Rotating speed control method and device of alternating current motor and alternating current motor

Abstract

The invention provides a rotating speed control method and device of an alternating current motor and the alternating current motor, and relates to the technical field of rotating speed control of the alternating current motor, and the method comprises the steps: monitoring the operation state of the motor in real time through a sensor, obtaining three-phase current, motor rotating speed and motor output torque data, carrying out the preprocessing of the data, determining the architecture of a neural network model, and obtaining a neural network model; the method is used for accurately predicting voltage adjustment amount. According to the same data, constructing a fuzzy control model, defining a fuzzy set and a membership function, formulating a fuzzy rule, combining the trained neural network with a fuzzy control system, and dynamically adjusting the fuzzy logic through the voltage adjustment amount predicted by the neural network, and combining the voltage adjustment amount output by the neural network with the torque adjustment amount output by the fuzzy controller, and calculating a final control voltage. The motor control can be optimized in real time, the output of the motor is accurately adjusted, and the stability and efficiency of motor operation are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of speed control of AC motors, and in particular to a speed control method and device for an AC motor and the AC motor. Background Art

[0002] In the past few decades, AC motors have been widely used in various types of machinery, automation equipment and power systems as core driving components of industrial and modern equipment. However, existing technologies mostly use motor drive methods based on fixed control strategies, which are often unable to flexibly adapt to the needs of different working environments and load changes. In traditional AC motor control, the control algorithm usually relies on preset models and parameters, which makes the system rigid when dealing with rapidly changing working conditions. For example, when the motor load changes, if the control strategy cannot be adjusted in time, it will directly affect the operating efficiency of the motor and the safety of the equipment. In addition, with the development of renewable energy and intelligent manufacturing, the requirements for motor control are also increasing. Simple switch control or PID control algorithms can no longer meet the real-time adjustment requirements under complex working conditions. Therefore, a single control strategy cannot effectively cope with diverse motor application scenarios.

[0003] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0004] The object of the present invention is to provide a speed control method and device for an AC motor and an AC motor to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A method for controlling the speed of an AC motor, the specific steps comprising:

[0007] Step 1: collecting the running status data of the motor to be tested during normal operation, wherein the running status data includes the running load data, input power data and speed data of the motor, and preprocessing the collected running status data to generate a historical data set;

[0008] Step 2: Divide the historical data set into a training set, a validation set, and a test set, and build a speed deep learning model. The motor's operating load data and input power data are used as the input of the model, and the motor speed data is used as the output of the model. The speed deep learning model is trained using the training set, validation set, and test set.

[0009] Step 3: Collect the running status data of the motor to be monitored in real time, input the real-time collected running load data and input power data into the trained speed deep learning model to obtain the predicted motor speed, compare the real-time collected motor speed with the predicted motor speed, and determine the adjustment range;

[0010] Step 4: Select the PID controller as the core of the calibration, design the PID controller, set the initial proportional, integral, and differential gains according to the required adjustment amplitude, integrate the PID controller with the motor control system, and use the controller output to adjust the input voltage and frequency of the motor, thereby adjusting the speed. Further, the input characteristics are: three-phase current I a ,I b ,I c is the three-phase current value of the motor at each time step; ω is the real-time speed of the motor; T is the motor speed torque; the output target is: voltage adjustment ΔV;

[0011] The speed measurement formula is:

[0012]

[0013] Where ω(t) represents the speed of the motor, θ(t) is the angular displacement of the motor at time t;

[0014] The torque measurement formula is:

[0015] T(t)=K·I(t)

[0016] Among them, T(t) is the torque at time t, K is a constant, and I(t) is the total current, which can be calculated by the vector sum of the three-phase currents.

[0017] Furthermore, after the data collection is completed, the raw data needs to be preprocessed to clean the data, remove or correct sensor errors, signal noise and missing values, and perform data normalization to convert data of different dimensions to the same scale range. The specific formula is:

[0018]

[0019] Among them, X is the three-phase current data, motor speed data, motor output torque, and voltage adjustment data as the original data, X min and X max They are the minimum and maximum values ​​of three-phase current data, motor speed data, motor output torque, and voltage adjustment data respectively.

[0020] Furthermore, the logic of setting the number of input layer nodes, hidden layer activation function and output layer nodes is:

[0021] The number of nodes in the input layer is equal to the number of input features, which are: three-phase current I a ,I b ,I c ; Real-time speed of the motor ω; Motor speed torque T; The number of input layer nodes is 5;

[0022] Set the input feature vector to be x = [I a ,I b ,I c ,ω,T], the output is y, which represents the predicted voltage adjustment;

[0023] The calculation formula for the hidden layer is:

[0024] Use a simple feed-forward neural network with one hidden layer and one output layer:

[0025] h=f(W1x+b1)

[0026] Among them, W1 is the weight matrix from the input layer to the hidden layer, b1 is the bias of the hidden layer, and f is the activation function;

[0027] Using the Tanh activation function, the specific formula is as follows:

[0028]

[0029] Among them, x is the data of the previous input layer;

[0030] The number of nodes in the output layer is equal to the number of target variables to be predicted. The output layer has only one node, which is used to predict the voltage adjustment ΔV. The calculation formula of the output layer is:

[0031] y=g(W2h+b2)

[0032] Among them, W2 is the weight matrix from the hidden layer to the output layer, b2 is the bias of the output layer, and g is the activation function of the output layer. The specific formula is:

[0033] g(x)=x

[0034] Furthermore, the selection loss function formula is:

[0035]

[0036] Among them, y i is the actual voltage adjustment, is the voltage adjustment predicted by the neural network, and N is the number of samples;

[0037] The optimization algorithm is used to adjust the weights of the neural network to minimize the loss function. The stochastic gradient descent formula of the optimization algorithm is:

[0038]

[0039] Among them, θ is the weight parameter, η is the learning rate, is the gradient of the loss function with respect to the weights.

[0040] Furthermore, the logic of training the neural network model is as follows: the training process includes two stages: forward propagation and back propagation. Forward propagation is to calculate the input data through the network to obtain output predictions; calculating the loss is to use the loss function to calculate the difference between the predicted value and the true value; back propagation is to calculate the gradient of each weight according to the loss value, and adjust the weight using an optimization algorithm; iterative training is to repeat the process of forward propagation and back propagation until the loss function converges or the maximum number of iterations is reached.

[0041] Furthermore, in the fuzzy control model, the input variables are: error e(t): the difference between the current speed value and the actual value. Error change rate Δe(t): the rate of change of the speed error; output variables are: torque adjustment T(t): the torque adjustment to be applied;

[0042] The fuzzy set is defined for each input and output variable, and a suitable membership function is selected. The membership function of the error e(t) can be: e (x): large, negative small, zero, positive small, positive large; the membership function of the difference change rate Δe(t) can be: μ Δe (y): large, negative small, zero, positive small, positive large; the fuzzy rule is formulated as follows: if e(t) is "negative large" and Δe(t) is "negative large", then T(t) is "negative large". If e(t) is "zero" and Δe(t) is "positive small", then T(t) is "positive small";

[0043] Furthermore, the voltage adjustment value predicted by the neural network is combined with the torque adjustment value of the fuzzy controller to calculate the final control voltage formula:

[0044] ΔV control =α·ΔV+β·Δτ

[0045] Where, ΔV control The final control voltage; ΔV is the voltage adjustment; Δτ is the torque adjustment; α and β are the adjustment coefficients.

[0046] The present invention further provides a speed control device for an AC motor, the speed control device is used to execute the above-mentioned speed control method for the AC motor, comprising:

[0047] Acquisition module: collects the running status data of the motor to be detected during normal operation, wherein the running status data includes the running load data, input power data and speed data of the motor, and pre-processes the collected running status data to generate a historical data set;

[0048] Prediction module: divide the historical data set into training set, validation set and test set, build a speed deep learning model, use the motor's operating load data and input power data as the model input, and the motor speed data as the model output. Use the training set, validation set and test set to train the speed deep learning model.

[0049] Error module: collects the running status data of the motor to be monitored in real time, inputs the real-time collected running load data and input power data into the trained speed deep learning model to obtain the predicted motor speed, compares the real-time collected motor speed with the predicted motor speed, and determines the adjustment range required;

[0050] Correction module: Select PID controller as the core of calibration, design PID controller, set initial proportional, integral and differential gains according to the adjustment amplitude, integrate PID controller with motor control system, and use controller output to adjust input voltage and frequency of motor, so as to adjust speed.

[0051] The present invention also provides an AC motor, which includes the above-mentioned speed control device of the AC motor.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] Ben Amin has significantly improved the speed control capability of AC motors by combining neural networks with fuzzy algorithms to form an adaptive neural fuzzy inference system ANFIS. Specifically, the system can monitor the operating status of the motor in real time, and dynamically adjust the control parameters of the motor based on the collected data such as three-phase current, motor speed and output torque, thereby effectively improving the operating efficiency of the motor and reducing energy consumption. In addition, through intelligent control algorithms, the system can flexibly respond to changes in motor load and environmental fluctuations, ensuring that the motor can maintain efficient and safe operation under various working conditions. Compared with the prior art, the present invention not only improves the adaptive ability and response speed of the motor control system, but also enhances the reliability of the system in dealing with complex application scenarios, and ultimately achieves efficient and accurate motor control. The implementation of this innovative technology can bring significant technological progress to the fields of industrial automation and motor drive, and meet the needs of future industrial systems for efficient and intelligent control. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is a schematic diagram of the overall method flow of the present invention;

[0055] Figure 2 It is a schematic diagram of the overall device design of the present invention. DETAILED DESCRIPTION

[0056] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0057] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0058] Example:

[0059] See also Figure 1 , the present invention provides a technical solution:

[0060] A method for controlling the speed of an AC motor, the specific steps comprising:

[0061] Step 1: collecting the running status data of the motor to be tested during normal operation, wherein the running status data includes the running load data, input power data and speed data of the motor, and preprocessing the collected running status data to generate a historical data set;

[0062] Input characteristics: The three-phase current value I of the motor at each time step a ,I b ,I c ; Real-time motor speed ω; Motor speed torque T; Output target: voltage adjustment ΔV;

[0063] The speed measurement formula is:

[0064]

[0065] Where ω(t) represents the speed of the motor, θ(t) is the angular displacement of the motor at time t;

[0066] The torque measurement formula is:

[0067] T(t)=K·I(t)

[0068] Among them, T(t) is the torque at time t, K is a constant, and I(t) is the total current, which can be calculated by the vector sum of the three-phase currents.

[0069] It should be noted that the three-phase current value is detected by a current transformer, which generates a secondary current proportional to the current in the inductive current coil, and the secondary current can be detected by a digital acquisition device.

[0070] After the data collection is completed, the raw data needs to be preprocessed to clean the data, remove or correct sensor errors, signal noise and missing values, and perform data normalization to convert data of different dimensions to the same scale range. The specific formula is:

[0071]

[0072] Among them, X is the three-phase current data, motor speed data, motor output torque, and voltage adjustment data as the original data, X min and X max They are the minimum and maximum values ​​of three-phase current data, motor speed data, motor output torque, and voltage adjustment data respectively.

[0073] The logic of normalization is as follows:

[0074] In practical applications, the values ​​of different features may have different orders of magnitude or units. If normalization is not performed, these differences in dimensions may cause the model to favor features with larger values ​​and ignore features with smaller values ​​during training, thus affecting the performance of the model. Through normalization, the influence of different dimensions can be eliminated so that all features are processed by the model at the same scale.

[0075] Step 2: Divide the historical data set into a training set, a validation set, and a test set, build a speed deep learning model, use the motor's operating load data and input power data as the input of the model, and the motor speed data as the output of the model, and use the training set, validation set, and test set to train the speed deep learning model; the logic of setting the number of input layer nodes, hidden layer activation function, and output layer nodes is:

[0076] The number of nodes in the input layer is equal to the number of input features, which are: three-phase current I a ,I b ,I c ; Real-time speed of the motor ω; Motor speed torque T; The number of input layer nodes is 5;

[0077] Set the input feature vector to be x = [I a ,Ib ,I c ,ω,T], the output is y,, which shows the predicted voltage adjustment; the calculation formula of the hidden layer is:

[0078] Use a simple feed-forward neural network with one hidden layer and one output layer:

[0079] h=f(W1x+b1)

[0080] Among them, W1 is the weight matrix from the input layer to the hidden layer, b1 is the bias of the hidden layer, and f is the activation function;

[0081] Using the Tanh activation function, the specific formula is as follows:

[0082]

[0083] Among them, x is the data of the previous input layer;

[0084] The number of nodes in the output layer is equal to the number of target variables to be predicted. The output layer has only one node, which is used to predict the voltage adjustment ΔV. The calculation formula of the output layer is:

[0085] y=g(W2h+b2)

[0086] Among them, W2 is the weight matrix from the hidden layer to the output layer, b2 is the bias of the output layer, and g is the activation function of the output layer. The specific formula is:

[0087] g(x)=x.

[0088] The selection loss function formula is:

[0089]

[0090] Among them, y i is the actual voltage adjustment, is the voltage adjustment predicted by the neural network, and N is the number of samples;

[0091] The optimization algorithm is used to adjust the weights of the neural network to minimize the loss function. The stochastic gradient descent formula of the optimization algorithm is:

[0092]

[0093] Among them, θ is the weight parameter, η is the learning rate, is the gradient of the loss function with respect to the weights.

[0094] The logic of training the neural network model is as follows: the training process includes two stages: forward propagation and back propagation. Forward propagation is to calculate the input data through the network to obtain the output prediction; calculating the loss is to use the loss function to calculate the difference between the predicted value and the true value; back propagation is to calculate the gradient of each weight according to the loss value, and adjust the weight using the optimization algorithm; iterative training is to repeat the process of forward propagation and back propagation until the loss function converges or the maximum number of iterations is reached.

[0095] Step 3: Collect the running status data of the motor to be monitored in real time, input the real-time collected running load data and input power data into the trained speed deep learning model to obtain the predicted motor speed, compare the real-time collected motor speed with the predicted motor speed, and determine the adjustment range;

[0096] In the fuzzy control model, the input variables are: error e(t): the difference between the current speed value and the actual value. Error change rate Δe(t): the rate of change of the speed error; output variables are: torque adjustment T(t): the torque adjustment to be applied;

[0097] The fuzzy set is defined for each input and output variable, and a suitable membership function is selected. The fuzzy rule is formulated as follows: the membership function of the error e(t) can be: μe(x): negative, negative small, zero, positive small, positive large; the membership function of the rate of change Δe(t) can be: negative:, negative small, zero, positive small, positive large;

[0098] The fuzzy rule is defined as: if (t) is "negatively large" and Δe(t) is "negatively large", then T(t) is "negatively large". If e(t) is "zero" and Δe(t) is "positively small", then T(t) is "positively small";

[0099] Step 4: Select the PID controller as the core of the calibration, design the PID controller, set the initial proportional, integral, and differential gains according to the required adjustment range, integrate the PID controller with the motor control system, and use the controller output to adjust the motor's input voltage and frequency, thereby adjusting the speed. The data that needs to be collected include:

[0100] Furthermore, the voltage adjustment value predicted by the neural network is combined with the torque adjustment value of the fuzzy controller to calculate the final control voltage formula:

[0101] ΔV control =α·ΔV+β·Δτ

[0102] Where, ΔV control The final control voltage; ΔV is the voltage adjustment; Δτ is the torque adjustment; α and β are the adjustment coefficients.

[0103] See also Figure 2The present invention further provides a speed control device for an AC motor, the speed control device is used to execute the above-mentioned speed control method for an AC motor, comprising:

[0104] Acquisition module: collects the running status data of the motor to be detected during normal operation, wherein the running status data includes the running load data, input power data and speed data of the motor, and pre-processes the collected running status data to generate a historical data set;

[0105] Prediction module: divide the historical data set into training set, validation set and test set, build a speed deep learning model, use the motor's operating load data and input power data as the model input, and the motor speed data as the model output. Use the training set, validation set and test set to train the speed deep learning model.

[0106] Error module: collects the running status data of the motor to be monitored in real time, inputs the real-time collected running load data and input power data into the trained speed deep learning model to obtain the predicted motor speed, compares the real-time collected motor speed with the predicted motor speed, and determines the adjustment range required;

[0107] Correction module: Select PID controller as the core of calibration, design PID controller, set initial proportional, integral and differential gains according to the adjustment amplitude, integrate PID controller with motor control system, and use controller output to adjust input voltage and frequency of motor, so as to adjust speed.

[0108] The present invention also provides an AC motor, which includes the above-mentioned speed control device of the AC motor.

[0109] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0110] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. Those skilled in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0111] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0112] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A method for controlling the speed of an AC motor, characterized in that: The specific steps include: Step 1: collecting the running status data of the motor to be tested during normal operation, wherein the running status data includes the running load data, input power data and speed data of the motor, and preprocessing the collected running status data to generate a historical data set; Step 2: Divide the historical data set into a training set, a validation set, and a test set, and build a speed deep learning model. The motor's operating load data and input power data are used as the input of the model, and the motor speed data is used as the output of the model. The speed deep learning model is trained using the training set, validation set, and test set. Step 3: Collect the running status data of the motor to be monitored in real time, input the real-time collected running load data and input power data into the trained speed deep learning model to obtain the predicted motor speed, compare the real-time collected motor speed with the predicted motor speed, and determine the adjustment range; Step 4: Select the PID controller as the core of the calibration, design the PID controller, set the initial proportional, integral, and derivative gains according to the required adjustment amplitude, integrate the PID controller with the motor control system, and use the controller output to adjust the input voltage and frequency of the motor, thereby adjusting the speed.

2. The method for controlling the speed of an AC motor according to claim 1, characterized in that: The motor speed is measured by a photoelectric encoder. The formula for calculating the motor speed according to the photoelectric encoder is: V is the actual speed of the motor, N is the number of pulses output by the encoder, T is the time interval, and M is the number of pulses per revolution of the encoder.

3. The method for controlling the speed of an AC motor according to claim 1, characterized in that: After the data collection is completed, the raw data needs to be preprocessed to clean the data, remove or correct sensor errors, signal noise and missing values, and perform data normalization to convert data of different dimensions to the same scale range. The specific formula is: Among them, X' is the data point of the motor speed data and the normalized motor load data, X is the motor speed data and the motor load data as the original data, X min and X max They are respectively the minimum and maximum values ​​of the motor speed data and the motor operating load data.

4. The method for controlling the speed of an AC motor according to claim 1, characterized in that: The motor load data and motor speed data in the historical data are used as the training set and label of the deep learning model respectively. The motor load data and motor speed data in the historical data are divided into 70% training set and 30% test set, and a 15% validation set is further divided to monitor the model performance. The LSTM deep learning model for processing time series data is selected, and the input layer is constructed according to the number of features of the motor load data. Add an LSTM layer and select the activation function ReLU to construct the hidden layer; Output the predicted motor speed data in the output layer; select the mean square error (MSE) loss function, use Adam to update the parameters, use the validation set to monitor the performance of the model, check whether overfitting occurs, and adjust the learning rate, batch size, and number of network layers according to the performance of the validation set to improve the model.

5. The method for controlling the speed of an AC motor according to claim 4, characterized in that: The selection loss function formula is: Among them, y i is the actual motor speed, is the motor speed predicted by the neural network, and K is the number of samples.

6. The method for controlling the speed of an AC motor according to claim 1, characterized in that: The formula for calculating the error between the motor speed predicted by deep learning and the actual motor speed is: Where, e(t) is the speed error, y i is the actual motor speed, is the motor speed predicted by the neural network.

7. The method for controlling the speed of an AC motor according to claim 1, characterized in that: The control formula of the PID is: Among them, u(t) is the PID output signal, K p is the proportional control gain, K i is the integral control gain, K d is the differential control gain, ∫e(t)dt is the integral of the error, is the derivative of the error.

8. The method for controlling the speed of an AC motor according to claim 7, characterized in that: The output signal of the PID controller will be converted into the adjustment of the motor input voltage and input frequency. By controlling the input voltage and frequency of the motor, the motor speed can be directly affected. The voltage and frequency adjustment formula is: Among them, V(t) is the motor input voltage at time t, is the reference voltage, ΔV is the voltage adjustment amplitude, u(t) is the output signal of the PID controller, and U max is the maximum value of the control signal; f(t) is the motor input frequency at time t, f0 is the reference frequency, and Δf is the frequency adjustment amplitude; Designing an adaptive PID controller can enhance the adaptability of the system. Its gain K p , K i , K d It can be adjusted according to the size of the error and the dynamic characteristics of the system. The specific gain adjustment formula is: Among them, K p0 , K i0 , K d0 are the initial proportional, integral and differential gains, |e(t)| is the absolute value of the error, α, β, γ are the preset adaptive adjustment coefficients, |e(t)| is the absolute value of the error, is the preset target speed.

9. A speed control device for an AC motor, characterized in that: The speed control device is used to execute the speed control method of the AC motor according to any one of claims 1 to 8, comprising: Acquisition module: collects the running status data of the motor to be detected during normal operation, wherein the running status data includes the running load data, input power data and speed data of the motor, and pre-processes the collected running status data to generate a historical data set; Prediction module: divide the historical data set into training set, validation set and test set, build a speed deep learning model, use the motor's operating load data and input power data as the model input, and the motor speed data as the model output. Use the training set, validation set and test set to train the speed deep learning model. Error module: collects the running status data of the motor to be monitored in real time, inputs the real-time collected running load data and input power data into the trained speed deep learning model to obtain the predicted motor speed, compares the real-time collected motor speed with the predicted motor speed, and determines the adjustment range required; Correction module: Select PID controller as the core of calibration, design PID controller, set initial proportional, integral and differential gains according to the adjustment amplitude, integrate PID controller with motor control system, and use controller output to adjust input voltage and frequency of motor, so as to adjust speed.

10. An AC motor, characterized in that: The AC motor includes the speed control device for the AC motor according to claim 9.

Images