Intelligent control method and system of a biological 3D printing device and computer device

Abstract

The application relates to an intelligent control method of a biological 3D printing device, which comprises the following steps: receiving a natural language task requirement, analyzing and parameter reasoning on the natural language task requirement based on a pre-constructed knowledge base, and generating an initial printing scheme; matching a corresponding standard operation skill template according to the natural language task requirement, and converting the initial printing scheme and the standard operation skill template into operation instructions through a model context protocol interface; in the process of executing the operation instructions, acquiring multi-modal monitoring data of the biological 3D printing device, and performing abnormality identification; in the case of identifying an abnormality, generating a dynamic intervention strategy based on the knowledge base and the multi-modal monitoring data, and dynamically adjusting by issuing the dynamic intervention strategy. The application can solve the technical problems that biological 3D printing is strongly dependent on artificial trial and error and knowledge and experience are difficult to reuse, and improve the success rate, safety and operation consistency of biological 3D printing.

Description

Technical Field

[0001] This application relates to the field of bio-3D printing technology, and in particular to an intelligent control method, system and computer device for a bio-3D printing device. Background Technology

[0002] Bioprinting technologies (such as tissue engineering scaffold printing, cell-loaded bio-ink printing, and microfluidic structure printing) are highly sensitive to material conditions, environmental conditions, and printing processes.

[0003] Existing bioprinting control software primarily relies on manual parameter configuration and manual operation processes. Operators need to complete model import, slicing, material selection, parameter entry, and equipment debugging in multiple independent interfaces. During actual printing, if defects such as filament breakage, over-extrusion, under-extrusion, collapse, poor interlayer adhesion, pore size deviation, or nozzle clogging occur, it usually requires continuous manual observation and manual pause of the equipment or trial adjustment of parameters based on experience. This operating mode suffers from high trial-and-error costs, delayed anomaly detection, and the knowledge scattered in experimental records, historical printing tasks, and expert experience is difficult to effectively reuse.

[0004] Currently, no effective solutions have been proposed to address the technical challenges of high operational barriers for bio-3D printing equipment and the difficulty in standardizing and reusing process knowledge. Summary of the Invention

[0005] This application provides an intelligent control method, system, and computer device for bio-3D printing equipment, which at least solves the problem of high operating threshold for bio-3D printing equipment in related technologies.

[0006] In a first aspect, embodiments of this application provide an intelligent control method for a bio-3D printing device, the method comprising: Receive natural language task requirements, parse and reason about the natural language task requirements based on a pre-built knowledge base, and generate an initial printing scheme; According to the natural language task requirements, a corresponding standard operation skill template is matched. Based on the workflow steps defined in the standard operation skill template, the initial printing scheme is dynamically orchestrated to obtain operation instructions. The operation instructions are then gradually sent to the bio-3D printing device through the model context protocol interface. During the execution of the operation instructions, multimodal monitoring data of the bio-3D printing equipment is acquired, and spatial alignment is performed between the multimodal monitoring data and the target three-dimensional model and slicing path for anomaly identification. Upon detecting an abnormal trend or identifying an anomaly, a dynamic intervention strategy is generated based on the knowledge base and the multimodal monitoring data. The dynamic intervention strategy is then issued to instruct the bio-3D printing device to dynamically adjust the current printing content and / or printing parameters.

[0007] In some embodiments, the knowledge base employs a structured storage architecture and a progressive disclosure mechanism. Based on the pre-built knowledge base, it parses and infers parameters for the natural language task requirements, generating an initial printing scheme including: The natural language task requirements are analyzed using a large language model to extract the printing intent; Using a retrieval enhancement mechanism, a background knowledge set matching the printing intention is retrieved from the knowledge base, wherein the background knowledge set includes: material and process knowledge, equipment and historical experience, and multimodal image knowledge; The initial printing scheme is generated based on the background knowledge set, the equipment capabilities and current status of the bio-3D printing equipment, and the model features of the target 3D model.

[0008] In some embodiments, based on the workflow steps defined in the standard operation skill template, the initial printing scheme is dynamically orchestrated to obtain operation instructions, and these operation instructions are progressively issued to the bio-3D printing device through the model context protocol interface, including: The validity of the parameter configuration of the initial printing scheme is verified by using the preset process constraint rules and safety verification conditions in the standard operation skill template. After the verification is passed, the execution intention of the current workflow step is combined with the specific parameters in the initial printing scheme according to the workflow step sequence defined in the standard operation skill template to generate the operation instruction for the current work step. The model context protocol interface is invoked to convert the operation instructions of the current work step into a tool interface call request corresponding to the bio-3D printing device and send it out. After receiving the execution status feedback of the bio-3D printing device for the current work step, the workflow step is updated to drive the generation and sending of operation instructions for subsequent work steps.

[0009] In some embodiments, acquiring multimodal monitoring data of the bio-3D printing device and spatially aligning the multimodal monitoring data with the target 3D model and slice path for anomaly identification includes: The multimodal monitoring data is obtained by acquiring real-time image data and sensor data collected by the bio-3D printing equipment. Based on the multimodal monitoring data, spatial alignment calculations are performed on the current nozzle position of the bio-3D printing device, the real-time image data, the outline of the target 3D model, and the current slice layer path. Based on the results of the spatial alignment calculation, anomalies are identified, including: over-extrusion anomaly, under-extrusion anomaly, localized wire breakage anomaly, and structural collapse anomaly.

[0010] In some embodiments, dynamically adjusting the current print content and / or print parameters includes: Upon detecting the over-extrusion and under-extrusion anomalies, fine-tune the printing speed and / or extrusion pressure parameters according to the current deviation status. If the local filament breakage anomaly is detected, the detected filament breakage image location is reverse-mapped to the slicing path to determine the filament breakage start point and the missing path segment. The current printing task is paused and the printhead is moved to the filament breakage start point to reprint the missing path segment. If the structural collapse anomaly is detected, the unprinted subsequent slice layers are replanned, and boundary reinforcement paths are dynamically added to the subsequent slice layers while the local printing speed is reduced. For unrecoverable local defect areas, the printing path of the corresponding area can be skipped by modifying the underlying control commands.

[0011] In some embodiments, after generating the dynamic intervention strategy, the method further includes: The dynamic intervention strategy is subjected to parameter physical range verification, platform anti-collision boundary verification, and model printability verification. After the physical range of the parameters, the anti-collision boundary of the platform, and the printability of the model are verified, access control is performed according to the preset risk classification mechanism to execute the adjustment instructions corresponding to the dynamic intervention strategy. After executing the adjustment command, if the monitoring index exceeds the preset range, a dynamic rollback strategy is executed to roll back the printing status to the previous safety parameter or implement a safety pause.

[0012] In some embodiments, low-risk fine-tuning instructions are executed automatically under a managed system. For medium-risk instructions, the system will forcibly suspend the process and push a confirmation-pending solution, then execute the adjustment instructions based on the user's confirmation result. High-risk commands that exceed security limits will be intercepted and prohibited from being issued.

[0013] In some embodiments, the method further includes: After the printing task is completed, the initial printing plan, interface call logs, dynamic intervention strategies, multimodal monitoring data and user evaluations are merged into historical case data and stored in the knowledge base. The effectiveness of the dynamic intervention strategy is verified. If the strategy is effective, the corresponding strategy action is used as reusable experience and updated to the standard operation skill template of the corresponding material.

[0014] Secondly, embodiments of this application provide an intelligent control system for a bio-3D printing device, the system comprising: The requirement response module is used to receive natural language task requirements, parse and reason about the natural language task requirements based on a pre-built knowledge base, and generate an initial printing scheme. The task generation module is used to dynamically arrange the instructions of the initial printing scheme to obtain operation instructions based on the workflow steps defined in the standard operation skill template, and gradually issue the operation instructions to the bio-3D printing device through the model context protocol interface. The monitoring and identification module is used to collect multimodal monitoring data of the bio-3D printing equipment during the execution of the operation instructions, and to spatially align the multimodal monitoring data with the target three-dimensional model and the slice path to identify anomalies. The optimization and adjustment module is used to generate a dynamic intervention strategy based on the knowledge base and the multimodal monitoring data when an abnormal development trend is detected or an anomaly is identified, and to instruct the bio-3D printing device to dynamically adjust the current printing content and / or printing parameters by issuing the dynamic intervention strategy.

[0015] Thirdly, embodiments of this application provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the method described in the first aspect.

[0016] Compared to related technologies, the intelligent control method and system for a bio-3D printing device provided in this application receives natural language task requirements, combines them with a knowledge base for parsing and parameter reasoning to generate an initial printing plan, and uses standard operation skill templates and model context protocol interfaces to convert the plan into operation instructions and issue them for execution. During the printing process, multimodal monitoring data is acquired and spatially aligned with the target 3D model and slicing path to achieve millisecond-level anomaly identification. Under the hierarchical management of safety risks, a dynamic intervention strategy is generated to instruct the device to make real-time dynamic adjustments to the printing content and parameters, while the entire cycle of data is written back to the knowledge base to form an iterative closed loop. This application solves the technical problems of bio-3D printing, such as heavy reliance on manual trial and error, difficulty in reusing knowledge and experience, lagging process monitoring, and lack of dynamic content correction capabilities. It realizes knowledge-driven parameter optimization, intelligent scheduling and orchestration, and real-time monitoring and dynamic safety intervention of the printing status, which can improve the success rate, safety, and operational consistency of bio-3D printing. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart of an intelligent control method for a bio-3D printing device according to an embodiment of this application; Figure 2 This is a flowchart illustrating an embodiment of the present application for generating an initial printing scheme; Figure 3 This is a flowchart illustrating the generation of operation instructions and their issuance to a bio-3D printing device for execution, according to an embodiment of this application. Figure 4 This is a flowchart illustrating an anomaly identification based on multimodal monitoring data according to an embodiment of this application; Figure 5 This is a flowchart of another intelligent control method for a bio-3D printing device according to an embodiment of this application; Figure 6 This is a structural block diagram of an intelligent control system for a bio-3D printing device according to an embodiment of this application; Figure 7 This is a schematic diagram of the internal structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the application is described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application.

[0019] Obviously, the accompanying drawings described below are merely some examples or embodiments of this application. Those skilled in the art can apply this application to other similar scenarios based on these drawings without any inventive effort. Furthermore, it is understood that although the efforts made in this development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, any changes to design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as insufficient disclosure of the content of this application.

[0020] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that is mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0021] Unless otherwise defined, the technical or scientific terms used in this application shall have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms “a,” “an,” “an,” “the,” and similar words used in this application do not indicate quantity limitation and may indicate singular or plural. The terms “comprising,” “including,” “having,” and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units not listed, or may include other steps or units inherent to these processes, methods, products, or devices. The terms “connected,” “linked,” “coupled,” and similar words used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. “Multiple” used in this application refers to two or more. “And / or” describes the relationship between related objects, indicating that three relationships may exist; for example, “A and / or B” can represent: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following objects are in an "or" relationship. The terms "first," "second," and "third" used in this application are merely to distinguish similar objects and do not represent a specific ordering of the objects.

[0022] This embodiment provides an intelligent control method for a bio-3D printing device. Figure 1 This is a flowchart of an intelligent control method for a bio-3D printing device according to an embodiment of this application, such as... Figure 1 As shown, the process includes the following steps: Step S101: Receive the natural language task requirements, parse and reason the natural language task requirements based on the pre-built knowledge base, and generate an initial printing scheme.

[0023] Specifically, Figure 2 This is a flowchart illustrating an embodiment of the present application for generating an initial printing scheme, such as... Figure 2 As shown, this step includes the following sub-steps: S101a: Collect knowledge related to biological 3D printing and build a knowledge base; In this embodiment, scattered knowledge related to bio-3D printing is pre-organized and constructed into a knowledge base. This knowledge base contains static documents, employs a structured storage architecture and a progressive disclosure mechanism, and is configured as a knowledge graph or structured database covering multiple dimensions of deep correlation, including problems, causes, parameters, devices, materials, models, and results.

[0024] Specifically, the knowledge base includes: material knowledge (including GelMA concentration, alginate concentration, viscosity range, crosslinking method, and temperature control requirements, used to determine applicable material parameters and adjustment direction), equipment knowledge (including equipment model, working space, nozzle specifications, pressure range, temperature control range, and speed range, used to define parameter boundaries), process knowledge (including recommended layer height, path spacing, filling density, and pressure matching rules, used to generate initial parameter schemes), anomaly knowledge (including defect types and cause analysis such as filament breakage, over-extrusion, under-extrusion, collapse, head blockage, and interlayer misalignment), image knowledge (including printing process images, defect annotation diagrams, micrographs, and finished product comparison diagrams, used as a visual diagnostic benchmark), historical task data (covering models, materials, parameters, environment, result scores, and failure reasons), and expert rules (such as constraints such as "pressure adjustment should not exceed a preset ratio each time" and "high cell density materials should prioritize reducing shear risk"). This knowledge base is linked using vector retrieval, keyword indexing, structured databases, or knowledge graphs.

[0025] S101b utilizes a retrieval enhancement mechanism to retrieve a background knowledge set matching the printing intent from the knowledge base. The background knowledge set includes: material and process knowledge, equipment and historical experience, and multimodal image knowledge. In specific application scenarios, after receiving the user's natural language task requirements, the AI-based task understanding module identifies the task objective, equipment object, model file, material information, time requirements, parameter preferences, and safety constraints. If any information is missing, a follow-up questioning mechanism is triggered, or the current task context is automatically retrieved. Subsequently, through a retrieval enhancement mechanism, a background knowledge set matching the printing intent is retrieved from the knowledge base.

[0026] S101c generates an initial printing plan based on the background knowledge set, the equipment capabilities and current status of the bio-3D printing equipment, and the model characteristics of the target 3D model.

[0027] Through the parameter design module, combined with the retrieved background knowledge set, current equipment capabilities, material properties, and model features of the target 3D model, multi-factor parameter design is performed to infer and generate one or more initial printing schemes that include pressure, speed, layer height, temperature, and crosslinking parameters.

[0028] As an exemplary implementation scenario in the prepress parameter design stage of this step, when receiving a natural language request such as "Printing a scaffold using GelMA material, the previous step had edge collapse, current pressure 35kPa, speed 8mm / s, layer height 0.25mm, how should it be adjusted?", the task understanding module extracts material properties, defect properties (edge ​​collapse), and current process parameters. The knowledge retrieval module extracts process knowledge and historical collapse cases for this material. The parameter design module further generates specific optimization solutions, such as outputting instructions like "It is recommended to first reduce the layer height to 0.18 to 0.2mm, appropriately reduce the speed to 6mm / s, and check the crosslinking delay conditions." Simultaneously, a risk assessment is performed. If the image history database reveals an over-extrusion risk under the aforementioned pressure, a risk warning is output, explicitly blocking further pressure increases, and suggesting the generation of a small-size verification model.

[0029] Furthermore, as another optional interactive scenario in this application embodiment, the intelligent control method also includes an interactive knowledge question-and-answer session. Users can input free natural language questions related to bio-3D printing processes, material rheological properties, or equipment maintenance rules through the interactive interface. After parsing the question-and-answer request, the task understanding module converts it into a knowledge retrieval vector, and retrieves matching process mechanisms or operation guidelines across modalities from the knowledge graph, static documents, and historical case data in the knowledge base. Subsequently, the large language model combines the retrieved background knowledge to generate a structured and interpretable response text, which is then pushed to the user's terminal. Further, during the printing task execution, this knowledge question-and-answer session can be linked in real-time with the status reporting module, allowing users to retrieve the cause analysis of current anomalies through instant question-and-answer (e.g., a user asks online, "Why is it now showing over-extrusion?"). The system can combine the multimodal monitoring data of the current layer with the anomaly knowledge in the knowledge base to perform correlation reasoning, and output targeted qualitative diagnostic reports and troubleshooting suggestions online.

[0030] It is understandable that step S101 involves building a knowledge base, using a natural language processing module to parse user needs and retrieve matching knowledge, and then performing multi-factor parameter reasoning to generate an initial printing plan. This approach overcomes the drawback of bio-3D printing's heavy reliance on manual trial and error, directly outputting interpretable parameter optimization guidance through knowledge-driven methods, thus lowering the barrier to equipment operation.

[0031] Step S102: Match the corresponding standard operation skill template according to the requirements of the natural language task, and dynamically arrange the operation instructions for the initial printing scheme based on the workflow steps defined in the standard operation skill template. Then, issue the operation instructions to the bio-3D printing device step by step through the model context protocol interface.

[0032] Figure 3This is a flowchart illustrating the generation of operation instructions and their issuance to a bio-3D printing device for execution, according to an embodiment of this application. Figure 3 As shown, the process includes the following steps: S102a, using the preset process constraint rules and safety verification conditions in the standard operation skill template, performs a legality verification of the parameter configuration of the initial printing scheme; In this step, to seamlessly integrate AI natural language understanding capabilities with the underlying device interface, standard operation skill templates (Skills) and a Model Context Protocol (MCP) interface are introduced. The Standard Operation Skill Templates serve as editable carriers storing specific device tasks and standard process flows. In this embodiment, their internal definitions include: applicable scenarios, input requirements, parameter constraints (e.g., layer height range, speed range), safety constraints (e.g., nozzle calibration and platform leveling must be performed before printing), and the order of MCP tool calls. The Model Context Protocol interface provides standardized external tool call specifications, mapping the underlying hardware control capabilities of the bio-3D printer into a set of instructions that can be directly invoked by artificial intelligence.

[0033] Furthermore, the corresponding standard operation skill template is matched based on the parsed task type. The constraints in the template are then used to perform safety checks on the initial printing scheme, specifically including parameter physical range checks, platform anti-collision boundary checks, and model printability checks.

[0034] S102b, after verification, according to the workflow steps defined in the standard operation skill template, the execution intent of the current workflow step is combined with the specific parameters in the initial printing scheme to generate the operation instructions for the current work step; After verification, the execution intent of the current workflow step is combined with the specific parameters in the initial printing scheme according to the workflow step sequence defined in the standard operation skill template to generate the operation instruction for the current work step. It should be noted that this embodiment supports multiple scheduling modes, including: immediate execution, scheduled execution (parsing time nodes in natural language and registering trigger times), periodic execution (used for equipment status inspection or printhead cleaning), conditional triggering (triggered when the temperature reaches the standard or the equipment is idle), queue scheduling (sorted according to task priority, material type and expected duration), and failure retry strategy.

[0035] In one exemplary embodiment, regarding task orchestration and scheduling, taking a periodic equipment inspection scenario as an example, if the received natural language request is "Check the status of the bioprinter once every morning at 8:00 AM, including connection, temperature control, platform position, and nozzle status; notify if any abnormalities are found," the system will automatically match the corresponding "Standard Operating Skill Template for Bioprinting Equipment Inspection" after parsing, and convert the task into a periodic execution task registered with the task scheduling module. At the set time each day, the system automatically calls the Model Context Protocol interface to obtain the underlying equipment status. If all indicators are normal, the task is marked as completed and silently recorded; if the temperature sensor data is unresponsive or image recognition detects residual gel clogging the nozzle, a diagnostic description including images of the abnormal area is generated, pushed to an external collaboration platform, and the equipment is suspended for standby, avoiding material waste caused by invalid printing.

[0036] S102c, after the orchestration is completed, calls the model context protocol interface to convert the operation instructions of the current working step into the tool interface call request corresponding to the bio-3D printing device and sends it out. After receiving the execution status feedback of the bio-3D printing device for the current working step, it triggers the update of the workflow steps to drive the generation and sending of operation instructions for subsequent working steps.

[0037] Specifically, through the MCP device adaptation module, standardized tools of the 3D printing equipment (including status query, model loading, slice path generation, parameter setting, preview generation, and print start) are invoked to sequentially complete the instruction issuance and drive the bio-3D printing equipment to execute the printing task.

[0038] Step S102, by matching standard operation skill templates and performing parameter boundary and printability safety checks, orchestrates the initial scheme and standardized control sequences into operation instructions, which are then issued through the model context protocol interface. This approach achieves the connection between natural language intent and the underlying hardware interface, reuses standard process flows, and ensures the safety and controllability of equipment operation task scheduling.

[0039] Step S103: During the execution of the operation instructions, acquire the multimodal monitoring data of the bio-3D printing equipment, and spatially align the multimodal monitoring data with the target 3D model and the slice path to identify anomalies.

[0040] Figure 4 This is a flowchart illustrating an anomaly identification based on multimodal monitoring data according to an embodiment of this application, such as... Figure 4 As shown, the process includes the following steps: S103a collects multimodal monitoring data through multidimensional sensors and image acquisition devices; After the printing task is started, a real-time monitoring channel is established to obtain multimodal monitoring data. The obtained multimodal monitoring data includes: equipment status data (connection status, task progress, printhead position, platform position, alarm code), sensor data (temperature, pressure, speed, flow rate, humidity, crosslinking strength), and image data (real-time camera images, magnified views, and interlayer images).

[0041] S103b, based on multimodal monitoring data, performs spatial alignment calculations between the current nozzle position and real-time image data of the bio-3D printing equipment and the outline of the target 3D model and the current slice layer path; Furthermore, after acquiring multi-mode monitoring data, the image understanding module reads the three-dimensional spatial coordinates of the current nozzle in real time and establishes a virtual projection in the digital space. It performs multi-dimensional spatial overlap and homogeneous coordinate alignment calculations on the pixel areas of the two-dimensional image captured by the real-time macro camera, the geometric line segments of the slice path being executed in the current layer, and the original boundary mesh of the target three-dimensional model.

[0042] S103c, based on the results of spatial alignment calculation, identifies whether there are any abnormal phenomena, including: over-extrusion anomaly, under-extrusion anomaly, local wire breakage anomaly, and structural collapse anomaly.

[0043] Based on the spatial alignment calculation results obtained in step S103b above, anomaly identification is performed.

[0044] Specifically, for material extrusion printing processes, the identified anomalies include, but are not limited to: filament breakage (discontinuous lines in image recognition, inferred from the knowledge base as insufficient extrusion or slight blockage), over-extrusion / material accumulation (abnormally thicker line width, inferred as high pressure or low speed), under-extrusion (line width gradually narrows), structural collapse (deformation and sagging in the overhanging area), and interlayer misalignment. For photopolymerization (such as DLP) printing processes, the abnormal phenomena that can be identified can be: over-curing abnormality (cured area exceeds the outline of the slice target), under-curing abnormality (blurred edges of the exposed area, insufficient interlayer curing strength leading to layer peeling), and local uncured collapse defects (specific suspended or bridging structures fail to form a solid support with sufficient strength, resulting in local sagging, local missing parts, or floating of the formed object).

[0045] It should be noted that in real-world scenarios, anomalies can also include warping and shrinkage deformation, filament tearing, internal air bubbles, scratches and abrasions, layer misalignment, and nozzle blockage leading to filament failure. Those skilled in the art should understand that adaptive adjustments or equivalent substitutions made based on the monitoring, identification, and dynamic intervention mechanisms disclosed in this application to address other anomalies that may occur in bio-3D printing scenarios do not depart from the core technical concept of this application. The specific anomaly types and corresponding strategies listed above should not be construed as specific limitations on the scope of protection of this solution. In addition, it should be noted that during the above-mentioned detection and identification of anomalies, the status reporting module actively pushes status summaries to users according to the set reporting strategy (periodic reporting, phased reporting, immediate anomaly reporting, or user Q&A query reporting). The report content includes the current printing progress, quality trend, deviation status, and potential risks.

[0046] Step S103 involves real-time acquisition of multimodal monitoring data, including equipment status, sensor readings, and images. This data is then used to perform homogeneous coordinate overlap and alignment calculations with the target 3D model and slice path in digital space for anomaly identification. This approach changes the traditional reliance on continuous manual observation, enabling high-precision digital visual monitoring during the printing process and improving the timeliness and accuracy of anomaly detection.

[0047] Step S104: When an abnormal development trend is detected or an anomaly is identified, a dynamic intervention strategy is generated based on the knowledge base and multimodal monitoring data. The dynamic intervention strategy is then issued to instruct the bio-3D printing device to dynamically adjust the current printing content and / or printing parameters.

[0048] When printing defects or abnormal trends are identified, the dynamic adjustment decision module combines real-time multimodal feedback, target paths, and expert rules from the knowledge base to generate specific dynamic intervention strategies. These dynamic adjustments are not limited to global parameter modifications but also include fine-grained modifications to the printed content. Specific modifications include, but are not limited to: Targeted adjustments for over-extrusion / under-extrusion anomalies: When an over-extrusion or under-extrusion trend is detected, dynamic parameter fine-tuning is performed, slightly changing the printing speed or extrusion pressure parameters according to the current deviation status.

[0049] Targeted adjustment for filament breakage anomalies: When a local filament breakage anomaly is detected, a local re-printing strategy is executed: the detected filament breakage image location is reverse-mapped to the slicing path to determine the filament breakage start point and the missing path segment; the current main task is paused via the MCP interface, the printhead is moved to the filament breakage start point, and the missing path segment is re-printed at a preset safe low speed, seamlessly resuming the original task after completion. As a specific application scenario of the intervention strategy, when the image understanding module detects a local line interruption during printing, the target alignment module maps the filament breakage location space on the current image to the slice path data of the current layer to determine whether the location is suitable for re-printing. If the conditions are met, a dynamic intervention strategy containing a local re-printing path plan is generated. Optionally, given that path adjustment is a medium-risk operation, a pop-up window is used to report to the user terminal and request confirmation. After obtaining the authorization confirmation instruction, the model context protocol interface is called to execute the action sequence in sequence: pause the main path, control the printhead to rise and move to the filament breakage start point coordinates, start extruding a predetermined amount of material and re-print the missing path at a set safe low speed, and then return to the original interruption coordinates to continue the original printing task. After the reprinting operation is completed, the monitoring module continuously checks the image continuity of the reprinted area and archives the above processing process in the task log.

[0050] Targeted adjustments for structural collapse anomalies: Upon identifying an abnormal trend in structural collapse, predictive support and reinforcement strategies are implemented: The subsequent slice layers that have not yet been printed are replanned, and boundary reinforcement paths are dynamically added to these layers while reducing local printing speed to protect the completed structure. For this predictive intervention scenario under this structural collapse trend, when printing complex features such as cantilevered structures of tissue engineering scaffolds, if real-time camera monitoring data and the target model show slight sagging at the cantilever edge, it is determined that if subsequent layers continue to be extruded along the original path, the structural stability will face extremely high risks. Multiple contingency plans are generated instantly: for example, Plan 1 pauses the equipment and waits for the current layer material to complete gel curing; Plan 2 replans the printing paths for the next two layers, dynamically adding a boundary reinforcement path around the cantilever and reducing local printing speed. Upon receiving the user's instruction to select Plan 2, the data for the subsequent slice layers that have not yet been printed is immediately modified and sent to the equipment. This dynamic intervention only applies to the content of subsequent layers, ensuring no secondary damage to the already formed structure.

[0051] For local defect areas that have failed and cannot be recovered, a skip strategy is implemented: the subsequent printing path of the corresponding area is skipped directly by modifying the underlying control commands to avoid interfering with the surrounding normal structure.

[0052] It should be noted that the specific modifications and adjustments described above are merely illustrative. Those skilled in the art should understand that in actual bio-3D printing control processes, for other abnormal phenomena, such as shrinkage deformation, filament pulling, internal air bubbles, scratches, layer misalignment, and nozzle blockage leading to filament failure, the equivalent process parameter compensation or printing content path changes performed for other types of printing defects, or the conventional equivalent replacements, trimmings, and combinations of the action sequences, execution order, safety judgment threshold boundaries, and control parameter adjustment steps in the aforementioned dynamic intervention strategies, are all within the scope of the core technical concept of this application and should not be construed as specific limitations on this solution.

[0053] In addition, to ensure the stability and safety of equipment operation, strict access control and security constraints are implemented according to a preset risk classification mechanism before issuing dynamic intervention strategies. For low-risk operations (such as parameter fine-tuning within ±5%), automatic parameter tuning is performed under managed authorization; for medium- to high-risk operations (such as local path replanning, local re-printing, and significant temperature adjustments), the task is forcibly suspended, a confirmation plan is generated and a notification is pushed, awaiting manual confirmation; for high-risk operations (such as exceeding equipment physical thresholds, serious congestion, or other hardware alarms), the task is directly terminated, a safety pause is implemented, and manual intervention is prompted.

[0054] Taking the high-risk anomaly-triggered termination mechanism as an example, when printing high-concentration cell bio-ink, if the equipment returns a high-pressure alarm to the extruder at the bottom layer, and the multimodal image simultaneously shows a complete interruption of material output at the nozzle, the anomaly is determined to be a hard nozzle blockage of an extremely high risk level. Bypassing the dynamic intervention strategy generation stage, an emergency pause command is immediately invoked to forcibly stop the execution plan, intercept any command attempting to increase extrusion pressure, and push a red warning to the user terminal containing the alarm code, the number of layers affected, and image evidence, suggesting standard printhead cleaning or replacement operations to ensure the physical safety of the precision hardware and high-value biological cell materials. Furthermore, this embodiment also provides an action throttling and dynamic rollback strategy. If, after executing the adjustment command, the monitoring indicators continue to deteriorate or trigger new alarms, an automatic rollback is executed, reverting the printing status to the previous safe parameter or forcibly pausing the process.

[0055] Step S104, based on identified anomalies and knowledge base rules, generates and distributes dynamic intervention strategies, including parameter fine-tuning or local path reprinting, under strict security risk classification control and action throttling mechanisms. The data is then written back to the knowledge base. This method enables the printing device to dynamically correct content during execution, preventing out-of-control risks and avoiding or repairing scrap, forming a self-iterable intelligent closed loop.

[0056] Figure 5This is a flowchart of another intelligent control method for a bio-3D printing device according to an embodiment of this application, such as... Figure 5 As shown, after generating and dynamically adjusting the printing scheme, an iterative optimization step is also included. To achieve continuous iteration and optimization of the system, the following steps are performed after the task is completed: S105 summarizes the initial printing plan, interface call logs, multimodal monitoring data, dynamically adjusted actions, image evidence, and user evaluation results, and merges them into historical case data stored in the knowledge base.

[0057] S106 verifies and calculates the actual contribution and effectiveness of the dynamic intervention strategy. If the verification is effective, the corresponding strategy action is transformed into reusable experience and automatically updated to the standard operation skill template of the corresponding material, forming a closed-loop calibration mechanism for system self-evolution.

[0058] Through steps S101 to S106 above, this embodiment provides an AI-based control method for bio-3D printing. First, it parses natural language requirements and uses a multi-dimensional knowledge base to perform parameter reasoning to generate an initial printing plan. Then, it matches a standard operation skill template and performs rigorous security checks. The plan is then converted into operation commands and sent to the underlying device using a model context protocol interface. During printing, multimodal monitoring data is collected, and defects are identified at the millisecond level by spatially aligning it with the target model and slicing path. If an abnormal trend is detected, a dynamic intervention strategy covering parameter fine-tuning, local re-printing, or path replanning is generated under a risk-level control mechanism. This strategy instructs the device to correct the printed content in real time, and finally, the entire lifecycle execution data is written back to the knowledge base, driving the continuous iterative evolution of system control experience. This improves the success rate, safety, and operational consistency of bio-3D printing.

[0059] This application also provides an intelligent control system for a bio-3D printing device. Figure 6 This is a structural block diagram of an intelligent control system for a bio-3D printing device according to an embodiment of this application, such as... Figure 6 As shown, the system includes: The requirement response module 60 is used to receive natural language task requirements, parse and reason about the natural language task requirements based on a pre-built knowledge base, and generate an initial printing scheme. In specific application scenarios, upon receiving the user's natural language task requirements, the AI-based task understanding module identifies the task objective, equipment object, model file, material information, time requirements, parameter preferences, and safety constraints. If any information is missing, a follow-up questioning mechanism is triggered, or the current task context is automatically retrieved. Subsequently, a background knowledge set matching the printing intent is retrieved from the knowledge base. Through the parameter design module, combining the retrieved background knowledge set, current equipment capabilities, material properties, and model features of the target 3D model, multi-factor parameter design is performed, generating one or more initial printing schemes that include pressure, speed, layer height, temperature, and crosslinking parameters.

[0060] This demand response module 60 overcomes the shortcomings of bio-3D printing, which relies heavily on manual trial and error. It directly outputs interpretable parameter optimization guidance through knowledge-driven approaches, thereby reducing the operational threshold of the equipment.

[0061] The task generation module 61 is used to match the corresponding standard operation skill template according to the natural language task requirements, and to convert the initial printing scheme and standard operation skill template into operation instructions through the model context protocol interface and send them to the bio-3D printing device for execution. In one exemplary embodiment, regarding task orchestration and scheduling, taking a periodic equipment inspection scenario as an example, if the received natural language request is "Check the status of the bioprinter once every morning at 8:00 AM, including connection, temperature control, platform position, and nozzle status; notify if any abnormalities are found," the system will automatically match the corresponding "Standard Operating Skill Template for Bioprinting Equipment Inspection" after parsing, and convert the task into a periodic execution task registered with the task scheduling module. At the set time each day, the system automatically calls the Model Context Protocol interface to obtain the underlying equipment status. If all indicators are normal, the task is marked as completed and silently recorded; if the temperature sensor data is unresponsive or image recognition detects residual gel clogging the nozzle, a diagnostic description including images of the abnormal area is generated, pushed to an external collaboration platform, and the equipment is suspended for standby, avoiding material waste caused by invalid printing.

[0062] The task generation module 61 can connect natural language intents with the underlying hardware interface and reuse standard process flows to ensure the safety and controllability of equipment operation task scheduling.

[0063] The monitoring and identification module 62 is used to collect multimodal monitoring data of the bio-3D printing equipment during the execution of operation instructions, and to spatially align the multimodal monitoring data with the target 3D model and the slice path to identify anomalies. Specifically, the identified anomalies include, but are not limited to: broken wires (discontinuous lines in image recognition, inferred from the knowledge base as insufficient extrusion or slight blockage), over-extrusion / material accumulation (abnormally thicker line width, inferred as high pressure or low speed), under-extrusion (line width gradually narrows), structural collapse (deformation and sagging in the cantilever area), and interlayer offset.

[0064] This monitoring and identification module 62 changes the traditional mode of relying on continuous manual observation, and can realize high-precision digital visual monitoring during the printing process, improving the timeliness and accuracy of anomaly detection.

[0065] The optimization and adjustment module 63 is used to generate dynamic intervention strategies based on the knowledge base and multimodal monitoring data when an anomaly is detected, and to instruct the bio-3D printing equipment to dynamically adjust the current printing content and / or printing parameters by issuing the dynamic intervention strategies.

[0066] When printing defects or abnormal trends are identified, the dynamic adjustment decision module combines real-time multimodal feedback, target paths, and expert rules from the knowledge base to generate specific dynamic intervention strategies. These dynamic adjustments are not limited to global parameter modifications but also include fine-grained modifications to the printed content. The types of abnormal phenomena / defects include, but are not limited to, over-extrusion anomalies, under-extrusion anomalies, localized filament breakage anomalies, and structural collapse anomalies.

[0067] Finally, to achieve continuous iteration and optimization of the system, after the task is completed, the initial printing plan, interface call logs, multimodal monitoring data, dynamically adjusted actions, image evidence, and user evaluation results are summarized and merged into historical case data and stored in the knowledge base. The system verifies and calculates the actual contribution and effectiveness of the dynamic intervention strategy. If the verification is effective, the corresponding strategy actions are transformed into reusable experience and automatically updated to the standard operation skill template of the corresponding materials, forming a self-evolving closed-loop calibration mechanism for the system.

[0068] In addition to the aforementioned printing process defects, it should also be noted that the abnormal phenomena also include equipment errors and warning messages from the bio-3D printing equipment. It is understood that the identification method for this type of anomaly is not limited to image recognition and alignment calculations. Specifically, by real-time analysis of alarm codes or communication logs in the equipment status data, it is possible to identify hardware-level errors and warnings online, including but not limited to: low-level hardware communication connection timeouts, stepper motor stalls, temperature control module heating timeouts, and sensor disconnections. When such equipment errors or warning messages are identified, corresponding hardware-level safety protection actions are triggered (such as cutting off the heating power supply and locking the air valve).

[0069] The system addresses the technical challenges of bio-3D printing, such as its heavy reliance on manual trial and error, difficulty in reusing knowledge and experience, lagging process monitoring, and lack of dynamic content correction capabilities. It enables knowledge-driven parameter optimization, intelligent scheduling and orchestration, and real-time monitoring and dynamic safety intervention of the printing status, thereby improving the success rate, safety, and operational consistency of bio-3D printing.

[0070] In one embodiment, Figure 7 This is a schematic diagram of the internal structure of an electronic device according to an embodiment of this application, such as... Figure 7 As shown, an electronic device is provided, which can be a server, and its internal structure diagram can be as follows. Figure 7 As shown, the electronic device includes a processor, a network interface, internal memory, and non-volatile memory connected via an internal bus. The non-volatile memory stores an operating system, computer programs, and a database. The processor provides computing and control capabilities, the network interface communicates with external terminals via a network, the internal memory provides an environment for the operating system, the computer programs are executed by the processor to implement an intelligent control method for a bio-3D printing device, and the database stores data.

[0071] Those skilled in the art will understand that Figure 7 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the electronic device to which the present application is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0072] The technical features of the embodiments described above can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification. The embodiments described above only illustrate several implementation methods of this application, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the invention patent. It should be noted that for those skilled in the art, several modifications and improvements can be made without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A smart control method for a bio-3D printing device, characterized in that, The method includes: Receive natural language task requirements, parse and reason about the natural language task requirements based on a pre-built knowledge base, and generate an initial printing scheme; According to the natural language task requirements, a corresponding standard operation skill template is matched. Based on the workflow steps defined in the standard operation skill template, the initial printing scheme is dynamically orchestrated to obtain operation instructions. The operation instructions are then gradually sent to the bio-3D printing device through the model context protocol interface. During the execution of the operation instructions, multimodal monitoring data of the bio-3D printing equipment is acquired, and spatial alignment is performed between the multimodal monitoring data and the target three-dimensional model and slicing path for anomaly identification. Upon detecting an abnormal trend or identifying an anomaly, a dynamic intervention strategy is generated based on the knowledge base and the multimodal monitoring data. The dynamic intervention strategy is then issued to instruct the bio-3D printing device to dynamically adjust the current printing content and / or printing parameters.

2. The method according to claim 1, characterized in that, The knowledge base employs a structured storage architecture and a progressive disclosure mechanism. Based on the pre-built knowledge base, it parses and infers parameters for the natural language task requirements, generating an initial printing scheme including: The natural language task requirements are analyzed using a large language model to extract the printing intent; Using a retrieval enhancement mechanism, a background knowledge set matching the printing intention is retrieved from the knowledge base, wherein the background knowledge set includes: material and process knowledge, equipment and historical experience, and multimodal image knowledge; The initial printing scheme is generated based on the background knowledge set, the equipment capabilities and current status of the bio-3D printing equipment, and the model features of the target 3D model.

3. The method according to claim 2, characterized in that, Based on the workflow steps defined in the standard operation skill template, the initial printing scheme is dynamically orchestrated to obtain operation instructions, which are then progressively issued to the bio-3D printing device via the model context protocol interface, including: The validity of the parameter configuration of the initial printing scheme is verified by using the preset process constraint rules and safety verification conditions in the standard operation skill template. After the verification is passed, the execution intention of the current workflow step is combined with the specific parameters in the initial printing scheme according to the workflow step sequence defined in the standard operation skill template to generate the operation instruction for the current work step. The model context protocol interface is invoked to convert the operation instructions of the current work step into a tool interface call request corresponding to the bio-3D printing device and send it out. After receiving the execution status feedback of the bio-3D printing device for the current work step, the workflow step is updated to drive the generation and sending of operation instructions for subsequent work steps.

4. The method according to claim 1, characterized in that, Acquiring multimodal monitoring data from the bio-3D printing equipment, and spatially aligning the multimodal monitoring data with the target 3D model and slice path for anomaly identification, includes: The multimodal monitoring data is obtained by acquiring real-time image data and sensor data collected by the bio-3D printing equipment. Based on the multimodal monitoring data, spatial alignment calculations are performed on the current nozzle position of the bio-3D printing device, the real-time image data, the outline of the target 3D model, and the current slice layer path. Based on the results of the spatial alignment calculation, anomalies are identified, including: over-extrusion anomaly, under-extrusion anomaly, localized wire breakage anomaly, and structural collapse anomaly.

5. The method according to claim 4, characterized in that, Dynamically adjusting the current print content and / or print parameters includes: Upon detecting the over-extrusion and under-extrusion anomalies, fine-tune the printing speed and / or extrusion pressure parameters according to the current deviation status. If the local filament breakage anomaly is detected, the detected filament breakage image location is reverse-mapped to the slicing path to determine the filament breakage start point and the missing path segment. The current printing task is paused and the printhead is moved to the filament breakage start point to reprint the missing path segment. If the structural collapse anomaly is detected, the unprinted subsequent slice layers are replanned, and boundary reinforcement paths are dynamically added to the subsequent slice layers while the local printing speed is reduced. For unrecoverable local defect areas, the printing path of the corresponding area can be skipped by modifying the underlying control commands.

6. The method according to claim 1, characterized in that, After generating the dynamic intervention strategy, the method further includes: The dynamic intervention strategy is subjected to parameter physical range verification, platform anti-collision boundary verification, and model printability verification. After the physical range of the parameters, the anti-collision boundary of the platform, and the printability of the model are verified, access control is performed according to the preset risk classification mechanism to execute the adjustment instructions corresponding to the dynamic intervention strategy. After executing the adjustment command, if the monitoring index exceeds the preset range, a dynamic rollback strategy is executed to roll back the printing status to the previous safety parameter or implement a safety pause.

7. The method according to claim 6, characterized in that, Automatic management is executed for low-risk fine-tuning instructions; For medium-risk instructions, the system will forcibly suspend the process and push a confirmation-pending solution, then execute the adjustment instructions based on the user's confirmation result. High-risk commands that exceed security limits will be directly intercepted and prohibited from being issued.

8. The method according to claim 1, characterized in that, The method further includes: After the printing task is completed, the initial printing plan, interface call logs, dynamic intervention strategies, multimodal monitoring data and user evaluations are merged into historical case data and stored in the knowledge base. The effectiveness of the dynamic intervention strategy is verified. If the strategy is effective, the corresponding strategy action is used as reusable experience and updated to the standard operation skill template of the corresponding material.

9. An intelligent control system for a bio-3D printing device, characterized in that, The system includes: The requirement response module is used to receive natural language task requirements, parse and reason about the natural language task requirements based on a pre-built knowledge base, and generate an initial printing scheme. The task generation module is used to dynamically arrange the instructions of the initial printing scheme to obtain operation instructions based on the workflow steps defined in the standard operation skill template, and gradually issue the operation instructions to the bio-3D printing device through the model context protocol interface. The monitoring and identification module is used to collect multimodal monitoring data of the bio-3D printing equipment during the execution of the operation instructions, and to spatially align the multimodal monitoring data with the target three-dimensional model and the slice path to identify anomalies. The optimization and adjustment module is used to generate a dynamic intervention strategy based on the knowledge base and the multimodal monitoring data when an abnormal development trend is detected or an anomaly is identified, and to instruct the bio-3D printing device to dynamically adjust the current printing content and / or printing parameters by issuing the dynamic intervention strategy.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1 to 8.

Images