Control parameter prediction method of proportional-integral disturbance observer and PMSM system

Abstract

The present invention discloses a control parameter prediction method of a proportional integral disturbance observer and a PMSM system, the method comprising: constructing a motor voltage equation containing a permanent magnet synchronous motor system parameter disturbance; constructing a hyperlocal model with a motor parameter disturbance term according to the motor voltage equation; constructing a proportional integral observer model according to the hyperlocal model; and constructing a control parameter prediction model according to the proportional integral observer model using the parameters of the proportional integral observer as variables to predict the control parameters of the proportional integral observer model at the current moment. The control parameter prediction method of the proportional integral disturbance observer and the PMSM system disclosed by the present invention can improve the accuracy of the observer and the robustness of the PMSM system.

Description

Technical Field

[0001] The invention relates to the technical field of permanent magnet synchronous motors, and in particular to a control parameter prediction method of a proportional integral disturbance observer and a PMSM system. Background Art

[0002] Permanent Magnet Synchronous Motor (PMSM) is the most advanced motor technology at present. It has the advantages of high efficiency, low cost, low noise, high power factor, etc., so it is widely used in automobile, aerospace, electronics, industrial control, household appliances and other fields. The traditional control methods of PMSM include vector control and direct torque control. However, the coordinate change of vector control is complex and the adjustment of proportional integral (Proportional Integral) parameters is cumbersome, while the low-speed performance of direct torque control is poor and the real-time requirement is high. Therefore, a model predictive current control (MPCC) with simple structure, multi-objective optimization and fast response speed has attracted extensive attention from researchers.

[0003] In the drive of permanent magnet synchronous motor, the current loop in the control structure plays a very important role, directly affecting the dynamic and steady-state performance of the motor drive system. At present, predictive control is a control method with rapid response, simple concept and easy algorithm implementation, and has gradually become the mainstream method of permanent magnet synchronous motor control. MPCC predicts the current state at the next moment through the mathematical model of the motor, and then outputs the optimal switching state of the control system through the rolling optimization of the value function to ensure the control performance of the system. The MPCC scheme requires a relatively accurate mathematical model to ensure the accuracy of the predicted current. However, the operating environment of permanent magnet synchronous motor is complex. Under the long-term effects of high temperature, load shock and working condition conversion, the permanent magnet synchronous traction system will have electrical parameter perturbations such as resistance, inductance and permanent magnet flux in actual operation, and will be affected by mechanical parameter perturbations and external disturbances; these phenomena will aggravate speed fluctuations, current harmonics, increase torque pulsation, and the fault-tolerant control capability and robustness of the motor drive system are difficult to guarantee, which in turn affects safe and stable operation. Due to the uncertainty of parameter perturbations and external disturbances, the model-based fault-tolerant control method depends on the motor parameters and cannot ensure that the system still has good control performance under parameter perturbations and severe external disturbances. Therefore, in order to ensure the stable operation of permanent magnet synchronous motors, it is necessary to seek new control methods to achieve efficient and reliable operation of motors under parameter perturbations and external disturbances.

[0004] In the prior art, the patent publication number is CN113783484A, which provides a model-free control method for a permanent magnet synchronous motor based on disturbance observation. The method uses a Romberg observer as a disturbance observer of a hyperlocal local model, which increases the amount of calculation and makes the debugging of the two parameters cumbersome, making it impossible to optimize the observation performance of the observer, thereby causing the generation of motor disturbance observation errors and making the motor control inaccurate.

[0005] The patent publication number is CN116846271A, which provides a model-free continuous fast terminal sliding mode fault-tolerant control method and system for permanent magnet synchronous motors. The extended sliding mode disturbance observer is used to estimate the unknown part of the new super-local model. However, although the sliding mode observer has a certain robustness to parameter changes, it still relies on accurate modeling of the system dynamics. If there is a large error in the system model, the performance of the sliding mode observer will be affected, and the high-frequency switching characteristics of the sliding mode observer may make it more sensitive to measurement noise and system noise. These noises may aggravate the chattering phenomenon, and the extended sliding mode does not have adaptive capabilities, affecting the robustness of the observer. Summary of the invention

[0006] The technical problem to be solved by the present invention is to solve the problem that the control of a permanent magnet synchronous motor depends on motor parameters and cannot ensure that the system still has good control performance under parameter perturbations and severe external interferences.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] A control parameter prediction method for a proportional-integral disturbance observer, comprising:

[0009] Construct the motor voltage equation containing the permanent magnet synchronous motor system parameter disturbance;

[0010] According to the motor voltage equation, a hyperlocal model with motor parameter disturbance terms is constructed;

[0011] According to the hyperlocal model, a proportional integral observer model is constructed;

[0012] According to the proportional-integral observer model, the parameters of the proportional-integral observer are used as variables to construct a control parameter prediction model to predict the control parameters of the proportional-integral observer model at the current moment.

[0013] In one embodiment of the present invention, the hyperlocal model is obtained by:

[0014] Convert the motor voltage equation to the motor current equation;

[0015] Establish motor parameter disturbance terms;

[0016] Substitute the motor parameter disturbance term into the motor current equation to obtain a hyperlocal model;

[0017] Among them, the hyperlocal model is obtained by the following formula:

[0018]

[0019] Where, α1 and α2 are the d-axis and q-axis voltage coefficients of the permanent magnet synchronous motor, respectively. u d are the derivative of the current and voltage on the d-axis, u q are the derivative of the q-axis current and voltage, F d 、F q They are respectively the sum of the voltage disturbances caused by the magnetic field and resistance on the d-axis and q-axis of the permanent magnet synchronous motor.

[0020] In one embodiment of the present invention, the proportional integral observer model is obtained by:

[0021] Assume that at time K-1, the initial disturbance variable is F d0 、F q0 , then at time K, the actual disturbance variable F d 、F q for:

[0022]

[0023] in, In the formula, is the estimated value of the q-axis current, is the estimated value of the d-axis current, Δ(Δi q ) is the unknown function of the difference between the estimated value and the actual value of the q-axis current; Δ(Δi d ) is an unknown function of the difference between the estimated value and the actual value of the d-axis current;

[0024] The d-axis and q-axis currents and the voltage disturbances generated by the resistance and magnetic field on the d-axis and q-axis of the permanent magnet synchronous motor are selected as variables to construct a proportional integral observer model, including a d-axis proportional integral observer model and a q-axis proportional integral observer model;

[0025] Among them, the d-axis proportional integral observer model is:

[0026]

[0027] In the formula, is the estimated value of the sum of the voltage disturbances caused by the resistance and magnetic field on the d-axis, is the derivative of the estimated value of the d-axis current, K dp is the proportional gain of the d-axis proportional-integral observer, K di is the integral gain of the d-axis proportional-integral observer, ξ(Δid ) is a constraint Δ(Δi d ), sgn is the sign function, α1 is the d-axis voltage coefficient, u d is the d-axis voltage; where |Δ(Δi d )|≤ξ(Δi d );

[0028] The q-axis proportional integral observer model is:

[0029]

[0030] In the formula, is the estimated value of the sum of the voltage disturbances caused by the resistance and magnetic field on the q-axis, K qp is the proportional gain of the q-axis proportional-integral observer, K qi is the integral gain of the q-axis proportional-integral observer, ξ(Δi q ) is a constraint Δ(Δi q ), is the derivative of the estimated value of the q-axis current, α2 is the q-axis voltage coefficient, and u q is the q-axis voltage; where |Δ(Δi q )|≤ξ(Δi q ).

[0031] In one embodiment of the present invention, a control parameter prediction model is constructed to predict the control parameters of the proportional integral observer model at the current moment, including the following steps:

[0032] The proportional-integral observer model is discretized, and the proportional gain and integral gain of the proportional-integral observer model are set; and the proportional gain and integral gain of the set proportional-integral observer model are brought into the proportional-integral observer model; and the proportional-integral observer model is simplified according to the average value principle;

[0033] The cost function of the simplified proportional integral observer model is set; and the partial derivative of the proportional gain increment and the integral gain increment in the simplified proportional integral observer model is calculated and solved to obtain the proportional gain increment and the integral gain increment;

[0034] A control parameter prediction model is obtained according to the proportional gain increment and the integral gain increment.

[0035] In one embodiment of the present invention, the control parameter prediction model is obtained by the following formula:

[0036]

[0037] Where:

[0038] K qp[k] is the proportional gain of the q-axis proportional-integral observer at the current moment, K qp [k-1] is the proportional gain of the q-axis proportional-integral observer at time k-1, ΔK qp [k] is the proportional gain increment of the q-axis proportional-integral observer at the current moment, K qi [k] is the integral gain of the q-axis proportional integral observer at the current moment, K qi [k-1] is the integral gain of the q-axis proportional integral observer at time k-1, ΔK qi [k] is the integral gain increment of the q-axis proportional-integral observer at the current moment;

[0039] K dp [k] is the proportional gain of the d-axis proportional-integral observer at the current moment, K dp [k-1] is the proportional gain of the d-axis proportional-integral observer at time k-1, ΔK dp [k] is the proportional gain increment of the d-axis proportional-integral observer at the current moment, K di [k] is the integral gain of the d-axis proportional integral observer at the current moment, K di [k-1] is the integral gain of the d-axis proportional-integral observer at time k-1, ΔK di [k] is the integral gain increment of the d-axis proportional-integral observer at the current moment.

[0040] In one embodiment of the present invention, the proportional gain and integral gain of the proportional-integral observer model at the current moment are substituted into the proportional-integral observer model, and the estimated current and estimated disturbance at the next moment are observed and output, and the reference voltage input to the proportional-integral disturbance observer is obtained in combination with the hyperlocal model.

[0041] In one embodiment of the present invention, the reference voltage input to the proportional-integral disturbance observer is obtained by the following formula:

[0042]

[0043] In the formula, u dref is the d-axis reference voltage, i dref [k+2] is the d-axis reference current at time k+2, is the estimated value of the d-axis current at time k+1, is the estimated value of the sum of the voltage disturbances caused by the resistance and magnetic field on the d-axis at time k+1, α1 is the d-axis voltage coefficient, and T is the sampling period;

[0044]

[0045] In the formula, u qref is the q-axis reference voltage, i qref [k+2] is the q-axis reference current at time k+2, is the estimated value of the q-axis current at time k+1, is the estimated value of the sum of the voltage disturbances caused by the resistance and magnetic field on the q-axis at time k+1, and α2 is the q-axis voltage coefficient.

[0046] In one embodiment of the present invention, a proportional-integral observer model used to construct a control parameter prediction model is proven to be stable; wherein, the stability of the proportional-integral observer model is proven by sampling a Lyapunov function.

[0047] The present invention also provides a PMSM system, including a proportional-integral disturbance observer 10 and a model-free predictive current controller 20, wherein the proportional-integral disturbance observer 10 executes the control parameter prediction method of the proportional-integral disturbance observer described above; the output voltage of the model-free predictive current controller 20 serves as an input reference voltage of the proportional-integral disturbance observer 10.

[0048] In one embodiment of the present invention, the PMSM system further includes a maximum torque-to-current ratio table module 00, an IPark module 30, an SVPWM module 40, an Inverter module 50, a Clarke module 60, a Park module 70, an angle detection module 82, and a speed calculation module 83;

[0049] Among them, the maximum torque current ratio table module 00, the model-free predictive current controller 20, the IPark module 30, the SVPWM module 40, the Inverter module 50, the Clarke module 60, the Park module 70 and the proportional-integral disturbance observer 10 are connected in sequence; the output end of the model-free predictive current controller 20 is connected to the input end of the proportional-integral disturbance observer 10; the IPark module 30 is connected to the Park module 70;

[0050] The inverter module 50 is connected to the permanent magnet synchronous motor; the angle detection module 82 is connected to the permanent magnet synchronous motor, and the speed calculation module 83 is connected to the angle detection module 82, and the output end is connected to the input end of the maximum torque current ratio table module 00;

[0051] With a given torque target demand value T e and the current motor speed ω e As input, the current reference value i of the current d-axis and q-axis is obtained from the maximum torque current ratio table module 00 through two-dimensional interpolation method. dref 、i qref , substituted into the model-free predictive current controller 20, and the output d-axis and q-axis voltage reference values ​​u dref 、u qref The reference voltage u of the α-axis and β-axis of the permanent magnet synchronous motor is obtained by performing inverse Park transformation through IPark module 30 α 、uβ , the three-phase duty ratios pwma, pwmb, and pwmc calculated by the SVPWM module 40 through the spatial pulse width vector modulation algorithm are used to control the three-phase full-bridge inverter circuit in the Inverter module 50 to output the motor three-phase voltages U, V, and W to drive the permanent magnet synchronous motor;

[0052] The motor U and W two-phase feedback current i collected by the current sensor a 、i c The feedback current i of the α-axis and the β-axis is calculated by Clarke transformation using Clarke module 60. α 、i β , the d-axis and q-axis current feedback values ​​i are calculated by Park module 70 through Park transformation dfdb 、i qfdb , and as the reference current of the proportional-integral disturbance observer 10, and, with the d-axis and q-axis voltage reference values ​​u dref 、u qref As the input of the proportional-integral disturbance observer 10, the estimated current values ​​of the d and q axes and the estimated voltage disturbances generated by the magnetic field on the d and q axes are calculated by the proportional-integral disturbance observer 10. As an input to the model-free predictive current controller 20;

[0053] The angle detection module 82 detects the calculated motor electrical angle θ through a position sensor and uses it as the input for the transformation between the IPark module 30 and the Park module 70; and uses the motor electrical angle θ as the input of the speed calculation module 83 to obtain the real-time speed of the motor and feed it back to the maximum torque current table module 00.

[0054] Compared with the prior art, the beneficial effects of the present invention are: establishing a super-local model with motor parameter disturbance terms, using a proportional-integral observer model, which is independent of any motor parameters, and overcoming the shortcomings of the prior art that is susceptible to parameter drift and model mismatch caused by factors such as temperature, magnetic field saturation, and operating status, and then using a control parameter prediction model to predict the control parameters of the proportional-integral observer model, thereby improving the robustness of the observer and making it insensitive to the accuracy of the model and measurement noise. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 The present invention is a flowchart of a control parameter prediction method of a proportional-integral disturbance observer according to an embodiment of the present invention.

[0056] Figure 2 Schematic diagram of a PMSM system according to an embodiment of the present invention.

[0057] Figure 3 This is a control flow chart of a model-free predictive current controller based on a super-local model of the present invention. DETAILED DESCRIPTION

[0058] In order to facilitate those skilled in the art to understand the technical solution of the present invention, the technical solution of the present invention is further described in conjunction with the accompanying drawings of the specification.

[0059] The terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0060] Example 1

[0061] See also Figure 1 As shown, the present invention provides a control parameter prediction method of a proportional-integral disturbance observer, comprising:

[0062] S10, constructing the motor voltage equation containing the permanent magnet synchronous motor system parameter disturbance.

[0063] In this embodiment, the motor voltage equation is obtained by the following formula:

[0064]

[0065] Where, L d , L q are the d-axis and q-axis inductance parameters of the permanent magnet synchronous motor, i d 、i q 、u d 、u q are the d-axis and q-axis current and voltage respectively, R s is the stator resistance, ω e is the motor speed, For magnetic link.

[0066] S20, based on the motor voltage equation, a hyperlocal model with motor parameter disturbance terms is constructed.

[0067] In this embodiment, the motor voltage equation of formula (1.1) is converted into a motor current equation:

[0068]

[0069] Since the permanent magnet synchronous motor is a nonlinear system, the motor parameters will inevitably change and cause disturbances during the operation of the motor. In order to improve the control robustness of the motor under various working conditions, it is necessary to consider the parameter disturbance term and establish a super-local model of the motor as follows:

[0070]

[0071] Where, α1 and α2 are the d-axis and q-axis voltage coefficients of the permanent magnet synchronous motor, respectively. u d are the derivative of the current and voltage on the d-axis, u q The derivative of the q-axis current and voltage, F d 、F q They are respectively the sum of the voltage disturbances caused by the magnetic field and resistance on the d-axis and q-axis of the permanent magnet synchronous motor.

[0072] S30, constructing a proportional integral observer model according to the hyperlocal model.

[0073] In one embodiment of the present invention, the motor disturbance F is estimated according to the super-local model established by the above formula: d 、F q , assuming that at time K-1, the initial disturbance variable is F d0 、F q0 , and secondly, at time K, the actual disturbance variable is F d 、F q It can be expressed by the following formula:

[0074]

[0075] in, In the formula, is the estimated value of the q-axis current, is the estimated value of the d-axis current, Δ(Δi q ) is the unknown function of the difference between the estimated value and the actual value of the q-axis current; Δ(Δi d ) is the unknown function of the difference between the estimated value and the actual value of the d-axis current. And Δ(Δi q )、Δ(Δi d ) are unknown functions, which can be respectively expressed as known functions ξ(Δi q ),ξ(Δi d ) is constrained, where |Δ(Δi q )|≤ξ(Δi q )、|Δ(Δi d )|≤ξ(Δi d ).

[0076] The proportional integral observer with good performance in parameter disturbance observation is introduced to observe the disturbance of the system, and the dq axis current and system disturbance are selected as the system state variables to construct the state equation. Among them, the proportional integral observer model includes the d-axis proportional integral observer model and the q-axis proportional integral observer model.

[0077] The d-axis proportional integral observer model is:

[0078]

[0079] In the formula, is the estimated value of the sum of the voltage disturbances caused by the resistance and magnetic field on the d-axis, is the derivative of the estimated value of the d-axis current, K dp is the proportional gain of the d-axis proportional-integral observer, K di is the integral gain of the d-axis proportional-integral observer, ξ(Δi d ) is a constraint Δ(Δi d ), sgn is the sign function, α1 is the d-axis voltage coefficient, u d is the d-axis voltage.

[0080] The q-axis proportional integral observer model is:

[0081]

[0082] In the formula, is the estimated value of the sum of the voltage disturbances caused by the resistance and magnetic field on the q-axis, K qp is the proportional gain of the q-axis proportional-integral observer, K qi is the integral gain of the q-axis proportional-integral observer, ξ(Δi q ) is a constraint Δ(Δi q ), is the derivative of the estimated value of the q-axis current, α2 is the q-axis voltage coefficient, and u q is the q-axis voltage.

[0083] In this embodiment, the principle of the designed proportional integral observer model is to eliminate the deviation between the predicted current and the measured current to obtain an accurate disturbance state value. In other words, when the predicted current roughly matches the measured current, this indicates that the output of the proportional integral observer model is close to the true value of the motor dq axis disturbance.

[0084] In this embodiment, a proportional integral observer model is established in step S30 to estimate the motor disturbance. Before step S40, stability verification can be performed. Specifically, the stability of the above proportional integral observer model is proved by using the Lyapunov function. Take the stability proof of the q-axis proportional integral observer model as an example:

[0085] Let σ = K qi ∫Δi q dt, according to equations (1.4), (1.5), (1.7), the error equation is established:

[0086]

[0087] In the formula, is the derivative of σ.

[0088] In the above formula, let Then the above formula (1.8) can be rewritten as:

[0089]

[0090] The Lyapunov function is established as follows:

[0091]

[0092] Where V is the Lyapunov function, and s is a selected function expression with no specific meaning.

[0093] When K qi ≥0, V≥0 always holds true;

[0094]

[0095] In the formula, is the derivative of V, for The derivative of .

[0096] Among them, it is known that |Δ(Δi q )|≤ξ(Δi q ), from which we can get (Δ(Δi q )-ξ(Δi q )sgn(Δi q ))Δi q ≤0 is always true, then when there exists K qp ≥0, the following formula is always true:

[0097]

[0098] From the above, it can be concluded that the established proportional integral observer model satisfies Lyapunov stability. In addition, it can be concluded from the above formula:

[0099]

[0100] Where Δ=Δ(Δi q ), based on Lasalle invariant set theorem, we can get:

[0101] When time t→∞, Therefore, we can combine equations (1.5) and (1.7) to obtain, Δi q =0, F q0 +Δ=σ, so the disturbance estimated by the proportional integral observer model is:

[0102]

[0103] The above formula shows that in the invariant set, when t→∞,

[0104] Similarly, the stability of the d-axis proportional-integral observer model can also be derived.

[0105] S40, according to the proportional-integral observer model, the parameters of the proportional-integral observer are used as variables to construct a control parameter prediction model to predict the control parameters of the proportional-integral observer model at the current moment.

[0106] In this embodiment, the proportional integral observer model includes a d-axis proportional integral observer model and a q-axis proportional integral observer model, and the corresponding control parameter prediction model also includes a d-axis control parameter prediction model and a q-axis control parameter prediction model. The control parameter prediction model can be understood as a model for obtaining the optimal control parameters of the proportional integral observer model.

[0107] The control parameter prediction model is designed as a two-step model to predict the control parameters of the proportional integral disturbance observer. Since the motor disturbance will change due to the influence of temperature and environment, if the parameters of the proportional integral observer model are fixed, the system will lack adaptability. Therefore, we use the proportional integral parameters as variables and use model predictive control to calculate them. The proportional integral observer model includes the d-axis proportional integral observer model and the q-axis proportional integral observer model, so the control parameters of the d-axis proportional integral observer model and the control parameters of the q-axis proportional integral observer model must be obtained separately.

[0108] In this embodiment, the q-axis proportional integral observer model is discretized:

[0109]

[0110] In the formula, K qp [k] is the proportional gain of the q-axis proportional-integral observer at the current moment, K qi [k] is the integral gain of the q-axis proportional-integral observer at the current moment, is the estimated value of the q-axis current at time k+1, is the estimated value of the q-axis current at time k, T is the sampling period, u q [k] is the q-axis voltage at time k.

[0111] Substituting the perturbation into:

[0112]

[0113] Let K qp [k] = K qp [k-1]+ΔK qp [k], K qi [k] = Kqi [k-1]+ΔK qi [k], we can get:

[0114]

[0115] In the formula, ΔK qp [k] is the proportional gain increment of the q-axis proportional-integral observer at the current moment, ΔK qi [k] is the integral gain increment of the q-axis proportional-integral observer at the current moment.

[0116] Among them, the definition is:

[0117] H=(K qp [k-1]+ΔK qp [k])Δiq+(K qi [k-1]+ΔK qi [k])∫Δiqdt+ξ(Δi q )sgn(Δi q )

[0118] Simplified to:

[0119]

[0120] According to the average value principle, ΔK qp [k+1]=ΔK qp [k], that is:

[0121]

[0122] The two-step state prediction method is used to calculate the control parameters of the q-axis proportional integral observer model, and the cost function of the design model prediction is:

[0123]

[0124] In the formula, is the cost function of the q-axis proportional-integral observer model, δ p is the weight coefficient of the q-axis current, p is the accumulated parameter variable, β1 is the weight coefficient of the proportional gain of the q-axis proportional-integral observer, and β2 is the weight coefficient of the integral gain of the q-axis proportional-integral observer.

[0125] ΔK qp [k], ΔK qi [k] Find the partial derivative, which can be simplified to:

[0126]

[0127]

[0128] In the formula, is the cost function of the q-axis proportional-integral observer model for ΔK qp [k] Find the partial derivative, δ1 is the weight coefficient of a1, and δ2 is the weight coefficient of a2.

[0129] By solving the above two equations, we can obtain the proportional gain increment and integral gain increment of the q-axis proportional-integral observer model, namely:

[0130]

[0131] Simplifying formula (1.26), we get:

[0132]

[0133] Substitute equation (1.27) into equation K qp [k] = K qp [k-1]+ΔK qp [k], K qi [k] = K qi [k-1]+ΔK qi [k], the q-axis control parameter prediction model can be obtained, and the optimal control parameters of the q-axis proportional integral observer model can be obtained:

[0134]

[0135] Discretize the d-axis proportional integral observer model:

[0136]

[0137] Substituting the perturbation into:

[0138]

[0139] Let K dp [k] = K dp [k-1]+ΔK dp [k], K di [k] = K di [k-1]+ΔK di [k], we can get:

[0140]

[0141] in:

[0142] H1=((K dp [k-1]+ΔK dp [k])Δi d +(K di [k-1]+ΔK di [k])∫Δi d dt+ξ(Δid )sgn(Δi d ))

[0143] Simplified to:

[0144]

[0145] According to the average value principle, ΔK dp [k+1]=ΔK dp [k], that is:

[0146]

[0147] Simplified to:

[0148]

[0149] The two-step state prediction method is used to calculate the control parameters of the d-axis proportional integral observer model, and the cost function of the design model prediction is:

[0150]

[0151] In the formula, is the cost function of the d-axis proportional-integral observer model, γ p is the weight coefficient of the d-axis current, λ1 is the weight coefficient of the proportional gain of the d-axis proportional-integral observer, λ2 is the weight coefficient of the integral gain of the d-axis proportional-integral observer, ΔK dp [k] is the proportional gain of the d-axis proportional-integral observer at the current moment, ΔK di [k] is the integral gain of the d-axis proportional-integral observer at the current moment.

[0152] ΔK dp [k], ΔK di [k] Find the partial derivative, which can be simplified to:

[0153]

[0154] In the formula, is the cost function of the d-axis proportional-integral observer model for ΔK dp [k] Find the partial derivative, γ1 is the weight coefficient of b1, and γ2 is the weight coefficient of b2.

[0155] Solving the above two equations, we can get the proportional gain increment and integral gain increment of the d-axis proportional-integral observer model, namely:

[0156]

[0157] Substitute equation (1.41) into K dp [k] = K dp [k-1]+ΔKdp [k], K di [k] = K di [k-1]+ΔK di [k], the d-axis control parameter prediction model can be obtained, and the optimal control parameters of the d-axis proportional-integral observer model can be obtained:

[0158]

[0159] See also Figure 1 As shown, in one embodiment of the present invention, the proportional gain and integral gain of the proportional-integral observer model at the current moment are substituted into the proportional-integral observer model, and the estimated current and estimated disturbance at the next moment are observed and output, and the reference voltage input to the proportional-integral disturbance observer is obtained in combination with the hyperlocal model.

[0160] After the proportional integral observer model is discretized, considering that the sampling period T is small enough, it can be considered that Therefore, the d-axis and q-axis reference voltages can be calculated according to the hyperlocal model formula:

[0161] D-axis reference voltage:

[0162]

[0163] Q-axis reference voltage:

[0164]

[0165] In the formula, u dref is the d-axis reference voltage, i dref [k+2] is the d-axis reference current at time k+2, is the estimated value of the d-axis current at time k+1, is the estimated value of the sum of the voltage disturbances caused by the resistance and magnetic field on the d-axis at time k+1, α1 is the d-axis voltage coefficient, and T is the sampling period. qref is the q-axis reference voltage, i qref [k+2] is the q-axis reference current at time k+2, is the estimated value of the q-axis current at time k+1, is the estimated value of the sum of the voltage disturbances caused by the resistance and magnetic field on the q-axis at time k+1, and α2 is the q-axis voltage coefficient.

[0166] Example 2

[0167] See also Figures 1 to 3As shown, the present invention provides a PMSM system, including a proportional-integral disturbance observer 10 and a model-free predictive current controller 20. The proportional-integral disturbance observer 10 executes the control parameter prediction method of the proportional-integral disturbance observer described in Example 1. The output voltage of the model-free predictive current controller 20 is used as an input reference voltage of the proportional-integral disturbance observer 10.

[0168] In one embodiment of the present invention, the PMSM system further includes a maximum torque-to-current ratio table module 00 , an IPark module 30 , an SVPWM module 40 , an Inverter module 50 , a Clarke module 60 , a Park module 70 , an angle detection module 81 and a speed calculation module 82 .

[0169] Among them, the maximum torque current ratio table module 00, the model-free predictive current controller 20, the IPark module 30, the SVPWM module 40, the Inverter module 50, the Clarke module 60, the Park module 70 and the proportional-integral disturbance observer 10 are connected in sequence. The output end of the model-free predictive current controller 20 is connected to the input end of the proportional-integral disturbance observer 10, and the IPark module 30 is connected to the Park module 70.

[0170] After the inverter module 50 is connected to the permanent magnet synchronous motor, the angle detection module 81 is connected to the permanent magnet synchronous motor, and the speed calculation module 82 is connected to the angle detection module 81 , the output end is connected to the input end of the maximum torque current ratio table module 00 .

[0171] In this embodiment, given the torque target demand value T e and the current motor speed ω e As input, the current reference value i of the current d-axis and q-axis is obtained from the maximum torque current ratio table module 00 through two-dimensional interpolation method. dref 、i qref , substituted into the model-free predictive current controller 20, and the output d-axis and q-axis voltage reference values ​​u dref 、u qref The reference voltage u of the α-axis and β-axis of the permanent magnet synchronous motor is obtained by performing inverse Park transformation through IPark module 30 α 、u β The three-phase duty ratios pwma, pwmb, and pwmc calculated by the SVPWM module 40 through the spatial pulse width vector modulation algorithm are used to control the three-phase full-bridge inverter circuit in the Inverter module 50 to output the motor three-phase voltages U, V, and W to drive the permanent magnet synchronous motor.

[0172] The motor U and W two-phase feedback current i collected by the current sensor a 、i cThe feedback current i of the α-axis and the β-axis is calculated by Clarke transformation using Clarke module 60. α 、i β , the d-axis and q-axis current feedback values ​​i are calculated by Park module 70 through Park transformation dfdb 、i qfdb , and as the reference current of the proportional-integral disturbance observer 10, and, with the d-axis and q-axis voltage reference values ​​u dref 、u qref As the input of the proportional-integral disturbance observer 10, the estimated current values ​​of the d and q axes and the estimated voltage disturbances generated by the magnetic field on the d and q axes are calculated by the proportional-integral disturbance observer 10. As the input of the model-free predictive current controller 20.

[0173] The angle detection module 81 detects the calculated motor electrical angle θ through the position sensor and uses it as the input for the IPark module 30 and the Park module 70. The motor electrical angle θ is used as the input of the speed calculation module 82 to obtain the real-time speed of the motor and feed it back to the maximum torque current table module 00.

[0174] It is obvious to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential features of the present invention. Therefore, the embodiments should be regarded as exemplary and non-limiting from any point of view, and the scope of the present invention is defined by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention, and any reference numerals in the claims should not be regarded as limiting the claims involved.

[0175] The above-described embodiments merely represent implementation methods of the invention. The protection scope of the present invention is not limited to the above-described embodiments. For those skilled in the art, several modifications and improvements may be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A control parameter prediction method for a proportional-integral disturbance observer, characterized in that: include: Construct the motor voltage equations for systems containing permanent magnet synchronous motors; According to the motor voltage equation, a hyperlocal model with motor parameter disturbance terms is constructed; According to the hyperlocal model, a proportional integral observer model is constructed; According to the proportional-integral observer model, the parameters of the proportional-integral observer are used as variables to construct a control parameter prediction model to predict the control parameters of the proportional-integral observer model at the current moment; Among them, the hyperlocal model is obtained by: Convert the motor voltage equation to the motor current equation; Establish motor parameter disturbance terms; Substitute the motor parameter disturbance term into the motor current equation to obtain a hyperlocal model; The proportional-integral observer model is obtained by: Assume that at time K-1, the initial disturbance variable is F d0 、F q0 , then at time K, the actual disturbance variable F d 、F q for: in, In the formula, are the estimated values ​​of the q-axis and d-axis currents, Δ(Δi q )、Δ(Δi d ) are the unknown functions of the difference between the estimated current value and the actual current value of the q-axis and d-axis respectively; The d-axis and q-axis currents and the voltage disturbances generated by the resistance and magnetic field on the d-axis and q-axis of the permanent magnet synchronous motor are selected as variables to construct a proportional integral observer model, including a d-axis proportional integral observer model and a q-axis proportional integral observer model; The d-axis and q-axis proportional integral observer models are: In the formula, are the estimated values ​​of the sum of the voltage disturbances caused by the resistance and magnetic field on the d-axis and q-axis, respectively. are the derivatives of the estimated values ​​of the d-axis and q-axis currents, respectively, K dp Kdi and Kqi are the integral gains of the d-axis and q-axis proportional-integral observers, ξ(Δid) and ξ(Δi q ) are constraints Δ(Δi d )、Δ(Δi q ), sgn is the sign function, α1 and α2 are the d-axis and q-axis voltage coefficients, ud and uq are the d-axis and q-axis voltages, respectively; where |Δ(Δi d )|≤ξ(Δid);|Δ(Δi q )|≤ξ(Δi q ).

2. The control parameter prediction method of the proportional-integral disturbance observer according to claim 1, characterized in that: Constructing a control parameter prediction model to predict the control parameters of the proportional integral observer model at the current moment includes the following steps: The proportional-integral observer model is discretized, and the proportional gain and integral gain of the proportional-integral observer model are set; and the proportional gain and integral gain of the set proportional-integral observer model are brought into the proportional-integral observer model; and the proportional-integral observer model is simplified according to the average value principle; The cost function of the simplified proportional integral observer model is set; and the partial derivative of the proportional gain increment and the integral gain increment in the simplified proportional integral observer model is calculated and solved to obtain the proportional gain increment and the integral gain increment; A control parameter prediction model is obtained according to the proportional gain increment and the integral gain increment.

3. The control parameter prediction method of the proportional-integral disturbance observer according to claim 2, characterized in that: The control parameter prediction model is obtained by the following formula: Where: K qp [k] is the proportional gain of the q-axis proportional-integral observer at the current moment, K qp [k-1] is the proportional gain of the q-axis proportional-integral observer at time k-1, ΔK qp [k] is the proportional gain increment of the q-axis proportional-integral observer at the current moment, K qi [k] is the integral gain of the q-axis proportional integral observer at the current moment, K qi [k-1] is the integral gain of the q-axis proportional integral observer at time k-1, ΔK qi [k] is the integral gain increment of the q-axis proportional-integral observer at the current moment; K dp [k] is the proportional gain of the d-axis proportional-integral observer at the current moment, K dp [k-1] is the proportional gain of the d-axis proportional-integral observer at time k-1, ΔK dp [k] is the proportional gain increment of the d-axis proportional-integral observer at the current moment, K di [k] is the integral gain of the d-axis proportional integral observer at the current moment, K di [k-1] is the integral gain of the d-axis proportional-integral observer at time k-1, ΔK di [k] is the integral gain increment of the d-axis proportional-integral observer at the current moment.

4. The control parameter prediction method of the proportional-integral disturbance observer according to claim 2, characterized in that: The proportional gain and integral gain of the proportional-integral observer model at the current moment are substituted into the proportional-integral observer model, and the estimated current and estimated disturbance output at the next moment are observed and output. Combined with the hyperlocal model, the reference voltage input to the proportional-integral disturbance observer is obtained.

5. The control parameter prediction method of the proportional-integral disturbance observer according to claim 1, characterized in that: The proportional-integral observer model used to construct the control parameter prediction model has been proven to be stable; among them, the sampling Lyapunov function proves the stability of the proportional-integral observer model.

6. A PMSM system, characterized in that: The invention comprises a proportional-integral disturbance observer (10) and a model-free predictive current controller (20), wherein the proportional-integral disturbance observer (10) executes the control parameter prediction method of the proportional-integral disturbance observer described in any one of claims 1 to 5; and the output voltage of the model-free predictive current controller (20) serves as the input reference voltage of the proportional-integral disturbance observer (10).

7. The PMSM system according to claim 6, characterized in that: The PMSM system also includes a maximum torque current ratio table module (00), an IPark module (30), an SVPWM module (40), an Inverter module (50), a Clarke module (60), and a Park module (70); Wherein, the maximum torque current ratio table module (00), the model-free predictive current controller (20), the IPark module (30), the SVPWM module (40), the Inverter module (50), the Clarke module (60), the Park module (70) and the proportional integral disturbance observer (10) are connected in sequence; the output end of the model-free predictive current controller (20) is connected to the input end of the proportional integral disturbance observer (10); the IPark module (30) and the Park module (70) are connected; and the Inverter module (50) is connected to the permanent magnet synchronous motor; With a given torque target demand value T e With the current motor speed ω e As input, the current reference value i of the current d-axis and q-axis is obtained from the maximum torque current ratio table module (00) through two-dimensional interpolation method. dref 、i qref , substituted into the model-free predictive current controller (20), the output d-axis and q-axis voltage reference values ​​ud ref 、u qref , the reference voltage u of the α-axis and β-axis of the permanent magnet synchronous motor obtained by inverse Park transformation through IPark module (30) α 、u β , the three-phase duty ratios pwma, pwmb, and pwmc calculated by the spatial pulse width vector modulation algorithm of the SVPWM module (40) are used to control the three-phase full-bridge inverter circuit in the Inverter module (50) to output the three-phase voltages U, V, and W of the motor to drive the permanent magnet synchronous motor; The motor U and W two-phase feedback current i collected by the current sensor a 、i c The feedback current i of the α-axis and β-axis is calculated by Clarke transformation through Clarke module (60) α 、i β , the d-axis and q-axis current feedback values ​​id calculated by Park module (70) through Park transformation f db、i qf db, and as the reference current of the proportional-integral disturbance observer (10), and, with the d-axis and q-axis voltage reference values ​​u dref 、u qref As the input of the proportional-integral disturbance observer (10), the estimated current values ​​of the d and q axes and the estimated voltage disturbances generated by the magnetic field on the d and q axes are calculated by the proportional-integral disturbance observer (10). As the input of the model-free predictive current controller (20).

8. The PMSM system according to claim 7, characterized in that: The PMSM system further comprises an angle detection module (81) and a speed calculation module (82); after the angle detection module (81) is connected to the permanent magnet synchronous motor, and the speed calculation module (82) is connected to the angle detection module (81), an output end is connected to an input end of a maximum torque current ratio table module (00); The angle detection module (81) detects the calculated motor electrical angle θ through a position sensor and uses it as an input for transformation between the IPark module (30) and the Park module (70); The motor electrical angle θ is used as an input of a speed calculation module (82) to obtain the real-time speed of the motor and feed it back to the maximum torque current table module (00).

Images