A hall motor starting method based on time difference feedforward and error self-learning

Abstract

The application discloses a Hall motor starting method based on time difference feedforward and error self-learning, and comprises the following steps: starting a Hall capture channel, calculating a motor electric angle through a feedback value in a motor closed loop interruption, starting a motor starting process, and entering a predetermined positioning stage; positioning a rotor to a sector center line by reading a current sector index, then switching to an I / F open loop blind mode, and completing a three-edge observer preprocessing by using the first three Hall signal jumps, so that smooth switching from the open loop to the closed loop is realized; capturing a Hall signal jump time, checking signal legality, calculating a time difference change rate to judge a motor motion state; calculating an electric angle in a motor running process based on phase error dynamic correction speed; installing error compensation online self-learning, performing self-learning only when the motor is stably running, estimating an instantaneous angle error in real time, updating a sector compensation value, and preventing total angle drift through zero and constraint. The application can solve the problem of Hall motor closed loop control.

Description

Technical Field

[0001] This invention belongs to the field of motor control technology, specifically a Hall motor starting method based on time difference feedforward and error self-learning. Background Technology

[0002] Permanent magnet synchronous motors (PMSMs) have been widely used in key fields such as industrial automation, home appliances, electric vehicles, and aerospace due to their advantages of high efficiency, high power density, high reliability, and maintenance-free operation. However, aerospace applications place extremely stringent requirements on the size, weight, and energy efficiency of the drive system, necessitating a motor control solution that combines high control precision with extreme lightweight design.

[0003] Compared to high-precision position sensors such as rotary transformers or photoelectric encoders, Hall effect sensors offer significant advantages such as simple structure, small size, light weight, low cost, and ease of integration, making them the preferred position feedback solution in aerospace and other applications highly sensitive to system lightweighting. However, limited by their discrete 60° electrical angle resolution, installation deviations, and susceptibility to electromagnetic interference, traditional control methods based on Hall effect sensors generally suffer from poor startup performance and low operating efficiency, severely restricting their application potential in high-performance applications. Specifically:

[0004] High energy consumption and poor stability during startup: Traditional open-loop startup strategies lack precise pre-positioning of the rotor's initial position, and coupled with insufficient suppression of mechanical oscillations, the startup current often experiences large overshoots, which not only wastes energy but may also cause thermal stress damage.

[0005] Phase error accumulation during operation: The inherent installation error of the Hall sensor and signal noise are amplified during high-speed operation, causing continuous commutation phase deviation, reducing electromagnetic torque output efficiency, and even inducing torque pulsation;

[0006] The system complexity is passively increased: In order to compensate for the above defects, existing solutions often rely on additional hardware filtering circuits, redundant sensors or complex software post-processing mechanisms, which weakens the lightweight and low-cost advantages that the Hall effect solution should have, resulting in a decrease in overall system efficiency instead of an increase.

[0007] Therefore, there is an urgent need for a dedicated startup and operation control method based on the characteristics of Hall sensors. This method should achieve highly robust startup, dynamically compensate for installation errors, suppress the influence of signal noise, and effectively integrate timing information with a self-learning mechanism without introducing additional hardware. This would fully leverage the core advantages of Hall sensors in terms of size, weight, and integration, and meet the comprehensive needs of aerospace and other fields for efficient, lightweight, and reliable motor drive systems. Summary of the Invention

[0008] The purpose of this invention is to provide a Hall motor starting method based on time difference feedforward and error self-learning to solve the above-mentioned Hall motor closed-loop control problem.

[0009] The technical solution to achieve the purpose of this invention is as follows:

[0010] A Hall motor starting method based on time difference feedforward and error self-learning includes:

[0011] Step 1: When the driver receives the start command, it opens the Hall capture channel, calculates the motor electrical angle through the feedback value of the Hall sensor in the motor closed-loop interrupt, starts the motor start process, and enters the pre-positioning stage.

[0012] Step 2: By reading the current sector index, the rotor is positioned to the sector center line, then switched to the I / F open-loop blind transmission mode, and the three edge observer preprocessing is completed by using the first three Hall signal transitions to achieve a smooth switch from open loop to closed loop.

[0013] Step 3: Capture the Hall signal transition time, verify the signal validity, and calculate the time difference change rate to determine the motor's motion state;

[0014] Step 4: Calculate the electrical angles during motor operation based on the dynamic speed correction of phase error;

[0015] Step 5: Install online self-learning for error compensation. Self-learning is only performed when the motor is running stably. It estimates the instantaneous angle error in real time, updates the sector compensation value, and prevents total angle drift through zero-sum constraints.

[0016] The significant advantages of this invention compared to existing technologies are:

[0017] (1) By introducing a rotor initial position self-correction mechanism and a dynamic current ramp control strategy, this invention effectively suppresses current overshoot during the start-up phase, significantly reduces energy loss, and avoids the risk of step loss or reversal caused by mechanical oscillation, thus achieving a smooth and reliable soft start.

[0018] (2) The present invention adopts an online phase deviation identification and compensation algorithm based on multi-cycle commutation signal, which can correct the commutation angle error caused by the installation offset of Hall element in real time without relying on external calibration equipment, fundamentally solving the problem of error accumulation with running time in traditional methods, and ensuring efficient commutation in the full speed domain.

[0019] (3) By integrating Hall edge timing information, the present invention achieves virtual position interpolation at the sub-60° electrical angle level without increasing hardware costs, effectively reducing commutation lag under high-speed conditions, improving electromagnetic torque output efficiency, and significantly suppressing torque pulsation.

[0020] (4) The algorithm of this invention has strong anti-interference capability and has a fault-tolerant processing mechanism for Hall signal jitter, glitches and brief loss. It is particularly suitable for extreme application scenarios such as aerospace where there is strong electromagnetic interference, vibration shock and power consumption limitation. Attached Figure Description

[0021] Figure 1 This is a flowchart of the present invention.

[0022] Figure 2 This is a measured diagram of the motor starting A-phase current according to the present invention. Detailed Implementation

[0023] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0024] The flowchart of this invention is as follows Figure 1 As shown in the accompanying drawings and specific embodiments, the present invention will be further described in detail below:

[0025] Step 1: When the driver receives the start command, it opens the Hall capture channel, calculates the motor electrical angle through the feedback value of the Hall sensor in the motor closed-loop interrupt, starts the motor start process, and enters the pre-positioning stage.

[0026] Step 2: By reading the current sector index, the rotor is positioned to the sector centerline, then switched to the I / F open-loop blind transmission mode, and the three edge observer preprocessing is completed using the first three Hall signal transitions to achieve a smooth switch from open loop to closed loop.

[0027] Step 2.1: Rotor pre-positioning

[0028] When the start command is issued, the current sector index N of the motor is read, and the target positioning angle is determined. Defined as the center line of the sector, the controller forces current injection. The electrical angle is set as .

[0029] To prevent rotor vibration, current is injected. The instruction uses an S-shaped curve to increase the injected current from 0 to 20% of the rated current and maintain the current injection for 500ms. Wait for the mechanical oscillations to decay.

[0030] After the 500ms holding period ends, the rotor is considered to have stabilized and stopped. At this location, the current electrical angle is Then switch to I / F open-loop blind acceleration state.

[0031] Step 2.2: I / F open-loop blind firing drag

[0032] At this point, Hall effect signals are not used for feedback; the software internally generates a signal with a fixed acceleration. Increased virtual angle :

[0033]

[0034] According to virtual perspective A constant rotating current vector is applied along the dq axis: , Apply a constant current command to the q-axis. Apply a constant current command to the d-axis. The q-axis current is driven by the open-loop blind generator of I / F.

[0035] During the dragging process, continuously monitor the Hall signal pin; if If the change value exceeds 90 electrical degrees and no Hall signal level flip is detected, it is determined to be a stall or mechanical jam, and the machine is immediately stopped and an error is reported.

[0036] Step 2.3: Preprocessing of the three-edge observer

[0037] When the motor is running in I / F open-loop blind drive mode, the historical data is gradually recorded by using the first three Hall signal transitions, thus realizing the switch from open-loop to closed-loop.

[0038] Step 2.3.1 When the Hall signal flips, i.e., the first transition edge is detected, position calibration is triggered:

[0039] Step 2.3.1.1 Record the current timestamp .

[0040] Step 2.3.1.2 Immediately change the virtual angle in the software. Forced correction to the starting angle of this sector .

[0041] Step 2.3.1.3 Continue to maintain the I / F drag mode from step 2.2.

[0042] Step 2.3.2 When the Hall signal flips again, i.e., at the second transition edge, the trigger speed is calculated:

[0043] Step 2.3.2.1 Record the timestamp .

[0044] Step 2.3.2.2 Calculate the actual sector latency at the first moment: .

[0045] Step 2.3.2.3 Continue to maintain the I / F drag mode from step 2.2.

[0046] Step 2.3.3 When the transition edge of the third Hall signal is detected, the acceleration calculation is triggered:

[0047] Step 2.3.3.1 Record the timestamp ,calculate , The time consumed for the actual sector at the second moment.

[0048] Step 2.3.3.2 If the Hall effect switching requirement of step 3 is met, switch to motor closed-loop control function; otherwise, set the fault flag and trigger a safety shutdown.

[0049] Step 3: Capture the Hall signal transition time, verify the signal validity, and calculate the time difference change rate to determine the motor's motion state.

[0050] Step 3.1: Interrupt Trigger and Signal Validity Verification

[0051] Filter out electrical noise to ensure that only the actual rotor position changes.

[0052] Step 3.1.1 Event Capture:

[0053] When a level transition is detected on any pin of the Hall interface, an interrupt is immediately triggered, and the current hardware timer count value is latched. and the starting angle of the current sector .

[0054] Step 3.1.2 Remove illegal states:

[0055] Read the current three-phase Hall status value. If the status is 000 or 111, immediately set the fault flag and trigger a safety shutdown. If the new sector index is not adjacent to the old sector index, it is determined to be Hall interference or loss of synchronization, and a safety shutdown is triggered.

[0056] Step 3.1.3 Filtering out burrs:

[0057] Calculate instantaneous time difference ,in The actual sector time at time k, where k=1,2,…; This is the count value of the previous hardware timer.

[0058] like , If the minimum instantaneous time difference is found, the interrupt is determined to be an electromagnetic interference glitch, and the interrupt is exited directly without updating any state.

[0059] Step 3.2: Calculate acceleration using time difference

[0060] Step 3.2.1 Calculate the rate of change of time difference :

[0061]

[0062] in, This represents the actual sector latency at time k-1. If the sector time is shortening, the motor is accelerating; if The sector time is increasing, and the motor is slowing down; if If the sector times are equal, the motor will run at a constant speed.

[0063] Step 3.2.2 Numerical Limitation

[0064] like If the value is 0 or the maximum value of the acquisition channel, then a forced setting will be applied. .

[0065] Step 4: Calculate the electrical angle during motor operation based on the dynamic speed correction of phase error.

[0066] Calculate the motor speed when the Hall signal changes. Using errors to correct speed For use by the next sector. The calculation formula is as follows:

[0067]

[0068] in, For integration time, This is the phase error; For integral gain; The proportional gain is calculated using the following formula:

[0069]

[0070] in, Reference gain; Saturation gain; This is the sensitivity adjustment coefficient; This is a limiting function, ensuring that the product within the parentheses does not exceed 1.

[0071] The calculation formula is as follows:

[0072]

[0073] in, The damping ratio is kept constant and is set to 0.707.

[0074] The calculation formula is as follows:

[0075]

[0076] in, The electrical angle is calculated by interpolation when the Hall signal changes. The formula for calculating the electrical angle using interpolation is as follows:

[0077]

[0078] in, To calculate the time interval for electrical angles, The real-time electrical angle calculated at time t is used for the vector closed-loop control of the Hall motor. The measured graph of the motor starting A-phase current is shown below. Figure 2 As shown; The anchor point angle is the sector starting angle at the time of the last Hall signal transition. ; The acceleration estimate is calculated using the sector time difference method:

[0079]

[0080] By reading the theoretical starting angle corresponding to the current Hall state and superimposing the installation error compensation array... get:

[0081]

[0082] in, The theoretical starting angle for installing the Hall sensors at an ideal 120° angle (three sensors spaced 120° electrical angles apart). This is the installation error compensation array corresponding to sector N.

[0083] Step 5: Install online self-learning for error compensation. Self-learning is only performed when the motor is running stably. It estimates the instantaneous angle error in real time, updates the sector compensation value, and prevents total angle drift through zero-sum constraints.

[0084] Step 5.1: Learn the admission criteria

[0085] Ensure parameter updates are performed during stable motor operation to prevent acceleration / deceleration processes from being misjudged as installation errors. The judgment condition is: if If the speed is between 30% and 90% of the rated speed, then the following steps are activated to achieve self-learning of the Hall sector angle error of the motor.

[0086] Step 5.2: Instantaneous angle error estimation

[0087] First, calculate the actual physical angle. The calculation formula is as follows:

[0088]

[0089] in, The actual time consumed in the current sector is calculated from step 3; The speed is calculated from step 4.

[0090] Then read the current compensated angle of the sector. :

[0091]

[0092] And calculate the deviation:

[0093]

[0094] in, The angle after compensation for the current sector N; This is the installation error compensation array corresponding to sector N.

[0095] when When the error is greater than 0.1 degrees, the sector is updated using the following formula:

[0096]

[0097] Finally, a zero-sum constraint check is performed on the error. The sum of the angles of the six sectors must be strictly equal to 360 degrees. Without this constraint, the cumulative error of each sector may lead to a drift in the total angle. Calculate the total error:

[0098]

[0099] if Then it will be deducted in equal proportions:

[0100]

[0101] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any other way. Any modifications or equivalent changes made based on the technical essence of the present invention shall still fall within the scope of protection claimed by the present invention.

Claims

1. A Hall motor starting method based on time difference feedforward and error self-learning, characterized in that, include: Step 1: When the driver receives the start command, it opens the Hall capture channel, calculates the motor electrical angle through the feedback value of the Hall sensor in the motor closed-loop interrupt, starts the motor start process, and enters the pre-positioning stage. Step 2: By reading the current sector index, the rotor is positioned to the sector center line, then switched to the I / F open-loop blind transmission mode, and the three edge observer preprocessing is completed by using the first three Hall signal transitions to achieve a smooth switch from open loop to closed loop. Step 3: Capture the Hall signal transition time, verify the signal validity, and calculate the time difference change rate to determine the motor's motion state; Step 4: Calculate the electrical angles during motor operation based on the dynamic speed correction of phase error; Step 5: Install online self-learning for error compensation. Self-learning is only performed when the motor is running stably. It estimates the instantaneous angle error in real time, updates the sector compensation value, and prevents total angle drift through zero-sum constraints.

2. The Hall motor starting method based on time difference feedforward and error self-learning according to claim 1, characterized in that, Step 2 specifically includes: Step 2.1, Rotor Pre-positioning: Read the current motor sector index and set the target positioning angle. Defined as the centerline of the sector, the controller forces current injection, and the electrical angle is set to... ; After maintaining the set injection current for a certain time, switch to the I / F open-loop blind acceleration state; Step 2.2: I / F open-loop blind drive: At this time, Hall signals are not used for feedback, and a signal with a fixed acceleration is generated. Increased virtual angle According to virtual perspective A constant rotating current vector is applied along the dq axis; Step 2.3: Three-edge observer preprocessing: When the motor is running in I / F open-loop blind drive mode, the historical data is recorded step by step using the first three Hall signal transitions. If the Hall transition requirements of Step 3 are met, the system switches from open-loop to closed-loop.

3. The Hall motor starting method based on time difference feedforward and error self-learning according to claim 2, characterized in that, The preprocessing procedure for the three-edge observer is as follows: Step 2.3.1: When the Hall signal flips, i.e., the first transition edge is detected, position calibration is triggered. Record the current timestamp ; virtual angle Forced correction to the starting angle of this sector ; Continue using the I / F drag mode; Step 2.3.2: When the Hall signal flips again, i.e., the second transition edge, the trigger speed is calculated: Record timestamp ; Calculate the actual sector latency at the first moment: ; Continue using the I / F drag mode; Step 2.3.3: When the transition edge of the third Hall signal is detected, the acceleration calculation is triggered. Record timestamp Calculate the actual sector latency at the second time step. ; If the Hall effect switching requirement is met, the system switches to motor closed-loop control; otherwise, the fault flag is set, triggering a safety shutdown.

4. The Hall motor starting method based on time difference feedforward and error self-learning according to claim 1, characterized in that, The rate of change of time zone is calculated using the following formula: in This represents the actual sector latency at time k; This is the count value from the previous timer; This is the current timer count value; if , To minimize the instantaneous time difference, exit the interrupt; like To determine if the motor is accelerating; if To determine if the motor is decelerating; if Determine if the motor is running at a constant speed; if If the value is 0 or the maximum value of the acquisition channel, then a forced setting will be applied. .

5. The Hall motor starting method based on time difference feedforward and error self-learning according to claim 1, characterized in that, The electrical angle during motor operation is calculated using the following formula: in, To calculate the time interval for electrical angles, The real-time electrical angle calculated at time t. For the anchor point angle, The speed of the motor rotation. For time, This is the phase error; This is the integral gain; For the proportion to increase, benefit, The rate of change of time zone. Here are the estimated acceleration values, where This represents the actual sector time at time k.

6. The Hall motor starting method based on time difference feedforward and error self-learning according to claim 5, characterized in that, proportional gain The calculation formula is as follows: Integral gain The calculation formula is as follows: Phase error The calculation formula is as follows: in, Reference gain; Saturation gain; This is the sensitivity adjustment coefficient; For the amplitude limiting function, The damping ratio is constant. This is the starting angle of the current sector.

7. The Hall motor starting method based on time difference feedforward and error self-learning according to claim 6, characterized in that, The starting angle of the current sector is obtained by reading the theoretical starting angle corresponding to the current Hall state and superimposing it with the installation error compensation array: in, The theoretical starting angle for installing the Hall sensor at an ideal 120° angle. This is the installation error compensation array corresponding to sector N.

8. The Hall motor starting method based on time difference feedforward and error self-learning according to claim 1, characterized in that, Instantaneous angle error is estimated using the following formula: The formula for sector compensation value is: Zero-sum constraints are achieved by distributing and deducting the accumulated errors of each sector equally: in, The angle after compensation for the current sector N; This is the installation error compensation array corresponding to the current sector N. From a practical physics perspective, As the current compensated angle, The speed of the motor rotation. This represents the actual sector time at time k.

Images