Asynchronous motor optimization control method based on multi-scalar loss model method

Abstract

The invention provides an asynchronous motor optimization control method based on a multi-scalar loss model method, and the stability and robustness of a motor control system are improved on the basis of the model. According to the control method, an induction motor loss model based on a multi-scalar control method is established by combining an asynchronous motor loss model method under an existing stator field directional control method, and a dynamic optimal stator flux linkage under the model is calculated on the basis of realizing variable excitation decoupling control, so that real-time efficiency optimization of a motor system is realized, and the efficiency of the motor system is improved. On the basis, a sliding-mode observer for changing the reaching law is provided, more accurate tracking control over the stator flux linkage and the rotating speed of the motor is achieved, and the stability and robustness of a control system are improved. According to the method, the problems that the efficiency of the asynchronous motor is too low and the control performance is poor under the low-speed or light-load condition can be solved, and on the basis that the efficiency of the motor is improved, rapid torque dynamic response and stable flux linkage adjustment within each rotating speed range of the motor are achieved.

Description

Technical Field

[0001] The present invention relates to the field of motor control technology, and specifically to an asynchronous motor optimization control method based on a multi-scalar loss model method, which aims to optimize the low-speed and light-load operation efficiency of the asynchronous motor by constructing a multi-scalar loss model of the asynchronous motor, and improve the robustness of the control system by eliminating the speed sensor. It is suitable for optimizing the efficiency of asynchronous motors in electric vehicle drive systems and various application scenarios with high requirements on motor performance. Background Art

[0002] In terms of high-performance motor control, existing methods include field-oriented control, direct torque control, intelligent control technology, and dual-torque control methods. The most commonly used method is field-oriented control. Although this method is superior, it also has many disadvantages:

[0003] (1) Slow torque response speed. Field-oriented control indirectly controls torque by controlling current. In the control of electric vehicle drive systems, fast torque response is required. The use of this method will impose certain limitations on the improvement of this performance.

[0004] (2) It depends on the magnetic flux angle and requires the conversion of the reference coordinate system, making the control system design more complicated.

[0005] (3) There is a nonlinear coupling term consisting of flux and current in the electromagnetic torque formula. When the excitation changes, the flux and torque are still coupled, and the system control performance is reduced.

[0006] In terms of motor efficiency optimization control, commonly used methods include search method and loss model method. The search method uses artificial intelligence to find the optimal flux or excitation current under different working conditions, but this method is time-consuming and complex to design, and is generally rarely used. The current loss model methods are based on field-oriented control or direct torque control, and the inherent problems of these two methods cannot be avoided, and there are limitations in motor performance control.

[0007] At present, traditional speed estimation strategies for induction motors mainly include open-loop estimation, model reference adaptive method, adaptive reduced-order observer, adaptive full-order observer, Kalman filter algorithm, etc. Traditional linear observers estimate the speed based on the mathematical model of the induction motor. The observation accuracy depends on the stability of the motor parameters and load torque. The induction motor itself is a complex system with nonlinear, time-varying, multivariable, and strong coupling characteristics. In addition, during the operation of the motor, the motor parameters will be perturbed to varying degrees due to factors such as winding temperature rise. Therefore, the linear observer method has the defect of poor robustness.

[0008] The multi-scalar model control method can achieve fast and stable torque and flux regulation in various speed ranges of electric vehicle drive systems, but there is currently a lack of research on efficiency optimization under this control method. Therefore, an asynchronous motor efficiency optimization control method based on a multi-scalar loss model method is proposed to solve the comprehensive problem of asynchronous motor control and efficiency optimization, and to improve the torque pulsation problem under direct torque control. The sliding mode observer is used to improve the poor robustness of the traditional linear observer.

[0009] Asynchronous motors are widely used in industrial production, but their energy consumption has always been a focus of attention. Traditional motor control methods, such as constant voltage-frequency ratio control and slip frequency control, can achieve speed regulation to a certain extent, but they are insufficient in terms of energy consumption and dynamic performance. The emergence of vector control technology has significantly improved the dynamic performance of AC drive systems by decoupling the control of the motor's flux and torque. However, for dual-motor systems, how to minimize motor losses while maintaining power balance remains a challenge.

[0010] In the prior art, the method of optimizing torque distribution based on the motor efficiency distribution diagram relies on the accurate measurement of motor parameters, and the preparation time is long. There are problems such as "big dragging small" and "motor no-load dragging" in the case of motor coupling. Therefore, a more accurate and efficient asynchronous motor optimization control method is needed.

[0011] Based on the inherent variable structure characteristics of asynchronous motors, the sliding mode observer is less dependent on the accuracy of the motor model. Sliding mode control is essentially a typical nonlinear control concept that cuts the state variables with a spatial sliding mode surface to make the state variables move within the sliding mode region. It can overcome the uncertainty of system parameters and has better robustness and adaptability to external interference. Summary of the invention

[0012] The present invention proposes an asynchronous motor optimization control method based on a multi-scalar loss model method, which combines a multi-scalar control method with better motor performance control and an optimal flux control of a loss model method for high performance and efficiency optimization control of an asynchronous motor, and proposes a sliding mode observer, which can realize fast and stable torque and flux adjustment of a control system, improve the real-time efficiency of the asynchronous motor under various working conditions, and improve the system robustness and save costs; the steps in the use process are:

[0013] Step (1): construct a mathematical model of induction motor loss based on multi-scalar control;

[0014] When iron loss is considered, for a circuit consisting of an iron loss branch and an excitation branch, according to Kirchhoff's voltage law, the α-axis component of the iron loss current can be obtained as:

[0015]

[0016] Among them, i Feα is the α-axis component of the iron loss current, L m is the mutual inductance of the stator and rotor coaxial equivalent windings, ω e is the synchronous speed of the asynchronous motor, R Fe is the equivalent iron loss resistance, L s is the stator equivalent self-inductance, ψ sβ is the β-axis component of the stator flux;

[0017] The β-axis component of the iron loss current is:

[0018]

[0019] Among them, i Feβ is the β-axis component of the iron loss current, ψ sα is the α-axis component of the rotor flux;

[0020] Taking the stator current and stator flux as state variables, the state equation of the motor system in the αβ coordinate system can be derived as follows:

[0021]

[0022] Among them, i sα is the α-axis component of the stator current, di sα is the change of the α-axis component of the stator current, dt is the change of time, is the derivative of the α-axis component of the stator current, i sβ is the β-axis component of the stator current, di sβ is the change of the β-axis component of the stator current, is the derivative of the β-axis component of the stator current, u sα is the α-axis component of the input stator voltage, u sβ is the β-axis component of the input stator voltage, dψ sα is the change of the α-axis component of the stator current, dψ sβ is the change of the β-axis component of the stator current, is the derivative of the α-axis component of the stator flux, is the derivative of the stator flux β-axis component, R r is the rotor equivalent resistance, R s is the stator equivalent resistance, ω r is the motor rotor speed, and defines the current coefficient Magnetic flux coefficient Magnetic flux coefficient Magnetic flux coefficient Defining Voltage Coefficient σ is the leakage inductance coefficient, L r is the rotor equivalent self-inductance;

[0023] The electromagnetic torque considering iron loss is:

[0024]

[0025] Among them, T e is the electromagnetic torque, n p is the pole pair number;

[0026] Without considering the iron consumption, multi-scalar control is used to define four state variables [ω r ,ψ s 2 ,τ s ,η s ] T , [] T represents the transpose of the row vector, where τ s is the cross product of the stator flux and the stator current vector, is called normalized torque; η s is the dot product of the stator flux and the stator current vector, It is called normalized reactive torque; represents vector cross product, ⊙ represents vector dot product, is the stator flux vector form, is the stator current vector form, is the square of the stator flux modulus;

[0027] The system control signal is defined as:

[0028]

[0029] Among them, u τs is the input signal controlled by the defined electromagnetic torque loop, u ηs The input signal controlled by the defined reactive torque loop is: is the input stator voltage vector form,

[0030] Considering the iron loss, according to formula (4), the normalized active torque is redefined as:

[0031]

[0032] Redefine the normalized reactive torque as:

[0033] η s =ψ sα (i sα -i Feα )+ψ sβ (i sβ -iFeβ ) (7)

[0034] Substituting into equations (1) and (2), the reactive torque is converted to:

[0035]

[0036] The reactive torque obtained from equation (8) when considering iron loss is consistent with the original definition, because iron loss has nothing to do with reactive power;

[0037] The state equation after calculation is:

[0038]

[0039] Among them, dω r is the change in rotor speed, is the derivative of the rotor speed, is the change of stator flux modulus, is the derivative of the square of the stator flux modulus, dτ s is the change of normalized torque, is the derivative of the normalized torque, dη s is the normalized reactive torque change, is the derivative of the normalized reactive torque, J is the moment of inertia, T L is the load torque;

[0040] Step (2): The model described in equation (9) in step (1) is a nonlinear system, and linear decoupling is adopted;

[0041] In order to realize the dual closed-loop control of the mechanical system and the electromagnetic system, a linear decoupling process is performed to obtain a decoupled linear system; the nonlinear feedforward control quantity is defined as follows:

[0042]

[0043] Among them, u τs_f is the input signal controlled by the feedforward compensation electromagnetic torque loop, u ηs_f is the input signal controlled by the feedforward compensation reactive torque loop, L σ is the compensation coefficient, and equation (10) is substituted into equation (9). The state equation after linear decoupling is:

[0044]

[0045] The first and third equations in equation (11) are for the mechanical subsystem, and the other two equations are for the electromagnetic subsystem. Based on the motor loss model, this control method can divide the motor system into two parts for closed-loop control. Compared with the field-oriented control, the current inner loop is replaced by the dual torque inner loop, which directly controls the torque, which can speed up the system response speed, and the two systems are decoupled from each other and do not affect each other.

[0046] The dual torque inner loop output is the system control signal defined in step (1), and the control component of the input stator voltage αβ coordinate system is:

[0047]

[0048] Step (3): Under different steady-state conditions, the controllable loss of the motor is calculated, that is, the motor loss including the stator and rotor copper loss and the stator iron loss, ignoring the extremely small rotor iron loss, which is briefly referred to as the total loss, accounting for about 80% of the system loss; at the same time, the optimal stator flux that minimizes the total loss is calculated, and it is input as the reference stator flux into the double closed-loop control system described in step (2) to achieve efficiency optimization control; according to the equivalent circuit considering the iron loss, the asynchronous motor power diagram is obtained, and the total motor loss is calculated as:

[0049]

[0050] Under different working conditions, in steady-state operation, the motor speed and torque are constant, and the influence of motor parameter changes is ignored. When the derivative of total loss to stator flux is 0, the optimal stator flux that minimizes total loss is obtained as:

[0051]

[0052] Under different operating conditions of the system, the reference flux value is adjusted according to the optimal stator flux, thereby minimizing the controllable loss of the motor and improving the motor efficiency;

[0053] Step (4): Propose a motor control system observer considering iron loss; select stator current and stator flux as state variables, and the mathematical model of induction motor considering iron loss is as shown in formula (3); based on formula (3), the sliding mode observer is designed as:

[0054]

[0055] In the formula, is the change in the estimated value of the stator current α axis, is the derivative of the estimated value of the stator current about the α axis, is the estimated value of the stator current α axis, is the change in the estimated value of the stator current β axis, is the derivative of the estimated value of the stator current β axis, is the estimated value of the stator current β axis, is the change in the estimated value of the stator flux α axis, is the derivative of the estimated value of the stator flux α axis, is the estimated value of the stator flux α axis, is the change in the estimated value of the stator flux β axis, is the derivative of the estimated value of the stator flux β axis, is the estimated value of stator flux β axis, is the estimated value of the motor rotor speed, and the sliding mode control term is:

[0056]

[0057] Where V α1 is the sliding mode control term of the estimated value of the stator current α axis, V β1 is the sliding mode control term of the stator current β-axis estimate, V α2 is the sliding mode control term of the estimated value of the stator flux α axis, V β2 is the sliding mode control term of the stator flux β-axis estimation value, K1 is the observer gain of the stator current α-axis estimation value, K2 is the observer gain of the stator current β-axis estimation value, K3 is the observer gain of the stator flux α-axis estimation value, K4 is the observer gain of the stator flux β-axis estimation value, K1, K2, K3, K4 are all constants greater than zero; and are the stator current estimation errors of the α-axis and β-axis respectively, and the linear sliding surface is selected as S α =e isα and S β =e isβ , the specific form of the symbolic function sgn() is:

[0058]

[0059] Applying the power reaching law to the motor observation model, we get the sliding mode observer equation as shown in equation (15). The sliding mode control law V after application is α1 、V β1 、V α2 、V β2 for:

[0060]

[0061] Where ε1, ε2, ε3, ε4 are the gains of the current observer α axis, the current observer β axis, the flux observer α axis, and the flux observer β axis, respectively; k is the gain of the current and flux observers, all of which are constants greater than 0. The range of their values ​​is obtained by analyzing the stability of the system. The design of this control law makes the sliding surface approach zero more quickly, and optimizes the estimated stator current and stator flux. By adjusting the appropriate parameters, the dynamic response and stability of the system are improved, and the torque pulsation is reduced. If the motor parameters are accurate, equation (15) minus equation (3) can obtain the differential equation group of the stator current and stator flux error terms:

[0062]

[0063] In the formula, de isα is the change in the stator current α-axis error, is the derivative of the stator current α-axis error, de isβ is the change in the stator current β-axis error, is the derivative of the stator current β-axis error, de ψsα is the change in the stator flux α-axis error, is the derivative of the stator flux α-axis error, de ψsβ is the change in the stator flux β-axis error, is the derivative of the stator flux β-axis error, are the stator flux estimation errors of the α-axis and β-axis respectively, is the rotor speed estimation error;

[0064] Step (5): Adaptive identification of motor speed; the motor speed is estimated by the stability principle of the Lyapunov function; the following Lyapunov function is selected from equation (19):

[0065]

[0066] Where V is the constructed Lyapunov function, is the derivative of the constructed Lyapunov function, e ψs T =[e ψsα e ψsβ ] T , is the derivative of the stator flux error, is the derivative of the motor rotor speed error; to ensure is negative definite and bounded, simplifying the analysis and calculation, so and At the same time; since the motor mechanical time constant is much smaller than the electrical time constant, the motor speed changes much slower than the stator current and stator flux and other electrical quantities, so the speed remains unchanged, and the differential of the speed error term can be obtained, that is, The derivative form of the estimated speed can be solved by combining equations (19) and (20):

[0067]

[0068] Where e and γ are constants greater than 0, L T =[sgn(e isα )sgn(e isβ )], is the error of the estimated stator current, is the estimated value of stator flux; in order to improve the accuracy of the speed dynamic estimation method, the proportional integral control rate is introduced in formula (21) to improve the response speed and response time of the speed, so that the final speed estimation method is:

[0069] ω r =(K p +∫K i dt)|ψ sα e isβ -ψ sβ e isα | (22)

[0070] In the formula, K p and K i is the gain of the proportional-integral regulator; the overall structure of the sliding mode observer using the power reaching law is to estimate the stator current and stator flux through equation (19), and finally realize the speed estimation by equation (22).

[0071] The existing asynchronous motor field-oriented control method has a slow response speed, and the torque and flux decoupling is not good enough during variable excitation control; the multi-scalar model control method can achieve better decoupling of torque and flux and its stability and rapid response, but there is little research on efficiency optimization. The present invention combines dual torque control with the efficiency optimization idea, proposes an asynchronous motor efficiency optimization control method based on the multi-scalar loss model method, and proposes a sliding mode observer. The variable containing the stator current estimation information is designed as the power reaching law gain to achieve the adaptation of the system state. Then the stator current and flux observers are constructed, and finally the speed adaptation law is derived based on the Lyapunov stability theorem to achieve the performance optimization of the induction motor speed estimation. The proposed control strategy can accurately estimate the speed and further weaken the jitter based on the traditional observer. While improving the motor efficiency, it achieves faster dynamic response and flux stability than the field-oriented control method, especially for the efficient, fast and stable control of electric vehicle drive motors in the full speed range, which has certain practical significance. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 It is a control system block diagram based on the multi-scalar loss model method of the present invention. DETAILED DESCRIPTION

[0073] An asynchronous motor optimization control method based on a multi-scalar loss model method proposed in the present invention is described in detail as follows in conjunction with the accompanying drawings:

[0074] Figure 1 This is a control system block diagram based on the multi-scalar loss model method of the present invention. The present invention is a dual closed-loop control of the mechanical system and the electromagnetic system. The outer loop is a speed and flux loop. The reference speed is determined by the different working conditions of the application scenario. The reference flux value is calculated by the optimal flux square calculation module, instead of fixed excitation. The feedback is the stator flux and speed estimated by the sliding mode observer. The inner loop is an electromagnetic torque and reactive torque loop, which directly controls the system output torque, can speed up the response speed, and after obtaining the voltage signal required for the control, the modulation signal required by the inverter is obtained through space vector pulse width control. Then, the three-phase AC is output through the three-phase inverter to power the motor.

[0075] The parameters of a three-phase asynchronous motor are as follows: rated power 2.2kW, rated voltage 380V, rated frequency 50Hz, rated speed 1422r / min, n p =2, R s =3.4Ω, R r =2.444Ω, L s =0.2724H,L r =0.2715H,L m =0.2631H, R Fe=187.6Ω, J = 0.005kg·m 2 . Figure 1 The given rotor speed is where n * is the desired speed, rotor speed ω r =n×2×2π / 60, n is the motor output speed.

[0076] Step (1): Construct a mathematical model of induction motor losses based on dual torque control in this embodiment.

[0077] According to the decoupled state equation proposed in the present invention, the known motor parameters are substituted to construct the state space equation of this embodiment:

[0078]

[0079] Among them, ω r is the motor rotor speed, τ s is the normalized electromagnetic torque, T L is the load torque, η s is the normalized reactive torque, is the square of the stator flux, u τs is the input signal controlled by the defined electromagnetic torque loop, u ηs is the input signal controlled by the defined reactive torque loop. The state equations correspond to Figure 1 Inner and outer loop PI control.

[0080] Figure 1 The feedforward compensation is:

[0081]

[0082] Among them, ω e is the synchronous speed of the motor, L σ is the compensation coefficient, Figure 1 In the equation (1), the dual torque inner loop output is the system control signal defined in equation (1), and the control component of the input stator voltage αβ coordinate system is:

[0083]

[0084] Step (2): Under different steady-state conditions, calculate the total motor loss and the optimal stator flux that minimizes the total loss. The total motor control loss is:

[0085]

[0086] Among them, τ s is the electromagnetic torque.

[0087] Figure 1 The reference flux of the middle flux ring, that is, the optimal stator flux, is:

[0088]

[0089] Step (3): Select stator current and stator flux as state variables and design a sliding mode observer:

[0090]

[0091] Defining the current coefficient Magnetic flux coefficient Magnetic flux coefficient Magnetic flux coefficient L s =0.2724H is the stator equivalent self-inductance, σ is the leakage inductance coefficient, The power reaching law is applied to the motor observation model, where the sliding mode control law after application is:

[0092]

[0093] Among them, ε1, ε2, ε3, and ε4 are the gains of the current observer α axis, the current observer β axis, the flux observer α axis, and the flux observer β axis respectively. k is the gain of the current and flux observers, all of which are constants greater than 0. Their value ranges are obtained by analyzing the stability of the system.

[0094] Step (4): Speed ​​estimation. The proportional-integral control rate is introduced to improve the speed response time and response speed, so that the final speed estimation algorithm is:

[0095] ω r =(K p +∫K i dt)|ψ sα e isβ -ψ sβ e isα | (8)

[0096] A control system model was built according to the method, and experimental verification was carried out on a hardware-in-the-loop real-time simulation system experimental platform with dSPACE DS1104 as the main controller. The motor rotor speed and stator flux were estimated by a sliding mode observer, and the load was adjusted by a magnetic powder brake.

[0097] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. An asynchronous motor optimization control method based on a multi-scalar loss model method, characterized in that: Combined with the asynchronous motor loss model method under the stator field oriented control method, an induction motor loss model based on the multi-scalar control method is established, and on this basis, a sliding mode observer with improved reaching law is proposed; the operating efficiency of the asynchronous motor under light load conditions is improved, and the stability and robustness of the system are improved; the steps in the use process are: Step (1): construct a mathematical model of induction motor loss based on multi-scalar control; When iron loss is considered, for a circuit consisting of an iron loss branch and an excitation branch, according to Kirchhoff's voltage law, the α-axis component of the iron loss current can be obtained as: Among them, i Feα is the α-axis component of the iron loss current, L m is the mutual inductance of the stator and rotor coaxial equivalent windings, ω e is the synchronous speed of the asynchronous motor, R Fe is the equivalent iron loss resistance, L s is the stator equivalent self-inductance, ψ sβ is the β-axis component of the stator flux; The β-axis component of the iron loss current is: Among them, i Feβ is the β-axis component of the iron loss current, ψ sα is the α-axis component of the rotor flux; Taking the stator current and stator flux as state variables, the state equation of the motor system in the αβ coordinate system can be derived as follows: Among them, i sα is the α-axis component of the stator current, di sα is the change of the α-axis component of the stator current, dt is the change of time, is the derivative of the α-axis component of the stator current, i sβ is the β-axis component of the stator current, di sβ is the change of the β-axis component of the stator current, is the derivative of the β-axis component of the stator current, u sα is the α-axis component of the input stator voltage, u sβ is the β-axis component of the input stator voltage, dψ sα is the change of the α-axis component of the stator current, dψ sβ is the change of the β-axis component of the stator current, is the derivative of the α-axis component of the stator flux, is the derivative of the β-axis component of the stator flux, R r is the rotor equivalent resistance, R s is the stator equivalent resistance, ω r is the motor rotor speed, and defines the current coefficient Magnetic flux coefficient Magnetic flux coefficient Magnetic flux coefficient Defining Voltage Coefficient σ is the leakage inductance coefficient, L r is the rotor equivalent self-inductance; The electromagnetic torque considering iron loss is: Among them, T e is the electromagnetic torque, n p is the pole pair number; Without considering iron consumption, four state variables are defined using multi-scalar control [] T represents the transpose of the row vector, where τ s is the cross product of the stator flux and the stator current vector, is called normalized torque; η s is the dot product of the stator flux and the stator current vector, It is called normalized reactive torque; represents vector cross product, ⊙ represents vector dot product, is the stator flux vector form, is the stator current vector form, is the square of the stator flux modulus; The system control signal is defined as: Among them, u τs is the input signal controlled by the defined electromagnetic torque loop, u ηs The input signal controlled by the defined reactive torque loop is: is the input stator voltage vector form, Considering the iron loss, according to formula (4), the normalized active torque is redefined as: Redefine the normalized reactive torque as: or s =ψ sα (i sα -i Feα )+ψ sβ (i sβ -i Feβ ) (7) Substituting into equations (1) and (2), the reactive torque is converted to: The reactive torque obtained from equation (8) when considering iron loss is consistent with the original definition, because iron loss has nothing to do with reactive power; The state equation after calculation is: Among them, dω r is the change in rotor speed, is the derivative of the rotor speed, is the change of stator flux modulus, is the derivative of the square of the stator flux modulus, dτ s is the change of normalized torque, is the derivative of the normalized torque, dη s is the normalized reactive torque change, is the derivative of the normalized reactive torque, J is the moment of inertia, T L is the load torque; Step (2): The model described in equation (9) in step (1) is a nonlinear system, and linear decoupling is adopted; In order to realize the dual closed-loop control of the mechanical system and the electromagnetic system, a linear decoupling process is performed to obtain a decoupled linear system; the nonlinear feedforward control quantity is defined as follows: Among them, u τs_f is the input signal controlled by the feedforward compensation electromagnetic torque loop, u ηs_f is the input signal controlled by the feedforward compensation reactive torque loop, L σ is the compensation coefficient, and equation (10) is substituted into equation (9). The state equation after linear decoupling is: The first and third equations in equation (11) are for the mechanical subsystem, and the other two equations are for the electromagnetic subsystem. Based on the motor loss model, this control method can divide the motor system into two parts for closed-loop control. Compared with the field-oriented control, the current inner loop is replaced by the dual torque inner loop, which directly controls the torque, which can speed up the system response speed, and the two systems are decoupled from each other and do not affect each other. The dual torque inner loop output is the system control signal defined in step (1), and the control component of the input stator voltage αβ coordinate system is: Step (3): Under different steady-state conditions, the controllable loss of the motor is calculated, that is, the motor loss including the stator and rotor copper loss and the stator iron loss, ignoring the extremely small rotor iron loss, which is briefly referred to as the total loss, accounting for about 80% of the system loss; at the same time, the optimal stator flux that minimizes the total loss is calculated, and it is input as the reference stator flux into the double closed-loop control system described in step (2) to achieve efficiency optimization control; according to the equivalent circuit considering the iron loss, the asynchronous motor power diagram is obtained, and the total motor loss is calculated as: Under different working conditions, in steady-state operation, the motor speed and torque are constant, and the influence of motor parameter changes is ignored. When the derivative of total loss to stator flux is 0, the optimal stator flux that minimizes total loss is obtained as: Under different operating conditions of the system, the reference flux value is adjusted according to the optimal stator flux, thereby minimizing the controllable loss of the motor and improving the motor efficiency; Step (4): Propose a motor control system observer considering iron loss; select stator current and stator flux as state variables, and the mathematical model of induction motor considering iron loss is as shown in formula (3); based on formula (3), the sliding mode observer is designed as: In the formula, is the change in the estimated value of the stator current α axis, is the derivative of the estimated value of the stator current about the α axis, is the estimated value of the stator current α axis, is the change in the estimated value of the stator current β axis, is the derivative of the estimated value of the stator current β axis, is the estimated value of the stator current β axis, is the change in the estimated value of the stator flux α axis, is the derivative of the estimated value of the stator flux α axis, is the estimated value of the stator flux α axis, is the change in the estimated value of the stator flux β axis, is the derivative of the estimated value of the stator flux β axis, is the estimated value of stator flux β axis, is the estimated value of the motor rotor speed, and the sliding mode control term is: Where V α1 is the sliding mode control term of the estimated value of the stator current α axis, V β1 is the sliding mode control term of the stator current β-axis estimate, V α2 is the sliding mode control term of the estimated value of the stator flux α axis, V β2 is the sliding mode control term of the stator flux β-axis estimation value, K1 is the observer gain of the stator current α-axis estimation value, K2 is the observer gain of the stator current β-axis estimation value, K3 is the observer gain of the stator flux α-axis estimation value, K4 is the observer gain of the stator flux β-axis estimation value, K1, K2, K3, K4 are all constants greater than zero; and are the stator current estimation errors of the α-axis and β-axis respectively, and the linear sliding surface is selected as S α =e isα and S β =e isβ , the specific form of the symbolic function sgn() is: Applying the power reaching law to the motor observation model, we get the sliding mode observer equation as shown in equation (15). The sliding mode control law V after application is α1 、V β1 、V α2 、V β2 for: Where ε1, ε2, ε3, ε4 are the gains of the current observer α axis, the current observer β axis, the flux observer α axis, and the flux observer β axis, respectively; k is the gain of the current and flux observers, all of which are constants greater than 0. The range of their values ​​is obtained by analyzing the stability of the system. The design of this control law makes the sliding surface approach zero more quickly, and optimizes the estimated stator current and stator flux. By adjusting the appropriate parameters, the dynamic response and stability of the system are improved, and the torque pulsation is reduced. If the motor parameters are accurate, equation (15) minus equation (3) can obtain the differential equation group of the stator current and stator flux error terms: In the formula, de isα is the change in the stator current α-axis error, is the derivative of the stator current α-axis error, de isβ is the change in the stator current β-axis error, is the derivative of the stator current β-axis error, de ψsα is the change in the stator flux α-axis error, is the derivative of the stator flux α-axis error, de ψsβ is the change in the stator flux β-axis error, is the derivative of the stator flux β-axis error, are the stator flux estimation errors of the α-axis and β-axis respectively, is the rotor speed estimation error; Step (5): Adaptive identification of motor speed; the motor speed is estimated by the stability principle of the Lyapunov function; the following Lyapunov function is selected from equation (19): Where V is the constructed Lyapunov function, is the derivative of the constructed Lyapunov function, e ψs T =[e ψsα e ψsβ ] T , is the derivative of the stator flux error, is the derivative of the motor rotor speed error; to ensure is negative definite and bounded, simplifying the analysis and calculation, so and At the same time; since the motor mechanical time constant is much smaller than the electrical time constant, the motor speed changes much slower than the stator current and stator flux and other electrical quantities, so the speed remains unchanged, and the differential of the speed error term can be obtained, that is, The derivative form of the estimated speed can be solved by combining equations (19) and (20): Where e and γ are constants greater than 0, L T =[sgn(e isα ) sgn(e isβ )], is the error of the estimated stator current, is the estimated value of stator flux; in order to improve the accuracy of the speed dynamic estimation method, the proportional integral control rate is introduced in formula (21) to improve the response speed and response time of the speed, so that the final speed estimation method is: ω r =(K p +∫K i dt)|ψ sα e isβ -ψ sβ e isα | (22) In the formula, K p and K i is the gain of the proportional-integral regulator; the overall structure of the sliding mode observer using the power reaching law is to estimate the stator current and stator flux through equation (19), and finally realize the speed estimation by equation (22).

Images