Permanent magnet synchronous motor sensorless current reconstruction method suitable for electric aircraft

Abstract

The invention discloses a permanent magnet synchronous motor sensorless current reconstruction method suitable for an electric aircraft, which comprises the following steps: switching modes of a motor control system of the electric aircraft through a fault detection module and a switch, the modes comprising normal operation and fault tolerance; whether a current sensor fails or not is detected in real time through a fault detection module, if the current sensor fails, a direct-current bus current signal is obtained through single-resistor sampling, another phase of current and rotor position and rotating speed information are reconstructed in real time through an EKF method, third phase of current is calculated in combination with the Kirchhoff's current law, and the third phase of current is calculated. And position-sensorless vector control is realized. According to the invention, by combining the single-resistor sampling technology and the EKF algorithm, the three-phase current, the rotor position and the rotating speed of the motor can be reconstructed at the same time only through a single sampling resistor, the feedback current of the motor can be normally obtained under the condition that the current sensor fails, and the safe flight of an airplane is ensured.

Description

Technical Field

[0001] This invention relates to the field of motor control technology, and more specifically to a sensorless current reconfiguration method for permanent magnet synchronous motors suitable for electric aircraft. Background Technology

[0002] Permanent magnet synchronous motors (PMSMs) are widely used in electric aircraft, electric vehicles, and industrial drives due to their high efficiency, high power density, and excellent dynamic performance. In electric aircraft propulsion systems, the control accuracy and reliability of PMSMs are crucial. Traditional vector control (FOC) typically relies on current sensors (such as Hall effect sensors or shunt resistors) and position sensors (such as encoders or resolvers) to obtain three-phase current and rotor position information. However, in harsh environments such as aviation, sensors may fail due to vibration, high temperatures, or electromagnetic interference, leading to control shutdown and directly threatening flight safety. Therefore, researching how to ensure flight safety in the event of current sensor failure is of great significance. Summary of the Invention

[0003] Therefore, the purpose of this invention is to provide a sensorless current reconfiguration method for permanent magnet synchronous motors in electric aircraft, so as to improve the operational safety of electric aircraft.

[0004] The technical solution provided by this invention is: a sensorless current reconfiguration method for permanent magnet synchronous motors in electric aircraft, comprising:

[0005] Through the fault detection module and switch The mode switching of the electric aircraft motor control system includes normal operation and fault tolerance.

[0006] The fault detection module detects whether the current sensor is malfunctioning in real time. If the current sensor is malfunctioning, the DC bus current signal is obtained by sampling with a single resistor. The EKF method is used to reconstruct the current of another phase and the rotor position and speed information in real time. Then, the third phase current is calculated by combining Kirchhoff's current law to realize sensorless vector control.

[0007] Preferably, the switch Set to 1 for normal operation mode; the current acquisition module collects the three-phase current of the motor. , , The current is converted into a two-phase stationary coordinate system by Clark transformation. , Then, the two-phase rotating coordinate system current is obtained through the Park transformation. , , as current closed-loop feedback;

[0008] The voltage acquisition module acquires two-phase voltages. , Input the EKF module, which outputs the rotor position based on the motor model. and rotational speed This provides position and velocity feedback for Park transformation and velocity closed loop.

[0009] Preferably, the switch Setting it to 2 enables fault-tolerant mode. The fault detection module monitors for abnormal or lost current signals from the current acquisition module to determine if the current acquisition module has failed. If the current sensor fails, single-resistor sampling is initiated. The single-resistor acquisition module is placed on the DC side of the inverter and acquires the DC bus current. Combined with the two-phase voltage acquired by the voltage acquisition module , The PWM drive signal is input to the EKF module to achieve accurate reconstruction of the three-phase current;

[0010] The EKF module realizes the three-phase current of the motor through state estimation. , , The reconfiguration process involves the reconfiguration current being transformed by Clark and Park to replace the failed current feedback, and the EKF still outputs the rotor position. and rotational speed .

[0011] Preferably, in both normal operation and fault-tolerant modes, the speed closed loop uses a reference rotational speed. With feedback speed The deviation is adjusted by PI to generate a q-axis current reference value. Combined with d-axis current reference value , and feedback current , After comparison, the voltage is generated by PI regulation. , The result obtained by inverse Park transform , The input SVPWM module generates drive signals to control the inverter, and the output three-phase AC power drives the permanent magnet synchronous motor.

[0012] Preferably, the single-resistor acquisition module acquires the DC bus current. Combined with the two-phase voltage acquired by the voltage acquisition module , The PWM drive signal is input to the EKF module to achieve accurate reconstruction of the three-phase current, including: establishing a mapping relationship between the sampling resistor current and the switch state based on the one-to-one correspondence between the bus current and the voltage vector, wherein the positive direction of the current is defined according to the motor convention;

[0013] When the 100 vector is applied, the bus current is: The specific relationship is shown in the following formula:

[0014]

[0015] In the formula, , and These are the three-phase currents. , and These are the activation statuses of the three-phase bridge arms. It is the bus current;

[0016] The relationship between the three-phase currents of a motor can be obtained from Kirchhoff's laws:

[0017]

[0018] The switch state and phase current are obtained through the one-to-one correspondence between the bus current and the voltage vector. During the first sampling, the bus current flows through the current... During the second sampling, the current flowing through the busbar was - Similarly, by combining EKF and voltage signals, another current can be reconstructed;

[0019] The last current term was derived based on Kirchhoff's laws, and the three-phase current was successfully reconstructed by time-sharing sampling within one PWM cycle.

[0020] Preferably, the construction of the EKF modular equations includes:

[0021] Using a recursive calculation method, based on the optimal estimate at the current moment, and combining the system dynamic model, state equations, observation equations, and noise statistics, the system accurately predicts the motor speed and rotor position at the next moment.

[0022] The Taylor expansion method is used to approximate the linearization of the nonlinear system, and the continuous nonlinear control system is described by the following equations;

[0023]

[0024] Equation (10) is the state-space equation of the system, and equation (11) is the measurement equation of the system, where process noise... and observation noise They still follow a Gaussian white noise distribution, with corresponding covariance matrices Q and R, respectively;

[0025] make:

[0026]

[0027] when When the size is sufficiently small, performing a Taylor expansion at the optimal state estimation point, retaining the first-order terms and discarding the higher-order terms, yields the approximately linearized system state-space equations and measurement equations.

[0028]

[0029] The state estimate is:

[0030]

[0031] Subtracting the corresponding values ​​from the four equations above, we get:

[0032]

[0033] make and Then F and H are Jacobian determinants, as shown in equations (15) and (16) respectively:

[0034]

[0035] Then, by combining equation (15) with equations (16) and (17), we can obtain:

[0036]

[0037] Once the linearity condition is met, the above two equations are discretized.

[0038]

[0039] In the formula, represents the system's transition matrix under discrete conditions. When the overall frequency of the algorithm is sufficiently high, the following relationship holds:

[0040]

[0041] In the formula, It is the identity matrix. The algorithm execution cycle;

[0042]

[0043] The Kalman filter equation for the error between the estimated and true values ​​is derived through the above derivation, and its specific form is shown below:

[0044]

[0045] beg and The ultimate goal is to ensure that the filtered estimated value remains consistent with the true value.

[0046] From this, we can derive the five formulas for the nonlinear, discrete extended Kalman filter as follows:

[0047] .

[0048] Preferably, the fault detection module determines that the current acquisition module has failed by monitoring for abnormal or lost current signals.

[0049] The fault detection module compares the given current with the actual value collected by the current sensor. , , Comparison:

[0050]

[0051] Provide three-phase current to the motor. This represents the actual three-phase current of the motor; To set a threshold, if the current deviation of a certain phase is... If the current sensor fails, the current sensor for that phase is deemed to be faulty. Set to 2; if all three-phase currents satisfy If the current sensor is normal, then the current sensor is considered to be functioning correctly. Set to 1.

[0052] Preferably, in the single-resistance acquisition module, the duration of the effective vector is greater than the minimum sampling time. ; To obtain the precise phase current of the motor under different switching states:

[0053]

[0054] In the formula, Indicates dead time. The duration of the transient process during the switching of the transistor state ( ), This represents the total time required for the microprocessor sampling module to convert the analog signal into a digital signal from sampling.

[0055] Preferably, the EKF module outputs the rotor position based on the motor model. and rotational speed It provides position and velocity feedback for Park transformation and velocity closed loop, including:

[0056] To predict the state vector, the previous optimal estimate is used as the independent variable, and the non-linearized state transition matrix is ​​used to... The increment within the time period is calculated and added to the value of the previous best estimate to obtain the prior estimate of the current prediction.

[0057]

[0058] Using the already discretized state transition matrix To estimate the covariance matrix, we obtain the prior covariance matrix:

[0059]

[0060] Calculate the Kalman gain:

[0061]

[0062] The state vector estimate is updated using Kalman gain to obtain the posterior estimate, which is also the optimal estimate.

[0063]

[0064] Update the covariance matrix using Kalman gain:

[0065]

[0066] Preferably, the EKF module realizes the three-phase current of the motor through state estimation. , , The reconfiguration process involves the reconfiguration current being transformed by Clark and Park to replace the failed current feedback, and the EKF still outputs the rotor position. and rotational speed ,include:

[0067] The current in the three-phase stationary coordinate system (abc) obtained from EKF is converted to a two-phase orthogonal stationary coordinate system. ):

[0068]

[0069] After obtaining the mathematical model in the two-phase stationary coordinate system, the model is transformed into a model in the synchronous rotating coordinate system (dq) using the Park transformation:

[0070]

[0071] In the formula The angle between the synchronous rotating coordinate system and the two-phase stationary coordinate system is the electrical angle through which the rotor rotates.

[0072] This invention proposes a sensorless current reconstruction method for permanent magnet synchronous motors in electric aircraft. By combining single-resistor sampling with extended Kalman filtering (EKF) technology, current reconstruction under sensorless operating conditions is achieved. This method can identify current faults in real time when a current sensor failure is detected and reconstruct the current signal required by the motor vector control system. This ensures the continuous and stable operation of the motor vector control system, effectively avoiding system downtime caused by current sensor failure, thereby ensuring no loss of aircraft power and improving the operational safety of electric aircraft. Attached Figure Description

[0073] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0074] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0075] Figure 1 This is a block diagram of the motor control system of the electric aircraft of the present invention when there is no fault.

[0076] Figure 2 This is a block diagram of the motor control system of the electric aircraft of the present invention when there is no fault.

[0077] Figure 3 This is a schematic diagram illustrating the single-resistor sampling principle of the present invention;

[0078] Figure 4 This is a schematic diagram of the current flow path under the action of 100 vectors according to the present invention;

[0079] Figure 5 This is the basic structure of the Kalman filter of the present invention;

[0080] Figure 6 This is a flowchart of the Kalman filter process of the present invention;

[0081] Figure 7 This is a flowchart illustrating the single-resistor sampling and EKF-based sensorless and current-free control process of the present invention. Detailed Implementation

[0082] The present invention will be further explained below with reference to specific implementation schemes, but this explanation does not limit the scope of the invention.

[0083] Current sensors are key components in the motor control system of electric aircraft, responsible for providing real-time three-phase current feedback signals for the vector control algorithm of permanent magnet synchronous motors. To address the problem of current sensor failure, various sensorless control technologies based on software algorithms have been proposed in recent years. For example, current estimation methods based on sliding mode observers utilize the mathematical model of the motor to achieve sensorless control. However, these methods are sensitive to changes in motor parameters and are prone to system misalignment due to sudden changes in motor parameters. Traditional observers and filtering algorithms require high computational resources from the controller, and as the motor ages, its resistance, inductive reactance, and other parameters change, leading to a decrease in estimation accuracy. Noise and external interference can also affect the stability and accuracy of the estimation, and the high computational complexity makes real-time implementation in embedded systems difficult. This patent proposes a sensorless full-state observation method based on single-resistor sampling combined with EKF (Electronic Kinematic Filter). It only requires a single sampling resistor to detect the DC bus current, and uses an optimized EKF algorithm to simultaneously and accurately estimate two-phase current, rotor position, and speed. Then, it combines Kirchhoff's laws to calculate the third-phase current, achieving sensorless FOC (Field-Oriented Control) control. Compared to existing technologies, EKF effectively suppresses the effects of measurement noise and motor parameter variations through adaptive adjustment of noise covariance, ensuring stable observation over a wide speed range. Furthermore, it can be implemented in real-time on an embedded controller, meeting the demands of high-performance control. This method first detects current sensor failure through a fault diagnosis module. If a current sensor failure occurs, the current can be reconstructed through single-resistor sampling and EKF, ensuring that the motor does not stop and the aircraft continues to fly normally even in the event of a current sensor failure, thus improving the system's reliability and safety.

[0084] Traditional vector control relies on multiple current and position sensors, which are prone to failure and serious accidents under harsh operating conditions. This invention innovatively combines single-resistor sampling technology with the EKF algorithm, enabling the simultaneous reconstruction of the motor's three-phase current, rotor position, and speed using only a single sampling resistor. This ensures that the motor's feedback current can be obtained normally even if the current sensor fails, guaranteeing the safe flight of the aircraft and achieving sensorless control of the permanent magnet synchronous motor without current or position sensors.

[0085] This invention proposes a current reconstruction and fault-tolerant control method for electric aircraft motor control that integrates single-resistor sampling and EKF (Extended Kernel Factor) analysis. The electric aircraft motor control system mainly consists of a current acquisition module, a single-resistor acquisition module, a voltage acquisition module, a fault detection module, an EKF module, a Clark converter module, a Park converter module, a PI control module, an SVPWM module, an inverter module, and a permanent magnet synchronous motor, as shown in the figure. Figure 1 , Figure 2 As shown.

[0086] The system has two modes: normal operation and fault tolerance. This is achieved through a fault detection module and switches. To switch over:

[0087] Normal operating mode ( Set to 1, such as Figure 1 At this time, the three-phase current of the motor is collected by the current acquisition module. , , The current is converted into a two-phase stationary coordinate system by Clark transformation. , Then, the two-phase rotating coordinate system current is obtained through the Park transformation. , This serves as a current closed-loop feedback; simultaneously, the voltage acquisition module acquires the two-phase voltage. , Input the EKF module, which outputs the rotor position based on the motor model. and rotational speed This provides position and velocity feedback for Park transformation and velocity closed loop.

[0088] Fault-tolerant mode ( Set to 2, such as Figure 2 When the fault detection module detects a failure in the current acquisition module (such as an abnormal or lost current signal), the system automatically switches to fault-tolerant mode and initiates the single-resistor sampling scheme. At this time, the DC bus current is acquired through the single-resistor acquisition module on the DC side of the inverter. Combined with the two-phase voltage acquired by the voltage acquisition module , The PWM drive signal and the EKF module are input together; the EKF module realizes the motor's three-phase current through state estimation. , , The reconfiguration process involves replacing the failed current feedback with a reconfigured current after Clark and Park transformations, while the EKF still outputs the rotor position. and rotational speed This ensures the proper functioning of vector control.

[0089] In both modes, the speed closed loop uses a reference speed. With feedback speed The deviation is adjusted by PI to generate a q-axis current reference value. Combined with d-axis current reference value , and feedback current , After comparison, the voltage is generated by PI regulation. , The result obtained by inverse Park transform , The input SVPWM module generates drive signals to control the inverter, and outputs three-phase AC power to drive the permanent magnet synchronous motor, achieving stable and efficient operation of the motor.

[0090] The single-resistor sampling method used in this invention is based on the following principle:

[0091] Generally, PWM inverters employ a three-phase bridge structure, with each arm containing two power switching elements. Traditional solutions use Hall sensors or current transformers in each phase circuit to measure the three-phase current, while single-resistor sampling technology places the sole current sensing device on the DC bus, using high-frequency sampling within the PWM cycle to achieve accurate reconstruction of the three-phase current (see...). Figure 3 ).

[0092] This solution not only effectively reduces the cost and size of the control system, but also avoids the measurement asymmetry problem inherent in traditional multi-sensor solutions. The working principle of the single-resistor current reconstruction technology will be explained in detail below.

[0093] The DC bus current path changes under different switching states of the inverter. Since each basic voltage vector corresponds to a specific current path, there is a one-to-one correspondence between the bus current and the voltage vector. By studying the current path when each basic vector is in effect, the mapping relationship between the sampling resistor current and the switching state shown in Table 1 can be established, where the positive current direction is defined according to the motor convention.

[0094] Taking vector 100 as an example ( =1、 =0、 =0), such as Figure 4 As shown, the bus current is when a 100 vector is applied. The specific relationship is shown in the following formula:

[0095]

[0096] In the formula, , and These are the three-phase currents. , and These are the activation statuses of the three-phase bridge arms. It is the bus current.

[0097] The relationship between the three-phase currents of a motor can be obtained from Kirchhoff's laws:

[0098]

[0099] To solve for three unknowns, at least three mathematical equations are needed. Therefore, at least two current equations are obtained. Combining these with Kirchhoff's laws, the complete three-phase current is derived. Because the bus current does not contain current information during zero vector action, it is necessary to sample the current of any term during the effective vector action period. As shown in Table 1, during the first sampling, the bus current flows through the current... During the second sampling, the current flowing through the busbar was - By analogy, another current is reconstructed by combining EKF and voltage signals, and finally the last current is derived according to Kirchhoff's laws. The three-phase current is successfully reconstructed by time-division sampling within one PWM cycle.

[0100] Table 1 Switching Status and Phase Current

[0101]

[0102] The control principle of EKF involved in this implementation plan is as follows:

[0103] The theoretical basis of Kalman filtering can be traced back to the linear minimum variance estimation method. Compared with other observers, this algorithm has significant advantages in suppressing system noise and measurement interference, effectively enhancing the robustness of the control system. Its basic principle is to dynamically correct the system model using actual measurement data; when the deviation between the model output and the actual measurement approaches zero, the optimal state estimate is obtained. This algorithm employs a recursive calculation method, using the optimal estimate at the current moment as a benchmark, combined with the system dynamic model, state equations, observation equations, and noise statistics, to achieve accurate prediction of the state variables at the next moment. In the field of permanent magnet synchronous motor control, these state variables typically refer to motor speed and rotor position. Figure 5 A block diagram illustrating the basic working principle of a Kalman filter is shown.

[0104] A discrete linear control system can typically be represented by the following equation:

[0105]

[0106] Equation (3) is the system state equation, and equation (4) is the system measurement equation. , , These correspond to the state variables, output variables, and input variables in the control system, respectively. and Let be the system noise and the measurement noise, respectively. Both of them conform to the characteristics of Gaussian white noise, follow a normal distribution, and are independent of each other. The covariance estimation matrices for the system noise and the measurement noise are Q and R, respectively, where Q is a non-negative definite matrix and R is a positive definite matrix. Let be the state transition matrix of the system. For the input matrix, This is the output matrix.

[0107] The Kalman filter consists of two main steps: First, it uses the state equation to predict the current estimated value based on the data from the previous time step; this is generally called the prior estimate, denoted as $\begin{p}{\infty}$. And the prior covariance matrix is ​​denoted as The second step is to correct the prior value using the measured value to obtain the posterior estimate. and posterior covariance matrix In the specific implementation process, it is broken down into five steps: predicting prior estimates, predicting errors, calculating Kalman gain, correcting prior estimates using measurement results, and updating the optimal estimate error, as shown in equations (5), (6), (7), (8), and (9), respectively.

[0108]

[0109] In the formula, This is the system state transition matrix.

[0110]

[0111] In the formula, This is the system noise matrix.

[0112]

[0113] In the formula, R is the Kalman gain, and R is the measurement noise matrix.

[0114]

[0115] The specific process is shown in Figure 6.

[0116] The above is the basic principle of Kalman filtering, but it is only applicable to linear control systems. Extending it to nonlinear control systems, namely the PMSM control system in this paper, becomes its extended version EKF.

[0117] To enable the Kalman filter to be applied to nonlinear systems, the Taylor expansion method is commonly used to approximate the nonlinear system as linear. A continuous nonlinear control system can be described by the following equation.

[0118]

[0119] Equation (10) is the state-space equation of the system, and equation (11) is the measurement equation of the system. The definitions of the parameters in these two equations are basically the same as those described in the pre-Kalman filter control system. The main difference is that this system has nonlinear characteristics, so all parameters change with time. Among them, process noise and observation noise They still follow a Gaussian white noise distribution, with corresponding covariance matrices Q and R, respectively.

[0120] make:

[0121]

[0122] when When the size is sufficiently small, performing a Taylor expansion at the optimal state estimation point, retaining the first-order terms and discarding the higher-order terms, yields the approximately linearized system state-space equations and measurement equations.

[0123]

[0124] The state estimate is:

[0125]

[0126] Subtracting the corresponding values ​​from the four equations above, we get:

[0127]

[0128] make and Then F and H are Jacobian determinants, as shown in equations (15) and (16) respectively.

[0129]

[0130] Then, by combining equation (14) with equations (15) and (16), we can obtain:

[0131]

[0132] Once the linearity condition is met, since data in actual microcontroller systems is acquired and used discretely, but when the frequency is high enough, it can achieve approximately continuous accuracy, the above two equations are discretized.

[0133]

[0134] In the formula, represents the system's transition matrix under discrete conditions. When the overall frequency of the algorithm is sufficiently high, the following relationship holds:

[0135]

[0136] In the formula, It is the identity matrix. This represents the algorithm's execution cycle.

[0137]

[0138] The Kalman filter equation for the error between the estimated and true values ​​can be obtained through the above derivation, and its specific form is shown below:

[0139]

[0140] The ultimate goal of Kalman filtering is to ensure that the filtered estimated value remains consistent with the true value. Therefore, we need to find... and From this, we can derive the five formulas for the nonlinear, discrete extended Kalman filter as follows:

[0141]

[0142] When designing filters in practice, the following key aspects need to be paid special attention to:

[0143] (1) Reasonable setting of noise matrix: The values ​​of system noise matrix Q and measurement noise matrix R usually need to be repeatedly adjusted based on experience. If the noise matrix is ​​set improperly, the estimated result of EKF may deviate from the true value, thereby affecting the stability of the entire control system.

[0144] (2) Setting the initial state: The initial state of the filter needs to be given in advance, including the initial state variables. And its covariance matrix P.

[0145] (3) Optimization of computation frequency: When using EKF, the computation frequency should be increased as much as possible to ensure the state increment. It is small enough, which is an important condition for ensuring the effectiveness of linearization approximation.

[0146] The aforementioned fault detection module will compare the given current with the actual value collected by the current sensor. , , Comparison:

[0147]

[0148] Provide three-phase current to the motor. This represents the actual three-phase current of the motor; To set a threshold, if the current deviation of a certain phase is... If the current sensor fails, the current sensor for that phase is deemed to be faulty. Set to 2; if all three-phase currents satisfy If the current sensor is normal, then the current sensor is considered to be functioning correctly. Set to 1.

[0149] The aforementioned single-resistor sampling module, under ideal conditions, allows for negligible sampling process and analog-to-digital conversion time, enabling effective sampling even with a very short effective voltage vector duration. However, in practical motor control systems, due to factors such as switching transients and analog-to-digital conversion time, the effective vector duration must be greater than the minimum sampling time to achieve stable current sampling. :

[0150]

[0151] In the formula, Indicates dead time. The duration of the transient process during the switching of the transistor state ( ), This represents the total time required for the microprocessor sampling module to convert the analog signal into a digital signal from sampling. This ensures that the accurate phase current of the motor can be obtained under different switching states (as shown in Table 1).

[0152] The above-mentioned EKF-based current and position reconstruction module

[0153] Traditional EKF observers use a PMSM mathematical model in the abc coordinate system to establish the state-space equations. The voltage equation for a surface-mounted PMSM in the abc coordinate system is:

[0154]

[0155] , , This refers to the three-phase voltage of the motor; , , It is a three-phase current; Stator resistance; Stator inductance (mutual inductance ignored); , , It is a three-way back electromotive force.

[0156] The back electromotive force is determined by the change in the permanent magnet flux linkage with the rotor position, and is usually expressed as:

[0157]

[0158] The back electromotive force coefficient; Electric angular velocity; This is the electric angle position.

[0159] Substituting the above equation into the voltage equation, and using Kirchhoff's laws, we can transform it into a two-term current equation:

[0160]

[0161] Besides the current equation

[0162]

[0163] The state-space equations of the PMSM control system are as follows:

[0164]

[0165] In the formula: It is a state vector; It is the input vector; , or (Depending on the input PWM signal) is the output vector; , These are the system noise and the measurement noise, respectively, both of which are Gaussian white noise. B is the input matrix, with the following specific form:

[0166]

[0167] The general expression for H is derived from PWM (or gate signal):

[0168]

[0169] in:

[0170]

[0171] If the measurement is H=[1,0,0,0]; if the measurement is H = [0,1,0,0]; if the measurement is... H = [-1, -1, 0, 0], because .

[0172] From equation (28), we can see that for:

[0173]

[0174] System noise Noise measurement The corresponding covariances Q and R are shown below:

[0175]

[0176] Discretize the state space of the PMSM:

[0177]

[0178] definition For Jacobian determinants:

[0179]

[0180] The state transition matrix is ​​shown below:

[0181]

[0182] The specific workflow of EKF is as follows:

[0183] (1) Predict the state vector, using the previous optimal estimate as the independent variable, and use the non-linearized state transition matrix to... The increment within the time period is calculated and added to the previous optimal estimate to obtain the prior estimate for this prediction. The non-linearized state transition matrix is ​​used for more accurate estimation, since higher-order terms are discarded after Taylor expansion, introducing some error.

[0184]

[0185] (2) Use the already discretized state transition matrix To estimate the covariance matrix, we obtain the prior covariance matrix:

[0186]

[0187] (3) Calculate the Kalman gain:

[0188]

[0189] (4) Update the state vector estimate using Kalman gain to obtain the posterior estimate, which is also the optimal estimate:

[0190]

[0191] (5) Update the covariance matrix using Kalman gain:

[0192]

[0193] The complete EKF design is now finished. During operation, it utilizes the motor's own parameters and the information collected during the control process, such as motor voltage and current, to iterate multiple times and finally output the correct three-phase current information, speed information, and rotor angle information of the motor.

[0194] The Clark transform and Park transform involved in this implementation scheme are as follows:

[0195] Transform the current in the three-phase stationary coordinate system (abc) obtained from EKF to a two-phase orthogonal stationary coordinate system. ):

[0196]

[0197] After obtaining the mathematical model in the two-phase stationary coordinate system, the model can be transformed into a model in the synchronous rotating coordinate system (dq) using the Park transformation:

[0198]

[0199] In the above formula The angle between the synchronous rotating coordinate system and the two-phase stationary coordinate system is the electrical angle through which the rotor rotates.

[0200] The motor speed in the above speed PI can be expressed as:

[0201]

[0202] Where n is the motor speed.

[0203] The speed PI controller is the outer loop, and its goal is to regulate the speed of the motor.

[0204]

[0205] Calculate speed error:

[0206] in That is the actual rotational speed. It is a given rotational speed. It is the difference between the given speed and the actual speed.

[0207] Calculate the reference torque using a PI controller:

[0208]

[0209] in This is the reference torque. It is the proportional gain of speed. It is the integral gain of speed.

[0210] Convert the torque reference value to a current reference value based on the motor parameters:

[0211]

[0212] in It is the desired q-axis current. It is the torque constant of the motor.

[0213] The current PI involved in this implementation scheme is the inner loop, and its control objective is to adjust the stator current so that it reaches the target value on the dq axis.

[0214] Calculate the current error:

[0215]

[0216] in It is the desired current along the d-axis. and These are the actual currents along the d-axis and q-axis, respectively. and These represent the differences between the expected current and the actual current along the d-axis and q-axis, respectively.

[0217] Calculate the voltage control quantity using a PI controller:

[0218]

[0219] in and , which are the voltages on the d-axis and q-axis, used to generate SVPWM signals. and These are the proportional gains of the d-axis and q-axis current errors, respectively. and These are the integral gains of the d-axis and q-axis current errors, respectively.

[0220] get and Then, through the inverse Park transformation:

[0221]

[0222] get and The signal is then input into SVPWM to obtain the duty cycle signal used to control the inverter switch. Finally, the inverter obtains the three-phase AC voltage of the motor and controls the motor. The angle of rotation of the coordinate axes.

[0223] Based on the above modules, the working principle of the sensorless control method for permanent magnet synchronous motors based on single-resistor sampling and EKF provided by this invention is as follows: During motor operation, the fault detection module detects the failure of the current sensor in real time. If the current sensor fails, the DC bus current signal is obtained through single-resistor sampling. Then, the EKF algorithm is used to reconstruct the current of another phase and the rotor position and speed information in real time. Finally, the third phase current is calculated by combining Kirchhoff's current law, realizing complete sensorless vector control. This method effectively integrates the motor dynamic model and actual measurement data through the EKF state prediction and measurement update mechanism. Even under changes in motor parameters or external interference, it can still maintain high-precision current and position estimation, ensuring the continuous and stable operation of the motor control system. It is particularly suitable for applications with extremely high reliability requirements, such as aerospace. Its control flow is as follows: Figure 7 As shown.

[0224] This invention proposes a sensorless current reconstruction method for permanent magnet synchronous motors (PMSMs) in electric aircraft, which offers significant advantages over traditional sensorless control of PMSMs. In terms of hardware, only a single sampling resistor is needed to accurately reconstruct the three-phase current, rotor position, and speed, greatly simplifying the system structure and reducing complexity. Simultaneously, the reduced number of sensors significantly improves system reliability, making it particularly suitable for applications with stringent safety requirements. Regarding control performance, the employed EKF algorithm offers higher reconstruction accuracy and dynamic response speed, adapting to a wider speed range. Through an adaptive noise adjustment mechanism, the system exhibits stronger robustness to changes in motor parameters and external disturbances. It can be implemented in real-time on mainstream embedded controllers, meeting the demands of high-performance control.

[0225] By applying this invention, when a sensor malfunctions, the system can switch to a reconfiguration mode, effectively ensuring the continuous and reliable operation of critical equipment. Simultaneously, this solution lowers the application threshold for advanced control technologies, contributing to the development of industrial automation. The system's excellent performance in harsh environments makes it a promising candidate for applications in aerospace, new energy vehicles, and other fields.

[0226] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1A device that provides the functions specified in one or more boxes.

[0227] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0228] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0229] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0230] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A sensorless current reconfiguration method for permanent magnet synchronous motors in electric aircraft, characterized in that, include: Through the fault detection module and switch The mode switching of the electric aircraft motor control system includes normal operation and fault tolerance. The fault detection module detects whether the current sensor is malfunctioning in real time. If the current sensor is malfunctioning, the DC bus current signal is obtained by sampling with a single resistor. The EKF method is used to reconstruct the current of another phase and the rotor position and speed information in real time. Then, the third phase current is calculated by combining Kirchhoff's current law to realize sensorless vector control.

2. The sensorless current reconstruction method for permanent magnet synchronous motors in electric aircraft according to claim 1, characterized in that, The switch Set to 1 for normal operation mode; the current acquisition module collects the three-phase current of the motor. , , The current is converted into a two-phase stationary coordinate system by Clark transformation. , Then, the two-phase rotating coordinate system current is obtained through the Park transformation. , , as current closed-loop feedback; The voltage acquisition module acquires two-phase voltages. , Input the EKF module, which outputs the rotor position based on the motor model. and rotational speed This provides position and velocity feedback for Park transformation and velocity closed loop.

3. The sensorless current reconstruction method for permanent magnet synchronous motors in electric aircraft according to claim 1, characterized in that, The switch Setting it to 2 enables fault-tolerant mode. The fault detection module monitors for abnormal or lost current signals from the current acquisition module to determine if the current acquisition module has failed. If the current sensor fails, single-resistor sampling is initiated. The single-resistor acquisition module is placed on the DC side of the inverter and acquires the DC bus current. Combined with the two-phase voltage acquired by the voltage acquisition module , The PWM drive signal is input to the EKF module to achieve accurate reconstruction of the three-phase current; The EKF module realizes the three-phase current of the motor through state estimation. , , The reconfiguration process involves the reconfiguration current being transformed by Clark and Park to replace the failed current feedback, and the EKF still outputs the rotor position. and rotational speed .

4. The sensorless current reconstruction method for permanent magnet synchronous motors in electric aircraft according to claim 1, characterized in that, In both normal operation and fault-tolerant modes, the speed closed loop uses a reference speed. With feedback speed The deviation is adjusted by PI to generate a q-axis current reference value. Combined with d-axis current reference value , and feedback current , After comparison, the voltage is generated by PI regulation. , The result obtained by inverse Park transform , The input SVPWM module generates drive signals to control the inverter, and the output three-phase AC power drives the permanent magnet synchronous motor.

5. The sensorless current reconstruction method for permanent magnet synchronous motors in electric aircraft according to claim 3, characterized in that, The single-resistor acquisition module acquires the DC bus current. Combined with the two-phase voltage acquired by the voltage acquisition module , The PWM drive signal is input to the EKF module to achieve accurate reconstruction of the three-phase current, including: establishing a mapping relationship between the sampling resistor current and the switch state based on the one-to-one correspondence between the bus current and the voltage vector, wherein the positive direction of the current is defined according to the motor convention; When the 100 vector is applied, the bus current is: The specific relationship is shown in the following formula: In the formula, , and These are the three-phase currents. , and These are the activation statuses of the three-phase bridge arms. It is the bus current; The relationship between the three-phase currents of a motor can be obtained from Kirchhoff's laws: The switch state and phase current are obtained through the one-to-one correspondence between the bus current and the voltage vector. During the first sampling, the bus current flows through the current... During the second sampling, the current flowing through the busbar was - Similarly, by combining EKF and voltage signals, another current can be reconstructed; The last current term was derived based on Kirchhoff's laws, and the three-phase current was successfully reconstructed by time-sharing sampling within one PWM cycle.

6. The sensorless current reconstruction method for permanent magnet synchronous motors in electric aircraft according to claim 3, characterized in that, The construction of the EKF module equations includes: Using a recursive calculation method, based on the optimal estimate at the current moment, and combining the system dynamic model, state equations, observation equations, and noise statistics, the system accurately predicts the motor speed and rotor position at the next moment. The Taylor expansion method is used to approximate the linearization of the nonlinear system, and the continuous nonlinear control system is described by the following equations; Equation (10) is the state-space equation of the system, and equation (11) is the measurement equation of the system, where process noise... and observation noise They still follow a Gaussian white noise distribution, with corresponding covariance matrices Q and R, respectively; make: when When the size is sufficiently small, performing a Taylor expansion at the optimal state estimation point, retaining the first-order terms and discarding the higher-order terms, yields the approximately linearized system state-space equations and measurement equations. The state estimate is: Subtracting the corresponding values ​​from the four equations above, we get: make and Then F and H are Jacobian determinants, as shown in equations (16) and (17) respectively: Then, by combining equation (15) with equations (16) and (17), we can obtain: Once the linearity condition is met, the above two equations are discretized. In the formula, represents the system's transition matrix under discrete conditions. When the overall frequency of the algorithm is sufficiently high, the following relationship holds: In the formula, It is the identity matrix. The algorithm execution cycle; The Kalman filter equation for the error between the estimated and true values ​​is derived through the above derivation, and its specific form is shown below: beg and The ultimate goal is to ensure that the filtered estimated value remains consistent with the true value. From this, we can derive the five formulas for the nonlinear, discrete extended Kalman filter as follows: 。 7. The sensorless current reconstruction method for permanent magnet synchronous motors in electric aircraft according to claim 3, characterized in that, The fault detection module determines the failure of the current acquisition module by monitoring for abnormal or lost current signals. The fault detection module compares the given current with the actual value collected by the current sensor. , , Comparison: Provide three-phase current to the motor. This represents the actual three-phase current of the motor; To set a threshold, if the current deviation of a certain phase is... If the current sensor fails, the current sensor for that phase is deemed to be faulty. Set to 2; if all three-phase currents satisfy If the current sensor is normal, then the current sensor is considered to be functioning correctly. Set to 1.

8. The sensorless current reconstruction method for permanent magnet synchronous motors in electric aircraft according to claim 3, characterized in that, In the single-resistance acquisition module, the duration of the effective vector is greater than the minimum sampling time. ; To obtain the precise phase current of the motor under different switching states: In the formula, Indicates dead time. The duration of the transient process during the switching of the transistor state ( ), This represents the total time required for the microprocessor sampling module to convert the analog signal into a digital signal from sampling.

9. The sensorless current reconstruction method for permanent magnet synchronous motors in electric aircraft according to claim 2, characterized in that, The EKF module outputs the rotor position based on the motor model. and rotational speed It provides position and velocity feedback for Park transformation and velocity closed loop, including: To predict the state vector, the previous optimal estimate is used as the independent variable, and the non-linearized state transition matrix is ​​used to... The increment within the time period is calculated and added to the value of the previous best estimate to obtain the prior estimate of the current prediction. Using the already discretized state transition matrix To estimate the covariance matrix, we obtain the prior covariance matrix: Calculate the Kalman gain: The state vector estimate is updated using Kalman gain to obtain the posterior estimate, which is also the optimal estimate. Update the covariance matrix using Kalman gain: 。 10. The sensorless current reconstruction method for permanent magnet synchronous motors in electric aircraft according to claim 3, characterized in that, The EKF module realizes the three-phase current of the motor through state estimation. , , The reconfiguration process involves the reconfiguration current being transformed by Clark and Park to replace the failed current feedback, and the EKF still outputs the rotor position. and rotational speed ,include: The current in the three-phase stationary coordinate system (abc) obtained from EKF is converted to a two-phase orthogonal stationary coordinate system. ): After obtaining the mathematical model in the two-phase stationary coordinate system, the model is transformed into a model in the synchronous rotating coordinate system (dq) using the Park transformation: In the formula The angle between the synchronous rotating coordinate system and the two-phase stationary coordinate system is the electrical angle through which the rotor rotates.

Images