Motor control method and device and top air outlet outdoor unit

Abstract

The invention discloses a motor control method and device and a top air outlet outdoor unit, and belongs to the field of motor control. According to the technical scheme provided by the embodiment of the invention, multiple input current data and multiple rotation data of the motor for driving the fan blades of the top air-out outdoor unit are obtained. And processing the plurality of input current data and the plurality of rotation data by using a fan blade state determination model so as to determine the fan blade state of the fan blade, thereby realizing identification of the fan blade state of the fan blade. The motor of the top air-out outdoor unit is controlled based on the fan blade state, and the probability that the fan blades and the motor are damaged due to unbalance of the fan blades is reduced.

Description

Technical Field

[0001] The present application relates to the field of motor control, and in particular to a motor control method, device and top-outlet outdoor unit. Background Art

[0002] For the top-discharge outdoor unit, under the influence of extreme weather such as severe cold and freezing rain, the fan blades of the top-discharge outdoor unit may freeze on the surface due to factors such as water vapor condensation. If the ice on the fan blades is unevenly distributed at this time, after the motor is started, the uneven mass distribution and unbalanced fan blades may lead to unstable motor operation, over-current shutdown and other problems. The fan blades may even be damaged when the motor is running at high speed.

[0003] Therefore, how to identify the blade status of the fan blade and control the motor according to the blade status is a hot topic of research. Summary of the invention

[0004] The embodiment of the present application provides a motor control method, device and top-outlet outdoor unit, which can identify the state of a fan blade and control the motor according to the state of the fan blade. The technical solution is as follows:

[0005] In one aspect, a method for controlling a motor is provided, the method comprising:

[0006] Acquire an input current data set and a rotation data set of a motor of a top-outlet outdoor unit, wherein the motor is used to drive the fan blades of the top-outlet outdoor unit to rotate, the input current data set includes a plurality of input current data, and the rotation data set includes a plurality of rotation data;

[0007] Inputting the input current data set and the rotation data set into a fan blade state determination model, and processing the input current data set and the rotation data set by the fan blade state determination model to obtain a fan blade state of the fan blade, wherein the fan blade state includes balance and unbalance;

[0008] The motor is controlled based on the blade status of the blade.

[0009] In a possible implementation manner, the rotation data includes a rotation speed and a rotation position of a rotor of the motor, and the step of processing the input current data set and the rotation data set by the fan blade state determination model to obtain the fan blade state of the fan blade includes:

[0010] By using the fan blade state determination model, corresponding input current data in the input current data set are transformed based on a plurality of rotation positions in the rotation data set to obtain a plurality of first current pairs of the motor, wherein one of the first current pairs includes a set of d-axis current and q-axis current;

[0011] The blade state of the blade is determined by the blade state determination model based on a plurality of first current pairs of the motor and a plurality of rotation speeds in the rotation data set.

[0012] In a possible implementation, the fan blade state determination model is used to transform corresponding input current data in the input current data set based on multiple rotation positions in the rotation data set to obtain multiple first current pairs of the motor, including:

[0013] Using the wind blade state determination model, performing Clarke transformation on each input current data in the input current data set to obtain a plurality of second current pairs, wherein one second current pair includes a set of α current and β current in a two-phase stationary coordinate system;

[0014] The fan blade state determination model is used to perform Park transformation on corresponding second current pairs among the plurality of second current pairs based on a plurality of rotation positions in the rotation data set to obtain a plurality of first current pairs of the motor.

[0015] In a possible implementation manner, determining the blade state of the blade using the blade state determination model based on a plurality of first current pairs of the motor and a plurality of rotation speeds in the rotation data set includes:

[0016] By using the fan blade state determination model, feature extraction is performed on the multiple first current pairs and the multiple rotation speeds in the rotation data set to obtain the balance feature of the fan blade;

[0017] The balance feature is fully connected and normalized by the fan blade state determination model to obtain a fan blade state prediction score of the fan blade;

[0018] When the fan blade state prediction score is greater than or equal to a state score threshold, determining the fan blade state of the fan blade as balanced;

[0019] When the predicted score of the fan blade state is less than the state score threshold, the fan blade state of the fan blade is determined to be unbalanced.

[0020] In a possible implementation manner, the step of extracting features of the plurality of first current pairs and the plurality of rotation speeds in the rotation data set by using the fan blade state determination model to obtain the balance features of the fan blade includes:

[0021] By using the fan blade state determination model, the plurality of first current pairs and the plurality of rotation speeds in the rotation data set are normalized and time-frequency transformed to obtain current spectra of the plurality of first current pairs and rotation speed spectra of the plurality of rotation speeds;

[0022] By using the fan blade state determination model, feature extraction is performed on the current spectrum and the rotation speed spectrum to obtain a first spectrum feature of the rotation speed spectrum and a second spectrum feature of the rotation speed spectrum;

[0023] The first frequency spectrum feature and the second frequency spectrum feature are fused through the fan blade state determination model to obtain the balance feature of the fan blade.

[0024] In a possible implementation manner, the step of determining the model of the fan blade state and fusing the first spectrum feature and the second spectrum feature to obtain the balance feature of the fan blade includes:

[0025] Determine, by means of the fan blade state determination model, a first fusion weight corresponding to the first spectrum feature and a second fusion weight corresponding to the second spectrum feature, wherein the first fusion weight is used to indicate the degree of association between the first current pair and the fan blade state of the fan blade, and the second fusion weight is used to indicate the degree of association between the rotation speed and the fan blade state of the fan blade;

[0026] The first fusion weight and the second fusion weight are used to fuse the first spectrum feature and the second spectrum feature to obtain a balance feature of the wind blade.

[0027] In a possible implementation manner, the step of extracting features of the plurality of first current pairs and the plurality of rotation speeds in the rotation data set by using the fan blade state determination model to obtain the balance features of the fan blade includes:

[0028] By using the fan blade state determination model, the plurality of first current pairs and the plurality of rotation speeds are time-series encoded to obtain current time-series characteristics of the plurality of first currents and rotation speed time-series characteristics of the plurality of rotation speeds;

[0029] The current time series characteristics and the rotation speed time series characteristics are integrated through the fan blade state determination model to obtain the balance characteristics of the fan blade.

[0030] In a possible implementation manner, the controlling the motor based on the fan blade state of the fan blade includes:

[0031] When the fan blade is in a balanced state, controlling the motor to continue rotating according to the current control parameters;

[0032] When the fan blade is in an unbalanced state, the motor is controlled to reduce the rotation speed or stop rotating.

[0033] In a possible implementation manner, when the state of the fan blade of the fan blade is unbalanced, controlling the motor to reduce the rotation speed or stop rotating includes:

[0034] When the blade state of the fan blade is unbalanced, determining the number of times the blade state of the fan blade is continuously determined to be unbalanced;

[0035] When the number is greater than or equal to a first preset number, controlling the motor to reduce the rotation speed;

[0036] When the number is greater than or equal to a second preset number, the motor is controlled to stop rotating, and the second preset number is greater than the first preset number.

[0037] In a possible implementation manner, the training method of the wind blade state determination model includes:

[0038] Acquire multiple sample data sets of the top-outlet outdoor unit and the annotated fan blade states corresponding to each of the sample data sets, wherein the sample data sets include a sample input current data set and a sample rotation data set;

[0039] Inputting the plurality of sample data sets into the fan blade state determination model, processing each of the sample data sets by the fan blade state determination model, and obtaining a predicted fan blade state corresponding to each of the sample data sets;

[0040] The fan blade state determination model is trained based on the difference information between the predicted fan blade state and the labeled fan blade state corresponding to each of the sample data sets.

[0041] In a possible implementation manner, the acquiring of multiple sample data sets of the top-outlet outdoor unit and the annotated fan blade states corresponding to each of the sample data sets includes:

[0042] Acquire a plurality of first sample data sets of the top-outlet outdoor unit, wherein the plurality of first sample data sets are collected when the fan blades are in a balanced state, and different first sample data sets are collected when the top-outlet outdoor unit is in different blocking and return air ratios;

[0043] Acquire multiple second sample data sets of the top-outlet outdoor unit, wherein the multiple second sample data sets are sample data sets collected when the fan blade state of the fan blade is unbalanced, and different second sample data sets are collected when the top-outlet outdoor unit is under different unbalanced loads.

[0044] In one aspect, a control device for a motor is provided, the device comprising:

[0045] A data acquisition module, used to acquire an input current data set and a rotation data set of a motor of a top-outlet outdoor unit, wherein the motor is used to drive the fan blades of the top-outlet outdoor unit to rotate, the input current data set includes a plurality of input current data, and the rotation data set includes a plurality of rotation data;

[0046] a fan blade state determination module, configured to input the input current data set and the rotation data set into a fan blade state determination model, and process the input current data set and the rotation data set through the fan blade state determination model to obtain a fan blade state of the fan blade, wherein the fan blade state includes balance and unbalance;

[0047] A control module is used to control the motor based on the status of the fan blades.

[0048] In a possible implementation, the rotation data includes a rotation speed and a rotation position of a rotor of the motor, and the fan blade state determination module is used to transform corresponding input current data in the input current data set based on a plurality of rotation positions in the rotation data set through the fan blade state determination model to obtain a plurality of first current pairs of the motor, wherein a first current pair includes a set of d-axis current and q-axis current; and determine the fan blade state of the fan blade based on the plurality of first current pairs of the motor and a plurality of rotation speeds in the rotation data set through the fan blade state determination model.

[0049] In a possible implementation, the fan blade state determination module is used to perform Clarke transform on each input current data in the input current data set through the fan blade state determination model to obtain multiple second current pairs, and each second current pair includes a set of α current and β current in a two-phase stationary coordinate system; and perform Park transform on corresponding second current pairs in the multiple second current pairs based on multiple rotation positions in the rotation data set through the fan blade state determination model to obtain multiple first current pairs of the motor.

[0050] In a possible implementation, the fan blade state determination model is used to extract features of the multiple first current pairs and the multiple rotation speeds in the rotation data set to obtain balance features of the fan blade; the fan blade state determination model is used to fully connect and normalize the balance features to obtain a fan blade state prediction score of the fan blade; when the fan blade state prediction score is greater than or equal to a state score threshold, the fan blade state of the fan blade is determined to be balanced; when the fan blade state prediction score is less than the state score threshold, the fan blade state of the fan blade is determined to be unbalanced.

[0051] In a possible implementation, the fan blade state determination module is used to normalize and time-frequency transform the multiple first current pairs and the multiple rotation speeds in the rotation data set through the fan blade state determination model to obtain current spectra of the multiple first current pairs and rotation speed spectra of the multiple rotation speeds; to extract features of the current spectrum and the rotation speed spectrum through the fan blade state determination model to obtain first spectral features of the rotation speed spectrum and second spectral features of the rotation speed spectrum; and to fuse the first spectral features and the second spectral features through the fan blade state determination model to obtain a balance feature of the fan blade.

[0052] In a possible implementation, the fan blade state determination module is used to determine a first fusion weight corresponding to the first spectral feature and a second fusion weight corresponding to the second spectral feature through the fan blade state determination model, the first fusion weight being used to indicate the degree of association between the first current pair and the fan blade state of the fan blade, and the second fusion weight being used to indicate the degree of association between the rotation speed and the fan blade state of the fan blade; the first fusion weight and the second fusion weight are used to fuse the first spectral feature and the second spectral feature to obtain the balance feature of the fan blade.

[0053] In a possible implementation, the fan blade state determination module is used to perform time encoding on the multiple first current pairs and the multiple rotation speeds through the fan blade state determination model to obtain current timing characteristics of the multiple first currents and rotation speed timing characteristics of the multiple rotation speeds; and to fuse the current timing characteristics and the rotation speed timing characteristics through the fan blade state determination model to obtain the balance characteristics of the fan blade.

[0054] In a possible implementation, the control module is used to control the motor to continue rotating according to current control parameters when the blade state of the fan blade is balanced; and to control the motor to reduce the rotation speed or stop rotating when the blade state of the fan blade is unbalanced.

[0055] In a possible embodiment, the control module is used to determine the number of times the blade state of the fan blade is continuously determined to be unbalanced when the blade state of the fan blade is unbalanced; when the number is greater than or equal to a first preset number, control the motor to reduce the speed; when the number is greater than or equal to a second preset number, control the motor to stop rotating, and the second preset number is greater than the first preset number.

[0056] In a possible embodiment, the device also includes a training module for obtaining multiple sample data sets of the top-exhaust air outdoor unit and the labeled fan blade status corresponding to each of the sample data sets, the sample data sets including a sample input current data set and a sample rotation data set; inputting the multiple sample data sets into the fan blade status determination model, processing each of the sample data sets through the fan blade status determination model to obtain a predicted fan blade status corresponding to each of the sample data sets; and training the fan blade status determination model based on the difference information between the predicted fan blade status corresponding to each of the sample data sets and the labeled fan blade status.

[0057] In a possible implementation, the training module is used to obtain multiple first sample data sets of the top-outlet air outdoor unit, the multiple first sample data sets are collected when the fan blade state of the fan blade is balanced, and different first sample data sets are collected when the top-outlet air outdoor unit is in different blocking and return air ratios; obtain multiple second sample data sets of the top-outlet air outdoor unit, the multiple second sample data sets are sample data sets collected when the fan blade state of the fan blade is unbalanced, and different second sample data sets are collected when the top-outlet air outdoor unit is in different unbalanced loads.

[0058] On the one hand, a top-air outlet outdoor unit is provided, which includes one or more processors and one or more memories, wherein at least one computer program is stored in the one or more memories, and the computer program is loaded and executed by the one or more processors to implement the control method of the motor.

[0059] In one aspect, a computer-readable storage medium is provided, wherein at least one computer program is stored in the computer-readable storage medium, and the computer program is loaded and executed by a processor to implement the motor control method.

[0060] On the one hand, a computer program product or computer program is provided, which includes a program code, the program code is stored in a computer-readable storage medium, a processor of a computer device reads the program code from the computer-readable storage medium, and the processor executes the program code, so that the computer device executes the above-mentioned motor control method.

[0061] Through the technical solution provided in the embodiment of the present application, multiple input current data and multiple rotation data of the motor driving the fan blade of the top-outlet outdoor unit are obtained. The multiple input current data and multiple rotation data are processed using a fan blade state determination model to determine the fan blade state of the fan blade, thereby realizing the identification of the fan blade state of the fan blade. Based on the fan blade state, the motor of the top-outlet outdoor unit is controlled to reduce the probability of damage to the fan blade and the motor due to fan blade imbalance. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0063] Figure 1 It is a structural schematic diagram of a top-outlet outdoor unit provided in an embodiment of the present application;

[0064] Figure 2 is a flow chart of a motor control method provided in an embodiment of the present application;

[0065] Figure 3 is a flow chart of another motor control method provided in an embodiment of the present application;

[0066] Figure 4 This is a schematic diagram of data comparison before and after normalization provided in an embodiment of the present application;

[0067] Figure 5 is a flow chart of another motor control method provided in an embodiment of the present application;

[0068] Figure 6 is a flow chart of a method for training a wind blade state determination model provided in an embodiment of the present application;

[0069] Figure 7 is a schematic diagram of a hyperplane provided in an embodiment of the present application;

[0070] Figure 8 is a schematic diagram of a processing result of a kernel function provided in an embodiment of the present application;

[0071] Fig. 9 is a schematic diagram of a processing result of a Gaussian kernel function provided in an embodiment of the present application;

[0072] Fig.10 is a schematic diagram of the structure of a control device for a motor provided in an embodiment of the present application;

[0073] Fig.11 It is a structural schematic diagram of another top-outlet outdoor unit provided in an embodiment of the present application. DETAILED DESCRIPTION

[0074] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.

[0075] In this application, the terms "first", "second", etc. are used to distinguish identical or similar items with basically the same effects and functions. It should be understood that there is no logical or temporal dependency between "first", "second", and "nth", nor is there any limitation on quantity and execution order.

[0076] In order to illustrate the technical solution provided by the embodiments of the present application, some terms involved in the embodiments of the present application are introduced below.

[0077] Top-discharge outdoor unit: The top-discharge outdoor unit is part of the central air conditioning system and is mainly used for cooling and heating needs of large buildings. This type of outdoor unit is usually installed on the top of the building, such as the roof or balcony, because its air outlet faces upward. The main function of the top-discharge outdoor unit is to achieve heat exchange between indoor and outdoor through the circulation of refrigerant, providing a suitable temperature environment for the room.

[0078] Fan blades: The main function of the fan blades is to accelerate air flow and help the outdoor unit to exchange heat. When cooling, the fan is used to accelerate the air through the condenser, take away the heat in it, and condense the refrigerant into liquid; when heating, it helps increase the air flow rate and assists the condenser to release heat.

[0079] D-axis current / q-axis current: In a permanent magnet synchronous motor, the d-axis and q-axis are virtual coordinate systems based on the rotor magnetic field. The d-axis current is consistent with the direction of the rotor magnetic field, while the q-axis current is perpendicular to the direction of the rotor magnetic field. The currents of these two axes control the magnetic field and torque of the motor to achieve precise control of the motor performance.

[0080] Clark transform: also known as Clark's transform, is a very important technology in motor control, mainly used to convert the current or voltage in a three-phase stationary coordinate system into an equivalent quantity in a two-phase stationary coordinate system.

[0081] Park transform: It is a coordinate transformation method commonly used in motor control, which is used to convert the time domain components of three-phase current or voltage into two components in an orthogonal stationary coordinate system, namely the d-axis and q-axis components.

[0082] Machine Learning (ML) is a multi-disciplinary subject involving probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory, etc. It specializes in studying how computers simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge sub-models to continuously improve their performance.

[0083] Support Vector Machine (SVM): It is a supervised learning binary classification model. Its basic model is a linear classifier with the largest interval defined in the feature space. In two-dimensional space, a hyperplane can be a straight line; in three-dimensional space, it is a plane; and in higher-dimensional space, it is a hyperplane. The core idea of ​​SVM is to find a hyperplane that maximizes the interval between two types of samples. When the data set is linearly separable, SVM can find an optimal hyperplane; when the data set is approximately linearly separable or nonlinearly separable, SVM uses a kernel function to map the data to a higher-dimensional space, making the data linearly separable in the space, thereby achieving classification.

[0084] Fourier Transform: The Fourier transform is an important mathematical tool that converts a signal from the time domain to the frequency domain, allowing us to analyze the frequency characteristics of a signal.

[0085] Normalization: Mapping sequences with different value ranges to the interval (0, 1) or (-1, 1) makes it easier to process data. In some cases, the normalized values ​​can be directly implemented as probabilities.

[0086] The control method for the motor of the top-outlet outdoor unit in the related art is usually unable to identify whether the fan blades are balanced, which easily causes damage to the fan blades or the motor during the operation of the top-outlet outdoor unit. After adopting the technical solution provided in the embodiment of the present application, the fan blade state of the fan blade can be identified, so that the motor of the top-outlet outdoor unit can be controlled according to the fan blade state of the fan blade, reducing the probability of damage to the fan blade and the motor due to the imbalance of the fan blade.

[0087] Figure 1 is a schematic diagram of a top-outlet outdoor unit provided in an embodiment of the present application, see Figure 1The top-outlet outdoor unit includes a side panel 1, an air inlet 2, and a fan blade 3. The side panel 1 plays a supporting and protective role. The air inlet 2 is used to suck air into the top-outlet outdoor unit. The fan blade 3 is used to accelerate the air flow and help the top-outlet outdoor unit to perform heat exchange. Since the fan blade 3 of the top-outlet outdoor unit is vertically upward, under the influence of extreme weather such as severe cold and freezing rain, the fan blade 3 of the top-outlet outdoor unit may freeze on the surface due to factors such as water vapor condensation. If the distribution of ice on the fan blade is uneven, the fan blade will be unbalanced.

[0088] Figure 2 is a flow chart of a motor control method provided in an embodiment of the present application, see Figure 2 Taking the motor controller of the top-outlet outdoor unit as an example, the method includes the following steps.

[0089] 201. A motor controller obtains an input current data set and a rotation data set of a motor of a top-blowing outdoor unit. The motor is used to drive the fan blades of the top-blowing outdoor unit to rotate. The input current data set includes a plurality of input current data, and the rotation data set includes a plurality of rotation data.

[0090] Among them, the motor is connected to the fan blade, and the fan blade rotates under the drive of the motor. The connection between the motor and the fan blade includes direct connection and indirect connection. Direct connection refers to the direct connection between the motor shaft and the fan blade shaft of the fan blade, for example, using flange connection or bevel gear connection. Direct connection has the advantages of simple structure, compactness and high reliability; indirect connection refers to connecting the motor to the fan blade through components such as couplings. Indirect connection can reduce and balance the vibration and impact between the motor and the fan blade through the elastic buffer of the coupling, and improve stability. The embodiment of the present application does not limit the connection method between the motor and the fan blade. Input current data refers to the relevant data of the three-phase electricity directly input to the motor. For example, the input current data includes the current of the three phases in the three-phase electricity. Rotation data refers to the relevant data describing the rotation of the motor. For example, active data includes the rotation speed of the motor and the rotation position of the rotor. The input current data set is a set of multiple input current data, and the multiple input current data are collected within a preset time when the top-outlet air-conditioning outdoor unit is running. Similarly, the rotation data set is a set of multiple rotation data, and the multiple rotation data are collected within a preset time when the top-outlet air-conditioning outdoor unit is running. The preset time is set by technical personnel according to actual conditions, and the embodiments of the present application do not limit this.

[0091] 202. The motor controller inputs the input current data set and the rotation data set into a fan blade state determination model, processes the input current data set and the rotation data set through the fan blade state determination model, and obtains a fan blade state of the fan blade, wherein the fan blade state includes balance and imbalance.

[0092] The fan blade state determination model is a two-classification model, which is trained based on multiple sample data sets of the top-outlet outdoor unit and the annotated fan blade states corresponding to each sample data set, and the sample data set includes a sample input current data set and a sample rotation data set. After training, the fan blade state determination model can learn the relationship between the input current data set and the rotation data set and the fan blade state of the style, so it can classify the fan blade as balanced or unbalanced according to the input multiple input current data and multiple rotation data.

[0093] 203. The motor controller controls the motor based on the blade state of the blade.

[0094] Herein, control refers to controlling the rotation of the motor. For example, controlling the motor to reduce the speed, stop rotating, or continue rotating according to current control parameters can be called control. The current control parameters refer to the parameters used to control the motor determined under conventional control logic.

[0095] Through the technical solution provided in the embodiment of the present application, multiple input current data and multiple rotation data of the motor driving the fan blade of the top-outlet outdoor unit are obtained. The multiple input current data and multiple rotation data are processed using a fan blade state determination model to determine the fan blade state of the fan blade, thereby realizing the identification of the fan blade state of the fan blade. Based on the fan blade state, the motor of the top-outlet outdoor unit is controlled to reduce the probability of damage to the fan blade and the motor due to fan blade imbalance.

[0096] The above steps 201-203 are a brief description of the motor control method provided in the embodiment of the present application. The motor control method provided in the embodiment of the present application will be described in more detail below with reference to some examples. Figure 3 Taking the motor controller of the top-outlet outdoor unit as an example, the method includes the following steps.

[0097] 301. A motor controller obtains an input current data set and a rotation data set of a motor of a top-blowing outdoor unit. The motor is used to drive the fan blades of the top-blowing outdoor unit to rotate. The input current data set includes a plurality of input current data, and the rotation data set includes a plurality of rotation data.

[0098] Among them, the motor is connected to the fan blade, and the fan blade rotates under the drive of the motor. The connection between the motor and the fan blade includes direct connection and indirect connection. Direct connection refers to the direct connection between the motor shaft of the motor and the fan blade shaft of the fan blade, for example, using flange connection or bevel gear connection. Direct connection has the advantages of simple structure, compactness and high reliability. Indirect connection refers to connecting the motor to the fan blade through components such as couplings. Indirect connection can reduce and balance the vibration and impact between the motor and the fan blade through the elastic buffer of the coupling, and improve stability. The embodiment of the present application does not limit the connection method between the motor and the fan blade. The top-outlet outdoor unit is powered by three-phase electricity. The input current data refers to the relevant data of the three-phase electricity directly input to the motor. For example, the input current data includes the current of the U phase, V phase and W phase in the three-phase electricity. Rotation data refers to the relevant data describing the rotation of the motor. For example, active data includes the rotation speed of the motor and the rotation position of the rotor.

[0099] In a possible implementation, the motor controller obtains a plurality of input current data and a plurality of rotation data of the motor within a preset time period before the current moment, and packages the plurality of input current data into an input current data set and the plurality of rotation data into a rotation data set.

[0100] Among them, under this implementation mode, the preset time length is a period of time before the current moment, for example, the preset time length is 30s, 60s or 180s before the current moment, etc., and the embodiment of the present application does not limit this. Packing multiple input current data into the input current data set means packing multiple input current data into an input current data sequence in the order of acquisition time, and the input current data sequence can reflect the change of the input current data of the motor over time. Similarly, packing the multiple rotation data into a rotation data set means packing multiple rotation data into a rotation data sequence in the order of acquisition time, and the rotation data sequence can reflect the change of the rotation data of the motor over time. In some embodiments, the number of input current data in the input current data set is the same as the number of rotation data in the rotation data set, and any input current data in the input current data set has corresponding rotation data in the rotation data set, and the correspondence here means that the acquisition time is the same or close.

[0101] Under this implementation, multiple input current data and multiple rotation data of the motor within a preset time period before the current moment are packaged into an input current data set and a rotation data set, which can not only reflect the changes in the data history, but also eliminate the errors caused by single data, which is helpful for the subsequent use of the input current data and rotation data to realize the identification of the status of the fan blades.

[0102] For example, the input current data includes the current of the three phases in the three-phase electricity, and the rotation data includes the rotation speed and the rotation position of the rotor (or called the rotation angle). The motor controller uses the preset time length to query in the storage medium to obtain multiple input current data and multiple rotation data within the preset time length before the current moment. The motor controller packs the multiple input current data into an input current data set in the order of acquisition time, and packs the multiple rotation data into a rotation data set in the order of acquisition time.

[0103] It should be noted that since the sum of the currents of the three phases in three-phase electricity is 0, obtaining the currents of any two phases of the three-phase electricity is actually equivalent to obtaining the currents of the three phases separately. For example, the motor controller can calculate the current of the W phase by obtaining the current of the U phase and the current of the V phase.

[0104] Another implementation of the above step 301 is described below.

[0105] In a possible implementation, the motor controller obtains a plurality of input current data and a plurality of rotation data of the motor within a preset time period after the current moment, and packages the plurality of input current data into an input current data set and the plurality of rotation data into a rotation data set.

[0106] Among them, under this implementation mode, the preset duration is a period of time after the current moment, for example, the preset duration is 30s, 60s or 180s after the current moment, etc., and the embodiment of the present application does not limit this.

[0107] Under this implementation, multiple input current data and multiple rotation data of the motor acquired within a preset time length after the current moment are packaged into an input current data set and a rotation data set, which can not only reflect the latest changes in the data, but also eliminate the errors caused by single data, which is helpful for the subsequent use of the input current data and rotation data to realize the identification of the status of the fan blades.

[0108] For example, the input current data includes the current of the three phases of the three-phase electricity, and the rotation data includes the rotation speed and the rotation position of the rotor. The motor controller continuously acquires the input current data and the rotation data within a preset time period starting from the current moment, and obtains multiple input current data and multiple rotation data within a preset time period after the current moment. The motor controller packs the multiple input current data into an input current data set in the order of acquisition time, and packs the multiple rotation data into a rotation data set in the order of acquisition time.

[0109] Optionally, before obtaining the input current data set and the rotation data set through the above-mentioned implementation method, the motor controller can also pre-process the acquired multiple input current data and multiple rotation data to eliminate erroneous data and data glitches in the multiple input current data and multiple rotation data, thereby improving the accuracy of subsequent determination of the status of the wind blades.

[0110] 302. The motor controller inputs the input current data set and the rotation data set into a fan blade state determination model, processes the input current data set and the rotation data set through the fan blade state determination model, and obtains a fan blade state of the fan blade, wherein the fan blade state includes balance and imbalance.

[0111] Among them, the fan blade state determination model is a two-classification model, which is obtained by training based on multiple sample data sets of the top-outlet outdoor unit and the labeled fan blade states corresponding to each sample data set, and the sample data set includes a sample input current data set and a sample rotation data set. After training, the fan blade state determination model can learn the relationship between the input current data set and the rotation data set and the fan blade state of the style, so it can classify the fan blade as balanced or unbalanced according to the input multiple input current data and multiple rotation data. The training process of the fan blade state determination model will be described in other subsequent embodiments.

[0112] In a possible implementation, the rotation data includes a rotation speed and a rotation position of the rotor of the motor, and the motor controller inputs the input current data set and the rotation data set into a fan blade state determination model, and transforms the corresponding input current data in the input current data set based on the multiple rotation positions in the rotation data set through the fan blade state determination model to obtain multiple first current pairs of the motor, wherein one of the first current pairs includes a set of d-axis current and q-axis current. The motor controller determines the fan blade state of the fan blade based on the multiple first current pairs of the motor and the multiple rotation speeds in the rotation data set through the fan blade state determination model.

[0113] Among them, since the input current data is the current of three phases respectively, the purpose of transforming the input current data is to transform the input current into d-axis current and q-axis current, so as to facilitate subsequent processing. That is to say, in the process of determining the state of the fan blade, the d-axis current, q-axis current and rotation speed are actually used. The original collected input current data and the rotation position of the rotor are used to obtain the d-axis current and q-axis current. The principle of using the d-axis current, q-axis current and rotation speed to determine the state of the fan blade is because the d-axis current is consistent with the direction of the rotor magnetic field, while the q-axis current is perpendicular to the direction of the rotor magnetic field. When the fan blade is unbalanced, the magnetic field direction and rotation speed of the motor's rotor may change slightly, and this change can be used to identify the state of the fan blade.

[0114] In this implementation, the fan blade state determination model uses the rotation position in the rotation data to transform the input current data set into multiple d-axis currents and q-axis currents corresponding to each d-axis current, thereby helping the fan blade state determination model to determine the fan blade state. The fan blade state determination model uses multiple d-axis currents, q-axis currents corresponding to each d-axis current, and multiple rotation speeds to classify the fan blade state, which is more efficient.

[0115] In order to explain the above implementation more clearly, the above implementation will be explained in several parts below.

[0116] The first part describes the method of obtaining multiple first current pairs of the motor through the fan blade state determination model.

[0117] In a possible implementation, the motor controller performs Clarke transformation on each input current data in the input current data set through the fan blade state determination model to obtain multiple second current pairs, wherein one second current pair includes a set of α current and β current in a two-phase stationary coordinate system. The motor controller performs Parke transformation on corresponding second current pairs in the multiple second current pairs based on multiple rotation positions in the rotation data set through the fan blade state determination model to obtain multiple first current pairs of the motor.

[0118] Among them, Clark transformation and Park transformation are both ways of coordinate system transformation, and three-phase current can be converted into d-axis current and q-axis current through Clark transformation and Park transformation.

[0119] For example, the input current data includes the current of the U phase, the current of the V phase, and the current of the W phase. For any input current data in the input current data set, the motor controller determines the model through the fan blade state and uses the following formula (1) to transform the current of the U phase, the current of the V phase, and the current of the W phase into the α current and the β current, thereby obtaining the first second current pair. Formula (1) is the transformation formula corresponding to the Clarke transformation. The motor controller determines the model through the fan blade state and uses the following formula (2) to transform the α current and the β current into the d-axis current and the q-axis current, thereby obtaining a first current pair.

[0120]

[0121] Among them, I U is the U phase current, I V is the V phase current, I W is the W phase current, I α is the α current, I β is the β current, I qis the q-axis current, I d is the d-axis current and θ is the rotational position.

[0122] The second part describes a method for determining the status of the fan blade.

[0123] In a possible implementation, the motor controller extracts features of the multiple first current pairs and the multiple rotation speeds in the rotation data set through the fan blade state determination model to obtain a balance feature of the fan blade. The motor controller fully connects and normalizes the balance feature through the fan blade state determination model to obtain a fan blade state prediction score of the fan blade. When the fan blade state prediction score is greater than or equal to the state score threshold, the motor controller determines the fan blade state of the fan blade as balanced. When the fan blade state prediction score is less than the state score threshold, the motor controller determines the fan blade state of the fan blade as unbalanced.

[0124] Among them, the balance feature is a high-dimensional expression of multiple first current pairs and multiple rotation speeds. Compared with multiple first current pairs and multiple rotation speeds, the balance feature can reflect the state of the fan blade from a higher dimension. The state score threshold is set by the technician according to the actual situation, and the embodiment of the present application does not limit this.

[0125] In order to explain the above implementation more clearly, the feature extraction method in the above implementation will be further explained below.

[0126] In a possible implementation, the motor controller determines the model through the fan blade state to normalize and perform time-frequency transformation on the multiple first current pairs and the multiple rotation speeds in the rotation data set to obtain the current spectrum of the multiple first current pairs and the rotation speed spectrum of the multiple rotation speeds. The motor controller determines the model through the fan blade state to extract features from the current spectrum and the rotation speed spectrum to obtain the first spectrum feature of the rotation speed spectrum and the second spectrum feature of the rotation speed spectrum. The motor controller determines the model through the fan blade state to fuse the first spectrum feature and the second spectrum feature to obtain the balance feature of the fan blade.

[0127] The purpose of normalization is to unify the data with different dimensions and value ranges into one range after eliminating the dimensions, so as to eliminate the influence of the dimensions and value range on the processing of the wind blade state determination model and improve the universality and accuracy of the wind blade state determination model. In some embodiments, the normalized data are all within the range of (-1, 1), see Figure 4Before normalization, the U-phase current IU, the V-phase current IV, the speed, the d-axis current Id, the q-axis current Iq, and the rotation position theta all have different value ranges. After normalization, they all fall within the interval (-1, 1), which is convenient for subsequent processing. The purpose of time-frequency transformation is to convert multiple first current pairs and multiple rotation speeds in the time domain into the frequency domain, reduce the difficulty of feature extraction, and improve the efficiency of feature extraction.

[0128] In this embodiment, a wind blade state determination model is used to normalize and time-frequency transform multiple first current pairs and multiple rotation speeds, thereby obtaining a current spectrum and a rotation speed spectrum that eliminate the influence of dimension and value range. The current spectrum and the rotation speed spectrum are used to determine the balance feature, and the extraction efficiency and accuracy of the balance feature are high.

[0129] For example, the motor controller determines the model through the fan blade state to perform subsequent steps, normalizes the multiple first current pairs and the multiple rotation speeds, and obtains the normalized multiple first current pairs and the normalized multiple rotation speeds. Perform Fourier transform on the normalized multiple first current pairs and the normalized multiple rotation speeds to obtain the current spectrum and the rotation speed spectrum. Perform multiple convolutions and multiple full connections on the current spectrum to obtain the first spectrum feature. Perform multiple convolutions and multiple full connections on the rotation speed spectrum to obtain the second spectrum feature. Determine the first fusion weight corresponding to the first spectrum feature and the second fusion weight corresponding to the second spectrum feature, the first fusion weight is used to represent the degree of association between the first current pair and the fan blade state of the fan blade, and the second fusion weight is used to represent the degree of association between the rotation speed and the fan blade state of the fan blade. The first fusion weight and the second fusion weight are used to fuse the first spectrum feature and the second spectrum feature to obtain the balance feature of the fan blade.

[0130] The first fusion weight and the second fusion weight are set by the technician according to the actual situation, and the embodiment of the present application does not limit this. The convolution kernel used in multiple convolutions and the full connection matrix used in multiple full connections are gradually determined in the process of training the wind blade state determination model.

[0131] For example, the motor controller determines the subsequent steps of the model through the state of the fan blade, and uses a normalization function to process the multiple first current pairs and the multiple rotation speeds to obtain the normalized multiple first current pairs and the normalized multiple rotation speeds. Perform fast Fourier transform on the normalized multiple first current pairs and the normalized multiple rotation speeds to obtain the current spectrum and the rotation speed spectrum. Perform multiple convolutions and multiple full connections on the current spectrum to obtain the first spectrum feature. Perform multiple convolutions and multiple full connections on the rotation speed spectrum to obtain the second spectrum feature. Determine the first fusion weight and the second fusion weight. Multiply the first fusion weight with the first spectrum feature to obtain the first fusion feature. Multiply the second fusion weight with the second spectrum feature to obtain the second fusion feature. Add the first fusion feature and the second fusion feature to obtain the balance feature of the fan blade.

[0132] Another feature extraction method is described below.

[0133] In a possible implementation, the motor controller determines the model through the fan blade state to perform time-series encoding on the multiple first current pairs and the multiple rotation speeds to obtain current time-series characteristics of the multiple first currents and rotation speed time-series characteristics of the multiple rotation speeds. The motor controller determines the model through the fan blade state to fuse the current time-series characteristics and the rotation speed time-series characteristics to obtain the balance characteristics of the fan blade.

[0134] Among them, time series coding refers to using the time series relationship of data to extract features. Different from the previous feature extraction method, this feature extraction method is performed in the time domain.

[0135] In this implementation, the timing relationship of the data is used to encode multiple first current pairs and multiple rotation speeds to obtain current timing characteristics and rotation speed timing characteristics, thereby fully utilizing the temporal changes of the data to perform feature extraction, and the final balance feature has a higher accuracy.

[0136] For example, the motor controller determines the model through the fan blade state to execute subsequent steps, and sequentially encodes the multiple first current pairs based on the gating mechanism to obtain the current timing characteristics. Sequentially encode the multiple rotation speeds based on the gating mechanism to obtain the rotation speed timing characteristics. Determine the third fusion weight corresponding to the current timing characteristics and the fourth fusion weight corresponding to the rotation speed timing characteristics, the third fusion weight is used to indicate the degree of association between the first current pair and the fan blade state of the fan blade, and the fourth fusion weight is used to indicate the degree of association between the rotation speed and the fan blade state of the fan blade. The third fusion weight and the fourth fusion weight are used to fuse the current timing characteristics and the rotation speed timing characteristics to obtain the balance characteristics of the fan blade.

[0137] Among them, the gating mechanism includes being realized by a gating unit of a fan blade state determination model, the fan blade state determination model includes multiple cells, one cell includes multiple gating units, and the multiple gating units include input gates, forget gates, and output gates, which are used to control the flow of data in cells and between cells. The parameters of each gating unit are gradually determined during the training of the fan blade state determination model. The third fusion weight and the fourth fusion weight are set by technicians according to actual conditions, and the embodiments of the present application do not limit this.

[0138] For example, the motor controller executes subsequent steps through the fan blade state determination model, inputs the first first current pair of the multiple first current pairs into the first cell of the first group of cells of the fan blade state determination model, processes the first first current pair through the multiple gate control units in the first cell of the first group of cells, and obtains the cell state of the first cell in the first group of cells and the hidden features of the first first current pair. The cell state of the first cell in the first group of cells of the first group of cells, the hidden features of the first first current pair, and the second first current pair of the multiple first current pairs are input into the second cell of the first group of cells of the fan blade state determination model, and the cell state of the first cell, the hidden features of the first first current pair, and the second first current pair are processed through the multiple gate control units in the second cell of the first group of cells to obtain the cell state of the second cell in the first group of cells and the hidden features of the first first current pair. By analogy, the cell state output by the last cell in the first group of cells of the fan blade state determination model is determined as the current timing features of the multiple first current pairs. Similarly, the first rotation speed among the multiple rotation speeds is input into the first cell of the second group of cells in the fan blade state determination model, and the first rotation speed is processed by the multiple gate control units in the first cell of the second group of cells to obtain the cell state of the first cell in the second group of cells and the hidden feature of the first rotation speed. The cell state of the first cell in the second group of cells, the hidden feature of the first rotation speed, and the second rotation speed among the multiple rotation speeds are input into the second cell of the second group of cells in the fan blade state determination model, and the cell state of the first cell in the second group of cells, the hidden feature of the first rotation speed, and the second rotation speed are processed by the multiple gate control units in the second cell of the second group of cells to obtain the cell state of the second cell in the second group of cells and the hidden feature of the first rotation speed. By analogy, the cell state output by the last cell in the second group of cells in the fan blade state determination model is determined as the rotation speed time series feature of the multiple rotation speeds. The third fusion weight and the fourth fusion weight are determined. The third fusion weight is multiplied by the current time series feature to obtain the third fusion feature. The fourth fusion weight is multiplied by the rotation speed time series feature to obtain the fourth fusion feature. The third fusion feature and the fourth fusion feature are added together to obtain the balance feature of the wind blade.

[0139] 303. When the fan blade is in a balanced state, the motor controller controls the motor to continue rotating according to the current control parameters.

[0140] The current control parameter refers to a parameter determined under conventional control logic for controlling the motor, and the control method provided in step 303 means that the rotation of the motor is not interfered with.

[0141] 304. When the fan blade is in an unbalanced state, the motor controller controls the motor to reduce the rotation speed or stop rotating.

[0142] In a possible implementation, when the blade state of the fan blade is unbalanced, the motor controller determines the number of times the blade state of the fan blade is continuously determined to be unbalanced. When the number is greater than or equal to a first preset number, the motor controller controls the motor to reduce the rotation speed. When the number is greater than or equal to a second preset number, the motor controller controls the motor to stop rotating, and the second preset number is greater than the first preset number.

[0143] The first preset number and the second preset number are set by technicians according to actual conditions, and the present application embodiment does not limit this. The first preset number and the second preset number are configured to avoid the top-outlet outdoor unit being shut down by mistake due to a single fan blade state recognition error.

[0144] In this embodiment, when the fan blade is in an unbalanced state, the number of times the fan blade is continuously determined to be unbalanced is used to control the motor to reduce the speed or stop rotating, thereby avoiding an error in a single fan blade state identification causing the top air outlet outdoor unit to be erroneously shut down, thereby affecting the cooling or heating effect.

[0145] In some embodiments, when the state of the fan blade of the fan blade is unbalanced, the motor controller sends a first alarm message to the associated device of the top-outlet outdoor unit, and the first alarm message is used to indicate that the fan blade of the top-outlet outdoor unit may be unbalanced. In addition, when the number is greater than or equal to the first preset number of times, the motor controller sends a second alarm message to the associated device of the top-outlet outdoor unit, and the second alarm message is used to indicate that the fan blade of the top-outlet outdoor unit is unbalanced and the motor has been slowed down. Alternatively, when the number is greater than or equal to the second preset number of times, the motor controller sends a third alarm message to the associated device of the top-outlet outdoor unit, and the third alarm message is used to indicate that the fan blade of the top-outlet outdoor unit is unbalanced and the motor has been shut down.

[0146] The associated device is a device used by maintenance personnel of the top-outlet air-conditioning outdoor unit, and a binding relationship exists between the associated device and the top-outlet air-conditioning outdoor unit.

[0147] By sending the first alarm information, the second alarm information and the third alarm information, the user can be reminded to handle the fan blade in time to eliminate the imbalance.

[0148] In order to illustrate the technical solution provided by the embodiment of the present application, the following will be combined with Figure 5 For an explanation of the above steps 301-304, see Figure 5, the top-exhaust outdoor unit is started, and the input current set and the rotation data set are obtained. The input current set and the rotation data set are input into the fan blade state determination model, and the fan blade state determination model is used to perform data conversion on the input current set based on multiple rotation positions in the rotation data set to obtain multiple first current pairs. The multiple first current pairs and the multiple rotation speeds in the rotation data set are normalized and time-frequency transformed to obtain a current spectrum and a rotation speed spectrum. The current spectrum and the rotation speed spectrum are input into the fan blade state determination model, and the fan blade state determination model determines the blade state of the fan blade through the current spectrum and the rotation speed spectrum. When the blade state of the fan blade is balanced, the motor continues to rotate. When the blade state of the fan blade is unbalanced, the motor reduces the speed or stops rotating.

[0149] All the above optional technical solutions can be arbitrarily combined to form optional embodiments of the present application, which will not be described one by one here.

[0150] Through the technical solution provided in the embodiment of the present application, multiple input current data and multiple rotation data of the motor driving the fan blade of the top-outlet outdoor unit are obtained. The multiple input current data and multiple rotation data are processed using a fan blade state determination model to determine the fan blade state of the fan blade, thereby realizing the identification of the fan blade state of the fan blade. Based on the fan blade state, the motor of the top-outlet outdoor unit is controlled to reduce the probability of damage to the fan blade and the motor due to fan blade imbalance.

[0151] The following is an explanation of the training method of the fan blade state determination model provided in the embodiment of the present application. Figure 6 Taking the execution subject as an electronic device as an example, the method includes the following steps.

[0152] 601. The electronic device obtains a plurality of sample data sets of the top-outlet outdoor unit and a marked fan blade state corresponding to each sample data set, wherein the sample data set includes a sample input current data set and a sample rotation data set.

[0153] Among them, the sample input current data set includes multiple sample input current data, and the sample input current data is the historical input current data collected during the operation of the top-outlet outdoor unit, and is the real input current data of the motor of the top-outlet outdoor unit. Correspondingly, the sample rotation data set includes multiple sample rotation data, and the sample rotation data is the historical rotation data collected during the operation of the top-outlet outdoor unit, and is the real rotation data of the motor of the top-outlet outdoor unit. In addition, for a sample data set, the sample input current data and the sample rotation data in the sample data set exist in pairs, that is, each sample input current data has corresponding sample rotation data, and the collection time here is the same or similar.

[0154] In a possible implementation, the electronic device obtains multiple first sample data sets of the top-outlet outdoor unit, the multiple first sample data sets are collected when the fan blade state of the fan blade is balanced, and different first sample data sets are collected when the top-outlet outdoor unit is at different blocking return air ratios. The electronic device obtains multiple second sample data sets of the top-outlet outdoor unit, the multiple second sample data sets are sample data sets collected when the fan blade state of the fan blade is unbalanced, and different second sample data sets are collected when the top-outlet outdoor unit is at different unbalanced loads.

[0155] Among them, see Figure 2 , blocking the return air refers to artificially applying a blockage on the air outlet 2 to hinder the air intake of the top-outlet outdoor unit. This is to simulate the situation where the air intake pipeline is blocked to varying degrees when the top-outlet outdoor unit is working normally. In some embodiments, the return air blocking ratio includes 0%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80% and 90%, etc., which is not limited to the embodiments of the present application. The unbalanced load is achieved by technicians placing weights on the fan blades of the top-outlet outdoor unit. This is to simulate the working conditions of the top-outlet outdoor unit under different fan blade states. In the embodiments of the present application, the unbalanced load is represented by the mass of the weight. For example, the unbalanced load includes 5g, 10g and 20g, etc., which is not limited to the embodiments of the present application. Multiple first sample data sets can be regarded as positive samples for training the fan blade state determination model, and multiple second sample data sets can be regarded as negative samples for training the fan blade state determination model. The fan blade state determination model is trained by positive samples and negative samples, so that the fan blade state determination model can learn knowledge related to the two fan blade states of balance and imbalance, thereby realizing the binary classification of balance and imbalance. The first sample data set is collected under different blocking and return air ratios, and the second sample data set is collected under different balanced loads in order to cover as many working scenarios of the top-outlet outdoor unit as possible, so as to improve the generalization ability of the fan blade state determination model at the training location.

[0156] 602. The electronic device inputs the plurality of sample data sets into the fan blade state determination model, processes each sample data set through the fan blade state determination model, and obtains a predicted fan blade state corresponding to each sample data set.

[0157] Among them, step 602 and the above-mentioned step 302 belong to the same inventive concept, and the implementation process refers to the relevant description of the above-mentioned step 302, which will not be repeated here.

[0158] 603. The electronic device trains the fan blade state determination model based on the difference information between the predicted fan blade state and the labeled fan blade state corresponding to each sample data set.

[0159] In a possible embodiment, during any round of training, the electronic device trains the wind blade state determination model based on the first difference information between the predicted wind blade state and the labeled wind blade state corresponding to the first sample data set used in the round, and the second difference information between the predicted wind blade state and the labeled wind blade state corresponding to the second sample data set used in the round.

[0160] For example, in any round of training, the electronic device substitutes the first difference information and the second difference information into the loss function of the fan blade state determination model to determine the loss value of the round. The electronic device uses the loss value to perform back propagation in the fan blade state determination model using the gradient descent method to adjust the model parameters of the fan blade state determination model, thereby achieving a round of training for the fan blade state determination model.

[0161] For example, the loss function is a cross entropy loss function. In any round of training, the electronic device substitutes the first difference information and the second difference information into the cross entropy loss function of the fan blade state determination model to obtain the loss value of the round. The electronic device uses the Adam optimizer to perform back propagation in the fan blade state determination model based on the loss value to adjust the model parameters of the fan blade state determination model, thereby realizing a round of training for the fan blade state determination model.

[0162] Among them, the Adam (Adaptive Moment Estimation) optimizer is an adaptive learning rate optimization algorithm. It combines the ideas of momentum and adaptive learning rate, and adjusts the learning rate by performing exponentially weighted moving average on the first-order moment estimate and the second-order moment estimate of the gradient.

[0163] In order to more clearly illustrate the above model training process, the principle of the wind blade state determination model is explained below by taking the wind blade state determination model that is trained based on the principle of support vector machine as an example.

[0164] See also Figure 7, the current spectrum (processed data A) and the rotation speed spectrum (processed data B) are regarded as a data point, the corresponding data point when the blade state of the fan blade is balanced is white, and the corresponding data point when the blade state is unbalanced is black, wherein the processing includes normalization and time-frequency transformation. Assuming that the two groups of data points (a group of white data points and a group of black data points) are linearly separable, there must be a hyperplane y=w*x+b=0, so that the two groups of data points are completely separated, and in order to make this hyperplane more robust, the best hyperplane will be found, that is, the hyperplane that can separate the two types of data points with the maximum interval. In some embodiments, the best hyperplane is also called the maximum interval hyperplane. The goal of training the fan blade state determination model is to find this hyperplane, thereby realizing the binary classification of data points. In addition, some data points in the sample that are closest to the hyperplane are called support vectors.

[0165] Among them, the maximum margin hyperplane should meet the following two conditions: ① The two types of data points are distributed on both sides of the hyperplane; ② The distance from the data points on both sides closest to the hyperplane to the hyperplane should be the largest.

[0166] Assume that there is a hyperplane y = w*x + b = 0. For any data point in space, its distance to the hyperplane is:

[0167] |w*x+b| / ||w||

[0168] The distance from the support vector to the hyperplane is specified as d, so the following relationship holds:

[0169] (w*x+b) / ||w||≥d,y=1

[0170] (w*x+b) / ||w||≤-d,y=-1

[0171] The two equations can be combined and simplified as y*(w*x+b)≥1

[0172] Therefore, this condition satisfies the linear separability constraint. At the same time, under the condition of ensuring the constraint, the distance d should be as large as possible so that the interval between the two data points is larger, that is, the following optimization problem is satisfied:

[0173] max(d)styi*(w*xi+b)≥1,d=|w*x+b| / ||w||

[0174] If the data satisfies linear separability, there must be a maximum interval hyperplane that allows the data points to be completely separated. The above formula is simplified by derivation to obtain the following optimization problem:

[0175] min(1 / 2*||w||^2)styi*(w*xi+b)≥1

[0176] For the convex quadratic programming problem of 1 / 2*||w||^2s.t.yi*(w*xi+b)≥1, the Lagrangian duality is used to transform it into an optimization problem of dual variables. Through the equivalence of the two problems, the optimal solution of the original problem is obtained. The Lagrangian duality transformation is to add the Lagrangian operator α to the constraint conditions and define the Lagrangian function:

[0177] L(w,b,α)=1 / 2*||w||^2-∑αi*(yi*(w*xi+b)-1)

[0178] Let θ(w)=max(αi>0)(L(w,b,α)), then the original problem is minimized by 1 / 2*||w||^2, which is equivalent to directly minimizing θ(w), and the objective function is transformed into:

[0179] min(w,b)(θ(w))=min(w,b)max(αi>0)(L(w,b,α))

[0180] The maximum and minimum problems are interchangeable, and the problem eventually becomes:

[0181] max(αi>0)min(w,b)(L(w,b,α))

[0182] Solving this problem yields the maximum margin hyperplane.

[0183] However, the reality is that most classification problems are not linearly separable. For nonlinear separable problems, there is no such hyperplane. Figure 8 As shown, the solution is to map low-dimensional linearly inseparable samples to high-dimensional space, so that the sample points are linearly separable in high-dimensional space. For samples that are linearly inseparable in finite-dimensional vector space, map them to higher-dimensional vector space, and then learn to obtain support vector machines by maximizing the interval, which is nonlinear SVM. The so-called process of mapping to a higher-dimensional vector space is also the feature extraction process in the embodiment of the present application. Let φ(x) represent the new vector after x is mapped to the new feature space, then the hyperplane is y=w*φ(x)+b=0.

[0184] After the low-dimensional space is mapped to the high-dimensional space, the dimension will be very large. If the dot products of all samples are calculated, the amount of calculation is too large. Therefore, it is necessary to use the kernel function k(xi,xj)=(φ(xi),φ(xj)). The inner product of xi and xj in the feature space is equal to the result calculated by the function k(xi,xj) in the original sample space. Therefore, it is no longer necessary to calculate the inner product of high-dimensional or even infinite-dimensional space.

[0185] Therefore, if Fig. 9As shown, the Gaussian kernel function is selected as the mapping form of the data, and its expression is:

[0186] k(xi,xj)=exp(-(||xi-xj|| / 2δ^2))

[0187] Based on the kernel function, the model parameters are continuously optimized through the processed sample data set, and a fan blade state determination model based on support vector machine is trained.

[0188] Fig.10 is a schematic diagram of a motor control device provided in an embodiment of the present application, see Fig.10 The device includes: a data acquisition module 1001, a wind blade state determination module 1002 and a control module 1003.

[0189] The data acquisition module 1001 is used to acquire the input current data set and the rotation data set of the motor of the top-outlet outdoor unit. The motor is used to drive the fan blades of the top-outlet outdoor unit to rotate. The input current data set includes multiple input current data, and the rotation data set includes multiple rotation data.

[0190] The fan blade state determination module 1002 is used to input the input current data set and the rotation data set into the fan blade state determination model, and process the input current data set and the rotation data set through the fan blade state determination model to obtain the fan blade state of the fan blade, which includes balanced and unbalanced.

[0191] The control module 1003 is used to control the motor based on the status of the fan blade.

[0192] It should be noted that: the control device for the motor provided in the above embodiment only uses the division of the above functional modules as an example to illustrate when controlling the motor. In actual applications, the above functional distribution can be completed by different functional modules as needed, that is, the internal structure of the top-outlet outdoor unit is divided into different functional modules to complete all or part of the functions described above. In addition, the control device for the motor provided in the above embodiment and the control method embodiment of the motor belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0193] Through the technical solution provided in the embodiment of the present application, multiple input current data and multiple rotation data of the motor driving the fan blade of the top-outlet outdoor unit are obtained. The multiple input current data and multiple rotation data are processed using a fan blade state determination model to determine the fan blade state of the fan blade, thereby realizing the identification of the fan blade state of the fan blade. Based on the fan blade state, the motor of the top-outlet outdoor unit is controlled to reduce the probability of damage to the fan blade and the motor due to fan blade imbalance.

[0194] Fig.11 The top-outlet outdoor unit 1100 includes one or more processors 1101 and one or more memories 1102 .

[0195] The processor 1101 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 1101 may be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor 1101 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 1101 may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 1101 may also include an AI (Artificial Intelligence) processor, which is used to process computing operations related to machine learning.

[0196] The memory 1102 may include one or more computer-readable storage media, which may be non-transitory. The memory 1102 may also include a high-speed random access memory, and a non-volatile memory, such as one or more disk storage devices, flash memory storage devices. In some embodiments, the non-transitory computer-readable storage medium in the memory 1102 is used to store at least one computer program, which is executed by the processor 1101 to implement the motor control method provided in the method embodiment of the present application.

[0197] Those skilled in the art will understand that Fig.11 The structure shown in the figure does not constitute a limitation on the top-discharge outdoor unit 1100, and may include more or fewer components than shown in the figure, or combine certain components, or adopt a different component arrangement.

[0198] In an exemplary embodiment, a computer-readable storage medium is also provided, such as a memory including a computer program, and the computer program can be executed by a processor to complete the control method of the motor in the above embodiment. For example, the computer-readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), a magnetic tape, a floppy disk, an optical data storage device, etc.

[0199] In an exemplary embodiment, a computer program product or a computer program is also provided, which includes a program code, and the program code is stored in a computer-readable storage medium. A processor of a computer device reads the program code from the computer-readable storage medium, and the processor executes the program code, so that the computer device executes the above-mentioned motor control method.

[0200] In some embodiments, the computer program involved in the embodiments of the present application may be deployed and executed on a computer device, or on multiple computer devices located at one location, or on multiple computer devices distributed at multiple locations and interconnected by a communication network. Multiple computer devices distributed at multiple locations and interconnected by a communication network may constitute a blockchain system.

[0201] A person skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware or by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk or an optical disk, etc.

[0202] The above are only optional embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A method for controlling a motor, characterized in that: The method comprises: Acquire an input current data set and a rotation data set of a motor of a top-outlet outdoor unit, wherein the motor is used to drive the fan blades of the top-outlet outdoor unit to rotate, the input current data set includes a plurality of input current data, and the rotation data set includes a plurality of rotation data; Inputting the input current data set and the rotation data set into a fan blade state determination model, and processing the input current data set and the rotation data set by the fan blade state determination model to obtain a fan blade state of the fan blade, wherein the fan blade state includes balance and unbalance; The motor is controlled based on the blade status of the blade.

2. The method according to claim 1, characterized in that The rotation data includes a rotation speed and a rotation position of a rotor of the motor, and the input current data set and the rotation data set are processed by the fan blade state determination model to obtain the fan blade state of the fan blade, including: By using the fan blade state determination model, corresponding input current data in the input current data set are transformed based on a plurality of rotation positions in the rotation data set to obtain a plurality of first current pairs of the motor, wherein one of the first current pairs includes a set of d-axis current and q-axis current; The blade state of the blade is determined by the blade state determination model based on a plurality of first current pairs of the motor and a plurality of rotation speeds in the rotation data set.

3. The method according to claim 2, characterized in that The step of determining the model of the fan blade state and transforming the corresponding input current data in the input current data set based on the multiple rotation positions in the rotation data set to obtain the multiple first current pairs of the motor includes: Using the wind blade state determination model, performing Clarke transformation on each input current data in the input current data set to obtain a plurality of second current pairs, wherein one second current pair includes a set of α current and β current in a two-phase stationary coordinate system; The fan blade state determination model is used to perform Park transformation on corresponding second current pairs among the plurality of second current pairs based on a plurality of rotation positions in the rotation data set to obtain a plurality of first current pairs of the motor.

4. The method according to claim 2, characterized in that: The determining of the blade state of the blade by the blade state determination model based on a plurality of first current pairs of the motor and a plurality of rotation speeds in the rotation data set comprises: By using the fan blade state determination model, feature extraction is performed on the multiple first current pairs and the multiple rotation speeds in the rotation data set to obtain the balance feature of the fan blade; The balance feature is fully connected and normalized by the fan blade state determination model to obtain a fan blade state prediction score of the fan blade; When the fan blade state prediction score is greater than or equal to a state score threshold, determining the fan blade state of the fan blade as balanced; When the predicted score of the fan blade state is less than the state score threshold, the fan blade state of the fan blade is determined to be unbalanced.

5. The method according to claim 4, characterized in that The step of extracting features of the plurality of first current pairs and the plurality of rotation speeds in the rotation data set by using the fan blade state determination model to obtain the balance features of the fan blade includes: By using the fan blade state determination model, the plurality of first current pairs and the plurality of rotation speeds in the rotation data set are normalized and time-frequency transformed to obtain current spectra of the plurality of first current pairs and rotation speed spectra of the plurality of rotation speeds; By using the fan blade state determination model, feature extraction is performed on the current spectrum and the rotation speed spectrum to obtain a first spectrum feature of the rotation speed spectrum and a second spectrum feature of the rotation speed spectrum; The first frequency spectrum feature and the second frequency spectrum feature are fused through the fan blade state determination model to obtain the balance feature of the fan blade.

6. The method according to claim 5, characterized in that The step of determining the fan blade state model and fusing the first spectrum feature and the second spectrum feature to obtain the balance feature of the fan blade includes: Determine, by means of the fan blade state determination model, a first fusion weight corresponding to the first spectrum feature and a second fusion weight corresponding to the second spectrum feature, wherein the first fusion weight is used to indicate the degree of association between the first current pair and the fan blade state of the fan blade, and the second fusion weight is used to indicate the degree of association between the rotation speed and the fan blade state of the fan blade; The first fusion weight and the second fusion weight are used to fuse the first spectrum feature and the second spectrum feature to obtain a balance feature of the wind blade.

7. The method according to claim 4, characterized in that The step of extracting features of the plurality of first current pairs and the plurality of rotation speeds in the rotation data set by using the fan blade state determination model to obtain the balance features of the fan blade includes: By using the fan blade state determination model, the plurality of first current pairs and the plurality of rotation speeds are time-series encoded to obtain current time-series characteristics of the plurality of first currents and rotation speed time-series characteristics of the plurality of rotation speeds; The current time series characteristics and the rotation speed time series characteristics are integrated through the fan blade state determination model to obtain the balance characteristics of the fan blade.

8. The method according to claim 1, characterized in that The controlling of the motor based on the state of the fan blade comprises: When the fan blade is in a balanced state, controlling the motor to continue rotating according to the current control parameters; When the fan blade is in an unbalanced state, the motor is controlled to reduce the rotation speed or stop rotating.

9. The method according to claim 8, characterized in that When the state of the fan blade is unbalanced, controlling the motor to reduce the rotation speed or stop rotating includes: When the blade state of the fan blade is unbalanced, determining the number of times the blade state of the fan blade is continuously determined to be unbalanced; When the number is greater than or equal to a first preset number, controlling the motor to reduce the rotation speed; When the number is greater than or equal to a second preset number, the motor is controlled to stop rotating, and the second preset number is greater than the first preset number.

10. The method according to claim 1, characterized in that The training method of the wind blade state determination model includes: Acquire multiple sample data sets of the top-outlet outdoor unit and the annotated fan blade states corresponding to each of the sample data sets, wherein the sample data sets include a sample input current data set and a sample rotation data set; Inputting the plurality of sample data sets into the fan blade state determination model, processing each of the sample data sets by the fan blade state determination model, and obtaining a predicted fan blade state corresponding to each of the sample data sets; The fan blade state determination model is trained based on the difference information between the predicted fan blade state and the labeled fan blade state corresponding to each of the sample data sets.

11. The method according to claim 10, characterized in that The obtaining of a plurality of sample data sets of the top-outlet outdoor unit and the annotated fan blade states corresponding to each of the sample data sets includes: Acquire a plurality of first sample data sets of the top-outlet outdoor unit, wherein the plurality of first sample data sets are collected when the fan blades are in a balanced state, and different first sample data sets are collected when the top-outlet outdoor unit is in different blocking and return air ratios; Acquire multiple second sample data sets of the top-outlet outdoor unit, wherein the multiple second sample data sets are sample data sets collected when the fan blade state of the fan blade is unbalanced, and different second sample data sets are collected when the top-outlet outdoor unit is under different unbalanced loads.

12. A control device for a motor, characterized in that: The device comprises: A data acquisition module, used to acquire an input current data set and a rotation data set of a motor of a top-outlet outdoor unit, wherein the motor is used to drive the fan blades of the top-outlet outdoor unit to rotate, the input current data set includes a plurality of input current data, and the rotation data set includes a plurality of rotation data; a fan blade state determination module, configured to input the input current data set and the rotation data set into a fan blade state determination model, and process the input current data set and the rotation data set through the fan blade state determination model to obtain a fan blade state of the fan blade, wherein the fan blade state includes balance and unbalance; A control module is used to control the motor based on the status of the fan blades.

13. A top-outlet outdoor unit, characterized in that: The top-outlet outdoor unit includes one or more processors and one or more memories, wherein at least one computer program is stored in the one or more memories, and the computer program is loaded and executed by the one or more processors to implement the control method of the motor as described in any one of claims 1 to 11.

Images