Method for compensating sector commutation based on direct current brushless motor control

Abstract

The application relates to the technical field of motor control and discloses a sector commutation compensation method based on direct-current brushless motor control, which estimates the motor electrical angle in real time through an extended Kalman filter state observer; when the Hall position sensor signal jumps, the error between the Hall theoretical angle at the moment and the estimated electrical angle is used to adjust the motor model parameters in the observer online to resist parameter drift; meanwhile, the current differential profile is calculated based on the phase current signal, and an asymmetry index capable of quickly representing load change is extracted; through fusion of feedback information provided by the estimated electrical angle and feedforward information provided by the asymmetry index, a dynamic target commutation phase angle is calculated, and commutation is performed accordingly. The application effectively suppresses the influence of parameter drift, significantly enhances the response speed and robustness of the system under dynamic load, and improves the comprehensive control performance of the motor.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of motor control, in particular to a sector commutation compensation method based on a direct-current brushless motor. BACKGROUND

[0002] Direct-current brushless motors are widely used in industrial automation, electric vehicles and household appliances due to their high efficiency, high power density and good reliability. In the conventional control scheme of direct-current brushless motors, a six-step commutation method based on Hall position sensors is usually adopted. By detecting the rotor magnetic pole position, the energization state of the stator winding is switched at the preset six discrete angle points to generate continuous driving torque.

[0003] However, due to the existence of motor winding inductance, the establishment of current requires a certain time, resulting in a lag of the actual current phase relative to the ideal back electromotive force phase when the motor is running at high speed, which will weaken the output torque and reduce the running efficiency. To solve this problem, the existing technology generally adopts a commutation advance angle compensation strategy, that is, according to the motor speed, a pre-off-line calibrated data table is consulted to perform commutation operation in advance.

[0004] This compensation method based on fixed lookup table highly depends on the constancy of motor model parameters. However, in actual operation, the resistance of motor winding will significantly increase due to the temperature rise effect, and its inductance value will also nonlinearly decrease under heavy load conditions due to saturation of the stator core magnetic circuit. The real-time changes of these motor physical parameters make the off-line calibrated fixed advance angle unable to continue to match the optimal commutation point under the current working condition, resulting in deviation of the control performance from the design target, causing unnecessary energy loss and torque pulsation.

[0005] In addition, this method shows inherent lag when responding to dynamic load changes. When the load changes suddenly, the system can only passively wait for the change to be reflected as a fluctuation in speed, and then adjust the compensation angle through speed feedback. This adjustment method that relies on post-error has a long response chain, inevitably leading to a large speed drop and a long recovery time when the motor is subjected to load shock, and the dynamic robustness of the system is insufficient. To improve the position estimation accuracy, some control methods use state observers, but the performance of these observers is also based on accurate motor models. If the model parameters are inaccurate or drift, the estimation accuracy of the observer will decrease, and ultimately accurate commutation control cannot be achieved. SUMMARY

[0006] The technical problem to be solved by the present application is to provide a direct-current brushless motor adaptive commutation compensation method that can simultaneously cope with motor model parameter drift and dynamic load shock, to improve the running accuracy, efficiency and dynamic response performance of the motor control system under complex working conditions.

[0007] To solve the above technical problems, the application provides a sector commutation compensation method based on a direct-current brushless motor control, comprising:

[0008] In each PWM control cycle, the phase current signal of the direct-current brushless motor is synchronously collected;

[0009] An extended Kalman filter state observer is run, and the motor electrical angle is estimated in real time according to the phase current signal and the current effective motor model parameter;

[0010] When the signal of the Hall position sensor is detected to jump, the motor model parameter is adjusted online according to the error between the Hall theoretical angle at the moment and the motor electrical angle;

[0011] Based on the phase current signal, the current differential profile is calculated and the asymmetry index representing the load change is extracted;

[0012] The motor electrical angle and the asymmetry index are fused to calculate a dynamic target commutation phase angle;

[0013] The motor electrical angle and the target commutation phase angle are compared, and when a preset commutation condition is met, a commutation operation is performed.

[0014] In one specific embodiment of the application, the online adjustment mechanism of the motor model parameter is based on the technical principle of using the discrete but absolutely accurate position reference provided by the Hall position sensor to calibrate the continuous angle estimation based on the model. When the Hall signal jumps, the system captures the accurate moment, and compares the Hall theoretical angle at the moment with the motor electrical angle output by the extended Kalman filter state observer. The deviation between the two is regarded as the model prediction error, which directly reflects the mismatch degree between the current motor model parameter and the actual motor physical parameter. The model prediction error is used as the input of one or more proportional-integral (PI) controllers, and the output of the controller is directly used to correct the estimated resistance value and the estimated inductance value in the motor model. This closed-loop adjustment mechanism enables the motor model parameter to dynamically track the actual parameter drift caused by factors such as temperature change or magnetic circuit saturation, thereby ensuring the long-term accuracy of the motor electrical angle output by the extended Kalman filter state observer.

[0015] In another specific embodiment of the present application, the extraction of the asymmetry index is based on the micro-structure of the current waveform within a single PWM cycle. Specifically, the current differential profile is obtained by calculating the rising slope of the current during the power tube on-state and the falling slope of the current during the freewheeling state. When the motor load changes dramatically, the amplitude of the winding current changes accordingly, which in turn causes a nonlinear change in the degree of magnetic saturation of the stator core, and this change directly leads to a transient change in the effective inductance value of the winding. According to the motor voltage equation, the change in inductance will cause a change in the rate of current change, resulting in a difference in amplitude between the rising slope and the falling slope. The asymmetry index is a quantitative representation of this difference, and therefore it can be used as a high-time-efficiency indicator to detect load changes before they have a significant impact on the motor speed.

[0016] In another specific embodiment of the present application, the calculation of the dynamic target commutation phase angle combines feedback control and feedforward control. The target commutation phase angle is composed of three linearly superimposed parts: a basic sector angle, which defines the theoretical boundary of the next commutation sector; a state feedback-based advance angle, whose value is determined based on current motor speed and current and other steady-state operating parameters, to optimize output torque and efficiency in steady-state conditions; and a feedforward advance angle based on the asymmetry index. The feedforward advance angle is directly related to the asymmetry index, and when a load mutation is detected, this feedforward component can quickly and actively adjust the target commutation phase angle to compensate for the impending system dynamics.

[0017] The present application provides a method for controlling sector commutation compensation based on a direct-current brushless motor. The method has the following advantages:

[0018] 1. The present application uses the absolute position information at the time of Hall signal jump to periodically calibrate the motor model parameters relied on by the extended Kalman filter state observer, effectively suppressing the accumulation of angle estimation errors caused by parameter drift due to motor temperature rise or changes in operating conditions, and ensuring the stability and high precision of the control system in the entire operating range and long operating period.

[0019] 2. The present application extracts the asymmetry index in the current differential profile to establish a fast feedforward channel from load transients to commutation angle compensation. This mechanism can compensate for the angle at the first time of load impact, much faster than the traditional speed loop feedback adjustment method, thereby effectively reducing the speed drop under load impact and shortening the recovery time.

[0020] 3、The application combines high-precision continuous angle estimation, online self-tuning of model parameters, and fast dynamic feedforward compensation organically, so that the motor control system can not only work at the optimal efficiency point in steady-state operation, but also show stronger anti-disturbance ability in dynamic process, meeting the stringent requirements of high-performance servo drive, electric vehicles, and other application scenarios for precision, efficiency, and dynamic performance. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 A schematic diagram of the hardware architecture of the motor control system of an embodiment of the application;

[0022] Figure 2 A system function module architecture diagram deployed in a microcontroller of an embodiment of the application;

[0023] Figure 3 A flowchart of the adaptive commutation compensation method of an embodiment of the application;

[0024] Figure 4 A comparison curve diagram of the operating efficiency of the embodiment and the comparative embodiment under different loads;

[0025] Figure 5 A comparison diagram of the phase current waveforms of the embodiment and the comparative embodiment when the motor operates under the 80% rated speed and 75% rated load working conditions;

[0026] Figure 6 A comparison curve diagram of the motor speed response of the embodiment and the comparative embodiment under load mutation conditions.

[0027] 1, microcontroller; 2, three-phase inverter bridge; 3, DC brushless motor; 4, Hall position sensor; 5, phase current sampling circuit; 10, high-frequency electrical signal acquisition and differential profile analysis module; 20, back electromotive force phase observer module based on extended Kalman filter; 30, motor model online self-tuning module based on Hall signal reference; 40, hybrid commutation decision module fusing feedforward and feedback; 50, PWM generation and inverter driving module. DETAILED DESCRIPTION

[0028] To make the purpose, technical solutions, and advantages of the application clearer and more apparent, the DC brushless motor adaptive commutation compensation method and system provided by the application with model self-tuning and feedforward control will be described in detail below in combination with the drawings and one or more specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the application, and do not constitute a limitation on the application.

[0029] Referring to the drawings Figure 1 , Figure 1This is a schematic diagram of the hardware architecture of a motor control system according to an embodiment of the present invention. The system includes: a microcontroller 1, a three-phase inverter bridge 2, a brushless DC motor 3, a Hall position sensor 4, and a phase current sampling circuit 5. The control output terminal of the microcontroller 1 is connected to the control input terminal of the three-phase inverter bridge 2, and the power output terminal of the three-phase inverter bridge 2 is connected to the stator winding of the brushless DC motor 3.

[0030] The brushless DC motor 3 integrates a Hall position sensor 4, and the signal output terminal of the Hall position sensor 4 is connected to the input interface of the microcontroller 1. The phase current sampling circuit 5 is set on the power path of the three-phase inverter bridge 2 to detect the phase current of the brushless DC motor 3, and its output terminal is connected to the analog-to-digital converter (ADC) input terminal of the microcontroller 1.

[0031] See attached document Figure 2 , Figure 2 This is a system functional module architecture diagram deployed in a microcontroller 1 according to an embodiment of the present invention. The system functionally includes: a high-frequency electrical signal acquisition and differential profile analysis module 10, a back EMF phase observer module 20 based on extended Kalman filtering, an online self-tuning module for a motor model based on Hall signal reference 30, a hybrid commutation decision module 40 integrating feedforward and feedback, and a PWM generation and inverter drive module 50.

[0032] The high-frequency electrical signal acquisition and differential profile analysis module 10 receives the phase current signal from the phase current sampling circuit 5, processes it, and outputs the current differential profile and asymmetry index.

[0033] The back EMF phase observer module 20 based on extended Kalman filtering receives the current differential profile, system parameters from the motor model online self-tuning module 30 based on Hall signal reference, and other electrical signals to estimate the real-time motor electrical angle.

[0034] The online self-tuning module 30 of the motor model based on the Hall signal reference receives the signal from the Hall position sensor 4 and the estimated electrical angle from the back EMF phase observer module 20 based on the extended Kalman filter, calculates the model error, and outputs the corrected motor model parameters to the back EMF phase observer module 20 based on the extended Kalman filter.

[0035] The hybrid commutation decision module 40, which integrates feedforward and feedback, receives the estimated electrical angle and asymmetry index, calculates the target commutation angle, and issues a commutation command.

[0036] The PWM generation and inverter drive module 50 receives the commutation command, generates the corresponding PWM control signal, and applies it to the three-phase inverter bridge 2.

[0037] See attached document Figure 3 ,Figure 3 is a flow chart of an adaptive commutation compensation method according to an embodiment of the present application. The present application provides a method for commutation compensation of a direct current brushless motor control sector, comprising the following steps:

[0038] S10, system initialization is performed, and initial parameters of a motor model and hardware peripherals required by a motor control system are configured;

[0039] S20, in each PWM control cycle, phase current signals of the direct current brushless motor are synchronously collected;

[0040] S30, based on the phase current signals, current differential profiles are calculated, and an asymmetry index representing load change is extracted;

[0041] S40, a preset extended Kalman filter state observer is run, and a state vector including a motor electrical angle is estimated in real time according to the phase current signals and currently effective motor model parameters;

[0042] S50, a signal state of a Hall position sensor is monitored, when a signal jump is detected, an error between a Hall theoretical angle at the moment and an estimated electrical angle is used to adjust motor model parameters online;

[0043] S60, the estimated electrical angle and the asymmetry index are fused, and a dynamic target commutation phase angle is calculated;

[0044] S70, the real-time estimated electrical angle is compared with the target commutation phase angle, when a preset commutation condition is met, a commutation operation is performed, and a PWM output state for driving the motor is updated.

[0045] The technical details and specific implementation modes involved in the above steps will be described in detail below.

[0046] In step S10, the motor control system is initialized, and the step establishes necessary hardware working environments and software initial states for subsequent real-time control cycles.

[0047] The initialization step can specifically include the following steps:

[0048] S101, configure the hardware peripherals of the microcontroller 1. The configuration includes but is not limited to: configure the GPIO port for receiving the Hall position sensor 4 signal as an input mode with external interrupt function; configure the timer module to generate a center-aligned mode PWM waveform, and set its switching frequency, pre-division coefficient and auto-reload value; configure the ADC module, associate its trigger source to a specific event (e.g. update event or comparison match event) of the timer module, to establish the synchronization relationship between ADC sampling and PWM cycle. The configuration method of the specific registers of the microcontroller can be completed by the person skilled in the art according to the chip manual, which belongs to the prior art and will not be described here.

[0049] S102, set the initial value of the motor model parameter. To provide the calculation basis for the back-EMF phase observer module 20 based on extended Kalman filter and the motor model online self-tuning module 30 based on Hall signal reference, the resistance and inductance parameters in the motor model need to be initialized. Specifically, the estimated resistance value and the estimated inductance value are set to the nominal value determined by experiment in advance or provided by the motor manufacturer.

[0050] S103, initialize the state variable of the extended Kalman filter (EKF) state observer. To ensure the convergence of the EKF algorithm and the stability in the initial stage, the state vector and the error covariance matrix need to be set.

[0051] The initial state vector is set as follows:

[0052] ;

[0053] wherein, is the initial phase current estimation value, which is assumed to be 0 before the motor starts; is the initial electrical angular velocity estimation value, which is assumed to be 0 before the motor starts; is the initial electrical angle estimation value, which is set to , is obtained by looking up the preset Hall state-angle correspondence table according to the initial state combination of the Hall position sensor 4 read at power-on; is the vector transpose operator.

[0054] The initial error covariance matrix is set as a diagonal matrix to reflect the uncertainty of each initial state estimation value:

[0055] ;

[0056] wherein, is the variance of initial current estimation, which is directly measured and has a small initial uncertainty, so it can be set to a small value; is the variance of initial angular velocity estimation, which is a deterministic state at start-up when the velocity is zero, so it can also be set to a small value; is the variance of initial angle estimation, which has the largest uncertainty because the initial angle resolution provided by the Hall sensor is low (within a 60-degree electrical angle sector), so it needs to be set to a relatively large value to allow the EKF to quickly converge on the angle in the initial stage.

[0057] In step S20, the phase current signals of the brushless DC motor 3 are synchronously sampled in each PWM control cycle to obtain the instantaneous current values that can reflect the electrical characteristics of the motor and are used for subsequent calculation and state estimation.

[0058] The synchronous sampling step specifically can include:

[0059] S201, a synchronous relationship between the sampling time and the PWM carrier is established. In a specific embodiment, to avoid electromagnetic interference introduced by the switching transient process of the power tubes in the three-phase inverter bridge 2, the sampling time point of the phase current is set in the stable interval of the PWM cycle. Specifically, when the timer module in the microcontroller 1 is configured to generate a center-aligned PWM waveform, the interval of the current ripple is relatively stable at the peak point and the valley point of the PWM cycle. Therefore, the timer module is configured to automatically generate a hardware trigger signal when the counter reaches the top of the count value (peak point) or the zero point of the count value (valley point).

[0060] S202, analog-to-digital conversion is performed. The hardware trigger signal is directly used as the conversion start source of the ADC module inside the microcontroller 1. After receiving the trigger signal, the ADC module performs analog-to-digital conversion on the analog voltage signal from the phase current sampling circuit 5. Through this direct linkage at the hardware level, it is ensured that the time of each current sampling is synchronized with the PWM carrier, eliminating the uncertainty caused by software delay.

[0061] S203, the current data is obtained and stored. After the ADC module completes the conversion, the digital result obtained is stored in the data register. The microcontroller 1 reads the digital result and calibrates it to a physical quantity in amperes (A) according to the gain coefficient of the phase current sampling circuit 5 and the reference voltage of the ADC module. The obtained current value will be passed to subsequent steps S30 and S40 for processing. In an embodiment, the phase current sampling circuit 5 uses a single sampling resistor, which is connected in series to the negative terminal of the DC bus of the three-phase inverter bridge 2. At this time, the current value collected is the current flowing through the power tube of the lower bridge arm that is turned on, and the size of the current value is equal to the size of the phase current of the current conduction phase, and the sign is determined according to the specific commutation sector state.

[0062] In step S30, based on the phase current signal acquired in step S20, a process is performed to calculate the current differential profile, and from it an asymmetry index is extracted which is capable of characterizing the dynamic changes in the load of the electric machine.

[0063] This step can specifically include:

[0064] S301, obtaining a plurality of current sampling values for calculating the slope. To calculate the current rate of change in a PWM cycle, at least two current values at different times are needed. In a specific embodiment, the ADC module is configured to sample at the start and end of the PWM on period to obtain current values and ; likewise, sampling at the start and end of the freewheeling period to obtain current values and .

[0065] S302, calculating the current differential profile. The current differential profile is composed of the current rising slope of the power tube on phase and the current falling slope of the freewheeling phase. Its calculation formula is as follows:

[0066] ;

[0067] ;

[0068] wherein, and are the phase current sampling values at the start and end of the PWM on period, respectively; and are the phase current sampling values at the start and end of the PWM freewheeling period, respectively; is the on duration of the PWM, whose value is determined by the PWM duty cycle and the PWM cycle , that is ; is the freewheeling duration of the PWM, that is . The calculated current slope (for example, one of or ) will be used in the operation of the state observer in step S40.

[0069] S303, extracting the asymmetry index. Using the calculated rising slope and falling slope, the current differential asymmetry index is further calculated, and in a specific embodiment, its calculation formula is:

[0070] ;

[0071] asymmetry index The physical basis of the extraction is that when the motor load increases, its phase current increases, causing the magnetic circuit of the motor stator core to tend to be saturated. The direct consequence of magnetic circuit saturation is that the effective inductance value of the motor winding occurs nonlinearly decreases. According to the motor voltage equation, the decrease of inductance will directly lead to an increase in the absolute value of the current rate of change Since such inductance changes are related to the size and direction of the current, the dynamic responses of the current rising process and the falling process appear differences, which are reflected by the amplitude difference between and Therefore, can be used as an indicator to quickly reflect the drastic changes in motor load, providing a key basis for the subsequent step S60 of feedforward control.

[0072] In step S40, in order to estimate the internal state variables of the motor that cannot be directly measured from the measurement signal containing noise, an extended Kalman filter (EKF) state observer is run. The effective operation of the observer is based on a mathematical model that can accurately describe the dynamic characteristics of the brushless DC motor.

[0073] The specific process of establishing the state space model of the motor system is as follows:

[0074] S401, define the state vector of the system. In order to comprehensively describe the electrical and mechanical dynamics of the motor, the state vector of the system is selected as:

[0075] ;

[0076] Among them, is the phase current of the currently conducting two phases; is the electrical angular velocity of the motor; is the electrical angle of the motor, that is, the angle between the rotor magnetic field and the axis of a certain phase winding of the stator.

[0077] S402, establish the continuous-time state space equation. The equation is composed of a set of differential equations that describe the physical laws of the motor.

[0078] The electrical subsystem equation of the motor describes the relationship between the voltage and current applied to the stator winding and the back electromotive force. Within any 60-degree commutation sector, two phase windings are conducting, and the equivalent line voltage equation is:

[0079] ;

[0080] Among them, ​is the line voltage applied between the two conducting phases, which is related to the DC bus voltage and the PWM duty cycle; and are the equivalent line resistance and line inductance, respectively, which are twice the phase resistance and phase inductance . In actual operation, the estimated values after line tuning in step S50 are used and ; is the line back EMF, whose amplitude is proportional to the rotor speed and whose waveform is related to the rotor position. In this embodiment, for the convenience of establishing a general model, the fundamental component of the line back EMF is approximated as a sine function: where is the back EMF coefficient of the motor.

[0081] The mechanical subsystem equation of the motor describes the torque balance relationship acting on the rotor:

[0082] ;

[0083] where, is the moment of inertia of the rotor; is the electromagnetic torque generated by the motor, which is related to the current and the back EMF, and can be expressed as where is the torque coefficient; is the load torque borne by the motor; is the viscous friction coefficient.

[0084] The kinematic relationship between the electrical angle and the electrical angular velocity of the motor is:

[0085] ;

[0086] The above equations are integrated to obtain the continuous-time nonlinear state equation describing the entire motor system , which is the basis for subsequent discretization and linearization of the extended Kalman filter algorithm; where, is the input vector of the system, mainly including the line voltage applied between the two conducting phases and the load torque borne by the motor .

[0087] After establishing the state space model of the motor system, the EKF algorithm performs optimal estimation on the state vector of the system through a recursive prediction-update cycle.

[0088] S403, discretization of the state equation and the observation equation. Since the EKF algorithm operates in the discrete time domain, the continuous-time state equation established in S401 needs to be converted into a discrete-time form. The first-order Euler method is adopted, and the control period T is used as the sampling period (i.e. PWM period ) is discretized to obtain the discrete-time state transition equation:

[0089] ;

[0090] where and are the state vectors at the current time and the last time , respectively.

[0091] The observation equation of the system describes the relationship between the measurable physical quantity and the state vector. In the present embodiment, the only directly measurable state quantity is the phase current . Therefore, the observation equation is:

[0092] ;

[0093] where is the measurement value at time , i.e. , is the actual phase current measurement value at the discrete time , which is obtained by synchronous acquisition and calibration in step S20; is the observation matrix, since only the first state quantity is observed, so ; is the measurement noise, which is assumed to be white noise following a Gaussian distribution.

[0094] S404, the prediction step of EKF is performed. This step is based on the optimal estimation value at the last time, to predict the state at the current time. The state prediction equation is:

[0095] ;

[0096] The error covariance prediction equation is:

[0097] ;

[0098] where is the state prediction value from time to time ; is the corresponding prediction error covariance matrix; is the optimal estimation error covariance matrix at the last time; is the Jacobian matrix obtained by linearizing the state transition function at ; is the Jacobian matrix The transpose of the matrix; It is the process noise covariance matrix, used to describe the uncertainty of the model itself, and is a diagonal matrix that needs to be tuned.

[0099] S405, Perform the EKF update step. This step utilizes the actual measurements at the current time. This is used to correct the predicted value and obtain the optimal estimate for the current moment.

[0100] First, calculate the Kalman gain. :

[0101] ;

[0102] in, It is the observation matrix The transpose of the matrix; It is the measurement noise covariance matrix, which is a scalar in this embodiment, representing the measurement noise variance of the current sensor, and is a parameter that needs to be tuned. It is the matrix inversion operator.

[0103] Then, the state is updated to obtain the optimal state estimate at the current moment. :

[0104] ;

[0105] The physical meaning of this equation is: the optimal estimate equals the predicted value plus a correction term. The magnitude of this correction term is determined by the Kalman gain. and measurement residuals ( (To be decided jointly)

[0106] Finally, update the error covariance matrix. :

[0107] ;

[0108] in, It is an identity matrix.

[0109] By cyclically executing S404 and S405 within each control cycle, the EKF observer can continuously output the optimal state estimation vector. .in, At discrete time The optimal phase current estimate; At discrete time The optimal electrical angular velocity estimate; At discrete time The optimal electrical angle estimate; It is a vector transpose operator. The electrical angles in this optimal state estimation vector. It has a much higher accuracy than the 60-degree resolution of Hall sensors, providing feedback signals for subsequent commutation control.

[0110] In step S50, to counteract the drift of motor parameters caused by changes in operating conditions (such as temperature) and to ensure the long-term accuracy of the EKF state observer model, the system uses the absolute position information provided by the Hall position sensor to perform periodic online calibration of the model. The core of this step is to extract the model prediction error.

[0111] The specific implementation process for extracting model prediction errors is as follows:

[0112] The S501 precisely captures the transition moments of Hall signals. The GPIO port on microcontroller 1 connected to the signal output of Hall position sensor 4 is configured as an external interrupt input, sensitive to both rising and falling edges of the signal. When the state of any Hall signal changes, an interrupt service routine is immediately triggered. This interrupt mechanism ensures that the system can accurately capture the transition moments of the Hall signals. Respond.

[0113] S502 determines the theoretical electrical angle corresponding to the Hall signal. Within the interrupt service routine, microcontroller 1 immediately reads the current state combination of the three Hall sensors (e.g., a three-bit binary number). A static Hall state-angle mapping table is pre-stored within the system. By querying this mapping table, the theoretical electrical angle uniquely corresponding to the current Hall state combination can be directly obtained. These theoretical angles are discrete; for example, for a standard three-phase motor, these angles are typically 0°, 60°, 120°, 180°, 240°, and 300°, which represent the boundaries of the six commutation sectors.

[0114] S503, calculate the model prediction error. Within the same interrupt service routine that captures the Hall transition and determines the theoretical angle, the system obtains the current time value from the EKF observer in step S40. Output real-time electrical angle estimate Then, the deviation between the estimated value and the theoretical value is calculated to obtain the model prediction error. :

[0115] ;

[0116] in, The model prediction error, the magnitude and sign of which reflect the degree to which the EKF model leads or lags behind the true position of the physical system; The estimated electrical angle of the EKF at the moment of Hall signal transition; This represents the theoretical electrical angle corresponding to this Hall effect transition.

[0117] In obtaining the model prediction error Subsequently, the system uses a parameter adaptive adjustment unit to process the parameters, thereby updating and correcting the resistance and inductance parameters in the motor model online.

[0118] S504 performs adaptive adjustment of the motor model's resistance parameters. This adjustment process utilizes a proportional-integral (PI) controller. The PI controller is based on the model prediction error calculated in S503. As input, its output is directly used to update the estimated resistance value. Its discretization update algorithm is specifically expressed as follows:

[0119] ;

[0120] ;

[0121] in, It is in the The estimated resistance value is updated after the Hall transition; It is in the The model prediction error calculated during the Hall transition; and These are the proportional gain and integral gain of the PI controller used for resistance regulation. These two parameters need to be experimentally tuned according to the specific motor system. and These are the accumulated integrals of the PI controller in this update and the last update, respectively.

[0122] The S505 performs adaptive adjustment of the inductance parameters of the motor model. Similar to resistance adjustment, this process also uses an independent PI controller with the same model prediction error. As input, update the estimated inductance value. Its discretization update algorithm is as follows:

[0123] ;

[0124] ;

[0125] in, It is in the The estimated inductance value is updated after the Hall effect transition; and These are the proportional gain and integral gain of the PI controller used for inductor regulation. and These are the integral accumulation items for the current and previous updates of the inductance-controlled PI controller.

[0126] In one specific embodiment, to prevent integral saturation and unreasonable drastic fluctuation of parameter values, upper and lower limits are set for the integral accumulation term in both PI controllers and , and reasonable value ranges are also set for the final output and .

[0127] Through the above steps, whenever the Hall sensor state jumps, the system will use this absolute position reference point to correct the motor model parameters relied on by the EKF observer. The closed-loop regulation mechanism formed in this way enables the model parameters to dynamically track the actual parameter changes of the motor, thereby ensuring the long-term accuracy and reliability of the EKF observer output angle. The updated parameters and will be used in the EKF operation of step S40 in all subsequent control cycles.

[0128] In step S60, the system fuses feedback and feedforward information from different modules to calculate a dynamic target commutation phase angle that can adapt to the current working condition. This target commutation angle is the precise angle reference for triggering the next commutation.

[0129] The specific process of calculating the dynamic target commutation phase angle is as follows:

[0130] S601, Synthesize the target commutation phase angle. The final target commutation angle is composed of three parts: the basic sector angle , the feedback-based lead angle , and the feedforward-based lead angle . Its synthesis formula is:

[0131] ;

[0132] S602, Determine the basic sector angle. The basic sector angle is a discrete fixed value that defines the standard theoretical position of the next commutation event. The value of this angle is only related to the current commutation sector. For example, when the system is in the first sector, the basic sector angle corresponding to the next commutation point is 60°; when in the second sector, the basic sector angle corresponding to the next commutation point is 120°, and so on. The determination of this value is a well-known technique in the art and will not be repeated here.

[0133] S603, Calculate the feedback lead angle. The feedback lead angle is to compensate the current response lag caused by the motor winding inductance, and to actively advance the phase in order to obtain greater output torque at high speed. The size of this angle is closely related to the real-time operating state of the motor. In a specific embodiment, is a function of the estimated electrical angular velocity of the motor and the estimated phase current . This function can be implemented by a pre-calibrated two-dimensional look-up table (LUT). The microcontroller 1 takes as input the and output by the EKF observer in step S40, and through table lookup and interpolation operations, obtains the most suitable feedback advance angle under the current operating conditions.

[0134] S604, calculate the feedforward advance angle. The role of the feedforward advance angle is to respond to the system dynamics caused by sudden changes in load, providing rapid compensation. This angle is directly related to the current differential asymmetry index calculated in step S30. In an embodiment, the two are in a linear proportional relationship:

[0135] ;

[0136] wherein, is the feedforward gain coefficient, which needs to be experimentally calibrated according to the specific motor and load characteristics. Its physical meaning is to map the change in the current differential profile to the required angle compensation.

[0137] Through the calculations of S601 to S604, the system combines the feedback control reflecting the steady-state operating characteristics of the motor with the feedforward control reflecting the transient changes in the load of the motor, generating a target commutation angle that can both guarantee optimal steady state and quickly respond to dynamic changes.

[0138] In step S70, the system compares the real-time estimated electrical angle with the dynamically calculated target commutation angle, and decides whether to perform commutation operation according to the comparison result, to update the PWM output state of the driving motor.

[0139] The specific implementation process of this step is as follows:

[0140] S701, make a judgment on the commutation condition. In each real-time processing task of the control cycle, the system compares the current optimal electrical angle estimate value output by the EKF observer in step S40 with the dynamically calculated target commutation phase angle in step S60. The preset commutation condition is:

[0141] ;

[0142] The satisfaction of the condition indicates that the motor rotor has reached or passed the ideal commutation position point in the current operating condition.

[0143] S702, performing commutation operation. When the commutation condition in S701 is satisfied, the microcontroller 1 immediately executes the commutation instruction. The specific execution of the instruction is to update the related register configuration of the PWM generation and inverter drive module 50 to change the control signal output to the gate drive circuit of the three-phase inverter bridge 2. This operation will cause the switch state combination of the six power tubes in the three-phase inverter bridge 2 to change, so as to switch the excitation current of the stator winding to the corresponding phase winding in the next sector. The specific switch state combination switching follows the standard six-step commutation logic table, which is a well-known technology in the art and will not be described here. After the commutation operation is completed, the system updates the basic sector angle to the boundary angle of the next sector.

[0144] S703, maintaining the current commutation state. If the commutation condition in S701 is not satisfied, the system does not perform any commutation operation. The PWM generation and inverter drive module 50 will maintain the current output state, the switch combination of the three-phase inverter bridge 2 remains unchanged, and the motor continues to operate in the current sector. The system will directly end the judgment of the current control period and enter the next control period to repeat the comparison process described above.

[0145] To further verify the actual effect of the method proposed in the present application, a motor control experimental platform is built, and the performance of the embodiment of the present application and the conventional comparison embodiment based on the Hall sensor signal and superimposed fixed speed lookup table lead angle are tested and compared.

[0146] The hardware configuration of the experimental platform is as follows: the main control unit is a microcontroller; the three-phase inverter bridge is composed of a gate drive chip and six power MOSFETs with a rated value of 100V / 80A; the test motor is a 400W rated power, 48V rated voltage, 4-pole direct current brushless motor; the load is provided by a magnetic powder brake for simulating different load conditions. All experimental data are collected by an oscilloscope and a power analyzer.

[0147] The hardware of the experimental platform used in the comparison embodiment includes a microcontroller, a three-phase inverter bridge, a test motor, and a load, which is exactly the same as the embodiment of the present application. The specific implementation of the control method is as follows:

[0148] 1. Position sensing and commutation triggering: The control core of this comparative example is based on directly using the signals of the Hall position sensor 4 to make commutation decisions. The GPIO port of the microcontroller 1 is configured in external interrupt mode, and every time a jump in the Hall signal state is detected (i.e. a "Hall event"), a commutation-related calculation and execution task is triggered. The basic resolution of its position sensing is limited to the 60-degree electrical angle interval of the Hall sensor.

[0149] 2. Speed calculation: In the interrupt service routine triggered by each Hall event, the controller first calculates the average electrical angular speed of the motor in the 60-degree sector just passed based on the time interval between the current Hall event and the last Hall event, using the time stamps captured by the high-precision timer .

[0150] 3. Commutation lead angle compensation: To compensate for the phase lag at high speeds, this comparative example uses a fixed lead angle compensation strategy based on a speed lookup table. The specific steps are as follows:

[0151] Using the average electrical angular speed calculated in step 2 as input, a one-dimensional lookup table that has been calibrated offline and solidified in the program memory is queried.

[0152] Through table lookup and linear interpolation operations, a fixed lead angle compensation value corresponding to the current speed is obtained.

[0153] This lead angle is converted into a specific time delay , and the calculation formula is: .

[0154] After detecting the occurrence of a Hall event, the controller does not immediately commutate, but starts a hardware timer, and after a precise delay , an interrupt is generated to perform the final commutation operation, i.e. update the PWM output state to the next sector.

[0155] In subsequent experimental comparisons, the "comparative example" mentioned refers to the control method implemented in the above manner.

[0156] Referring to the attached Figure 4 , Figure 4 is the efficiency comparison curve of the inventive example and the comparative example under different loads. In the experiment, the motor is made to run stably at 80% of the rated speed, and the load is gradually increased from 20% of the rated load to 100% of the rated load. As Figure 4As shown, the operating efficiency of the embodiment of the present application (solid line in the figure) is higher than that of the comparative example (dashed line in the figure) at all test load points. Especially in the medium and high load region, the efficiency can be increased by more than 3% due to the adaptive commutation compensation of the present application, which enables the motor to always work in the optimal field-oriented state, thereby significantly reducing the energy loss of the system.

[0157] Referring to the accompanying drawings Figure 5 , Figure 5 is a comparison chart of phase current waveforms of the embodiment and the comparative example when the motor operates at 80% rated speed and 75% rated load. Among them, Figure 5 (a) is the phase current waveform of the comparative example, Figure 5 (b) is the phase current waveform of the embodiment of the present application. As can be clearly seen from the figure, the current waveform of the comparative example has obvious distortion, while the current waveform of the embodiment of the present application is smoother and closer to the ideal trapezoidal waveform. This shows that the precise commutation control of the present application effectively suppresses the current overshoot and oscillation in the commutation process, thereby reducing the torque ripple, vibration and electromagnetic noise of the motor.

[0158] Referring to the accompanying drawings Figure 6 , Figure 6 is a comparison chart of motor speed responses of the embodiment and the comparative example under the condition of load mutation. In the experiment, the motor first operates stably at 70% rated speed and 40% rated load, and at t1, the load is instantaneously increased to 90% rated load by a magnetic powder brake. As shown in Figure 6 , after suffering the same load shock, the motor speed (dashed line in the figure) using the comparative example has a large drop of about 250 RPM and takes about 140 ms to recover to stable. In comparison, the motor (solid line in the figure) using the embodiment of the present application, thanks to the feedforward compensation mechanism based on the current differential profile, can quickly predict and compensate for load changes, and its speed drop is only about 90 RPM, and the time to recover to stable is shortened to about 55 ms. The experimental results prove that the present application can significantly enhance the robustness of the motor control system under dynamic load.

[0159] Although embodiments of the present application have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and alterations can be made thereto without departing from the principles and spirit of the present application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for sector commutation compensation based on DC brushless motor control, characterized in that, The method comprises the following steps: In each PWM control cycle, the phase current signal of the DC brushless motor is synchronously collected; A extended Kalman filter state observer is run to estimate the motor electrical angle in real time according to the phase current signal and the currently effective motor model parameters; When the signal of the Hall position sensor is detected to jump, the motor model parameters are adjusted online according to the error between the Hall theoretical angle at the moment and the motor electrical angle; Based on the phase current signal, the current differential profile is calculated, which includes calculating the current rising slope in the power tube conduction stage and the current falling slope in the freewheeling stage; and based on the amplitude difference between the rising slope and the falling slope, the asymmetry index representing the load change is extracted; The motor electrical angle and the asymmetry index are fused to calculate the dynamic target commutation phase angle; The motor electrical angle and the target commutation phase angle are compared, and when the preset commutation condition is met, the commutation operation is performed.

2. The DC brushless motor control sector commutation compensation method of claim 1, wherein, The step of calculating the dynamic target commutation phase angle comprises: The basic sector angle, the lead angle based on state feedback and the feedforward lead angle based on the asymmetry index are linearly superimposed to obtain the target commutation phase angle.

3. The DC brushless motor control sector commutation compensation method of claim 2, wherein, The feedforward lead angle and the asymmetry index are in linear proportional relationship.

4. The DC brushless motor control sector commutation compensation method of claim 1, wherein, The step of adjusting the motor model parameters online comprises: At the moment when the signal of the Hall position sensor jumps, the motor electrical angle output by the extended Kalman filter state observer is obtained; The model prediction error between the motor electrical angle and the Hall signal jump corresponding Hall theoretical angle is calculated, and the motor model parameters are updated based on the model prediction error.

5. The DC brushless motor control sector commutation compensation method of claim 4, wherein, The motor model parameters are updated by using one or more proportional-integral controllers with the model prediction error as input.

6. The DC brushless motor control sector commutation compensation method of claim 4, wherein, The motor model parameters include estimated resistance value and estimated inductance value.

7. The DC brushless electric motor control sector commutation compensation method of claim 1 wherein, The state vector estimated by the extended Kalman filter state observer comprises: Phase current, motor electrical angular velocity and the motor electrical angle.

8. The DC brushless electric motor control sector commutation compensation method of claim 1 wherein, The method further comprises the step of initializing the extended Kalman filter state observer, and the initialization step comprises: The initial error covariance matrix is set, wherein the variance set for the initial angle estimation value is greater than the variance set for the initial phase current estimation value and the initial angular velocity estimation value.

9. The DC brushless electric motor control sector commutation compensation method of claim 1 wherein, The collection of the phase current signal is performed in the stable interval of the PWM cycle to avoid the electromagnetic interference introduced by the switching transient process of the power tube in the three-phase inverter bridge.

Images