Motor dynamic driving control method, device and system based on deep learning prediction

Abstract

The invention relates to a motor dynamic driving control method, device and system based on deep learning prediction. The method comprises the following steps: S1, acquiring current motor operation data; s2, predicting future motor operation data according to the current motor operation data through a prediction model; s3, judging a future motor working condition according to the future motor operation data; s4, selecting a control algorithm according to the future motor working condition; and S5, by taking the current motor operation data as an input condition, generating a control signal through the control algorithm. According to the method, the future change of the working condition of the motor can be mastered in time through the prediction model, so that a proper control algorithm is actively switched according to the future change, the dynamic response capability is relatively high, and relatively high motor efficiency and relatively good stability can also be obtained under special working conditions.

Description

Technical Field

[0001] The present invention relates to the field of motor technology, and in particular to a motor dynamic drive control method, device and system based on deep learning prediction. Background Art

[0002] With the development of industrial automation and intelligent manufacturing, permanent magnet synchronous servo motors (PMSM) are widely used in various mechanical equipment due to their high efficiency, precision and reliability. Permanent magnet synchronous servo motors include stators and rotors. The stator consists of an iron core and a coil to generate a rotating magnetic field. The rotor consists of permanent magnets and bearings, and the permanent magnets are used to generate a constant magnetic field.

[0003] According to the principle that like poles repel each other and opposite poles attract each other, the magnetic poles of the stator rotating magnetic field and the rotor constant magnetic field will attract each other. Therefore, by controlling the current in the stator coil, the size and direction of the rotating magnetic field can be controlled, thereby driving the rotor to rotate together. The purpose of the motor drive control method is to generate a control signal to control the current in the stator coil of the permanent magnet synchronous servo motor, and then control the permanent magnet synchronous servo motor to complete the specified task.

[0004] However, existing motor drive control methods are usually based on ideal operating conditions with stable load as models. When the motor operating conditions change, they can only passively adapt and cannot actively make adaptive adjustments. The dynamic response capability is insufficient, and the motor efficiency is low and the stability is poor under special motor operating conditions. Summary of the invention

[0005] Based on this, the object of the present invention is to provide a motor dynamic drive control method, device and system based on deep learning prediction.

[0006] The present invention provides a motor dynamic drive control method based on deep learning prediction, comprising the steps of: S1, acquiring current motor operation data; S2, predicting future motor operation data according to the current motor operation data through a prediction model; S3, judging future motor operating conditions according to the future motor operation data; S4, selecting a control algorithm according to the future motor operating conditions; S5, generating a control signal through the control algorithm with the current motor operation data as input conditions.

[0007] The present invention predicts future motor data in advance through a prediction model based on deep learning. It can timely grasp the future changes of the motor operating conditions according to the prediction results, and actively switch the appropriate control algorithm according to the future changes. Compared with the traditional PID control algorithm, it has stronger dynamic response capability and can also achieve higher motor efficiency and better stability under special working conditions.

[0008] Furthermore, the prediction model is a deep learning model trained by historical motor operation data; when new current motor operation data is obtained, the prediction model calculates the loss function by comparing the difference between the future motor operation data predicted at the previous moment and the current motor operation data, thereby optimizing its own parameters according to the loss function.

[0009] Furthermore, in step S3, the future motor operating conditions include: stable load conditions, sudden load changes and complex nonlinear conditions; in step S4, the control algorithm corresponding to the stable load conditions is a PID control algorithm, the control algorithm corresponding to the sudden load changes is an MPC control algorithm, and the control algorithm corresponding to the complex nonlinear conditions is a sliding mode control algorithm.

[0010] Further, step S3 includes: S31, when the future motor operation data meets the first preset condition, judging that the future motor operating condition is a load stable condition; the first preset condition is specifically: in the future motor operation data, the absolute value of the motor speed change rate is less than the set first speed change rate threshold, the absolute value of the motor torque change rate is less than the set first torque change rate threshold, the absolute value of the motor load change rate is less than the set first load change rate threshold, and the absolute value of the motor stator current change rate is less than the set first current change rate threshold; S32, in the future motor operation When the data meets the second preset condition, the future motor operating condition is judged to be a load sudden change condition; the second preset condition is specifically: in the future motor operation data, the absolute value of the motor speed change rate is greater than the set second speed change rate threshold, the absolute value of the motor torque change rate is greater than the set second torque change rate threshold, the absolute value of the motor load change rate is greater than the set second load change rate threshold, the absolute value of the motor stator current change rate is greater than the set second current change rate threshold, and the absolute value of the change rate of the difference between the input power and the output power of the motor is greater than the set first power difference change rate threshold.

[0011] Further, step S3 also includes: S33, when the future motor operation data meets the third preset condition, judging that the future motor operating condition is a complex nonlinear operating condition; the third preset condition is specifically: in the future motor operation data, the motor speed change rate The set speed change rate is within the normal range, the second-order derivative of the motor torque is greater than the set second-order derivative threshold, the predicted confidence of the motor load is less than the set confidence threshold, and the change rate of the motor stator current is The set normal range of the current change rate and the absolute value of the change rate of the difference between the input power and the output power of the motor are greater than the set second power difference change rate threshold.

[0012] Furthermore, step S6 is included after step S5: repeating steps S1-S5, continuously predicting the future motor operating conditions according to the current motor operating data, and switching the corresponding control algorithm according to the future motor operating conditions to generate a control signal.

[0013] Furthermore, the method further includes the following steps before step S1: S01, before the motor starts to run, predicting the initial motor operation data when the motor starts to run based on the historical motor operation data through the prediction model; S02, judging the initial motor operating condition based on the predicted initial motor operation data, and then selecting an initial control algorithm based on the initial motor operating condition; S03, using the predicted initial motor operation data as an input condition, preparing to generate an initial control signal through the initial control algorithm.

[0014] Further, step S4 specifically includes: S41, judging whether the future motor operating condition changes; S42, if the future motor operating condition changes, switching the control algorithm to the control algorithm corresponding to the future motor operating condition; step S42 switches the control algorithm to the control algorithm corresponding to the future motor operating condition, and step S5 generates a control signal through the control algorithm, specifically including: within a preset buffer time, generating a first control signal through the original control algorithm before switching, and generating a second control signal through the new control algorithm after switching; making the first control signal and the second control signal weighted superposition to obtain the final output control signal; after the preset buffer time ends, generating a control signal through the new control algorithm after switching; wherein, within the preset buffer time, the calculation formula of the control signal is: S=α1·S1+α2·S2, wherein S represents the final output control signal, S1 represents the first control signal, S2 represents the second control signal, α1 represents the weight corresponding to the first control signal, and α2 represents the weight corresponding to the second control signal; the calculation formula of the weight α1 corresponding to the first control signal is: Wherein T represents the length of the preset buffer time, t represents the current time starting from the instant of switching the control algorithm; the calculation formula of the weight α2 corresponding to the second control signal is: Wherein T represents the length of the preset buffer time, and t represents the current time starting from the instant of switching the control algorithm.

[0015] Based on the same inventive concept, the present invention also provides a motor dynamic drive control device based on deep learning prediction, including: a data acquisition module, which is used to acquire current motor operation data; a prediction model, which is used to predict future motor operation data based on the current motor operation data; a future motor operating condition judgment module, which is used to judge the future motor operating condition based on the future motor operation data; a control algorithm selection module, which is used to select a control algorithm based on the future motor operating condition; a control signal generation module, which is used to generate a control signal through the control algorithm with the current motor operation data as input conditions.

[0016] Based on the same inventive concept, the present invention also provides a motor dynamic drive system based on deep learning prediction, comprising: a sensor module, a control device and a drive module; the sensor module monitors the motor body to collect motor operation data; the control device executes any of the above-mentioned motor dynamic drive control methods based on deep learning prediction to generate a control signal; the drive module controls the drive current at the input end of the motor body according to the control signal, thereby dynamically controlling the operation of the motor body.

[0017] For better understanding and implementation, the present invention is described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a module schematic diagram of the motor dynamic drive system based on deep learning prediction of the present invention;

[0019] Figure 2 It is a module schematic diagram of a motor dynamic drive control device based on deep learning prediction of the present invention;

[0020] Figure 3 It is a flow chart of the motor dynamic drive control method based on deep learning prediction of the present invention. DETAILED DESCRIPTION

[0021] It should be clear that the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the embodiments of the present application.

[0022] The terms used in the embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to limit the embodiments of the present application. The singular forms of "a", "said" and "the" used in the embodiments of the present application and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings.

[0023] When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims. In the description of the present application, it should be understood that the terms "first", "second", "third", etc. are only used to distinguish similar objects, and do not have to be used to describe a specific order or sequence, nor can they be understood as indicating or implying relative importance. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to the specific circumstances.

[0024] In addition, in the description of this application, unless otherwise specified, "plurality" means two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the associated objects are in an "or" relationship.

[0025] The existing motor drive control method is generally a PID control algorithm based on current data, which performs poorly under special working conditions. For example, under the condition of sudden load change, the PID control algorithm cannot predict the future changes in the load, and therefore cannot adjust the motor speed in time. Not only is the dynamic response slow, but it is also prone to overshoot, affecting production efficiency and equipment stability. The present invention predicts the future motor working conditions of the motor through a prediction model, and actively and in advance switches the appropriate control algorithm according to the prediction results, which can also achieve higher motor efficiency and better stability under special working conditions.

[0026] See also Figure 1 , Figure 1 It is a module schematic diagram of the motor dynamic drive system based on deep learning prediction of the present invention. The motor dynamic drive system based on deep learning prediction of the present invention comprises: a sensor module 1, a control device 2 and a drive module 3. The sensor module 1 monitors the motor body to collect motor operation data. The control device 2 is the motor dynamic drive control device based on deep learning prediction of the present invention, which generates a dynamically adjusted control signal according to the motor operation data by executing the motor dynamic drive control method based on deep learning prediction of the present invention. The drive module 3 controls the drive current at the input end of the motor body according to the control signal, and then dynamically controls the operation of the motor body. Specifically, the drive module 3 can be an inverter, which controls the operation of the motor body by controlling the magnitude and frequency of the drive current at the input end of the motor body.

[0027] See also Figure 2 and Figure 3 , Figure 2 is a module schematic diagram of a motor dynamic drive control device based on deep learning prediction of the present invention, Figure 3 It is a flow chart of the motor dynamic drive control method based on deep learning prediction of the present invention.

[0028] Specifically, the control device 2 includes: an initialization module M0, a data acquisition module M1, a prediction model M2, a future motor operating condition judgment module M3, a control algorithm selection module M4, a control signal generation module M5 and a loop module M6.

[0029] The initialization module M0 is used to execute step S0: obtaining an initial motor operating condition, an initial control algorithm and an initial control signal.

[0030] The data acquisition module M1 is used to execute step S1: when the motor starts to run, obtain the current motor operation data collected by the sensor module 1. The motor operation data includes: the motor speed, torque, load, stator current, input power and output power.

[0031] The prediction model M2 is used to execute step S2: predicting future motor operation data based on the current motor operation data.

[0032] The prediction model M2 is used to extract features from historical motor operation data and current motor operation data, and predict future motor operation data. The prediction model M2 is a deep learning model trained by historical motor operation data. When the sensor module 1 uploads new current motor operation data, the prediction model M2 can calculate the loss function by comparing the difference between the prediction result of the future motor operation data predicted at the previous moment and the current motor operation data, thereby automatically optimizing its own parameters according to the loss function. Therefore, when the sensor module 1 uploads new current motor operation data, the current motor operation data at the previous moment is merged with the historical motor operation data to generate new historical motor operation data, and the parameters of the prediction model M2 are also continuously updated accordingly, so that it can adapt to complex and changeable motor operating conditions.

[0033] The prediction model M2 can be a deep learning model such as RNN (Recurrent Neural Network), LSTM (Long Short-Term Memory), GRU (Gate Recurrent Unit), Bi-LSTM (Bidirectional Long Short-Term Memory), etc. that can learn historical data features.

[0034] In this embodiment, the prediction model M2 is preferably a Bi-LSTM model. The Bi-LSTM model can capture historical and future sequence features at the same time; it has a self-attention mechanism and pays high attention to key time step features; it has an online optimization mechanism based on feedback errors and can adapt to changes in motor operating conditions. In a highly dynamic environment, the Bi-LSTM model can identify complex load changes in advance and provide effective control references; in nonlinear load scenarios, the Bi-LSTM model is also highly adaptable to complex systems and has good robustness. The Bi-LSTM model has excellent performance in terms of prediction accuracy, dynamic response speed, and nonlinear adaptability.

[0035] Furthermore, considering that the prediction model based on deep learning has high requirements on computing power, in one embodiment, the prediction model M2 can be deployed on the cloud. Other modules of the control device 2 communicate remotely with the prediction model M2 deployed on the cloud via a remote communication protocol such as the MQTT protocol or the HTTP protocol.

[0036] The future motor working condition judgment module M3 is used to execute step S3: judging the future motor working condition according to the future motor operation data. The future motor working condition includes: a stable load condition, a sudden load condition and a complex nonlinear condition.

[0037] The future motor working condition judgment module M3 specifically includes: a load stable working condition judgment unit, a load sudden change working condition judgment unit, a complex nonlinear working condition judgment unit and other judgment units.

[0038] The load stable working condition judgment unit is used to execute step S31: when the future motor operation data meets the first preset condition, the future motor working condition is judged to be a load stable working condition. The first preset condition is specifically: in the future motor operation data, the absolute value of the motor speed change rate is less than the set first speed change rate threshold, the absolute value of the motor torque change rate is less than the set first torque change rate threshold, the absolute value of the motor load change rate is less than the set first load change rate threshold, and the absolute value of the motor stator current change rate is less than the set first current change rate threshold. In this embodiment, the first speed change rate threshold is set to 1% of the rated speed, the first torque change rate threshold is set to 1% of the rated torque, the first load change rate threshold is set to 1% of the rated load, and the first current change rate threshold is set to 2% of the rated stator current.

[0039] The load mutation condition judgment unit is used to execute step S32: when the future motor operation data meets the second preset condition, the future motor operation condition is judged to be a load mutation condition. The second preset condition is specifically: in the future motor operation data, the absolute value of the motor speed change rate is greater than the set second speed change rate threshold, the absolute value of the motor torque change rate is greater than the set second torque change rate threshold, the absolute value of the motor load change rate is greater than the set second load change rate threshold, the absolute value of the motor stator current change rate is greater than the set second current change rate threshold, and the absolute value of the change rate of the difference between the motor input power and the output power is greater than the preset first power difference change rate threshold. In this embodiment, the second speed change rate threshold is set to 5% of the rated speed, the second torque change rate threshold is set to 10% of the rated torque, the second load change rate threshold is set to 10% of the rated load, the second current change rate threshold is set to 20% of the rated stator current, and the first power difference change rate threshold is set to 5% of the rated power.

[0040] The complex nonlinear working condition judgment unit is used to execute step S33: when the future motor operation data meets the third preset condition, judge that the future motor working condition is a complex nonlinear working condition. The third preset condition is specifically: in the future motor operation data, the motor speed change rate The set speed change rate is within the normal range, the second-order derivative of the motor torque is greater than the preset second-order derivative threshold, the predicted confidence of the motor load is less than the set confidence threshold, and the change rate of the motor stator current is The set current change rate normal range and the absolute value of the change rate of the difference between the input power and the output power of the motor are greater than the preset second power difference change rate threshold. In this embodiment, the second power difference change rate threshold is set to 5% of the rated power.

[0041] The other judgment unit is used to execute step S34: when the future motor operation data does not meet any of the above preset conditions, judge that the future motor operating condition has not changed.

[0042] The control algorithm selection module M4 is used to execute step S4: select a control algorithm according to the future motor operating condition. Step S4 specifically includes: S41, judging whether the future motor operating condition has changed. S42, if the future motor operating condition has changed, switching the control algorithm to the control algorithm corresponding to the future motor operating condition. S43, if the future motor operating condition has not changed, maintaining the current control algorithm unchanged.

[0043] In this embodiment, the control algorithm corresponding to the stable load condition is the PID control algorithm, the control algorithm corresponding to the sudden load condition is the MPC control algorithm, and the control algorithm corresponding to the complex nonlinear condition is the sliding mode control algorithm. This is because the PID control algorithm has high efficiency and low consumption, and the use of the PID control algorithm under the stable load condition has a high cost performance; the MPC control algorithm has strong dynamic response capability and is suitable for the sudden load condition; the sliding mode control algorithm has strong robustness to system parameter changes and external disturbances, and has significant advantages under the nonlinear load condition.

[0044] The control signal generating module M5 is used to execute step S5: using the current motor operation data as input conditions, and generating a control signal in real time through the control algorithm.

[0045] The loop module M6 is used to execute the steps: repeatedly call the data acquisition module M1, the prediction model M2, the future motor operating condition judgment module M3, the control algorithm selection module M4 and the control signal generation module M5, so as to repeatedly execute steps S1-S5, continuously predict the future motor operating condition based on the current motor operation data, and switch the corresponding control algorithm according to the future motor operating condition, so as to dynamically generate a control signal.

[0046] Furthermore, the inventors found that in the initial stage of motor operation, due to the lack of motor operation data, the initial control signal often overshoots, resulting in unstable and inaccurate motor operation. For this, step S0 also includes: S01, before the motor starts to run, using the prediction model, based on historical motor operation data, predicting the initial motor operation data when the motor starts to run; S02, judging the initial motor operating conditions based on the predicted initial motor operation data, and then selecting an initial control algorithm based on the initial motor operating conditions; S03, using the predicted initial motor operation data as input conditions, preparing to generate an initial control signal through the initial control algorithm. Then, in the initial stage of the motor starting to run, the motor body is controlled to run according to the initial control signal. Step S0 prepares to generate a suitable initial control signal based on the predicted initial motor operation data before the motor runs, and even if there is a lack of current motor operation data, the initial control signal will not overshoot, thereby ensuring the stability and accuracy of the motor in the initial stage of operation.

[0047] Furthermore, the inventors discovered that the problem of control signal overshoot or oscillation may occur at the moment of switching the control algorithm in step S42. This may be due to the different initial conditions of different control algorithms, and the input parameters cannot be continuously transmitted when switching the control algorithm; or the hardware computing power of the control device is limited, resulting in a decrease in the response speed at the moment of switching the algorithm. To this end, the control method of the present invention also includes the steps of: before switching the control algorithm in step S42, using a data buffer to save the current motor operation data; after switching the control algorithm in step S42, using the motor operation data of the previous moment saved in the data buffer as the input condition of the new control algorithm, maintaining the preset buffer time, and then allowing the new control algorithm to generate a control signal according to the current motor operation data.

[0048] Furthermore, step S42 switches the control algorithm to a control algorithm corresponding to the future motor operating condition, and step S5 generates a control signal through the control algorithm, specifically: within a preset buffer time, a first control signal is generated by the original control algorithm before switching, and a second control signal is generated by the new control algorithm after switching; the first control signal and the second control signal are weightedly superimposed to obtain a final output control signal; after the preset buffer time ends, a control signal is generated by the new control algorithm after switching.

[0049] Among them, within the preset buffer time, the calculation formula of the control signal is: S = α1·S1+α2·S2, where S represents the final output control signal, S1 represents the first control signal, S2 represents the second control signal, α1 represents the weight corresponding to the first control signal, and α2 represents the weight corresponding to the second control signal. The weight α1 corresponding to the first control signal and the weight α2 corresponding to the second control signal satisfy the following relationship: α1+α2=1, and α1 decreases with time, and α2 increases with time. In this embodiment, the calculation formula of the weight α1 corresponding to the first control signal is: Where T represents the length of the preset buffer time, and t represents the current time starting from the instant of switching the control algorithm. The calculation formula of the weight α2 corresponding to the second control signal is: Wherein T represents the length of the preset buffer time, and t represents the current time starting from the instant of switching the control algorithm.

[0050] The present invention has the following technical effects: 1. The present invention predicts future motor data in advance through a prediction model based on deep learning, and can timely grasp the future changes in the motor operating conditions according to the prediction results, so as to actively switch the appropriate control algorithm according to the future changes. Compared with the traditional PID control algorithm, it has a stronger dynamic response capability, and can also achieve higher motor efficiency and better stability under special working conditions. 2. The prediction model of the present invention can continuously update its own parameters according to the updated historical motor operation data, so that it can still maintain a high prediction accuracy under the condition of complex and changeable motor operating conditions. 3. The prediction model of the present invention can predict the initial motor operation data before the motor runs, so that a more appropriate initial control signal can be prepared based on the predicted initial motor operation data, ensuring the stability and accuracy of the initial stage of motor operation. 4. The present invention also designs a buffer mechanism for the unstable phenomenon at the moment of control algorithm switching, taking into account motor efficiency and stability.

[0051] The above-mentioned embodiments only express several implementation methods of the present invention, and the description is relatively specific and detailed, but it cannot be understood as limiting the scope of the invention patent. It should be pointed out that for ordinary technicians in this field, several modifications and improvements can be made without departing from the concept of the present invention, and the present invention is also intended to include these modifications and modifications.

Claims

1. A motor dynamic drive control method based on deep learning prediction, characterized in that: Includes steps: S1, obtain the current motor operation data; S2, predicting future motor operation data according to the current motor operation data through a prediction model; S3, judging the future motor operating condition according to the future motor operation data; S4, selecting a control algorithm according to the future motor operating condition; S5, using the current motor operation data as input conditions, generating a control signal through the control algorithm.

2. The motor dynamic drive control method based on deep learning prediction according to claim 1, characterized in that: The prediction model is a deep learning model trained by historical motor operation data; When new current motor operation data is obtained, the prediction model calculates the loss function by comparing the difference between the future motor operation data predicted at the last moment and the current motor operation data, thereby optimizing its own parameters according to the loss function.

3. The motor dynamic drive control method based on deep learning prediction according to claim 1, characterized in that: In step S3, the future motor operating conditions include: stable load condition, sudden load change condition and complex nonlinear condition; In step S4, the control algorithm corresponding to the stable load condition is a PID control algorithm, the control algorithm corresponding to the sudden load change condition is an MPC control algorithm, and the control algorithm corresponding to the complex nonlinear condition is a sliding mode control algorithm.

4. The motor dynamic drive control method based on deep learning prediction according to claim 3 is characterized in that: Step S3 includes: S31, when the future motor operation data satisfies a first preset condition, determining that the future motor operating condition is a load stable condition; the first preset condition is specifically: in the future motor operation data, the absolute value of the motor speed change rate is less than a set first speed change rate threshold, the absolute value of the motor torque change rate is less than a set first torque change rate threshold, the absolute value of the motor load change rate is less than a set first load change rate threshold, and the absolute value of the motor stator current change rate is less than a set first current change rate threshold; S32, when the future motor operation data meets the second preset condition, judge that the future motor operating condition is a load mutation condition; the second preset condition is specifically: in the future motor operation data, the absolute value of the motor speed change rate is greater than the set second speed change rate threshold, the absolute value of the motor torque change rate is greater than the set second torque change rate threshold, the absolute value of the motor load change rate is greater than the set second load change rate threshold, the absolute value of the motor stator current change rate is greater than the set second current change rate threshold, and the absolute value of the change rate of the difference between the input power and the output power of the motor is greater than the set first power difference change rate threshold.

5. The motor dynamic drive control method based on deep learning prediction according to claim 4 is characterized in that: Step S3 also includes: S33, when the future motor operation data meets the third preset condition, determine that the future motor operating condition is a complex nonlinear operating condition; the third preset condition is specifically: in the future motor operation data, the motor speed change rate The set speed change rate is within the normal range, the second-order derivative of the motor torque is greater than the set second-order derivative threshold, the predicted confidence of the motor load is less than the set confidence threshold, and the change rate of the motor stator current is The set normal range of the current change rate and the absolute value of the change rate of the difference between the input power and the output power of the motor are greater than the set second power difference change rate threshold.

6. The motor dynamic drive control method based on deep learning prediction according to claim 5, characterized in that: The method further includes step S6 after step S5: repeating steps S1-S5, continuously predicting the future motor operating conditions according to the current motor operating data, and switching the corresponding control algorithm according to the future motor operating conditions to generate a control signal.

7. The motor dynamic drive control method based on deep learning prediction according to claim 6 is characterized in that: Also include the steps before step S1: S01, before the motor starts to run, using the prediction model and historical motor running data to predict initial motor running data when the motor starts to run; S02, determining an initial motor operating condition according to the predicted initial motor operating data, and then selecting an initial control algorithm according to the initial motor operating condition; S03, using the predicted initial motor operation data as an input condition, preparing to generate an initial control signal through the initial control algorithm.

8. The motor dynamic drive control method based on deep learning prediction according to claim 7, characterized in that: Step S4 specifically includes: S41, determining whether the future motor operating condition changes; S42, if the future motor operating condition changes, switching the control algorithm to a control algorithm corresponding to the future motor operating condition; Step S42 switches the control algorithm to the control algorithm corresponding to the future motor working condition, and step S5 generates a control signal through the control algorithm, specifically including: Within the preset buffer time, the first control signal is generated by the original control algorithm before the switch, and the second control signal is generated by the new control algorithm after the switch; the first control signal and the second control signal are weighted and superimposed to obtain the final output control signal; after the preset buffer time ends, the control signal is generated by the new control algorithm after the switch; Wherein, within the preset buffer time, the calculation formula of the control signal is: S=α1·S1+α2·S2, where S represents the final output control signal, S1 represents the first control signal, S2 represents the second control signal, α1 represents the weight corresponding to the first control signal, and α2 represents the weight corresponding to the second control signal; The calculation formula of the weight α1 corresponding to the first control signal is: Wherein T represents the length of the preset buffer time, t represents the current time starting from the instant of switching the control algorithm; the calculation formula of the weight α2 corresponding to the second control signal is: Wherein T represents the length of the preset buffer time, and t represents the current time starting from the instant of switching the control algorithm.

9. A motor dynamic drive control device based on deep learning prediction, characterized in that: include: A data acquisition module, which is used to obtain current motor operation data; A prediction model, which is used to predict future motor operation data based on the current motor operation data; A future motor operating condition judgment module, which is used to judge the future motor operating condition according to the future motor operation data; A control algorithm selection module, which is used to select a control algorithm according to the future motor operating conditions; A control signal generating module is used to generate a control signal through the control algorithm using the current motor operation data as an input condition.

10. A motor dynamic drive system based on deep learning prediction, characterized in that: include: Sensor module, control device and drive module; The sensor module monitors the motor body to collect motor operation data; The control device executes the motor dynamic drive control method based on deep learning prediction according to any one of claims 1 to 8 to generate a control signal; The driving module controls the driving current at the input end of the motor body according to the control signal, thereby dynamically controlling the operation of the motor body.

Images