3d printing warehouse replenishment control method and system combining laser ranging and ai algorithm

Abstract

The present application relates to the field of 3D printing technology, and more particularly to a 3D printing warehouse replenishment control method and system combining laser ranging and AI algorithm, which first collects laser material level, nozzle coordinates and path layering data, constructs a full-state perception data set through denoising fusion; secondly, based on the deviation of material level and material consumption rate, a layered response link is dynamically constructed, and the transmission priority of the interrupt signal and the continuous printing parameters is adjusted according to the emergency level; then, the priority is converted into a weight parameter, a continuous printing matching instruction set is generated combined with the breakpoint position, and the state after replenishment is synchronized across modules; finally, the extrusion system is driven to perform breakpoint alignment and parameter adaptation action. This method uses non-contact laser ranging to solve the monitoring problem of high-viscosity materials, realizes active early warning and full-automatic high-precision continuous printing through AI dynamic regulation and control, effectively eliminates manual reset errors, ensures the continuity and interlayer bonding quality of large-size component printing, and significantly reduces the scrap rate.

Description

Technical Field

[0001] This invention relates to the field of 3D printing technology, and in particular to a 3D printing chamber feeding control method and system that combines laser ranging and AI algorithms. Background Technology

[0002] With the rapid development of additive manufacturing technology, 3D printing of large-size concrete structural components has shown broad application prospects in infrastructure fields such as wind turbine towers, buildings, and bridges. However, in the long-cycle, continuous UHPC (Ultra-High Performance Concrete) 3D printing process, the stability of the feeding system directly affects the printing quality and forming accuracy. In existing technologies, the feeding hopper usually uses weighing sensors or ultrasonic sensors to monitor the material level. However, UHPC materials have high viscosity and easy adhesion characteristics, which makes traditional sensors susceptible to interference, resulting in poor measurement accuracy and difficulty in accurately reflecting the actual material level changes in real time. Once the feeding is interrupted, the system lacks a precise identification and recording mechanism for the breakpoint location. After restarting, manual intervention is required to locate the breakpoint and adjust parameters. This is not only inefficient but also prone to problems such as interlayer misalignment and mismatched extrusion parameters, leading to insufficient interlayer bonding strength and even scrapping of the entire part.

[0003] Furthermore, most current 3D printing control systems employ a centralized control architecture. When faced with multi-site collaborative material feeding and complex printing paths, they exhibit significant response lag and control delays, making it difficult to achieve rapid and intelligent resume printing after material loss. Especially in unattended or automated production lines, the lack of a breakpoint resume matching mechanism based on real-time sensing and intelligent prediction severely restricts the application of 3D printing technology in industrial and large-scale production. Therefore, there is an urgent need for a control method capable of high-precision material level sensing, intelligent assessment of material loss risk, and automatic breakpoint location and resume parameter matching to improve the continuity, stability, and intelligence of the 3D printing process.

[0004] The above content is only used to help understand the technical solution of the present invention and does not represent an admission that the above content is prior art. Summary of the Invention

[0005] The main objective of this invention is to provide a 3D printing chamber feeding control method and system that combines laser ranging and AI algorithms. This aims to solve the technical problems of existing large-size UHPC 3D printing technology, which suffers from inaccurate traditional material level monitoring due to highly viscous materials, lack of intelligent breakpoint recognition and adaptive resuming mechanism, resulting in reliance on manual intervention and low reset accuracy after material supply interruption, which can easily lead to interlayer bonding failure or even scrapping of the entire piece.

[0006] To achieve the above objectives, this invention provides a 3D printing chamber replenishment control method combining laser ranging and AI algorithms, the method comprising: Real-time material level data from laser ranging in the 3D printing hopper, real-time coordinate parameters of the printing nozzle, and path layering parameters of the printed components are collected. The material level and printing status data collected from multiple sources are denoised and fused to obtain a full-state perception dataset for printing. Based on the difference between the material level change rate and the material consumption rate in the full-state perception dataset of printing, a hierarchical response link is dynamically constructed between the printing control system and the material supply drive system. The transmission priority of the interruption signal and the resume printing parameter in the hierarchical response link is adjusted according to the urgency of the material level falling below the threshold, so as to obtain the adjusted signal transmission priority. The adjusted signal transmission priority is converted into a response weight parameter, and combined with the position parameter of the current breakpoint to generate a continuation matching instruction set; The continuous printing matching instruction set is synchronized across modules with the real-time material supply status after the replenishment is completed to obtain the synchronized continuous printing matching instruction set. The synchronized continue printing matching instruction set drives the printing extrusion system to perform continue printing actions with breakpoint alignment and parameter adaptation, so as to obtain a high-precision 3D printing breakpoint continue printing control scheme.

[0007] Optionally, the step of converting the adjusted signal transmission priority into a response weight parameter and generating a continuation matching instruction set in combination with the position parameter of the current breakpoint includes: The hierarchical response link is divided into intervals based on the adjusted signal transmission priority level value, and the level value is converted within the interval to obtain the response weight parameter corresponding to the hierarchical response link. Extract the printed layer thickness deviation parameter and the extruder cumulative displacement deviation parameter at the current breakpoint position, and combine the printed layer thickness deviation and the extruder displacement deviation into a breakpoint position error coefficient; The response weight parameters and the breakpoint position error coefficient are jointly constrained to obtain the weight parameter distribution result. Based on the weight parameter distribution result, a continuation call generation template corresponding to each level in the hierarchical response link is constructed. The breakpoint position error coefficient is input into the continuous printing instruction generation template to generate a continuous printing matching instruction set that includes breakpoint alignment compensation and extrusion flow correction constraints.

[0008] Optionally, the step of jointly constraining the response weight parameters and the breakpoint position error coefficient to obtain the weight parameter distribution result includes: The range of variation of the breakpoint position error coefficient is divided into continuous sub-intervals, and each sub-interval corresponds to a coordinate transformation function; Based on the hierarchical position of the response weight parameter in the hierarchical response link, obtain the initial alignment conversion factor associated with the hierarchical position; Within the sub-interval, the breakpoint position error coefficient and the initial alignment transformation factor are correlated and calculated according to the coordinate transformation function to obtain the correlated sub-interval alignment parameters; The alignment parameters of the associated sub-intervals are fused by the interval boundaries to obtain the fused alignment parameters. The inter-level precision balance constraint is then applied to the fused alignment parameters to obtain the weight parameter distribution result.

[0009] Optionally, within the sub-interval, the process of calculating the correlation between the breakpoint position error coefficient and the initial alignment transformation factor based on the coordinate transformation function to obtain the correlated sub-interval alignment parameters includes: The slope adjustment factor of the coordinate transformation function is determined according to the span range of the sub-interval, and the breakpoint position error coefficient is scaled according to the slope adjustment factor to generate the breakpoint deviation value. Within the sub-interval, the alignment accuracy adjustment factor is obtained according to the hierarchical position corresponding to the response weight parameter, and the breakpoint deviation value is cross-correlated with the alignment accuracy adjustment factor to obtain the initial alignment conversion factor. The initial alignment conversion factor is subjected to interval accumulation processing, and the accumulation processing result is superimposed with the breakpoint deviation value to generate the associated sub-interval alignment parameters.

[0010] Optionally, the process of collecting real-time material level data from the laser ranging system in the 3D printing hopper, real-time coordinate parameters of the printing nozzle, and path layering parameters of the printed components, and performing noise reduction and fusion processing on the multi-source collected material level and printing status data, yields a full-state printing perception dataset, including: The scanning coverage area of ​​the laser ranging sensor is divided into blind spot supplementation areas, and the laser ranging material level data is segmented into a discrete point cloud dataset containing spatial identifiers. Within the discrete point cloud dataset, duplicate feature matching is performed on the overlapping area data of adjacent scan frames, and redundant point cloud data points exceeding the set duplicate threshold are deleted to generate a deredundant material level dataset. Extract dynamic material consumption feature points and static material supply silo structure feature points related to material level changes from the deredundant material level dataset, separate the dynamic material consumption feature points from the static material supply silo structure feature points, and generate dynamic feature datasets and static feature datasets. The dynamic feature dataset is extended and predicted according to the material consumption rate of the current printing path to obtain an extended breakpoint coordinate sequence. The extended breakpoint coordinate sequence is then spatially concatenated with the static feature dataset to generate a full-state printing awareness dataset.

[0011] Optionally, the step of extending and predicting the breakpoint coordinates of the dynamic feature dataset according to the material consumption rate of the current printing path to obtain an extended breakpoint coordinate sequence includes: Extract the measured coordinate values ​​of the print head for continuous time periods from the dynamic feature dataset, and calculate the trend vector of print head movement based on the change in the measured coordinate values ​​of the print head; Using the measured coordinates of the print head at the moment of shutdown as the reference point, along the opposite direction of the trend vector, and based on the preset alignment compensation step size, the coordinate prediction interval of the breakpoint position is obtained, and a sequence of predicted breakpoint coordinate points is generated. The predicted breakpoint coordinate sequence is subjected to prediction confidence verification processing. The verified predicted breakpoint coordinate sequence is then merged with the dynamic feature dataset by timestamp alignment to obtain the extended breakpoint coordinate sequence.

[0012] Optionally, the step of driving the printing extrusion system to perform breakpoint alignment and parameter adaptation follow-up actions through the synchronized follow-up matching instruction set to obtain a high-precision 3D printing breakpoint follow-up control scheme includes: The synchronized continuous printing matching instruction set is decomposed into a continuous printing action sequence corresponding to the printing extrusion system. The continuous printing action sequence includes the breakpoint target coordinates, the alignment compensation amount, and the extrusion parameter correction label. Based on the extrusion parameter correction label, the AI-calculated resource occupancy ratio is allocated to the printing control system, and the coordinate search range of the breakpoint alignment search is determined according to the resource occupancy ratio. The boundary conditions of the coordinate search range are dynamically adjusted based on the target coordinates of the breakpoint and the resource occupancy ratio, and the error buffer of the nozzle positioning is calculated based on the adjusted coordinate search range. The predicted breakpoint coordinate parameters and the current material supply status of 3D printing are aligned and then input into the error buffer area to generate the 3D printing chamber replenishment breakpoint continuation matching control scheme that combines laser ranging and AI algorithm.

[0013] Furthermore, to achieve the above objectives, the present invention also provides a 3D printing silo replenishment control system combining laser ranging and AI algorithms, the system comprising: The multi-source fusion module is used to collect real-time material level data from laser ranging in the 3D printing hopper, real-time coordinate parameters of the printing nozzle, and path layering parameters of the printed components. It performs noise reduction and fusion processing on the material level and printing status data collected from multiple sources to obtain a full-state printing perception dataset. The dynamic optimization module is used to dynamically construct a hierarchical response link between the printing control system and the material feeding drive system based on the difference between the material level change rate and the material consumption rate in the printing full-state perception dataset, and adjust the transmission priority of the interruption signal and the continuation printing parameter in the hierarchical response link according to the urgency of the material level falling below the threshold, so as to obtain the adjusted signal transmission priority. The instruction generation module is used to convert the adjusted signal transmission priority into response weight parameters, and generate a continuation matching instruction set in combination with the position parameters of the current breakpoint. The cross-domain synchronization module is used to synchronize the continuous printing matching instruction set with the real-time material supply status after the replenishment is completed, so as to obtain the synchronized continuous printing matching instruction set. The precision continue printing module is used to drive the printing extrusion system to perform breakpoint alignment and parameter adaptation continue printing actions through the synchronized continue printing matching instruction set, so as to obtain a high-precision 3D printing breakpoint continue printing control scheme.

[0014] Furthermore, to achieve the above objectives, the present invention also provides a 3D printing chamber replenishment control device combining laser ranging and AI algorithms. The device includes: a memory, a processor, and a 3D printing chamber replenishment control program combining laser ranging and AI algorithms stored in the memory and executable on the processor. The 3D printing chamber replenishment control program combining laser ranging and AI algorithms is configured to implement the steps of the 3D printing chamber replenishment control method combining laser ranging and AI algorithms as described above.

[0015] Furthermore, to achieve the above objectives, the present invention also provides a computer-readable storage medium storing a 3D printing chamber replenishment control program that combines laser ranging and AI algorithms. When the 3D printing chamber replenishment control program that combines laser ranging and AI algorithms is executed by a processor, it implements the steps of the 3D printing chamber replenishment control method that combines laser ranging and AI algorithms as described above.

[0016] This invention provides a 3D printing chamber replenishment control method combining laser ranging and AI algorithms. Addressing the issue of sensor failure caused by the adhesion of high-viscosity concrete such as UHPC, which leads to the use of traditional weighing or ultrasonic ranging, this method employs non-contact laser ranging technology, effectively avoiding material adhesion interference. By multi-source denoising and fusion of laser material level data, nozzle coordinates, and path layering parameters, a high-fidelity full-state printing perception dataset is constructed, ensuring millisecond-level accurate control of remaining material and printing progress under complex working conditions, providing a reliable data foundation for intelligent decision-making. Furthermore, a dynamic layered response link based on the deviation between the material level change rate and the material consumption rate is innovatively introduced. The system no longer relies on fixed low liquid level alarms. Instead, it can predict material supply gaps based on real-time printing speed and dynamically adjust the transmission priority of interruption signals and resume printing parameters. This proactive early warning mechanism can trigger a shutdown within the optimal window before material depletion, completely eliminating unexpected material interruptions and nozzle blockages caused by material supply delays, ensuring the continuity of long-cycle printing tasks. Through AI algorithms, signal priorities are converted into response weights, and a resume printing matching instruction set containing precise coordinate compensation and flow correction is automatically generated. After material replenishment, the system can automatically complete nozzle alignment, layer thickness compensation, and extrusion parameter adaptation without manual intervention. This not only solves the industry pain points of large errors and low reset success rates in manually finding breakpoints, but also ensures seamless integration of the resume printing point with the original components in terms of microstructure, significantly improving interlayer bonding strength. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating an embodiment of the 3D printing silo replenishment control method combining laser ranging and AI algorithms of the present invention. Figure 2 This is a structural block diagram of an embodiment of the 3D printing silo replenishment control system that combines laser ranging and AI algorithms according to the present invention.

[0018] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0019] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0020] Reference Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of the 3D printing chamber replenishment control method combining laser ranging and AI algorithms of the present invention. An embodiment of the 3D printing chamber replenishment control method combining laser ranging and AI algorithms of the present invention is presented.

[0021] In one embodiment, the 3D printing chamber replenishment control method combining laser ranging and AI algorithms includes: Step S100: Collect real-time material level data from laser ranging in the 3D printing hopper, real-time coordinate parameters of the printing nozzle, and path layering parameters of the printed components. Perform noise reduction and fusion processing on the multi-source collected material level and printing status data to obtain a full-state perception dataset for printing.

[0022] The 3D printing feed hopper can be a container device for storing and continuously supplying highly viscous concrete materials such as UHPC, and can be used as a physical carrier and temporary material storage unit for laser ranging. In this embodiment, the characteristics of the inner wall of the 3D printing feed hopper determine the necessity of non-contact measurement. The real-time material level data of laser ranging can be a sequence of instantaneous distance values ​​on the surface of UHPC material in the feed hopper obtained by a non-contact laser sensor, which can be used to provide a physical layer material level reference that is not affected by material adhesion. For example, the real-time material level data of laser ranging can be obtained by measuring the distance from the sensor to the material surface using the laser time-of-flight or phase difference principle and outputting a digital signal after analog-to-digital conversion. Furthermore, the real-time material level data of laser ranging can be coordinated with the real-time coordinate parameters of the printing nozzle and converted from one-dimensional distance to three-dimensional volume through spatial mapping, or it can be coordinated with noise reduction and fusion processing and used as the core input source of a multi-source sensing dataset. In an exemplary embodiment, the real-time material level data of laser ranging can include, but is not limited to, raw echo signal data, filtered distance point cloud data, and time-series smoothed material level curve data. Real-time coordinate parameters of the printhead can be the instantaneous position and attitude data of the printhead in three-dimensional space. These parameters can be fused with material level data to calculate the actual material consumption rate and provide a spatial reference for breakpoint positioning. The path layering parameters of the printed component can be structured data describing the geometric path, layer height, and process settings of the completed printing layers. These parameters can be used to infer the theoretical material consumption and compare it with actual material level changes to identify material supply anomalies.

[0023] Collecting real-time material level data from the 3D printing hopper using laser ranging, real-time coordinate parameters of the printing nozzle, and path layering parameters of the printed components can be achieved by simultaneously acquiring three heterogeneous data streams through the interfaces of a laser sensor, encoder, and controller. Furthermore, collecting these data allows for the establishment of a comprehensive perception foundation covering materials, motion, and processes, thus replacing single-sensor monitoring. Denoising and fusing the multi-source material level and printing status data can be performed by applying filtering algorithms to remove outliers and aligning the data with timestamps to integrate the multi-source data into a unified dataset. In one specific embodiment, this denoising and fusing can be achieved by using Kalman filtering combined with a sliding window mean method for noise suppression, or by using wavelet transform to separate high-frequency noise from effective trend signals. This eliminates interference factors such as UHPC wall adhesion and vibration, resulting in a high-fidelity, full-state perception dataset for printing.

[0024] The printing full-state perception dataset can be a high-fidelity, multi-dimensional time-series data set that integrates material level, nozzle position, and path parameters and has undergone noise reduction processing. It can be used to characterize the true coupling state between material consumption and component formation during the printing process. For example, the printing full-state perception dataset can be obtained by performing spatiotemporal alignment and outlier removal on laser material level data, nozzle coordinates, and path layer parameters, and constructing a structured data stream with a unified timestamp. Furthermore, the printing full-state perception dataset can collaborate with the layered response link and serve as the sole data source for calculating the difference between the material level change rate and the material consumption rate deviation. It can also collaborate with AI algorithms to support the automatic generation of continuous printing matching instruction sets. In this embodiment, the printing full-state perception dataset may include, but is not limited to, subsets of material inventory status, subsets of nozzle movement trajectories, and subsets of inter-layer process parameters.

[0025] Step S200: Based on the difference between the material level change rate and the material consumption rate in the full-state perception dataset of printing, a hierarchical response link is dynamically constructed between the printing control system and the material feeding drive system. The transmission priority of the interrupt signal and the resume printing parameter in the hierarchical response link is adjusted according to the urgency of the material level falling below the threshold, so as to obtain the adjusted signal transmission priority.

[0026] The material level change rate can be the actual reduction in the volume or height of material in the feeding hopper per unit time. It can be used to reflect the actual material consumption and serve as a measured input variable for constructing the hierarchical response link. The material consumption rate deviation difference can be the difference between the actual material level change rate and the theoretical material consumption rate calculated based on path parameters. It can be used to quantify the degree of supply-demand imbalance in the feeding system and serve as a core criterion for triggering dynamic response. The printing control system can be a central processing unit responsible for path planning, motion control, and process execution. It can be used to receive signals from the hierarchical response link and coordinate the movement of the printhead and the execution of continuous printing commands. The feeding drive system can be a power execution unit that controls the operation of the material conveying pump or screw feeder. It can be used to respond to the replenishment or shutdown commands of the hierarchical response link and maintain the continuity of material supply.

[0027] Based on the deviation between the material level change rate and the material consumption rate from the full-state printing sensing dataset, a hierarchical response link is dynamically constructed between the printing control system and the material feeding drive system. This can be achieved by calculating the deviation between the measured and theoretical material consumption rates in real time and dynamically configuring the hierarchical structure and communication strategy of the response link accordingly. For example, this dynamic construction of a hierarchical response link between the printing control system and the material feeding drive system based on the deviation between the material level change rate and the material consumption rate from the full-state printing sensing dataset can be achieved by automatically increasing the link sampling frequency and switching to a high-priority channel when the deviation exceeds a safety threshold, or by adaptively adjusting the link bandwidth allocation ratio according to the deviation change rate. This enables a shift from passive alarm to proactive risk prediction and ensures that the response mechanism matches the real-time operating conditions.

[0028] A hierarchical response link can be a multi-level dynamic communication and logic judgment channel connecting the printing control system and the feeding drive system. It can be used to replace the fixed threshold alarm mode and realize proactive prediction and hierarchical response to feeding risks. In an exemplary embodiment, the hierarchical response link can dynamically configure the transmission queue, processing priority, and feedback mechanism of interruption signals and resume printing parameters based on the rate deviation in real-time sensing data. Furthermore, the hierarchical response link can coordinate with signal transmission priority and dynamically adjust the link bandwidth allocation according to the urgency level. It can also coordinate with response weight parameters and convert the physical layer urgency into the control layer execution weight. In this embodiment, the hierarchical response link may include, but is not limited to, an emergency stop response layer, an early warning speed adjustment response layer, and a regular feeding response layer. The interruption signal can be a control message indicating that printing is paused or feeding has stopped. It can be used to trigger a safety stop and prevent empty squeezing or blockage when the material level falls below the threshold. The resume printing parameters can be a set of process settings required to resume printing. They can be used to ensure the consistency of the resume printing process with the state before the interruption in terms of flow rate, speed, etc.

[0029] Adjusting the transmission priority of interrupt signals and resume parameters in the hierarchical response link based on the urgency of the material level falling below a threshold can be achieved by dynamically rearranging the execution order of the signal queue according to the magnitude and rate at which the material level falls below the preset threshold. Furthermore, adjusting the transmission priority of interrupt signals and resume parameters in the hierarchical response link based on the urgency of the material level falling below the threshold ensures zero-delay transmission of emergency signals, thereby preventing unexpected material shortages or equipment damage caused by communication congestion. The adjusted signal transmission priority can be considered the scheduling order of interrupt signals and resume parameters in the hierarchical response link, which can be used to ensure priority transmission of high-urgency signals and avoid material shortage accidents caused by response delays.

[0030] Step S300: The adjusted signal transmission priority is converted into response weight parameters, and a continuation matching instruction set is generated by combining the position parameters of the current breakpoint.

[0031] The process of converting the adjusted signal transmission priority into response weight parameters can be achieved by using a mapping function to transform discrete priority levels into continuous numerical weights. Furthermore, this conversion allows the AI ​​algorithm to quantitatively utilize the abstract urgency level, thus supporting the generation of refined continuation printing parameters. The response weight parameters can be numerical execution coefficients derived from the quantification of signal transmission priority, serving as weighting factors in the AI ​​algorithm's generation of the continuation printing matching instruction set and influencing the compensation intensity. The position parameter of the current breakpoint can be the precise three-dimensional coordinates and layer number information of the printhead at the moment of printing interruption, providing spatial anchors for the continuation printing matching instruction set and ensuring reset accuracy. Generating the continuation printing matching instruction set by combining the position parameter of the current breakpoint involves the AI ​​algorithm taking the response weights and breakpoint coordinates as input and outputting a complete sequence of continuation printing instructions including compensation. For example, generating a set of continuing-painting matching instructions by combining the position parameters of the current breakpoint can be achieved by training a regression model based on historical successful continuing-painting samples to predict the optimal compensation parameters, or by using a physical simulation model to solve the process parameter combination that maximizes the interlayer bonding strength in real time. This can automatically generate a high-precision continuing-painting scheme and eliminate reliance on human experience to ensure the quality of interlayer bonding.

[0032] The continuous printing matching instruction set can be an automated continuous printing control command sequence containing parameters for breakpoint coordinate compensation, flow correction, and layer thickness adaptation. It can be used to drive the printing extrusion system to automatically complete precise alignment and parameter adaptation after material replenishment. In one specific embodiment, the continuous printing matching instruction set can be generated by an AI algorithm based on response weight parameters and current breakpoint position parameters, and then encapsulated into standard G-code or proprietary protocol instructions after cross-module synchronous verification. Furthermore, the continuous printing matching instruction set can coordinate with the real-time material feeding status and ensure the material feeding system is ready when the instruction is executed through cross-module synchronization. It can also coordinate with the printing extrusion system and directly drive the actuator to complete the continuous printing action. In this embodiment, the continuous printing matching instruction set may include, but is not limited to, spatial positioning compensation instructions, extrusion flow correction instructions, and interlayer transition process instructions.

[0033] Step S400: Synchronize the continuous printing matching instruction set with the real-time material supply status after the replenishment is completed across modules to obtain the synchronized continuous printing matching instruction set.

[0034] The real-time feeding status can be the pressure, flow rate, and material level confirmation information of the feeding system after replenishment. This can be used as a verification condition for cross-module synchronization and to prevent continued printing from being executed when the feeding is not stable. Synchronizing the continued printing matching instruction set with the real-time feeding status after replenishment can be done by marking and releasing the continued printing instruction set after the feeding system provides a ready signal. Furthermore, synchronizing the continued printing matching instruction set with the real-time feeding status after replenishment ensures that continued printing is only executed after the feeding is stable, thus preventing secondary breakpoints or molding defects caused by feeding fluctuations. The synchronized continued printing matching instruction set can be the final instruction sequence that has been verified by the feeding status and marked as executable. This can be used to eliminate the risk of timing discrepancies between multiple systems and ensure the safe initiation of continued printing.

[0035] In step S500, the printing extrusion system is driven to perform breakpoint alignment and parameter adaptation continuous printing actions through the synchronized continuous printing matching instruction set, so as to obtain a high-precision 3D printing breakpoint continuous printing control scheme.

[0036] The printing extrusion system can be a material forming actuator including a nozzle, extruder, and temperature control unit. It can receive synchronized print reflow matching instruction sets and perform physical actions for breakpoint alignment and parameter adaptation. Driving the printing extrusion system to perform breakpoint alignment and parameter adaptation via the synchronized print reflow matching instruction set allows the system to parse coordinates and process parameters from the instruction set, automatically move the nozzle to the breakpoint, adjust extrusion parameters, and then resume printing. Furthermore, driving the printing extrusion system to perform breakpoint alignment and parameter adaptation via the synchronized print reflow matching instruction set enables precise unattended print reflow and ensures seamless fusion of the microstructure at the reflow point, thereby improving the continuity of long-cycle printing.

[0037] For example, in the scenario of UHPC 3D printing of wind turbine towers, the 3D printing chamber replenishment control method combining laser ranging and AI algorithms in this embodiment can be implemented as follows: During the continuous printing of wind turbine towers, the laser sensor continuously collects the material level in the chamber, while simultaneously recording the nozzle coordinates and printed path parameters; if the system detects that the material level drop rate is significantly lower than the theoretical material consumption rate, it determines that there is a risk of material supply lag, immediately constructs a high-priority layered response link, triggers an early warning, and adjusts the interruption signal transmission sequence; printing is automatically paused during the optimal window period before material depletion, and the AI ​​algorithm generates a set of resume printing instructions including flow compensation based on the breakpoint location and response weight; after replenishment is completed and the material supply status is stable, the system automatically synchronizes the instructions and drives the nozzle to accurately reset to the breakpoint, adjusts the extrusion parameters, and seamlessly resumes printing, without the need for manual intervention throughout the process, thus avoiding problems such as interlayer misalignment and insufficient bonding strength.

[0038] In one embodiment, the adjusted signal transmission priority is converted into a response weight parameter, and combined with the position parameter of the current breakpoint to generate a continuation matching instruction set, including: The hierarchical response links are divided into intervals based on the adjusted signal transmission priority level values. Within each interval, the level values ​​are converted to obtain the response weight parameters of the corresponding hierarchical response links.

[0039] The adjusted signal transmission priority level value can be a discrete or continuous value, dynamically adjusted to quantify the urgency of interruption signals and continuation parameters. This value can serve as the initial input for interval division and weight conversion, characterizing the current urgency level of supply risk. Dividing the hierarchical response link into intervals based on the adjusted signal transmission priority level value can be achieved by dividing the hierarchical response link into several functional sub-intervals, each corresponding to a specific response strategy. This allows for refined segmented management of the response link and differentiated processing logic for signals of different urgency levels. In this embodiment, converting the level value within an interval to obtain the corresponding hierarchical response link's response weight parameter can be achieved by applying a preset mapping function within each interval to convert the level value into a continuous response weight parameter. Furthermore, converting the level value within an interval to obtain the corresponding hierarchical response link's response weight parameter can be achieved by using linear interpolation for smooth conversion within the interval or by using a piecewise nonlinear function to achieve an adaptive mapping between urgency and weight. This transforms discrete priority levels into continuous weights that can participate in mathematical operations to support subsequent joint constraint processing.

[0040] Extract the printed layer thickness deviation parameter and the extruder cumulative displacement deviation parameter at the current breakpoint position, and combine the printed layer thickness deviation and the extruder displacement deviation into a breakpoint position error coefficient.

[0041] The printed layer thickness deviation parameter at the current breakpoint position can be the difference between the actual printed thickness of the layer at the breakpoint and the theoretically designed thickness. This parameter reflects the cumulative vertical error during single-layer forming and serves as a key component of the breakpoint position error coefficient. The extruder head cumulative displacement deviation parameter represents the cumulative deviation between the actual movement trajectory of the extruder head and the theoretical path from the printing start point to the current breakpoint position. This parameter characterizes the spatial positioning drift introduced by the mechanical system during long-cycle printing and its impact on the accuracy of subsequent printing alignment. For example, extracting the printed layer thickness deviation parameter and the extruder head cumulative displacement deviation parameter at the current breakpoint position can be achieved by retrieving the measured layer thickness value and the cumulative error value of the nozzle trajectory at the breakpoint moment from the full-state printing perception dataset. This allows for the acquisition of the original physical quantities constituting the breakpoint position error coefficient and the realization of a precise digital characterization of the breakpoint state.

[0042] The breakpoint position error coefficient can be a composite quantitative index that comprehensively characterizes the geometric forming deviation and cumulative motion error at the current breakpoint. It can be used to accurately describe the deviation of the actual physical state at the breakpoint and provide a targeted compensation benchmark for the continuous printing instruction generation template to solve the problem that a single coordinate cannot reflect the actual forming defects. In an exemplary embodiment, the breakpoint position error coefficient can be generated by merging the printed layer thickness deviation parameter and the extruder head cumulative displacement deviation parameter through weighted superposition or vector synthesis algorithm. Furthermore, the breakpoint position error coefficient can be coordinated with the response weight parameter and generate the weight parameter distribution result through joint constraint processing. The breakpoint position error coefficient can also be coordinated with the continuous printing instruction generation template and used as a core input variable to drive the calculation of compensation amount and correction conditions. In a specific embodiment, the breakpoint position error coefficient may include, but is not limited to, the axial height cumulative error coefficient, the radial path offset error coefficient, and the interlayer thickness fluctuation error coefficient. The superposition and merging of the printed layer thickness deviation and the extruder head displacement deviation into the breakpoint position error coefficient can be achieved by normalizing the two types of deviation parameters and then generating a composite error coefficient through weighted summation or vector synthesis algorithm. Furthermore, the error coefficient of the printed layer thickness deviation and the extruder displacement deviation can be superimposed and merged into the breakpoint position error coefficient by using a fixed weight coefficient for linear superposition or by dynamically allocating weights based on training an error fusion model based on historical printing data. This integrates multi-dimensional error information and forms a unified and comprehensive description of the breakpoint state to avoid the limitations of single-dimensional compensation.

[0043] The response weight parameters and breakpoint position error coefficients are jointly constrained to obtain the weight parameter distribution results. Based on the weight parameter distribution results, a continuation instruction generation template corresponding to each level in the hierarchical response link is constructed.

[0044] The weight parameter distribution result can be a multi-dimensional parameter mapping set formed by jointly constraining the response weight parameters and the breakpoint position error coefficient. This can be used to establish a dynamic correlation between control urgency and physical error, and guide the differentiated configuration of the continuation call generation template. The weight parameter distribution result obtained by jointly constraining the response weight parameters and the breakpoint position error coefficient can be achieved by trimming or scaling the breakpoint position error coefficient with the response weight parameters as the constraint boundary, generating a parameter distribution adapted to the current operating condition. This establishes a dynamic coupling between the control strategy and the physical state, ensuring that the compensation intensity meets accuracy requirements without exceeding the system safety margin. The continuation call generation template can be a parameterized instruction construction framework with pre-set response strategies and error compensation logic at different levels. It can be used to standardize and adaptively generate the continuation call matching instruction set, replacing manual experience adjustments to ensure that the optimal continuation call strategy can be output under different urgency levels and error states. In this embodiment, the continuation call generation template can dynamically configure its structure based on the weight parameter distribution result and internally encapsulates the corresponding level of compensation algorithms and constraint rules. Furthermore, the continuation call generation template can coordinate with the breakpoint position error coefficient, receive the error coefficient, and map it to specific compensation parameters. The continuation call generation template can also coordinate with the hierarchical response link, maintaining a one-to-one correspondence between its structure and the link hierarchy to ensure the consistency of the control strategy. For example, the continuation call generation template may include, but is not limited to, emergency stop reset templates, early warning speed adjustment transition templates, and regular material replenishment connection templates. Constructing continuation call generation templates corresponding to each level in the hierarchical response link based on the weight parameter distribution results can involve selecting or dynamically generating instruction template structures and parameter slots that match each response level, thereby enabling hierarchical adaptive configuration of the continuation call strategy and ensuring that structurally reasonable and parameter-adaptive instruction sets are generated under different risk levels.

[0045] Input the breakpoint position error coefficient into the continuation printing instruction generation template to generate a continuation printing matching instruction set that includes breakpoint alignment compensation amount and extrusion flow correction constraints.

[0046] The breakpoint alignment compensation amount can be a correction value used in the continuous firing matching instruction set to correct the offset of the nozzle reset coordinate. It can be used to eliminate the spatial deviation represented by the breakpoint position error coefficient and ensure that the continuous firing start point is precisely aligned with the original component. The extrusion flow rate correction constraint can be a process boundary parameter in the continuous firing matching instruction set that limits the extrusion rate, pressure, or time window. It can be used to compensate for changes in material rheological properties caused by downtime and to prevent over-extrusion, under-extrusion, or poor bonding at the continuous firing interface. Inputting the breakpoint position error coefficient into the continuous firing instruction generation template generates a continuous firing matching instruction set containing the breakpoint alignment compensation amount and extrusion flow rate correction constraint. This can be achieved by substituting the error coefficient into the compensation algorithm and constraint rules in the template and automatically calculating the alignment compensation amount and flow rate correction condition, which is then encapsulated into an instruction set. This enables the automated and precise generation of continuous firing parameters and eliminates human intervention errors to ensure seamless fusion of the microstructure of the continuous firing interface.

[0047] Taking UHPC continuous printing of bridge piers as an example, the 3D printing chamber feeding control method combining laser ranging and AI algorithm in this embodiment can be as follows: during the printing process of bridge piers, if the system detects a feeding delay, it triggers a medium-priority interruption signal, and the corresponding level value is classified into the warning speed adjustment response range and converted into a medium response weight parameter. At the same time, the 0.3mm layer thickness positive deviation and the 1.2mm radial cumulative offset of the extruder head at the current breakpoint are extracted and merged to generate a breakpoint position error coefficient. After the coefficient is constrained by the medium response weight, a warning-level continuing printing instruction generation template is constructed. The template automatically calculates the alignment compensation amount of -0.25mm and the correction constraint condition of reducing the extrusion flow rate by 8% and generates a continuing printing matching instruction set. After the feeding is completed, the system automatically executes the instruction to accurately compensate for the geometric deviation and adapt to the material rheological state, so that the interlayer bonding strength at the continuing printing point meets the design requirements and no manual adjustment is required throughout the process.

[0048] In one embodiment, the response weight parameters and the breakpoint position error coefficient are jointly constrained to obtain the weight parameter distribution result, including: The range of variation of the breakpoint position error coefficient is divided into continuous sub-intervals, and each sub-interval corresponds to a coordinate transformation function; The variation range of the breakpoint position error coefficient can be the complete numerical domain of the breakpoint position error coefficient that may take values ​​during the actual printing process, and can be used as the basis for defining the domain of sub-interval division. A continuous sub-interval can be a series of consecutive, non-overlapping numerical segments formed by dividing the variation range of the error coefficient, which can be used to achieve discretized segmented management of the error space. In this embodiment, the continuous sub-intervals work in conjunction with the coordinate transformation function, so that each segment can be matched with a dedicated coordinate transformation function. The coordinate transformation function can be a mathematical expression describing the mapping relationship between the error value and the alignment compensation amount within a specific sub-interval of the breakpoint position error coefficient, and can be used to achieve segmented and refined modeling of nonlinear errors. For example, the coordinate transformation function can be pre-calibrated or generated online based on the UHPC material rheological properties and mechanical system response model. Furthermore, the coordinate transformation function can work in conjunction with the initial alignment transformation factor to jointly calculate the associated sub-interval alignment parameters within the sub-interval. In an exemplary embodiment, the coordinate transformation function can include, but is not limited to, linear scaling transformation functions, polynomial fitting transformation functions, piecewise spline interpolation transformation functions, etc. Dividing the range of variation of the fault location error coefficient into continuous sub-intervals can be achieved by dividing the complete numerical domain into several interconnected sub-intervals based on error distribution characteristics or process experience. This allows for the technical effect of segmented and refined management of the nonlinear error space.

[0049] Based on the hierarchical position of the response weight parameter in the hierarchical response link, obtain the initial alignment conversion factor associated with the hierarchical position; The hierarchical position can be the logical hierarchical identifier of the response weight parameter in the hierarchical response link, and can be used as an index key to obtain the initial alignment conversion factor. In this embodiment, the hierarchical position and the initial alignment conversion factor work together, and their value is uniquely determined by the hierarchical level of the current response weight. The initial alignment conversion factor can be a basic alignment sensitivity coefficient bound to a specific hierarchical position in the hierarchical response link, and can be used to convert the urgency of the control layer into the compensation benchmark strength of the physical layer. For example, the initial alignment conversion factor can be retrieved from a preset parameter table or dynamically calculated based on the hierarchical position corresponding to the response weight parameter. Furthermore, the initial alignment conversion factor can work together with the coordinate transformation function, serving as an input parameter for the function operation and influencing the generation of the sub-interval alignment parameter. In a specific embodiment, the initial alignment conversion factor may include, but is not limited to, an emergency stop reset factor, a warning speed regulation transition factor, and a regular material replenishment connection factor. Obtaining the initial alignment conversion factor based on the hierarchical position corresponding to the response weight parameter in the hierarchical response link can be done by reading from a preset parameter library using the hierarchical position as an index or by calculating the corresponding initial alignment conversion factor using a formula. Furthermore, the initial alignment conversion factor can be obtained by directly mapping the level and factor value according to the hierarchical position of the response weight parameter in the hierarchical response link. Alternatively, the factor can be dynamically generated by substituting the level number into a preset function. This can establish a dynamic correlation between the control of urgency and the strength of the compensation benchmark.

[0050] Within the sub-interval, the error coefficient of the breakpoint position is correlated with the initial alignment transformation factor according to the coordinate transformation function to obtain the correlation sub-interval alignment parameters; The associated sub-interval alignment parameters can be the original compensation amount calculated within a specific sub-interval using a coordinate transformation function combined with an initial alignment transformation factor. These parameters can characterize the unsmoothed theoretical alignment correction value within the error interval. In this embodiment, the associated sub-interval alignment parameters and the fused alignment parameters work together, with the latter serving as input for boundary processing. Within the sub-interval, the error coefficient of the breakpoint position is correlated with the initial alignment transformation factor based on the coordinate transformation function. This can be achieved by substituting the error coefficient and the initial alignment transformation factor into the coordinate transformation function corresponding to the current sub-interval and calculating the original alignment parameters, thereby achieving the technical effect of a nonlinear and precise mapping from error to compensation amount.

[0051] The alignment parameters of the associated sub-intervals are fused by the interval boundaries to obtain the fused alignment parameters. The inter-level precision balance constraint is then applied to the fused alignment parameters to obtain the weight parameter distribution results.

[0052] The fused alignment parameters can be continuous, step-free alignment compensation intermediate quantities obtained after interval boundary smoothing, which can be used to eliminate parameter abrupt changes introduced by segmented processing. For example, the fused alignment parameters can be generated by performing weighted averaging, gradual transition, or continuity constraint algorithms on the associated alignment parameters of adjacent sub-intervals at the boundaries. Furthermore, the fused alignment parameters can be coordinated with inter-level precision balance constraints as the direct object of constraint operations. In an exemplary embodiment, the fused alignment parameters can include, but are not limited to, linear gradual fusion parameters, sigmoid smoothing fusion parameters, and adaptive weight fusion parameters. The weight parameter distribution result can be the final compensation parameter set output after inter-level precision balance constraints, which can be used as direct input for generating continuation instructions. The fused alignment parameters are obtained by fusing the associated sub-interval alignment parameters at interval boundaries, which can be achieved by performing a smooth transition algorithm on the original alignment parameters on both sides at the boundary of adjacent sub-intervals to eliminate step discontinuities. Furthermore, by fusing the alignment parameters of the associated sub-intervals at the interval boundaries, the resulting fused alignment parameters can be achieved through linear interpolation to gradually transition within the boundary neighborhood or by using higher-order spline functions to ensure the continuity of the first or second derivative. This ensures the continuous differentiability of the compensation parameters throughout the entire error domain. Applying inter-level precision balance constraints to the fused alignment parameters yields the weighted parameter distribution. This can be achieved by applying consistency checks and tolerance limits to the fused parameters corresponding to different levels and then adjusting the output to ensure the stability of the compensation accuracy during cross-level switching.

[0053] For example, in the scenario of UHPC 3D printing of high-rise building walls, the 3D printing chamber feeding control method combining laser ranging and AI algorithms in this embodiment can be as follows: During continuous printing of high-rise walls, the system detects abnormal feeding and triggers an early warning response. The corresponding layer position obtains a medium initial alignment transformation factor. The error coefficient of the current breakpoint position is 1.8mm, falling into the preset second sub-interval, which uses a quadratic polynomial coordinate transformation function. Substituting the error coefficient and the initial factor into the function, the original alignment parameter is calculated to be 2.1mm. Since this value is at the boundary between the first and second sub-intervals, the system uses a Sigmoid smoothing fusion algorithm to fuse it with the parameters of the adjacent intervals to 2.05mm. After inter-layer accuracy balance constraint verification, it is confirmed that the value is within the allowable tolerance of the early warning level, and the final output is the weight parameter distribution result. The continuing printing instruction generation template generates a precise alignment compensation amount based on this. The nozzle reset deviation is less than 0.1mm, the inter-layer bonding of the continuing printing is good, and no manual calibration is required throughout the process.

[0054] In one embodiment, within a sub-interval, the breakpoint position error coefficient and the initial alignment transformation factor are correlated according to the coordinate transformation function to obtain the correlated sub-interval alignment parameters, including: The slope adjustment factor of the coordinate transformation function is determined based on the span range of the sub-interval, and the breakpoint position error coefficient is scaled according to the slope adjustment factor to generate the breakpoint deviation value. The span of the sub-interval can be the numerical width covered by the current sub-interval within the range of the error coefficient change at the breakpoint position, and can be used to characterize the scale characteristics of error change within the interval. In this embodiment, the span of the sub-interval can be used as the basis for determining the slope adjustment factor. The slope adjustment factor can be a proportional coefficient determined based on the span of the sub-interval to adjust the sensitivity of the coordinate transformation function. It can be used to achieve adaptive adjustment of compensation sensitivity under different error magnitudes and resolve the contradiction that fixed proportional transformation cannot simultaneously take into account the fine correction of small errors and the rapid convergence of large errors. For example, the slope adjustment factor can be calculated in reverse or obtained by looking up a table based on the numerical width of the sub-interval, and the wider the interval, the smaller the factor to achieve normalized scaling. Furthermore, the slope adjustment factor can be coordinated with the breakpoint position error coefficient and generated through scaling to produce the breakpoint deviation value, or it can be coordinated with the span of the sub-interval, and its value is uniquely determined by the interval width. In an exemplary embodiment, the slope adjustment factor can include, but is not limited to, a linear inverse proportional adjustment factor, a logarithmic compression adjustment factor, and a piecewise constant adjustment factor. The breakpoint deviation value can be a normalized error representation obtained by scaling the breakpoint position error coefficient with a slope adjustment factor. It can be used to eliminate the differences in the original error dimensions and magnitudes and to provide standardized input for subsequent cross-correlation and superposition operations.

[0055] The slope adjustment factor of the coordinate transformation function is determined based on the span of the sub-interval. This can be achieved by calculating the corresponding slope adjustment factor based on the numerical width of the current sub-interval using preset rules or functions. Furthermore, determining the slope adjustment factor of the coordinate transformation function based on the span of the sub-interval allows for adaptive changes in compensation sensitivity with the error scale, thus avoiding oversensitivity in large intervals or sluggish response in small intervals. The breakpoint position error coefficient is scaled according to the slope adjustment factor to generate the breakpoint deviation value. This can be achieved by multiplying the original error coefficient by the slope adjustment factor to obtain a normalized deviation value. In a specific embodiment, scaling the breakpoint position error coefficient according to the slope adjustment factor to generate the breakpoint deviation value can be achieved through linear scaling using direct multiplication or compression scaling using a nonlinear mapping function. This unifies the error expression scale within different sub-intervals and provides a consistent basis for subsequent cross-interval calculations.

[0056] Within the sub-interval, the alignment accuracy adjustment factor is obtained according to the hierarchical position corresponding to the response weight parameter, and the breakpoint deviation value is cross-correlated with the alignment accuracy adjustment factor to obtain the initial alignment conversion factor. The alignment accuracy adjustment factor can be a correction coefficient bound to the hierarchical position of the response weight parameter, used to regulate the accuracy of alignment calculation. It can be used to establish a dynamic coupling between the control urgency and physical compensation accuracy, ensuring automatic improvement of alignment response accuracy at high urgency levels to avoid sacrificing continuous firing quality due to urgent material supply. For example, the alignment accuracy adjustment factor can be retrieved from a preset accuracy configuration table or dynamically calculated based on the hierarchical position of the current response weight. Furthermore, the alignment accuracy adjustment factor can be coordinated with the breakpoint deviation value and generated through cross-correlation to form an initial alignment conversion factor, or it can be coordinated with the hierarchical position and its value can be adaptively adjusted according to the response priority. In this embodiment, the alignment accuracy adjustment factor can include, but is not limited to, a high-precision fine-tuning factor, a standard precision balancing factor, and a fast-convergence coarse-tuning factor. The initial alignment conversion factor can be a basic compensation strength parameter generated by cross-correlation between the breakpoint deviation value and the alignment accuracy adjustment factor. It can be used to integrate error magnitude and control accuracy requirements to form an intermediate compensation amount that combines adaptive sensitivity and hierarchical adaptability.

[0057] The alignment accuracy adjustment factor is obtained based on the hierarchical position corresponding to the response weight parameter. This can be achieved by reading the alignment accuracy adjustment factor from the accuracy configuration library using the hierarchical position as an index or by calculating the alignment accuracy adjustment factor in real time. Furthermore, obtaining the alignment accuracy adjustment factor based on the hierarchical position corresponding to the response weight parameter allows the compensation accuracy requirement to be dynamically matched with the control urgency, thereby preventing accuracy degradation under high priority. The initial alignment conversion factor is obtained by cross-correlating the breakpoint deviation value with the alignment accuracy adjustment factor. This can be achieved by performing a product, weighted, or function combination operation on the normalized deviation value and the accuracy adjustment factor to generate the initial alignment conversion factor. In an exemplary embodiment, the initial alignment conversion factor is obtained by cross-correlating the breakpoint deviation value with the alignment accuracy adjustment factor by directly coupling the two using a product method or by using a lookup table to achieve a nonlinear cross-mapping, thereby integrating the error amplitude and accuracy requirements and generating basic compensation parameters that are both adaptive and hierarchically adaptable.

[0058] The initial alignment conversion factor is subjected to interval accumulation processing. The accumulation processing result is superimposed with the breakpoint deviation value to generate the associated sub-interval alignment parameters.

[0059] The cumulative processing result can be a historical state memory obtained by integrating or summing the initial alignment conversion factor within a sub-interval. This memory can be used to introduce a smoothing effect in the time dimension and suppress parameter jumps caused by sensor noise or instantaneous disturbances to ensure the continuity of the compensation trajectory. The associated sub-interval alignment parameters can be the final sub-interval compensation output generated by superimposing the cumulative processing result and the breakpoint deviation value. This output can be used to integrate the instantaneous error response and historical state memory to provide stable and accurate local alignment correction values. Performing interval cumulative processing on the initial alignment conversion factor can be done by integrating or sliding the initial alignment conversion factor in the time or spatial dimension within the current sub-interval. Furthermore, performing interval cumulative processing on the initial alignment conversion factor can introduce historical state memory and smooth instantaneous disturbances, thereby enhancing the temporal stability of the compensation parameters. Superimposing the cumulative processing result and the breakpoint deviation value to generate the associated sub-interval alignment parameters can be done by adding or weightedly fusing the cumulative processing result with the current breakpoint deviation value to output the final sub-interval alignment parameters. In this embodiment, the cumulative processing result and the breakpoint deviation value are superimposed to generate the associated sub-interval alignment parameter, which can be achieved by balancing the immediate response and historical trend, thereby ensuring that the compensation amount reflects the current error while maintaining the continuity of the trajectory.

[0060] For example, in the scenario of UHPC 3D printing of offshore wind power foundations, the 3D printing chamber feeding control method combining laser ranging and AI algorithms in this embodiment can be as follows: during continuous printing of offshore wind power foundations, the system detects feeding fluctuations, triggers an early warning response, and obtains the standard accuracy adjustment factor at the corresponding layer position. The current breakpoint position error coefficient is 2.5mm and falls within a sub-interval with a span of 3mm, and the slope adjustment factor is determined to be 0.33. The error coefficient is scaled to obtain a breakpoint deviation value of 0.825, which is then cross-correlated with the standard accuracy adjustment factor of 1.2 to obtain an initial alignment conversion factor of 0.99. The factor is then subjected to sliding window accumulation processing to obtain an accumulated result of 0.15. Finally, it is superimposed with the breakpoint deviation value to generate the associated sub-interval alignment parameter of 0.975. After boundary fusion and accuracy balancing, this parameter is used to generate the continuous printing command, and the nozzle reset is accurate and jitter-free, ensuring good interlayer bonding during continuous printing and adapting to the marine vibration environment without manual intervention.

[0061] In one embodiment, real-time material level data from laser ranging in the 3D printing feed hopper, real-time coordinate parameters of the printing nozzle, and path layering parameters of the printed components are collected. The multi-source collected material level and printing status data are then denoised and fused to obtain a full-state printing perception dataset, including: The scanning coverage area of ​​the laser ranging sensor is divided into blind spot supplementation areas, and the laser ranging material level data is segmented into a discrete point cloud dataset containing spatial identifiers.

[0062] The laser rangefinder sensor can be a non-contact measurement device that emits a laser beam and receives the echo to obtain target distance information. It can be used as the physical front end of UHPC level sensing and overcome the interference of highly viscous materials on traditional sensors. The scanning coverage area can be the maximum spatial volume that the laser rangefinder sensor can reach in a single or periodic scan. It can be used to define the effective data acquisition boundary and serve as a prerequisite for blind spot identification and supplementary scanning planning. The blind spot supplementary scanning area can be a supplementary scanning space range defined for the sensing gaps caused by the inherent field-of-view limitations of the laser rangefinder sensor or the obstruction of the silo structure. It can be used to eliminate sensing blind spots in complex silos and ensure the spatial integrity and continuity of level data acquisition. In this embodiment, the blind spot supplementary scanning area can be identified by ray projection simulation based on the 3D model of the feeding silo and the sensor installation pose to generate corresponding supplementary scanning path planning. In an exemplary embodiment, the blind spot supplementary scanning area can be used in conjunction with the discrete point cloud dataset as a spatial index for data segmentation and can be used in conjunction with static feeding silo structural feature points for calibration and removal of fixed background interference. Furthermore, the blind spot compensation area may include, but is not limited to, structural occlusion blind spots, specular reflection failure areas, and edge attenuation compensation areas.

[0063] Spatial identifiers can be unique spatial location codes or region labels assigned to discrete point cloud data points, supporting rapid data point retrieval, region association, and cross-frame matching. A discrete point cloud dataset can be a set of three-dimensional coordinate points with location labels, segmented from continuous laser ranging signals according to spatial identifiers. It can be used to achieve structured representation of material level data and support subsequent refined redundancy removal and feature separation processing based on spatial location. For example, a discrete point cloud dataset can spatially mesh the original laser echo data and assign each data point a unique region code and timestamp. In a specific embodiment, the discrete point cloud dataset can collaboratively provide the data foundation for overlapping region matching with adjacent scan frames, and can collaboratively carry the spatial attribution information of supplementary scan data with blind zone supplementary scan areas. Furthermore, the discrete point cloud dataset can include, but is not limited to, high-density core area point clouds, low-density edge area point clouds, and transitional connection area point clouds.

[0064] To divide the scanning coverage area of ​​a laser rangefinder sensor into blind spot supplementation regions, a supplementation strategy can be generated based on the silo structure and sensor pose model to identify areas with missing perception. Furthermore, this blind spot supplementation can be achieved using ray tracing simulation algorithms, ensuring no blind spots in level sensing and improving data spatial integrity. Segmenting the laser rangefinder level data into a discrete point cloud dataset with spatial identifiers can be achieved by spatially meshing the original rangefinder data and adding region encoding and time labels. Further, this segmentation can be achieved through adaptive meshing using an octree spatial segmentation algorithm or by uniformly slicing the data using a pre-defined regular grid and labeling the regions with IDs, thus achieving data structuring and supporting subsequent fine-grained processing.

[0065] Within the discrete point cloud dataset, duplicate feature matching is performed on the overlapping area data of adjacent scan frames to delete redundant point cloud data points that exceed the set duplicate threshold, thereby generating a deredundant material level dataset.

[0066] Adjacent scan frames can be two or more sets of point cloud data that are temporally continuous and spatially overlapping. These can be used to provide temporal redundancy information and for duplicate feature matching and noise suppression. Overlapping region data can be a subset of point clouds corresponding to the spatially overlapping parts of adjacent scan frames. This can be used as the computational region for duplicate feature matching and to ensure the accuracy of redundancy removal operations. Duplicate features can be redundant data generated by scanning the same physical point multiple times within the overlapping region. These can be used to identify redundant information to be removed and improve the signal-to-noise ratio of the dataset. Setting a duplicate threshold can be a quantitative criterion for determining whether a data point is redundant. This can be used to control the sensitivity of redundancy removal operations and balance data fidelity and computational efficiency. Redundant point cloud data points can be invalid or repeatedly collected data points that exceed the duplicate threshold after matching. These can be used as objects to be deleted to clean up the dataset and improve the accuracy of subsequent processing. The deredundant material location dataset can be the set of valid material location point clouds retained after removing redundant point cloud data points. This can be used to provide a clean data foundation and support the accurate extraction of dynamic and static features. Performing duplicate feature matching on overlapping regions of adjacent scan frames can be achieved by comparing the spatial location and attribute similarity of data points from adjacent frames within the overlapping region. Furthermore, this duplicate feature matching can be implemented by statistically analyzing the overlap after point cloud registration based on the ICP algorithm, or by using local descriptors for feature-level similarity measurement. This helps identify redundant information and provides a basis for redundancy removal. Deleting redundant point cloud data points exceeding a set duplication threshold can be achieved by comparing the matching results with the threshold and removing data points deemed redundant. Further, deleting redundant point cloud data points exceeding the set duplication threshold can be achieved using an adaptive threshold adjustment algorithm, thereby improving the signal-to-noise ratio of the dataset and reducing subsequent computational load.

[0067] Extract dynamic material consumption feature points and static material supply silo structure feature points related to material level changes from the redundant material level dataset. Separate the dynamic material consumption feature points from the static material supply silo structure feature points to generate dynamic feature datasets and static feature datasets.

[0068] The dynamic material consumption feature points can be data points characterizing the geometric or intensity features of the material surface that change in real time during the printing process. These points can accurately reflect the actual material consumption behavior and provide clean dynamic input for breakpoint prediction and material consumption rate calculation. In this embodiment, dynamic material consumption feature points can be selected from the redundant material level dataset using time-series difference analysis or curvature change detection to identify points with significant time-varying attributes. In an exemplary embodiment, dynamic material consumption feature points can be used in conjunction with the extended breakpoint coordinate sequence as the starting reference for extended prediction, and can be used in conjunction with static material supply silo structural feature points through a separation operation to achieve dynamic-static decoupling. Furthermore, dynamic material consumption feature points can include, but are not limited to, points indicating a downward trend in liquid level, points indicating local collapse anomalies, and points responding to flow rate fluctuations. Static material supply silo structural feature points can be data points characterizing the geometry of fixed components such as the inner wall of the supply silo and the stirring blades, and can remain unchanged over time. These points can be used to construct a stable spatial reference system and to correct dynamic data while preventing fixed structures from being misjudged as material level changes. For example, static material supply silo structural feature points can be extracted using multi-frame cumulative averaging or background modeling methods to extract long-term stable spatial structural features. In one specific embodiment, the static material supply silo structural feature points can work in conjunction with the dynamic material consumption feature points to improve the purity of dynamic data through a separation operation, and can also work in conjunction with the extended breakpoint coordinate sequence to provide an absolute coordinate framework for spatial splicing. Furthermore, the static material supply silo structural feature points may include, but are not limited to, silo wall contour feature points, internal component feature points, and installation reference feature points.

[0069] Dynamic feature datasets can be structured datasets containing only dynamic material consumption feature points, which can be used to focus on material consumption behavior and eliminate static background interference. Static feature datasets can be structured datasets containing only static silo structure feature points, which can be used to provide a stable spatial reference frame and assist in dynamic data correction and stitching. Extracting dynamic material consumption feature points and static silo structure feature points related to material level changes from the deredundant material level dataset can be achieved by applying time series analysis or background modeling methods to distinguish between time-varying and constant features. Furthermore, extracting dynamic material consumption feature points and static silo structure feature points related to material level changes from the deredundant material level dataset can be achieved by separating high-frequency dynamic components from low-frequency static components using principal component analysis or by using sliding window variance detection to extract significant change points, thereby achieving decoupling of dynamic and static features and improving the purity of each data. Separating dynamic material consumption feature points from static silo structure feature points can be achieved by assigning data points to independent datasets based on feature classification results. Furthermore, separating dynamic material consumption feature points from static silo structure feature points can be achieved through multidimensional feature space clustering algorithms, thereby constructing dedicated data channels and avoiding cross-interference.

[0070] The dynamic feature dataset is extended and predicted according to the material consumption rate of the current printing path to obtain the extended breakpoint coordinate sequence. The extended breakpoint coordinate sequence is then spatially concatenated with the static feature dataset to generate a full-state printing awareness dataset.

[0071] The current printing path can be the printhead's motion trajectory planning data in the ongoing printing task, which can be used to provide a theoretical material consumption benchmark and support the prediction of breakpoint location extension. The material consumption rate can be the theoretical or measured volume of material consumed per unit time, which can be used as the slope parameter for extension prediction and determine the trend of the extended breakpoint coordinate sequence. The extended breakpoint coordinate sequence can be a predicted coordinate trajectory generated by extrapolating the breakpoint location at several future times based on the current material consumption trend, which can be used to predict the spatial location of the material supply gap in advance and provide forward-looking data support for proactive early warning and pre-generation of continuous printing parameters. In this embodiment, the extended breakpoint coordinate sequence can be calculated linearly or nonlinearly using a dynamic feature dataset as input and combined with the material consumption rate of the current printing path. In an exemplary embodiment, the extended breakpoint coordinate sequence can inherit the dynamic change trend in conjunction with dynamic material consumption feature points, and can be integrated into the global coordinate system through spatial splicing in conjunction with a static feature dataset. Furthermore, the extended breakpoint coordinate sequence can include, but is not limited to, short-term real-time prediction sequences, medium-term trend prediction sequences, and long-term boundary prediction sequences.

[0072] Extending and predicting breakpoint coordinates from the dynamic feature dataset based on the current material consumption rate of the printing path can be achieved by extrapolating the dynamic features in conjunction with the material consumption rate. Furthermore, this extension can be implemented using a linear regression model for short-term trend extrapolation or by employing an LSTM network to learn historical material consumption patterns for nonlinear prediction, thus generating forward-looking breakpoint coordinates and supporting proactive early warning. Spatially concatenating the extended breakpoint coordinate sequence with the static feature dataset can be done by fusing the prediction sequence and static structural data in a unified coordinate system. Further, this spatial concatenation can be achieved using a rigid body transformation matrix alignment algorithm, thereby constructing a high-fidelity sensing dataset that combines real-time status with future trends.

[0073] For example, in the scenario of continuous UHPC printing of bridge piers, the 3D printing chamber replenishment control method combining laser ranging and AI algorithms in this embodiment can be as follows: During the 3D printing process of bridge piers, when the laser sensor scans the feeding chamber, it finds a blind spot below the mixing paddle at the bottom of the chamber. The system automatically divides this area into a blind spot replenishment area and adjusts the scanning angle to collect data. After segmenting the full chamber point cloud according to spatial identifiers, ICP registration and matching are performed on the overlapping areas of two consecutive frames to remove redundant points caused by dust reflection. The liquid level decline trend points are extracted from the purified data as dynamic material consumption feature points, and the chamber wall contour is separated as static feature points. Based on the material consumption rate of the current printing path, the breakpoint position extension prediction of the dynamic feature points in the next 5 seconds is performed to generate an extended breakpoint coordinate sequence. Finally, this sequence is spliced ​​with the static chamber features in a unified coordinate system to form a complete perception dataset, enabling the system to predict the possible material supply gap 30 seconds later and actively trigger the replenishment process to avoid printing interruption.

[0074] In one embodiment, the dynamic feature dataset is extended and predicted according to the material consumption rate of the current printing path to obtain an extended breakpoint coordinate sequence, including: Extract the measured coordinate values ​​of the printhead over a continuous time period from the dynamic feature dataset, and calculate the trend vector of printhead movement based on the changes in the measured coordinate values ​​of the printhead.

[0075] The continuous time period can be a continuous and equally spaced time window used to extract the printhead motion trend. It provides sufficient time span to smooth out instantaneous jitter and accurately reflect the macroscopic motion trend of the printhead. The measured coordinate values ​​of the printhead can be the actual spatial position of the printhead at a specific moment, collected by an encoder or position sensor. This can be used as the original data source for trend vector calculation and reflects the actual motion state rather than the command-planned state. In this embodiment, extracting the continuous time period of the measured printhead coordinate values ​​from the dynamic feature dataset can be achieved by reading printhead position records from the dynamic feature dataset at fixed time intervals, thereby obtaining time-series samples reflecting the actual motion state and supporting trend analysis. The change can be the displacement difference between the measured coordinate values ​​of the printhead at adjacent time points. This can be used to quantify the instantaneous speed and direction of the printhead motion and serve as a basic intermediate variable for deriving the trend vector.

[0076] A trend vector can be a directed geometric quantity characterizing the direction and magnitude of the printhead's movement within a continuous time period. It can be used to digitally describe the dynamic printing state before shutdown and provide kinematic basis for predicting the reverse backtracking of breakpoint locations. For example, the trend vector can be obtained by performing differential operations or linear fitting on the measured coordinate values ​​of the printhead within a continuous time period to extract the displacement change rate and direction components. Furthermore, the trend vector can be used in conjunction with the predicted breakpoint coordinate sequence as a directional reference for reverse backtracking, and can be used in conjunction with a preset alignment compensation step size to jointly determine the range of the coordinate prediction interval. In an exemplary embodiment, the trend vector can include, but is not limited to, a two-dimensional planar trend vector, a three-dimensional spatial trend vector, a curvature correction trend vector, etc. Calculating the printhead movement trend vector based on the change in the measured coordinate values ​​of the printhead can be achieved by performing differential or fitting operations on the extracted coordinate sequence to solve for the displacement direction and rate. Furthermore, the trend vector of printhead movement can be calculated based on the change in the measured coordinate values ​​of the printhead. This can be achieved by using the least squares method to perform linear regression on the coordinates of multiple points to obtain the average trend, or by using the sliding window difference method to calculate the instantaneous trend and then weighted average. This transforms discrete coordinates into a continuous motion representation and overcomes the limitations of single-point static description.

[0077] Using the measured coordinates of the printhead at the moment of shutdown as the reference point, the coordinate prediction interval of the breakpoint position is obtained by backtracking along the opposite direction of the trend vector and based on the preset alignment compensation step size, and a sequence of predicted breakpoint coordinate points is generated.

[0078] The shutdown moment can be the instantaneous point in time when the material supply is interrupted or the system triggers a pause command. It can be used as the zero point of time and the spatial reference anchor point for reverse backtracking prediction. The reference point can be the spatial position determined by the measured coordinates of the printhead at the shutdown moment, and can be used as the starting reference origin of the predicted interval for the breakpoint location coordinates. In this embodiment, the reverse direction can be a geometric direction completely opposite to the trend vector, which can be used to indicate the correct path for the breakpoint backtracking search and conform to the physical law that material extrusion lags behind printhead movement. The preset alignment compensation step size can be a pre-set spatial backtracking unit length used to generate the predicted breakpoint coordinate point sequence. It can be used to control the resolution and coverage of the prediction interval and balance positioning accuracy and computational efficiency. The predicted interval for the breakpoint location coordinates can be a spatial range that may contain the actual breakpoint, defined by the predicted breakpoint coordinate point sequence. It can be used to transform single-point positioning into interval search and enhance the system's robustness to mechanical backlash and response delay. For example, using the measured coordinates of the printhead at the moment of shutdown as a reference point, and backtracking along the opposite direction of the trend vector based on a preset alignment compensation step size, the coordinate prediction interval for the breakpoint position can be obtained by iteratively generating a series of backtracking coordinate points from the reference point in the opposite direction with a fixed step size. Furthermore, this step can be achieved by generating a uniform sequence through constant step size equidistant backtracking, or by dynamically adjusting the step size according to the trend vector magnitude, increasing the step size in high-speed segments and decreasing the step size in low-speed segments. This allows for the construction of a candidate region covering the actual breakpoint and effectively compensates for positioning deviations caused by system response delays.

[0079] The predicted breakpoint coordinate sequence can be an ordered set of coordinates containing multiple candidate breakpoint locations, generated based on the reverse backtracking of the trend vector. This sequence can be used to construct a search space covering the theoretical breakpoint region and to offset overshoot errors caused by system response delays, thereby improving the fault tolerance rate of breakpoint localization. In one specific embodiment, the predicted breakpoint coordinate sequence can be obtained by iteratively calculating a series of discrete coordinate points along the reverse direction of the trend vector, starting from the stop time coordinates, at a preset alignment compensation step size. Furthermore, the predicted breakpoint coordinate sequence can be used as input for verification operations in conjunction with prediction confidence verification processing, and can be used in conjunction with dynamic feature datasets to achieve virtual-real data fusion through timestamp alignment. The predicted breakpoint coordinate sequence can include, but is not limited to, equal-step backtracking sequences, variable-step adaptive sequences, and weighted probability distribution sequences. Generating the predicted breakpoint coordinate sequence can be achieved by arranging all the coordinate points obtained from the backtracking calculation into a sequence in temporal or spatial order, thereby forming a structured breakpoint search space and facilitating subsequent verification and fusion processing.

[0080] The predicted breakpoint coordinate sequence is subjected to prediction confidence verification processing. The verified predicted breakpoint coordinate sequence is then merged with the dynamic feature dataset after timestamp alignment to obtain the extended breakpoint coordinate sequence.

[0081] The prediction credibility verification process can be a screening mechanism for the reasonableness verification and anomaly removal of the predicted breakpoint coordinate sequence. This can be used to ensure the accuracy and reliability of the extended breakpoint coordinate sequence and prevent erroneous predictions from causing misalignment or collision risks. In this embodiment, the prediction credibility verification process can be obtained by combining path constraints, physical boundaries, or historical data consistency to perform multi-dimensional evaluation of the predicted points to filter out invalid coordinates that deviate from the actual working conditions. Furthermore, the prediction credibility verification process can work synergistically with the predicted breakpoint coordinate sequence to improve its data quality, and can work synergistically with timestamp alignment and merging to ensure the spatiotemporal validity of the fused data. For example, the prediction credibility verification process can include, but is not limited to, geometric constraint verification, temporal consistency verification, and statistical outlier verification.

[0082] Performing prediction confidence verification on the predicted breakpoint coordinate sequence can involve applying preset rules or models to screen the reasonableness of each point in the sequence. Furthermore, this verification can be achieved through geometric boundary constraint verification based on a printed path CAD model or statistical distribution consistency checks using a historical breakpoint database. This process eliminates abnormal prediction points and ensures the engineering usability of the breakpoint location results. Timestamps can be unified time labels marking the time of data point acquisition or generation, enabling precise temporal alignment between predicted coordinates and dynamically sensed data and ensuring the synchronization of multi-source data fusion. Timestamp alignment and merging of the verified predicted breakpoint coordinate sequence with the dynamic feature dataset can be achieved by inserting verified prediction points into their corresponding temporal positions in the dynamic feature dataset using the timestamp as the key. This allows for spatiotemporal unification of virtual predicted data and real sensed data, and the construction of a high-fidelity extended breakpoint coordinate sequence.

[0083] Taking the UHPC printing of high-rise building walls as an example, the 3D printing chamber feeding control method combining laser ranging and AI algorithms in this embodiment can be used to prevent the machine from stopping due to material supply fluctuations during the printing process of high-rise walls. The system immediately extracts the measured coordinates of the nozzle every 10 milliseconds within 2 seconds before the stop from the dynamic feature dataset, and calculates the trend vector of the nozzle moving along the positive X-axis at a speed of 50mm / s. Using the coordinates at the stop time as the reference point, the system backtracks along the negative X-axis in 2mm steps to generate a sequence of predicted breakpoint coordinates containing 25 candidate points. The system performs geometric constraint verification on the sequence and removes 3 abnormal points that exceed the wall design boundary. The remaining 22 valid predicted points are inserted into the dynamic feature dataset according to the timestamp to form an extended breakpoint coordinate sequence. Based on this, the system automatically identifies the actual breakpoint located at the 8th backtracking point, drives the nozzle to accurately reset and adjusts the extrusion parameters before seamlessly resuming printing. No manual intervention is required throughout the process, and the interlayer bonding surface is flat and without misalignment.

[0084] In one embodiment, the printing extrusion system is driven by a synchronized continue-printing matching instruction set to perform continue-printing actions that align breakpoints and adapt parameters, thereby obtaining a high-precision 3D printing breakpoint continue-printing control scheme, including: The synchronized continuous printing matching instruction set is broken down into the continuous printing action sequence corresponding to the printing extrusion system. The continuous printing action sequence includes the breakpoint target coordinates, alignment compensation amount, and extrusion parameter correction labels.

[0085] The continuous printing action sequence can be an ordered set of atomic operations executed by the printing extrusion system after parsing the synchronized continuous printing matching instruction set. It can be used to achieve fine-grained mapping of high-level control instructions to low-level execution actions and support step-by-step precise control of breakpoint alignment and parameter adaptation. In this embodiment, the continuous printing action sequence can be obtained by semantically decomposing and protocol-converting the continuous printing matching instruction set and extracting structured execution units containing breakpoint target coordinates, alignment compensation amounts, and extrusion parameter correction labels. For example, the continuous printing action sequence can be coordinated with the resource occupancy ratio calculated by AI, with the extrusion parameter correction labels serving as the basis for resource allocation. Simultaneously, the continuous printing action sequence can be coordinated with the error buffer, with the breakpoint target coordinates serving as the benchmark for search range adjustment. In an exemplary embodiment, the continuous printing action sequence may include, but is not limited to, spatial positioning action subsequences, process parameter loading subsequences, and status verification feedback subsequences.

[0086] The breakpoint target coordinates can be a three-dimensional spatial point in space that represents the theoretical printing interruption position specified in the re-printing action sequence. This point can serve as an anchor point between the center reference of the coordinate search range and the error buffer, guiding the target position for the physical repositioning of the printhead. The alignment compensation amount can be a preset spatial correction vector for mechanical transmission errors or interlayer offsets. It can be used to superimpose fine-tuning amounts on the breakpoint target coordinates to improve the overlap between the actual printhead position and the theoretical breakpoint. The extrusion parameter correction label can be metadata tags that identify the required flow rate, pressure, or temperature adjustment range for re-printing. This can be used to characterize the complexity of the re-printing process and serve as a basis for allocating AI computing resources. Decomposing the synchronized re-printing matching instruction set into the corresponding re-printing action sequence of the printing extrusion system can involve parsing and semantically segmenting the instruction set to extract atomic operation units such as breakpoint target coordinates, alignment compensation amounts, and extrusion parameter correction labels. Furthermore, this decomposition process decouples and structures the control instructions, providing operable input elements for subsequent resource allocation and search range determination.

[0087] Based on the extrusion parameter correction label, the AI-calculated resource occupancy ratio is allocated to the printing control system, and the coordinate search range of the breakpoint alignment search is determined according to the resource occupancy ratio.

[0088] The resource occupancy ratio for AI computing can be a dynamically allocated share of processor computing power to support the breakpoint alignment algorithm based on extrusion parameter correction tags. This can be used to establish a dynamic adaptation mechanism between computing resources and the criticality of the re-printing task, ensuring that the high-precision alignment algorithm receives priority computing power in resource-constrained environments. In this embodiment, the resource occupancy ratio for AI computing can be obtained by parsing the process complexity represented by the extrusion parameter correction tags and determining the CPU or GPU time slice or memory quota of the AI ​​module in the printing control system through a preset mapping strategy. For example, the resource occupancy ratio for AI computing can be coordinated with the re-printing action sequence and trigger resource reallocation based on the correction tags therein. Simultaneously, the resource occupancy ratio for AI computing can be coordinated with the coordinate search range, and the resource ratio directly determines the initial size and boundary elasticity of the search range. In a specific embodiment, the resource occupancy ratio for AI computing may include, but is not limited to, time slice rotation ratio, memory bandwidth reservation ratio, and dedicated accelerator core utilization rate.

[0089] The coordinate search range can be a spatial search area defined around the target coordinates of the breakpoint for locating the actual breakpoint. This can be used to limit the search space of the AI ​​algorithm and balance alignment accuracy and computational efficiency to avoid blind global searching. Allocating the AI ​​computing resource ratio to the printing control system based on the extrusion parameter correction label can be achieved by reading the process complexity identifier in the correction label and looking up a table or by calculating the computing power quota that the AI ​​module should receive through a function. Furthermore, this allocation process can be implemented by using a priority queue scheduler to dynamically adjust the time slice weight of the AI ​​thread according to the label level, or by using a hardware-level quality of service mechanism to reserve a fixed proportion of memory bandwidth and cache resources for the AI ​​computing unit. This ensures the supply of computing resources for critical printing cycles and avoids delays or failures in the alignment algorithm due to system load fluctuations. Determining the coordinate search range for breakpoint alignment based on the resource occupancy ratio can be achieved by using the resource occupancy ratio as an input variable and calculating the initial size and resolution of the coordinate search range through a preset mapping relationship. Furthermore, this determination process allows the search strategy to adaptively match the available computing power, thereby expanding the search range to improve robustness when resources are sufficient and shrinking the range to ensure real-time performance when resources are scarce.

[0090] The boundary conditions of the coordinate search range are dynamically adjusted based on the target coordinates of the breakpoint and the resource occupancy ratio, and the error buffer of the nozzle positioning is calculated based on the adjusted coordinate search range.

[0091] The boundary conditions can be geometric constraint parameters of the coordinate search range, including radius, shape, and elasticity coefficients. These can be used to dynamically scale the search area based on resource occupancy ratio and breakpoint target coordinates to adapt to positioning requirements under different working conditions. The error buffer can be a tolerance verification space constructed based on the adjusted coordinate search range, accommodating mechanical and material nonlinear errors. It can be used to absorb positioning deviations caused by uncertainties such as transmission gaps and material rheology, and to provide a robust matching verification window for breakpoint continuation printing. In this embodiment, the error buffer can be centered on the breakpoint target coordinates, combined with the search boundary determined by the resource occupancy ratio, and superimposed with the alignment deviation between the feeding state and the predicted coordinates to generate a multi-dimensional tolerance domain. For example, the error buffer can be coordinated with the predicted breakpoint coordinate parameters and the current feeding state of 3D printing, and after alignment, used as input to verify the validity of the interval. Simultaneously, the error buffer can be coordinated with the 3D printing chamber replenishment breakpoint continuation printing matching control scheme, and this interval is the constraint boundary for generating the final control scheme. In an exemplary embodiment, the error buffer may include, but is not limited to, a mechanical return error buffer, a material extrusion hysteresis buffer, and a thermal deformation compensation buffer.

[0092] The boundary conditions of the coordinate search range are dynamically adjusted based on the breakpoint target coordinates and resource occupancy ratio. This can be achieved by using the breakpoint target coordinates as the geometric center and combining them with the resource occupancy ratio to real-time correct the radius, shape, or flexible boundary of the search range. Furthermore, when the resource ratio is below a threshold, this adjustment process can compress the circular search range into an elliptical domain along the printing path to reduce computation, or dynamically set asymmetric boundaries based on the geometric characteristics of the target coordinate area to avoid known interference zones. This optimizes the geometry of the search space and minimizes invalid search areas while ensuring breakpoint coverage, thereby improving alignment efficiency and accuracy. The error buffer space for nozzle positioning is calculated based on the adjusted coordinate search range. This can be achieved by overlaying mechanical tolerances, material rheological models, and historical error statistics within the adjusted search range to generate a multi-dimensional tolerance verification space. Furthermore, this calculation process constructs a fault-tolerant mechanism that integrates physical characteristics and real-time status, enabling the system to still determine effective alignment within a certain deviation range and improving the success rate of subsequent printing.

[0093] After aligning the predicted breakpoint coordinate parameters with the current material supply status of 3D printing, the error buffer area is input to generate a 3D printing bin replenishment breakpoint continuation matching control scheme that combines laser ranging and AI algorithms.

[0094] The predicted breakpoint coordinate parameters can be theoretical interruption position data calculated based on historical motion trajectories and the feeding model. This data can be used to align with the real-time feeding status and input into the error buffer, providing a priori position references to narrow the verification range. The current feeding status of 3D printing can be the real-time pressure, flow stability, and material continuity indicators of the feeding system at the time of resuming printing. This reflects the actual readiness of the extrusion system and aligns with the predicted coordinates to ensure reliable material supply during resuming printing. The 3D printing silo replenishment breakpoint resuming matching control scheme can be a final resuming execution strategy set integrating laser ranging, AI algorithms, and error buffer verification. This can be used to output highly reliable control commands verified by tolerance and drive the system to complete high-precision breakpoint resuming printing without human intervention. Aligning the predicted breakpoint coordinate parameters with the current feeding status of 3D printing and inputting them into the error buffer area can involve synchronizing the predicted coordinates with timestamps and normalizing the units, then using this as a joint input vector to verify its effectiveness within the error buffer. Furthermore, the input process can be implemented by using a sliding time window to align the predicted coordinates with the material feeding status sampling points and then calculating their membership degree within the error buffer, or by using Bayesian estimation to fuse the uncertainties of both and then determining whether they fall into the confidence region between the buffers. This allows for cross-validation between theoretical predictions and physical reality, and ensures that the generated continuous firing control scheme simultaneously meets the dual conditions of accurate positioning and reliable material feeding.

[0095] Taking UHPC continuous printing of bridge piers as an example, the 3D printing chamber replenishment control method combining laser ranging and AI algorithm in this embodiment can be used to decompose the resuming instruction set into an action sequence containing breakpoint coordinates, compensation amount, and high flow correction label after the bridge pier printing is paused due to replenishment. Given that the correction label indicates a high-complexity process, the proportion of AI computing resources is increased to 80%, and a larger initial coordinate search range is determined accordingly. Then, with the breakpoint target coordinates as the center and combined with the high resource ratio, the search boundary is expanded into a more inclusive spherical domain. Within this range, a three-dimensional buffer zone covering mechanical return and material hysteresis error is calculated. Finally, the breakpoint coordinates based on path prediction and the stable feeding pressure value after replenishment are aligned and input into this range. After verification, a resuming control scheme is automatically generated and the nozzle is driven to accurately reset and adapt to high flow parameters to complete seamless resuming printing, without the need for manual calibration throughout the process.

[0096] In addition, refer to Figure 2 To achieve the above objectives, the present invention also provides a 3D printing silo replenishment control system combining laser ranging and AI algorithms, the system comprising: The multi-source fusion module 10 is used to collect real-time material level data from laser ranging in the 3D printing hopper, real-time coordinate parameters of the printing nozzle, and path layering parameters of the printed components. It performs noise reduction and fusion processing on the material level and printing status data collected from multiple sources to obtain a full-state perception dataset of printing. The dynamic optimization module 20 is used to dynamically construct a hierarchical response link between the printing control system and the material supply drive system based on the difference between the material level change rate and the material consumption rate in the printing full-state perception dataset, and adjust the transmission priority of the interruption signal and the continuation printing parameter in the hierarchical response link according to the urgency of the material level falling below the threshold, so as to obtain the adjusted signal transmission priority. The instruction generation module 30 is used to convert the adjusted signal transmission priority into response weight parameters, and generate a continuation matching instruction set in combination with the position parameters of the current breakpoint. The cross-domain synchronization module 40 is used to synchronize the continuous printing matching instruction set with the real-time material supply status after the replenishment is completed across modules to obtain the synchronized continuous printing matching instruction set. The precision continue printing module 50 is used to drive the printing extrusion system to perform breakpoint alignment and parameter adaptation continue printing actions through the synchronized continue printing matching instruction set, so as to obtain a high-precision 3D printing breakpoint continue printing control scheme.

[0097] Other embodiments or specific implementations of the 3D printing silo replenishment control system combining laser ranging and AI algorithms described in this invention can be referred to the above-mentioned method embodiments, and will not be repeated here.

[0098] Furthermore, to achieve the above objectives, the present invention also provides a 3D printing chamber replenishment control device combining laser ranging and AI algorithms. The device includes: a memory, a processor, and a 3D printing chamber replenishment control program combining laser ranging and AI algorithms stored in the memory and executable on the processor. The 3D printing chamber replenishment control program combining laser ranging and AI algorithms is configured to implement the steps of the 3D printing chamber replenishment control method combining laser ranging and AI algorithms as described above.

[0099] Furthermore, to achieve the above objectives, the present invention also provides a computer-readable storage medium storing a 3D printing chamber replenishment control program that combines laser ranging and AI algorithms. When the 3D printing chamber replenishment control program that combines laser ranging and AI algorithms is executed by a processor, it implements the steps of the 3D printing chamber replenishment control method that combines laser ranging and AI algorithms as described above.

[0100] The above are merely preferred embodiments of the present invention and do not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for controlling material replenishment in a 3D printing chamber by combining laser ranging and AI algorithms, characterized in that, The method includes: Real-time material level data from laser ranging in the 3D printing hopper, real-time coordinate parameters of the printing nozzle, and path layering parameters of the printed components are collected. The material level and printing status data collected from multiple sources are denoised and fused to obtain a full-state perception dataset for printing. Based on the difference between the material level change rate and the material consumption rate in the full-state perception dataset of printing, a hierarchical response link is dynamically constructed between the printing control system and the material supply drive system. The transmission priority of the interruption signal and the resume printing parameter in the hierarchical response link is adjusted according to the urgency of the material level falling below the threshold, so as to obtain the adjusted signal transmission priority. The adjusted signal transmission priority is converted into a response weight parameter, and combined with the position parameter of the current breakpoint to generate a continuation matching instruction set; The continuous printing matching instruction set is synchronized across modules with the real-time material supply status after the replenishment is completed to obtain the synchronized continuous printing matching instruction set. The synchronized continue printing matching instruction set drives the printing extrusion system to perform continue printing actions with breakpoint alignment and parameter adaptation, so as to obtain a high-precision 3D printing breakpoint continue printing control scheme.

2. The 3D printing chamber replenishment control method combining laser ranging and AI algorithm as described in claim 1, characterized in that, The step of converting the adjusted signal transmission priority into response weight parameters and generating a continuation matching instruction set in combination with the current breakpoint position parameters includes: The hierarchical response link is divided into intervals based on the adjusted signal transmission priority level value, and the level value is converted within the interval to obtain the response weight parameter corresponding to the hierarchical response link. Extract the printed layer thickness deviation parameter and the extruder cumulative displacement deviation parameter at the current breakpoint position, and combine the printed layer thickness deviation and the extruder displacement deviation into a breakpoint position error coefficient; The response weight parameters and the breakpoint position error coefficient are jointly constrained to obtain the weight parameter distribution result. Based on the weight parameter distribution result, a continuation call generation template corresponding to each level in the hierarchical response link is constructed. The breakpoint position error coefficient is input into the continuous printing instruction generation template to generate a continuous printing matching instruction set that includes breakpoint alignment compensation and extrusion flow correction constraints.

3. The 3D printing chamber replenishment control method combining laser ranging and AI algorithm as described in claim 2, characterized in that, The step of jointly constraining the response weight parameters and the breakpoint position error coefficient to obtain the weight parameter distribution result includes: The range of variation of the breakpoint position error coefficient is divided into continuous sub-intervals, and each sub-interval corresponds to a coordinate transformation function; Based on the hierarchical position of the response weight parameter in the hierarchical response link, obtain the initial alignment conversion factor associated with the hierarchical position; Within the sub-interval, the breakpoint position error coefficient and the initial alignment transformation factor are correlated and calculated according to the coordinate transformation function to obtain the correlated sub-interval alignment parameters; The alignment parameters of the associated sub-intervals are fused by the interval boundaries to obtain the fused alignment parameters. The inter-level precision balance constraint is then applied to the fused alignment parameters to obtain the weight parameter distribution result.

4. The 3D printing chamber replenishment control method combining laser ranging and AI algorithm as described in claim 3, characterized in that, Within the sub-interval, the alignment parameters of the sub-interval are obtained by correlating the breakpoint position error coefficient with the initial alignment transformation factor according to the coordinate transformation function, including: The slope adjustment factor of the coordinate transformation function is determined according to the span range of the sub-interval, and the breakpoint position error coefficient is scaled according to the slope adjustment factor to generate the breakpoint deviation value. Within the sub-interval, the alignment accuracy adjustment factor is obtained according to the hierarchical position corresponding to the response weight parameter, and the breakpoint deviation value is cross-correlated with the alignment accuracy adjustment factor to obtain the initial alignment conversion factor. The initial alignment conversion factor is subjected to interval accumulation processing, and the accumulation processing result is superimposed with the breakpoint deviation value to generate the associated sub-interval alignment parameters.

5. The 3D printing chamber replenishment control method combining laser ranging and AI algorithm as described in claim 1, characterized in that, The system collects real-time material level data from the laser ranging system in the 3D printing hopper, real-time coordinate parameters of the printing nozzle, and path layering parameters of the printed components. It then performs noise reduction and fusion processing on the multi-source collected material level and printing status data to obtain a full-state printing perception dataset, including: The scanning coverage area of ​​the laser ranging sensor is divided into blind spot supplementation areas, and the laser ranging material level data is segmented into a discrete point cloud dataset containing spatial identifiers. Within the discrete point cloud dataset, duplicate feature matching is performed on the overlapping area data of adjacent scan frames, and redundant point cloud data points exceeding the set duplicate threshold are deleted to generate a deredundant material level dataset. Extract dynamic material consumption feature points and static material supply silo structure feature points related to material level changes from the deredundant material level dataset, separate the dynamic material consumption feature points from the static material supply silo structure feature points, and generate dynamic feature datasets and static feature datasets. The dynamic feature dataset is extended and predicted according to the material consumption rate of the current printing path to obtain an extended breakpoint coordinate sequence. The extended breakpoint coordinate sequence is then spatially concatenated with the static feature dataset to generate a full-state printing awareness dataset.

6. The 3D printing chamber replenishment control method combining laser ranging and AI algorithm as described in claim 5, characterized in that, The step of extending and predicting the breakpoint coordinates of the dynamic feature dataset according to the material consumption rate of the current printing path to obtain the extended breakpoint coordinate sequence includes: Extract the measured coordinate values ​​of the print head for continuous time periods from the dynamic feature dataset, and calculate the trend vector of print head movement based on the change in the measured coordinate values ​​of the print head. Using the measured coordinates of the print head at the moment of shutdown as the reference point, along the opposite direction of the trend vector, and based on the preset alignment compensation step size, the coordinate prediction interval of the breakpoint position is obtained, and a sequence of predicted breakpoint coordinate points is generated. The predicted breakpoint coordinate sequence is subjected to prediction confidence verification processing. The verified predicted breakpoint coordinate sequence is then merged with the dynamic feature dataset by timestamp alignment to obtain the extended breakpoint coordinate sequence.

7. The 3D printing chamber replenishment control method combining laser ranging and AI algorithm as described in claim 1, characterized in that, The method of driving the printing extrusion system to perform breakpoint alignment and parameter adaptation continuous printing actions through the synchronized continuous printing matching instruction set to obtain a high-precision 3D printing breakpoint continuous printing control scheme includes: The synchronized continuous printing matching instruction set is decomposed into a continuous printing action sequence corresponding to the printing extrusion system. The continuous printing action sequence includes the breakpoint target coordinates, the alignment compensation amount, and the extrusion parameter correction label. Based on the extrusion parameter correction label, the AI-calculated resource occupancy ratio is allocated to the printing control system, and the coordinate search range of the breakpoint alignment search is determined according to the resource occupancy ratio. The boundary conditions of the coordinate search range are dynamically adjusted based on the target coordinates of the breakpoint and the resource occupancy ratio, and the error buffer of the nozzle positioning is calculated based on the adjusted coordinate search range. The predicted breakpoint coordinate parameters and the current material supply status of 3D printing are aligned and then input into the error buffer area to generate the 3D printing chamber replenishment breakpoint continuation matching control scheme that combines laser ranging and AI algorithm.

8. A 3D printing silo replenishment control system combining laser ranging and AI algorithms, characterized in that, The system includes: The multi-source fusion module is used to collect real-time material level data from laser ranging in the 3D printing hopper, real-time coordinate parameters of the printing nozzle, and path layering parameters of the printed components. It performs noise reduction and fusion processing on the material level and printing status data collected from multiple sources to obtain a full-state printing perception dataset. The dynamic optimization module is used to dynamically construct a hierarchical response link between the printing control system and the material feeding drive system based on the difference between the material level change rate and the material consumption rate in the printing full-state perception dataset, and adjust the transmission priority of the interruption signal and the continuation printing parameter in the hierarchical response link according to the urgency of the material level falling below the threshold, so as to obtain the adjusted signal transmission priority. The instruction generation module is used to convert the adjusted signal transmission priority into response weight parameters, and generate a continuation matching instruction set in combination with the position parameters of the current breakpoint. The cross-domain synchronization module is used to synchronize the continuous printing matching instruction set with the real-time material supply status after the replenishment is completed, so as to obtain the synchronized continuous printing matching instruction set. The precision continue printing module is used to drive the printing extrusion system to perform breakpoint alignment and parameter adaptation continue printing actions through the synchronized continue printing matching instruction set, so as to obtain a high-precision 3D printing breakpoint continue printing control scheme.

9. A 3D printing chamber feeding control device combining laser ranging and AI algorithms, characterized in that, The device includes: a memory, a processor, and a 3D printing chamber replenishment control program that combines laser ranging and AI algorithms, stored in the memory and executable on the processor, the 3D printing chamber replenishment control program that combines laser ranging and AI algorithms is configured to implement the steps of the 3D printing chamber replenishment control method that combines laser ranging and AI algorithms as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a 3D printing chamber replenishment control program that combines laser ranging and AI algorithms. When the 3D printing chamber replenishment control program that combines laser ranging and AI algorithms is executed by a processor, it implements the steps of the 3D printing chamber replenishment control method that combines laser ranging and AI algorithms as described in any one of claims 1 to 7.

Images