Magnetic suspension bearing PI current controller parameter setting method for differential current control

Abstract

The invention discloses a magnetic suspension bearing PI current controller parameter setting method for differential current control, and belongs to the technical field of electrical control. According to the method, the bandwidth of the differential current of the electromagnetic coil is taken as an optimization target, and the parameter setting of the PI current controller is realized by adopting a simulated annealing algorithm. Specifically, through continuous search optimization of parameters of a magnetic suspension bearing PI current controller in a parameter space, the differential current of an electromagnetic coil reaches a preset bandwidth and a dynamic response index. According to the method, the parameters of the magnetic suspension bearing PI current controller adopting the differential control strategy are set based on the simulated annealing algorithm, and the method has the advantages of being visual in evaluation index, high in search speed, good in target performance matching degree and the like.

Description

Technical Field

[0001] The invention belongs to the technical field of electrical control, and more specifically, relates to a parameter setting method of a magnetic suspension bearing PI current controller for differential current control. Background Art

[0002] Compared with traditional mechanical bearings, electromagnetic bearings (also known as active magnetic bearings) use the electromagnetic force generated by the stator to suspend the motor rotor and realize non-contact operation between the stator and the rotor. The controller is the core part of the magnetic bearing. Its main function is to receive signals such as displacement and current feedback from the sensor, and give reference current instructions to control the stable suspension of the magnetic bearing. Due to the negative stiffness characteristics of the magnetic bearing, the rotor system supported by the magnetic bearing is open-loop unstable. Only through the closed-loop control of the controller can the rotor be stably suspended. The current controller is mainly a digital controller with a microcontroller as the core. Different control algorithms can be implemented inside the controller to control the working characteristics of the magnetic bearing under different working conditions.

[0003] However, the nonlinear, high-order, and time-varying characteristics of the magnetic suspension system make it difficult to accurately model the dynamic behavior of the system. The sensitivity and coupling effect of the system, as well as the influence of external environmental disturbances and uncertainties, increase the complexity of the magnetic suspension bearing controller parameter setting. In recent years, the emergence and rise of intelligent optimization algorithms have provided new ideas for this problem. Summary of the invention

[0004] In view of the defects of the prior art, the purpose of the present invention is to provide a method for adjusting the parameters of a PI current controller of a magnetic bearing for differential current control, aiming to solve the problems of poor differential current control performance of the existing magnetic bearing and complex process of the PI current controller parameter adjustment method.

[0005] To achieve the above objectives, the present invention applies a simulated annealing algorithm (SA) to an active magnetic bearing system that adopts differential control and a PI current controller, and optimizes the PI current controller parameters of the magnetic bearing controller, which is of great significance for obtaining the corresponding optimal control parameters.

[0006] The present invention first determines that the control strategy of the magnetic suspension bearing is differential current control and current loop closed-loop control, determines that the optimization target of the algorithm is the differential current bandwidth of the electromagnetic coil of the magnetic suspension bearing, and determines that the optimization object of the algorithm is the PI current controller parameter of the magnetic suspension bearing; in the current loop control of the magnetic suspension bearing controller, the differential current is defined as the current difference between two ports of a group of electromagnetic coils of the magnetic suspension bearing, and the differential current bandwidth is defined as the frequency range of the -3dB point where the amplitude drops to the maximum value in the frequency response curve of the differential current.

[0007] The method for setting parameters of a magnetic bearing PI current controller for differential current control provided by the present invention comprises the following steps:

[0008] At the current moment, the position signal of the magnetic bearing is collected, the position signal is converted into a voltage signal, the target current is calculated according to the voltage signal, the target current is subtracted from the current electromagnetic coil current, and the error current obtained is input into the PI current controller. The current signal output by the PI current controller is then input into the power electronic converter, which is converted into the electromagnetic coil current at the next moment and outputs the power supply to realize the initial suspension of the magnetic bearing;

[0009] Taking the differential current bandwidth of the electromagnetic coil as the optimization target, the parameters of the PI current controller are optimized and the simulated annealing algorithm is used to realize the tuning of the PI current controller parameters.

[0010] Before the simulated annealing algorithm runs, it is necessary to give the initial state, state generation mechanism, conditions for accepting new solutions, cooling strategy, and termination conditions. The initial state includes the initial solution (Initial Solution) and the initial temperature (Initial Temperature, T0): the initial solution is a set of PI current controller parameters obtained based on previous experience or randomly generated PI current controller parameters; the initial temperature determines the degree of exploration at the beginning of the algorithm. The higher the temperature, the easier it is for the algorithm to accept solutions of poor quality to explore the global solution space. The state generation mechanism is the neighborhood generation function, which generates the current candidate solution by applying random small perturbations to the current solution. The specific formula is explained in the next paragraph. The condition for accepting a new solution is: when the new solution is better than the current solution, it is always accepted; when the new solution is worse than the current solution, it is accepted with a certain probability according to the Metropolis criterion to avoid falling into the local optimum. The specific formula is explained in the next paragraph. The cooling strategy is to reduce the temperature at a certain rate after each iteration. The termination conditions include the minimum temperature and the maximum number of iterations. The minimum temperature condition means that the iteration is stopped when the temperature drops below a certain threshold, and the maximum number of iterations condition means that the iteration is stopped after the preset number of iterations is reached.

[0011] The domain generation function of the current controller P and I parameters is:

[0012] Pa new =Pa current +(rand-0.5)×(K p ×T÷T initial )

[0013] Ia new =Ia current +(rand-0.5)×(K i ×T÷Tinitial )

[0014] The conditions for accepting the new solution are:

[0015] ΔΕ<0or rand <e -ΔΕ / T

[0016] Where T is the current temperature, K p is the proportional coefficient of random fluctuation of Pa parameter, K i is the proportional coefficient of the random fluctuation of the Ia parameter. The proportional coefficient is modified and determined according to the actual working conditions. ΔΕ=Value new -Value current is the difference in objective function value, rand <e -ΔΕ / T is the Metropolis criterion, and rand is a generated uniformly distributed random number rand~U(0,1). The Metropolis criterion is a rule used to decide whether to accept the transition from the current state x to the candidate state x′, so that in the search process, new areas can be explored (including temporarily accepting worse solutions to escape the local optimum) and gradually converge to the global optimal solution. The probability given by the Metropolis criterion is P(x→x′)=e -ΔΕ / T .

[0017] Furthermore, the use of simulated annealing algorithm to achieve the tuning of PI current controller parameters includes: determining initial values ​​of the PI current controller parameters and a domain generating function, iteratively optimizing the PI current controller parameters according to the simulated annealing algorithm, and obtaining a set of PI current controller parameters that meet the optimization target.

[0018] Furthermore, the differential current bandwidth value corresponding to the PI current controller parameters (x_p, x_i) in the current state is f(x_p, x_i), and the differential current bandwidth value corresponding to the newly generated PI current controller parameters (x_p', x_i') ​​in each iteration is f(x_p', x_i').

[0019] When f(x_p',x_i') ​​is not less than the preset target value, the iteration ends;

[0020] When f(x_p',x_i') ​​is less than the preset target value, calculate Δf = f(x_p',x_i')-f(x_p,x_i). When Δf ≥ 0, set x ‘As the current PI current controller parameter, simultaneously judge f(x_p',x_i') ​​and the current maximum differential current bandwidth, select the larger value of the two as the new maximum differential current bandwidth to store, and store the corresponding optimal PI current controller parameter; when Δf<0, determine whether to use (x_p',x_i') ​​as the current PI current controller parameter according to the Metropolis criterion; enter the next iteration, generate new PI current controller parameters based on the current PI current controller parameters, and the disturbance size is proportional to the current temperature;

[0021] If the maximum number of iterations is reached or the temperature drops to the set minimum temperature, the iteration ends; otherwise, the iteration continues.

[0022] Furthermore, the measurement of the differential current bandwidth refers to injecting a sinusoidal signal of a specific frequency and amplitude into the differential current command signal of the magnetic bearing, and using an adaptive filter to measure the amplitude of the same frequency component of the actual differential current fed back by the current sensor. The frequency of the injected signal increases successively until the actual differential current amplitude is 0.707 times (-3dB) of the injected differential current amplitude.

[0023] Furthermore, the formula used by the adaptive filter is:

[0024] x=w 1x ×sin(ωt)+w 2x ×cos(ωt)

[0025] y=w 1d ×sin(ωt)+w 2d ×cos(ωt)

[0026]

[0027] w 1d (n+1)=w 1d (n)+2×ε×μ×sin(ωt)

[0028] w 2d (n+1)=w 2d (n)+2×ε×μ×cos(ωt)

[0029] Where x is the sinusoidal excitation signal at the input of the adaptive filter, y is the actual output signal at the input of the adaptive filter, is the output signal of the adaptive filter, w 1x is the sinusoidal weight of signal x, w 2x is the cosine weight of signal x, w 1d is the signal y and the signal The sine weight, w 2d is the output signal y and the signal The cosine weight of the signal, X(s) is the Laplace transform function of the signal x(t), and Y(s) is the signal Laplace transform function, H(s) is the adaptive filter transfer function, |H(s)| is the amplitude-frequency response function, and ε is the signal The error signal w with signal y 1d (n+1) is the corrected signal The sine weight, w 2d (n+1) is the corrected signal The cosine weight of .

[0030] The present invention also provides an electronic device, comprising: a computer-readable storage medium and a processor;

[0031] The computer-readable storage medium is used to store executable instructions;

[0032] The processor is used to read the executable instructions stored in the computer-readable storage medium to execute the above method.

[0033] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to execute the above method.

[0034] The present invention also provides a computer program product, comprising a computer program or instructions, wherein the computer program or instructions implement the above method when executed by a processor.

[0035] In general, the method proposed by the present invention can achieve the following beneficial effects:

[0036] A magnetic bearing differential current control performance evaluation method and a magnetic bearing PI current controller parameter self-tuning algorithm based on a simulated annealing algorithm are established. Through continuous search and optimization of the magnetic bearing PI current controller parameters, the magnetic bearing differential current control achieves the target performance, and the bearing is stably suspended. The evaluation index is intuitive, the search speed is fast, and the target performance matching is good. In order to solve the above problems, the present invention proposes a method for tuning the parameters of a magnetic bearing controller PI current controller for an active magnetic bearing that adopts differential control and a PI current controller. After the PI current controller parameters are brought into the current loop of the magnetic bearing controller, the differential current bandwidth value is output for feedback. The SA algorithm is used to optimize the PI current controller parameters of the current loop, and a set of parameters that make the differential current bandwidth of the electromagnetic coil of the magnetic bearing meet the standard are found. The method has the characteristics of intuitive evaluation index, fast search speed, and good target performance matching. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1A flow chart of a method for setting parameters of a PI current controller of a magnetic suspension bearing for differential current control provided by the present invention;

[0038] Figure 2 This is the current control block diagram of the magnetic bearing controller system;

[0039] Figure 3 To simplify the adaptive filter structure block diagram. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in each embodiment of the present invention described below can be combined with each other and reordered as long as they do not conflict with each other.

[0041] The method for setting parameters of a magnetic bearing PI current controller for differential current control provided by the present invention comprises the following steps:

[0042] At the current moment, the position signal of the magnetic bearing is collected, the position signal is converted into a voltage signal, the target current is calculated according to the voltage signal, the target current is subtracted from the current electromagnetic coil current, and the error current obtained is input into the PI current controller. The current signal output by the PI current controller is then input into the power electronic converter, which is converted into the electromagnetic coil current at the next moment and outputs the power supply to realize the initial suspension of the magnetic bearing;

[0043] Taking the differential current bandwidth of the electromagnetic coil as the optimization target, the parameters of the PI current controller are further optimized, and the simulated annealing algorithm is used to realize the tuning of the PI current controller parameters.

[0044] Given the algorithm's initial state, state generation mechanism, conditions for accepting new solutions, cooling strategy, and termination conditions.

[0045] The flow chart of the method for setting parameters of a magnetic suspension bearing PI current controller for differential current control provided by the present invention is as follows: Figure 1 As shown, the method includes:

[0046] The first step is to determine the control strategy, optimization goal and optimization object of the magnetic bearing system, determine the objective function f(x, y) obtained using a simplified adaptive filter, take the magnetic bearing PI current controller parameters as the optimization object, and take the bandwidth of the differential current of the magnetic bearing electromagnetic coil as the optimization goal. The objective function value is the differential current bandwidth of the electromagnetic coil measured by the adaptive filter with the PI current controller parameters as the input state.

[0047] In the second step, the initial temperature is selected, the state generation mechanism, i.e., the domain generation function, is determined, and a set of initial PI current controller parameters are determined as the current solution (x_p, x_i). In the preliminary experiment, an adaptive filter is used to measure the differential current bandwidth corresponding to the initial PI current controller parameters as the current objective function value f(x_p, x_i), and the value of the optimal PI parameters is stored as the value of the initial PI current controller parameters (x_p, x_i), and the value of the maximum bandwidth is stored as the corresponding differential current bandwidth f(x_p, x_i).

[0048] The third step is to generate new PI current controller parameters as a new solution (x_p', x_i') ​​through the domain generation function determined in the previous step under the current state (x_p, x_i) and f(x_p, x_i), and calculate its objective function value f(x_p', x_i'). At the same time, if the new objective function value f(x_p', x_i') ​​is greater than the differential current target bandwidth, the operation is terminated directly, and the new PI current controller parameters (x_p', x_i') ​​are output as the final parameters obtained by tuning.

[0049] Step 4: Update the current solution according to the result of the previous step and the Metropolis acceptance criterion. If f(x_p', x_i') ​​is larger than f(x_p, x_i), that is, the differential current bandwidth obtained by the newly generated PI current controller parameters is larger than the differential current bandwidth of the current state, then accept the new solution, that is, update the current state to the new PI current controller parameters (x_p', x_i') ​​and the differential current bandwidth f(x_p', x_i') ​​corresponding to the new parameters. At the same time, if the differential current bandwidth of the new solution is larger than the differential current bandwidth of the current optimal PI parameters (that is, the maximum bandwidth), then update the optimal PI parameters to the new PI current controller parameters (x_p', x_i'), and the maximum bandwidth is the differential current bandwidth f(x_p', x_i') ​​corresponding to the new parameters; if f(x_p', x_i') ​​is smaller than f(x_p, x_i), then accept the new solution according to the Metropolis criterion, that is, accept the new solution with a certain probability.

[0050] Step 5: After the update, determine whether the predetermined maximum number of iterations and minimum temperature have been reached. If the maximum number of iterations has been reached, lower the temperature and reset the number of iterations. If the stop temperature has been reached, stop the algorithm and end the operation. Output the optimal PI parameters found by the algorithm this time, which can be used as the initial solution for the next algorithm to increase the possibility of finding parameters that meet the conditions. If the stop temperature has not been reached, jump to step 3 to continue searching and iterative updating.

[0051] Based on this algorithm, the present invention can be regarded as constructing an optimization fitting model, whose independent variables are the actual parameters in the magnetic bearing system (such as bearing rotor displacement, electromagnetic coil current fluctuation, etc.). The model continuously searches and optimizes in the parameter space through the simulated annealing algorithm, and outputs a set of PI current controller parameters that meet the predetermined performance indicators. Specifically, the optimization fitting function realizes the mapping relationship between the controller parameters and the actual state of the system for the system parameters with current fluctuations, thereby ensuring that the appropriate PI controller parameters can be adjusted under various actual working conditions, so that the differential current bandwidth of the electromagnetic coil of the magnetic bearing reaches the operating effect required by international standards.

[0052] Figure 2 The block diagram of the current loop control of the magnetic bearing controller is shown in Figure 1. The current control part uses a current loop system using a PI current controller. The input of the PI current controller is the difference between the reference electromagnetic coil current signal and the actual electromagnetic coil current signal obtained by the power electronic converter. The output current signal enters the power electronic converter and is converted by the power electronic converter into the actual electromagnetic coil current for control.

[0053] Figure 3 To simplify the adaptive filter structure diagram, the adaptive filter formula used is:

[0054] x=w 1x ×sin(ωt)+w 2x ×cos(ωt)

[0055] y=w 1d ×sin(ωt)+w 2d ×cos(ωt)

[0056]

[0057]

[0058] w 1d (n+1)=w 1d (n)+2×ε×μ×sin(ωt)

[0059] w 2d (n+1)=w 2d (n)+2×ε×μ×cos(ωt)

[0060] Where x is the sinusoidal excitation signal at the input of the adaptive filter, y is the actual output signal at the input of the adaptive filter, is the output signal of the adaptive filter, w 1x is the sinusoidal weight of signal x, w 2x is the cosine weight of signal x, w 1d is the signal y and the signal The sine weight, w 2d is the output signal y and the signal The cosine weight of the signal, X(s) is the Laplace transform function of the signal x(t), and Y(s) is the signal Laplace transform function, H(s) is the adaptive filter transfer function, |H(s)| is the amplitude-frequency response function, and ε is the signal The error signal w with signal y 1d (n+1) is the corrected signal The sine weight, w 2d (n+1) is the corrected signal The cosine weight of .

[0061] The filter can be used to obtain the frequency response of a certain frequency signal in the signal, thereby obtaining the differential current bandwidth and constructing the objective function. The method includes:

[0062] The adaptive filter is applied at the feedback and differential positions. By modifying the ω of the input sinusoidal signal of the adaptive filter, the adaptive filter traverses a sufficiently large range of ω, and the amplitude corresponding to all frequencies under certain precision conditions in the corresponding interval is obtained, that is, the frequency response of the system. When used in the parameter setting of the PI current controller of the magnetic bearing controller, the Bode diagram of the differential current of the electromagnetic coil can be obtained, so the required differential current bandwidth can be further obtained by analyzing the Bode diagram.

[0063] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for setting parameters of a PI current controller of a magnetic bearing for differential current control, characterized in that: The following steps are involved: At the current moment, the position signal of the magnetic bearing is collected, the position signal is converted into a voltage signal, the target current is calculated according to the voltage signal, the target current is subtracted from the current electromagnetic coil current, and the error current obtained is input into the PI current controller. The current signal output by the PI current controller is then input into the power electronic converter, which is converted into the electromagnetic coil current at the next moment and outputs the power supply to realize the initial suspension of the magnetic bearing; Taking the bandwidth of the differential current of the electromagnetic coil as the optimization target, the parameters of the PI current controller are optimized and the simulated annealing algorithm is used to realize the tuning of the PI current controller parameters.

2. The method according to claim 1, characterized in that The method of using a simulated annealing algorithm to adjust the parameters of a PI current controller includes: determining initial values ​​and an objective function of the parameters of the PI current controller, iteratively optimizing the parameters of the PI current controller according to the simulated annealing algorithm, and obtaining a set of PI current controller parameters that meet the optimization target.

3. The method according to claim 2, characterized in that The differential current bandwidth value corresponding to the PI current controller parameters (x_p, x_i) in the current state is f(x_p, x_i), and the differential current bandwidth value corresponding to the newly generated PI current controller parameters (x_p', x_i') ​​in each iteration is f(x_p', x_i'). When f(x_p', x_i') ​​is not less than the preset target value, the iteration ends; When f(x_p', x_i') ​​is less than the preset target value, calculate Δf=f(x_p', x_i')-f(x_p, x_i). When Δf>0, use (x_p', x_i') ​​as the current PI current controller parameter, and judge f(x_p', x_i') ​​and the current maximum differential current bandwidth at the same time, select the larger value of the two as the new maximum differential current bandwidth to store, and store the corresponding optimal PI current controller parameter; when Δf≤0, determine whether to use (x_p', x_i') ​​as the current PI current controller parameter according to the Metropolis criterion; Entering the next iteration, new PI current controller parameters are generated based on the current PI current controller parameters, and the disturbance size is proportional to the current temperature; If the temperature drops to the set minimum temperature, the iteration ends; otherwise, the temperature is reduced proportionally, the number of iterations is reset, and the iteration continues.

4. The method according to claim 1, characterized in that: The differential current bandwidth is measured by an adaptive filter, and the measurement method includes: injecting a sinusoidal signal into the differential current command signal; Using an adaptive filter to measure the amplitude of the same frequency component as the injected sinusoidal signal in the actual differential current; The frequency of the injected sinusoidal signal is gradually increased according to a predetermined step size until the amplitude of the actual differential current drops to a preset multiple of the amplitude of the injected sinusoidal signal. The difference between the frequency of the sinusoidal signal at this time and the frequency of the initially injected sinusoidal signal is the differential current bandwidth.

5. The method according to claim 4, characterized in that The formula used for the adaptive filter is: x=w 1x ×sin(ωt)+w 2x ×cos(ωt) y=w 1d ×sin(ωt)+w 2d ×cos(ωt) w 1d (n+1)=w 1d (n)+2×ε×μ×sin(ωt) w 2d (n+1)=w 2d (n)+2×ε×μ×cos(ωt) Where x is the sinusoidal excitation signal at the input of the adaptive filter, y is the actual output signal at the input of the adaptive filter, is the output signal of the adaptive filter, w 1x is the sinusoidal weight of signal x, w 2x is the cosine weight of signal x, w 1d is the signal y and the signal The sine weight, w 2d is the output signal y and the signal The cosine weight of the signal, X(s) is the Laplace transform function of the signal x(t), and Y(s) is the signal Laplace transform function, H(s) is the adaptive filter transfer function, |H(s)| is the amplitude-frequency response function, and ε is the signal The error signal w with signal y 1d (n+1) is the corrected signal The sine weight, w 2d (n+1) is the corrected signal The cosine weight of .

6. An electronic device, characterized in that: include: A computer readable storage medium and a processor; The computer-readable storage medium is used to store executable instructions; The processor is used to read the executable instructions stored in the computer-readable storage medium and execute the method according to any one of claims 1 to 5.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to execute the method according to any one of claims 1 to 5.

8. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images