub-areas;

[0007] Performing video monitoring on the multiple air conditioning sub-areas based on an infrared camera to obtain infrared videos of the multiple air conditioning sub-areas, and performing image interception on the infrared videos according to a preset sampling period to obtain an infrared video frame sequence;

[0008] Performing target detection on the infrared video frame sequence based on the trained target detection model to obtain the number of users corresponding to the multiple air conditioning sub-areas, and assigning node weights of the air conditioning area map based on the number of users corresponding to the multiple air conditioning sub-areas to obtain a weighted air conditioning area map;

[0009] generating a first control signal for the multi-connected air conditioning unit based on the weighted air conditioning area map, and performing preliminary control on the multi-connected air conditioning unit based on the first control signal;

[0010] Based on the weighted air conditioning area map and the user's behavior characteristics, the load demand of the multi-connected air conditioning unit is obtained, energy consumption modeling of the multi-connected air conditioning unit is performed to obtain an energy consumption model of the multi-connected air conditioning unit, and control optimization of the energy consumption model of the multi-connected air conditioning unit is performed based on the load demand to obtain a control optimization model of the multi-connected air conditioning unit;

[0011] The control optimization model of the multi-split air-conditioning unit is solved based on a swarm particle optimization algorithm to obtain a second control signal, and the multi-split air-conditioning unit after preliminary control is finely controlled based on the second control signal.

[0012] Preferably, the air conditioning sub-area corresponds to an indoor unit in the multi-split air-conditioning unit; the nodes of the air conditioning area map correspond to the air conditioning sub-areas, and the associated edges of the air conditioning area map correspond to the connectivity relationship between the two air conditioning sub-areas.

[0013] Preferably, the target detection model adopts an improved YOLOv5 model, and the YOLOv5 model is improved based on the spatiotemporal feature extraction layer, wherein the spatiotemporal feature extraction module includes a spatiotemporal feature extraction layer and a convolution layer, and the spatiotemporal feature extraction layer is used for feature mapping, and the expression of the feature mapping is as follows:

[0014]

[0015] Where X represents any given original feature map, scale represents the scaling factor, and f x,y Represents all entries X of the feature map that are proportionally divisible by i+x and i+y i+y Formation, S represents the scaling parameter.

[0016] Preferably, generating a first control signal for the multi-connected air conditioning unit based on the weighted air conditioning area map, and performing preliminary control on the multi-connected air conditioning unit based on the first control signal, comprises:

[0017] Determining whether there is a user in the air conditioning sub-area based on the node weights of the weighted air conditioning area graph;

[0018] If there is a user in the air conditioning sub-area, determining that the air conditioning sub-area is a first air conditioning sub-area;

[0019] If there is no user in the air conditioning sub-area, determining that the air conditioning sub-area is a second air conditioning sub-area;

[0020] generating an opening control command for the multi-connected air conditioning unit based on the determined first air conditioning sub-area and the second air conditioning sub-area, wherein the opening control command is a first control signal;

[0021] The indoor units in the first air conditioning sub-zone are controlled to turn on based on the turn-on control command.

[0022] Preferably, the expression of the user's behavior characteristics is as follows:

[0023] B a =B u ×B t ;

[0024] Among them, B a Indicates the user's behavioral characteristics, B u represents the probability of opening the indoor unit in the multi-split air-conditioning unit, B t represents the time probability of using the indoor unit of the multi-split air-conditioning unit, and u represents the number of indoor units in the multi-split air-conditioning unit that are turned on. total represents the total number of indoor units in a multi-split air conditioning unit, and t represents the usage time of the indoor unit in the multi-split air-conditioning unit, t total Indicates the total detection time of the indoor units in a multi-split air conditioning unit.

[0025] Preferably, obtaining the load demand of the multi-connected air conditioning unit based on the weighted air conditioning area map and the user's behavior characteristics includes:

[0026] A user behavior feature value is obtained based on the node weights in the weighted air conditioning area map and the user's behavior features, wherein the calculation formula of the user behavior feature value is as follows:

[0027]

[0028] Among them, x i represents the user behavior feature value of the i-th node in the weighted air conditioning area graph, N i represents the maximum number of users that can be accommodated by the i-th node in the weighted air conditioning area graph, s i represents the weight of the i-th node in the weighted air conditioning area graph, H i represents the running time of the i-th node in the weighted air conditioning area graph;

[0029] Based on the trained load prediction model, load prediction is performed according to the user behavior characteristic value to obtain the load demand of the multi-split air-conditioning unit, wherein the load prediction model adopts a long short-term memory neural network.

[0030] Preferably, the energy consumption model of the multi-connected air-conditioning unit includes a chiller energy consumption model, a fan energy consumption model, a cooling tower energy consumption model and a chilled water pump energy consumption model, and the chiller energy consumption model is as follows:

[0031] P chiller =P ref_chiller CAP(T chws ,T cws )EIR(T chws ,T cws )EPLR(T chws ,T cws );

[0032] Among them, P chiller Indicates the power of the chiller, P ref_chiller Indicates the power of the chiller under rated conditions, CAP(T chws ,T cws ) represents the ratio of the actual cooling capacity of the chiller to the cooling capacity under rated conditions. EPLR(T chws ,T cws ) represents the partial load efficiency of the chiller, PLR(T chws ,T cws ) represents the partial load rate of the chiller, T chws Indicates the chilled water outlet temperature of the chiller, T cws represents the cooling water inlet temperature of the chiller, and and and Among them, a i , b i 、c i ,d i 、e i and f i represents the fitting coefficient of the quadratic polynomial, g i represents the fitting coefficient of the Gaussian function, PLR represents the partial load ratio, and Q represents the actual cooling capacity of the chiller. ref Indicates the rated cooling capacity of the chiller;

[0033] The fan energy consumption model is as follows:

[0034]

[0035] Among them, P fanrepresents the power of the fan, r fan Indicates the fan speed ratio, m fan Indicates the actual air volume flowing through the fan, h i represents the fitting coefficients of the quadratic polynomial;

[0036] The cooling tower energy consumption model is as follows:

[0037]

[0038] Among them, P tower Indicates the power of the cooling tower, Q w Indicates the amount of cooling water flowing through the cooling tower, C p represents the main load rate of the chilled water pump, ΔT represents the difference between the water temperature entering the cooling tower and leaving the cooling tower, and η represents the efficiency of the cooling tower;

[0039] The energy consumption model of the chilled water pump is as follows:

[0040]

[0041] Among them, P pump Indicates the power of the chilled water pump, j i represents the chilled water pump characteristic fitting coefficient, β e represents the main load factor of the chilled water pump, and m e-L Indicates the rated chilled water flow rate of the chilled water pump, P e-L Indicates the rated power of the chilled water pump.

[0042] Preferably, the control optimization model of the multi-connected air conditioning unit is as follows:

[0043]

[0044] Wherein, L1 represents the first objective function of the multi-split air-conditioning unit, L2 represents the second objective function of the multi-split air-conditioning unit, and P pre represents the load demand, P chiller_max , P fan_max , P pump_max and P tower_max Respectively represent the maximum power of the chiller, fan, chilled water pump and cooling tower.

[0045] Preferably, solving the multi-connected air conditioning unit control optimization model based on a swarm particle optimization algorithm to obtain a second control signal includes:

[0046] Randomly generate an initial population, set the population size N, initialize the fitness of each particle, where the fitness is the first objective function value and the second objective function value, and set the maximum number of iterations T;

[0047] Calculate z for each particle i The fitness value of the particle z is obtained and the optimal particle z in the current population is identified. best , whose fitness is f best , calculate the other particles in the population and the optimal particle z best The distance between them is calculated as follows:

[0048] d i =||z i -z best ||;

[0049] Among them, d i Represents the distance between the i-th particle in the population and the optimal particle;

[0050] The position is updated according to the position of the particle and the distance between the particles, wherein the position update formula is as follows:

[0051]

[0052] in, and represents the position of the particle with coordinates (i, j) in the population in the t-th and t+1-th iterations in the d-dimensional solution space, r1 represents the first random number, r2 represents the second random number, represents the optimal position of the current particle at the tth iteration, y1 and y2 represent the segmentation coefficients;

[0053] Calculate the fitness of the new position. If the fitness of the new position is better than that of the old position, update the position and fitness of the particle. Repeat the above operation until the maximum number of iterations T is reached or the fitness converges to obtain the particle with the best fitness in the current population and its fitness, thereby obtaining the second control signal.

[0054] The technical solution adopted to solve the above technical problem is: a control device for a multi-split air-conditioning unit, which is applicable to the control method for a multi-split air-conditioning unit, comprising:

[0055] A map construction unit, the map construction unit is used to divide the air conditioning area into regions to obtain a plurality of air conditioning sub-regions, and construct an air conditioning area map based on the plurality of air conditioning sub-regions;

[0056] A video monitoring unit, the video monitoring unit is used to perform video monitoring on the multiple air conditioning sub-areas based on an infrared camera to obtain infrared videos of the multiple air conditioning sub-areas, and to capture images of the infrared videos according to a preset sampling period to obtain an infrared video frame sequence;

[0057] A people counting unit, the people counting unit is used to perform target detection on the infrared video frame sequence based on a trained target detection model to obtain the number of users corresponding to the multiple air conditioning sub-areas, and assign node weights of the air conditioning area map based on the number of users corresponding to the multiple air conditioning sub-areas to obtain a weighted air conditioning area map;

[0058] a first control unit, the first control unit being configured to generate a first control signal for the multi-connected air-conditioning unit based on the weighted air-conditioning area map, and to perform preliminary control on the multi-connected air-conditioning unit based on the first control signal;

[0059] A control optimization unit, the control optimization unit is used to obtain the load demand of the multi-connected air-conditioning unit based on the weighted air conditioning area map and the user's behavior characteristics, perform energy consumption modeling on the multi-connected air-conditioning unit to obtain an energy consumption model of the multi-connected air-conditioning unit, and perform control optimization on the energy consumption model of the multi-connected air-conditioning unit based on the load demand to obtain a control optimization model of the multi-connected air-conditioning unit;

[0060] The second control unit is used to solve the control optimization model of the multi-split air-conditioning unit based on a swarm particle optimization algorithm to obtain a second control signal, and to perform fine control on the multi-split air-conditioning unit after preliminary control based on the second control signal.

[0061] The beneficial effects of the present invention are as follows: (1) The present invention divides the air conditioning area into a plurality of air conditioning sub-areas by reasonably dividing the air conditioning area, so that the air conditioning system can be more flexible in control. This division can not only be adjusted according to the actual usage, but also better meet the needs of different numbers of users. By constructing a weighted air conditioning area map, the weight of each sub-area can reflect its actual usage frequency and importance, which is helpful to realize a more refined control strategy; (2) The present invention can obtain the number of people in each sub-area in real time by using an infrared camera for video monitoring, which not only improves the timeliness of data collection, but also effectively reduces the need for manual intervention and reduces management costs. By intercepting the infrared video frame sequence according to a preset sampling period and performing target detection, the usage of each area can be quickly and accurately reflected to achieve dynamic feedback. This monitoring method enables the air conditioning system to adjust the operating status at any time to cope with environmental changes and real-time changes in user behavior; (3) The present invention can accurately evaluate the load demand of each sub-area based on user number information and behavior characteristics. By constructing an energy consumption model and combining it with the actual load demand, the gap between the current energy consumption state and the target can be identified and then optimized. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 A schematic flow chart of the steps of an overall method in an embodiment of the present invention;

[0063] Figure 2 A schematic diagram of the device architecture of an overall device in an embodiment of the present invention.

[0064] Figure numerals: 1. Graph construction unit; 2. Video monitoring unit; 3. People counting unit; 4. First control unit; 5. Control optimization unit; 6. Second control unit. DETAILED DESCRIPTION

[0065] Embodiment 1, as Figure 1 As shown, the control method for a multi-split air conditioning unit proposed by the present invention comprises:

[0066] S1. Dividing the air conditioning area into regions to obtain a plurality of air conditioning sub-regions, and constructing an air conditioning area map based on the plurality of air conditioning sub-regions;

[0067] S2. Performing video monitoring on the multiple air conditioning sub-areas based on an infrared camera to obtain infrared videos of the multiple air conditioning sub-areas, and performing image interception on the infrared videos according to a preset sampling period to obtain an infrared video frame sequence;

[0068] S3, performing target detection on the infrared video frame sequence based on the trained target detection model to obtain the number of users corresponding to the multiple air conditioning sub-areas, and assigning node weights of the air conditioning area map based on the number of users corresponding to the multiple air conditioning sub-areas to obtain a weighted air conditioning area map;

[0069] S4, generating a first control signal for the multi-connected air conditioning unit based on the weighted air conditioning area map, and performing preliminary control on the multi-connected air conditioning unit based on the first control signal;

[0070] S5. Based on the weighted air conditioning area map and the user's behavior characteristics, the load demand of the multi-split air-conditioning unit is obtained, and the energy consumption model of the multi-split air-conditioning unit is modeled to obtain the energy consumption model of the multi-split air-conditioning unit. Based on the load demand, the energy consumption model of the multi-split air-conditioning unit is controlled and optimized to obtain the control optimization model of the multi-split air-conditioning unit.

[0071] S6. Solve the control optimization model of the multi-split air-conditioning unit based on the swarm particle optimization algorithm to obtain a second control signal, and perform fine control on the multi-split air-conditioning unit after preliminary control based on the second control signal.

[0072] In the present invention, an air-conditioning area refers to an area in a building or space that is demarcated in order to achieve specific temperature and humidity control. There may be multiple different air-conditioning sub-areas in each area; an infrared camera is a sensor that can capture images in the infrared band and is often used to monitor heat (such as heat emitted by the human body) and can work under low light conditions; image capture refers to extracting static images at a specific moment from a video to form an image sequence for subsequent analysis; a target detection model refers to a model trained using a machine learning algorithm that can identify specific objects in an image (such as people) and count the number of these objects; a swarm particle optimization algorithm is an algorithm based on swarm intelligence and particle swarm optimization ideas, which is often used to solve complex optimization problems and find the optimal solution by simulating information sharing and collaboration between individuals in a group.

[0073] Embodiment 2, the control method for a multi-split air-conditioning unit proposed in the present invention, compared with embodiment 1, this embodiment also includes: an air conditioning sub-area corresponds to an indoor unit in the multi-split air-conditioning unit; the nodes of the air conditioning area map correspond to the air conditioning sub-areas, and the associated edges of the air conditioning area map correspond to the connectivity relationship between the two air conditioning sub-areas.

[0074] In an optional embodiment, the target detection model adopts an improved YOLOv5 model, and improves the YOLOv5 model based on the spatiotemporal feature extraction layer, wherein the spatiotemporal feature extraction module includes a spatiotemporal feature extraction layer and a convolution layer, and the spatiotemporal feature extraction layer is used for feature mapping, and the expression of feature mapping is as follows:

[0075]

[0076] Where X represents any given original feature map, scale represents the scaling factor, and f x,y Represents all entries X of the feature map that are proportionally divisible by i+x and i+y i+y Formation, S represents the scaling parameter.

[0077] In an optional embodiment, generating a first control signal of the multi-connected air conditioning unit based on the weighted air conditioning area map, and performing preliminary control on the multi-connected air conditioning unit based on the first control signal includes:

[0078] A1. Determine whether there is a user in the air conditioning sub-area based on the node weights of the weighted air conditioning area map;

[0079] A2. If there is a user in the air conditioning sub-area, determining that the air conditioning sub-area is the first air conditioning sub-area;

[0080] A3. If there is no user in the air conditioning sub-area, the air conditioning sub-area is determined to be the second air conditioning sub-area;

[0081] A4. Generate a start control command for the multi-connected air conditioning unit based on the determined first air conditioning sub-area and the second air conditioning sub-area, where the start control command is a first control signal;

[0082] A5. Based on the start control command, the indoor unit in the first air conditioning sub-area is controlled to start. In an optional embodiment, the expression of the user's behavior characteristic is as follows:

[0083] B a =B u ×B t ;

[0084] Among them, B a Indicates the user's behavioral characteristics, B u represents the probability of opening the indoor unit in the multi-split air-conditioning unit, B t represents the time probability of using the indoor unit of the multi-split air-conditioning unit, and u represents the number of indoor units in the multi-split air-conditioning unit that are turned on. total represents the total number of indoor units in a multi-split air conditioning unit, and t represents the usage time of the indoor unit in the multi-split air-conditioning unit, t total Indicates the total detection time of the indoor units in a multi-split air conditioning unit.

[0085] In an optional embodiment, obtaining the load demand of the multi-connected air conditioning unit based on the weighted air conditioning area map and the user's behavior characteristics includes:

[0086] B1. Obtaining a user behavior feature value based on the node weights in the weighted air conditioning area graph and the user's behavior features, wherein the calculation formula of the user behavior feature value is as follows:

[0087]

[0088] Among them, x i represents the user behavior feature value of the i-th node in the weighted air conditioning area graph, N i represents the maximum number of users that can be accommodated by the i-th node in the weighted air conditioning area graph, s i represents the weight of the i-th node in the weighted air conditioning area graph, H i represents the running time of the i-th node in the weighted air conditioning area graph;

[0089] B2. Based on the trained load prediction model, load prediction is performed according to the user behavior characteristic values ​​to obtain the load demand of the multi-split air-conditioning unit, wherein the load prediction model adopts a long short-term memory neural network.

[0090] It should be noted that the long short-term memory neural network is a special type of recurrent neural network (RNN), which is specially designed to process and predict sequence data. LSTM can effectively capture long-term and short-term dependencies and is suitable for time series data analysis, such as energy consumption prediction.

[0091] In an optional embodiment, the energy consumption model of the multi-split air-conditioning unit includes a chiller energy consumption model, a fan energy consumption model, a cooling tower energy consumption model and a chilled water pump energy consumption model. The chiller energy consumption model is as follows:

[0092] P chiller =P ref_ c hiller CAP(T chws ,T cws )EIR(T chws ,T cws )EPLR(T chws ,T cws );

[0093] Among them, P chiller Indicates the power of the chiller, P ref_chiller Indicates the power of the chiller under rated conditions, CAP(T chws ,T cws ) represents the ratio of the actual cooling capacity of the chiller to the cooling capacity under rated conditions. EPLE(T chws ,T cws ) represents the partial load efficiency of the chiller, PLR(T chws ,T cws ) represents the partial load rate of the chiller, T chws Indicates the chilled water outlet temperature of the chiller, T cws represents the cooling water inlet temperature of the chiller, and and and Among them, a i , b i 、c i ,d i 、e i and f i represents the fitting coefficient of the quadratic polynomial, g i represents the fitting coefficient of the Gaussian function, PLR represents the partial load ratio, and Q represents the actual cooling capacity of the chiller. ref Indicates the rated cooling capacity of the chiller;

[0094] The fan energy consumption model is as follows:

[0095]

[0096] Among them, P fan represents the power of the fan, r fan Indicates the fan speed ratio, m fan Indicates the actual air volume flowing through the fan, h i represents the fitting coefficients of the quadratic polynomial;

[0097] The cooling tower energy consumption model is as follows:

[0098]

[0099] Among them, P tower Indicates the power of the cooling tower, Q w Indicates the amount of cooling water flowing through the cooling tower, C p represents the main load rate of the chilled water pump, ΔT represents the difference between the water temperature entering the cooling tower and leaving the cooling tower, and η represents the efficiency of the cooling tower;

[0100] The energy consumption model of the chilled water pump is as follows:

[0101]

[0102] Among them, P pump Indicates the power of the chilled water pump, j i represents the chilled water pump characteristic fitting coefficient, β e represents the main load factor of the chilled water pump, and m e-L Indicates the rated chilled water flow rate of the chilled water pump, P e-L Indicates the rated power of the chilled water pump.

[0103] It should be noted that the chiller absorbs heat inside the building by circulating cooling water and sends the cooled water into the air-conditioning system; the fan is responsible for circulating the air in the air-conditioning system to ensure that the cold / hot air is evenly distributed; the cooling tower reduces the temperature of the cooling water through evaporative cooling to support the efficient operation of the chiller; the function of the chilled water pump is to transport the cooling water from the chiller to the air-conditioning equipment.

[0104] In an optional embodiment, the multi-connected air conditioning unit control optimization model is as follows:

[0105]

[0106] Wherein, L1 represents the first objective function of the multi-split air-conditioning unit, L2 represents the second objective function of the multi-split air-conditioning unit, and P pre represents the load demand, P chiller_max , P fan_max , P pump_max and P tower_maxRespectively represent the maximum power of the chiller, fan, chilled water pump and cooling tower.

[0107] In an optional embodiment, solving the multi-connected air conditioning unit control optimization model based on a swarm particle optimization algorithm to obtain a second control signal includes:

[0108] C1. Randomly generate an initial population, set the population size N, initialize the fitness of each particle, where the fitness is the value of the first objective function and the value of the second objective function, and set the maximum number of iterations T;

[0109] C2. Calculate z for each particle i The fitness value of the particle z is obtained and the optimal particle z in the current population is identified. best , whose fitness is f best , calculate the other particles in the population and the optimal particle z best The distance between them is calculated as follows:

[0110] d i =||z i -z best ||;

[0111] Among them, d i Represents the distance between the i-th particle in the population and the optimal particle;

[0112] C3. Update the position according to the position of the particle and the distance between particles. The position update formula is as follows:

[0113]

[0114] in, and represents the position of the particle with coordinates (i, j) in the population in the t-th and t+1-th iterations in the d-dimensional solution space, r1 represents the first random number, r2 represents the second random number, represents the optimal position of the current particle at the tth iteration, y1 and y2 represent the segmentation coefficients;

[0115] C4. Calculate the fitness of the new position. If the fitness of the new position is better than that of the old position, update the position and fitness of the particle. Repeat the above operation until the maximum number of iterations T is reached or the fitness converges to obtain the particle with the best fitness in the current population and its fitness, thereby obtaining the second control signal.

[0116] Embodiment three, as Figure 2 As shown, the control device for a multi-split air-conditioning unit proposed by the present invention is applicable to the control method for a multi-split air-conditioning unit, comprising:

[0117] A map construction unit 1, which is used to divide the air conditioning area into regions to obtain a plurality of air conditioning sub-regions, and to construct an air conditioning area map based on the plurality of air conditioning sub-regions;

[0118] The video monitoring unit 2 is used to perform video monitoring on the multiple air conditioning sub-areas based on an infrared camera to obtain infrared videos of the multiple air conditioning sub-areas, and to capture images of the infrared videos according to a preset sampling period to obtain an infrared video frame sequence;

[0119] The population counting unit 3 is used to perform target detection on the infrared video frame sequence based on the trained target detection model to obtain the number of users corresponding to the multiple air conditioning sub-areas, and assign node weights of the air conditioning area map based on the number of users corresponding to the multiple air conditioning sub-areas to obtain a weighted air conditioning area map;

[0120] A first control unit 4, the first control unit 4 is used to generate a first control signal of the multi-connected air conditioning unit based on the weighted air conditioning area map, and perform preliminary control on the multi-connected air conditioning unit based on the first control signal;

[0121] The control optimization unit 5 is used to obtain the load demand of the multi-connected air-conditioning unit based on the weighted air conditioning area map and the user's behavior characteristics, perform energy consumption modeling on the multi-connected air-conditioning unit to obtain the energy consumption model of the multi-connected air-conditioning unit, and perform control optimization on the energy consumption model of the multi-connected air-conditioning unit based on the load demand to obtain the control optimization model of the multi-connected air-conditioning unit;

[0122] The second control unit 6 is used to solve the multi-split air-conditioning unit control optimization model based on the swarm particle optimization algorithm to obtain a second control signal, and to perform fine control on the multi-split air-conditioning unit after preliminary control based on the second control signal.

[0123] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited thereto, and various changes can be made within the knowledge scope of technicians in the relevant technical field without departing from the purpose of the present invention.

Claims

1. A control method for a multi-connected air conditioning unit, characterized in that: include: Dividing the air conditioning area into multiple air conditioning sub-areas, and constructing an air conditioning area map based on the multiple air conditioning sub-areas; Performing video monitoring on the multiple air conditioning sub-areas based on an infrared camera to obtain infrared videos of the multiple air conditioning sub-areas, and performing image interception on the infrared videos according to a preset sampling period to obtain an infrared video frame sequence; Performing target detection on the infrared video frame sequence based on the trained target detection model to obtain the number of users corresponding to the multiple air conditioning sub-areas, and assigning node weights of the air conditioning area map based on the number of users corresponding to the multiple air conditioning sub-areas to obtain a weighted air conditioning area map; generating a first control signal for the multi-connected air conditioning unit based on the weighted air conditioning area map, and performing preliminary control on the multi-connected air conditioning unit based on the first control signal; Based on the weighted air conditioning area map and the user's behavior characteristics, the load demand of the multi-connected air conditioning unit is obtained, energy consumption modeling of the multi-connected air conditioning unit is performed to obtain an energy consumption model of the multi-connected air conditioning unit, and control optimization of the energy consumption model of the multi-connected air conditioning unit is performed based on the load demand to obtain a control optimization model of the multi-connected air conditioning unit; The control optimization model of the multi-split air-conditioning unit is solved based on a swarm particle optimization algorithm to obtain a second control signal, and the multi-split air-conditioning unit after preliminary control is finely controlled based on the second control signal.

2. The control method for a multi-split air conditioning unit according to claim 1, characterized in that: The air conditioning sub-area corresponds to an indoor unit in the multi-split air-conditioning unit; the nodes of the air conditioning area map correspond to the air conditioning sub-areas, and the associated edges of the air conditioning area map correspond to the connectivity relationship between two air conditioning sub-areas.

3. The control method for a multi-split air conditioning unit according to claim 1, characterized in that: The target detection model adopts an improved YOLOv5 model, and improves the YOLOv5 model based on a spatiotemporal feature extraction layer, wherein the spatiotemporal feature extraction module includes a spatiotemporal feature extraction layer and a convolution layer, and the spatiotemporal feature extraction layer is used for feature mapping, and the expression of the feature mapping is as follows: Where X represents any given original feature map, scale represents the scaling factor, and f x,y Represents all entries X of the feature map that are proportionally divisible by i+x and i+y i+y Formation, S represents the scaling parameter.

4. The control method for a multi-split air conditioning unit according to claim 1, characterized in that: Generating a first control signal for the multi-connected air conditioning unit based on the weighted air conditioning area map, and performing preliminary control on the multi-connected air conditioning unit based on the first control signal, comprising: Determining whether there is a user in the air conditioning sub-area based on the node weights of the weighted air conditioning area graph; If there is a user in the air conditioning sub-area, determining that the air conditioning sub-area is a first air conditioning sub-area; If there is no user in the air conditioning sub-area, determining that the air conditioning sub-area is a second air conditioning sub-area; generating an opening control command for the multi-connected air conditioning unit based on the determined first air conditioning sub-area and the second air conditioning sub-area, wherein the opening control command is a first control signal; The indoor units in the first air conditioning sub-zone are controlled to turn on based on the turn-on control command.

5. The control method for a multi-split air conditioning unit according to claim 1, characterized in that: The expression of the user's behavior characteristics is as follows: B a =B u ×B t ; Among them, B a Indicates the user's behavioral characteristics, B u represents the probability of opening the indoor unit in the multi-split air-conditioning unit, B t represents the time probability of using the indoor unit of the multi-split air-conditioning unit, and u represents the number of indoor units in the multi-split air-conditioning unit that are turned on. total represents the total number of indoor units in a multi-split air conditioning unit, and t represents the usage time of the indoor unit in the multi-split air-conditioning unit, t total Indicates the total detection time of the indoor units in a multi-split air conditioning unit.

6. The control method for a multi-split air conditioning unit according to claim 5, characterized in that: Obtaining the load demand of the multi-connected air conditioning unit based on the weighted air conditioning area map and the user's behavior characteristics includes: A user behavior feature value is obtained based on the node weights in the weighted air conditioning area map and the user's behavior features, wherein the calculation formula of the user behavior feature value is as follows: Among them, x i represents the user behavior feature value of the i-th node in the weighted air conditioning area graph, N i represents the maximum number of users that can be accommodated by the i-th node in the weighted air conditioning area graph, s i represents the weight of the i-th node in the weighted air conditioning area graph, H i represents the running time of the i-th node in the weighted air conditioning area graph; Based on the trained load prediction model, load prediction is performed according to the user behavior characteristic value to obtain the load demand of the multi-split air-conditioning unit, wherein the load prediction model adopts a long short-term memory neural network.

7. The control method for a multi-split air conditioning unit according to claim 1, characterized in that: The energy consumption model of the multi-connected air conditioning unit includes a chiller energy consumption model, a fan energy consumption model, a cooling tower energy consumption model and a chilled water pump energy consumption model. The energy consumption model of the chiller is as follows: P chiller =P ref_chiller CAP(T chws ,T cws )EIR(T chws ,T cws )EPLR(T chws ,T cws ); Among them, P chiller Indicates the power of the chiller, P ref_chiller Indicates the power of the chiller under rated conditions, CAP(T chws ,T cws ) represents the ratio of the actual cooling capacity of the chiller to the cooling capacity under rated conditions. EPLR(T chws ,T cws ) represents the partial load efficiency of the chiller, PLR(T chws ,T cws ) represents the partial load rate of the chiller, T chws Indicates the chilled water outlet temperature of the chiller, T cws represents the cooling water inlet temperature of the chiller, and and and Among them, a i 、b i 、c i ,d i 、e i and f i represents the fitting coefficient of the quadratic polynomial, g i represents the fitting coefficient of the Gaussian function, PLR represents the partial load ratio, and Q represents the actual cooling capacity of the chiller. ref Indicates the rated cooling capacity of the chiller; The fan energy consumption model is as follows: Among them, P fan represents the power of the fan, r fan Indicates the fan speed ratio, m fan Indicates the actual air volume flowing through the fan, h i represents the fitting coefficients of the quadratic polynomial; The cooling tower energy consumption model is as follows: Among them, P tower Indicates the power of the cooling tower, Q w Indicates the amount of cooling water flowing through the cooling tower, C p represents the main load rate of the chilled water pump, ΔT represents the difference between the water temperature entering the cooling tower and leaving the cooling tower, and η represents the efficiency of the cooling tower; The energy consumption model of the chilled water pump is as follows: Among them, P pump Indicates the power of the chilled water pump, j i represents the chilled water pump characteristic fitting coefficient, β e represents the main load factor of the chilled water pump, and m e-L Indicates the rated chilled water flow rate of the chilled water pump, P e-L Indicates the rated power of the chilled water pump.

8. The control method for a multi-split air conditioning unit according to claim 7, characterized in that: The control optimization model of the multi-connected air-conditioning unit is as follows: Wherein, L1 represents the first objective function of the multi-split air-conditioning unit, L2 represents the second objective function of the multi-split air-conditioning unit, and P pre represents the load demand, P chiller_max , P fan_max , P pump_max and P tower_max Respectively represent the maximum power of the chiller, fan, chilled water pump and cooling tower.

9. The control method for a multi-split air conditioning unit according to claim 8, characterized in that: The multi-connected air conditioning unit control optimization model is solved based on a swarm particle optimization algorithm to obtain a second control signal, including: Randomly generate an initial population, set the population size N, initialize the fitness of each particle, where the fitness is the first objective function value and the second objective function value, and set the maximum number of iterations T; Calculate z for each particle i The fitness value of the particle z is obtained and the optimal particle z in the current population is identified. best , whose fitness is f best , calculate the other particles in the population and the optimal particle z best The distance between them is calculated as follows: d i =|| z i-z best ||; Among them, d i Represents the distance between the i-th particle in the population and the optimal particle; The position is updated according to the position of the particle and the distance between the particles, wherein the position update formula is as follows: in, and represents the position of the particle with coordinates (i, j) in the population in the t-th and t+1-th iterations in the d-dimensional solution space, r1 represents the first random number, r2 represents the second random number, represents the optimal position of the current particle at the tth iteration, y1 and y2 represent the segmentation coefficients; Calculate the fitness of the new position. If the fitness of the new position is better than that of the old position, update the position and fitness of the particle. Repeat the above operation until the maximum number of iterations T is reached or the fitness converges to obtain the particle with the best fitness in the current population and its fitness, thereby obtaining the second control signal.

10. A control device for a multi-split air-conditioning unit, which is applicable to the control method for a multi-split air-conditioning unit according to any one of claims 1 to 9, characterized in that: include: A map construction unit (1), the map construction unit (1) is used to divide the air conditioning area into regions to obtain a plurality of air conditioning sub-regions, and to construct an air conditioning area map based on the plurality of air conditioning sub-regions; A video monitoring unit (2), the video monitoring unit (2) being used to perform video monitoring on the multiple air conditioning sub-areas based on an infrared camera to obtain infrared videos of the multiple air conditioning sub-areas, and to perform image interception on the infrared videos according to a preset sampling period to obtain an infrared video frame sequence; A population counting unit (3), the population counting unit (3) being used to perform target detection on the infrared video frame sequence based on a trained target detection model to obtain the number of users corresponding to the multiple air conditioning sub-areas, and to assign node weights of the air conditioning area map based on the number of users corresponding to the multiple air conditioning sub-areas to obtain a weighted air conditioning area map; A first control unit (4), the first control unit (4) being used to generate a first control signal for the multi-connected air conditioning unit based on the weighted air conditioning area map, and to perform preliminary control on the multi-connected air conditioning unit based on the first control signal; A control optimization unit (5), the control optimization unit (5) being used to obtain the load demand of the multi-connected air conditioning unit based on the weighted air conditioning area map and the user's behavior characteristics, perform energy consumption modeling on the multi-connected air conditioning unit to obtain the multi-connected air conditioning unit energy consumption model, and perform control optimization on the multi-connected air conditioning unit energy consumption model based on the load demand to obtain the multi-connected air conditioning unit control optimization model; A second control unit (6), the second control unit (6) is used to solve the control optimization model of the multi-split air-conditioning unit based on a swarm particle optimization algorithm to obtain a second control signal, and to perform fine control on the multi-split air-conditioning unit after preliminary control based on the second control signal.

Images