A method based on gantry double-drive deflection force compensation

Abstract

The present application relates to gantry double drive motion control technical field, specifically said is a kind of method based on gantry double drive deflection force compensation, including calibration stage and running compensation stage: calibration stage will the effective motion stroke of gantry be discretized into multiple calibration position points, adjust the relative position of double-side drive at each point, with the minimum of double-side motor current, torque, load rate or its combination as criterion, determine the optimal relative bias quantity that makes the minimum stress of beam structure, establish the minimum deflection force position mapping table;Running compensation stage real-time obtains target position, running direction and double-side actual position, obtains ideal relative bias quantity by table lookup and interpolation, calculates the deviation of actual relative bias quantity and ideal value, converts the deviation into current, torque, speed or position compensation quantity, and applies to double-drive control system in feedforward mode.The present application takes the minimization of structural stress as control target, without additional sensor, can effectively reduce beam deflection force and internal stress.

Description

Technical Field

[0001] This invention relates to the field of gantry dual-drive motion control technology, specifically a method based on gantry dual-drive yaw force compensation. Background Technology

[0002] Gantry cranes, with their high load-bearing capacity, large span, and excellent dynamic response characteristics, are widely used in modern industrial manufacturing, logistics handling, and precision testing. To drive the smooth movement of the large-span beam, gantry cranes generally employ a dual-side independent drive structure. This means that two motors, a transmission mechanism, and a guiding mechanism, respectively located at both ends of the beam, jointly drive the beam along the guide rails. However, due to unavoidable errors in machining and assembly, such as misalignment of the dual-side motors, straightness and parallelism errors in the guide rails, pitch errors in the ball screws or racks, and accumulated clearance in the transmission chain, coupled with load bias, temperature rise deformation, and component wear after long-term use, the beam may not necessarily be in the position of minimum mechanical stress on both sides of the dual drive, even when nominally in the same position.

[0003] Existing gantry dual-drive control technologies mainly focus on improving synchronization accuracy. Common solutions include master-slave synchronization control, dual-position-loop synchronization control, cross-coupling synchronization control, torque difference closed-loop control, and decoupling control based on yaw angle estimation. Master-slave synchronization control uses one motor as the master shaft, with the other motor following the master shaft's position command. It has a simple structure but poor disturbance rejection capability, as disturbances in the master shaft are directly transmitted to the slave shaft. Dual-position-loop synchronization control establishes independent position closed loops for each motor, achieving synchronization through a unified position command, but it cannot effectively suppress internal stress caused by mechanical structural asymmetry. Cross-coupling synchronization control introduces cross-feedback of position errors on both sides, correcting speed commands in real time, which can accelerate the convergence of synchronization errors. However, it is essentially still feedback control, only correcting after a deviation occurs, and cannot pre-compensate for inherent nonlinear errors in the system. Torque difference closed-loop control uses the torque difference between the two motors as feedback, adjusting the current command to balance the output on both sides, which can suppress internal stress to some extent. However, it is usually only debugged at a single point and does not consider the nonlinear changes in error over the entire stroke range. Decoupled control based on deflection angle estimation decomposes the beam motion into translation and deflection by establishing a kinematic model of the system and controlling them separately. However, it requires accurate system model parameters, and some schemes require additional angle or attitude sensors, which increases the hardware complexity and implementation cost of the system.

[0004] The aforementioned existing technologies generally suffer from a core flaw: they all assume that equal geometric positions on both sides constitute the optimal operating state, taking the minimization of bilateral positional errors as the sole control objective. However, in actual industrial scenarios, due to factors such as installation deviations, guide rail parallelism errors, cumulative transmission chain errors, and structural deformation, "geometric equality" and "minimum stress corresponding position" often do not coincide. When the control system forces both motors to operate at the same target position for an extended period, the crossbeam will continuously bear internal torque and yaw force, leading to abnormal wear of the guide rails, loosening of connecting parts, and micro-deformation of the crossbeam, severely affecting the positioning accuracy and service life of the equipment. Furthermore, most existing technologies employ feedback correction mechanisms, lacking pre-identification and pre-compensation for the inherent optimal correspondence throughout the entire stroke of the dual-drive system, making it difficult to avoid the generation of long-term accumulated stress. Solutions that partially rely on multi-sensor fusion are not only costly but also require high on-site installation and adjustment conditions and model accuracy, hindering rapid promotion and application in conventional industrial equipment. Meanwhile, existing solutions generally lack compensation mechanisms for different operating conditions, such as direction of operation, load changes, temperature rise drift, and long-term wear, making it difficult for the same set of synchronization parameters to adapt to structural characteristic changes under different operating conditions and at different stages of use. Summary of the Invention

[0005] To address the problems in existing technologies, this invention provides a method for gantry dual-drive yaw force compensation. This method does not aim for absolute equality of geometric positions on both sides, but rather focuses on minimizing the stress on the gantry structure. It establishes a minimum yaw force position mapping table through full-stroke discretization calibration, and employs table-lookup feedforward compensation during operation. This effectively reduces the yaw force and internal stress of the crossbeam, minimizes abnormal wear on guide rails and transmission components, and extends the equipment's service life. Another objective of this invention is to provide a yaw force compensation method that requires no additional complex sensors for structural stress and attitude, is simple to implement, adaptable to various operating conditions, and supports dynamic updates.

[0006] The technical solution adopted by the present invention to solve its technical problem is: a method based on gantry double drive yaw force compensation, including a calibration stage and an operation compensation stage. The method is applicable to a gantry double drive mechanism with a first drive motor and a second drive motor. The first drive motor and the second drive motor are respectively arranged on both sides of the gantry beam to drive the beam to move along the direction of motion.

[0007] The calibration phase consists of the following steps:

[0008] S11: Increases the effective travel distance of the dual-drive gantry system. , Discretized into multiple calibration location points , ,and ;

[0009] S12: At each calibration location point Using one side of the drive as the reference side, a relative position fine-tuning amount is applied to the other side. Multiple sets of candidate relative position relationships are formed, and after each set of candidate relationships is stable, the criteria quantity reflecting the output or mutual torsion of the two motors is collected.

[0010] S13: Based on the predetermined minimum output criterion, select the optimal relative offset that minimizes the stress on the beam structure from multiple candidate relative positional relationships. ;

[0011] S14: Put all calibration points and its corresponding optimal relative bias Associative storage is used to establish a minimum yaw force position mapping table;

[0012] The operational compensation phase consists of the following steps:

[0013] S21: Real-time acquisition of current target position s, running direction d, and actual position on the first side. Second side actual position ;

[0014] S22: Based on the target position s, look up the minimum yaw force position mapping table and calculate the ideal relative offset corresponding to the current position through interpolation. ;

[0015] S23: Calculate the actual relative offset Thus, the deviation is obtained. ;

[0016] S24: Adjust the deviation It is converted into at least one of current compensation, torque compensation, speed compensation, and position fine-tuning, and applied to the first drive motor and / or the second drive motor in a feedforward manner so that the actual relative bias is close to the ideal relative bias.

[0017] Specifically, in step S11, the discretization method consists of fixed step size discretization or adaptive step size discretization; the spacing between adjacent calibration position points in the fixed step size discretization is a preset fixed value L; the adaptive step size discretization uses a smaller step size in sections where the guide rail error changes rapidly, the assembly error is large, the transmission chain error changes abruptly, or the historical compensation amount changes greatly, and uses a larger step size in sections where the error changes gradually.

[0018] Specifically, in step S13, the minimum output criterion comprises at least one of the following: minimum total current criterion, minimum current difference criterion, minimum total torque criterion, minimum torque difference criterion, minimum load rate difference criterion, and weighted comprehensive criterion; the evaluation function of the weighted comprehensive criterion is: ,in , These are the operating currents of the first and second motors, respectively. This is the equivalent quantity of the beam deflection angle. , , , For weighting coefficients, the optimal relative bias is... .

[0019] Specifically, in step S14, the storage format of the minimum yaw force position mapping table consists of at least one of the following: an absolute position correspondence table, a relative offset table, a unidirectional mapping table indexed by one side position, and a multidimensional mapping table containing position and operating condition parameters; when the running direction affects the optimal relative position relationship, a forward running mapping table T is established respectively. + Mapping table T with reverse operation − When load, temperature, or speed affect the optimal relative position relationship, a multi-condition mapping table is established.

[0020] Specifically, in step S22, the interpolation method comprises at least one of linear interpolation, piecewise linear interpolation, smooth interpolation based on the slope constraint of adjacent intervals, a hybrid method of table lookup and fitting, and a nearest neighbor table lookup method; when the target position Exactly equal to the calibration location point At that time, take directly When the target location Located at adjacent calibration points and When the interval is between, linear interpolation is used for calculation: .

[0021] Specifically, in step S24, the conversion method of the compensation amount consists of at least one of the following:

[0022] Current compensation amount ;

[0023] Torque compensation ;

[0024] Speed ​​compensation amount ;

[0025] Position fine adjustment amount ;

[0026] in , , , The preset proportional conversion coefficient is used; the compensation amount is distributed in a manner that consists of at least one of the following: symmetrical distribution, asymmetrical distribution that acts only on the driven shaft, and unequal distribution based on the load ratio / stiffness ratio / driving capacity ratio on both sides.

[0027] Specifically, it also includes the step of dynamically updating the minimum yaw force position mapping table. The dynamic update strategy consists of at least one of the following: periodically recalibrating the entire stroke or a specified segment, locally recalibrating key position points during power-on self-test, fine-tuning table entries online under stable operating conditions, performing recursive averaging correction on long-term residual deviations, and updating the corresponding entries in the multi-condition mapping table respectively; the recursive averaging correction formula is: ;in For updating coefficients, 0 < ≤1, This is the optimal relative bias before the update. This is the optimal relative bias obtained from observation.

[0028] Specifically, it also includes safety constraint and protection logic, which consists of at least one of the following: setting deviation amount. The system has a safety threshold that triggers speed reduction, shutdown, or alarm when the threshold is exceeded; it sets the upper limits for compensation current, compensation speed, and compensation displacement; it employs a compensation suppression strategy during emergency stop, zero return, reversal, over-limit, or servo re-enabling phases; and it automatically switches to safe synchronization mode when the calibration table is missing, the encoder is abnormal, the current / torque sampling channel is abnormal, or the data exceeds the limit.

[0029] Specifically, in step S12, the relative position fine-tuning amount The range of values ​​is , The scanning method is determined based on the upper limit of assembly error, guide rail span, beam stiffness, transmission clearance, and safety limit; the scanning method consists of at least one of the following: unidirectional incremental scanning, bidirectional scanning centered on zero deviation, coarse scanning plus fine scanning, and adaptive search combining the golden section or bisection approximation.

[0030] Specifically, the feedforward compensation based on the calibration trajectory can also work in conjunction with conventional position loops, velocity loops, or current loops. The feedforward compensation is located in the outer coordination layer, and its control objective is to minimize the stress on the beam structure, rather than just minimizing the bilateral synchronization error.

[0031] The beneficial effects of this invention are:

[0032] Innovation in control objectives: This invention no longer simply controls the positions on both sides to be completely equal, but takes minimizing the stress of the gantry structure as the core control objective. By obtaining the minimum stress correspondence at each position point through calibration, it can more directly and effectively reduce the torque and sway force inside the crossbeam.

[0033] Nonlinear compensation throughout the entire stroke: This invention establishes a minimum yaw force position mapping table covering the entire motion stroke by discretizing and calibrating the entire stroke. It can compensate for nonlinear positional relationships caused by non-parallel installation, misalignment of assembly, guide rail error, transmission error and directional hysteresis. The compensation effect is far better than a single fixed compensation value.

[0034] Feedforward compensation mechanism: This invention belongs to the feedforward compensation method based on calibration trajectory, which can pre-correct the dual drive command before the deviation is significantly amplified, reduce the lag of feedback control, and avoid the generation of long-term accumulated stress.

[0035] Simple to implement in engineering: This invention directly uses readily available signals such as motor current, torque, and load rate as the minimum output criterion, without the need for additional complex structural stress sensors or attitude sensors. It has low hardware costs and is easy to modify and upgrade existing industrial equipment.

[0036] High adaptability: This invention can be expanded into forward and reverse table partitioning, multi-condition table partitioning, and dynamic update table partitioning, which can adapt to different operating directions, different loads, different temperatures, and characteristic changes after long-term wear, while taking into account both high accuracy and long-term stability.

[0037] Improving equipment reliability: This invention can effectively reduce abnormal wear of guide rails and transmission components, reduce uneven stress on the crossbeam, and improve the positioning stability, structural reliability and service life of the equipment. Attached Figure Description

[0038] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0039] Figure 1 This is an overall flowchart of the gantry dual-drive yaw force compensation method of the present invention;

[0040] Figure 2 This is a detailed flowchart of the calibration stage of the present invention;

[0041] Figure 3 This is a detailed flowchart of the compensation phase of the present invention. Detailed Implementation

[0042] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0043] like Figures 1-3As shown, the method for gantry dual-drive yaw force compensation described in this invention includes a calibration stage and an operational compensation stage. It is applicable to gantry dual-drive mechanisms with a first drive motor and a second drive motor, which are respectively arranged on both sides of the gantry beam to drive the beam to move along the direction of motion. The system includes at least a dual-side position acquisition unit, a force criterion acquisition unit, a calibration table generation unit, a lookup table compensation unit, a compensation command conversion unit, and a dual-drive control unit.

[0044] (a) Calibration stage

[0045] The core task of the calibration phase is to establish a minimum yaw force position mapping table for the gantry dual-drive system throughout its full stroke, which specifically includes the following steps:

[0046] 1. Stroke discretization:

[0047] The effective travel distance of the gantry dual-drive system [ , Discretized into multiple calibration location points ,in ,and Effective travel range refers to the full range of travel that a gantry dual-drive system is allowed to operate during normal processing, handling, inspection, or positioning.

[0048] Discretization can be performed using either a fixed step size or an adaptive step size. In fixed step size discretization, the distance between adjacent calibration points is a preset fixed value L, which can be set to 5cm, 10cm, 20cm, or even smaller values ​​such as 5mm or 10mm, depending on the equipment's accuracy requirements. In adaptive step size discretization, based on the guide rail error distribution, assembly error variations, sudden changes in transmission chain errors, and historical compensation trend, a smaller step size is used in sections with rapid error changes, while a larger step size is used in sections with gradual error changes, thereby improving calibration efficiency while ensuring calibration accuracy.

[0049] 2. Calibration point benchmark in place:

[0050] At each calibration location point First, bring both drives into the reference position. The reference positioning method can be any of the following:

[0051] (1) Using the first motor as the reference axis, make the first motor reach the nominal position first. ;

[0052] (2) Using the second motor as the reference axis, make the second motor reach the nominal position first. ;

[0053] (3) Make both motors reach the same nominal position simultaneously. ;

[0054] (4) First, use conventional synchronous control to bring the crossbeam into the current calibration window, and then perform the subsequent fine-tuning calibration steps.

[0055] 3. Relative position scanning:

[0056] At the current calibration location Using one of the first or second side drives as a reference side, a relative position fine-tuning amount is applied to the other side. This creates a set of candidate relative positional relationships on both sides. The candidate relative positional relationships can be represented as...

[0057] ,or .

[0058] The range of values ​​for the relative position fine-tuning amount δ is as follows: ,in The scanning step length is determined comprehensively based on the equipment's assembly error limit, guide rail span, beam stiffness, transmission clearance, and safety limits. The relative position offset of the candidates is taken in sequence. .

[0059] The scanning method can employ unidirectional incremental scanning, bidirectional scanning centered on zero deviation, coarse scanning plus fine scanning, or an adaptive search method combining optimization strategies such as the golden section and binary approximation. Under each candidate relative positional relationship, after the dual-side drive enters a stable operating state, criteria quantities such as current, torque, load rate, or attitude deflection angle are collected, and the candidate relative positional relationship is associated with and stored with the corresponding criteria quantity for subsequent determination of the optimal relative offset.

[0060] 4. Minimum output criterion:

[0061] For each candidate relative position, the output information of both motors is collected, and the candidate relationship is determined to be closer to the minimum yaw force state based on a pre-determined minimum output criterion. The minimum output criterion is determined before calibration and remains consistent throughout the same full-stroke calibration process; different criterion types can be selected for different equipment, different working conditions, or different accuracy requirements.

[0062] Minimum output criteria include at least one of the following:

[0063] (1) Criterion for minimum total current: operating current of the first motor With the operating current of the second motor The sum of their absolute values ​​is the smallest, that is... Minimum;

[0064] (2) Criterion for minimum current difference: The absolute value of the difference between the operating current of the first motor and the operating current of the second motor is the minimum, i.e. Minimum;

[0065] (3) Criterion for minimum total torque: Output torque of the first motor With the output torque of the second motor The sum of their absolute values ​​is the smallest, that is... Minimum;

[0066] (4) Minimum Torque Difference Criterion: The absolute value of the difference between the output torque of the first motor and the output torque of the second motor is minimized, i.e. Minimum;

[0067] (5) Minimum load rate difference criterion: The absolute value of the difference between the load rate of the first driver and the load rate of the second driver is the smallest;

[0068] (6) Weighted comprehensive criterion: The weighted evaluation function composed of current, torque, load rate, beam attitude deflection angle or a combination thereof is minimized.

[0069] In the preferred embodiment, a weighted comprehensive criterion is adopted, and the evaluation function is defined as follows:

[0070] ;

[0071] in This represents the equivalent quantity of the beam yaw angle or attitude error. , , , This is a weighting coefficient, which can be adjusted according to the specific characteristics of the equipment and control requirements. For the first... Given a set of calibration points, take the evaluation function. The smallest relative position adjustment is taken as the optimal relative compensation amount at that position. ,Right now: When using other criterion types, the relative position adjustment amount that minimizes the corresponding criterion quantity is also taken as... .

[0072] 5. Establish a minimum yaw force mapping table:

[0073] For each calibration location point Record the corresponding optimal dual-drive position relationship and construct a minimum yaw force position mapping table. The mapping table should include at least the calibration position points. and optimal relative bias Two types of data; preferably, it may also include the ideal position on the first side. Ideal position on the second side Information such as operating direction, load conditions, temperature conditions, calibration time, criterion type, and minimum criterion value.

[0074] The mapping table can be stored in any of the following forms:

[0075] (1) Absolute position correspondence table: stores the one-to-one correspondence between the ideal position on the first side and the ideal position on the second side, i.e. ;

[0076] (2) Relative bias table: stores the correspondence between the calibration location point and the optimal relative bias, i.e. ,in ;

[0077] (3) A one-way mapping table with one side's position as the index and the other side's ideal position as the value;

[0078] (4) A multidimensional mapping table that simultaneously includes location, orientation, load conditions, and temperature conditions.

[0079] When the direction of movement affects the optimal relative position relationship, a forward movement mapping table T is established respectively. + Mapping table T with reverse operation − When load, temperature, or speed range affects the optimal relative position relationship, a multi-condition mapping table is established, and the corresponding table entry is selected according to the real-time conditions.

[0080] (II) Operational Compensation Phase

[0081] The core task of the operational compensation phase is to implement lookup-based feedforward compensation for the dual-side drive based on the minimum yaw force position mapping table obtained from calibration. This includes the following steps:

[0082] 1. Obtain the target position and motion parameters:

[0083] The system acquires the current target position s, current running direction d, and actual position on the first side in real time. and the actual position of the second side The running direction d includes at least two states: forward and reverse. In the implementation using a direction-based table, the system selects the corresponding mapping table T based on the running direction d. + or T − In implementations using multi-condition mapping tables, the system also needs to obtain current load, temperature, speed, and other condition parameters in order to select the corresponding table entry.

[0084] 2. Obtain the ideal dual-drive position relationship by referring to the table:

[0085] According to the target location Find the current ideal relative bias in the selected mapping table. Or ideal corresponding position pair ( , When the target location Exactly equal to a certain calibration point At that time, take directly When the target position s is located at an adjacent calibration point and When the interval is between, the interpolation method is used to calculate the result. .

[0086] The interpolation method can be selected from linear interpolation, piecewise linear interpolation, smooth interpolation based on the slope constraint of the nearest interval, a hybrid method of lookup table plus fitting, or nearest neighbor lookup table method. The formula for linear interpolation is as follows:

[0087] ;

[0088] By using the above lookup table and interpolation rules, we can ensure that the ideal dual-drive position relationship corresponding to the calibration result can be obtained at any target position during the operation phase.

[0089] 3. Calculate the deviation:

[0090] During the operational phase, the system no longer simply determines whether the actual positions on both sides are equal, but rather assesses the deviation between the actual relative relationship on both sides and the calibrated optimal relative relationship. The actual relative offset is defined as:

[0091] ;

[0092] The ideal relative bias obtained from the mapping table is = Then the deviation is:

[0093] ;

[0094] when When, it indicates that the current dual-drive relative relationship is consistent with the calibrated minimum yaw force relationship; when When >0, it indicates that the second side is biased in the positive direction relative to the calibration optimal relationship; when When <0, it indicates that the second side is biased in the negative direction relative to the calibrated optimal relationship. The control system is based on... The size and direction generate corresponding compensation amounts, so that the dual-side drive returns to the ideal relative position relationship defined by the mapping table.

[0095] 4. Equivalent characterization of pendulum force:

[0096] In this invention, "yawing force" refers to an equivalent mechanical quantity caused by mismatch in the two-sided drive, resulting in a torsional tendency or additional structural stress in the crossbeam. Since it is difficult to directly measure the internal stress of the crossbeam in real time during engineering applications, this invention uses position difference, current difference, torque difference, or a combination thereof to equivalently characterize the yawing force. Preferably, the equivalent quantity of the yawing force... Defined as: or ,in , Conversion factors are set based on calibration or experience.

[0097] 5. Convert and apply compensation:

[0098] According to the deviation or yaw force equivalent quantity Generate compensation quantities, which can be in the form of current compensation quantities. Torque compensation amount Speed ​​compensation amount Position fine-tuning amount , or a combination of two or more of them.

[0099] The conversion method for compensation amount is as follows:

[0100] (1) Current compensation method: ,or ,in This is the proportional conversion factor from position deviation to current compensation. For linear or nonlinear conversion functions;

[0101] (2) Torque compensation method: ,in This is the proportional conversion factor from position deviation to torque compensation.

[0102] (3) Speed ​​compensation method: ,in This is the proportional conversion factor from position deviation to velocity compensation.

[0103] (4) Position fine-tuning method: ,in This is the proportional conversion factor from position deviation to position fine-tuning.

[0104] The compensation amount can be applied simultaneously to both drives in a symmetrical distribution manner, for example:

[0105] or or .

[0106] The compensation amount can also be applied asymmetrically only to the driven shaft, or distributed unequally according to the load ratio, stiffness ratio, and driving capacity ratio on both sides. The positive and negative directions of the compensation amount are determined by... The sign of the value is determined, and the compensation objective is to reduce the value of the value. This makes the actual relative bias amount Gradually approaching the ideal relative bias .

[0107] The compensation method of this invention belongs to feedforward compensation based on a calibrated trajectory. Its core lies in obtaining the ideal dual-drive correspondence in advance based on the current position before or during motion, and then pre-correcting the commands on both sides according to this correspondence, rather than simply relying on position error closed loops, cross-coupling error closed loops, or traditional PID synchronization strategies to correct deviations after they occur. This invention can work in conjunction with conventional position loops, velocity loops, or current loops. The lookup-table compensation is located in the outer coordination layer, and its control objective is to minimize structural stress, not just minimize synchronization error.

[0108] (III) Dynamically updating calibration

[0109] After long-term operation of the equipment, the optimal dual-drive correspondence may change due to temperature rise, wear, load changes, maintenance disassembly and assembly, or changes in transmission clearance. Therefore, this invention also includes a step of dynamically updating the minimum yaw force position mapping table. The dynamic update strategy includes, but is not limited to:

[0110] 1. Periodic recalibration: Recalibrate the entire stroke or a specified section according to the preset running time, number of processing operations, or maintenance cycle;

[0111] 2. Local recalibration during power-on self-test: During the power-on, zeroing, or low-speed testing phases of the equipment, local calibration is performed on key locations.

[0112] 3. Online fine-tuning of meter entries: Under conditions of low speed, low acceleration, stable load and no triggering of safety thresholds, the meter entries are fine-tuned based on real-time collected current, torque, load rate or position deviation;

[0113] 4. Recursive Average Correction: The residual deviation obtained from long-term operational statistics is recursively averaged to correct the optimal relative bias at the corresponding location point. The correction formula is as follows: ,in For updating coefficients, 0 < ≤1, This is the optimal relative bias before the update. This is the optimal relative bias obtained by back-calculation based on recent operational data;

[0114] 5. Multi-condition table update: Create multiple calibration tables according to temperature range, load range, speed range or direction of operation, and update the corresponding table entries respectively.

[0115] To ensure the stability of dynamic updates, online fine-tuning or recursive averaging correction is only performed when the equipment is in a stable operating state. When an emergency stop, collision, limit violation, servo alarm, encoder malfunction, current / torque sampling channel malfunction, load change, or compensation exceeding the threshold is detected, dynamic updates are paused, and the original mapping table is maintained or switched to safe synchronization mode. The updated mapping table can overwrite the original entries or be stored separately as a new version mapping table. The system can select valid entries based on version number, update time, operating parameters, or verification results.

[0116] (iv) Security constraints and protection logic

[0117] To avoid overcompensation or mechanical impact under abnormal conditions, this invention also incorporates comprehensive safety constraints and protection logic:

[0118] 1. Set the deviation amount The safety threshold, when When the threshold is exceeded, a speed reduction, shutdown, or alarm will be triggered.

[0119] 2. Set upper limits for compensation current, compensation speed, and compensation displacement to prevent excessive compensation from damaging the equipment;

[0120] 3. Use dedicated entries or compensation suppression strategies during emergency stop, return to zero, reversal, over-limit, or servo re-enabling phases to avoid erroneous compensation under unstable operating conditions;

[0121] 4. Automatically switch to safe synchronization mode when calibration table is missing, sensor is abnormal, or data exceeds limits to ensure basic operational safety of the equipment.

[0122] Example 1: A basic example based on fixed step size calibration and current compensation.

[0123] This embodiment provides a basic implementation method for gantry dual-drive yaw force compensation, suitable for gantry handling platforms with general accuracy requirements and relatively stable operating conditions. The effective travel of this gantry dual-drive platform is 0-500cm. The first drive motor is located on the left side of the gantry, and the second drive motor is located on the right side of the gantry. It adopts a ball screw transmission method and is mainly used for workpiece handling operations in the workshop, with small load variations and medium operating speed.

[0124] I. Calibration Phase

[0125] 1. Travel Discretization: Using a fixed step size discretization method, the effective travel distance [0 cm, 500 cm] is discretized into calibration points every 10 cm, resulting in a total of 51 calibration position points. , ,in =0 cm, =500 cm. This step size selection takes into account both calibration efficiency and compensation accuracy, and can meet the normal usage requirements of the handling platform.

[0126] 2. Calibration point benchmark placement: At each calibration point... The positioning method uses the left-side first motor as the reference axis. The specific operation is as follows: First, using the normal position control mode, the left-side first motor is moved at low speed to the nominal position. Once the position stabilizes, a position locking command is triggered to maintain the position of the left motor. At this time, the second motor on the right is in a free state, ready for fine-tuning. This reference positioning method is simple to operate, ensures the accuracy of the reference side position, and provides a stable reference for subsequent relative position scanning.

[0127] 3. Relative position scanning:

[0128] At the current calibration location Using the left motor as a reference, apply a relative position fine-tuning amount to the right motor. Based on the equipment's assembly error limit (±0.8mm) and guide rail span (300cm), set... =1 mm, scanning step size =0.1 mm, therefore the range of δ is [-1 mm, +1 mm], resulting in 21 candidate relative position relationships.

[0129] Employing a bidirectional scanning method centered on zero bias, i.e., from Starting from 0, gradually increase in the positive direction with a step size of 0.1 mm until δ = +1 mm, then return to the starting position. =0, then gradually decrease in the negative direction with a step size of 0.1mm until... =-1 mm. Under each candidate relative position relationship, wait for both sides of the drive to enter a stable state (stabilization time is set to 0.5s to ensure the current and torque signals are stable), and then collect the operating current of the left motor through the communication interface of the servo driver. and the operating current of the motor on the right ,Will Value and corresponding The value association is stored in the controller's temporary buffer.

[0130] 4. Minimum Output Criterion: This embodiment uses the minimum total current criterion, that is, for each calibration position point... Iterate through all candidate relative positional relationships. Find the value corresponding to the minimum value. The value is used as the optimal relative bias at that point. For example, at the calibration location point At a distance of 100 cm, when When =0.3 mm, The minimum value of 2.8A is obtained, while when When =0, The value is 3.5A, therefore the optimal relative offset at this point is... =0.3 mm, corresponding to the ideal dual-drive position relationship is =100 cm, =100.03 cm. This indicates that at this position, when the right motor is 0.3 mm ahead of the left motor, the structural stress of the crossbeam is minimal, and the total output of both motors is lowest.

[0131] 5. Establish a minimum yaw force mapping table: Store the calibration results in the form of a relative offset table. All calibration locations and their corresponding optimal relative offsets are written into the controller's non-volatile memory (such as Flash) to ensure that calibration data is not lost after power failure. The format of the mapping table is shown in the table below (only some entries are listed):

[0132]

[0133] II. Operational Compensation Phase

[0134] 1. Acquire target position and motion parameters: The system receives the target position data from the host computer in real time. Simultaneously, the actual position of the left side is obtained through dual-sided encoders at a sampling frequency of 1kHz. and the actual position on the right The running direction d is determined based on the difference between the target position and the current position. This embodiment does not use a separate direction table; therefore, the running direction does not affect the selection of the mapping table.

[0135] 2. Obtain the ideal relative offset by looking up the table: The target position 's' is looked up in the mapping table T. This is done when 's' is exactly equal to a certain calibration point. At that time, take directly When s is located at an adjacent calibration point and When the interval is between, linear interpolation is used for calculation. For example, when the target position s = 105 cm, it is located at... =100 cm and =between 110 cm, assuming =0.3 mm, =0.35 mm, then: =0.3+110-1000.35−0.3×(105-100)=0.325 mm. Linear interpolation is simple to calculate, has good real-time performance, and can meet the control requirements of the handling platform.

[0136] 3. Calculate the deviation: Calculate the actual relative offset. Then calculate the deviation. For example, when the target position s = 105 cm, =0.325 mm, if detected in real time =105 cm, =105.05 cm, then =0.5 mm, =0.5−0.325=0.175 mm. This indicates that the right motor is 0.175 mm off from its calibrated optimal position, and the crossbeam is bearing additional sway force, which needs to be compensated.

[0137] 4. Convert and apply compensation:

[0138] This embodiment uses current compensation, and the compensation amount conversion formula is as follows: ,in The preset proportional conversion factor is calibrated to 2A / mm based on the current-position characteristics of the equipment. This factor is obtained through offline calibration, i.e., by applying a known positional deviation and measuring the corresponding current change, the proportional relationship is determined.

[0139] The compensation is applied simultaneously to both motors in a symmetrical distribution manner, meaning the current command for the left motor is... The current command for the motor on the right is For the above example, =0.175 mm, therefore =2 × 0.175 = 0.35 A, the current command for the left motor decreases by 0.175 A, and the current command for the right motor increases by 0.175 A. This symmetrical distribution method ensures that the speed changes of the two motors are equal in magnitude and opposite in direction, thereby quickly reducing the relative position deviation and avoiding additional translational acceleration of the crossbeam.

[0140] The compensation command is sent to the current loop input of the servo driver at a frequency of 1kHz, and together with the original current command, they control the motor operation. Through continuous compensation adjustment, the actual relative offset gradually approaches the ideal value of 0.325mm, thereby keeping the crossbeam in a state of minimum stress.

[0141] III. Comparison with Example 1: Traditional PID Synchronous Control

[0142] To compare the effects of this invention, a traditional PID synchronous control method was used to test the same gantry dual-drive platform. Traditional PID synchronous control relies on the position error on both sides. As a feedback quantity, the speed commands on both sides are corrected in real time by a PID controller, and the control objective is to make... Approaching zero. The PID parameters were fine-tuned multiple times to achieve optimal synchronization accuracy and response speed.

[0143] Under the same operating conditions (no load, operating speed 0.5 m / s, full-stroke round trip), the methods of this invention and the traditional PID synchronous control method were tested respectively, and the test results are as follows:

[0144] Motor current comparison: When using traditional PID synchronous control, the average total current of both motors is 4.2A; when using the method of this invention, the average total current of both motors is 3.5A, a reduction of approximately 16.7%. This indicates that the method of this invention can effectively reduce the internal stress of the crossbeam and reduce the output of the motor.

[0145] Guide rail wear comparison: After 1000 hours of continuous operation, the surface wear of the guide rail using traditional PID synchronous control was approximately 0.08 mm, and the wear was uneven, with more severe wear at both ends of the stroke. The surface wear of the guide rail using the method of this invention was approximately 0.03 mm, and the wear was uniform. This demonstrates that the method of this invention can significantly reduce abnormal wear of the guide rail and extend the service life of the equipment.

[0146] Comparison of operational stability: When using traditional PID synchronous control, the vibration amplitude of the crossbeam is approximately 0.15 mm; when using the method of this invention, the vibration amplitude of the crossbeam is approximately 0.08 mm, resulting in smoother operation. This is because the method of this invention pre-compensates for the inherent error of the system, reducing the adjustment amount of feedback control, thereby reducing vibration.

[0147] Example 2: Dynamic Update Based on Adaptive Step Size Calibration and Forward / Reverse Table Separation

[0148] This embodiment provides a gantry double-drive yaw force compensation method suitable for high-precision CNC machine tools. The effective travel of the machine tool is 0-200cm. It adopts a linear motor drive and has extremely high requirements for positioning accuracy and running stability. Furthermore, due to the difference in the forward and reverse friction characteristics of the linear motor, the optimal relative position relationship is closely related to the running direction.

[0149] I. Calibration Phase

[0150] 1. Travel Discretization: An adaptive step-size discretization method is adopted. Based on the straightness test report of the guide rail and historical operating data, the effective travel [0 cm, 200 cm] is divided into three segments: 0-50 cm (larger guide rail installation error), 50-150 cm (gradual error change), and 150-200 cm (larger error at the end of the guide rail). A small step size of 5 mm is used in the 0-50 cm and 150-200 cm segments, and a large step size of 20 mm is used in the 50-150 cm segment. A total of 253 calibration position points are ultimately set. This ensures both the calibration accuracy in areas with larger errors and improves the overall calibration efficiency.

[0151] 2. Calibration point reference positioning: A reference positioning method is adopted, which first enters the calibration window through conventional cross-coupling synchronous control, and then performs fine-tuning calibration. At each calibration location point... First, cross-coupling synchronous control is used to bring both motors to their nominal positions. Near the calibration point, the positional error is controlled within ±0.1mm. Then, the position of the left motor is locked, and the right motor is fine-tuned. This method avoids excessive initial deviation at the calibration point and reduces the time for fine-tuning the scan.

[0152] 3. Relative position scanning: A coarse scan followed by a fine scan is used to improve calibration accuracy and efficiency. First, a coarse scan is performed in the range of [-2 mm, +2 mm] with a step size of 0.2 mm to find the approximate range that minimizes the criterion (e.g., 0.4-0.6 mm). Then, a fine scan is performed in this small range with a step size of 0.02 mm to accurately determine the optimal relative offset. The assembly precision of the machine tool and the rigidity of the crossbeam are set to 2mm to ensure that all possible assembly error ranges are covered.

[0153] 4. Minimum Output Criterion: This embodiment adopts a weighted comprehensive criterion, and the evaluation function is: Since this machine tool is not equipped with an attitude sensor, the crossbeam yaw angle term is not included. The weighting coefficients are adjusted according to the dynamic characteristics of the machine tool, with the current difference term having a higher weight because the current difference can more directly reflect the degree of imbalance in the output of the two motors, thus more accurately characterizing the internal torque state of the crossbeam.

[0154] 5. Establish a mapping table for minimum yaw force in both directions: Due to the significant difference in friction characteristics between the forward and reverse directions of a linear motor, the optimal relative positional relationship differs between forward and reverse operation. Therefore, a mapping table T+ for forward operation and a mapping table T for reverse operation are established separately. − First, complete the full-stroke calibration in the forward direction (from s=0 cm to s=200 cm) to obtain the optimal relative offset at each position point during forward movement, and store it in T. + Then, repeat the above calibration steps in the reverse direction (from s=200 cm to s=0 cm) to obtain the optimal relative offset at each position point during reverse operation, and store it in T. − .

[0155] II. Operational Compensation Phase

[0156] 1. Acquiring target position and motion parameters: The system acquires the target position s, running direction d, and actual positions on both sides in real time. and And the current load condition. Select the corresponding mapping table based on the running direction d: when d is positive, call T. + When d is reversed, T is called. − .

[0157] 2. Obtain the ideal relative bias by looking up a table: Calculate the ideal relative bias using a smooth interpolation method based on the slope constraint of adjacent intervals. Compared to linear interpolation, this interpolation method avoids abrupt slope changes at calibration points, resulting in smoother changes in compensation amounts and thus improving the stability of machine tool operation. For example, when the target position is located at... and In between, not only consider and The δ value at that location is also considered. and place The value constrains the interpolation slope to ensure the continuity of the first derivative of the interpolation curve.

[0158] 3. Calculate the deviation and conversion compensation:

[0159] This embodiment uses torque compensation, and the compensation amount conversion formula is as follows: ,in The torque conversion factor is calibrated to 5 N·m / mm. Torque compensation acts directly on the torque loop of the servo driver, resulting in a fast response speed and more effectively suppressing yaw force during dynamic processes.

[0160] The compensation amount is applied asymmetrically only to the right driven shaft, meaning the torque command of the right motor is... The torque command for the left spindle remains unchanged. This method simplifies the control logic while ensuring that the spindle's positional accuracy is unaffected by compensation.

[0161] III. Dynamic Calibration Update

[0162] This embodiment adopts a dynamic update strategy that combines power-on self-test local recalibration with recursive average correction to adapt to changes in equipment characteristics after long-term operation.

[0163] 1. Self-test and partial recalibration upon startup: After each power-on and zeroing cycle, the system automatically performs partial recalibration on five key points: the two ends of the stroke (0cm, 200cm), the boundary points of error abrupt change sections (50cm, 150cm), and the middle position (100cm), updating the corresponding entries. This allows for rapid compensation for structural deformation caused by temperature changes during equipment downtime.

[0164] 2. Recursive Average Correction: During normal equipment operation, under low speed (<0.1m / s) and stable load conditions, statistical analysis is performed on the real-time collected current and position data. When the average current difference over 100 consecutive sampling periods exceeds a threshold, the optimal relative bias is considered to have changed, and the mapping table is fine-tuned using a recursive average method. Update coefficients. Set to 0.01 to ensure the stability of the update process and avoid sudden changes in the compensation amount.

[0165] For example, for the calibration location s(20) = 40 cm, the original optimal relative offset... =0.25 mm, after 100 hours of operation, the statistically optimal relative offset was obtained. =0.27 mm, then the updated result is: This slow, incremental update process can gradually compensate for changes in equipment characteristics caused by long-term wear and tear, while ensuring the stability of equipment operation.

[0166] IV. Comparison with Example 2: Single Fixed Compensation Value Method

[0167] A single fixed compensation value method was used for comparative testing. This method involves obtaining a fixed relative offset of 0.2mm at the machine tool origin and using this compensation value uniformly throughout the entire stroke, without distinguishing the direction of travel or performing dynamic updates.

[0168] The test results are as follows:

[0169] 1. Comparison of current across the entire stroke: At the midpoint of the stroke (100cm), the total current of both motors using the single fixed compensation value method is approximately 2.8A, close to the 2.7A of the method described in this invention. However, at the two ends of the stroke (0cm and 200cm), the total current of both motors using the single fixed compensation value method reaches 4.5A, 50% higher than the 3.0A of the method described in this invention. This indicates that a single fixed compensation value cannot compensate for nonlinear errors across the entire stroke range, leading to a significant increase in motor output at locations with larger errors.

[0170] 2. Comparison of forward and reverse operation: When using a single fixed compensation value method, the average current difference between forward and reverse operation is 0.8A. However, the method of this invention, by using separate meters for forward and reverse operation, controls the average current difference between forward and reverse operation to within 0.1A. This demonstrates that the method of this invention can effectively compensate for characteristic differences related to the operating direction.

[0171] 3. Long-term operation comparison: After 500 hours of operation, the wear of the guide rails of the machine tool using the single fixed compensation value method reached 0.06mm at both ends of the stroke, showing obvious uneven wear, and the positioning accuracy decreased by 0.02mm; while the guide rails of the machine tool using the method of this invention wore evenly, with a wear of only 0.02mm, and the positioning accuracy remained within 0.005mm.

[0172] Example 3: Safety Protection Example Based on Multi-condition Metering and Combined Compensation

[0173] This embodiment provides a method for compensating for sway force in heavy-duty gantry cutting equipment (the heavy-duty gantry cutting equipment is already equipped with a weighing sensor and a guide rail temperature sensor for working condition monitoring. This embodiment directly reuses the signals of the existing sensors to achieve multi-working-condition compensation without adding any additional dedicated sensors). The effective stroke of this equipment is 0-800cm, it adopts a rack and pinion transmission method, the load variation range is large (0-5000kg), and the working environment temperature varies significantly (0-60℃), which places extremely high demands on the safety and reliability of the equipment.

[0174] I. Calibration Phase

[0175] 1. Stroke discretization and multi-condition calibration: A fixed step size discretization method was adopted, with the step size set to 20cm, resulting in 41 calibration position points. Considering the significant impact of load and temperature on the optimal relative position relationship, full-stroke calibration was performed under three typical operating conditions: no load, 2500kg load, and 5000kg load, and three typical temperature conditions: 20℃, 40℃, and 60℃. Before each calibration, the equipment was preheated to the target temperature and the corresponding counterweight was applied to ensure that the calibration conditions were consistent with the actual operating conditions.

[0176] 2. Relative position scanning and minimum output criterion: A unidirectional incremental scanning method is adopted. The setting is 3mm, and the scanning step size is 0.2mm. Since the internal torque of the crossbeam under heavy load is mainly determined by the torque difference between the two motors, the minimum torque difference criterion is adopted, i.e., taking the minimum torque difference criterion. The minimum relative position adjustment is used as the optimal relative offset. The torque signal is obtained through the torque observer of the servo driver, which offers high accuracy and good real-time performance.

[0177] 3. Establish a multidimensional minimum yaw force mapping table: The storage format of the mapping table is as follows: ,in For load conditions (0, 2500, 5000 kg). The operating conditions are 20, 40, and 60℃. The calibration results for all 3×3=9 operating conditions are stored in the controller's large-capacity memory, forming a multi-dimensional mapping table. This mapping table can comprehensively reflect the optimal dual-drive position relationship under different locations, loads, and temperatures.

[0178] II. Operational Compensation Phase

[0179] 1. Acquisition of multi-condition parameters: The system acquires the target position s, running direction d, and actual positions on both sides in real time. and Meanwhile, the current load is obtained through a load cell installed on the crossbeam. The temperature of the guide rail is obtained by a temperature sensor installed on the guide rail. .

[0180] 2. Multidimensional table lookup and interpolation: Based on the target location s and the current load and temperature A lookup is performed in the multidimensional mapping table. When the load or temperature is between two calibration conditions, bilinear interpolation is used to calculate the ideal relative offset. For example, if the current load is 3000 kg, the temperature is 30℃, and the target position is 100 cm, then first interpolate at 20℃ and 40℃ to obtain the values ​​for the 3000 kg load. Value, and then in these two Temperature interpolation is performed between the values ​​to obtain the final ideal relative bias.

[0181] 3. Combined compensation methods:

[0182] This embodiment employs a combined compensation method that integrates speed compensation and position fine-tuning to balance both speed and accuracy. For larger deviations... (∣) (|>0.5 mm), mainly using speed compensation to quickly reduce deviation; for smaller deviations (∣) (≤0.5 mm), precise compensation is mainly achieved through position fine-tuning.

[0183] The compensation amount is distributed unequally according to the stiffness ratio of the two sides. Due to the stiffness of the left guide rail... Slightly greater than the stiffness of the right guide rail Therefore, the compensation amount on the left is The compensation amount on the right is This unequal distribution method allows the beam to be subjected to more even stress, further reducing internal stress.

[0184] III. Security Constraints and Protection Logic

[0185] This embodiment incorporates robust security protection logic to ensure the safe operation of the device under abnormal conditions:

[0186] 1. Deviation threshold protection: Set the deviation amount The first-level safety threshold is 2mm, when If the speed exceeds 2mm, immediately reduce the operating speed to 50% of the rated speed and issue an audible and visual alarm; set the secondary safety threshold to 5mm. If the error exceeds 5mm, immediately stop the machine and disconnect the servo power supply to prevent equipment damage.

[0187] 2. Compensation Amount Limit Protection: The upper limit of compensation speed is set to 5mm / s and the upper limit of compensation displacement is 1mm to prevent excessive compensation from causing severe impact on the equipment.

[0188] 3. Operating Condition Switching Protection: During emergency stop, zeroing, and reversing phases, the feedforward compensation function is temporarily disabled, switching to the normal synchronization mode. Compensation is restored after the equipment enters a stable operating state. This is because the dynamic characteristics of the equipment change significantly under these operating conditions, and feedforward compensation may malfunction.

[0189] 4. Fault self-diagnosis and switching: When an encoder malfunction, current / torque sampling channel failure, or calibration table data corruption is detected, the system automatically switches to safe synchronization mode, disables automatic operation, and generates a fault log for manual inspection.

[0190] IV. Comparison with Example 3: Compensation method without multi-condition sub-meters

[0191] A comparative test was conducted using a compensation method without multiple operating conditions and separate meters. This method only performs a calibration under no-load and 20°C conditions once when the equipment leaves the factory, and then does not update the mapping table or consider the effects of load and temperature.

[0192] The test results are as follows:

[0193] 1. Comparison of load change adaptability: In a test where the load suddenly changed from no load to 5000kg, the compensation method without multi-condition sub-meters showed a large sway, with a deviation of 1.8mm, close to the first-level safety threshold, and the crossbeam vibrated significantly; while the method of this invention, by calling the mapping table of heavy-load conditions, had a deviation of only 0.4mm and operated smoothly.

[0194] 2. Temperature change adaptability comparison: When the ambient temperature rises from 20℃ to 60℃, the average total current of the dual motors without multi-condition sub-metering compensation method increases from 6.5A to 7.8A, an increase of 20%; while the method of the present invention, by calling the mapping table of 60℃ condition, the average total current only increases to 6.7A, an increase of 3.1%.

[0195] 3. Long-term operation comparison: After 2000 hours of operation, due to guide rail wear and increased transmission clearance, the average total current of the dual motors without dynamic update method increased from the initial 6.5A to 8.2A, an increase of 26.2%, and obvious operation vibration occurred; while the average total current of the dual motors using the dynamic update method of this invention only increased to 6.8A, an increase of 4.6%, and the equipment operation remained stable.

[0196] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of protection claimed by the present invention. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method based on gantry dual-drive yaw force compensation, characterized in that, The method includes a calibration phase and an operation compensation phase. It is applicable to a gantry dual-drive mechanism with a first drive motor and a second drive motor, wherein the first drive motor and the second drive motor are respectively arranged on both sides of the gantry beam to drive the beam to move along the direction of motion. The calibration phase consists of the following steps: S11: Increases the effective travel distance of the dual-drive gantry system. , Discretized into multiple calibration location points , ,and ; S12: At each calibration location point Using one side of the drive as the reference side, a relative position fine-tuning amount is applied to the other side. Multiple sets of candidate relative position relationships are formed, and after each set of candidate relationships is stable, the criteria quantity reflecting the output or mutual torsion of the two motors is collected. S13: Based on the predetermined minimum output criterion, select the optimal relative offset that minimizes the stress on the beam structure from multiple candidate relative positional relationships. ; S14: Put all calibration points and its corresponding optimal relative bias Associative storage is used to establish a minimum yaw force position mapping table; The operational compensation phase consists of the following steps: S21: Real-time acquisition of current target position s, running direction d, and actual position on the first side. Second side actual position ; S22: Based on the target position s, look up the minimum yaw force position mapping table and calculate the ideal relative offset corresponding to the current position through interpolation. ; S23: Calculate the actual relative offset Thus, the deviation is obtained. ; S24: Adjust the deviation It is converted into at least one of current compensation, torque compensation, speed compensation, and position fine-tuning, and applied to the first drive motor and / or the second drive motor in a feedforward manner so that the actual relative bias is close to the ideal relative bias.

2. The method based on gantry dual-drive yaw force compensation according to claim 1, characterized in that: In step S11, the discretization method consists of fixed step size discretization or adaptive step size discretization; the spacing between adjacent calibration position points in the fixed step size discretization is a preset fixed value L; the adaptive step size discretization uses a smaller step size in sections where the guide rail error changes rapidly, the assembly error is large, the transmission chain error changes abruptly, or the historical compensation amount changes greatly, and uses a larger step size in sections where the error changes gradually.

3. The method based on gantry dual-drive yaw force compensation according to claim 1, characterized in that: In step S13, the minimum output criterion comprises at least one of the following: minimum total current criterion, minimum current difference criterion, minimum total torque criterion, minimum torque difference criterion, minimum load rate difference criterion, and weighted comprehensive criterion; the evaluation function of the weighted comprehensive criterion is: ,in , These are the operating currents of the first and second motors, respectively. This is the equivalent quantity of the beam deflection angle. , , , For weighting coefficients, the optimal relative bias is... .

4. The method based on gantry dual-drive yaw force compensation according to claim 1, characterized in that: In step S14, the storage format of the minimum yaw force position mapping table consists of at least one of the following: an absolute position correspondence table, a relative offset table, a one-way mapping table indexed by one side position, and a multi-dimensional mapping table containing position and operating parameters; when the running direction affects the optimal relative position relationship, a forward running mapping table T is established respectively. + Mapping table T with reverse operation − When load, temperature, or speed affect the optimal relative position relationship, a multi-condition mapping table is established.

5. The method based on gantry dual-drive yaw force compensation according to claim 1, characterized in that: In step S22, the interpolation method comprises at least one of the following: linear interpolation, piecewise linear interpolation, smooth interpolation based on the slope constraint of the nearest interval, a hybrid method of table lookup and fitting, and a nearest neighbor table lookup method; when the target position Exactly equal to the calibration position point At that time, take directly When the target location Located at adjacent calibration points and When the interval is between, linear interpolation is used for calculation: 。 6. The method based on gantry dual-drive yaw force compensation according to claim 1, characterized in that: In step S24, the conversion method of the compensation amount is determined by at least one of the following: composition: Current compensation amount ; Torque compensation ; Speed ​​compensation amount ; Position fine adjustment amount ; in , , , The preset proportional conversion coefficient is used; the compensation amount is distributed in a way that consists of at least one of the following: symmetrical distribution, asymmetrical distribution that only acts on the driven shaft, and unequal distribution based on the load ratio / stiffness ratio / driving capacity ratio on both sides.

7. The method based on gantry dual-drive yaw force compensation according to claim 1, characterized in that: It also includes the step of dynamically updating the minimum yaw force position mapping table. The dynamic update strategy consists of at least one of the following: periodically recalibrating the entire stroke or a specified segment, locally recalibrating key position points during power-on self-test, fine-tuning table entries online under stable operating conditions, performing recursive averaging correction on long-term residual deviations, and updating the corresponding entries in the multi-condition mapping table respectively; the recursive averaging correction formula is: ;in For updating coefficients, 0 < ≤1, This is the optimal relative bias before the update. This is the optimal relative bias obtained from observation.

8. The method based on gantry dual-drive yaw force compensation according to claim 1, characterized in that: It also includes safety constraint and protection logic, which consists of at least one of the following: setting deviation amount The system has a safety threshold that triggers speed reduction, shutdown, or alarm when the threshold is exceeded; it sets the upper limits for compensation current, compensation speed, and compensation displacement; it employs a compensation suppression strategy during emergency stop, zero return, reversal, over-limit, or servo re-enabling phases; and it automatically switches to safe synchronization mode when the calibration table is missing, the encoder is abnormal, the current / torque sampling channel is abnormal, or the data exceeds the limit.

9. A method for gantry dual-drive yaw force compensation according to claim 1, characterized in that: In step S12, the relative position fine-tuning amount The range of values ​​is , The scanning method is determined based on the upper limit of assembly error, guide rail span, beam stiffness, transmission clearance, and safety limit; the scanning method consists of at least one of the following: unidirectional incremental scanning, bidirectional scanning centered on zero deviation, coarse scanning plus fine scanning, and adaptive search combining the golden section or bisection approximation.

10. A method for gantry dual-drive yaw force compensation according to claim 1, characterized in that: The feedforward compensation based on the calibration trajectory can also work in conjunction with conventional position loop, velocity loop or current loop. The feedforward compensation is located in the outer coordination layer, and its control objective is to minimize the stress of the beam structure, rather than just minimizing the bilateral synchronization error.

Images