Permanent magnet synchronous motor sensorless control method based on isoa optimized sliding mode observer

Abstract

This invention discloses a sensorless control method for permanent magnet synchronous motors (PMSMs) based on ISOA-optimized sliding mode observers. The dynamic observation equations of the sliding mode observer (SMO) are reconstructed, and dynamic compensation terms for current errors and adaptive saturation threshold adjustment logic are added. Four core parameters—sliding mode gain h, saturation parameter ε, damping coefficient k, and tracking gain L—are selected as optimization targets. A quantum behavior mechanism is incorporated into the traditional Sliding Ocean Algorithm (SOA), and the follower position update rule is optimized by designing dynamic quantum variance to control the search range. The sensorless control performance of the motor under fixed speed conditions is verified using the optimized SMO core parameters. Compared with existing technologies, this invention, relying on the synergistic design of equation reconstruction and ISOA optimization, significantly improves the accuracy of back EMF estimation, speed tracking effect, and system stability, providing reliable input for the phase-locked loop (PLL), achieving high-precision sensorless control of the PMSM, and adapting to various high-precision drive scenarios.

Description

Technical Field

[0001] This invention relates to the field of sensorless control of motors, and specifically to a sensorless control method for permanent magnet synchronous motors based on an ISOA-optimized sliding mode observer. Background Technology

[0002] Permanent magnet synchronous motors (PMSMs) have been widely used in electric vehicle drive systems due to their advantages such as high efficiency, high power density, and good speed regulation performance. To reduce system costs, decrease the size of the control system, and improve operational reliability, replacing mechanical position sensors with sensorless control methods has become an important research direction in the field of PMSM control. Sliding mode observers (SMOs) have found considerable application in sensorless control of PMSMs due to their relatively simple structure, good robustness, and ease of engineering implementation.

[0003] However, in electric vehicle drive systems, motor operation is often accompanied by changes in operating conditions such as start-stop, acceleration / deceleration, and load variations. This places high demands on the parameter matching of the control system, the accuracy of back EMF observation, and the stability of rotor position and speed calculations. Traditional sliding mode observers still have certain shortcomings in application: the coupling relationship between their key parameters is strong, and when relying on experience for manual tuning, it is often difficult to balance observation accuracy, dynamic response, and chattering suppression; at the same time, the back EMF estimation process is susceptible to parameter mismatch and noise disturbances, which increases the current observation error and further affects the rotor position and speed calculation results.

[0004] Therefore, it is necessary to propose an improved method for sensorless control of permanent magnet synchronous motors to improve the efficiency of sliding mode observer parameter tuning, improve the back EMF observation effect, and thus enhance the stability and practicality of the motor control system, providing support for its application in electric vehicle driving scenarios. Summary of the Invention

[0005] Purpose of the invention: To address the problems mentioned in the background art, this invention provides a sensorless control method for permanent magnet synchronous motors based on ISOA optimized sliding mode observers. By optimizing the key parameters of the sliding mode observer and combining it with observation equation reconstruction, the back EMF estimation effect is improved, the current observation error is reduced, and the accuracy of rotor position and speed calculation is increased. This enhances the stability and practicality of the sensorless control system for permanent magnet synchronous motors and provides technical support for its application in electric vehicle drive systems.

[0006] Technical solution: This invention discloses a sensorless control method for permanent magnet synchronous motors based on an ISOA-optimized sliding mode observer, comprising the following steps:

[0007] S1: Establish a coordinate coefficient mathematical model of the permanent magnet synchronous motor (PMSM) and build the basic control architecture of the sliding mode observer (SMO) and phase-locked loop (PLL).

[0008] S2: The back electromotive force calculation model of the sliding mode observer (SMO) output to the PLL is supplemented with a current error dynamic compensation term and an adaptive saturation threshold adjustment logic for accurate reconstruction.

[0009] S3: Select the SMO core control parameters associated with the optimized back EMF model in S2, determine them as the optimization targets of the intelligent optimization algorithm, and define the scope of objects for subsequent algorithm optimization;

[0010] S4: Iteratively improve the SOA algorithm of the tunicate, incorporate quantum behavior mechanism to construct the ISOA algorithm, and optimize the core control parameters of SMO;

[0011] S5: Verify the sensorless control performance of the motor under fixed speed conditions using the optimized SMO core parameters.

[0012] Furthermore, the current dynamic observation formula of the SMO back electromotive force calculation model in S2 is precisely reconstructed, specifically as follows:

[0013] ;

[0014] in, , For d-axis and q-axis inductance, , for shaft voltage, Electric angular velocity, For resistance, , for The observed current, Current observation error, For sliding mode gain, For saturation parameters, To suppress the damping coefficient of sliding surface oscillation, To track gain, It is a saturation function.

[0015] Furthermore, saturation function The details are as follows;

[0016] ;

[0017] Where δ is the saturation threshold, and the dynamic adjustment formula is:

[0018] ;

[0019] in, The number of simulation iterations. The total number of simulation steps is used to dynamically adjust the saturation threshold as the motor runs.

[0020] Furthermore, in S2, the back EMF calculation model updates the estimated back EMF value through integral tracking, introducing a tracking gain L. The formula for updating the estimated back EMF value is as follows:

[0021] ;

[0022] in, , for Estimated voltage, , This is the differential of the current observation error.

[0023] Furthermore, the four core control parameters selected in S3 are used as the optimization targets of the ISOA algorithm in S5, namely: sliding mode gain. saturation parameter That is, the saturation gain, the damping coefficient that suppresses sliding surface oscillations. Tracking gain .

[0024] Furthermore, the objective function for optimizing the SMO core control parameters in S3 is as follows:

[0025]

[0026] in, These are the weighting coefficients. For the maximum position estimation error, The root mean square of the stator current. For the maximum speed observation error, This represents the average back electromotive force observation error.

[0027] Furthermore, in S4, the SOA algorithm is iteratively improved by incorporating a quantum behavior mechanism to construct the ISOA algorithm. The specific quantum behavior mechanism is as follows:

[0028] ;

[0029] in, Let be the quantum variance of the parameters of the j-th sliding mode observer in the t-th generation. The initial quantum variance of the j-th parameter, It is a natural exponential function. Let the current iteration algebra be... Maximum number of iterations.

[0030] Furthermore, the update process of the ISOA algorithm in S4 is as follows:

[0031] ;

[0032] in, , It is the parameter value of the (t+1)th generation, the i-th follower in the t-th generation, and the j-th parameter value. It is the j-th component of the global optimal parameter in generation t, and it is the j-th parameter value of the i-th follower in generation t.

[0033] Furthermore, the specific process for verifying the sensorless control performance of the motor under fixed speed conditions in S5 is as follows:

[0034] The SMO core optimization parameters obtained from S4 are substituted into the PMSM sensorless control simulation model to generate a smooth back EMF signal E. The back EMF signal E is directly used as the core input signal of the subsequent phase-locked loop (PLL). After accurate calculation by the PLL, the rotor angle position and speed signals are output as the final detection output signals to achieve stable operation control of the motor under sensorless conditions.

[0035] Beneficial effects:

[0036] 1. Based on the traditional tunic algorithm, the ISOA algorithm is constructed by incorporating quantum behavior mechanism. By designing dynamic quantum variance to adjust the parameter search range and optimizing the follower position update rule, the search efficiency and optimization accuracy of the algorithm in high-dimensional space are enhanced, achieving global optimal matching and avoiding the subjectivity and limitations of traditional manual parameter tuning.

[0037] 2. The back EMF calculation model of the sliding mode observer is accurately reconstructed, and a dynamic compensation term for current error and an adaptive saturation threshold adjustment logic are added to effectively optimize the output characteristics of the sliding mode observer and reduce the basis of current observation error and back EMF ripple.

[0038] 3. Precise matching algorithm and observer enable high-precision state observation across the entire speed domain. Based on the SMO observation equations reconstructed using ISOA-optimized parameters, the convergence speed of the sliding surface and the damping suppression effect can be precisely controlled, significantly reducing speed observation errors and position estimation deviations to achieve accurate calculation and tracking of rotor position and speed.

[0039] 4. Enhanced System Robustness. Through the collaborative design of observation equation reconstruction and ISOA parameter optimization, the optimized SMO can effectively offset the current observation errors caused by external disturbances and load changes, ensuring that the back EMF estimate remains stable. This provides a continuous and reliable input to the phase-locked loop (PLL), avoids PLL calculation drift caused by signal distortion, significantly reduces the fluctuation amplitude of speed and position observations, and significantly improves anti-interference capability. Attached Figure Description

[0040] Figure 1 Flowchart of the control strategy for permanent magnet synchronous motor;

[0041] Figure 2 Graph illustrating the parameter optimization process for the ISOA algorithm;

[0042] Figure 3 A performance comparison chart of various algorithms under the fitness function;

[0043] Figure 4 Comparison of back EMF estimation for traditional SMO and ISOA-improved SMO;

[0044] Figure 5 Comparison of electrical angle errors between traditional SMO and ISOA improved SMO;

[0045] Figure 6 Comparison of the electric angular velocity effects of traditional SMO and ISOA improved SMO. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] This embodiment discloses a sensorless control method for a permanent magnet synchronous motor based on an ISOA-optimized sliding mode observer. The method steps are as follows:

[0048] S1: Establish a coordinate coefficient mathematical model of the permanent magnet synchronous motor (PMSM) and build the basic control architecture of sliding mode observer (SMO) and phase-locked loop (PLL).

[0049] S1.1: Establishing a permanent magnet synchronous motor in two-phase stationary states State-space equations in coordinate system and sliding mode brushless DC motor The equations for shaft voltage and current are:

[0050]

[0051]

[0052] in, , They are respectively , Shaft stator voltage, , They are respectively , Shaft stator current, For stator resistance, For stator equivalent inductance, Electric angular velocity, It is a permanent magnet flux linkage. The rotor electrical angle.

[0053] S1.2: Constructing the current observation equations and sliding surface functions of the SMO:

[0054] The equation for current observation is:

[0055]

[0056] in, , For inductance, , For voltage, Electric angular velocity, For resistance, , for The observed current, , for The observed voltage.

[0057] The sliding surface function is defined as the current observation error:

[0058]

[0059] S2: For the back EMF calculation model of the SMO output to the PLL, the formula is reconstructed to improve accuracy in back EMF estimation, taking into account control precision requirements. The current observation equation is:

[0060]

[0061] in, , For inductance, , For voltage, Electric angular velocity, For resistance, , for The observed current, Current observation error, For sliding mode gain, For saturation gain, The damping coefficient is used to suppress oscillations of the sliding surface (current error). To track gain. The specific formula for the saturation function is as follows;

[0062]

[0063] Where δ is the saturation threshold, and its dynamic adjustment formula is:

[0064]

[0065] in, The number of simulation iterations. The total number of simulation steps is given by this formula. The saturation threshold is dynamically adjusted as the motor runs, which solves the contradiction between chattering and convergence in the motor startup and steady-state operation stages when the traditional fixed threshold is used.

[0066] The back electromotive force estimate is updated via integral tracking. A tracking gain L is introduced to improve the convergence speed and tracking accuracy of the estimate. The update formula is as follows:

[0067]

[0068] in, , This is the differential of the current observation error.

[0069] The motor parameters simulated in this embodiment (adapted for brushless DC motors in the transportation field) are shown in Table 1 below:

[0070] Table 1 PMSM Parameter Table

[0071] parameter numerical values parameter numerical values Stator resistance 7.2Ω Stator inductor 5.5mh Extreme logarithm 4 Damping coefficient 1.2Nms Moment of inertia <![CDATA[0.8kg.m 2 ]]> Rated speed 1700rpm

[0072] S3: Select the SMO core control parameters associated with the optimized back EMF formula in step S2, and determine them as the optimization targets of the intelligent optimization algorithm, thus defining the scope of the target for subsequent algorithm optimization:

[0073] Based on the reconstructed SMO back EMF calculation model in S2, four core control parameters that play a decisive role in observation performance were selected as the optimization targets of the ISOA algorithm. These parameters are: For sliding mode gain, For saturation gain, The damping coefficient is used to suppress oscillations of the sliding surface (current error). To track gain.

[0074] The optimization range (upper and lower bounds) of each parameter are clearly defined, taking into account both engineering practice and optimization efficiency. The lower bound of the parameter is set as lb=[1,0.001,0.001,1], and the upper bound is set as ub=[20,2,2,1], that is, h∈[1,20], ε∈[0.001,2], k∈[0.001,2], L∈[1,20], which defines a clear range for subsequent ISOA algorithm parameter optimization.

[0075] S4: Improve and optimize the SMO core control parameters of the SOA algorithm;

[0076] S4.1: Improves traditional SOA by incorporating quantum behavior mechanisms to construct the ISOA algorithm:

[0077] Based on the traditional Leader-Follower Algorithm (SOA) of the Search Algorithm (SOA), a quantum behavior mechanism is incorporated to construct the Improved Search Algorithm (ISOA). This addresses the problems of low search efficiency, susceptibility to local optima, and insufficient convergence accuracy in the traditional SOA. The core improvements are as follows:

[0078] ISOA algorithm constructed by incorporating dynamic quantum variance control:

[0079]

[0080] in, Let be the quantum variance of the parameters of the j-th sliding mode observer in the t-th generation. The initial quantum variance of the j-th parameter, It is a natural exponential function. Let the current iteration algebra be... Maximum number of iterations. The quantum variance decreases with each iteration using a natural exponential function, allowing the algorithm to perform global exploration in the early stages and local development in the later stages, balancing search breadth and accuracy. The follower position update rule is optimized by introducing quantized follower position updates, with the following formula:

[0081]

[0082] in, It is the parameter value of the i-th follower in the (t+1)-th generation. It is the j-th component of the globally optimal parameter t.

[0083] S4.2: The fitness objective function for the ISOA algorithm is constructed as follows:

[0084]

[0085] in, These are the weighting coefficients. For the maximum position estimation error, The stator current is the root mean square (only the steady-state phase of the motor is considered). This represents the maximum speed observation error (compared to the reference speed). To average the back EMF observation error, the influence of each index on SMO performance is balanced by weight allocation, avoiding the dominance of a single index in the optimization results.

[0086] S4.3: Implementation of ISOA Algorithm Parameter Optimization:

[0087] The algorithm execution flow is as follows:

[0088] (1) Initialize the location of the salps population, and randomly generate 40 population individuals within the upper and lower bounds of the parameters. Each individual is a 1×4 parameter vector [h,ε,k,L];

[0089] (2) Calculate the fitness value of each individual in the population, and select the globally optimal individual X using the objective function as the fitness. best ;

[0090] (3) Iteratively update the positions of leaders and followers, and optimize the population positions generation by generation by combining dynamic quantum variance, quantized follower update rules and adaptive perturbation strategy;

[0091] (4) After each iteration, update the global best individual. If the fitness value of an individual in the current population is less than the global best value, then update X. best ;

[0092] (5) After reaching the maximum number of iterations, output the global optimal parameters [h,ε,k,L], which is the optimal solution of the SMO core control parameters.

[0093] In this embodiment, the optimal SMO parameters obtained after ISOA algorithm optimization are: h≈16, ε≈0.03, k≈0.2, L≈6 (the specific values ​​may fluctuate slightly with simulation iterations).

[0094] See Figure 3The graph compares the performance of various algorithms under the fitness function. As can be seen from the comparison of the average convergence curves of the algorithms, ISOA clearly outperforms SOA, PSO, GWO, and GA algorithms in all three core optimization performance dimensions: convergence accuracy, convergence speed, and convergence stability. In the early stages of iteration, ISOA can quickly reduce its performance index, approaching the optimal solution region more efficiently, demonstrating superior convergence speed. During iteration, its curve is smooth without significant oscillations, maintaining a stable downward trend, avoiding the drastic fluctuations of the GA algorithm, and exhibiting stronger convergence stability. At 300 iterations, ISOA achieves the lowest convergence performance index among all algorithms, indicating that it can find a better solution and its convergence accuracy is superior to other compared algorithms.

[0095] S5: Build a PMSM sensorless control simulation model on the MATLAB platform. The simulation parameters are consistent with the motor parameters. Set the sampling frequency to 10kHz (sampling time Ts=1×10−4s). The total simulation time is 0.5s, covering the motor start-up stage (0~0.2s) and steady-state operation stage (0.2~0.5s).

[0096] The core modules of the simulation model include: PMSM body model, current loop PI control module, reconstructed SMO module, PLL position / speed calculation module, and speed loop PI control module. The optimal parameters [h,ε,k,L] obtained by ISOA optimization in S4 are substituted into the SMO module to complete the model parameter configuration.

[0097] Using a fixed speed of 1000 rpm as the verification scenario, the performance of Particle Swarm Optimization (PSO), Gray Wolf Optimization (GWO), Genetic Algorithm (GA), traditional SOA algorithm, and the optimized ISOA algorithm were compared. The optimization advantages of the ISOA algorithm of this invention were verified from four dimensions: back EMF estimation accuracy, speed tracking effect, position estimation accuracy, and current chattering degree. The core verification results are as follows:

[0098] See Figure 4 Back EMF estimation accuracy: The SMO output optimized by ISOA has a very high degree of fit with the actual value, with an average back EMF observation error of ≤0.05V and no obvious distortion or chattering.

[0099] See Figure 5 Rotor angle error comparison: The phase deviation between the estimated rotor angle position and the actual value is extremely small, with the maximum position estimation error ≤0.25°, meeting the high-precision position requirements of PMSM sensorless control.

[0100] See Figure 6 Rotor speed tracking effect: The estimated speed of the motor after steady state completely coincides with the actual speed, which is far lower than the effect of traditional SMO control. There is no overshoot or steady-state error, and the tracking response speed is fast.

[0101] The foregoing description of the embodiments enables those skilled in the art to make or use the present invention. Various modifications to the embodiments will be readily apparent to those skilled in the art. The general principles of the invention may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention should not be limited to the embodiments shown herein, but should cover the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A sensorless control method for a permanent magnet synchronous motor based on an ISOA-optimized sliding mode observer, characterized in that, Includes the following steps: S1: Establish a coordinate coefficient mathematical model of the permanent magnet synchronous motor (PMSM) and build the basic control architecture of the sliding mode observer (SMO) and phase-locked loop (PLL). S2: The back electromotive force calculation model of the sliding mode observer (SMO) output to the PLL is supplemented with a dynamic current error compensation term and an adaptive saturation threshold adjustment logic for precise reconstruction. S3: Select the SMO core control parameters associated with the optimized back EMF model in S2, determine them as the optimization targets of the intelligent optimization algorithm, and define the scope of objects for subsequent algorithm optimization; S4: Iteratively improve the SOA algorithm of the tunicate, incorporate quantum behavior mechanism to construct the ISOA algorithm, and optimize the core control parameters of SMO; S5: Verify the sensorless control performance of the motor under fixed speed conditions using the optimized SMO core parameters.

2. The sensorless control method for a permanent magnet synchronous motor based on an ISOA-optimized sliding mode observer according to claim 1, characterized in that, The current dynamic observation formula of the SMO back electromotive force calculation model in S2 is reconstructed for precision, specifically as follows: ； in, , For d-axis and q-axis inductance, , for shaft voltage, Electric angular velocity, For resistance, , for The observed current, Current observation error, For sliding mode gain, For saturation parameters, To suppress the damping coefficient of sliding surface oscillation, To track gain, It is a saturation function.

3. The sensorless control method for a permanent magnet synchronous motor based on an ISOA-optimized sliding mode observer according to claim 2, characterized in that, Saturation function The details are as follows; ； Where δ is the saturation threshold, and the dynamic adjustment formula is: ； in, The number of simulation iterations. The total number of simulation steps is used to dynamically adjust the saturation threshold as the motor runs.

4. The sensorless control method for a permanent magnet synchronous motor based on an ISOA-optimized sliding mode observer according to claim 2, characterized in that, In S2, the back EMF calculation model updates the estimated back EMF value through integral tracking, introducing a tracking gain L. The formula for updating the estimated back EMF value is as follows: ； in, , for Estimated voltage, , This is the differential of the current observation error.

5. The sensorless control method for a permanent magnet synchronous motor based on an ISOA-optimized sliding mode observer according to claim 2, characterized in that, The four core control parameters selected in S3 are used as the optimization targets of the ISOA algorithm in S5, namely: sliding mode gain. saturation parameter That is, the saturation gain, the damping coefficient that suppresses sliding mode surface oscillations. Tracking gain .

6. The sensorless control method for a permanent magnet synchronous motor based on an ISOA-optimized sliding mode observer according to claim 1, characterized in that, The objective function for optimizing the core control parameters of the SMO using the ISOA algorithm in S4 is as follows: in, These are the weighting coefficients. For the maximum position estimation error, The root mean square of the stator current. For the maximum speed observation error, This represents the average back electromotive force observation error.

7. The sensorless control method for a permanent magnet synchronous motor based on an ISOA-optimized sliding mode observer according to claim 1, characterized in that, In S4, the SOA algorithm for tunicate sea snails is iteratively improved by incorporating a quantum behavior mechanism to construct the ISOA algorithm. Specifically, the quantum behavior mechanism is as follows: ； in, Let be the quantum variance of the parameters of the j-th sliding mode observer in the t-th generation. The initial quantum variance of the j-th parameter, It is a natural exponential function. Let the current iteration algebra be... Maximum number of iterations.

8. The sensorless control method for a permanent magnet synchronous motor based on an ISOA-optimized sliding mode observer according to claim 7, characterized in that, The introduction of the ISOA algorithm in S4 to quantize the follower position update process is as follows: ； in, , It is the parameter value of the (t+1)th generation, the i-th follower in the t-th generation, and the j-th parameter value. It is the j-th component of the global optimal parameter in generation t, and it is the j-th parameter value of the i-th follower in generation t.

9. The sensorless control method for a permanent magnet synchronous motor based on an ISOA-optimized sliding mode observer according to claim 1, characterized in that, The specific process for verifying the sensorless control performance of the motor under fixed speed conditions in S5 is as follows: The SMO core optimization parameters obtained from S4 are substituted into the PMSM sensorless control simulation model to generate a smooth back EMF signal E. The back EMF signal E is directly used as the core input signal of the subsequent phase-locked loop (PLL). After accurate calculation by the PLL, the rotor angle position and speed signals are output as the final detection output signals to achieve stable operation control of the motor under sensorless conditions.

Images