An intelligent control system and method for an additional deflector of an air conditioner air supply

Abstract

The present invention relates to an intelligent control system and method for an additional deflector of an air conditioner. The method includes the following steps: S1, collecting the three-dimensional space information Qp where the human body is located; S2, inputting the three-dimensional space information Qp where the human body is located into the air conditioner air supply control model, and performing inference calculation to obtain the target opening and closing sizes Si of the holes on the additional deflector and the target deflector solid angle w of the additional deflector; the additional deflector is installed at the air outlet of the air conditioner and has a number of holes with independently controllable opening and closing sizes, the area of the opened hole is Si, and the angle between the plane of the additional deflector and the horizontal plane is the target deflector solid angle w; S3, controlling the additional deflector according to the target opening and closing sizes Si of the holes on the additional deflector and the target deflector solid angle w. It can perform intelligent adjustment according to the indoor environment change and user needs to optimize the air flow distribution, improve the comfort level and reduce the energy consumption.

Description

Technical Field

[0001] The present invention relates to the intelligent control of a deflector, in particular to an intelligent control system and method for an additional deflector of an air conditioner for air supply. Background Art

[0002] With the improvement of people's living quality and the development of technology, the demand for indoor comfort and energy-saving efficiency of air conditioners is increasing day by day. In the existing air conditioning systems, the air supply deflectors usually adopt fixed or simple mechanical control methods, which are difficult to achieve precise air supply and match personalized requirements. Summary of the Invention

[0003] The present invention aims to provide an intelligent control system and method for an additional deflector of an air conditioner for air supply, which can perform intelligent adjustment according to indoor environmental changes and user needs to optimize the air flow distribution, improve comfort and reduce energy consumption, so as to solve the problems in the background art.

[0004] By collecting indoor environmental data in real time and combining with the user's usage habits, the angle and direction of the air supply deflector of the air conditioner are automatically adjusted to accurately control the direction and speed of the air flow from the air conditioner, so as to achieve the best cooling / heating effect and uniform indoor temperature and humidity distribution.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] An intelligent control system for an additional deflector of an air conditioner for air supply, characterized in that the additional deflector is rotatably installed at the air outlet of the air conditioner and has a plurality of holes with independently controllable opening and closing sizes;

[0007] The additional deflector is connected to a first rotating shaft through a connecting rod, the first rotating shaft is fixed to a bracket, and the bracket is installed on the air conditioner housing, so that the additional deflector is installed at the air outlet of the air conditioner. By rotating the first rotating shaft, the included angle w between the plane of the additional deflector and the horizontal plane is adjusted;

[0008] Each hole on the additional deflector can be blocked by an independently controllable flip-up diaphragm. The area of the diaphragm is slightly larger than the hole and is fixed to the back of the additional deflector through a second rotating shaft. By different rotation angles, the opening and closing size Si of the hole is controlled;

[0009] The system includes:

[0010] An electric drive device is used to control the included angle w between the plane of the additional deflector and the horizontal plane and the opening and closing size Si of the holes of the additional deflector;

[0011] The electric drive device includes a first motor for controlling the first rotating shaft and a second motor for controlling the second rotating shaft;

[0012] A multi-dimensional sensor module for collecting the three-dimensional space information Qp of the human body;

[0013] A microprocessor controller, which is respectively connected to the electric drive device and the multi-dimensional sensor module, and is used to input the three-dimensional space information Qp of the human body into the air-conditioning air supply control model in the intelligent algorithm module, perform inference and calculation to obtain the target opening size Si' of each hole on the additional deflector and the target deflector solid angle w'; according to the target opening size Si' of each hole on the additional deflector and the target deflector solid angle w', control the electric drive device, so as to adjust the included angle w between the plane of the additional deflector and the horizontal plane and the opening size Si of the holes on the additional deflector.

[0014] Furthermore, the three-dimensional space information Qp of the human body includes: crowd characteristic information P, temperature setting information Ts, three-dimensional temperature field information Tr, three-dimensional humidity field information Rr, three-dimensional air quality information Ar, and regional position information Qr.

[0015] Furthermore, the regional position information Qr includes the spatial site type, indoor area and items, and three-dimensional space model, where:

[0016] The spatial site type includes sitting work sites (such as study rooms, meeting rooms), human movement activity sites (such as gymnasiums, dance halls), personnel static rest sites (such as bedrooms, rest rooms), and composite sites (sites used in combination of multi-functional areas);

[0017] The indoor area and items include work areas (areas where people are mainly in a sitting position at the table and are relatively static), rest areas (areas where people are mainly in a lying or semi-reclining position on the bed or sofa), activity areas (areas where people stand, walk, move, etc. with dynamic changes); fixed furniture (such as cabinets, desks, beds), and movable furniture (such as stools, chairs).

[0018] The three-dimensional space model includes the shape, length, width and height information of the indoor space, air-conditioning outlet information (including height from the ground, the wall or ceiling to which it belongs, outlet width, etc.), indoor furniture placement area and length, width and height information, and length, width and height information of each area where indoor personnel are located.

[0019] Furthermore, the crowd characteristic information P includes gender, age, crowd composition, position (the distance between the person and the air conditioner, such as in the same small area of the same room or in different areas of the same building), number of people, and posture (mainly divided into three types of human postures: standing, sitting, lying).

[0020] Furthermore, the posture is obtained by the human posture estimation method of a monocular image.

[0021] Further, the human body image obtained by the camera is input into the pre-trained model of Lightweight Openpose to generate a human body pose skeleton framework diagram.

[0022] Further, it also includes an energy storage and charging module, which is electrically connected to the microprocessor controller. The energy storage and charging module includes a solar photovoltaic power generation device for generating electricity and an energy storage lithium battery pack for supplying power to the system.

[0023] Further, it also includes a timer, which is electrically connected to the microprocessor controller. The timer is used for timing and time calibration.

[0024] Further, it also includes a voice interaction module, which is electrically connected to the microprocessor controller. The voice interaction module is used for voice interaction. The voice interaction module includes a microphone and a speaker.

[0025] Further, it also includes a noise suppressor, which is electrically connected to the microprocessor controller. The noise suppressor is used for receiving noise and then emitting reverse sound waves through the reverse sound wave noise reduction algorithm to suppress the noise; the noise suppressor is used for receiving noise and then emitting reverse sound waves through the reverse sound wave noise reduction algorithm to suppress the noise. The noise suppressor includes a microphone, a speaker and an algorithm module.

[0026] Further, it also includes a wireless transceiver device, which is electrically connected to the microprocessor controller. The wireless transceiver device is used for communicating with intelligent terminals, and the intelligent terminals include mobile phones, remote controls, computers, etc.

[0027] A control method for an intelligent control system of an air-conditioning air supply additional deflector based on the above includes the following steps:

[0028] S1. Collect the three-dimensional space information Qp of the human body through the multi-dimensional sensor module;

[0029] S2. Input the three-dimensional space information Qp of the human body into the air-conditioning air supply control model in the intelligent algorithm module, and perform inference and calculation to obtain the target opening and closing sizes Si' of the holes on the additional deflector and the target deflector solid angle w' of the additional deflector;

[0030] S3. Control the electric drive device according to the target opening and closing sizes Si' of the holes on the additional deflector and the target deflector solid angle w' of the additional deflector, so as to adjust the included angle w between the plane of the additional deflector and the horizontal plane and the opening and closing sizes Si of the holes on the additional deflector.

[0031] Further, the air-conditioning air supply control model adopts a neural network multi-classification model. The training of the neural network multi-classification model includes the following steps:

[0032] (1)Construct an initial neural network multi-classification model:

[0033] (2)Train the initial neural network multi-classification model using a sample data set to obtain a trained neural network multi-classification model; the three-dimensional space information Qp of the human body is used as the input of the model, and the target opening sizes Si of the holes on the additional deflector and the target deflector solid angle w of the additional deflector are used as the outputs of the model; the trained network model through the sample data set should achieve the following goals:

[0034] The indoor target area temperature reaches the target fastest: adjust the target opening sizes Si of the holes on the additional deflector and the target deflector solid angle w of the additional deflector by ±a °C according to the set temperature of the target area, and the average achievement time of the sample data set is b minutes. Dynamically adjust the opening and closing sizes of the deflector hole matrix through the model algorithm, and calculate the optimal deflector solid angle based on the target area;

[0035] The best energy-saving target under the balanced state of the indoor target area temperature; based on the currently detected temperature value of the target area, using the regional population characteristics P and the indoor area position information Qr as parameters, adopt the model algorithm to control the deflector solid angle and the opening and closing of the holes of the deflector, and under the condition of not changing the air supply and temperature of the air conditioner, achieve the best energy-saving target by preferentially controlling the opening degree of the holes of the deflector, and the average hourly energy consumption of the sample data set is c kWh.

[0036] Generally, a is 0.5, b is 5, c is 0.7, and a, b, and c can also be set as needed.

[0037] After the training is completed, use the test data set to test the neural network multi-classification model. When the success rate of achieving the target is higher than the first threshold, it can be used; otherwise, increase the data volume of the sample data set and continue training until the success rate of achieving the target is higher than the first threshold.

[0038] An electronic device, comprising:

[0039] One or more processors;

[0040] A memory for storing one or more programs;

[0041] When the one or more programs are executed by the one or more processors, the one or more processors implement the above method.

[0042] A computer-readable storage medium, on which computer instructions are stored, and when the instructions are executed by a processor, the steps of the above method are implemented.

[0043] The beneficial effects of the present invention are:

[0044] 1) It solves the problems of the air supply position and angle of the air conditioner, realizes the optimal control of air flow, reduces energy waste, and improves the energy efficiency of the air conditioner.

[0045] 2) It adopts an intelligent control system, which can automatically adjust the position and angle of the air supply deflector according to factors such as indoor temperature, humidity, and personnel activities, to achieve a more uniform air distribution and a comfortable indoor environment.

[0046] 3) By automatically adjusting the angle of the additional deflector, it can reduce the frequent maintenance of air conditioning equipment and lower the operation and maintenance costs. In addition, the intelligent control system can monitor the operation status of the air conditioner, improving the reliability and stability of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 is a schematic structural diagram of the system of the present invention;

[0048] Figure 2 is a schematic side view structural diagram of the system of the present invention;

[0049] Figure 3 is a schematic structural diagram of the diaphragm of the additional deflector of the present invention;

[0050] Figure 4 is a schematic transmission structural diagram of the diaphragm of the additional deflector of the present invention;

[0051] Figure 5 is another schematic structural diagram of the diaphragm of the additional deflector of the present invention;

[0052] Figure 6 is a flowchart of the method of the present invention;

[0053] Among them, the additional deflector 1, the hole 1-1, the first rotating shaft 1-2, the diaphragm 1-3, the rotating shaft 1-4, the microprocessor controller 2, the multi-dimensional sensor module 3, the energy storage and charging module 4, the timer 5, the voice interaction module 6, the noise suppressor 7, the wireless transceiver 8, the intelligent algorithm module 9, the worm gear A, and the worm B. DETAILED DESCRIPTION OF THE INVENTION

[0054] To further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following describes in detail the specific implementation manners, structures, features, and their effects of the present invention in combination with the accompanying drawings and preferred embodiments.

[0055] As Figure 1 , an intelligent control system for an additional deflector of air supply of an air conditioner, includes a microprocessor controller 2 and an additional deflector 1. The microprocessor controller 2 controls an electric drive device, a multi-dimensional sensor module 3, an energy storage and charging module 4, a timer 5, a voice interaction module 6, a noise suppressor 7, a wireless transceiver 8, and an intelligent algorithm module 9;

[0056] The additional deflector 1 is installed at the air outlet of the air conditioner and has several holes 1-1 whose opening and closing sizes can be independently controlled;

[0057] Such as Figure 2 , the additional deflector 1 is connected to the first rotating shaft 1-2 through a connecting rod, the first rotating shaft 1-2 is fixed to the bracket, and the bracket is installed on the air conditioner housing, so that the additional deflector 1 is installed at the air outlet of the air conditioner. By rotating the first rotating shaft 1-2, the included angle w between the plane of the additional deflector 1 and the horizontal plane is adjusted.

[0058] Such as Figure 3 , each hole 1-1 on the additional deflector 1 can be blocked by a flip-up diaphragm 1-3 that can be independently controlled. The area of the diaphragm 1-3 is slightly larger than that of the hole 1-1 and is fixed to the back of the additional deflector 1 through the second rotating shaft 1-4. By different rotation angles, the opening and closing size Si of the hole 1-1 is controlled.

[0059] The electric drive device is used to control the included angle w between the plane of the additional deflector 1 and the horizontal plane and the opening and closing size Si of the holes 1-1 of the additional deflector 1;

[0060] The multi-dimensional sensor module 3 is used to collect the three-dimensional space information Qp where the human body is located; the three-dimensional space information Qp where the human body is located includes: crowd characteristic information P, temperature setting information Ts, three-dimensional temperature field information Tr, three-dimensional humidity field information Rr, three-dimensional air quality information Ar, and regional position information Qr;

[0061] The energy storage and charging module 4 includes a solar photovoltaic power generation device for generating electricity and an energy storage lithium battery pack for supplying power to the system;

[0062] The timer 5 is used for timing and time calibration;

[0063] The voice interaction module 6 is used for voice interaction;

[0064] The noise suppressor 7 is used to receive noise and then emit reverse sound waves through the reverse sound wave noise reduction algorithm to suppress the noise;

[0065] The reverse sound wave noise reduction algorithm is based on the principle of sound wave superposition. When two sound waves with the same frequency and opposite phases meet, they will cancel each other out and form a silent vacuum state. This principle is the basis of the reverse sound wave noise reduction algorithm;

[0066] Noise capture: First, capture the noise in the environment through a sound pickup device such as a microphone;

[0067] Signal processing: The captured noise signal is deeply analyzed through an algorithm to determine key parameters such as the frequency and phase of the noise;

[0068] Inverse acoustic wave generation: Based on the analyzed noise parameters, the algorithm generates an inverse acoustic wave that is precisely symmetric to the original noise;

[0069] Acoustic wave superposition and cancellation: The generated inverse acoustic wave is played through devices such as speakers, meets the original noise, and cancels each other out;

[0070] In this embodiment, the Speex algorithm is adopted. Speex is an open-source, free, and patent-free application collection mainly for speech, including codec, voice activity detection (VAD), discontinuous transmission (DTX), echo cancellation (AEC), noise suppression (NS), and other practical modules; The noise suppression algorithm of Speex has good performance and adopts a complex algorithm process including steps such as preprocessing, energy calculation, noise energy update, signal-to-noise ratio calculation, gain calculation, and post-processing.

[0071] The intelligent algorithm module 9 is used to calculate the angle w between the plane of the additional deflector and the horizontal plane and the opening size Si of the holes according to the three-dimensional space information Qp where the human body is located;

[0072] The wireless transceiver device 8 is used to communicate with the intelligent terminal.

[0073] In this embodiment, the electric drive device includes a first motor that controls the first rotating shaft 1-2 and a second motor that controls the second rotating shaft 1-4. The shaft of the first motor is coaxially connected to the first rotating shaft 1-2; The second motor uses a mini motor, thereby reducing the volume and preventing interference between adjacent mini motors. For example, Figure 4 , the shaft of the second motor is installed with a worm B, the shaft end of the second rotating shaft 1-4 is inserted into the shaft hole, and the shaft hole is arranged on the back of the additional deflector 1 and on both sides below the hole 1-1, so that the diaphragm 1-3 can block the hole 1-1. The shaft rod part of the second rotating shaft 1-4 is installed with a worm gear A, and the transmission is carried out through the meshing of the worm gear A and the worm B. The number of second motors is the same as the number of holes 1-1, that is, each hole is provided with a diaphragm 1-3 that can rotate through the second rotating shaft 1-4; In some embodiments, connection and transmission can also be carried out through transmission mechanisms such as gears, belt pulleys, universal joints, and magnetic couplings.

[0074] Similarly, for example, Figure 5 , in some embodiments, the diaphragm 1-3 for blocking the hole 1-1 can also be opened and closed by the rotation of the second rotating shaft 1-4 installed in the middle thereof, so as to control the opening size Si of the hole.

[0075] In some embodiments, the diaphragm 1-3 can also block the hole 1-1 by sliding.

[0076] For example, Figure 6 , an intelligent control method for an additional deflector of air conditioning air supply includes the following steps:

[0077] S1. Collect the three-dimensional space information Qp of the human body's location;

[0078] S2. Input the three-dimensional space information Qp of the human body's location into the air-conditioning air supply control model, perform inference calculations to obtain the target opening sizes Si of the holes on the additional deflector and the target deflector solid angle w of the additional deflector; the additional deflector is installed at the air outlet of the air conditioner and has a number of holes with independently controllable opening sizes, the opening area of the holes is Si, and the angle between the plane of the additional deflector and the horizontal plane is the target deflector solid angle w;

[0079] Since each hole has its corresponding opening area value, Si is a vector, and each hole has its own position number, which facilitates the control of the opening size of each hole;

[0080] By adjusting and setting the target opening sizes Si and the target deflector solid angle w, the air volume, air speed, temperature cooling rate, etc. in each direction are controlled;

[0081] S3. Control the additional deflector according to the target opening sizes Si of the holes on the additional deflector and the target deflector solid angle w of the additional deflector.

[0082] In some embodiments, a remote control function is adopted, and users can adjust the control parameters of the additional deflector at any time and place through mobile phones or other intelligent devices, flexibly meeting personal needs.

[0083] The three-dimensional space information Qp of the human body's location includes: crowd characteristic information P, temperature setting information Ts, three-dimensional temperature field information Tr, three-dimensional humidity field information Rr, three-dimensional air quality information Ar, and regional location information Qr.

[0084] The temperature setting information Ts is the target temperature information input by the personnel through remote control or voice.

[0085] The three-dimensional temperature field information Tr is the three-dimensional temperature field information obtained by arranging infrared thermal imaging temperature detectors on the surface of the indoor space facing the additional deflector and point-type temperature detectors distributed in the indoor activity area with wireless transmission functions.

[0086] The three-dimensional humidity field information Rr is the three-dimensional humidity field information obtained by arranging humidity detectors on the surface of the indoor space facing the additional deflector and point-type humidity detectors distributed in the indoor activity area with wireless transmission functions.

[0087] The three-dimensional air quality information Ar is obtained by deploying air quality sensors on the surface of the additional deflector facing the indoor space and distributed air quality sensors with wireless transmission function in the indoor activity area to detect pollutant indicators such as PM2.5 and CO2 concentration, so as to obtain the three-dimensional air quality information of the space. The sensor models used above are Sensirion - SHT85 + SPS30.

[0088] The crowd characteristic information P includes gender, age, crowd composition, location (the distance between the person and the air conditioner, such as in the same small area of the same room or in different areas of the same building), number of people, and posture (mainly divided into three types of human postures: standing, sitting, and lying).

[0089] Gender and age are obtained by using a camera to take pictures and through the ResNet + AgeNet / GenderNet multi-task learning model. The location and number of people are obtained by deploying personnel activity sensors on the surface of the additional deflector facing the indoor space and distributed personnel activity sensors with wireless transmission function in the indoor activity area to obtain the location and number information of personnel activities in the space. The sensor model used is Senion Indoor Positioning and People CountingSystem.

[0090] The posture is obtained by the human posture estimation method of monocular images. Specifically, a fully connected neural network model is used. The model includes an input layer, a hidden layer, and an output layer. The input layer corresponds to a human skeleton frame diagram of n×m pixels, and the output layer corresponds to the results of three types of human posture recognition models, which are 0, 1, and 2. Among them, 0 represents the lying / semi-reclining posture, 1 represents the sitting posture, and 2 represents the standing posture; in this embodiment, n×m is 128×128;

[0091] For the heat map Hi of joint point i with size D×H×W, it discretizes the space where the human body is located into a discrete space of D×H×W. Each element Hi(x,y,z) of the volumetric heat map represents the confidence of the position of joint point i at (x,y,z). In the network training stage, after knowing the pseudo-three-dimensional coordinates (x truth ,y truth ,z truth ) of joint point i, the corresponding volumetric heat map Hi can be obtained through a Gaussian distribution with a fixed σ:

[0092] Hi(x,y,z)=(1 / (2πσ 2 )×(e^(-((x - x truth )) 2 +(y - y truth )) 2 +(z - z truth )) 2 ) / (2πσ2 )));

[0093] Use a volumetric heatmap to replace the three-dimensional coordinate point supervised network training, which converts the highly non-linear problem of three-dimensional coordinate regression into the prediction of a three-dimensional matrix in discrete space, thus better leveraging the advantage of the neural network in capturing spatial features. In the test inference stage, after the neural network outputs the volumetric heatmap of the joint points, the pseudo 3D coordinates (x pre , y pre , z pre ) of the corresponding joint points are the coordinate indices of the maximum element in the volumetric heatmap. That is

[0094] (x pre , y pre , z pre ) = arg max(Hi(x, y, z));

[0095] Input the human body image obtained by the camera into the pre-trained model of Lightweight Openpose to generate a human body pose skeleton framework diagram.

[0096] The pre-trained model of Lightweight Openpose can be obtained from Daniil-Osokin on github.

[0097] Specifically, first perform data preprocessing. Before inputting the collected image into the model, preprocess the image by scaling it to a size of 224×224 and then normalizing the pixel values of the image to the interval [-1, 1]. Random rotation, translation, flipping, etc. can also be added to improve the generalization ability of the model.

[0098] Then determine the input and output. The input is a 224×224 RGB image. The output is the number of key points we want to identify, which is 17 (the common number of human key points), and the output will be 17×2 (each key point has two coordinates x and y).

[0099] Then the input layer receives the RGB image with a size of 224×224, and flattens the three-dimensional image data of 224×224×3 into a one-dimensional vector with a length of 150528.

[0100] Next, pass through the first hidden layer which contains 1024 neurons, and the input is a 150528-dimensional input vector. Each neuron has weights connected to the input vector, so the size of the weight matrix is (150528, 1024). Then perform a linear transformation to calculate the weighted sum of each neuron. Pass the result of the linear transformation through the activation function and output it to the next layer.

[0101] The second hidden layer contains 512 neurons. The input is the output of the first layer, with a size of 1024. Therefore, the size of the weight matrix is (1024, 512). Then, a linear transformation is performed to calculate the weighted sum of each neuron. After passing through the activation function, the output is sent to the next layer.

[0102] The third hidden layer contains 256 neurons. The input is the output of the second layer, with a size of 512. Therefore, the size of the weight matrix is (512, 256). Then, a linear transformation is performed to calculate the weighted sum of each neuron. After passing through the activation function, the output is sent to the next layer.

[0103] The output layer has 34 output nodes, and each node corresponds to the x and y coordinates of a key point. The size of the output vector is 34. The size of the weight matrix is (256, 34).

[0104] The input of each hidden layer is processed by weighted summation and the activation function. The output layer maps the output of the third layer to the final 34 coordinate outputs.

[0105] The indoor area location information Qr is for setting the spatial site type and indoor area items to construct a three-dimensional indoor information model, where:

[0106] The spatial site type includes sitting work sites (such as study rooms, meeting rooms), human movement activity sites (such as gyms, ballrooms), personnel static rest sites (such as bedrooms, rest rooms), and composite sites (sites used in combination of multi-functional areas).

[0107] The indoor area and items include work areas (areas where people mainly sit at the table and are relatively static), rest areas (areas where people mainly lie and semi-recline on the bed and sofa), activity areas (areas with dynamic changes such as people standing, walking, and exercising); fixed furniture (such as cabinets, desks, beds), and movable furniture (such as stools, chairs).

[0108] The three-dimensional indoor construction information includes the shape, length, width, and height information of the indoor space, air-conditioning outlet information (including height from the ground, the wall or ceiling it belongs to, outlet width, etc.), the placement area and length, width, and height information of indoor furniture, and the length, width, and height information of each area where indoor personnel are located.

[0109] The air-conditioning air supply control model uses a neural network multi-classification model. The training of the neural network multi-classification model includes the following steps:

[0110] (1) Construct an initial neural network multi-classification model:

[0111] (2) Train the initial neural network multi-classification model using the sample data set to obtain the trained neural network multi-classification model; the three-dimensional spatial information Qp of the human body is used as the input of the model, and the target opening sizes Si of the holes on the additional deflector and the target deflector solid angle w of the additional deflector are used as the outputs of the model; the trained network model through the sample data set should achieve the following goals:

[0112] The indoor target area temperature reaches the target fastest: adjust the target opening sizes Si of the holes on the additional deflector and the target deflector solid angle w of the additional deflector by ±0.5 °C according to the set temperature of the target area, and the average achievement time of the sample data set is 5 minutes. Dynamically adjust the opening and closing sizes of the deflector hole matrix through the model algorithm, and calculate the optimal deflector solid angle based on the target area;

[0113] The best energy-saving goal under the balanced state of the indoor target area temperature; based on the current value of the detected target area temperature, with the regional population characteristics P and the indoor area position information Qr as parameters, use the model algorithm to control the deflector solid angle and the opening and closing of the deflector holes, and without changing the air supply and temperature of the air conditioner, achieve the best energy-saving goal by preferentially controlling the opening degree of the deflector holes. The average power consumption of the sample data set is 0.7 kWh per hour.

[0114] After the training is completed, use the test data set to test the neural network multi-classification model. It can be used when the success rate of achieving the goal is higher than 90%, otherwise increase the data volume of the sample data set and continue training until the success rate of achieving the goal is higher than 90%.

[0115] In this embodiment, the first threshold is 90%, which can also be set according to requirements.

[0116] An electronic device, comprising:

[0117] One or more processors;

[0118] A memory for storing one or more programs;

[0119] When one or more programs are executed by one or more processors, one or more processors implement the above method.

[0120] In the embodiments provided in this application, it should be understood that the disclosed methods and systems can also be implemented in other ways. The method and system embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions, and operations of the methods, systems, methods, and computer program products according to multiple embodiments of this application. In this regard, each block in the flowchart or block diagram can represent a module, a program segment, or a part of code. A module, a program segment, or a part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks can actually be executed substantially in parallel, and they can sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, as well as the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0121] In addition, each functional module in various embodiments of this application can be integrated together to form an independent part, or each module can exist separately, or two or more modules can be integrated to form an independent part.

[0122] On the other hand, a computer-readable storage medium stores computer instructions, and when the instructions are executed by a processor, the steps of the above method are implemented. When the computer program is executed by a processor, the method as described in any one of the above first aspects is implemented. If the function is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in various embodiments of this application. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory 101 (ROM, Read-Only Memory), random access memory 101 (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0123] The above are only the preferred embodiments of the present invention, and do not impose any form of limitation on the present invention. Although the present invention has been disclosed above with the preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to equivalent embodiments by using the above-disclosed technical content without departing from the technical solution of the present invention. However, as long as it does not depart from the technical solution content of the present invention, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A control method for an intelligent control system of an additional deflector for air supply of an air conditioner, characterized in that, The additional deflector is rotatably mounted on the air outlet of the air conditioner and has a number of holes whose opening and closing sizes can be independently controlled; The system includes: An electric drive device for controlling the angle w between the plane of the additional deflector and the horizontal plane and the opening and closing sizes Si of the holes of the additional deflector; A multi-dimensional sensor module for collecting the three-dimensional space information Qp where the human body is located; A microprocessor controller, which is respectively connected to the electric drive device and the multi-dimensional sensor module, and is used to input the three-dimensional space information Qp where the human body is located into the air supply control model of the intelligent algorithm module for inference calculation to obtain the target opening and closing sizes Si' of the holes on the additional deflector and the target deflector solid angle w' of the additional deflector; According to the target opening and closing sizes Si' of the holes on the additional deflector and the target deflector solid angle w' of the additional deflector, control the electric drive device to adjust the angle w between the plane of the additional deflector and the horizontal plane and the opening and closing sizes Si of the holes of the additional deflector; The three-dimensional space information Qp where the human body is located includes: crowd characteristic information P, temperature setting information Ts, three-dimensional temperature field information Tr, three-dimensional humidity field information Rr, three-dimensional air quality information Ar, and regional position information Qr; The regional position information Qr includes the type of space site, indoor areas and items, and the three-dimensional space model; The crowd characteristic information P includes gender, age, crowd composition, position, number of people, and posture; The method includes the following steps: S1. Collect the three-dimensional space information Qp where the human body is located through the multi-dimensional sensor module; S2. Input the three-dimensional space information Qp where the human body is located into the air supply control model of the intelligent algorithm module for inference calculation to obtain the target opening and closing sizes Si' of the holes on the additional deflector and the target deflector solid angle w' of the additional deflector; S3. According to the target opening and closing sizes Si' of the holes on the additional deflector and the target deflector solid angle w' of the additional deflector, control the electric drive device to adjust the angle w between the plane of the additional deflector and the horizontal plane and the opening and closing sizes Si of the holes of the additional deflector; The air supply control model of the air conditioner adopts a neural network multi-classification model, and the training of the neural network multi-classification model includes the following steps: (1) Construct an initial neural network multi-classification model: (2) Train the initial neural network multi-classification model with a sample data set to obtain a trained neural network multi-classification model. The sample data set includes the three-dimensional space information Qp where the human body is located, the target opening and closing sizes Si of the holes on the additional deflector, and the target deflector solid angle w of the additional deflector; The three-dimensional space information Qp where the human body is located is used as the input of the model, and the target opening and closing sizes Si of the holes on the additional deflector and the target deflector solid angle w of the additional deflector are used as the output of the model; The trained network model should achieve the following goals through the training of the sample data set: Control the target opening and closing sizes Si of the holes on the additional deflector and the target deflector solid angle w according to the temperature adjustment of ±a °C in the target area, and the average achievement time of the sample data set is b minutes; Based on the current value of the temperature in the target area detected, with the regional population characteristics P and the regional location information Qr as parameters, under the condition of not changing the air supply and temperature of the air conditioner, the average hourly energy consumption of the sample data set is c kilowatt-hours; Use the test data set to test the neural network multi-classification model. It can be used when the achieved target success rate is higher than the first threshold. Otherwise, increase the data volume of the sample data set and continue training until the achieved target success rate is higher than the first threshold.

2. The method according to claim 1, characterized in that, The pose is obtained by the human pose estimation method of the monocular image.

3. The method according to claim 1, characterized in that, Input the human body image obtained by the camera into the pre-trained model of Lightweight Openpose to generate the human body pose skeleton frame diagram.

4. An electronic device, characterized in that, Including: One or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1-3.

5. A computer-readable storage medium having computer instructions stored thereon, characterized in that, When the instruction is executed by the processor, the steps of the method according to any one of claims 1-3 are implemented.

Images