A driving control system and control method of intelligent transfer equipment

Abstract

The application relates to the technical field of electric vehicle driving control, and particularly discloses a driving control system and a control method of intelligent transfer equipment, which are used for solving the driving control problem of the intelligent transfer equipment six-phase permanent magnet synchronous motor under complex working conditions of a port, i.e. large current tracking error, high torque ripple and noise vibration, unstable regeneration, poor low-speed micro-climbing and ramp restart response caused by sampling and inverter time lag, parameter drift, bus fluctuation and degradation fault, and the system comprises an AT-AGV, a PMSM, a six-phase inverter, a direct-current bus, a controller and a planning module. Through midpoint sampling, two-beat equivalent time lag modeling and limited error-free beat voltage optimization, in combination with xy harmonic compensation, online parameter identification, three-section type regeneration track and fault-tolerant SVPWM, low overshoot fast torque establishment, low ripple low NVH, robust stability and efficient recovery under bus fluctuation and fault, and accurate tracking current and smooth traction during micro-climbing and ramp restart are realized.

Description

Technical Field

[0001] This invention relates to the field of electric vehicle drive control technology, and in particular to a drive control system and control method for intelligent transfer equipment. Background Technology

[0002] In the publicly available CNKI document (Ye Xun. Research on Disturbance Suppression of Dual Three-Phase Permanent Magnet Synchronous Motors Based on Deadbeat Predictive Control [D]. Yangzhou University, 2025. DOI:10.27441 / d.cnki.gyzdu.2025.001678.), the current tracking error and torque ripple problems caused by sampling, calculation, inverter delay and parameter drift of dual three-phase permanent magnet synchronous motors (PMSM) are addressed. By combining space vector pulse width modulation (SVPWM), midpoint sampling and anti-aliasing filtering, full-link time delay equivalent modeling, and deadbeat predictive current control, the transition and tracking quality of current, torque and speed links and the ability to suppress parameter deviations are improved.

[0003] However, in the transshipment of vehicles at ports, intelligent transshipment equipment (AT-AGV, All Terrain Automated Guided Vehicle) needs to complete all-weather automatic platooning, precise parking and positioning, low-speed crawling and heavy-load start, hill parking and restart, regenerative braking energy recovery and fault degradation operation under complex conditions such as steel deck and asphalt mixed road surface, slope and bend, frequent berth changes, uneven wheel end adhesion, large tire pressure load changes, steering angle fluctuations and electromagnetic interference. Therefore, the motor drive control of AT-AGV needs to meet the rigid requirements of rapid torque build-up, low torque pulsation, resistance to bus fluctuations, resistance to parameter drift and redundancy fault tolerance.

[0004] Under the complex conditions of commercial vehicle transshipment in ports, the control strategies described in the aforementioned literature are insufficient to suppress harmonic subspace currents based on constant parameter discrete feedforward and fixed voltage limiting. This is due to the low-speed micro-crawling and frequent start-stop of AT-AGVs, the rapid switching between slope maintenance, stick-slip states of steel surfaces and rubber tires, and the fluctuation of bus voltage with regenerative power. The regenerative braking tail-end current rise is limited and torque tracking lags behind. When the inverter is degraded, the decoupling capability of the x and y channels of the harmonic subspace decreases, resulting in increased harmonic current coupling and torque pulsation. This leads to the deterioration of vehicle noise, vibration, and acoustic roughness, slows down the turnover of commercial vehicles, increases operating costs, and increases the risk of accidents. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a drive control system and control method for intelligent transfer equipment. By midpoint sampling, two-phase equivalent time delay modeling, and limited deadbeat voltage optimization, combined with xy harmonic compensation, online parameter identification, three-stage regenerative trajectory, and fault-tolerant SVPWM, it achieves low overshoot and rapid torque establishment, low ripple and low NVH, robust stability and high-efficiency recovery under bus fluctuations and faults, and maintains accurate current tracking and smooth traction in scenarios such as micro-crawling and ramp restart.

[0006] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0007] A drive control system for an intelligent transfer equipment includes an AT-AGV, which is connected to a PMSM, a DC bus, a controller, and a planning module. The PMSM and the DC bus are both connected to a six-phase inverter, and the six-phase inverter and the planning module are both connected to the controller.

[0008] The DC side of the six-phase inverter is connected to the power battery pack of the AT-AGV via the DC bus, and the AC side of the six-phase inverter is connected to the six-phase stator winding of the PMSM.

[0009] AT-AGV collects current, bus voltage, speed, acceleration, and slip ratio;

[0010] The controller, based on midpoint sampling of the modulation period and employing three-segment time-delay equivalent compensation, establishes a dq coordinate system and an xy harmonic subspace model, where d and q are the stator synchronous rotation axes, and x and y are the harmonic axes. Under SVPWM and bus voltage limiting, the solution is obtained... The voltage vector at each moment is mapped to the duty cycle of the six bridge arms and then sent to the gate drive of the six-phase inverter at the next cycle boundary. For discrete moments, stator resistance, d-axis and q-axis inductance, harmonics and leakage inductance, and rotor permanent magnet flux linkage are identified online. The weights of each axis of current tracking error, voltage regularization weights, and confidence weights of the parameters obtained online are adaptively weighted according to error sensitivity. Bus ripple observation and adaptive current creep shaping are performed at the tail end of regenerative braking. When the bridge arm fails, phase reconstruction and current redistribution are performed. Road surface adhesion is estimated based on slip ratio and acceleration. Torque limiting and field weakening switching are linked. Feedforward cancellation and xy weight scheduling are implemented for the second and sixth harmonics. Torque constraints and speed differential corrections are applied during low-speed micro-creep, ramp parking and restart, wet and stick-slip switching on steel surfaces, and bus fluctuations. Available torque-energy constraints are output to the planning module.

[0011] As one aspect of the system of the present invention, the controller sets the current sampling time of each modulation cycle as the midpoint of the cycle and introduces a dynamic offset. Align the sampling time with the midpoint of the period of the average voltage of the six-phase inverter, and set the three-segment time delay equivalent as follows: And all are unified as equivalent time delays , For the current modulation period, , , These are the three time delays: sampling midpoint offset, calculation or holding until the next sampling cycle update in the current sampling period, and the average voltage of the six-phase inverter. set up, The PWM switching frequency of the six-phase inverter, with equivalent time delay. The delay in the continuous domain is approximated by a first-order approximation, which is expressed as: It updates synchronously with the pulse width modulation switching frequency of the six-phase inverter. For the Laplace operator;

[0012] dynamic offset The formula for obtaining it is ; The dead time of the bridge arm is set by the gate drive of the six-phase inverter. To control computation time, the CPU cycle count is used by the controller; , All are dimensionless calibration coefficients, obtained by performing step or frequency sweep tests on AT-AGV and fitting the time difference between the planned sampling midpoint and the measured average voltage midpoint using the least squares criterion.

[0013] As one aspect of the system of the present invention, the timing and execution steps of the controller in the modulation period, including sampling, equivalent time delay modeling, voltage prediction solution, limiting, and SVPWM distribution, include:

[0014] Step S1, Define the time base and period: Define the modulation period Define the periodic boundary ;

[0015] Step S2, Midpoint Sampling Alignment: Set the sampling time of the current midpoint. Dynamic offset ;

[0016] Step S3, Time Delay Decomposition and Summary: Decompose the total time delay of the inner current loop into... These correspond to the sampling midpoint offset, the current calculation / waiting for the next cycle update, and the formation of the six-phase inverter average voltage, respectively, and are summarized into an equivalent time delay. ;

[0017] Step S4, Continuous Domain Approximation: In the continuous time domain, use the transfer function with pure time delay. The equivalent time delay is represented using a first-order approximation. ;

[0018] Step S5, Discrete-domain time delay implementation: A two-step delay operator is used in the discrete-time domain. , It is a discrete shift operator. Two-beat delay, and midpoint sampling sequence With voltage reference Establish index relationships, The discrete current quantity sampled at the midpoint. They are equal in the limiting sense.

[0019] Step S6, Modulation command generation and issuance: at time... The discrete voltage reference is mapped to a six-phase space vector pulse width modulation duty cycle. , The duty cycle vector of the six bridge arms is obtained through the SVPWM mapping algorithm and output to the gate drive of the six-phase inverter;

[0020] Step S7, Feedback Data Acquisition and Indexing: at time... Collect current samples , and the result obtained in step S5 A two-beat delay relationship is formed;

[0021] Step S8, Parameter Update and Range Constraints: When When changing Synchronous updates and the approximation coefficients of step S4, and for Using saturation constraints , These are the experimentally fitted values. , To allow upper and lower limits, obtained through security and hardware constraints, the updated limits will be... , , Write the timing and parameter table to the controller and use it for limiting, prediction, and SVPWM mapping in the next cycle.

[0022] As one aspect of the system of the present invention, the controller controls the six-phase current vector Angular correlation transformation matrix get , For the first The stator current vector in the dq and xy coordinate systems at each sampling time. , , , , , These are the component current vectors of the six-phase current at each point. This refers to the spatial phase shift angle of the winding. The electric angle position, It consists of two sets of Clark-Parker transformations, the first set being three-phase. Current angle transformation according to , , , The stator stationary coordinate system for the first group of three-phase currents , Components; second group of three phases Angle transformation according to , , , The stator stationary coordinate system for the second group of three-phase currents , Components; and establish the current update formula according to backward differential discretization:

[0023] ;

[0024] ;

[0025] ;

[0026] ;

[0027] Equivalently written in matrix form:

[0028] ;

[0029] in, Electric angular velocity, , , , The six-phase currents are respectively The d-axis, q-axis, x-axis, and y-axis components at the sampling time. , , , The six-phase currents are respectively The d-axis, q-axis, x-axis, and y-axis components at the sampling time. For the first The stator current vector in the dq and xy coordinate systems at each sampling time. For stator resistance, For d-axis inductance, It is the q-axis inductance. For harmonics or leakage inductance, For rotor permanent magnet flux linkage, Let dq and xy be the voltage vectors. For the reason , , , A defined coupling matrix, Multiply the admittance matrix by The coefficients of dispersion, Given a known disturbance vector introduced by the back electromotive force term, all trigonometric functions and matrix operations are calculated in real time within the controller and updated accordingly. , , renew.

[0030] As one aspect of the system of the present invention, the controller regulates the voltage vector under the constraints of DC bus circular limiting and modulation realizability. At discrete time Solve using the following formula:

[0031] ;

[0032] Constraints , , , And the optimal Inverse Transformation The voltage is mapped to a six-phase stator phase voltage reference, and then the duty cycle is generated by symmetrically center-aligned six-phase space vector pulse width modulation. ,in, ;

[0033] in, For the dq and xy current reference, The sampled current is the current at the midpoint of the current discrete moment. It is a diagonal weighted matrix. For voltage regularization weights, This is the DC bus voltage. , The upper limit of the xy channel voltage ratio, with values ​​between 0 and 1. This is the voltage solution from the previous sampling period. This is the upper limit of the voltage step. electric angle Phase shift angle with winding space The six-phase to dq and xy inverse transformation matrices For the first time after adding the zero-order component The phase voltage reference is obtained by applying the zero-sequence component symmetry according to the duty cycle and the modulation margin rule. Correction of device voltage drop step size and duty cycle;

[0034] For saturation operators, , To modulate the lower and upper limits.

[0035] As one aspect of the system of the present invention, the controller controls the parameter vector. A recursive least squares method with a forgetting factor and projection constraints are applied.

[0036] ;

[0037] in, , For bounded sets The projection operator is applied to the stator resistance, and temperature correction is used. , For parameter estimates, For covariance matrix, , These are the regression vector and the observation vector obtained by rearranging the discrete voltage-current equations, respectively. For the numerical value in The forgetting factor is set by the controller parameter table; , They are bounded sets The lower and upper limits are given by the safety constraints; This is the initial value of the stator resistance. This is the temperature correction factor for stator resistance. , All are determined by the material data of the stator resistance. This is the current stator resistance temperature. The reference temperature for stator resistance temperature correction.

[0038] As one aspect of the system of the present invention, the controller adopts cross-coupled step size and harmonic feedforward in the xy harmonic subspace and performs weighted scheduling.

[0039] As one aspect of the system of the present invention, the controller adopts a three-segment DC side current trajectory and torque slope plan under regenerative braking and low-speed micro-creep conditions, and performs fault mapping at the same time.

[0040] A drive control method for intelligent transfer equipment, applying the aforementioned drive control system for intelligent transfer equipment, the method comprising:

[0041] Step 1, System Initialization and Time Base Setting: Set the pulse width modulation switching frequency of the six-phase inverter to determine the modulation period; establish the start and end boundaries of each cycle and the trigger time of midpoint sampling in the cycle; calibrate the small offset of midpoint sampling based on the dead zone of the bridge arm and the control calculation time; and write the six-phase inverter PWM switching frequency, modulation base period, start and end boundaries of the cycle, carrier phase mark, duty cycle issuance time, midpoint sampling trigger time, dynamic offset of aligning the average phase voltage center, safety and hard constraints, equivalent time delay, and sampling and measurement index relationship into the controller.

[0042] Step 2, Midpoint Sampling and Data Acquisition: At the midpoint offset of each modulation cycle, the six-phase current, DC bus voltage, rotor position and speed, AT-AGV vehicle longitudinal acceleration, and wheel end slip ratio signals are simultaneously acquired, and the gate and carrier phases of the current period are recorded as the status input for the current cycle.

[0043] Step 3, Time Delay Modeling and Alignment: The three delays of sampling midpoint offset, current calculation wait, and six-phase inverter average output establishment are combined into two-step equivalent time delays, and time delay compensation parameters linked to the switching frequency are enabled in the controller to ensure that the model, sampling and modulation are aligned on the time axis.

[0044] Step 4, Coordinate Transformation and State Construction: Transform the six-phase current from the phasor domain to the direct axis d, quadrature axis q, and the two orthogonal axes x and y of the harmonic subspace of the stator synchronous rotating coordinate system to form the current state for prediction and optimization, and obtain the electric angular position and electric angular velocity at the same time.

[0045] Step 5, Online parameter identification and constraint: Recursively identify stator resistance, direct-axis inductance, quadrature-axis inductance, harmonic leakage inductance, and rotor permanent magnet flux linkage; correct resistance with temperature; apply safety boundaries and projection constraints to all parameters; and reject abnormal estimates.

[0046] Step 6, Current Prediction Forward: Based on the discrete motor-six-phase inverter model and two-phase equivalent time delay, the current response at the next sampling moment is predicted using the current state and historical voltage command, so as to obtain a calculable characterization of the current sampling error between two adjacent phases.

[0047] Step 7, Constrained optimization to solve for voltage reference values: With the goal of minimizing the current sampling error in the next step, the voltage reference values ​​of the four components of the direct axis d, quadrature axis q and harmonic subspace are obtained by combining the DC bus circular limit, the upper limit of the harmonic channel voltage ratio, and the voltage step limit constraint.

[0048] Step 8, Harmonic Compensation and Weight Scheduling: Cross-coupling compensation and second and sixth harmonic feedforward are superimposed in the harmonic subspace; the weights of the two harmonic axes are adaptively adjusted based on the attachment coefficient estimation, and an upper limit is set on the rate of weight change.

[0049] Step 9, Regeneration and Low-Speed ​​Micro-Crawling Control: Based on the bus voltage reference and allowable charge and discharge capacity provided by the battery management system, a three-segment DC side regeneration current trajectory is generated. The traction torque slope is planned according to the trajectory curvature and longitudinal slope angle. First-order filtering is performed on the speed difference and anti-saturation is set.

[0050] Step 10, Modulation and Distribution: The optimized voltage reference value is restored from the coordinate system to the six-phase stator voltage, the zero-sequence component is calculated and the dead zone and device voltage drop are compensated, the duty cycle of the six bridge arms is generated by space vector pulse width modulation, and the voltage is uniformly distributed to the gate driver at the next cycle boundary.

[0051] Step 11, Fault Detection and Fault Tolerance: Real-time monitoring of open circuit, overvoltage and overcurrent faults in bridge arms or phases. Once triggered, the current reference of each subspace is mapped to the remaining effective phase according to the preset reconstruction matrix, and the fault-tolerant modulation and current and voltage limiting strategy is switched to maintain the correctness of the coordinate index relationship.

[0052] Step 12, Feedback Closed Loop and Parameter Update: At the midpoint sampling point of the next cycle, obtain the current sample after the modulation effect, close the correspondence between the delay input and feedback of the two adjacent cycles, and automatically update the midpoint offset, equivalent time delay and related parameter table when the switching frequency, dead time, calculation time or stator resistance temperature changes. Feed back the available torque and energy constraints calculated in the current cycle to the planning module to make collaborative decisions on subsequent formation and path planning.

[0053] As one aspect of the method of the present invention, in step 1, based on the maximum allowable discharge current, the maximum allowable charging current and the bus voltage measurement issued by the battery management system, the upper limit of the DC side discharge current and the upper limit of the charging current are calculated and written into the constraint table, and the constraints are invoked in steps 7 and 9.

[0054] In step 2, press Calculate the midpoint offset and apply saturation constraints, then write the results to the timing parameter table. or When changes occur, the calculation is automatically recalculated and written back in step 12;

[0055] In step 3, the total time delay of the inner current loop is equivalent to a discrete two-step delay operator, and a first-order approximation is used in the continuous domain. When the switching frequency changes, it is updated synchronously in step 12. , and approximation coefficients;

[0056] Before proceeding to the solution in step 7, select the weight matrix and constraint parameters from the calibration table according to the vehicle speed, rotational speed, and DC bus voltage range. , , , , The maximum allowable discharge current, the maximum allowable charging current, and the DC bus circle limit from step 1 are combined to form an optimized constraint set.

[0057] In step 8, define and with , The rate of change limit is updated at a limited speed. This is the harmonic weighting matrix. , These are scalar weighting functions for the x and y harmonic axes, respectively.

[0058] In step 9, according to , , Generate a three-stage DC-side regenerative current reference value , , These are the maximum allowable charging current and maximum discharge current, respectively. The traction torque slope is set according to the trajectory curvature and longitudinal slope angle, and the speed differential filter coefficient is selected by the controller parameters according to the working conditions.

[0059] In step 11, current discontinuity, gate feedback consistency and DC bus and phase voltage abnormality criteria are used for detection. After triggering, a preset reconstruction matrix is ​​selected according to the fault type and switched to fault-tolerant SVPWM and current limiting and voltage limiting tables.

[0060] In step 12, the available torque, available power, and their respective effective ranges are summarized each cycle, written to the shared buffer, and reported to the planning module. Simultaneously, the factors... , , Or caused by temperature changes , , The approximation coefficients are updated synchronously.

[0061] Compared with the prior art, the present invention has the following technical effects:

[0062] This invention achieves rapid torque build-up under complex port vehicle transshipment conditions by introducing a collaborative strategy in a six-phase PMSM drive, including mid-cycle sampling and three-segment time-delay equivalent compensation, dqxy decoupling modeling and constrained quadratic optimization voltage solution, harmonic feedforward and cross-coupling compensation, adaptive parameter identification and bus circle limiting, regenerative braking three-segment current shaping, and fault reconstruction SVPWM. This results in low overshoot, low torque ripple and low current ripple, robust suppression of sampling and inverter delays and parameter drift, adaptive stabilization of DC bus fluctuations and adhesion changes, optimization of noise, vibration, and harshness (NVH), improved energy recovery efficiency, and consideration of safety boundaries and redundancy. This technology improves the tracking accuracy, response speed, available torque-energy constraint release, and turnover efficiency of vehicles during low-speed crawling, hill parking and restarting, precise parking alignment, and platooning. Furthermore, it maintains current loop bandwidth in harsh environments such as salt spray, high humidity, electromagnetic interference, slippery steel surfaces, and alternating slopes, reducing harmonic distortion rate and bus overvoltage risk, minimizing braking regeneration switching impact and wheel-end stick-slip vibration, suppressing device thermal peaks and ripple current, extending the lifespan of power devices and batteries, and reducing energy consumption and maintenance costs. Even with bridge arm degradation or phase open-circuit faults, it maintains XY channel decoupling and controllable torque strength, ensuring the AT-AGV is controllable and stopable. It also coordinates with path planning and platooning to achieve the release and scheduling of available power and torque, improving berth utilization and safety. Attached Figure Description

[0063] Figure 1 This is a system block diagram of the present invention;

[0064] Figure 2 This is a schematic diagram of the control scheme of the present invention;

[0065] Figure 3 This is a flowchart of the method of the present invention. Detailed Implementation

[0066] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0067] Example 1

[0068] like Figure 1As shown, the present invention proposes a drive control system for intelligent transfer equipment, including an AT-AGV. The AT-AGV is connected to a PMSM (Permanent Magnet Synchronous Motor), a DC bus, a controller, and a planning module. Both the PMSM and the DC bus are connected to a six-phase inverter. The six-phase inverter and the planning module are both connected to the controller. The DC side of the six-phase inverter is connected to the power battery pack of the AT-AGV via the DC bus, and the AC side of the six-phase inverter is connected to the six-phase stator windings of the PMSM.

[0069] The AT-AGV collects current, bus voltage, speed, acceleration, and slip ratio. The controller samples at the midpoint of each modulation cycle and uses three-segment time delay equivalent compensation of the sampling cycle to establish a dq coordinate system and an xy harmonic subspace model. d and q are the stator synchronous rotation coordinate axes, and x and y are the harmonic axes. Discrete moments are solved under SVPWM and DC bus voltage limiting. The voltage vector is mapped to the duty cycle of the six bridge arms and then sent to the gate drive of the six-phase inverter at the next cycle boundary. For the current discrete moment, the stator resistance, d-axis and q-axis inductance, harmonics and leakage inductance, and rotor permanent magnet flux linkage are identified online, and adaptive weighting is performed according to error sensitivity. Bus ripple is observed and adaptive current creep shaping is performed at the tail end of regenerative braking. When the bridge arm fails, phase reconstruction and current redistribution are performed. Road surface adhesion is estimated based on slip ratio and acceleration, and torque limiting and field weakening switching are linked. Feedforward cancellation and xy weight scheduling are implemented for the second and sixth harmonics. Torque constraints and speed differential correction are applied during low-speed micro-creep, ramp parking and restart, steel surface wet and stick-slip switching and bus fluctuation. Available torque-energy constraints are output to the planning module.

[0070] Figure 2 This is a schematic diagram of the control scheme of the present invention, as shown below. Figure 2As shown, through closed-loop coordination of "speed outer loop PI → voltage vector optimization → (fault-tolerant) SVPWM → six-phase inverter → dual three-phase PMSM", based on the modeling of "midpoint sampling alignment + two-step equivalent time delay", combined with current prediction, dq / xy transformation and xy coupling compensation, 2nd / 6th harmonic feedforward and weighted scheduling, the next step current is tracked without error, improving the torque build-up speed and suppressing torque pulsation and phase current ripple. At the same time, the introduction of bus observation and circular limiting, regenerative end current shaping and step constraint improves the stability and energy recovery efficiency under regenerative braking and bus fluctuation. When faults such as bridge arm / phase open circuit occur, fault detection and reconstruction realize fault-tolerant modulation and controllable degradation. With the torque limiting of speed derivative and adhesion estimation, field weakening switching and zero-sequence injection, the influence of parameter drift and time delay is weakened, NVH and slippage are reduced, and finally, a fast, low overshoot, low noise, low energy consumption and robust drive control effect is achieved in port conditions such as low-speed micro-crawling, ramp parking / restart and adhesion sudden change.

[0071] It should be noted that the controller sets the current sampling time of each modulation cycle as the midpoint of the cycle and introduces a dynamic offset. Align the sampling time with the midpoint of the period of the average voltage of the six-phase inverter, and set the three-segment time delay equivalent as follows: And all are unified as equivalent time delays , For the current modulation period, , , These are the three time delays: sampling midpoint offset, calculation or holding until the next sampling cycle update in the current sampling period, and the average voltage of the six-phase inverter. set up, The PWM switching frequency of the six-phase inverter, with equivalent time delay. The delay in the continuous domain is approximated by a first-order approximation, which is expressed as: It updates synchronously with the pulse width modulation switching frequency of the six-phase inverter. For the Laplace operator;

[0072] dynamic offset The formula for obtaining it is ; The dead time of the bridge arm is set by the gate drive of the six-phase inverter. To control computation time, the CPU cycle count is used by the controller; , All are dimensionless calibration coefficients, obtained by performing step or frequency sweep tests on AT-AGV and fitting the time difference between the planned sampling midpoint and the measured average voltage midpoint using the least squares criterion.

[0073] By fixing the current sampling at the midpoint of the PWM cycle and using Online calibration aligns the sampling time with the actual effective center of the average phase voltage of the six-phase inverter, while also... , , The three-stage lag unification is equivalent to and use Padé's first-order approximation follows Adaptive updates, compared to existing technologies that rely on periodic initial sampling, ignoring dead zones / computational lags, or using fixed time delays and fixed amplitude limits, reduce phase and amplitude errors in voltage-current models, suppress aliasing and sampling jitter, reduce sensitivity to stator resistance temperature rise and parameter drift, improve the reachability and adjustability of deadbeat prediction for the next current, expand the stable bandwidth, and reduce overshoot, harmonic current, and torque ripple. They are more stable and efficient in bus fluctuations and regenerative tail sections, and respond faster, have lower NVH, smoother energy recovery, and more predictable fault tolerance degradation in low-speed micro-crawling, ramp parking and restart, and frequent start-stop scenarios.

[0074] It should be noted that the timing and execution steps of the controller's sampling, equivalent time delay modeling, voltage prediction solution, limiting, and SVPWM distribution within the modulation period include:

[0075] Step S1, Define the time base and period: Define the modulation period Define the periodic boundary ;

[0076] Step S2, Midpoint Sampling Alignment: Set the sampling time of the current midpoint. Dynamic offset ;

[0077] Step S3, Time Delay Decomposition and Summary: Decompose the total time delay of the inner current loop into... These correspond to the sampling midpoint offset, the current calculation / waiting for the next cycle update, and the formation of the six-phase inverter average voltage, respectively, and are summarized into an equivalent time delay. ;

[0078] Step S4, Continuous Domain Approximation: In the continuous time domain, use the transfer function with pure time delay. The equivalent time delay is represented using a first-order approximation. ;

[0079] Step S5, Discrete-domain time delay implementation: A two-step delay operator is used in the discrete-time domain. , It is a discrete shift operator. Two-beat delay, and midpoint sampling sequence With voltage reference Establish index relationships, The discrete current quantity sampled at the midpoint. They are equal in the limiting sense.

[0080] Step S6, Modulation command generation and issuance: at time... The discrete voltage reference is mapped to a six-phase space vector pulse width modulation duty cycle. , The duty cycle vector of the six bridge arms is obtained through the SVPWM mapping algorithm and output to the gate drive of the six-phase inverter;

[0081] Step S7, Feedback Data Acquisition and Indexing: at time... Collect current samples , and the result obtained in step S5 A two-beat delay relationship is formed;

[0082] Step S8, Parameter Update and Range Constraints: When When changing Synchronous updates and the approximation coefficients of step S4, and for Using saturation constraints , These are the experimentally fitted values. , To allow upper and lower limits, obtained through security and hardware constraints, the updated limits will be... , , Write the timing and parameter table to the controller and use it for limiting, prediction, and SVPWM mapping in the next cycle.

[0083] By implementing steps S1-S8 in a time-sequential manner, sampling points, time delay modeling, discrete prediction, limiting, and SVPWM distribution are formed in a closed loop on the same time base, ensuring that the midpoint sampling of the current and the effective center of the average phase voltage of the six-phase inverter are aligned. Real-time alignment, three time delays are unified and equivalent to It is characterized by the Padé first-order approximation and the two-step discrete operator, combined with the bus circle limiting and step constraint. The output can achieve the duty cycle, and follow... Online updates , , This reduces model phase / amplitude mismatch and aliasing errors, suppresses sampling jitter and computational lag-induced oscillations, improves the reachability of deadbeat prediction for the next current and the feasible domain of SVPWM modulation, expands the current loop stability bandwidth, reduces overshoot, harmonic current and torque ripple, and provides smoother and more stable bus fluctuations and regenerative tail sections. It also enables faster torque build-up, lower NVH and energy consumption in port operating conditions such as low-speed micro-climb / frequent start-stop, ramp parking / restart.

[0084] It should be noted that the controller controls the six-phase current vectors. Angular correlation transformation matrix get , For the first The stator current vector in the dq and xy coordinate systems at each sampling time. , , , , , These are the component current vectors of the six-phase current at each point. This refers to the spatial phase shift angle of the winding. The electric angle position, It consists of two sets of Clark-Parker transformations, the first set being three-phase. Current angle transformation according to , , , The stator stationary coordinate system for the first group of three-phase currents , Components; second group of three phases Angle transformation according to , , , The stator stationary coordinate system for the second group of three-phase currents , Components; and establish the current update formula according to backward differential discretization:

[0085] ;

[0086] ;

[0087] ;

[0088] ;

[0089] Equivalently written in matrix form:

[0090] ;

[0091] in, Electric angular velocity, , , , The six-phase currents are respectively The d-axis, q-axis, x-axis, and y-axis components at the sampling time. , , , The six-phase currents are respectively The d-axis, q-axis, x-axis, and y-axis components at the sampling time. For the first The stator current vector in the dq and xy coordinate systems at each sampling time. For stator resistance, For d-axis inductance, It is the q-axis inductance. For harmonics or leakage inductance, For rotor permanent magnet flux linkage, Let dq and xy be the voltage vectors. For the reason , , , A defined coupling matrix, Multiply the admittance matrix by The coefficients of dispersion, Given a known disturbance vector introduced by the back electromotive force term, all trigonometric functions and matrix operations are calculated in real time within the controller and updated accordingly. , , renew.

[0092] To project the spatially misaligned fundamental and harmonic currents of the two sets of three-phase currents (a,b,c) and (g,e,f) of the dual three-phase stator onto a coordinate subspace that is easy to control, and to ensure that the prediction model is consistent with "midpoint sampling - zero-order hold - two-step equivalent time delay", the controller obtains the six-phase currents in real time through two sets of Clark-Parker transformations containing phase displacement angle γ. Then construct using backward differential discretization , , The linear time-varying state equation, thus under known parameters and rotational speed ,cycle Below, with Accurate prediction It also solves the optimal voltage that satisfies the bus circle limiting and modulation feasible region, achieving the technical effects of decoupling the fundamental torque control and xy harmonic suppression, reducing parameter drift and coupling error, improving numerical stability and bandwidth, reducing current ripple and torque pulsation, and enhancing robustness to bus fluctuations and low-speed micro-creep / regeneration conditions.

[0093] It should be noted that the controller, under the constraints of DC bus circular limiting and modulation realizability, adjusts the voltage vector... At discrete time Solve using the following formula:

[0094] ;

[0095] Constraints , , , And the optimal Inverse Transformation The voltage is mapped to a six-phase stator phase voltage reference, and then the duty cycle is generated by symmetrically center-aligned six-phase space vector pulse width modulation. ,in, ;

[0096] in, For the dq and xy current reference, The sampled current is the current at the midpoint of the current discrete moment. It is a diagonal weighted matrix. For voltage regularization weights, This is the DC bus voltage. , The upper limit of the xy channel voltage ratio, with values ​​between 0 and 1. This is the voltage solution from the previous sampling period. This is the upper limit of the voltage step. electric angle Phase shift angle with winding space The six-phase to dq and xy inverse transformation matrices For the first time after adding the zero-order component The phase voltage reference is obtained by applying the zero-sequence component symmetry according to the duty cycle and the modulation margin rule. Correction of device voltage drop step size and duty cycle;

[0097] For saturation operators, , To modulate the lower and upper limits.

[0098] To ensure that deadbeat control for "next-step arrival" is both solvable and implementable under the complex operating conditions of port AT-AGVs, where DC bus voltage is limited, modulation realizable domain is finite, and bus voltage fluctuates with sudden changes in regenerative power and load, this invention expresses the voltage solution as a weighted quadratic optimization while simultaneously applying circular limiting. upper limit of xy channel ratio , Step constraints Duty cycle saturation after zero-sequence injection Then, with the electric angle With phase shift The inverse transform of the optimal Mapped to six-phase reference values, with device voltage drop step size correction superimposed. And SVPWM is aligned with the center of symmetry for landing; predictive-constraint-modulation closed-loop control can eliminate distortion caused by overmodulation and clipping, ensure that the gate sequence can be consistent with the bus and device safety boundaries, limit voltage jumps between adjacent cycles, suppress current spikes and electromagnetic noise, and in It maintains model-execution consistency even when fluctuations, 2nd / 6th harmonics, and parameter drift are present, thereby achieving faster torque build-up, lower current ripple and torque pulsation, smaller NVH and switching losses, smoother regenerative tail-end energy recovery, and improved robust stability and vehicle energy efficiency in scenarios such as low-speed micro-crawling, hill-start / parking / restarting, and frequent start-stop.

[0099] It should be noted that the controller has a parameter vector A recursive least squares method with a forgetting factor and projection constraints are applied.

[0100] ;

[0101] in, , For bounded sets The projection operator is applied to the stator resistance, and temperature correction is used. , For parameter estimates, For covariance matrix, , These are the regression vector and the observation vector obtained by rearranging the discrete voltage-current equations, respectively. For the numerical value in The forgetting factor is set by the controller parameter table; , They are bounded sets The lower and upper limits are given by the safety constraints; This is the initial value of the stator resistance. This is the temperature correction factor for stator resistance. , All are determined by the material data of the stator resistance. This is the current stator resistance temperature. The reference temperature for stator resistance temperature correction.

[0102] By the parameter vector A recursive least squares approach with a forgetting factor is employed, and projection constraints are applied. The stator resistance is modeled according to a temperature model. Real-time correction enables model parameters to converge rapidly and be confined within the physically feasible range under conditions such as thermal drift, magnetic saturation, sudden changes in load and speed, and device aging. This avoids voltage solution distortion introduced by estimation divergence and infeasible solutions. Consequently, it improves the consistency of voltage-current predictions and the accessibility of beat-free solutions, expands the current loop stability bandwidth, reduces sensitivity to parameter mismatch and numerical jitter, suppresses harmonic current and torque ripple, noise and vibration, and enhances the stability of bus fluctuations and the tail end of regenerative braking, the smoothness of energy recovery, and the reduction of low-speed micro-climbing of the entire vehicle. The system enhances robustness in driving, hill start-up, and frequent start-stop operations, while maintaining timing consistency between online identification results and constraint optimization, SVPWM modulation, current midpoint sampling, and two-step equivalent time delay modeling. This ensures that the voltage solution for the next step is consistent with the actual gate execution. Simultaneously, the projection operator H limits the parameter vector within the manufacturer's parameters and safety boundaries, avoiding overfitting and estimation drift accumulation. This reduces distortion and saturation during extreme temperature rises and bus fluctuations, improving long-term operational stability, maintainability, and cross-platform portability, as well as enhancing energy efficiency and reliability.

[0103] It should be noted that the controller employs cross-coupled step size and harmonic feedforward, and performs weighted scheduling in the xy harmonic subspace:

[0104] ;

[0105] , ,and It is a 2×2 cross-coupling compensation matrix, whose coefficients are determined by the electric angular velocity. and Calculated and calibrated, , The amplitudes of the 2nd and 6th harmonics are respectively. , The phase offsets are for the 2nd and 6th harmonics, respectively. For the xy voltage components, For estimation based on adhesion coefficient The weight matrix, Estimated by fusion of wheel speed difference and acceleration, , It is a piecewise affine function, and the upper limit of the rate of change of the weights of the piecewise affine function is set by the controller. It is a sinusoidal feedforward.

[0106] By introducing electric angular velocity into the xy harmonic subspace With leakage Adaptive 2×2 cross-coupling compensation This is used to cancel the mutual interference between harmonic channels caused by leakage inductance coupling and non-ideal modulation, and to superimpose a sinusoidal feedforward with a fixed phase bias for the 2nd / 6th harmonics. To pre-eliminate major harmonic sources such as tooth grooves, dead zones, and biases, and to estimate the weight matrix based on the adhesion coefficient. Implementing segmented affine scheduling avoids overcompensation and modulation saturation under low-adhesion and regeneration conditions, thereby reducing current ripple and torque pulsation, noise and vibration (NVH) in the xy channel under bus fluctuations, parameter drift, and frequent start-stop conditions. It maintains the modulability and duty cycle smoothness of SVPWM, improves torque controllability, energy recovery smoothness, and system robustness during low-speed micro-crawling, hill start-up, and formation alignment, and reduces switching losses and thermal stress, thereby improving reliability.

[0107] It should be noted that the controller employs a three-stage DC-side current trajectory and torque slope design and includes fault mapping under regenerative braking and low-speed micro-creep conditions. The reference value for the DC-side regenerative current is:

[0108] ;

[0109] in, , , , , From bus voltage reference value Determined by its rate of change, the traction torque reference slope plan uses trajectory curvature and longitudinal slope angle as independent variables. The trajectory curvature is obtained through trajectory planning, and the longitudinal slope angle is obtained through slope estimation. The speed differential correction uses a first-order filter and incorporates anti-saturation logic. This refers to the DC-side target current of the intermediate constant current segment in the three-segment regenerative braking current trajectory. , The target regenerative braking power, converted from braking demand, is calculated from the deceleration request and AT-AGV parameters. The maximum allowable charging current given to the battery management system of the AT-AGV. This represents the upper limit of the DC-side discharge current. When an open-circuit fault is detected in a bridge arm or phase, a preset reconstruction matrix is ​​used to map the determinant of the current reference values ​​of each harmonic subspace to the remaining effective phases and switch to fault-tolerant space vector modulation and current-limiting / voltage-limiting configuration. The preset reconstruction matrix is ​​obtained through fault-tolerant configuration and six-phase inverter topology, and the current reference values ​​of each harmonic subspace are obtained by current outer loop allocation and inner loop limiting. The battery management system and energy management strategy of the AT-AGV are given.

[0110] By introducing a three-stage DC-side current trajectory under regenerative braking and low-speed micro-creep conditions With the torque slope plan, the connection between the rise-constant-fall three stages is determined by the bus voltage reference and its rate of change. First-order velocity differential estimation is used to suppress noise and integral drift using anti-saturation logic. The constant current section is limited to the battery's allowable range, and the duty cycle step constraint and circular amplitude limiter ensure the feasibility of modulation. When the bridge arm / phase is open, the preset reconstruction matrix maps the determinant of the current reference values ​​of each harmonic subspace to the remaining phase and switches the fault-tolerant SVPWM and current and voltage limiting, so that the bus power distribution, regenerative energy recovery and adjustable torque are coordinated and consistent, reducing bus fluctuations and current spikes in the regenerative tail section, reducing torque impact and NVH, suppressing low-adhesion slippage, improving energy recovery efficiency, smoothness and controllability of micro-speed crawling and hill parking / start-up, and maintaining continuous availability and robust stability in formation alignment, frequent start-stop and fault scenarios, while reducing switching losses and thermal stress.

[0111] Example 2

[0112] The difference between Embodiment 2 and Embodiment 1 is that this embodiment introduces a drive control method for intelligent transfer equipment.

[0113] like Figure 3 As shown, the present invention proposes a drive control method for intelligent transfer equipment, applying the drive control system for intelligent transfer equipment described in Embodiment 1. The method includes:

[0114] Step 1, System Initialization and Time Base Setting: Set the pulse width modulation switching frequency of the six-phase inverter to determine the modulation period; establish the start and end boundaries of each cycle and the trigger time of the midpoint sampling of the cycle; calibrate the small offset of the midpoint sampling based on the dead zone of the bridge arm and the control calculation time, and write these timing parameters into the controller.

[0115] Step 2, Midpoint Sampling and Data Acquisition: At the midpoint offset of each modulation cycle, the six-phase current, DC bus voltage, rotor position and speed, AT-AGV vehicle longitudinal acceleration, and wheel end slip ratio signals are simultaneously acquired, and the gate and carrier phases of the current period are recorded as the status input for the current cycle.

[0116] Step 3, Time Delay Modeling and Alignment: The three delays of sampling midpoint offset, current calculation wait, and six-phase inverter average output establishment are combined into two-step equivalent time delays, and time delay compensation parameters linked to the switching frequency are enabled in the controller to ensure that the model, sampling and modulation are aligned on the time axis.

[0117] Step 4, Coordinate Transformation and State Construction: Transform the six-phase current from the phasor domain to the direct axis d, quadrature axis q, and the two orthogonal axes x and y of the harmonic subspace of the stator synchronous rotating coordinate system to form the current state for prediction and optimization, and obtain the electric angular position and electric angular velocity at the same time.

[0118] Step 5, Online parameter identification and constraint: Recursively identify stator resistance, direct-axis inductance, quadrature-axis inductance, harmonic leakage inductance, and rotor permanent magnet flux linkage; correct resistance with temperature; apply safety boundaries and projection constraints to all parameters; and reject abnormal estimates.

[0119] Step 6, Current Prediction Forward: Based on the discrete motor-six-phase inverter model and two-step equivalent time delay, the current response at the next sampling moment is predicted using the current state and historical voltage command, so as to obtain a calculable characterization of the error of the next step.

[0120] Step 7, Constrained optimization to solve for voltage reference values: With the goal of minimizing the current deviation in the next step, and combining the DC bus circular limit, the upper limit of the harmonic channel voltage ratio, and the voltage step limit constraint, the voltage reference of the four components of the direct axis d, the quadrature axis q, and the harmonic subspace is obtained;

[0121] Step 8, Harmonic Compensation and Weight Scheduling: Cross-coupling compensation and second and sixth harmonic feedforward are superimposed in the harmonic subspace; the weights of the two harmonic axes are adaptively adjusted based on the attachment coefficient estimation, and an upper limit is set on the rate of weight change.

[0122] Step 9, Regeneration and Low-Speed ​​Micro-Crawling Control: Based on the bus voltage reference and allowable charge and discharge capacity provided by the battery management system, a three-segment DC side regeneration current trajectory is generated. The traction torque slope is planned according to the trajectory curvature and longitudinal slope angle. First-order filtering is performed on the speed difference and anti-saturation is set.

[0123] Step 10, Modulation and Distribution: The optimized voltage reference is restored from the coordinate system to the six-phase stator voltage, the zero-sequence component is calculated and the dead zone and device voltage drop are compensated, the duty cycle of the six bridge arms is generated by space vector pulse width modulation, and the voltage is uniformly distributed to the gate driver at the next cycle boundary.

[0124] Step 11, Fault Detection and Fault Tolerance: Real-time monitoring of open circuit, overvoltage and overcurrent faults in bridge arms or phases. Once triggered, the current reference of each subspace is mapped to the remaining effective phase according to the preset reconstruction matrix, and the fault-tolerant modulation and current and voltage limiting strategy is switched to maintain the correctness of the coordinate index relationship.

[0125] Step 12, Feedback closed loop and parameter update: At the midpoint sampling point of the next cycle, obtain the current sample after the modulation effect, close the two-step correspondence, and when the switching frequency, dead time, calculation time or stator resistance temperature change, automatically update the midpoint offset, equivalent time delay and related parameter table, and feed back the available torque and energy constraints calculated in the current cycle to the planning module for collaborative decision-making on subsequent formation and path planning.

[0126] Through the time-sequential coordination of steps 1 to 12, midpoint sampling, two-phase equivalent time delay, discrete prediction and constrained optimization, voltage modulation and zero-sequence injection, online parameter identification and temperature correction, xy coupling compensation and second / sixth harmonic feedforward, three-stage regeneration and torque slope planning, fault detection and reconstruction, and planning collaborative closed loop are unified under the same time base. This achieves consistency and feasibility between the current-voltage model and gate execution, expands the current loop stability bandwidth, and reduces overshoot, harmonic current and torque ripple, bus fluctuation and regeneration tail-end impact. This improves torque controllability, energy recovery efficiency, NVH and energy consumption performance during low-speed micro-crawling, hill start / restart and formation alignment, and maintains controllable degradation and continuous availability under faults such as bridge arm / phase open circuit. At the same time, it reduces the sensitivity to stator resistance temperature rise and inductance deviation, suppresses oscillations induced by sampling jitter and calculation delay, constrains duty cycle step and voltage jump between adjacent cycles, reduces electromagnetic noise and switching losses, ensures the safety boundary of bus, device and battery management, and improves the long-term stability, maintainability and all-condition robustness of the system.

[0127] It should be noted that in step 1, based on the maximum allowable discharge current, the maximum allowable charging current and the bus voltage measurement issued by the battery management system, the upper limit of the DC side discharge current and the upper limit of the charging current are calculated and written into the constraint table. The constraints are invoked in steps 7 and 9.

[0128] In step 2, press Calculate the midpoint offset and apply saturation constraints, then write the results to the timing parameter table. or When changes occur, the calculation is automatically recalculated and written back in step 12;

[0129] In step 3, the total time delay of the inner current loop is equivalent to a discrete two-step delay operator, and a first-order approximation is used in the continuous domain. When the switching frequency changes, it is updated synchronously in step 12. , and approximation coefficients;

[0130] Before proceeding to the solution in step 7, select the weight matrix and constraint parameters from the calibration table according to the vehicle speed, rotational speed, and DC bus voltage range. , , , , The maximum allowable discharge current, the maximum allowable charging current, and the DC bus circle limit from step 1 are combined to form an optimized constraint set.

[0131] In step 8, define and with , The rate of change limit is updated at a limited speed. This is the harmonic weighting matrix. , These are scalar weighting functions for the x and y harmonic axes, respectively.

[0132] In step 9, according to , , Generate a three-stage DC-side regenerative current reference value , , These are the maximum allowable charging current and maximum discharge current, respectively. The traction torque slope is set according to the trajectory curvature and longitudinal slope angle, and the speed differential filter coefficient is selected by the controller parameters according to the working conditions.

[0133] In step 11, current discontinuity, gate feedback consistency and DC bus and phase voltage abnormality criteria are used for detection. After triggering, a preset reconstruction matrix is ​​selected according to the fault type and switched to fault-tolerant SVPWM and current limiting and voltage limiting tables.

[0134] In step 12, the available torque, available power, and their respective effective ranges are summarized each cycle, written to the shared buffer, and reported to the planning module. Simultaneously, the factors... , , Or caused by temperature changes , , The approximation coefficients are updated synchronously.

[0135] Based on the charging and discharging limits and bus voltage given by the BMS, the current constraint is calculated and issued in real time, so that the optimization and regeneration planning are constrained by the battery / bus safety boundary throughout the process, avoiding overcurrent and thermal stress, and improving energy recovery and system stability;

[0136] according to Calculate the midpoint offset and apply saturation constraints to reduce phase / amplitude modeling errors and improve current prediction accuracy and effective current loop bandwidth.

[0137] The total time delay is equivalent to a two-beat discrete delay and characterized by a first-order approximate unified continuous domain. Parameters are updated online to ensure the model is consistent with the execution timeline, suppressing latency-induced oscillations and enhancing stability margin;

[0138] Select by vehicle speed / RPM / bus range before entering optimization. , , , , The current and circular limiting constraints from step 1 are combined to ensure that the voltage solution is both solvable and modulated, taking into account both fast torque response and low ripple.

[0139] by It also limits the speed of the updated weights, so that the harmonic compensation adapts to the adhesion without changing excessively, avoiding flutter and saturation, and improving NVH and traction stability under low adhesion and regeneration conditions.

[0140] in accordance with , , Generate a three-stage regenerative current and plan the torque slope according to curvature / longitudinal slope. Segmented settings smooth the regeneration and micro-creep processes, suppressing bus fluctuations and current spikes, and improving comfort and energy efficiency;

[0141] The fault-tolerant SVPWM and current and voltage limiting are quickly diagnosed based on current discontinuity, gate consistency and bus / phase voltage anomaly criteria, and the fault mapping matrix is ​​used to switch between fault-tolerant SVPWM and current and voltage limiting to ensure controllable degradation, device protection and continuous availability under fault conditions.

[0142] Each cycle of feedback provides available torque / power range for planning coordination, and also addresses the factors... , , Or caused by temperature changes , , Real-time synchronization with approximation coefficients ensures consistency between planning and execution, and that the model stays true to reality over the long term, improving robustness and efficiency.

[0143] In summary, this invention achieves rapid torque build-up under complex port vehicle transshipment conditions by introducing a collaborative strategy in a six-phase PMSM drive. This strategy includes midpoint sampling and three-segment time delay equivalent compensation, dqxy decoupling modeling and constrained quadratic optimization voltage solution, harmonic feedforward and cross-coupling compensation, adaptive parameter identification and bus circle limiting, regenerative braking three-segment current shaping, and fault reconstruction SVPWM. It also achieves robust suppression of low overshoot, low torque ripple and low current ripple, sampling and inverter delay, and parameter drift. Furthermore, it adaptively stabilizes DC bus fluctuations and adhesion changes, optimizes NVH, improves energy recovery efficiency, and balances safety boundaries and redundancy tolerance, enabling vehicles to perform well in low-speed crawling, hill-start parking, and restarting. This invention improves the tracking accuracy, response speed, available torque-energy constraint release, and turnaround efficiency of precise parking alignment and platooning. Furthermore, it maintains current loop bandwidth under harsh environments such as salt spray, high humidity, electromagnetic interference, slippery steel surfaces, and alternating slopes, reducing harmonic distortion rate and bus overvoltage risk, minimizing braking regeneration switching impact and wheel-end stick-slip vibration, suppressing device thermal peaks and ripple current, extending power device and battery life, and reducing energy consumption and maintenance costs. Even with bridge arm degradation or phase open circuit faults, it maintains XY channel decoupling and controllable torque strength, ensuring the AT-AGV is controllable and stopable. It also coordinates with path planning and platooning to achieve the release and scheduling of available power and torque, improving berth utilization and safety. The above description is merely a specific embodiment of the invention, but the scope of protection of the invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this invention should be included within the scope of protection of this invention. Therefore, the scope of protection of this invention should be determined by the scope of the claims.

[0144] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0145] The above provides a detailed description of the drive control system and control method for an intelligent transfer equipment provided by the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A drive control system for intelligent transfer equipment, characterized in that, This includes AT-AGVs, which are connected to PMSMs, DC buses, controllers, and planning modules. Both the PMSMs and DC buses are connected to six-phase inverters, and both the six-phase inverters and planning modules are connected to the controller. AT-AGV collects current, bus voltage, speed, acceleration, and slip ratio; The controller, based on midpoint sampling of the modulation period and employing three-segment time-delay equivalent compensation, establishes a dq coordinate system and an xy harmonic subspace model, where d and q are the stator synchronous rotation axes, and x and y are the harmonic axes. Under SVPWM and bus voltage limiting, the solution is obtained... The voltage vector at each moment is mapped to the duty cycle of the six bridge arms and then sent to the gate drive of the six-phase inverter at the next cycle boundary. For discrete moments, stator resistance, d-axis and q-axis inductance, harmonics and leakage inductance, and rotor permanent magnet flux linkage are identified online. The weights of each axis of current tracking error, voltage regularization weights, and confidence weights of parameters obtained online are adaptively weighted according to error sensitivity. Bus ripple observation and adaptive current creep shaping are performed at the tail end of regenerative braking. When the bridge arm fails, phase reconstruction and current redistribution are performed. Road surface adhesion is estimated based on slip ratio and acceleration. Torque limiting and field weakening switching are linked. Feedforward cancellation and xy weight scheduling are implemented for second and sixth harmonics. Torque constraints and speed differential correction are applied during low-speed micro-creep, ramp parking and restart, steel surface wet and stick-slip switching, and bus fluctuation. Available torque-energy constraints are output to the planning module. The controller sets the current sampling time of each modulation cycle as the midpoint of the cycle and introduces a dynamic offset. Align the sampling time with the midpoint of the period of the average voltage of the six-phase inverter, and set the three-segment time delay equivalent as follows: And all are unified as equivalent time delays , For the current modulation period, , , These are the three time delays: sampling midpoint offset, calculation or holding until the next sampling cycle update in the current sampling period, and the average voltage of the six-phase inverter. set up, The PWM switching frequency of the six-phase inverter, with equivalent time delay. The delay in the continuous domain is approximated by a first-order approximation, which is expressed as: It updates synchronously with the pulse width modulation switching frequency of the six-phase inverter. For the Laplace operator; dynamic offset The formula for obtaining it is ; The dead time of the bridge arm is set by the gate drive of the six-phase inverter. To control computation time, the CPU cycle count is used by the controller; , All are dimensionless calibration coefficients, obtained by performing step or frequency sweep tests on AT-AGV and fitting the time difference between the planned sampling midpoint and the measured average voltage midpoint using the least squares criterion; The timing and execution steps of the controller during the modulation period, including sampling, equivalent time delay modeling, voltage prediction solution, limiting, and SVPWM distribution, include: Step S1, Define the time base and period: Define the modulation period Define the periodic boundary ; Step S2, Midpoint Sampling Alignment: Set the sampling time of the current midpoint. Dynamic offset ; Step S3, Time Delay Decomposition and Summary: Decompose the total time delay of the inner current loop into... These correspond to the sampling midpoint offset, the current calculation / waiting for the next cycle update, and the formation of the six-phase inverter average voltage, respectively, and are summarized into an equivalent time delay. ; Step S4, Continuous Domain Approximation: In the continuous time domain, use the transfer function with pure time delay. The equivalent time delay is represented using a first-order approximation. ; Step S5, Discrete-domain time delay implementation: A two-step delay operator is used in the discrete-time domain. , It is a discrete shift operator. Two-beat delay, and midpoint sampling sequence With voltage reference Establish index relationships, For discrete current quantities sampled at the midpoint, They are equal in the limiting sense. Step S6, Modulation command generation and issuance: at time... The discrete voltage reference is mapped to a six-phase space vector pulse width modulation duty cycle. , The duty cycle vector of the six bridge arms is obtained through the SVPWM mapping algorithm and output to the gate drive of the six-phase inverter; Step S7, Feedback Data Acquisition and Indexing: at time... Acquiring current samples The result obtained in step S5 A two-beat delay relationship is formed; Step S8, Parameter Update and Range Constraints: When When changing Synchronous updates and the approximation coefficients of step S4, and for Using saturation constraints , These are the experimentally fitted values. , To allow upper and lower limits, obtained through security and hardware constraints, the updated limits will be... , , Write the timing and parameter table to the controller and use it for limiting, prediction, and SVPWM mapping in the next cycle.

2. The drive control system for an intelligent transfer equipment according to claim 1, characterized in that, The controller controls the six-phase current vector Angular correlation transformation matrix get , For the first The stator current vector in the dq and xy coordinate systems at each sampling time. , , , , , These are the component current vectors of the six-phase current at each point. The spatial phase shift angle of the winding. The electric angle position, It consists of two sets of Clark-Parker transformations, the first set being three-phase. Current angle transformation according to , , , The stator stationary coordinate system for the first group of three-phase currents , Components; second group of three phases Angle transformation according to , , , The stator stationary coordinate system for the second group of three-phase currents , Components; and establish the current update formula according to backward differential discretization: ； ； ； ； Equivalently written in matrix form: ； in, Electric angular velocity, , , , The six-phase currents are respectively The d-axis, q-axis, x-axis, and y-axis components at the sampling time. , , , The six-phase currents are respectively The d-axis, q-axis, x-axis, and y-axis components at the sampling time. For the first The stator current vector in the dq and xy coordinate systems at each sampling time. For stator resistance, For d-axis inductance, It is the q-axis inductance. For harmonics or leakage inductance, For rotor permanent magnet flux linkage, Let dq and xy be the voltage vectors. For the reason , , , A defined coupling matrix, Multiply the admittance matrix by The coefficients of dispersion, Given a known disturbance vector introduced by the back electromotive force term, all trigonometric functions and matrix operations are calculated in real time within the controller and updated accordingly. , , renew.

3. The drive control system for an intelligent transfer equipment according to claim 2, characterized in that, The controller operates on the voltage vector under the constraints of DC bus circular limiting and modulation realizability. At discrete time Solve using the following formula: ； Constraints , , , And the optimal Inverse Transformation The voltage is mapped to a six-phase stator phase voltage reference, and then the duty cycle is generated by symmetrically center-aligned six-phase space vector pulse width modulation. ,in, ; in, For the dq and xy current reference, It is a diagonal weighted matrix. For voltage regularization weights, This is the DC bus voltage. , The upper limit of the xy channel voltage ratio, with values ​​between 0 and 1. This is the voltage solution from the previous sampling period. This is the upper limit of the voltage step. electric angle Phase shift angle with winding space The six-phase to dq and xy inverse transformation matrices For the first time after adding the zero-order component The phase voltage reference is obtained by applying the zero-sequence component symmetry according to the duty cycle and the modulation margin rule. Correction of device voltage drop step size and duty cycle; For saturation operators, , To modulate the lower and upper limits.

4. The drive control system for an intelligent transfer equipment according to claim 3, characterized in that, Controller on parameter vector A recursive least squares method with a forgetting factor and projection constraints are applied. ; in, , For bounded sets The projection operator is applied to the stator resistance, and temperature correction is used. , For parameter estimates, For covariance matrix, , These are the regression vector and the observation vector obtained by rearranging the discrete voltage-current equations, respectively. For the numerical value in The forgetting factor is set by the controller parameter table; , They are bounded sets The lower and upper limits are given by the safety constraints; This is the initial value of the stator resistance. This is the temperature correction factor for stator resistance. , All are determined by the material data of the stator resistance. This is the current stator resistance temperature. The reference temperature for stator resistance temperature correction.

5. The drive control system for an intelligent transfer equipment according to claim 1, characterized in that, The controller employs cross-coupled step size and harmonic feedforward in the xy harmonic subspace and performs weighted scheduling.

6. The drive control system for an intelligent transfer equipment according to claim 3, characterized in that, The controller employs a three-segment DC-side current trajectory and torque slope planning under regenerative braking and low-speed micro-creep conditions, while simultaneously performing fault mapping.

7. A drive control method for intelligent transfer equipment, employing the drive control system for intelligent transfer equipment as described in claim 3, characterized in that, The method includes: Step 1, System Initialization and Time Base Setting: Set the pulse width modulation switching frequency of the six-phase inverter to determine the modulation period; establish the start and end boundaries of each cycle and the trigger time of midpoint sampling in the cycle; calibrate the small offset of midpoint sampling based on the dead zone of the bridge arm and the control calculation time; and write the six-phase inverter PWM switching frequency, modulation base period, start and end boundaries of the cycle, carrier phase mark, duty cycle issuance time, midpoint sampling trigger time, dynamic offset of aligning the average phase voltage center, safety and hard constraints, equivalent time delay, and sampling and measurement index relationship into the controller. Step 2, Midpoint Sampling and Data Acquisition: At the midpoint offset of each modulation cycle, the six-phase current, DC bus voltage, rotor position and speed, AT-AGV vehicle longitudinal acceleration, and wheel end slip ratio signals are simultaneously acquired, and the gate and carrier phases of the current period are recorded as the status input for the current cycle. Step 3, Time Delay Modeling and Alignment: The three delays of sampling midpoint offset, current calculation wait, and six-phase inverter average output establishment are combined into two-step equivalent time delays, and time delay compensation parameters linked to the switching frequency are enabled in the controller to ensure that the model, sampling and modulation are aligned on the time axis. Step 4, Coordinate Transformation and State Construction: Transform the six-phase current from the phasor domain to the direct axis d, quadrature axis q, and the two orthogonal axes x and y of the harmonic subspace of the stator synchronous rotating coordinate system to form the current state for prediction and optimization, and obtain the electric angular position and electric angular velocity at the same time. Step 5, Online parameter identification and constraint: Recursively identify stator resistance, direct-axis inductance, quadrature-axis inductance, harmonic leakage inductance, and rotor permanent magnet flux linkage; correct resistance with temperature; apply safety boundaries and projection constraints to all parameters; and reject abnormal estimates. Step 6, Current Prediction Forward: Based on the discrete motor-six-phase inverter model and two-phase equivalent time delay, the current response at the next sampling moment is predicted using the current state and historical voltage command, so as to obtain a calculable characterization of the current sampling error between two adjacent phases. Step 7, Constrained optimization to solve for voltage reference values: With the goal of minimizing the current sampling error in the next step, the voltage reference values ​​of the four components of the direct axis d, quadrature axis q and harmonic subspace are obtained by combining the DC bus circular limit, the upper limit of the harmonic channel voltage ratio, and the voltage step limit constraint. Step 8, Harmonic Compensation and Weight Scheduling: Cross-coupling compensation and second and sixth harmonic feedforward are superimposed in the harmonic subspace; the weights of the two harmonic axes are adaptively adjusted based on the attachment coefficient estimation, and an upper limit is set on the rate of weight change. Step 9, Regeneration and Low-Speed ​​Micro-Crawling Control: Based on the bus voltage reference and allowable charge and discharge capacity provided by the battery management system, a three-segment DC side regeneration current trajectory is generated. The traction torque slope is planned according to the trajectory curvature and longitudinal slope angle. First-order filtering is performed on the speed difference and anti-saturation is set. Step 10, Modulation and Distribution: The optimized voltage reference value is restored from the coordinate system to the six-phase stator voltage, the zero-sequence component is calculated and the dead zone and device voltage drop are compensated, the duty cycle of the six bridge arms is generated by space vector pulse width modulation, and the voltage is uniformly distributed to the gate driver at the next cycle boundary. Step 11, Fault Detection and Fault Tolerance: Real-time monitoring of open circuit, overvoltage and overcurrent faults in bridge arms or phases. Once triggered, the current reference of each subspace is mapped to the remaining effective phase according to the preset reconstruction matrix, and the fault-tolerant modulation and current and voltage limiting strategy is switched to maintain the correctness of the coordinate index relationship. Step 12, Feedback Closed Loop and Parameter Update: At the midpoint sampling point of the next cycle, obtain the current sample after the modulation effect, close the correspondence between the delay input and feedback of the two adjacent cycles, and automatically update the midpoint offset, equivalent time delay and related parameter table when the switching frequency, dead time, calculation time or stator resistance temperature changes. Feed back the available torque and energy constraints calculated in the current cycle to the planning module to make collaborative decisions on subsequent formation and path planning.

8. The drive control method for intelligent transfer equipment according to claim 7, characterized in that, In step 1, based on the maximum allowable discharge current, maximum allowable charging current and bus voltage measurement issued by the battery management system, the upper limit of DC side discharge current and upper limit of charging current are calculated and written into the constraint table. The constraints are invoked in steps 7 and 9. In step 2, press Calculate the midpoint offset and apply saturation constraints, then write the results to the timing parameter table. or When changes occur, the calculation is automatically recalculated and written back in step 12; In step 3, the total time delay of the inner current loop is equivalent to a discrete two-step delay operator, and a first-order approximation is used in the continuous domain. When the switching frequency changes, it is updated synchronously in step 12. , and approximation coefficients; Before proceeding to the solution in step 7, select the weight matrix and constraint parameters from the calibration table according to the vehicle speed, rotational speed, and DC bus voltage range. , , , , The maximum allowable discharge current, the maximum allowable charging current, and the DC bus circle limit from step 1 are combined to form an optimized constraint set. In step 8, define and with , The rate of change limit is updated at a limited speed. This is the harmonic weighting matrix. , These are scalar weighting functions for the x and y harmonic axes, respectively; In step 9, according to , , Generate a three-stage DC-side regenerative current reference value , This is the reference value for bus voltage. , These are the maximum allowable charging current and maximum discharge current, respectively. The traction torque slope is set according to the trajectory curvature and longitudinal slope angle, and the speed differential filter coefficient is selected by the controller parameters according to the working conditions. In step 11, current discontinuity, gate feedback consistency and DC bus and phase voltage abnormality criteria are used for detection. After triggering, a preset reconstruction matrix is ​​selected according to the fault type and switched to fault-tolerant SVPWM and current limiting and voltage limiting tables. In step 12, the available torque, available power, and their respective effective ranges are summarized each cycle, written to the shared buffer, and reported to the planning module. Simultaneously, the factors... , , Or caused by temperature changes , , The approximation coefficients are updated synchronously.

Images