Dynamic compensation and real-time synchronous control method and system for five-axis light-cured printing

Abstract

The present application relates to the field of additive manufacturing control engineering, embedded real-time system and motion control algorithm, and proposes a dynamic compensation and real-time synchronization control method and system for five-axis light curing printing, which comprises the following steps: obtaining real-time posture angle and rotational axis angular velocity of a surface to be cured, inputting a dynamic compensation model to obtain target exposure energy density and XYZ composite feed speed; mapping the target exposure energy density into light source driving parameters and mapping the XYZ composite feed speed into motion axis driving parameters; querying a backlash error mapping table to generate a feedforward compensation pulse and inject it into a motion control queue when the rotational axis reverses; and synchronizing the light source driving parameter update and the motion axis driving pulse generation through master-slave linkage of a hardware timer. The present application realizes real-time and accurate mapping of exposure dose and feed speed through a layer thickness-inclination-angle velocity dynamic compensation model, effectively solving the problem of uneven curing in inclined printing.

Description

Technical Field

[0001] This invention relates to the fields of additive manufacturing control engineering, embedded real-time systems and motion control algorithms, and in particular to a dynamic compensation and real-time synchronization control method and system for five-axis photopolymerization printing. Background Technology

[0002] Five-axis photopolymerization printing, as an important development direction in the field of additive manufacturing, has significant advantages in forming complex curved components, overhanging structures, and free-form parts. However, five-axis linkage introduces dynamic control challenges that are not present in traditional static Z-axis printing. Existing technologies still have significant limitations in areas such as curing energy matching under dynamic postures, transmission chain error compensation, and multi-axis synchronous control, which restricts the engineering application of five-axis photopolymerization printing technology in the field of precision manufacturing.

[0003] The problem of solidified energy mismatch under dynamic attitude: Traditional photopolymerization control methods calculate exposure dosage based on static Z-axis layer thickness, assuming a constant coating thickness for each layer and that the UV light incident direction is perpendicular to the coating surface. However, in five-axis simultaneous printing, the tilt of the rotation axis causes a geometric amplification of the actual effective coating thickness with the attitude angle, while changes in the UV light incident angle cause a cosine decay of the energy density. Existing open-loop control methods cannot detect and compensate for these geometric and optical changes in real time, leading to under-curing on overhanging surfaces and over-curing in planar areas due to excess energy, ultimately affecting the dimensional accuracy and mechanical property consistency of the parts. This problem is particularly prominent in the printing of precision features such as dental prosthesis shoulders, making it difficult to meet the accuracy requirements of clinical applications.

[0004] Harmonic drive commutation dead zone problem: When five-axis printing equipment tracks complex curved surfaces, the rotary axes need to frequently perform reversing operations. As the core component of the rotary axis transmission, the harmonic reducer exhibits nonlinear elastic deformation and microscopic backlash at low-speed reversing points, making it difficult for traditional PID control methods to eliminate the aforementioned reversing dead zone.

[0005] Multi-axis synchronization delay problem: In general five-axis control architectures, motion interpolation and UV light source control are typically synchronized using software polling, which inevitably introduces scheduling delays. When the feed rate changes drastically, the light intensity adjustment lags behind the motion commands, causing localized curing depth deviations and affecting interlayer bonding strength. Software-level synchronization mechanisms are affected by multiple factors, including operating system scheduling cycles, interrupt response times, and communication link delays, making it difficult to meet the stringent real-time requirements of high-speed continuous printing. Summary of the Invention

[0006] In view of the above, the main objective of this invention is to propose a dynamic compensation and real-time synchronization control method and system for five-axis photopolymerization printing to solve the above-mentioned technical problems.

[0007] This invention proposes a dynamic compensation and real-time synchronization control method for five-axis photopolymerization printing, the method comprising the following steps: Step 1: Obtain the real-time attitude angle data and the real-time angular velocity data of the rotation axis of the surface to be cured. Use the real-time attitude angle data and the real-time angular velocity data of the rotation axis as two input feature parameters and input them into the preset dynamic compensation model for calculation to obtain the target exposure energy density and XYZ synthesis feed rate, respectively. Step 2: Generate the corresponding light source pulse width modulation driving parameters based on the target exposure energy density mapping; perform five-axis inverse kinematics calculation based on the XYZ synthesis feed rate to generate the basic motion axis driving pulse sequence; Step 3: Monitor the direction of rotational axis speed in real time. When a change in the direction of rotational axis speed is detected, query the preset backlash error mapping table based on the current position data and current temperature data to generate feedforward compensation pulses. Then, superimpose the feedforward compensation pulses into the basic motion axis drive pulse sequence to generate the compensated final motion axis drive pulse sequence. Step 4: Through the hardware timer master-slave linkage mechanism, the compensated final motion axis drive pulse sequence is used as the external clock trigger source of the master timer, and the update operation of the slave timer's compare register is synchronously triggered by the update event of the master timer, so as to align the light source pulse width modulation drive parameters with the compensated final motion axis drive pulse sequence in timing and generate a synchronous execution control signal. Step 5: Based on the synchronous execution control signal, control the synchronous execution of the light source exposure action and the motion axis feed action to obtain the final control result.

[0008] This invention also proposes a dynamic compensation and real-time synchronization control system for five-axis photopolymerization printing, the system comprising: The host computer module is used for: The real-time attitude angle data and the real-time angular velocity data of the rotation axis of the surface to be cured are obtained. The real-time attitude angle data and the real-time angular velocity data of the rotation axis of the surface to be cured are used as two input feature parameters and input into the preset dynamic compensation model for calculation to obtain the target exposure energy density and XYZ synthesis feed rate, respectively. Real-time control module, used for: The corresponding light source pulse width modulation driving parameters are generated based on the target exposure energy density mapping; the five-axis inverse kinematics calculation is performed based on the XYZ composite feed rate to generate the basic motion axis driving pulse sequence; Motion drive module, used for: The rotation axis speed direction is monitored in real time. When a change in the rotation axis speed direction is detected, the preset backlash error mapping table is queried according to the current position data and the current temperature data to generate feedforward compensation pulses. The feedforward compensation pulses are then superimposed and injected into the basic motion axis drive pulse sequence to generate the compensated final motion axis drive pulse sequence. The light source driver module is used for: Through the hardware timer master-slave linkage mechanism, the compensated final motion axis drive pulse sequence is used as the external clock trigger source of the master timer, and the update operation of the slave timer's compare register is synchronously triggered by the update event of the master timer, so as to align the light source pulse width modulation drive parameters with the compensated final motion axis drive pulse sequence in timing and generate a synchronous execution control signal. Master-slave timer linkage unit, used for: Based on the synchronous execution control signal, the synchronous execution of the light source exposure action and the motion axis feed action is controlled to obtain the final control result.

[0009] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention establishes a continuous dynamic compensation model for layer thickness, tilt angle, and angular velocity, achieving real-time and accurate mapping between exposure dose and feed rate, effectively solving the problem of uneven curing in tilted printing. This model automatically adjusts the target exposure energy density and feed rate based on real-time changes in the attitude angle and rotation axis angular velocity. This results in a corresponding decrease in feed rate and a corresponding increase in dwell time when tilting or rotating more violently, thereby compensating for the geometric amplification of layer thickness and the cosine attenuation of light intensity, maintaining the consistency of the curing dose, and effectively suppressing under-curing on overhanging surfaces and over-curing on planar surfaces.

[0010] 2. The backlash feedforward compensation algorithm based on position-temperature mapping proposed in this invention triggers query and compensation only at the moment of rotational shaft speed reversal. This eliminates the influence of transmission chain backlash on continuous coating trajectory and avoids the MCU computational load caused by continuous query. The compensation pulse is distributed across multiple control cycles using an exponential decay smoothing algorithm. The largest share is injected in the first cycle to quickly compensate for dead zone positioning lag, and the amount is gradually reduced in subsequent cycles to avoid step impact, while the total compensation amount remains constant. In non-reversal cycles, only one sign comparison is performed without table lookup, resulting in extremely low computational load.

[0011] 3. The feed speed-light source intensity hardware-level proportional synchronization mechanism constructed in this invention achieves microsecond-level synchronization accuracy through master-slave timer linkage. When the speed changes, the light intensity duty cycle is updated immediately in the next clock cycle, effectively eliminating software lag and adapting to the needs of high-speed continuous printing.

[0012] 4. This invention supports multi-resin parameter library management. The system has built-in characteristic coefficients of different brands / models of resins. When the user selects a process formula, the corresponding parameters are automatically loaded. It also supports on-site fitting of new resin parameters through a quick calibration wizard. Process switching can be achieved without redeveloping firmware, which significantly improves the compatibility and engineering applicability of the equipment.

[0013] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by means of embodiments of the invention. Attached Figure Description

[0014] Figure 1 This is a flowchart of the dynamic compensation and real-time synchronization control method for five-axis photopolymerization printing proposed in this invention.

[0015] Figure 2 This is a schematic diagram of the overall control architecture of the five-axis photopolymerization printing dynamic compensation and real-time synchronization control method proposed in this invention.

[0016] Figure 3 This is an architecture diagram of the dynamic compensation and real-time synchronization control system for five-axis photopolymerization printing proposed in this invention. Detailed Implementation

[0017] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0018] These and other aspects of the embodiments of the present invention will become clear from the following description and accompanying drawings. In these descriptions and drawings, some specific embodiments of the present invention are specifically disclosed to illustrate some ways of implementing the principles of the embodiments of the present invention; however, it should be understood that the scope of the embodiments of the present invention is not limited thereto.

[0019] Please see Figure 1 This embodiment provides a dynamic compensation and real-time synchronization control method for five-axis photopolymerization printing, the method including the following steps: Step 1: Obtain the real-time attitude angle data and the real-time angular velocity data of the rotation axis of the surface to be cured. Use the real-time attitude angle data and the real-time angular velocity data of the rotation axis as two input feature parameters and input them into the preset dynamic compensation model for calculation to obtain the target exposure energy density and XYZ synthesis feed rate, respectively.

[0020] Please see Figure 2In step 1, the real-time attitude angle data and real-time angular velocity data of the rotation axis of the surface to be cured are obtained. These two data are used as two input feature parameters and input into a preset dynamic compensation model for calculation to obtain the target exposure energy density and XYZ composite feed rate, respectively. The specific steps include the following: Acquire real-time attitude angle data of the surface to be cured and real-time angular velocity data of the rotation axis; The real-time attitude angle data and the real-time angular velocity data of the rotation axis are used as two input feature parameters; Obtain the baseline set layer thickness, input the two input feature parameters into the preset dynamic compensation model for calculation, take the cosine value of the real-time attitude angle data, and divide the baseline set layer thickness by the cosine value to obtain the geometrically magnified layer thickness; calculate the product of the absolute value of the real-time angular velocity data of the rotation axis and the preset first resin characteristic coefficient to obtain the shear thinning correction value; calculate the product of the square of the real-time angular velocity data of the rotation axis and the preset second resin characteristic coefficient to obtain the fluid inertial accumulation correction value; add the numerical value 1, the shear thinning correction value and the fluid inertial accumulation correction value to obtain the angular velocity comprehensive correction factor; multiply the geometrically magnified layer thickness and the angular velocity comprehensive correction factor to obtain the effective layer thickness data; Obtain the reference exposure energy density, calculate the ratio of the effective layer thickness data to the reference set layer thickness to obtain the layer thickness magnification; take the cosine value of the real-time attitude angle data and calculate the reciprocal of the cosine value to obtain the light intensity projection attenuation compensation value; multiply the reference exposure energy density, the layer thickness magnification, and the light intensity projection attenuation compensation value to obtain the target exposure energy density. The XYZ synthesis feed rate is obtained based on the target exposure energy density and the preset control mode; wherein, the preset control mode includes: speed adjustment mode.

[0021] The process of obtaining the XYZ composite feed rate based on the target exposure energy density and the preset control mode specifically includes: When the preset control mode is speed regulation mode, the preset UV light source intensity is kept constant. The feed speed adjustment ratio is obtained by calculating the ratio of the reference exposure energy density to the target exposure energy density. The reference feed speed is obtained and multiplied by the feed speed adjustment ratio to obtain the XYZ composite feed speed. The baseline set layer thickness is obtained, and the two input feature parameters are input into a preset dynamic compensation model for calculation. The geometrically amplified layer thickness is obtained by taking the cosine value of the real-time attitude angle data and dividing the baseline set layer thickness by the cosine value. The absolute value of the real-time angular velocity data of the rotation axis is calculated and multiplied by a preset first resin characteristic coefficient to obtain a shear thinning correction value. The square of the real-time angular velocity data of the rotation axis is calculated and multiplied by a preset second resin characteristic coefficient to obtain a fluid inertial accumulation correction value. The numerical value 1, the shear thinning correction value, and the fluid inertial accumulation correction value are added together to obtain an angular velocity comprehensive correction factor. The geometrically amplified layer thickness is multiplied by the angular velocity comprehensive correction factor to obtain the effective layer thickness data. The corresponding relationship in this process is as follows: ; in, Indicates the effective layer thickness data. Indicates time, Indicates the reference layer thickness. Represents the cosine function. This represents real-time attitude angle data. Indicates the thickness of the geometrically enlarged layer. This indicates the preset first resin characteristic coefficient. This represents the absolute value of the real-time angular velocity data of the rotating axis. This represents the shear thinning correction value. This indicates the preset second resin characteristic coefficient. This represents the correction value for fluid inertial deposition. This represents the comprehensive correction factor for angular velocity.

[0022] The ratio of the effective layer thickness data to the reference set layer thickness is calculated to obtain the layer thickness magnification; the cosine value of the real-time attitude angle data is taken and the reciprocal of the cosine value is calculated to obtain the light intensity projection attenuation compensation value; the reference exposure energy density, the layer thickness magnification, and the light intensity projection attenuation compensation value are multiplied to obtain the target exposure energy density. The corresponding relationship in this process is as follows: ; in, Indicates the target exposure energy density. Indicates the baseline exposure energy density. Indicates the magnification factor of the layer thickness. This represents the light intensity projection attenuation compensation value.

[0023] Calculate the ratio of the reference exposure energy density to the target exposure energy density to obtain the feed rate adjustment ratio; obtain the reference feed rate, and multiply the reference feed rate by the feed rate adjustment ratio to obtain the XYZ composite feed rate. The corresponding relationship in this process is as follows: ; in, Indicates the XYZ composite feed rate. Indicates the reference feed rate. Indicates the feed rate adjustment ratio; The reference exposure energy density is calculated by multiplying the ratio between the reference set layer thickness and the reference feed rate by a preset constant UV light source intensity. The corresponding relationship is as follows: ; in, This indicates a preset, constant UV light source intensity.

[0024] In a specific embodiment of the present invention, assuming the current attitude angle is 30° and the reference set layer thickness is 0.05mm, the geometrically amplified layer thickness is 0.05 / cos(30°)≈0.0577mm, that is, when tilted at 30°, the effective coating thickness increases from 0.05mm to approximately 0.0577mm; assuming the current rotation axis angular velocity is 500° / s, the first resin characteristic coefficient... The second resin characteristic coefficient is 0.02. If the value is 0.001, then the shear thinning correction value is 0.02×8.73≈0.175, the fluid inertial packing correction value is 0.001×(8.73)²≈0.076, the angular velocity comprehensive correction factor is 1+0.175+0.076=1.251, and the effective layer thickness is 0.0577×1.251≈0.0722mm; assuming the reference exposure energy density is 30mJ / cm², then the layer thickness magnification is 0.0722 / 0.05≈1.444, the light intensity projection attenuation compensation value is 1 / cos(30°)≈1.155, and the target exposure energy density is 30×1.444×1.155≈50.0mJ / cm², that is, when tilted at 30° and rotated at an angular velocity of 500° / s, the target exposure energy density increases from 30mJ / cm² to about 50mJ / cm², an increase of about 67%; If a speed-adjustable mode is used and the base feed rate is 10 mm / s, then the feed rate adjustment ratio is 30 / 50 ≈ 0.6. The XYZ composite feed rate is 10 × 0.6 = 6 mm / s. This means that to maintain a constant light source intensity, the feed rate is reduced from 10 mm / s to 6 mm / s to increase the exposure time. For the two resin characteristic coefficients... and The system has a built-in library of characteristic coefficients for different brands and models of resin. When users use a new resin for the first time, they can print standard test blocks with different attitude angles and angular velocities according to the quick calibration wizard and measure the actual layer thickness. The system automatically uses the least squares method to complete the fitting and storage of coefficients.

[0025] In the dynamic compensation model of this invention, the first resin characteristic coefficient and the second resin characteristic coefficient are key parameters characterizing the rheological behavior of different photocurable resin materials under rotational motion conditions. The first characteristic coefficient mainly reflects the sensitivity of the resin's viscosity to changes in shear rate during shear flow, i.e., the shear thinning effect; the second characteristic coefficient mainly reflects the resin's tendency to accumulate fluid due to inertial forces under rotational acceleration. These two coefficients are related to the inherent properties of the resin, such as its chemical composition, filler content, and molecular weight distribution. The values ​​differ significantly for different brands or models of resin, therefore, they must be determined through a standardized calibration process.

[0026] This embodiment has a built-in resin parameter library, which supports the automatic loading of corresponding characteristic coefficients when users select different resins. It also provides a quick calibration wizard. Users can print standard test blocks and measure the actual layer thickness. The system automatically completes coefficient fitting and storage without manual programming.

[0027] (a) Geometric structure and printing parameters of standard test blocks The standard test block is designed as a disc-shaped structure with multiple inclined steps, each step corresponding to a specific curing surface attitude angle. The bottom of the test block is a flat base, and the top has six fan-shaped stepped areas evenly distributed along the circumference. The inclination angle of each area is fixed to a specific angle value, for example, starting from zero degrees and increasing to a larger angle, covering various tilting conditions that the equipment may encounter in actual printing.

[0028] On each inclined step surface, multiple sub-regions are also set along the circumference, each of which is subjected to a different rotational angular velocity during the printing process. The angular velocity setting starts from zero and increases at equal intervals to the maximum angular velocity allowed by the equipment, covering the hydrodynamic effects under different rotational speed conditions. Isolation grooves are set between all sub-regions to prevent mutual interference between adjacent regions during curing.

[0029] During calibration printing, all process parameters were fixed to baseline values: baseline layer thickness, baseline feed rate, and constant ultraviolet light source intensity were all kept uniform. The test block material used was the resin to be calibrated, and post-processing cleaning and curing were performed after printing.

[0030] (ii) Method for measuring the thickness of the cured layer After printing and cleaning the standard test blocks, and allowing them to dry completely, a high-precision contact profilometer is used to scan the surface profile along the centerline of each sub-region. The profilometer can accurately measure the actual cured layer thickness, with a resolution better than the micrometer level. Each sub-region is measured at least three times, with points taken at different locations. The arithmetic mean of the three measurements is taken as the actual layer thickness data under the given attitude angle and angular velocity conditions. Measurements should avoid edge areas to ensure the data reflects a stable cured layer thickness.

[0031] (III) The specific process of least squares fitting Using all measured actual layer thickness data as observations, the calculation formula for effective layer thickness in the dynamic compensation model of this invention is used as the theoretical prediction model. This model shows that the actual effective layer thickness depends on the product of the geometrically amplified layer thickness and the angular velocity comprehensive correction factor. The geometrically amplified layer thickness is obtained by dividing the baseline set layer thickness by the cosine of the attitude angle; the angular velocity comprehensive correction factor consists of the sum of a constant, a shear thinning correction term, and a fluid inertial accumulation correction term. The shear thinning correction term is equal to the first characteristic coefficient multiplied by the absolute value of the angular velocity, and the fluid inertial accumulation correction term is equal to the second characteristic coefficient multiplied by the square of the angular velocity.

[0032] The goal of the fitting process is to find an optimal set of first and second characteristic coefficients that minimizes the sum of squares of the differences between the theoretically predicted effective layer thickness and the actually measured layer thickness for all combinations of attitude angles and angular velocities. Furthermore, considering the physical meaning of the coefficients, both are non-negative values; therefore, a non-negative constraint is added during the fitting process.

[0033] To solve this optimization problem, a classic nonlinear least squares algorithm, such as the Levenberg-Marquardt method, is employed. This method is suitable for small- to medium-scale parameter estimation and has the advantages of fast convergence speed and low dependence on initial values. The initial value for iterative calculation can be set as an empirical small positive number. The iteration termination condition is set as the change in the objective function between two consecutive iterations being less than a preset minimum value, or the number of iterations reaching an upper limit. The algorithm automatically iterates until the termination condition is met, outputting the optimal coefficients as the characteristic parameters of the resin.

[0034] (iv) Physical dimensions and typical range of values ​​of characteristic coefficients According to the dimensional analysis in the model, the dimension of the first characteristic coefficient is time divided by angle (i.e., seconds per radian), and its value reflects the sensitivity of the resin viscosity to changes in shear rate; the dimension of the second characteristic coefficient is time squared divided by angle squared (i.e., seconds squared per radian squared), and its value reflects the degree of resin accumulation under the action of rotational inertia.

[0035] Experimental measurements show that the primary characteristic coefficient of common UV-cured resins is generally in the range of a few percent, while the secondary characteristic coefficient is around a few parts per thousand. Resins for different applications show significant differences: standard rigid resins have a moderate coefficient, while flexible resins, due to their lower viscosity and higher fluidity, exhibit a more pronounced shear thinning effect and thus a larger coefficient; dental resins typically contain a higher proportion of ceramic fillers, have higher viscosity, and are relatively less sensitive to changes in shear rate, resulting in a smaller coefficient.

[0036] (v) Verification and storage of calibration results The fitted characteristic coefficients are immediately stored in the system's resin parameter library, corresponding one-to-one with the brand and model of the resin. To verify the accuracy of the calibration results, users can print a set of verification test blocks with different attitude angles and angular velocities, measure the actual layer thickness, and compare it with the model's predicted value. Extensive calibration experiments have verified that the coefficients obtained using the calibration process of this invention can control the relative deviation between the actual layer thickness and the target layer thickness within ±3% under all test conditions, while the deviation can reach over ±25% without compensation, fully demonstrating the effectiveness of the calibration method and the significant effect of model compensation.

[0037] When users reuse the same resin, they only need to select the corresponding brand and model in the system, and the characteristic coefficients will be automatically loaded into the dynamic compensation model without the need for recalibration. If a new resin that has never been used before is used, the system will automatically complete the entire process fitting and storage by following the quick calibration wizard process described above, without requiring manual calculation by the user. This greatly simplifies the process changeover and improves the engineering applicability of the equipment.

[0038] Compared with existing technologies, the core innovation of this embodiment lies in establishing a continuously differentiable dynamic compensation model that includes a cosine correction term for the attitude angle and first and second correction terms for the angular velocity. Existing photopolymerization control methods calculate exposure dose based on static Z-axis layer thickness, assuming a constant layer thickness and that the UV incident direction is perpendicular to the coating surface. They fail to consider the geometric amplification effect of the actual effective coating thickness with attitude angle caused by the tilt of the rotation axis in five-axis linkage, nor do they consider the cosine attenuation of energy density caused by changes in the UV incident angle. More importantly, existing technologies neglect the influence of the rotation axis angular velocity on the resin's hydrodynamic characteristics—namely, the shear thinning effect and the fluid inertial accumulation effect. Although some existing technologies have proposed static parameter adjustment schemes based on model shape information or real-time curing degree compensation methods based on dielectric sensing, their compensation is based on offline presets or sensor signals, rather than physical models of attitude angle and angular velocity, thus failing to achieve real-time compensation under dynamic attitude conditions in five-axis linkage printing. This invention uses real-time attitude angle data and rotation axis angular velocity data as two input feature parameters, and introduces a first resin characteristic coefficient and a second resin characteristic coefficient to characterize the shear thinning effect and the fluid inertial stacking effect, respectively. For the first time, a comprehensive compensation model covering geometric amplification, light intensity attenuation and fluid dynamics effects is established, realizing real-time and accurate mapping of exposure energy density and feed rate under dynamic attitude.

[0039] Step 2: Generate the corresponding light source pulse width modulation driving parameters based on the target exposure energy density mapping; perform five-axis inverse kinematics calculation based on the XYZ synthesis feed rate to generate the basic motion axis driving pulse sequence.

[0040] In a specific embodiment of the present invention, the corresponding light source pulse width modulation driving parameters are generated based on the calculated target exposure energy density. That is, the required light source output power is calculated based on the target exposure energy density and the current feed speed. Then, the corresponding duty cycle value is determined based on the calibration curve of the light source output power and the pulse width modulation duty cycle and written into the comparison register of the pulse width modulation module. For example, assuming that the rated power of the ultraviolet light source is 100W and the pulse width modulation duty cycle is linearly related to the output power, if the required output power is determined to be 60W based on the target exposure energy density, then the corresponding pulse width modulation duty cycle is 60%. The five-axis mechanism configuration applicable to this embodiment is the XYZAB type five-axis configuration, which adopts a dual independent kinematic chain structure: the workpiece kinematic chain is bed → X linear axis → Y linear axis → A rotary axis (pitch) → B rotary axis (rotation) → workpiece, and the print head kinematic chain is bed → Z linear axis → nozzle; wherein the Z linear axis moves up and down relative to the workpiece, and the A rotary axis and B rotary axis carry the workpiece posture. In this configuration, the nozzle maintains a constant distance from the workpiece surface during printing. The Z-axis automatically adjusts its height based on changes in the workpiece's real-time posture and position, ensuring that even if rotation of the A and B axes causes changes in the workpiece surface height, the coating distance between the nozzle and the workpiece surface remains constant.

[0041] Under this XYZAB configuration, the process of generating the basic motion axis drive pulse sequence by performing five-axis inverse kinematics calculation based on the XYZ composite feed rate calculated in step 1 is as follows: First, based on the geometric information of the current five-axis linkage trajectory, the XYZ composite feed rate is decomposed into three linear axes: X-axis, Y-axis, and Z-axis. Then, based on the forward kinematics of this configuration, the rotation angle and angular velocity (i.e., workpiece posture) of the A-axis and B-axis at the current trajectory point are obtained, thereby determining the component velocity or component angular velocity of each of the five motion axes. Subsequently, combining the step resolution and drive microstepping of each axis motor, the component velocity or component angular velocity of each axis is converted into the drive pulse frequency and pulse quantity of the stepper motor, ultimately generating the basic motion axis drive pulse sequence containing five sets of stepper pulse signals. During the inverse kinematics process, it is necessary to ensure that the magnitude and direction of the resultant velocity of the five motion axes after synthesis are strictly consistent with the given XYZ composite feed rate. When the printing trajectory involves both spatial translation and workpiece posture changes, the velocity components of the linear axis and the angular velocity components of the rotary axis are solved in real time using the forward kinematics matrix. The velocity or angular velocity assigned to each axis must not exceed the maximum value allowed by the equipment to ensure smooth motion and trajectory accuracy.

[0042] It should be noted that for other five-axis configurations (such as XYZBC type, XYZAC type, etc.), due to their different kinematic models, the inverse kinematic algorithm needs to be re-derived according to their corresponding forward kinematics. However, regardless of the configuration, the core control process of this invention (dynamic compensation, backlash feedforward, hardware synchronization) remains unchanged. Those skilled in the art only need to replace the corresponding kinematic transformation matrix to port the solution of this invention to devices with different configurations.

[0043] For example, assuming the combined XYZ feed rate is 6 mm / s, the tangential vector of the current trajectory in the XY plane is (0.8, 0.6), and the Z-axis velocity is 0 (i.e., the nozzle-workpiece distance is constant at this instant and no Z-axis compensation is needed), and the forward kinematics show that the A-axis angular velocity and B-axis angular velocity at this trajectory point are both 0 (i.e., the workpiece attitude is constant at this instant and only translation is performed), then the X-axis velocity is 4.8 mm / s and the Y-axis velocity is 3.6 mm / s. If the step resolution of the X-axis and Y-axis motors is 0.01 mm / step and the control cycle is 1 ms, then the X-axis needs to output 0.48 steps and the Y-axis needs to output 0.36 steps in each control cycle. There are no drive pulses for the A-axis and B-axis in this cycle. Long-term positioning accuracy is ensured through cumulative error processing. The above example only shows the linear axis decomposition in the XY plane for simplicity. When the trajectory involves attitude changes, the angular velocity decomposition and pulse generation of the A-axis and B-axis are performed synchronously according to the above inverse kinematics process.

[0044] Step 3: Monitor the direction of rotational axis speed in real time. When a change in the direction of rotational axis speed is detected, query the preset backlash error mapping table based on the current position data and current temperature data to generate feedforward compensation pulses. Then, superimpose the feedforward compensation pulses into the basic motion axis drive pulse sequence to generate the compensated final motion axis drive pulse sequence.

[0045] In step 3, the direction of the rotation axis velocity is monitored in real time. When a change in the direction of the rotation axis velocity is detected, a preset backlash error mapping table is queried based on the current position data and the current temperature data to generate a number of feedforward compensation pulses. These feedforward compensation pulses are then superimposed and injected into the basic motion axis drive pulse sequence to generate the compensated final motion axis drive pulse sequence. The specific steps include the following: The rotation axis speed direction is monitored in real time. When the rotation axis speed direction is detected to change, a preset backlash error mapping table is queried according to the current position data and the current temperature data to obtain the backlash error estimate corresponding to the current position and temperature. The number of feedforward compensation pulses is calculated based on the ratio between the estimated backlash error value corresponding to the current position and temperature and the preset motor step resolution. A differential smoothing filter is applied to the number of feedforward compensation pulses to generate filtered feedforward compensation pulses. The filtered feedforward compensation pulses are superimposed and injected into the basic motion axis drive pulse sequence to generate the compensated final motion axis drive pulse sequence.

[0046] Based on the ratio between the estimated backlash error value corresponding to the current position and temperature and the preset motor step resolution, the number of feedforward compensation pulses is calculated. The corresponding relationship in this process is as follows: ; in, Indicates the number of feedforward compensation pulses. This represents the estimated backlash error value corresponding to the current location and temperature. This indicates the preset motor step pitch resolution.

[0047] In a specific embodiment of the present invention, the direction of the rotating shaft velocity is monitored in real time. The direction is determined by detecting the sign of the rotating shaft angular velocity. The rotating shaft angular velocity data is read in each control cycle and compared with the previous control cycle. When a change in the direction of the rotating shaft velocity is detected, a preset backlash error mapping table is consulted based on the current position data and current temperature data to obtain the estimated backlash error corresponding to the current position and temperature. This backlash error mapping table is a position-temperature two-dimensional mapping table, used to provide accurate feedforward compensation pulse counts when the rotating shaft velocity changes direction. This mapping table is pre-built through a standard process during the equipment factory calibration stage and supports online updates during long-term equipment operation to adapt to mechanical wear and performance degradation. The specific methods for its establishment and updating are as follows: (i) The calibration of the mapping table is performed in a precision constant temperature chamber. The temperature control range of the constant temperature chamber covers the entire range of ambient temperatures that the equipment may encounter in actual use, with points evenly distributed from low to high temperatures. At each calibration temperature node, the equipment needs to be kept at the constant temperature chamber for a sufficient period of time to ensure that all components of the equipment (especially the harmonic reducer) reach a sufficient thermal equilibrium state, and to avoid thermal transients interfering with the calibration data.

[0048] During the calibration process, the position feedback of the rotating shaft is measured using a high-precision circular grating encoder. The resolution of this encoder is much higher than the positioning accuracy required for equipment control, and it can accurately capture minute changes in backlash error.

[0049] (II) Division of Location Nodes and Temperature Nodes The entire range of motion of the rotating axes (including the A-axis and B-axis) is divided into several position nodes at equal intervals. The spacing between adjacent position nodes is uniform, ensuring that there are enough sampling points throughout the entire range of motion to characterize the spatial distribution of backlash error. In the temperature dimension, several temperature nodes are divided at equal intervals within the temperature range achievable by the constant temperature chamber. The spacing between adjacent temperature nodes is sufficient to capture the trend of backlash error changing with temperature, while ensuring that the calibration workload is within an acceptable range.

[0050] (III) Methods for collecting backlash error data Under the combined conditions of each location node and each temperature node, the following standardized backlash error measurement procedure is performed: First, the rotating shaft is controlled to approach the target position from the positive direction (i.e., the forward rotation direction). After the shaft comes to a complete stop, the actual position value reached is read and recorded by a circular encoder. Then, the rotating shaft is controlled to approach the same target position again from the opposite direction (i.e., the reverse rotation direction). Similarly, after the shaft comes to a complete stop, the actual position value reached is recorded.

[0051] Due to the inherent gear meshing clearance of the harmonic reducer, there is a fixed difference between the actual positions when approaching the same target position in both directions. Half of this difference is the backlash error value under the conditions of that position node and that temperature node. Its physical meaning is: when the rotating shaft reverses its speed at a certain position, due to the backlash in the transmission chain, the motor needs to rotate an additional angle to eliminate the backlash and restart the load movement; this additional angle is the backlash error that needs to be compensated for.

[0052] At each location-temperature combination node, the above forward and reverse approximation measurements need to be repeated multiple times. The arithmetic mean of the multiple measurement results is taken as the backlash error estimate for that node to eliminate the influence of random measurement noise.

[0053] (iv) Construction of mapping table and data storage After collecting backlash error data for all location and temperature node combinations, the data from all nodes are organized into a two-dimensional table based on both location and temperature dimensions. Rows in the table correspond to different location nodes, and columns correspond to different temperature nodes. Each element in the table represents the backlash error value measured for that location-temperature combination.

[0054] This two-dimensional mapping table is stored in the device's non-volatile memory in a structured data format, ensuring that data is not lost after the device is powered off. Each time the device is powered on, the system automatically loads the complete mapping table data from the non-volatile memory for real-time querying during runtime.

[0055] (v) Runtime mapping table lookup and interpolation methods In actual operation, the current position and temperature of the rotating shaft often do not fall exactly on a node in the mapping table. In this case, the system uses a two-dimensional linear interpolation method to calculate the backlash error estimate under the current conditions based on the backlash error values ​​of the four adjacent position-temperature nodes around the current position.

[0056] Specifically, the system first determines the interval between two adjacent position nodes based on the current position, and simultaneously determines the interval between two adjacent temperature nodes based on the current temperature, thus defining a rectangular region enclosed by the four corner nodes. Then, the system performs two linear interpolations in the temperature dimension to obtain the backlash error estimates for the current position at the two adjacent temperature nodes; finally, it performs one more linear interpolation in the position dimension to obtain the backlash error estimates for the current position and current temperature. This two-dimensional linear interpolation method has low computational cost and fast response, making it suitable for high-frequency execution in embedded real-time systems.

[0057] (vi) Online update and adaptive mechanism of mapping table During long-term operation of the equipment, the backlash error of the harmonic reducer gradually increases due to wear, and the pre-calibrated data in the mapping table may deviate from the actual state. To address this, the present invention introduces an online update mechanism for the mapping table.

[0058] During normal operation of the equipment, each time the rotating shaft completes a full forward and reverse motion cycle—that is, moving from the forward direction through a certain position to the reverse direction and back to that position—the system automatically records the actual backlash error value observed at that position. The system compares this measured value with the predicted value obtained by interpolation from the mapping table. If the deviation between the two exceeds a preset allowable threshold, it indicates that the data near that position in the mapping table is no longer accurate due to wear or performance degradation, and the system immediately triggers a partial update.

[0059] The specific method for local updates is as follows: Using the newly acquired measured value as the center, the values ​​of grid nodes within a small neighborhood of that location in the mapping table are corrected. The correction employs an inverse distance-weighted approach, meaning that nodes closer to the measured location are more affected by the new data, while nodes farther away are less affected. In this way, the mapping table can continuously optimize and adapt during use, always maintaining an effective representation of the current device state.

[0060] The updated mapping table data is persistently stored in non-volatile memory in real time, ensuring that the update results remain valid the next time the device is powered on. This online learning mechanism enables the backlash compensation scheme to remain effective throughout the entire lifecycle of the device, eliminating the need for periodic factory recalibration.

[0061] During the equipment's factory calibration phase, a mapping table is pre-built by measuring the half-difference values ​​of the forward and reverse positioning errors of each node under different temperature conditions. During operation, bilinear interpolation is performed based on the current position and temperature information to obtain an estimated backlash error. Then, the number of feedforward compensation pulses is calculated based on the ratio of the estimated backlash error to the preset motor step resolution. For example, assuming the estimated backlash error is 0.06° and the motor step resolution is 0.01° / step, the number of feedforward compensation pulses is 0.06 / 0.01 = 6 steps. Next, a differential smoothing filter is applied to the number of feedforward compensation pulses to generate filtered feedforward compensation pulses. The purpose of applying the differential smoothing filter to the number of feedforward compensation pulses is that although the compensation pulses accurately reflect the number of pulses required for the backlash error, injecting all of these pulses at once in the current control cycle would cause a sudden change in the frequency of the motion axis drive pulses, which in turn would cause a step change in the speed of the rotating axis. This sudden speed change would leave visible ripple marks on the cured surface of the printed part, affecting the surface quality. Therefore, it is necessary to distribute the compensation pulses into multiple control cycles through smoothing filtering, so that the speed change process is smooth and controllable.

[0062] This invention employs a first-order lag filtering algorithm to achieve the aforementioned smoothing process. The basic idea of ​​this algorithm is that the actual amount of compensation pulses injected in the current control cycle is a weighted sum of the actual amount injected in the previous control cycle and the newly calculated original compensation pulse amount. Specifically, the filtered output value of the current cycle is equal to the filtered output value of the previous cycle multiplied by a positive coefficient less than one, plus the original compensation pulse value of the current cycle multiplied by one minus this coefficient. This positive coefficient is called the filtering coefficient; the closer its value is to one, the stronger the filtering effect, the longer the dispersion time of the compensation pulses, and the smoother the velocity change; conversely, the closer the coefficient is to zero, the weaker the filtering effect, the faster the response speed, and the more concentrated the compensation pulses.

[0063] The filter coefficient needs to be set to match the control cycle. In the embedded real-time control system used in this invention, the control cycle is fixed at the millisecond level. The time constant of the filter determines the filter's response speed; the larger the time constant, the smoother the filtering effect. The formula for calculating the filter coefficient shows that the coefficient is equal to the natural constant raised to the power of the negative ratio of the control cycle to the time constant. In a preferred embodiment of this invention, the control cycle is one millisecond, and the filter time constant is ten milliseconds, resulting in a filter coefficient of approximately 0.905. Under this parameter configuration, the filter's cutoff frequency is approximately sixteen Hz, which is much lower than the natural frequency of the mechanical system and will not generate adverse excitation to the mechanical transmission system. Simultaneously, this cutoff frequency is much higher than the frequency range of encoder quantization noise and temperature sensor measurement noise, effectively filtering out the interference of these high-frequency noises on the compensation pulse calculation results. The filter's group delay is approximately ten milliseconds, which is negligible relative to the mechanical system's response time and will not affect the timeliness of commutation compensation.

[0064] It should be noted that the specific values ​​of the filter coefficients and filter time constant in the above first-order lag filtering algorithm are not uniquely fixed. Those skilled in the art can make adaptive adjustments based on the actual control cycle and the dynamic response characteristics of the mechanical system. In applications with shorter control cycles or where the mechanical system is more sensitive to sudden speed changes, the time constant can be appropriately increased to make the injection process of the compensation pulse smoother; while in applications with longer control cycles or where the system allows for faster response, the time constant can be appropriately decreased to make the compensation effect more timely. All of the above parameter adjustments do not depart from the basic concept and protection scope of this invention.

[0065] After the filtering calculation is completed, the resulting filtered output value is usually a decimal number, but the actual number of pulses injected into the motion control queue must be an integer. Therefore, the system rounds the filtered output value to the nearest integer, using the rounded integer value as the actual number of feedforward compensation pulses injected in the current control cycle. During the recursive process across multiple control cycles, the accumulated error of the decimal part will naturally be distributed into subsequent cycles, preventing systematic deviations.

[0066] The significant advantage of the first-order hysteresis filtering algorithm lies in its extreme computational simplicity. Each control cycle requires only two multiplications and one addition, resulting in minimal computational overhead. This makes it highly suitable for high-frequency execution at millisecond-level control cycles in the embedded real-time processor used in this invention, without increasing processor load or affecting the scheduling of other real-time tasks. Furthermore, this filtering algorithm does not rely on large-capacity data caches or complex state storage, consuming minimal resources, which aligns with the engineering realities of limited embedded computing resources in photopolymerization printing systems.

[0067] The filtered compensation pulses are superimposed onto the basic motion axis drive pulse sequence to generate the final compensated motion axis drive pulse sequence. This sequence is then used to drive the rotary axis motor to perform actual motion and serves as the external clock trigger source for the master timer, triggering the synchronous update of the light source pulse width modulation drive parameters. Through the above differential smoothing filtering process, the commutation compensation pulses achieve smooth injection, eliminating the impact of backlash error on trajectory accuracy and avoiding the damage to surface quality caused by sudden speed changes, thereby ensuring the forming quality of five-axis photopolymerization printing during the rotary axis commutation period.

[0068] It should be noted that the current embodiment uses a position-temperature two-dimensional mapping table, which is suitable for control scenarios using open-loop stepper motors. If a closed-loop stepper or servo motor is used, it can be expanded into a position-load-temperature three-dimensional table to further improve the compensation accuracy.

[0069] Compared with existing technologies, the core innovation of this embodiment lies in the introduction of a feedforward compensation mechanism for harmonic reducer backlash into a five-axis photopolymerization printing scenario for the first time, and the adoption of a "reversal-triggered" query strategy and a position-temperature two-dimensional mapping table to achieve precise compensation. In existing technologies, harmonic reducer backlash compensation is mainly studied in the field of CNC machine tools, but compensation schemes in the machine tool field primarily focus on positioning accuracy, while photopolymerization printing focuses on the surface quality of the continuous coating trajectory—any tiny positional deviation at the reversal moment will leave a mark on the cured surface. In the field of 3D printing, existing solutions mostly remain at the level of mechanical structure optimization, such as using zero-backlash harmonic reducers or double-plate preloaded gear structures, lacking a dynamic compensation mechanism based on temperature and position information; in terms of motion control, existing technologies have a lag in compensating for inter-axis position synchronization, and trajectory drift caused by differences in axis inertia at the moment of switching cannot be suppressed in time. This invention establishes a position-temperature two-dimensional backlash error mapping table and adopts a query strategy of "triggered only during commutation" to significantly reduce the computational load on the microcontroller. It also applies differential smoothing filtering to the compensation pulse to avoid sudden speed changes. This not only eliminates the influence of transmission chain backlash on the continuous coating trajectory, but also avoids the consumption of computational resources caused by continuous queries, thus adapting to the engineering reality of embedded computing resources limited in photopolymerization printing systems.

[0070] Step 4: Using the hardware timer master-slave linkage mechanism, the compensated final motion axis drive pulse sequence is used as the external clock trigger source of the master timer. The update event of the master timer synchronously triggers the comparison register update operation of the slave timer to align the light source pulse width modulation drive parameters with the compensated final motion axis drive pulse sequence in timing, thereby generating a synchronous execution control signal.

[0071] In this embodiment of the invention, hardware-level synchronization between the light source and motion is achieved. The system uses a hardware timer master-slave linkage mechanism to ensure microsecond-level synchronization accuracy between the update of the light source driving parameters and the generation of motion axis driving pulses. Specifically, the master timer is configured in external clock trigger mode, using the generated compensated final motion axis driving pulse sequence as the external clock trigger source. The master timer counts once for each step pulse generated by the motion axis; therefore, the counting frequency of the master timer is proportional to the feed speed of the motion axis. The slave timer is configured in pulse width modulation output mode to output the ultraviolet light source driving signal, and its comparison register update event is synchronously triggered by the master timer update event.

[0072] When the master timer overflows and generates an update event, this event synchronously triggers an update operation in the slave timer's compare register, ensuring that the update of the light source pulse width modulation drive parameters and the generation of the motion axis drive pulses are strictly aligned in timing. The entire synchronization process is completed autonomously by the hardware circuit, unaffected by operating system scheduling cycles, interrupt response times, and communication link delays. When the feed speed changes, the light intensity duty cycle can be updated immediately in the next clock cycle, thus completely eliminating the scheduling delays unavoidable in software polling.

[0073] Step 5: Based on the synchronous execution control signal, control the synchronous execution of the light source exposure action and the motion axis feed action to obtain the final control result.

[0074] In this embodiment of the invention, this is the final execution stage of the entire control method. Based on the generated synchronous execution control signal, the system controls the ultraviolet light source to output exposure according to the light source pulse width modulation driving parameters, while simultaneously controlling each motion axis to perform feed motion according to the compensated final motion axis drive pulse sequence. Because the synchronous execution control signal ensures that the update of the light source drive parameters and the generation of the motion axis drive pulses are strictly aligned in time, the exposure action of the ultraviolet light source and the feed action of the motion axes are kept synchronized at the microsecond level. When the posture of the surface to be cured changes in the printing trajectory, the dynamic compensation model calculates the new target exposure energy density and XYZ composite feed speed in real time. The hardware timer master-slave linkage mechanism ensures the synchronous adjustment of the light source intensity and feed speed, thereby achieving real-time and precise matching of curing energy during the five-axis linkage printing process, resulting in the final control result.

[0075] Please see Figure 3 The present invention also proposes a dynamic compensation and real-time synchronization control system for five-axis photopolymerization printing, the system comprising: The host computer module is used for: The real-time attitude angle data and the real-time angular velocity data of the rotation axis of the surface to be cured are obtained. The real-time attitude angle data and the real-time angular velocity data of the rotation axis of the surface to be cured are used as two input feature parameters and input into the preset dynamic compensation model for calculation to obtain the target exposure energy density and XYZ synthesis feed rate, respectively. Real-time control module, used for: The corresponding light source pulse width modulation driving parameters are generated based on the target exposure energy density mapping; the five-axis inverse kinematics calculation is performed based on the XYZ composite feed rate to generate the basic motion axis driving pulse sequence; Motion drive module, used for: The rotation axis speed direction is monitored in real time. When a change in the rotation axis speed direction is detected, the preset backlash error mapping table is queried according to the current position data and the current temperature data to generate feedforward compensation pulses. The feedforward compensation pulses are then superimposed and injected into the basic motion axis drive pulse sequence to generate the compensated final motion axis drive pulse sequence. The light source driver module is used for: Through the hardware timer master-slave linkage mechanism, the compensated final motion axis drive pulse sequence is used as the external clock trigger source of the master timer, and the update operation of the slave timer's compare register is synchronously triggered by the update event of the master timer, so as to align the light source pulse width modulation drive parameters with the compensated final motion axis drive pulse sequence in timing and generate a synchronous execution control signal. Master-slave timer linkage unit, used for: Based on the synchronous execution control signal, the synchronous execution of the light source exposure action and the motion axis feed action is controlled to obtain the final control result.

[0076] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0077] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A dynamic compensation and real-time synchronization control method for five-axis photopolymerization printing, characterized in that, The method includes the following steps: Step 1: Obtain the real-time attitude angle data and the real-time angular velocity data of the rotation axis of the surface to be cured. Use the real-time attitude angle data and the real-time angular velocity data of the rotation axis as two input feature parameters and input them into the preset dynamic compensation model for calculation to obtain the target exposure energy density and XYZ synthesis feed rate, respectively. Step 2: Generate the corresponding light source pulse width modulation driving parameters based on the target exposure energy density mapping; perform five-axis inverse kinematics calculation based on the XYZ synthesis feed rate to generate the basic motion axis driving pulse sequence; Step 3: Monitor the direction of rotational axis speed in real time. When a change in the direction of rotational axis speed is detected, query the preset backlash error mapping table based on the current position data and current temperature data to generate feedforward compensation pulses. Then, superimpose the feedforward compensation pulses into the basic motion axis drive pulse sequence to generate the compensated final motion axis drive pulse sequence. Step 4: Through the hardware timer master-slave linkage mechanism, the compensated final motion axis drive pulse sequence is used as the external clock trigger source of the master timer, and the update operation of the slave timer's compare register is synchronously triggered by the update event of the master timer, so as to align the light source pulse width modulation drive parameters with the compensated final motion axis drive pulse sequence in timing and generate a synchronous execution control signal. Step 5: Based on the synchronous execution control signal, control the synchronous execution of the light source exposure action and the motion axis feed action to obtain the final control result.

2. The dynamic compensation and real-time synchronization control method for five-axis photopolymerization printing according to claim 1, characterized in that, In step 1, the real-time attitude angle data and the real-time angular velocity data of the rotation axis of the surface to be cured are obtained. These two data are used as two input feature parameters and are input into a preset dynamic compensation model for calculation to obtain the target exposure energy density and XYZ composite feed rate, respectively. The specific steps include the following: Acquire real-time attitude angle data of the surface to be cured and real-time angular velocity data of the rotation axis; The real-time attitude angle data and the real-time angular velocity data of the rotation axis are used as two input feature parameters; Obtain the baseline set layer thickness, input the two input feature parameters into the preset dynamic compensation model for calculation, take the cosine value of the real-time attitude angle data, and divide the baseline set layer thickness by the cosine value to obtain the geometrically magnified layer thickness; calculate the product of the absolute value of the real-time angular velocity data of the rotation axis and the preset first resin characteristic coefficient to obtain the shear thinning correction value; calculate the product of the square of the real-time angular velocity data of the rotation axis and the preset second resin characteristic coefficient to obtain the fluid inertial accumulation correction value; add the numerical value 1, the shear thinning correction value and the fluid inertial accumulation correction value to obtain the angular velocity comprehensive correction factor; multiply the geometrically magnified layer thickness and the angular velocity comprehensive correction factor to obtain the effective layer thickness data; Obtain the baseline exposure energy density and calculate the ratio of the effective layer thickness data to the baseline set layer thickness to obtain the layer thickness magnification. The cosine value of the real-time attitude angle data is taken and the reciprocal of the cosine value is calculated to obtain the light intensity projection attenuation compensation value; the reference exposure energy density, the layer thickness magnification ratio and the light intensity projection attenuation compensation value are multiplied to calculate the target exposure energy density. The XYZ synthesis feed rate is obtained based on the target exposure energy density and the preset control mode; wherein, the preset control mode includes: speed adjustment mode.

3. The dynamic compensation and real-time synchronization control method for five-axis photopolymerization printing according to claim 2, characterized in that, The process of obtaining the XYZ composite feed rate based on the target exposure energy density and the preset control mode specifically includes: When the preset control mode is speed adjustment mode, the preset UV light source intensity is kept constant. The feed speed adjustment ratio is obtained by calculating the ratio of the reference exposure energy density to the target exposure energy density. The reference feed speed is obtained and multiplied by the feed speed adjustment ratio to obtain the XYZ composite feed speed.

4. The dynamic compensation and real-time synchronization control method for five-axis photopolymerization printing according to claim 3, characterized in that, The baseline set layer thickness is obtained, and the two input feature parameters are input into a preset dynamic compensation model for calculation. The geometrically amplified layer thickness is obtained by taking the cosine value of the real-time attitude angle data and dividing the baseline set layer thickness by the cosine value. The absolute value of the real-time angular velocity data of the rotation axis is calculated and multiplied by a preset first resin characteristic coefficient to obtain a shear thinning correction value. The square of the real-time angular velocity data of the rotation axis is calculated and multiplied by a preset second resin characteristic coefficient to obtain a fluid inertial accumulation correction value. The numerical value 1, the shear thinning correction value, and the fluid inertial accumulation correction value are added together to obtain an angular velocity comprehensive correction factor. The geometrically amplified layer thickness is multiplied by the angular velocity comprehensive correction factor to obtain the effective layer thickness data. The corresponding relationship in this process is as follows: ； in, Indicates the effective layer thickness data. Indicates time, Indicates the reference layer thickness. Represents the cosine function. This represents real-time attitude angle data. Indicates the thickness of the geometrically enlarged layer. This indicates the preset first resin characteristic coefficient. This represents the absolute value of the real-time angular velocity data of the rotating axis. This represents the shear thinning correction value. This indicates the preset second resin characteristic coefficient. This represents the correction value for fluid inertial deposition. This represents the comprehensive correction factor for angular velocity.

5. The dynamic compensation and real-time synchronization control method for five-axis photopolymerization printing according to claim 4, characterized in that, The ratio of the effective layer thickness data to the reference set layer thickness is calculated to obtain the layer thickness magnification; the cosine value of the real-time attitude angle data is taken and the reciprocal of the cosine value is calculated to obtain the light intensity projection attenuation compensation value; the reference exposure energy density, the layer thickness magnification, and the light intensity projection attenuation compensation value are multiplied to obtain the target exposure energy density. The corresponding relationship in this process is as follows: ； in, Indicates the target exposure energy density. Indicates the baseline exposure energy density. Indicates the magnification factor of the layer thickness. This represents the light intensity projection attenuation compensation value.

6. The dynamic compensation and real-time synchronization control method for five-axis photopolymerization printing according to claim 5, characterized in that, Calculate the ratio of the reference exposure energy density to the target exposure energy density to obtain the feed rate adjustment ratio; obtain the reference feed rate, and multiply the reference feed rate by the feed rate adjustment ratio to obtain the XYZ composite feed rate. The corresponding relationship in this process is as follows: ； in, Indicates the XYZ composite feed rate. Indicates the reference feed rate. Indicates the feed rate adjustment ratio; The reference exposure energy density is calculated by multiplying the ratio between the reference set layer thickness and the reference feed rate by a preset constant UV light source intensity. The corresponding relationship is as follows: ； in, This indicates a preset, constant UV light source intensity.

7. The dynamic compensation and real-time synchronization control method for five-axis photopolymerization printing according to claim 6, characterized in that, In step 3, the direction of the rotation axis speed is monitored in real time. When a change in the direction of the rotation axis speed is detected, a preset backlash error mapping table is queried based on the current position data and the current temperature data to generate a number of feedforward compensation pulses. The number of feedforward compensation pulses is then superimposed and injected into the basic motion axis drive pulse sequence to generate the compensated final motion axis drive pulse sequence. The specific steps include the following: The rotation axis speed direction is monitored in real time. When the rotation axis speed direction is detected to change, a preset backlash error mapping table is queried according to the current position data and the current temperature data to obtain the backlash error estimate corresponding to the current position and temperature. The number of feedforward compensation pulses is calculated based on the ratio between the estimated backlash error value corresponding to the current position and temperature and the preset motor step resolution. A differential smoothing filter is applied to the number of feedforward compensation pulses to generate filtered feedforward compensation pulses. The filtered feedforward compensation pulses are superimposed and injected into the basic motion axis drive pulse sequence to generate the compensated final motion axis drive pulse sequence.

8. The dynamic compensation and real-time synchronization control method for five-axis photopolymerization printing according to claim 7, characterized in that, Based on the ratio between the estimated backlash error value corresponding to the current position and temperature and the preset motor step resolution, the number of feedforward compensation pulses is calculated. The corresponding relationship in this process is as follows: ； in, Indicates the number of feedforward compensation pulses. This represents the estimated backlash error value corresponding to the current location and temperature. This indicates the preset motor step pitch resolution.

9. A dynamic compensation and real-time synchronization control system for five-axis photopolymerization printing, characterized in that, The system employs the dynamic compensation and real-time synchronization control method for five-axis photopolymerization printing as described in any one of claims 1-8, and the system comprises: The host computer module is used for: The real-time attitude angle data and the real-time angular velocity data of the rotation axis of the surface to be cured are obtained. The real-time attitude angle data and the real-time angular velocity data of the rotation axis of the surface to be cured are used as two input feature parameters and input into the preset dynamic compensation model for calculation to obtain the target exposure energy density and XYZ synthesis feed rate, respectively. Real-time control module, used for: The corresponding light source pulse width modulation driving parameters are generated based on the target exposure energy density mapping; the five-axis inverse kinematics calculation is performed based on the XYZ composite feed rate to generate the basic motion axis driving pulse sequence; Motion drive module, used for: The rotation axis speed direction is monitored in real time. When a change in the rotation axis speed direction is detected, the preset backlash error mapping table is queried according to the current position data and the current temperature data to generate feedforward compensation pulses. The feedforward compensation pulses are then superimposed and injected into the basic motion axis drive pulse sequence to generate the compensated final motion axis drive pulse sequence. The light source driver module is used for: Through the hardware timer master-slave linkage mechanism, the compensated final motion axis drive pulse sequence is used as the external clock trigger source of the master timer, and the update operation of the slave timer's compare register is synchronously triggered by the update event of the master timer, so as to align the light source pulse width modulation drive parameters with the compensated final motion axis drive pulse sequence in timing and generate a synchronous execution control signal. Master-slave timer linkage unit, used for: Based on the synchronous execution control signal, the synchronous execution of the light source exposure action and the motion axis feed action is controlled to obtain the final control result.

Images