Method and device for temperature control of multiple points using machine learning, and air conditioner

Abstract

The application provides a method and device for temperature control of multiple points by using machine learning and an air conditioner. The method comprises the following steps: updating the evaluation value of the first target combination data according to the temperature output value and the first actual temperature output value of the first target combination data; calculating the average value of the parameter control value of the first target combination data and the second target combination data, and obtaining the second actual temperature output value in a stable running state; and determining the newly added combination data according to the average value and the second actual temperature output value. In the embodiment of the application, the evaluation value of the existing combination data is updated based on the measured temperature value, and new combination data is continuously generated, so that the combination data gradually conforms to the installation environment through continuous updating, the air conditioner gradually adapts to the installation environment without the need of installing expensive sensors, and the temperature around the user is adjusted according to the user's demand.

Description

Technical Field

[0001] This invention relates to the field of automatic shutdown control technology for air conditioners, and more specifically, to a method, apparatus, and air conditioner for multi-point temperature control using machine learning. Background Technology

[0002] In recent years, to improve user comfort, some air conditioners have developed technologies that use thermal cameras to acquire and adjust the temperature around users. However, in commercial multi-split air conditioning applications, multiple indoor units may be connected and installed in a large space. Due to the interaction between these units, it is difficult to regulate the temperature around users. Furthermore, in such environments, multiple users may reside in the space, necessitating temperature adjustment for multiple locations.

[0003] Currently, there are methods for simulating fluids and controlling temperature using simulation technology. However, fluid simulations assume an unobstructed space, and conditions such as ceiling height are assumed, which may result in data that does not match actual installation conditions. Another method involves first identifying the installation environment using a high-performance camera before performing fluid simulation; however, the computing equipment used for simulation is already expensive, and adding a high-performance camera would significantly increase the cost. Summary of the Invention

[0004] To address the aforementioned problems, this invention provides a method for multi-point temperature control using machine learning. The method includes: acquiring a target temperature value at at least one location; determining a first target combination of data from multiple combinations based on the target temperature value; each combination of data includes a parameter control value, a temperature output value, and an evaluation value; the first target combination of data is the combination of data where the difference between the temperature output value and the target temperature value is the smallest; controlling an air conditioner to operate based on the parameter control value of the first target combination of data; acquiring a first actual temperature output value under stable operating conditions; updating the evaluation value of the first target combination of data based on the temperature output value and the first actual temperature output value; determining a second target combination of data from multiple combinations based on the target temperature value; the second target combination of data is the combination of data where the difference between the temperature output value and the target temperature value is the second smallest; calculating the average value of the parameter control values ​​of the first target combination of data and the second target combination of data; controlling the air conditioner to operate based on the average value; acquiring a second actual temperature output value under stable operating conditions; and determining newly added combination data based on the average value and the second actual temperature output value; the parameter control value of the newly added combination data is the average value, the temperature output value is the second actual temperature output value, and the evaluation value is an initial evaluation value.

[0005] The method for multi-point temperature control using machine learning provided in this invention has multiple sets of combined data consisting of control values, multi-point room temperature data, and evaluation values. When adjusting the indoor temperature at any location, the data whose output value is closest to the set temperature is selected from the combined data as the control parameter for the air conditioner. The evaluation values ​​of the existing combined data are updated based on the measured temperature values, and new combined data are continuously generated. This continuous updating allows the combined data to gradually conform to the installation environment. The air conditioner can gradually adapt to the installation environment without the need for expensive sensors, thereby adjusting the temperature around the user precisely according to the user's needs.

[0006] Optionally, the method further includes: if the number of the multiple combined data is greater than a preset threshold, then the combined data with the lowest evaluation value is deleted.

[0007] In this embodiment of the invention, the method of discarding combined data described above can be continuously optimized and continuously learned to adapt to the installation environment.

[0008] Optionally, the method further includes: if a temperature adjustment value for any one or more locations is received from a user input, then determining a third target combination data in the combination data based on the temperature adjustment value; the third target combination data is the combination data with the smallest difference between the temperature output value and the temperature adjustment value among all the combination data; and controlling the operation of the air conditioner based on the parameter control value of the third target combination data.

[0009] In this embodiment of the invention, when sufficient combined data is obtained based on the above-mentioned machine learning method, or when the number of learning iterations reaches a preset condition, the operation of the air conditioner can be controlled based on the above-mentioned combined data.

[0010] Optionally, the method further includes: if the number of multiple combined data is greater than a preset threshold and there are multiple combined data with the lowest evaluation value, then delete the combined data with the earliest date among the combined data with the lowest evaluation value.

[0011] In this embodiment of the invention, when there are multiple combinations of data with the lowest evaluation values, the combination data with the earliest generation date can be deleted to retain the more recently generated combination data, which is more in line with the current installation environment.

[0012] Optionally, each of the combined data includes temperature output values ​​for multiple locations; the difference between the temperature output value and the target temperature value is the sum of the differences between the target temperature value at each location and the corresponding temperature output value; or, the difference between the temperature output value and the temperature adjustment value is the sum of the differences between the temperature adjustment value at each location and the corresponding temperature output value.

[0013] In this embodiment of the invention, the user can adjust one or more locations, and the air conditioner can precisely control the temperature of multiple locations.

[0014] Optionally, the evaluation value G is calculated using the following formula: G =

[0015] in, Ts_i The target temperature value for temperature test point Ti. is the temperature output value of temperature test point Ti, and n is the number of temperature test points.

[0016] The present invention provides a formula for calculating the evaluation value G, which can be used to evaluate the applicability of combined data and thus select appropriate combined data.

[0017] Optionally, the formula for calculating the updated evaluation value is as follows: Z_after = Z_before +

[0018] Where Z_after is the updated evaluation value, and Z_before is the evaluation value before the update. Tm"_i This represents the actual temperature output value at temperature test point Ti. α The evaluation threshold is used.

[0019] This invention provides a feasible method for updating evaluation values, which can update the evaluation values ​​of combined data.

[0020] Optionally, each set of combined data further includes outdoor temperature output value and outdoor light intensity output value, and the formula for calculating the updated evaluation value is as follows: G =

[0021] Where Tout_m is the measured outdoor temperature. This is the outdoor temperature output value. Hm This is the measured light intensity. This is the outdoor light output value. F1 , F2 , Gi All are weighted coefficients. Ts_i The target temperature value for temperature test point Ti. is the temperature output value of temperature test point Ti, and n is the number of temperature test points.

[0022] This invention provides a feasible method for updating evaluation values, which can update the evaluation values ​​of combined data.

[0023] This invention provides an apparatus for multi-point temperature control using machine learning. The apparatus includes: a first acquisition module, configured to acquire a target temperature value at at least one location, and determine a first target combination data among multiple combination data based on the target temperature value; each combination data includes a parameter control value, a temperature output value, and an evaluation value, wherein the first target combination data is the combination data with the smallest difference between the temperature output value and the target temperature value among the combination data; a first operation module, configured to control the operation of an air conditioner based on the parameter control value of the first target combination data, and acquire a first actual temperature output value under stable operation; and an evaluation module, configured to update the first target combination data based on the temperature output value of the first target combination data and the first actual temperature output value. The system comprises: an evaluation value for a target combination of data; a second acquisition module, configured to determine a second target combination of data among multiple combination data based on the target temperature value, wherein the second target combination of data is the combination of data with the second smallest difference between the temperature output value and the target temperature value; a second operation module, configured to calculate the average value of the parameter control values ​​of the first target combination of data and the second target combination of data, and to control the operation of the air conditioner based on the average value to obtain a second actual temperature output value under stable operation; and a new addition module, configured to determine new combination data based on the average value and the second actual temperature output value; wherein the parameter control value of the new combination data is the average value, the temperature output value is the second actual temperature output value, and the evaluation value is an initial evaluation value.

[0024] This invention provides a multi-split air conditioner, including the aforementioned device for multi-point temperature control using machine learning and multiple temperature measuring devices; the temperature measuring devices are used to measure the actual temperature output value at each location.

[0025] This invention provides a computer-readable storage medium storing a computer program, which is read and executed by a processor to implement the above-described method.

[0026] The device and multi-split air conditioner for multi-point temperature control using machine learning of the present invention can achieve the same technical effect as the method for multi-point temperature control using machine learning described above. Attached Figure Description

[0027] Picture 1 This is a schematic diagram illustrating the basic principle of the method for multi-point temperature control using machine learning in an embodiment of the present invention. Picture 2 This is a schematic diagram of a two-dimensional horizontal plane model of a hypothetical installation scenario in an embodiment of the present invention; Picture 3 This is a schematic flowchart illustrating a method for multi-point temperature control using machine learning in an embodiment of the present invention. Picture 4 This is a schematic diagram of a device for multi-point temperature control using machine learning, as described in an embodiment of the present invention. Detailed Implementation

[0028] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0029] The method for multi-point temperature control using machine learning provided in this embodiment of the invention can control the temperature of multiple locations through machine learning.

[0030] To facilitate understanding of the scheme in this embodiment, its basic principles will be explained with examples below. Picture 1 This diagram illustrates the basic principle and flow of a method for multi-point temperature control using machine learning, as described in an embodiment of the present invention.

[0031] Let the control value be X, the output value be Y, and the evaluation value be Z. The following defines the relevant combinations: The theoretical correlation between control value and output value: Y = 2 x X 2 The measured correlation between the control value and the output value is Y = 2 x X 2 +3 The correlation formulas above are merely examples for illustrative purposes.

[0032] (1) Obtaining initial combination data Let the initial combination data be ①~④ (and let the initial evaluation value be 10 for each of them). ①(X=1,Y=2,Z=10), ②(X=3,Y=18,Z=10), ③(X=5,Y=50,Z=10), ④(X=8,Y=128,Z=10) (2) Select a combination of data that closely matches the user's requirements. Select the combination of output values ​​from the initial combination data above that best matches the user's required value. For example, if the user's required value is 30, then the combination whose output value is closest to the required value is ② (X=3, Y=18, Z=10).

[0033] (3) After the air conditioner has been running and has stabilized, measure the actual output value. The air conditioner operates according to the control value in the selected combination ②, and the value under steady-state conditions is used as the actual output value for calculation. Continuing with the previous example, if the control value X = 3, then the actual output value Y' = 21.

[0034] (4) Update the evaluation value of the combined data Based on the actual output values ​​described above, evaluate the reliability (accuracy) of the selected combination data ②, and update the evaluation value of combination data ②. The formula for updating the evaluation value is: Z_after = Z_before - |Y – Y'| + α, where α is the evaluation threshold (e.g., α=5). Continuing with the previous example, the evaluation value of ② (X=3, Y=18, Z=10) is: Z = 10 - |18 - 21| + 5 = 12 (5) Generate new control values.

[0035] Calculate the average of the control values ​​in the combination closest to the required value and the second closest combination, and generate a new control value. The closest combination ② (X=3, Y=18, Z=12), the second closest combination ③ (X=5, Y=50, Z=10), the new control value X = (3 + 5) / 2 = 4.

[0036] (6) Measure the output value and obtain new combination data. The system operates according to the control value obtained in (5), and the value under stable operating conditions is used as the actual output value for calculation. This value is then stored as a new combination of control value and output value. Continuing with the previous example, the new control value X = 4, the measured output value Y = 35, and the new combination ⑤ (X=4, Y=35, Z=10).

[0037] (7) Discarding combination data with low evaluation When the number of combined data exceeds a certain limit, the data with the lowest evaluation value is automatically discarded. If multiple combined data have the same evaluation value, the data with the lower label value (or earlier date) is discarded. Continuing with the previous example, given ①(X=1,Y=2,Z=10), ②(X=3,Y=18,Z=12), ③(X=5,Y=50,Z=10), ④(X=8,Y=128,Z=10), and ⑤(X=4,Y=35,Z=10), the combination ①(X=1,Y=2,Z=10) needs to be discarded.

[0038] (8) Repeat the above steps (2) to (7) to optimize the output value and make the output value gradually approach the required value. At the same time, continuously learn to adapt to the installation environment by discarding combinations with low evaluation values ​​and generating new combinations.

[0039] Take, for example, a commercial multi-split air conditioner that can control the temperature of multiple locations. Picture 2 This is a schematic diagram of a two-dimensional horizontal plane model assuming an installation scenario. It shows that there are multiple indoor units A, B, C, and D in a room defined by wall w, with temperature measurement and setting positions T1 to T9.

[0040] In addition, the control values ​​and output values ​​are defined as shown in Table 1.

[0041]

[0042] Table 1 Initial combination data for control and output values ​​can be obtained through preliminary experimental measurements, simulation calculations, or calculations based on post-installation operational data. Additionally, the user requires a set temperature value, and if this is not possible... Picture 2 By setting the temperature test point as shown, the intake temperature of the indoor unit can be used as the output value.

[0043] Regarding how to select a combination of data, the data can be evaluated using the applicability (accuracy) evaluation formula G shown below, and the combination with the smallest G value should be selected. If there is only one temperature measurement point, the difference is calculated directly; if there are multiple temperature measurement points, the absolute values ​​of the differences are calculated and then summed.

[0044] G =

[0045] Where Ts_i is the set temperature of temperature test point Ti.

[0046] As shown below, the actual operating output value is obtained using the control values ​​from the selected combination data. The evaluation value of the combination data is then updated based on the difference between the selected combination's output value and the actual output value. The updated evaluation value is as follows. Z_after = Z_before + + α Where Tm”_i is the measured temperature of temperature test point Ti, and α is the evaluation threshold.

[0047] Then, new control values ​​are generated, output values ​​are calculated, and new combination data are added. In addition, when the number of combination data exceeds a certain amount (e.g., 1000 sets of combination data), data with low evaluation values ​​are discarded.

[0048] Picture 3 The diagram illustrates a schematic flowchart of a method for multi-point temperature control using machine learning according to an embodiment of the present invention. The method includes the following steps: S302, acquire a target temperature value at at least one location, and determine a first target combination of data among multiple combination data based on the target temperature value.

[0049] Each set of combined data includes parameter control values, temperature output values, and evaluation values. The first target combined data is the set of combined data with the smallest difference between the temperature output value and the target temperature value. It should be noted that the temperature output value in the above combined data can be the temperature output value from multiple locations; if target temperature values ​​from multiple locations are obtained, the difference between the target temperature value and the temperature output value at each location can be calculated, and then the absolute values ​​of all differences can be summed. The set of combined data with the smallest sum is then used as the first target combined data.

[0050] Specifically, the above parameter control values ​​may include evaporation temperature, internal fan speed, etc.

[0051] Taking a multi-split air conditioner with multiple indoor units as an example, multiple sets of combined data consisting of control values, multi-point room temperature data, and evaluation values ​​are preset. When adjusting the indoor temperature at any location, the combined data with the output value closest to the target temperature is selected as the control parameter.

[0052] S304, control the operation of the air conditioner according to the parameter control value of the first target combination data mentioned above, and obtain the first actual temperature output value under stable operating conditions.

[0053] After the air conditioner has been running for a preset period of time with the parameter control values ​​of the first target combination data mentioned above and has reached a stable state, the measured temperature values ​​at each location can be obtained.

[0054] S306, update the evaluation value of the first target combination data based on the temperature output value of the first target combination data and the first actual temperature output value.

[0055] After obtaining the actual temperature output value as measured above, the degree of conformity between the combined data and the temperature output value in the combined data can be evaluated based on the difference between the actual temperature output value and the temperature output value in the combined data.

[0056] For example, the formula for calculating the evaluation value G is as follows: G =

[0057] in, Ts_i The target temperature value for temperature test point Ti. is the temperature output value of temperature test point Ti, and n is the number of temperature test points.

[0058] For example, the formula for calculating the updated evaluation value G is as follows: Z_after = Z_before +

[0059] Where Z_after is the updated evaluation value, and Z_before is the evaluation value before the update. Tm"_i This represents the actual temperature output value at temperature test point Ti. α The evaluation threshold is used.

[0060] S308, determine the second target combination data from multiple combination data based on the aforementioned target temperature value. This second target combination data is the combination data with the second smallest difference between the temperature output value and the target temperature value among all combination data.

[0061] Based on the average of the control values ​​in the first target combination data that is closest to the target temperature value and the second target combination data that is second closest, a new control value is generated, and then the corresponding combination data is obtained.

[0062] S310, calculate the average value of the parameter control values ​​of the first target combination data and the second target combination data, and control the air conditioner to operate according to the average value to obtain the second actual temperature output value under stable operating conditions.

[0063] S312, determine the newly added combined data based on the above average value and the second actual temperature output value. The parameter control value of the newly added combined data is the above average value, the temperature output value is the above second actual temperature output value, and the evaluation value is the initial evaluation value.

[0064] A new set of combined data is generated by combining the updated first target combined data with the second target combined data. The parameter control value of the new combined data is the average of the parameter control values ​​of the two sets of data, and the temperature output value is the measured output value obtained by running based on the average value. This combined data more accurately describes the current usable space, thereby enabling continuous learning and adaptation to the environment.

[0065] The method for multi-point temperature control using machine learning provided in this invention has multiple sets of combined data consisting of control values, multi-point room temperature data, and evaluation values. When adjusting the indoor temperature at any location, the data whose output value is closest to the set temperature is selected from the combined data as the control parameter for the air conditioner. The evaluation values ​​of the existing combined data are updated based on the measured temperature values, and new combined data are continuously generated. This continuous updating allows the combined data to gradually conform to the installation environment. The air conditioner can gradually adapt to the installation environment without the need for expensive sensors, thereby adjusting the temperature around the user precisely according to the user's needs.

[0066] Based on the generation of new combined data described above, the overall quantity limit also needs to be considered. If the number of combined data exceeds a certain amount, combined data with low evaluation values ​​can be discarded. Therefore, the method can also include: if the number of multiple combined data exceeds a preset threshold, then the combined data with the lowest evaluation value is deleted. By discarding combined data in this way, continuous optimization and adaptation to the installation environment can be achieved.

[0067] Furthermore, the above method may also include: if the number of multiple data combinations exceeds a preset threshold and there are multiple data combinations with the lowest evaluation values, then the data combination with the lowest evaluation value that was generated earliest is deleted. In the case of multiple data combinations with the lowest evaluation values, the data combination with the earliest generation date can be deleted to retain the more recently generated data combination, which is more suitable for the current installation environment.

[0068] Having obtained sufficient combined data based on the aforementioned machine learning method, or having completed a preset number of learning iterations, the operation of the air conditioner can be controlled based on this combined data. Therefore, the method may further include: if temperature adjustment values ​​for any one or more locations are received from user input, then determining a third target combined data set based on these temperature adjustment values; this third target combined data set is the combined data set with the smallest difference between the temperature output value and the temperature adjustment value; then, controlling the air conditioner's operation based on the parameter control values ​​of this third target combined data set.

[0069] Controlling the air conditioner's operation based on the combined data obtained in the above manner can more accurately regulate the temperature around the user and improve the user experience.

[0070] Optionally, each of the above combined data includes temperature output values ​​for multiple locations; the difference between the above temperature output values ​​and the target temperature value is the sum of the differences between the target temperature value and the corresponding temperature output value at each location; or, the difference between the above temperature output values ​​and the temperature adjustment value is the sum of the differences between the temperature adjustment value and the corresponding temperature output value at each location. Here, the summation of differences is the summation of the absolute values ​​of each difference. Users can adjust one or more locations, and the air conditioner can precisely control the temperature at multiple locations.

[0071] In addition, considering that indoor temperature is affected by outdoor temperature and solar radiation, this temperature influence is also reflected in the output value. Optionally, the combination of control value and output value is shown in Table 2. However, since outdoor temperature and solar radiation affect indoor temperature through heat conduction through the exterior walls, changes during windy or cloudy periods are ignored, and the average value over a period of time (e.g., one hour) is used.

[0072]

[0073] Table 2 The applicability evaluation formula G has a weighting coefficient. The evaluation value (fit value) of the outdoor temperature and solar radiation is set to be greater than the difference between it and the set temperature at each location (in order to limit the light intensity of the outdoor thermometer to similar conditions).

[0074] Based on this, each of the above combined data also includes outdoor temperature output value and outdoor light output value. Therefore, the formula for calculating the updated evaluation value G is as follows: G=

[0075] in, Tout_m This is the measured outdoor temperature. This is the outdoor temperature output value. Hm This is the measured light intensity. This is the outdoor light output value. F1 , F2 , Gi All are weighted coefficients. Ts_i The target temperature value for temperature test point Ti. is the temperature output value of temperature test point Ti, and n is the number of temperature test points.

[0076] Outdoor temperature and solar radiation can be measured by sensors installed on the equipment, or meteorological data can be obtained via the network. Alternatively, combined data can be stored in the cloud.

[0077] Since the indoor units are far from the users and only the intake temperature of each indoor unit is used without setting other temperature measurement points, resulting in overly coarse measurements, a temperature measuring device that can be installed near the user can be used. This temperature measuring device can also input the set temperature for that location and can transmit the data to the indoor unit via wired or wireless means.

[0078] In this embodiment of the invention, a multi-split air conditioner that connects multiple indoor units has multiple data consisting of control values ​​and multi-point room temperature data. When adjusting the indoor temperature at any location, the data whose output value is closest to the set temperature is selected from the combined data as the control parameter.

[0079] The consistency (accuracy) of the combined data in the project is evaluated, and the data with lower evaluation is discarded, while the new data is retained, so that the data can be gradually made consistent with the installation environment without updates.

[0080] This invention provides a multi-split air conditioner, including the aforementioned device for multi-point temperature control using machine learning and multiple temperature measuring devices; the temperature measuring devices are used to measure the actual temperature output value at each location.

[0081] The multi-split air conditioner includes a temperature measuring device that can acquire temperature data at any location and communicate with the air conditioner, and the temperature setting for that location can be entered into the temperature measuring device.

[0082] Picture 4 This diagram illustrates a device for multi-point temperature control using machine learning, according to an embodiment of the present invention. The device includes: The first acquisition module 401 is used to acquire a target temperature value at at least one location, and to determine a first target combination data among multiple combination data based on the target temperature value; each combination data includes a parameter control value, a temperature output value, and an evaluation value, and the first target combination data is the combination data with the smallest difference between the temperature output value and the target temperature value among all the combination data; The first operating module 402 is used to control the operation of the air conditioner according to the parameter control value of the first target combination data, and to obtain the first actual temperature output value under stable operating conditions. Evaluation module 403 is used to update the evaluation value of the first target combination data based on the temperature output value of the first target combination data and the first actual temperature output value; The second acquisition module 404 is used to determine a second target combination data among multiple combination data based on the target temperature value. The second target combination data is the combination data with the second smallest difference between the temperature output value and the target temperature value among all the combination data. The second operation module 405 is used to calculate the average value of the parameter control values ​​of the first target combination data and the second target combination data, and to control the air conditioner to operate according to the average value to obtain the second actual temperature output value under stable operation. A new module 406 is added to determine new combined data based on the average value and the second actual temperature output value; the parameter control value of the new combined data is the average value, the temperature output value is the second actual temperature output value, and the evaluation value is the initial evaluation value.

[0083] The device for multi-point temperature control using machine learning provided in this invention has multiple sets of combined data consisting of control values, multi-point room temperature data, and evaluation values. When adjusting the indoor temperature at any location, the device selects the data whose output value is closest to the set temperature from the combined data as the control parameter for the air conditioner. The device updates the evaluation value of the existing combined data based on the measured temperature value and continuously generates new combined data. This continuous updating allows the combined data to gradually adapt to the installation environment. The air conditioner can gradually adapt to the installation environment without the need for expensive sensors, thereby adjusting the temperature around the user precisely according to the user's needs.

[0084] As a possible approach, the method further includes: if the number of the multiple combined data is greater than a preset threshold, then the combined data with the lowest evaluation value is deleted.

[0085] As an alternative approach, the method further includes: if a user inputs a temperature adjustment value for any one or more locations, then determining a third target combination of data in the combination data based on the temperature adjustment value; the third target combination of data is the combination of data in which the difference between the temperature output value and the temperature adjustment value is the smallest; and controlling the operation of the air conditioner based on the parameter control value of the third target combination of data.

[0086] As a feasible approach, the method further includes: if the number of multiple combined data is greater than a preset threshold and there are multiple combined data with the lowest evaluation value, then delete the combined data with the earliest date among the combined data with the lowest evaluation value.

[0087] As one possible approach, each of the combined data includes temperature output values ​​for multiple locations; the difference between the temperature output value and the target temperature value is the sum of the differences between the target temperature value at each location and the corresponding temperature output value; or, the difference between the temperature output value and the temperature adjustment value is the sum of the differences between the temperature adjustment value at each location and the corresponding temperature output value.

[0088] As a feasible approach, the evaluation value G is calculated using the following formula: G =

[0089] in, Ts_i The target temperature value for temperature test point Ti. is the temperature output value of temperature test point Ti, and n is the number of temperature test points.

[0090] As a feasible approach, the formula for calculating the updated evaluation value G is as follows: Z_after = Z_before +

[0091] Where Z_after is the updated evaluation value, and Z_before is the evaluation value before the update. Tm"_i This represents the actual temperature output value at temperature test point Ti. α The evaluation threshold is used.

[0092] As a feasible approach, each set of combined data also includes outdoor temperature output and outdoor light intensity output, and the formula for calculating the updated evaluation value G is as follows: G =

[0093] Where Tout_m is the measured outdoor temperature. This is the outdoor temperature output value. Hm This is the measured light intensity. This is the outdoor light output value. F1 , F2 , Gi All are weighted coefficients. Ts_i The target temperature value for temperature test point Ti. is the temperature output value of temperature test point Ti, and n is the number of temperature test points.

[0094] This invention also provides a computer-readable storage medium storing a computer program. When the computer program is read and executed by a processor, it implements the method provided in the above embodiments and achieves the same technical effect. To avoid repetition, further details are omitted here. The computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0095] Of course, those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by computer-controlled devices. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the above method embodiments. The storage medium can be a memory, a disk, an optical disk, etc.

[0096] While the present invention has been disclosed above, it is not limited thereto. Any person skilled in the art can make various modifications and alterations without departing from the spirit and scope of the invention; therefore, the scope of protection of the present invention should be determined by the scope defined in the claims.

[0097] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0098] While the present invention has been disclosed above, it is not limited thereto. Any person skilled in the art can make various modifications and alterations without departing from the spirit and scope of the invention; therefore, the scope of protection of the present invention should be determined by the scope defined in the claims.

Claims

1. A method for multi-point temperature control using machine learning, characterized in that, The method includes: Acquire a target temperature value at at least one location, and determine a first target combination of data from multiple combinations based on the target temperature value; each combination of data includes a parameter control value, a temperature output value, and an evaluation value, and the first target combination of data is the combination of data with the smallest difference between the temperature output value and the target temperature value among all the combination of data; The air conditioner is controlled to operate based on the parameter control values ​​of the first target combination data, and the first actual temperature output value under stable operating conditions is obtained. The evaluation value of the first target combination data is updated based on the temperature output value of the first target combination data and the first actual temperature output value. Based on the target temperature value, a second target combination data is determined from multiple combination data. The second target combination data is the combination data with the second smallest difference between the temperature output value and the target temperature value among all the combination data. Calculate the average value of the parameter control values ​​of the first target combination data and the second target combination data, and control the air conditioner to operate according to the average value to obtain the second actual temperature output value under stable operating conditions; The newly added combined data is determined based on the average value and the second actual temperature output value; the parameter control value of the newly added combined data is the average value, the temperature output value is the second actual temperature output value, and the evaluation value is the initial evaluation value.

2. The method as described in claim 1, characterized in that, The method further includes: If the number of combined data sets exceeds a preset threshold, the combined data set with the lowest evaluation value will be deleted.

3. The method as described in claim 2, characterized in that, The method further includes: If a user inputs temperature adjustment values ​​for any one or more locations, then a third target combination of data is determined based on the temperature adjustment values; the third target combination of data is the combination of data with the smallest difference between the temperature output value and the temperature adjustment value among all the combination of data. The air conditioner is controlled to operate based on the parameter control values ​​of the third target combination data.

4. The method as described in claim 2, characterized in that, The method further includes: If the number of combined data sets exceeds a preset threshold and there are multiple combined data sets with the lowest evaluation values, then the combined data set with the lowest evaluation values ​​and the earliest date will be deleted.

5. The method as described in claim 3, characterized in that, Each of the combined data includes temperature output values ​​for multiple locations; The difference between the temperature output value and the target temperature value is the sum of the differences between the temperature values ​​at each location and the corresponding temperature output value at each location; or, The difference between the temperature output value and the temperature adjustment value is the sum of the differences between the temperature values ​​at each location of the temperature adjustment value and the corresponding temperature output value.

6. The method according to any one of claims 1-5, characterized in that, The formula for calculating the evaluation value G is as follows: G = in, Ts_i The target temperature value for temperature test point Ti. is the temperature output value of temperature test point Ti, and n is the number of temperature test points.

7. The method as described in claim 6, characterized in that, The formula for calculating the evaluation value of the updated first target combination data is as follows: Z_after = Z_before + Where Z_after is the updated evaluation value, and Z_before is the evaluation value before the update. Tm"_i This represents the actual temperature output value at temperature test point Ti. α The evaluation threshold is used.

8. The method according to any one of claims 1-5, characterized in that, Each set of combined data also includes outdoor temperature output value and outdoor light intensity output value. The calculation formula for the evaluation value G of updating the first target combined data is as follows: G = Where Tout_m is the measured outdoor temperature. This is the outdoor temperature output value. Hm This is the measured light intensity. This is the outdoor light output value. F1 , F2 , Gi All are weighted coefficients. Ts_i The target temperature value for temperature test point Ti. is the temperature output value of temperature test point Ti, and n is the number of temperature test points.

9. A device for multi-point temperature control using machine learning, characterized in that, The device includes: The first acquisition module is used to acquire a target temperature value at at least one location, and to determine a first target combination data among multiple combination data based on the target temperature value; each combination data includes a parameter control value, a temperature output value, and an evaluation value, and the first target combination data is the combination data with the smallest difference between the temperature output value and the target temperature value among all the combination data; The first operating module is used to control the operation of the air conditioner according to the parameter control value of the first target combination data, and to obtain the first actual temperature output value under stable operating conditions. The evaluation module is used to update the evaluation value of the first target combination data based on the temperature output value of the first target combination data and the first actual temperature output value. The second acquisition module is used to determine a second target combination data among multiple combination data based on the target temperature value. The second target combination data is the combination data with the second smallest difference between the temperature output value and the target temperature value among all the combination data. The second operating module is used to calculate the average value of the parameter control values ​​of the first target combination data and the second target combination data, and to control the air conditioner to operate according to the average value, so as to obtain the second actual temperature output value under stable operating conditions. A new module is added to determine new combined data based on the average value and the second actual temperature output value; the parameter control value of the new combined data is the average value, the temperature output value is the second actual temperature output value, and the evaluation value is the initial evaluation value.

10. A multi-split air conditioner, characterized in that, The device for multi-point temperature control using machine learning as described in claim 9, and a plurality of temperature measuring devices; The temperature measuring device is used to measure the actual temperature output value at each location.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when read and executed by a processor, implements the method as described in any one of claims 1-7.

Images