A printhead multi-material switching method and system

Abstract

This invention discloses a method and system for multi-material switching in a printhead. The method includes: receiving a printing task instruction; generating a material switching demand vector based on the printing task instruction; activating the preheating mechanism of a multi-channel feeding system based on the material switching demand vector using a predictive preheating algorithm to dynamically preheat the material channel to be used, generating a preheating ready state signal; triggering the material switching action of the multi-channel feeding system using a parallel switching control algorithm based on the preheating ready state signal, generating a switching instruction sequence; responding to the switching instruction sequence, executing material supply switching, and adjusting the channel temperature in real time through an integrated cooling mechanism to maintain a stable material switching state. Using this invention, efficient collaborative switching of multiple material channels can be achieved, improving the printhead's response speed and temperature stability.

Description

Technical Field

[0001] This invention belongs to the field of printing technology, specifically a method and system for switching multiple materials in a printhead. Background Technology

[0002] In existing multi-material 3D printing technologies, the material switching process in the print head often faces problems such as response delay and unstable temperature control, leading to decreased print quality or material waste. Traditional switching methods typically employ sequential preheating and switching mechanisms, which are difficult to adapt to the differences in thermal properties of different materials, especially prone to temperature fluctuations or channel blockages during high-speed printing. Furthermore, the collaborative control of multi-material switching lacks dynamic optimization capabilities, failing to adjust preheating and cooling strategies in real time according to task requirements, thus limiting printing efficiency and accuracy. Summary of the Invention

[0003] The purpose of this invention is to provide a method and system for switching multiple materials in a printhead, in order to overcome the shortcomings of the prior art, achieve efficient collaborative switching of multiple material channels, and improve the response speed and temperature stability of the printhead.

[0004] One embodiment of this application provides a method for switching multiple materials in a printhead, the method comprising:

[0005] Receive a print task instruction including a material type sequence, switching point coordinates and material property parameters, and generate a material switching requirement vector based on the print task instruction. The material switching requirement vector encodes the material switching order, switching time point and material thermal property requirements.

[0006] Based on the material switching demand vector, the preheating mechanism of the multi-channel feeding system is activated by a predictive preheating algorithm to dynamically preheat the material channels to be used and generate a preheating ready state signal. The predictive preheating algorithm optimizes the preheating temperature curve in real time to match the thermal characteristics of different materials.

[0007] Using the preheating ready status signal, the material switching action of the multi-channel feeding system is triggered by the parallel switching control algorithm to generate a switching instruction sequence. The parallel switching control algorithm adopts a multi-threaded scheduling mechanism to simultaneously execute the operation of the material channel before cooling and the operation of activating the new material channel.

[0008] In response to the switching command sequence, the material supply is switched, and the channel temperature is adjusted in real time through an integrated cooling mechanism to maintain a stable material switching state. The integrated cooling mechanism combines temperature sensor feedback to adaptively adjust the cooling rate to maintain the printhead temperature within a set threshold.

[0009] Optionally, the step of receiving a print task instruction including a material type sequence, switching point coordinates, and material property parameters, and generating a material switching requirement vector based on the print task instruction, wherein the material switching requirement vector encodes the material switching order, switching time point, and material thermal characteristic requirements, including:

[0010] Parse the material type sequence in the print task instruction, extract the unique identifier for each material, and generate a material identification sequence;

[0011] Based on the switching point coordinates and the printhead motion trajectory model, calculate the precise timestamp of each material switching action and output the set of switching time points;

[0012] Query the material property database, quantify the material thermal property requirements into feature values, and generate a material thermal property vector.

[0013] The material identification sequence, the set of switching time points, and the material thermal property vector are encoded into a fixed-dimensional tensor, and the material switching demand vector is output.

[0014] Optionally, based on the material switching demand vector, the preheating mechanism of the multi-channel feeding system is activated through a predictive preheating algorithm to dynamically preheat the material channels to be used, generating a preheating ready state signal. The predictive preheating algorithm optimizes the preheating temperature curve in real time to match the thermal characteristics of different materials, including:

[0015] Extract the next set of thermal property vector subsets of the materials to be switched from the material switching demand vector;

[0016] Input a subset of thermal property vectors into a preheating prediction model built on LSTM, and output the initial preheating temperature curves for each material channel;

[0017] By combining real-time temperature feedback from the printhead, the initial preheating temperature curve is dynamically optimized using a gradient descent algorithm to generate a dynamic preheating temperature curve.

[0018] The PTC heating element of the multi-channel feeding system is controlled according to the dynamic preheating temperature curve to dynamically preheat the material channel corresponding to the material to be switched.

[0019] When the difference between the preheated channel temperature and the target temperature enters the preset tolerance range, the preheating ready state signal is triggered.

[0020] Optionally, the step of using the preheating ready state signal to trigger the material switching action of the multi-channel feeding system through a parallel switching control algorithm to generate a switching command sequence, wherein the parallel switching control algorithm adopts a multi-threaded scheduling mechanism to synchronously execute the operations of the material channel before cooling and the activation of the new material channel, including:

[0021] Receive the preheating-ready status signal and activate the multi-threaded scheduling controller;

[0022] Create two threads in the controller, where thread 1 executes a cooling command sequence for the current material channel, and thread 2 executes an activation command sequence for the target new material channel to be switched to.

[0023] Synchronize the dual-thread states using shared memory and generate channel synchronization status codes;

[0024] When the channel synchronization status code reaches the preset threshold, the piezoelectric ceramic valve is triggered to open and close, and a material switching command sequence is output.

[0025] Optionally, in response to the switching command sequence, the material supply is switched, and the channel temperature is adjusted in real time through an integrated cooling mechanism to maintain a stable material switching state. The integrated cooling mechanism, combined with temperature sensor feedback, adaptively adjusts the cooling rate to maintain the printhead temperature within a set threshold, including:

[0026] The switching instruction sequence is analyzed, and the stepper motor is driven to perform the material supply path switching action.

[0027] After the material switching is completed, the real-time temperature distribution map of the printhead is collected through the microchannel temperature sensor array;

[0028] Input the real-time temperature distribution map into the PID-fuzzy composite controller to calculate the cooling rate compensation parameters;

[0029] The current of the semiconductor cooling chip is adjusted according to the cooling rate compensation parameter to maintain the printhead temperature within the set threshold.

[0030] Another embodiment of this application provides a printhead multi-material switching system, the system comprising:

[0031] The receiving module is used to receive a printing task instruction including a material type sequence, switching point coordinates and material property parameters, and generate a material switching requirement vector based on the printing task instruction. The material switching requirement vector encodes the material switching order, switching time point and material thermal property requirements.

[0032] The preheating module is used to activate the preheating mechanism of the multi-channel feeding system based on the material switching demand vector and through a predictive preheating algorithm to dynamically preheat the material channel to be used and generate a preheating ready status signal. The predictive preheating algorithm optimizes the preheating temperature curve in real time to match the thermal characteristics of different materials.

[0033] The generation module is used to trigger the material switching action of the multi-channel feeding system by means of the preheating ready state signal and the parallel switching control algorithm to generate a switching instruction sequence. The parallel switching control algorithm adopts a multi-threaded scheduling mechanism to synchronously execute the operation of the material channel before cooling and the operation of activating the new material channel.

[0034] The switching module is used to respond to the switching command sequence, perform material supply switching, and adjust the channel temperature in real time through an integrated cooling mechanism to maintain a stable material switching state. The integrated cooling mechanism combines temperature sensor feedback to adaptively adjust the cooling rate to maintain the printhead temperature within a set threshold.

[0035] Another embodiment of this application provides a storage medium storing a computer program, wherein the computer program is configured to execute the method described in any of the preceding claims when running.

[0036] Another embodiment of this application provides an electronic device including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the method described in any of the preceding claims.

[0037] Compared with existing technologies, the present invention provides a multi-material switching method for printheads, which receives printing task instructions and generates a material switching demand vector based on the printing task instructions; based on the material switching demand vector, a predictive preheating algorithm is used to activate the preheating mechanism of the multi-channel feeding system to dynamically preheat the material channel to be used, generating a preheating ready state signal; using the preheating ready state signal, a parallel switching control algorithm is used to trigger the material switching action of the multi-channel feeding system, generating a switching instruction sequence; responding to the switching instruction sequence, the material supply switching is executed, and the channel temperature is adjusted in real time through an integrated cooling mechanism to maintain a stable material switching state, thereby achieving efficient collaborative switching of multiple material channels and improving the response speed and temperature stability of the printhead. Attached Figure Description

[0038] Figure 1 A hardware structure block diagram of a computer terminal for a multi-material switching method for a printhead provided in an embodiment of the present invention;

[0039] Figure 2 This is a flowchart illustrating a method for switching multiple materials in a printhead according to an embodiment of the present invention.

[0040] Figure 3 This is a schematic diagram of a multi-material switching system for a printhead provided in an embodiment of the present invention. Detailed Implementation

[0041] The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0042] This invention first provides a method for switching multiple materials in a printhead, which can be applied to electronic devices, such as computer terminals, specifically ordinary computers.

[0043] The following detailed explanation uses a computer terminal as an example. Figure 1 This is a hardware structure block diagram of a computer terminal for a multi-material switching method for a printhead provided in an embodiment of the present invention. Figure 1 As shown, the computer device includes a processor, memory, and network interface connected via a system bus, wherein the memory may include non-volatile storage media and internal memory.

[0044] The non-volatile storage medium can store an operating system and a computer program. This computer program includes program instructions that, when executed, cause the processor to perform any printhead multi-material switching method.

[0045] The processor provides computing and control capabilities, supporting the operation of the entire computer device.

[0046] The internal memory provides an environment for the execution of computer programs in non-volatile storage media, which, when executed by a processor, enable the processor to perform any printhead multi-material switching method.

[0047] This network interface is used for network communication, such as sending assigned tasks. Those skilled in the art will understand that... Figure 1 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 computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0048] It should be understood that the processor can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among these, a general-purpose processor can be a microprocessor or any conventional processor.

[0049] See Figure 2 The present invention provides a method for switching multiple materials in a printhead, which may include the following steps:

[0050] S201, receive a printing task instruction including a material type sequence, switching point coordinates and material property parameters, and generate a material switching requirement vector based on the printing task instruction, wherein the material switching requirement vector encodes the material switching order, switching time point and material thermal property requirements;

[0051] Specifically, it can parse the material type sequence in the print task instruction, extract the unique identifier of each material, and generate a material identification sequence;

[0052] The print job command received by the system is a structured data packet, one of the core components of which is the Material Type Sequence (MTS). This sequence explicitly lists all the material types required by the print head at different printing stages, in the order they are printed. The parsing process is first executed by the Command Parsing Module (CPM) of the print control system. The CPM scans the entire command data packet and locates the MTS field. This field typically consists of a string of material names arranged in sequence, such as ["ABS_Black", "TPU_90A_Transparent", "PC_GF30"]. Internally, the system maintains a global Material Registry (MR), which is a database or configuration file that records all permitted materials and their corresponding unique identifiers. During parsing, the CPM reads the material name strings from the MTS item by item and then queries the MR. During the query, the system not only performs exact matching of the name strings but may also perform fuzzy matching or synonym processing (e.g., the user enters "Nylon 12" but the registry records it as "PA12"). Once a match is successful, the system extracts the material's unique identifier (UID) from the MR. This UID is a fixed-length code, such as a 128-bit UUID (Universally Unique Identifier), or a more concise but system-unique numeric / alphabetic combination code (such as MAT001). The UID is designed to ensure unambiguous reference to a specific material formulation and batch attribute, even in cases of similar material names or the existence of aliases.

[0053] The process of extracting the UID is not a simple table lookup. The system performs deep material authentication. For example, when parsing "TPU_90A_Transparent", the system not only extracts its basic UID (such as "TPU90A-TRANS-001"), but also queries the more detailed Material Attribute Database (MAD) based on the batch number (BN) or supplier code (SC) information that may be included in the print job instruction. The MAD stores the specific attribute certificates and measured parameters of the material in that batch. If the job instruction specifies a particular batch, the system verifies whether the batch of material is registered in the system and is available (e.g., not expired, sufficient stock). After successful verification, the extracted UID is actually a precise identifier bound to that specific batch (e.g., "TPU90A-TRANS-001-Batch#20240515-SupplierX"). This binding ensures that the material attribute data used in subsequent processes is absolutely accurate. All extracted UIDs are stored sequentially in an ordered list according to their original order in the MTS. This list is the generated Material Identification Sequence (MIS). The MIS is the basis for all subsequent material switching operations. The system records the MIS generation log, including the parsing time for each material, the matched UID, batch information, and any parsing warnings or errors (such as unregistered material names), for operator review or system self-diagnosis.

[0054] To improve robustness, the system also incorporates a fault-tolerance mechanism. If a material name not registered in the Material Name Mapping Table (MR) appears in the Material Type Switch (MTS), the Printing Manager (CPM) will attempt to infer it based on preset rules or trigger an alarm. For example, the system may have a Material Alias ​​Mapping Table (MAMT) that maps common aliases or abbreviations to standard names. If the inference fails, the system will send an "Unknown Material Error" (UME) alarm to the operator's console and suspend task parsing. The operator can choose to temporarily register the material in the MR (providing the necessary attribute parameters) or modify the print task instructions. Once all materials are successfully parsed and UIDs are generated, the final Material Name Switching Table (MIS) is encapsulated into a data structure, typically an array or linked list, with the element order strictly corresponding to the printing order. This MIS data structure is passed to the next stage of the process, becoming one of the core inputs driving the entire multi-material switching process. The system generates a unique Task ID (TID) for each print task and binds the MIS to this TID for storage, enabling end-to-end traceability.

[0055] Based on the switching point coordinates and the printhead motion trajectory model, calculate the precise timestamp of each material switching action and output the set of switching time points;

[0056] Another key parameter in the print job command is the Switch Point Coordinates (SPC). The SPC is an ordered list specifying the exact location in the 3D printing space (usually an X, Y, Z coordinate system) where the print head needs to switch materials, such as [(X1, Y1, Z1), (X2, Y2, Z2), ...]. These coordinates are typically calculated by slicing software based on the model geometry and predefined material regions. After obtaining the SPC list, the system needs to combine it with the print head's Motion Trajectory Model (MTM) to calculate the precise moment the material switching action occurs, i.e., the Precise Timestamp (PTS). The MTM is a mathematical model or data table describing the print head's motion. It includes dynamic parameters such as the print head's maximum acceleration (MA), maximum velocity (MV), and jerk in each axis, as well as the planned print path (PP). A print head (PP) typically consists of a series of continuous line segments (G01 instructions) or spline curves (such as NURBS), defining the detailed movement path of the print head from the start point to the end point.

[0057] The core of calculating accurate timestamps is Kinematic Forward Prediction (KFP). The system uses the dynamic parameters in the MTM (Motion Module) and the planned PP (Printer Probe) path to simulate the time process of the print head moving from the start of the task or the previous switch point to the current SPC (Site Module). The specific calculations are performed by the Motion Control Card (MCC) or the Trajectory Planning Module (TPM) on the host computer. The calculation process considers the following key factors:

[0058] Path Length (PL): Calculates the actual curve distance along the PP path from the current position (or starting point) to the target SPC point.

[0059] Velocity Profile (VP): Based on the MA, MV, Jerk parameters in MTM and preset segmented speed limits (such as corner deceleration), it generates an ideal speed-time curve from start-up (which may not be zero speed), acceleration, constant speed cruising, to deceleration and stopping at the SPC point.

[0060] Real-time Status (RTS): The calculation takes into account the current actual position (CP) and velocity (CV) of the printhead, which are usually provided by real-time feedback from the encoder.

[0061] Based on these inputs, the system uses numerical integration (such as displacement integration under trapezoidal or S-shaped velocity curves) or analytical methods to accurately calculate the Time of Arrival (TOA) required for the printhead to reach each SPC point. This TOA is an absolute time value relative to the start time of the printing task (T=0). This time value is the Material Switching Time (PTS) corresponding to that SPC point. The system iterates through all SPC points, calculating a PTS for each point.

[0062] The calculated PTS values ​​are collected into an ordered list, forming a SwitchTimepoint Set (STS). Each element in the STS is typically a structure containing the index (Index, IDX) of the SPC point, its three-dimensional coordinates (X, Y, Z), and the calculated PTS (units are usually milliseconds (ms) or microseconds (μs), depending on the system's accuracy requirements). Due to unforeseen minor disturbances during printing (such as motor step loss or platform vibration), the system typically introduces a safety margin (SM), for example, triggering the switching action a few milliseconds in advance to ensure that the switching is completed before reaching the precise coordinate point. The calculation results of the STS are correlated and verified with the previously generated MIS: each switching point corresponds to one switch in the material sequence in the MIS (i.e., switching from one material to the next). The system checks whether the number of elements in the STS is equal to the number of materials in the MIS minus one (because there are N-1 switches for N materials) and ensures that the order is consistent. The STS is finally encapsulated and output, and passed to the preheating and switching control module, becoming the time reference for triggering preheating and switching actions. The system records STS calculation logs, including the MTM parameters used, the calculation model, PL and VP details for each point, and the final PTS value, for debugging and optimization.

[0063] Query the material property database, quantify the material thermal property requirements into feature values, and generate a material thermal property vector.

[0064] The system possesses a core Material Attribute Database (MAD). The MAD is a structured storage system (like a relational database SQL or a time-series database) that stores detailed physical, chemical, and process property parameters for all registered materials (identified by their UIDs). For multi-material printing, especially when dealing with thermoplastics (such as PLA, ABS, PETG) or photosensitive resins, critical Material Thermal Property Requirements (MTPRs) are fundamental to successful switching and printing. These thermal properties typically include, but are not limited to:

[0065] Glass transition temperature (Tg): The critical temperature at which a material transitions from a glassy state to a rubbery state. It is crucial for preheating and cooling settings. Unit: degrees Celsius (°C).

[0066] Melting Temperature (Tm) or Recommended Printing Temperature (RTT): The nozzle / barrel temperature required for the material to melt and flow or reach the optimal printing viscosity, in degrees Celsius (°C).

[0067] Thermal Decomposition Temperature (Td): The upper limit of the temperature at which a material begins to undergo irreversible decomposition. This is to prevent overheating. The unit is degrees Celsius (°C).

[0068] Specific heat capacity (Cp): The amount of heat required to raise the temperature of a unit mass of material by one degree Celsius. It affects the preheating rate and energy demand. The unit is joules per kilogram of Kelvin (J / (kg·K)).

[0069] Thermal conductivity (TC): The ability of a material to conduct heat, which affects the internal temperature uniformity and cooling rate of the material. The unit is watts per meter Kelvin (W / (m·K)).

[0070] Viscosity-Temperature Coefficient (VTC): Describes the sensitivity of a material's viscosity to temperature changes, affecting its flowability and switching response.

[0071] Based on the material UID in the previously generated MIS, the system initiates a batch query (BQ) to the MAD to request the values ​​of the aforementioned key thermal characteristic parameters corresponding to these materials.

[0072] The query returns the raw parameter values. To facilitate subsequent processing by the preheating prediction model and control algorithms, these parameters, which have different physical meanings and dimensions, need to be quantized and combined into a unified, machine-readable feature set, i.e., feature values ​​(FV). The quantization process may include:

[0073] Normalization (Norm): Scaling values ​​of different dimensions and ranges to a uniform numerical range (such as 0 to 1 or -1 to 1). For example, temperature values ​​(Tg, Tm, Td) can be normalized based on the maximum allowable operating temperature of the system (MaxWorking Temp, MWT): Norm_Temp = (Temp_Value - Min_Temp) / (MWT -Min_Temp).

[0074] Feature Engineering (FE): Sometimes it is necessary to calculate derived features based on original parameters. For example, calculating the safe operating range of a material: Safe_Range = Td - RPT, and normalizing it. Or calculating the relative energy required for preheating: Relative_Energy = Cp * (RPT - Ambient_Temp), where Ambient_Temp is the ambient temperature.

[0075] Discretization / Binning (Bin): For certain continuous properties, they may be divided into several discrete levels (such as low, medium, and high thermal conductivity) based on experience and assigned numerical labels.

[0076] The system performs the above quantization operation for each material in the MIS sequence, ultimately generating a set of features containing a fixed number of features for each material. The selection of these features and the quantization rules are predefined and stored in the system's configuration file.

[0077] After quantization, the system arranges the set of feature values ​​corresponding to each material in a predefined, fixed order, forming an ordered numerical array. This array is the Material Thermal Property Vector (MTPV) for that material. For example, a simplified MTPV might be defined as: [Norm_Tg, Norm_Tm, Norm_Td, Norm_Cp, Norm_TC, Norm_VTC, Norm_Safe_Range]. Each element (dimension) in the vector corresponds to a specific thermal property. The MTPVs of all materials are collected in the order they appear in the MIS. Finally, the system outputs a Sequence of Vectors (SoV), which is the sequence of thermal property vectors for the new material (i.e., the material to be used after the switch) at each switching point. This SoV is the key input to the predictive preheating algorithm (subsequent steps), accurately describing the core temperature control requirements of each material to be used in a structured and numerical manner. The system records the MTPV generation process, including the original parameter values ​​queried, the applied quantization formula, and the final vector value, ensuring the traceability of the process.

[0078] The material identification sequence, the set of switching time points, and the material thermal property vector are encoded into a fixed-dimensional tensor, and the material switching demand vector is output.

[0079] At this point, the system has generated three core data components:

[0080] Material Identifier Sequence (MIS): An ordered list containing unique identifiers (UIDs) for the materials used in each printing stage;

[0081] Set of material switching time points (STS): An ordered list containing the precise timestamp (PTS) of each material switching action and its corresponding spatial coordinates (SPC);

[0082] Material Thermal Properties Vector (MTPV): An ordered sequence (corresponding to the MIS order), where each element is a vector containing the key quantitative thermal property features of the corresponding material.

[0083] To facilitate efficient processing by subsequent predictive preheating and parallel switching control algorithms, these three different types and structures of data need to be integrated and encoded into a unified, structured, fixed-size data container. This container is the Material Switch Demand Vector (MSDV). Although named a "vector," in multidimensional data processing, it is typically represented as a fixed-dimensional tensor (FDT). A tensor is a generalized concept of multidimensional arrays (scalars are 0-dimensional tensors, vectors are 1-dimensional tensors, and matrices are 2-dimensional tensors).

[0084] The key to the encoding process is designing a structured tensor format that can hold all the necessary information without loss. A typical design is to construct MSDV as a three-dimensional tensor:

[0085] Dimension 1: Sequence Index (Dim1_SI): This dimension corresponds to the stage sequence of the printing task. Its length equals the total number of material stages (TMP) used in the printing task, i.e., the length of the MIS. For example, if the task uses 3 materials (with 2 switching), then TMP = 3. Each index position i in Dim1_SI represents the i-th material stage in the task.

[0086] Dimension 2: Feature Slot (Dim2_FS): This dimension corresponds to the "slots" of all the different types of data that need to be stored at each material stage. The number and meaning of the slots are fixed. For example, a design might contain the following slots:

[0087] Slot 0: Material UID (stored as an integer hash value or an embedding vector);

[0088] Slot 1: Stage start time (i.e., the PTS when the material starts printing in this stage; it is 0 for the first stage and the PTS of the corresponding switching point in the STS for subsequent stages).

[0089] Slot 2: Stage End Time (i.e., the PTS when the material in this stage is finished printing; for the last stage, it is the total task time; for the other stages, it is the PTS of the next switching point in the STS).

[0090] Slot 3-9: The eigenvalues ​​(such as Norm_Tg, Norm_Tm, ...) of the thermal property vector (MTPV[i]) of the material in this stage;

[0091] Slot 10-12: Coordinates (X, Y, Z) of the switch point that occurs after this phase (fill with default values ​​such as NaN or 0 if a switch point exists);

[0092] Slot 13: Switch Action Flag (SAF), 0 indicates no switch, 1 indicates a switch (i.e., not the final stage).

[0093] Dimension 3: Batch / Redundancy (Optional) (Dim3_BR): This dimension can be expanded if the system supports processing multiple tasks at once or requires redundant storage. For a single task, this dimension is typically 1 in size.

[0094] The system fills the corresponding slot positions in the tensor with MIS[i], STS[i] (or the time calculated from it), MTPV[i], and related SPC coordinates in the order of Dim1_SI (i.e., the material stage order). For stages without switching points (the first and last stages), the switching point coordinate slots are filled with specific values ​​(such as NaN).

[0095] After encoding, the output MSDV is a tensor data structure with a fixed dimension of [TMP, Number_of_Slots, 1] (or [TMP, Number_of_Slots] if the third dimension is omitted). This fixed-dimensional design is crucial, ensuring that subsequent algorithms (such as LSTM-based preheating prediction models) can receive inputs of a uniform specification. The MSDV contains all the core information needed to drive the entire multi-material switching process:

[0096] What materials are used? (Through UID and MTPV)

[0097] When to switch? (Based on phase start / end time and switch flag SAF);

[0098] Where do I switch? (By switching point coordinates);

[0099] What are the temperature requirements for the material? (via MTPV)

[0100] This MSDV tensor is passed to the predictive preheating algorithm module as its core input, triggering subsequent preheating and switching control processes. The system records the MSDV encoding rules and the final generated tensor content (or stores its references) as a critical snapshot of the task execution state. This structured encoding method greatly improves the efficiency and reliability of the system in handling complex multi-material tasks.

[0101] This method first receives printing instructions containing key parameters for material switching. Through structured processing, it transforms discrete printing requirements into a unified mathematical expression. The material switching requirement vector, serving as the core input for subsequent processes, not only records the spatiotemporal information of material switching but also quantifies the thermodynamic requirements of different materials. This lays the data foundation for intelligent preheating and precise switching, realizing the transformation of printing task instructions into computable data and solving the problems of low efficiency and error-proneness in manual analysis in traditional methods. Through vectorized encoding, complex material switching requirements are transformed into a standardized format that can be processed by machines, significantly improving the planning efficiency and accuracy of multi-material printing.

[0102] S202, based on the material switching demand vector, the preheating mechanism of the multi-channel feeding system is activated by a predictive preheating algorithm to dynamically preheat the material channel to be used and generate a preheating ready state signal. The predictive preheating algorithm optimizes the preheating temperature curve in real time to match the thermal characteristics of different materials.

[0103] Specifically, a subset of the thermal property vectors of the next set of materials to be switched can be extracted from the material switching demand vector;

[0104] Once the system receives the fully encoded Material Switching Demand Vector (MSDV), this vector is a structured data container that integrates the material identifier sequence, the set of switching time points, and the material thermal property vector. The system first locates the next critical node in the current printing process—the Switch Time Point (STP) closest to the current time. Using a timestamp comparison algorithm (TCA), the system scans the Switch Time Point Set (STPS) stored in the MSDV to identify the next STP to be triggered and its corresponding Material Type Identifier (MTI). For example, if the current printing time is T=125.3 seconds, and the STPS contains time points T1=130.5 seconds (corresponding to material A switching to B) and T2=142.1 seconds (B switching to C), then T1=130.5 seconds is identified as the next switching point. Based on the MTI (e.g., "Material_B"), the system extracts a complete subset of thermal characteristics (STC) associated with "Material_B" from the Material Thermal Characteristic Vector (MTCV) region of the MSDV. This STC typically contains multiple quantified parameters, such as the material's glass transition temperature (Tg), melting temperature (Tm), specific heat capacity (Cp), thermal conductivity (k), and recommended printing temperature ranges for the material (e.g., Min_Print_Temp=195°C, Max_Print_Temp=220°C).

[0105] The extraction process is efficiently completed using the Memory-Mapped Access Technique (MMAT). The MSDV is organized in system memory as a multi-dimensional array structure, where the Material Identifier Sequence (MIS) and the Thermal Characteristic Vector (MTCV) are strictly aligned using an index key (IK). Once the target MTI is located, the system directly accesses the corresponding data block (DB) of the MTCV based on its index value (e.g., Index=2). The STC stored in this data block is not the original parameter value, but rather the feature value (FV) after normalization. For example, an actual Tm=205°C might be normalized to 0.82 (assuming a baseline range of 150-250°C), and k=0.25 W / m·K is converted to 0.15 (baseline range 0.1-1.0 W / m·K). This normalization process eliminates dimensional differences, facilitating unified processing by subsequent algorithms. The extracted STCs are encapsulated into a Thermal Characteristic Vector Subset (TCVS), with dimensions matching the number of material properties (e.g., 5 dimensions if there are 5 feature values). The system also records the Material Channel Number (MCN) associated with the TCVS, for example, channel 3 corresponds to "Material_B", ensuring that the preheating operation is accurately located to the physical hardware unit.

[0106] To ensure real-time performance, the extraction operation is executed in a Hard Real-Time Task Thread (HRT-TT). This thread has the highest scheduling priority, ensuring that the TCVS retrieval and encapsulation are completed within milliseconds. The system also performs a Data Integrity Check (DIC): verifying that the extracted TCVS has not been tampered with or transmitted incorrectly using a Cyclic Redundancy Check Code (CRCC). If the check fails, the system immediately triggers an exception handling process (such as re-extraction or an alarm); if successful, the TCVS is pushed to the input buffer (IB) of the Preheating Prediction Module (PPM) to prepare for subsequent preheating calculations. At this point, the system has accurately acquired all the key thermal property quantification data for the next material to be switched, laying the data foundation for dynamic preheating.

[0107] Input a subset of thermal property vectors into a preheating prediction model built on LSTM, and output the initial preheating temperature curves for each material channel;

[0108] The received thermal characteristic vector subset (TCVS) is fed into the core computing unit—the LSTM-based Preheating Prediction Model (LPPM). LSTM (Long Short-Term Memory) is a special type of recurrent neural network (RNN) that excels at processing time-series data. This LPPM model is trained using massive amounts of historical printing task data. Its network structure includes an input layer (IL), multiple LSTM hidden layers (HL), and a fully connected output layer (FCOL). The input layer receives fixed-dimensional TCVS (e.g., 5-dimensional vectors). Each LSTM hidden layer consists of several memory cells (MC), and each MC contains an input gate (IG), a forget gate (FG), and an output gate (OG), learning long-term dependencies through a gating mechanism. For example, when the input TCVS contains a high melting temperature (Tm) and low thermal conductivity (K), the model can "memorize" that such materials typically require longer preheating times and higher starting temperatures.

[0109] The LPPM processing flow for TCVS is as follows: First, the input vector is converted into a Time Step Sequence (TSS). Since TCVS is a static feature, the system uses Virtual Time Unfolding (VTU) to replicate it into a fixed-length sequence (e.g., 10 time steps) to simulate the time evolution process. This sequence is then sequentially input into the LSTM hidden layer. At each time step, the LSTM unit updates the activation values ​​(AV) of IG, FG, and OG based on the current input and the hidden state (HS) and cell state (CS) of the previous time step, thus determining which information to retain or discard. Finally, the HS of the last time step is passed to the fully connected output layer (FCOL). FCOL is a Multilayer Perceptron (MLP), and the number of its output neurons corresponds to the number of key parameters in the warm-up curve. For example, if the preheating curve is defined as a three-stage heating process (initial heating period, isothermal holding period, and final adjustment period), the output layer can be designed with 6 neurons, representing: the starting temperature (T_start), the first-stage heating rate (Ramp_Rate1), the first-stage duration (t1), the holding temperature (T_hold), the holding time (t_hold), and the second-stage heating rate (Ramp_Rate2). These parameters together constitute the Initial Preheating Temperature Curve (IPTC).

[0110] The model inference process is executed on dedicated Neural Network Accelerator (NNA) hardware to ensure low latency (e.g., <5ms). Taking an input TCVS=[0.82, 0.15, ...] (corresponding to high Tm, low k) as an example, LPPM might output IPTC: T_start=180°C, Ramp_Rate1=3°C / s, t1=8s, T_hold=200°C, t_hold=5s, Ramp_Rate2=1°C / s. This curve means: starting from 180°C, the temperature is increased at a rate of 3°C / s for 8 seconds to 204°C (180+3×8), held at 200°C for 5 seconds (due to the material's thermal inertia requiring stabilization), and then finely increased at 1°C / s to the target temperature of 205°C. IPTC is encapsulated as a data structure, associated with the target material channel number (MCN), and pushed to the Dynamic Optimization Module (DOM). Meanwhile, the model confidence (MC, ranging from 0 to 1) is recorded and used for subsequent optimization decisions.

[0111] By combining real-time temperature feedback from the printhead, the initial preheating temperature curve is dynamically optimized using a gradient descent algorithm to generate a dynamic preheating temperature curve.

[0112] The initial preheating temperature profile (IPTC) is only a theoretical prediction and needs to be corrected based on the real-time state of the physical system. The system acquires real-time temperature feedback (RTTF) through a micro thermocouple array (MTA) embedded in the printhead. The MTA contains multiple (e.g., 8) Type K thermocouples distributed in key hot zones (such as the nozzle wall, near the heating block, and material flow channels) to continuously measure temperature values ​​(TV) at a high sampling rate (e.g., 10Hz). These TVs are noise-reduced by a signal conditioning circuit (SCC) and then aggregated by the central processing unit (CPU) into a printhead temperature distribution vector (PTDV), for example, [198.2°C, 195.7°C, 201.1°C].

[0113] Dynamic optimization employs the Gradient Descent Algorithm (GDA). The algorithm aims to find the optimal preheating curve parameters that minimize the error between the predicted and actual temperatures. The loss function (LF) is defined as the mean squared error (MSE) between the predicted temperature (PT) and the actual temperature (AT) at each measurement point. The optimization process is as follows:

[0114] Initialization: Set the IPTC parameters (such as T_start, Ramp_Rate1, etc.) to the initial point θ0;

[0115] Forward calculation: Calculate the predicted temperature PT based on θ0 and thermodynamic models (such as discretization of the finite difference heat conduction equation);

[0116] Loss calculation: Compare PT with the current AT (e.g., PTDV mean) to calculate the LF value;

[0117] Gradient calculation: Calculate the partial derivative (PD) of LF with respect to each parameter in θ0 using the numerical differentiation method (NDM) or automatic differentiation (AD), which is the gradient vector ∇L;

[0118] Parameter update: Update the parameters according to the formula θ1 = θ0 - η × ∇L. Where η is the learning rate (LearningRate, LR, such as 0.01), which controls the step size.

[0119] For example, if the current AT = 198°C, but IPTC predicts it should be 205°C (LF is large), and the gradient shows that increasing Ramp_Rate1 can reduce LF, then the algorithm will increase the value of Ramp_Rate1. Iterate through steps 2-5 until LF is below the threshold or the maximum number of iterations (e.g., 20 times) is reached.

[0120] The optimization process is completed within the task cycle of the Real-Time Operating System (RTOS) (e.g., executed every 100ms). Updated curve parameters θ1 are generated after each iteration. When optimization terminates (e.g., LF < 0.5°C² or after 5 iterations), the final parameter set is used to construct the Dynamic Preheating Temperature Curve (DPTC). This curve not only includes the temperature setpoint but also incorporates a Dynamic Compensation Factor (DCF), such as a heating rate correction coefficient for the current printhead thermal inertia (e.g., Ramp_Rate1 = 3°C / s, adjusted to 3.2°C / s after optimization). The DPTC is converted into a Time-Temperature Command Sequence (TTCS), such as [(t=0s, T=180°C), (t=2s, T=186°C), ...], and associated with the target MCN. Thus, the system obtains a personalized preheating curve that integrates material property prediction and real-time thermal state feedback.

[0121] The PTC heating element of the multi-channel feeding system is controlled according to the dynamic preheating temperature curve to dynamically preheat the material channel corresponding to the material to be switched.

[0122] The execution of the Dynamic Preheating Temperature Profile (DPTC) relies on the Positive Temperature Coefficient Heater (PTCH) element in the Multi-Channel Feeding System (MCFS). The PTC is a self-limiting temperature heater whose resistance increases with temperature, providing overheat protection. Each material channel (e.g., channel 3) is equipped with an independent PTCH, installed close to the Material Flow Channel (MFC). The system generates a Temperature Control Command (TCC) based on the DPTC and drives the PTCH via a Pulse Width Modulation (PWM) signal. PWM is a method of controlling average power by adjusting the pulse width (the proportion of the high-level duration to the cycle, i.e., the duty cycle (DC)). For example, DC=70% means 70% of the time is powered on and 30% of the time is unpowered.

[0123] The control process achieves closed-loop feedback:

[0124] Target setting: The control system (such as a PID controller) reads the target temperature (TargetTemperature, TT) at the current moment in the DPTC. For example, at t = 3s, TT = 189°C.

[0125] Temperature acquisition: A dedicated thermocouple (Dedicated Thermocouple, DT) near the channel PTCH measures the actual temperature (Actual Channel Temperature, ACT).

[0126] Error calculation: Calculate the temperature error (Temperature Error, TE) = TT - ACT.

[0127] PID output: The PID controller (Proportional-Integral-Differential controller) calculates the required power percentage (Required Power Percentage, RPP) based on TE and its historical changes. The PID parameters (proportional gain Kp, integral time Ti, derivative time Td) are pre-tuned.

[0128] PWM generation: Convert RPP to the PWM duty cycle (DC). For example, RPP = 80% corresponds to DC = 80%.

[0129] Power drive: The PWM signal drives the PTCH to work through a Metal-Oxide-Semiconductor Field-Effect Transistor (MOSFET) switching circuit.

[0130] Taking the heating-up stage as an example: If ACT = 185°C < TT = 189°C, TE = +4°C, the PID outputs a high RPP (such as 95%), DC = 95%, and the PTCH heats up at full power; when ACT approaches TT, TE decreases, the PID reduces RPP, DC decreases, and the heating slows down until it stabilizes.

[0131] "Dynamic preheating" is reflected in the system's real-time tracking of the time-varying target of the DPTC. The DPTC may require non-linear heating (such as fast first and then slow), and the PID controller can quickly respond to changes in the target value. At the same time, the system monitors indirect indicators of the material state (such as the readings of the channel pressure sensor). If it detects that the material has poor fluidity (pressure increases) due to insufficient preheating, it will temporarily fine-tune the DPTC (such as increasing TT in advance). The temperature control of all channels is independent and executed in parallel, ensuring that when one channel is preheating, other channels can remain standby or cooling states. The preheating process data (ACT, DC, TT) are recorded in real time for subsequent analysis and model training.

[0132] When the difference between the preheated channel temperature and the target temperature enters the preset tolerance range, a preheating ready status signal is triggered.

[0133] The criterion for judging the completion of preheating is whether the difference (Temperature Difference, TD) between the actual channel temperature (ACT) and the current target temperature (TT) in the dynamic preheating temperature curve (DPTC) has stably entered the preset tolerance range (Preset Tolerance Range, PTR). PTR is usually a small interval centered on TT, such as TT ± ΔT, where ΔT is the tolerance threshold (Tolerance Threshold, TT), and its value is set according to the thermal sensitivity of the material (e.g., for high-precision materials, ΔT = ±0.5°C, and for ordinary materials, ΔT = ±1.5°C). The system continuously calculates TD = ACT - TT and monitors its absolute value |TD|.

[0134] To avoid false triggering caused by short-term fluctuations, the system introduces a stable time condition (Stable Time Condition, STC). That is, it is required that the state of |TD| ≤ TT lasts for a preset minimum stable time (Minimum Stable Time, MST, such as 3 seconds). At the same time, the system calculates the sliding window standard deviation (Sliding Window Standard Deviation, SWSD) of TD to evaluate the temperature stability. If within MST, SWSD is also lower than the fluctuation threshold (Fluctuation Threshold, FT, such as 0.2°C), the temperature is considered truly stable. For example: the target TT = 205°C, ΔT = 1°C, then PTR is [204°C, 206°C]. If ACT continuously remains between 204.2°C and 205.8°C within 3 seconds (satisfying |TD| ≤ 1°C), and SWSD = 0.15°C < FT = 0.2°C, the determination condition is met.

[0135] When the above conditions are met simultaneously, the system immediately generates a Preheating ReadyStatus Signal (PRSS). This signal is a high-level digital signal (such as TTL 5V) or a specific data packet (DP) containing key information: the ready channel number (MCN), the ready timestamp (TS), and the achieved temperature (AT). The PRSS is sent to the Parallel Switching Control Module (PSCM) to trigger subsequent material switching actions. Simultaneously, the system updates the Preheating Status Register (PSR), marking the corresponding channel as "ready," and displays a visual prompt on the Human Machine Interface (HMI) (e.g., the channel icon turns green). If preheating timeout occurs (e.g., exceeding the DPTC maximum time limit) or temperature abnormality occurs (e.g., ACT exceeds the safe range), a Fault Alarm Signal (FAS) is triggered, aborting the switching process.

[0136] An intelligent predictive algorithm is employed for proactive temperature management of the material channel, dynamically adjusting the preheating strategy based on the material's thermal properties. This step learns the printhead temperature change trend in real time, customizing optimal heating curves for different materials. This avoids the overheating or underheating problems caused by traditional fixed preheating modes, overcoming the limitations of traditional uniform preheating and achieving precise matching of material properties. Through dynamic temperature control, material flowability requirements are ensured while avoiding energy waste, providing a stable thermal environment for high-quality multi-material printing.

[0137] S203, using the preheating ready state signal, trigger the material switching action of the multi-channel feeding system through the parallel switching control algorithm to generate a switching instruction sequence, wherein the parallel switching control algorithm adopts a multi-thread scheduling mechanism to synchronously execute the operation of the material channel before cooling and the operation of activating the new material channel;

[0138] Specifically, it can receive a preheating-ready status signal and activate the multi-threaded scheduling controller;

[0139] When the multi-channel feeding system receives the Preheat Ready Status Signal (PRSS), generated by the predictive preheating module, it indicates that the target new material channel has been precisely heated to the preset tolerance range (e.g., ±2°C) by the PTC heating element (positive temperature coefficient thermistor heater). At this time, the Central Control Unit (CCU) immediately activates the built-in MultithreadScheduling Controller (MTSC). This controller is built on a real-time operating system (such as VxWorks or FreeRTOS) and has hard real-time task processing capabilities. The MTSC first performs initialization operations: allocating independent memory space to store the physical parameters (including channel ID, material viscosity, target temperature threshold, cooling rate limit, etc.) of the current material channel (labeled Channel_A) and the target new material channel (labeled Channel_B), and simultaneously loading the preset Thread Priority Configuration Table (TPCT). This configuration table defines the real-time priority of the cooling thread (Thread_Cool) and the activation thread (Thread_Activate) (e.g., setting the cooling thread priority to 90 and the activation thread priority to 85) to ensure that high-priority tasks can preempt system resources. After initialization, MTSC writes the activation flag (AF) to the system status register, changing its status value from "Standby" (StBY) to "Running" (RUN), indicating that the controller has officially entered the parallel control phase.

[0140] During activation, MTSC interacts with the underlying physical device through the Hardware Abstraction Layer (HAL). The HAL translates controller commands into specific hardware operations:

[0141] Send channel selection commands to the Field-Programmable Gate Array (FPGA) to enable the control ports corresponding to Channel_A and Channel_B;

[0142] Read the real-time clock (RTC) signal provided by the digital signal processor (DSP) to provide a microsecond-level time base for dual-thread operation;

[0143] Configure a Direct Memory Access (DMA) channel to ensure that the temperature sensor data stream (sampling rate 1kHz) can be directly read by the thread without CPU intervention.

[0144] Simultaneously, MTSC starts a watchdog timer (WTT) with a timeout threshold of 50 milliseconds (ms). If the two threads fail to synchronize their states within the time limit, the WDT will trigger a system reset (SRST) to prevent material solidification or leakage due to thread blocking.

[0145] To ensure control stability, MTSC performs Resource Conflict Detection (RCD) during the activation phase. This process is implemented through a Mutex Lock Manager (MLM): scanning for hardware resources that the cooling thread and the activation thread may contend for (such as shared temperature sensor buses, PWM output pins, etc.), and assigning exclusive access tokens (EAT) to conflicting resources. For example, if two threads need to access the same ADC (analog-to-digital converter) simultaneously, the MLM will force the activation thread to delay execution by 3 milliseconds (ms) to avoid data races. After resource arbitration is completed, MTSC writes an activation record to the System Event Log (SEL), including a timestamp (TS), a channel snapshot (CS), and a resource allocation map (RAM), providing a basis for tracing the source of subsequent fault diagnosis.

[0146] Create two threads in the controller, where thread 1 executes a cooling command sequence for the current material channel, and thread 2 executes an activation command sequence for the target new material channel to be switched to.

[0147] MTSC creates two independent real-time threads:

[0148] Thread 1 (Thread_Cool): Responsible for the cooling operation of the current material channel (Channel_A). This thread calls the Cooling Command Sequence Generator (CCSG), which dynamically generates three-step instructions based on the material's thermal property vector (including parameters such as specific heat capacity and glass transition temperature):

[0149] Step Cooling Command (SCC): Controls the Peltier Thermoelectric Cooler (PTEC) in the channel to cool down in a gradient manner (e.g., from 220°C→180°C→150°C, with each stage of cooling lasting 500ms).

[0150] Material Retraction Command (MRC): Drives a high-precision linear stepper motor (LSM) to retract residual material in the channel in 0.1 mm increments to prevent blockage during cooling;

[0151] Inert Gas Purging Command (IGPC): Open the solenoid valve (SV) to inject nitrogen (N2) into the channel at a flow rate of 5 liters / minute (L / min) for 200 ms to remove material vapors.

[0152] Thread 2 (Thread_Activate): Responsible for activating the target new material channel (Channel_B). This thread generates four instructions using the Activation Command Sequence Compiler (ACCC):

[0153] Pressure Balancing Command (PBC): Adjust the opening of the piezoelectric proportional valve (PPV) to 30% to make the internal pressure of Channel_B consistent with that of the printhead cavity (target pressure difference < 0.1 bar).

[0154] Channel Pre-fill Command (CPC): Start the microgear pump (MGP) to deliver material at a speed of 100 revolutions per minute (RPM) to fill the channel to the nozzle inlet;

[0155] Shear Stress Optimization Command (SSOC): Calculates the optimal shear rate based on the material viscosity (e.g., 3000 centipoise) and controls the MGP rotation speed to increase in a stepwise manner (100→300→500 RPM) to avoid material shear thinning;

[0156] Nozzle Wetting Command (NWC): Activate the ultrasonic vibrator (USV) at the nozzle end to operate at 40 kHz for 100 ms to eliminate contact hysteresis between the material and the nozzle wall.

[0157] The dual-thread execution employs Time-Slice Round Robin Scheduling (TSRRS). MTSC allocates a fixed time slice (TS) to each thread: Thread_Cool receives a 3-millisecond (ms) slice, and Thread_Activate receives a 2-millisecond (ms) slice. Each scheduling cycle (SC) is 5 milliseconds (ms), and thread switching is triggered by a Real-Time Clock Interrupt (RTCI). Critical status data (such as current temperature and pressure values) is backed up via thread-local storage (TLS) to ensure no data loss during context switching. Thread execution progress is tracked in real-time by a progress counter (PC): Thread_Cool increments PC by 1 for each completed instruction (maximum count 3), and Thread_Activate increments PC by 1 for each completed instruction (maximum count 4), providing quantitative data for subsequent synchronization.

[0158] Synchronize the dual-thread states using shared memory and generate channel synchronization status codes;

[0159] State synchronization between the two threads is achieved through a shared memory area (SMA). This area is located within the CCU's static random access memory (SRAM) and is divided into three data segments:

[0160] The Status Flag Segment (SFS) contains an 8-bit binary status word (SW). Bit 0 indicates whether Thread_Cool has completed residual material recirculation (1 = Completed), Bit 1 indicates the inert gas purging status (1 = In progress), and Bits 2-3 encode the pumping stage of Thread_Activate (00 = Not started, 01 = Pressure equalization, 10 = Prefilling, 11 = Shear optimization).

[0161] Progress Counter Segment (PCS): Stores two 16-bit counter values ​​(CV), recording the PC values ​​of Thread_Cool and Thread_Activate respectively;

[0162] Error Code Segment (ECS): Stores thread execution exception codes (such as Err_T for excessive temperature and Err_P for excessive pressure).

[0163] The synchronization process uses the publish-subscribe pattern (PSP):

[0164] When Thread_Cool completes the MRC instruction, it writes SW.Bit0=1 to SMA and forces the cached data to be updated through the Memory Barrier Instruction (MBI).

[0165] After Thread_Activate detects that SW.Bit0=1, it immediately executes the CPC instruction (flow channel prefilling), and then sets SW.Bit2=1 after completion;

[0166] Thread_Cool triggers the IGPC command (inert gas purging) after detecting SW.Bit2=1.

[0167] At the same time, a counter comparison (CC) is performed every millisecond (ms): when |PC_Cool-PC_Activate| ≤ 1 (i.e., the progress difference does not exceed 1 instruction), a synchronization pass flag (SPF=1) is written to the PCS.

[0168] The Channel Synchronization Status Code (CSSC) is generated by the Status Synchronization Engine (SSE). This engine scans the SMA every 2 milliseconds (ms) and performs three steps:

[0169] Validity Check (VC): Verify the integrity of SMA data using Cyclic Redundancy Check (CRC-16);

[0170] Logical AND Operation (LAO): Performs an AND operation on Bit0, Bit2, and SPF of SW (1&1&1=1).

[0171] Encoding Conversion (EC): Combines the result of the operation (1 bit), the PC difference (4 bits), and the error code (4 bits) into a 9-bit CSSC.

[0172] For example, when the progress of the two threads is synchronized and there are no errors, CSSC=0x111 (binary 00010001001) is generated, where the high 4 bits "0001" indicate that the PC difference is 1, the middle 4 bits "0001" indicate that there are no errors, and the lowest bit "1" indicates that the synchronization condition is met.

[0173] When the channel synchronization status code reaches the preset threshold, the piezoelectric ceramic valve is triggered to open and close, and a material switching command sequence is output.

[0174] The preset threshold (PT) is stored in non-volatile memory (NVM), with a typical value of CSSC ≥ 0x110 (binary 00010001000). A threshold comparator (TC) monitors CSSC in real time: when CSSC ≥ PT is detected, a trigger pulse (TP) is immediately sent to the piezoelectric valve driver (PVD). This pulse is a 5-volt (V) square wave signal with a width of 100 nanoseconds (ns), transmitted to the driver circuit via a high-speed optocoupler (HSO) for isolation. PVD is constructed using a high-voltage operational amplifier (HVOA) to amplify the input signal to 120 volts (V) and drive up to 2 amperes (A), ensuring that the response time (RT) of the piezoelectric ceramic valve (PCV) is ≤1 millisecond (ms).

[0175] Valve body movement is controlled by a four-phase cooperative control (FPCC):

[0176] To close the Channel_A main valve: Apply a 120V negative step voltage (NSV) to the PCV coil of Channel_A, and the valve core will seal the flow channel within 0.8 milliseconds (ms);

[0177] Open the Channel_B bypass valve: After a delay of 0.2 milliseconds (ms) after Channel_A is closed, apply an 80V positive voltage (PV) to the Channel_B bypass valve to establish a low-pressure auxiliary flow path;

[0178] Open Channel_B main valve: 1 millisecond (ms) after the bypass valve opens, apply 120V PV to the Channel_B main valve to fully open the flow path;

[0179] Close the bypass valve: Cut off the voltage to the bypass valve after the main valve stabilizes.

[0180] The entire process is monitored by a valve displacement sensor (VDS), which is based on the Hall effect principle and has an accuracy of 0.01 mm. If valve core jamming is detected (displacement change rate <0.1 mm / ms), a safety interlock (SI) is immediately triggered to cut off the material supply.

[0181] The Material Switching Command Sequence (MSCS) is output in a standardized Data Frame (DF) format, with each frame containing:

[0182] Frame Header (FH): 4-byte synchronization word (SW=0x55AA55AA);

[0183] Channel Status Word (CSW): A 2-byte encoding of the currently active channel ID (e.g., 0x0001=Channel_B);

[0184] Timestamp (TS): 8 bytes to record the absolute time when the switch was completed (precision 100 nanoseconds);

[0185] Checksum (CS): 2-byte CRC-16 checksum.

[0186] This sequence is broadcast to the entire system via the Controller Area Network Bus (CAN Bus). Upon receiving the sequence, the printhead motion controller immediately positions the coordinates to the Switching Point Coordinate (SPC) and initiates the new material deposition process. Simultaneously, the MSCS is written to the Black Box Memory (BBM) to support offline analysis of the switching performance.

[0187] Seamless switching of material channels is achieved through a parallelized control architecture, innovatively synchronizing cooling and activation operations. A multi-threaded scheduling mechanism ensures precise timing of transitions between old and new material channels, eliminating printing interruptions caused by traditional serial switching, significantly shortening material switching transition time, and achieving truly continuous printing. Parallel processing reduces switching time, making it particularly suitable for high-precision multi-material alternating printing scenarios, ensuring consistent printed surface quality.

[0188] S204, in response to the switching command sequence, execute the material supply switching, and adjust the channel temperature in real time through an integrated cooling mechanism to maintain a stable material switching state. The integrated cooling mechanism, combined with temperature sensor feedback, adaptively adjusts the cooling rate to maintain the printhead temperature within a set threshold.

[0189] Specifically, it can parse the switching instruction sequence and drive the stepper motor to perform material supply path switching actions;

[0190] When the system receives the Switch Command Sequence (SCS) generated by the parallel switching control algorithm, the Central Motion Control Module (CMCM) first parses the commands. The SCS is an ordered data packet containing specific operational instructions. Its typical structure includes: the Target Material Channel ID (TMC_ID, such as channel A, B, or C), the Switch Action Type (SAT, such as "full switch" or "gradual switch"), the Stepper Motor Target Angle (SM_TA, in degrees °), and the Switching Velocity Profile (SVP, such as acceleration and maximum speed). The parsing process uses a Command Decoder (CD), which converts the digital commands in the SCS into specific mechanical control signals by looking up a preset Command-Action Mapping Table (CAMT). For example, when the parsing shows TMC_ID as "B", SM_TA as "120°", and SVP as "Profile_3 (medium speed slow start-stop)", it indicates that the feeding path needs to be switched to channel B. The corresponding rotary valve needs to be rotated to the 120-degree position and rotated according to the preset Profile_3 speed curve.

[0191] The parsed control signal is sent to the stepper motor (SM) drive system. This stepper motor is typically a two-phase hybrid stepper motor (TPH-SM), whose shaft is directly connected to a precision-machined rotary valve disk (RVD) via a coupling. The RVD has multiple arc-shaped flow guiding grooves (FGG) corresponding to the material channels. The drive system includes a dedicated stepper motor driver (SMD) that receives pulse signals (PS) and direction signals (DS) from the CMCM. For example, after receiving a "120°" target angle command, the driver calculates the number of pulses to be sent (Pulse Count, PC = 120 / 1.8 = approximately 67 pulses) based on the motor step angle (e.g., 1.8° / step), and generates a corresponding pulse frequency variation curve according to SVP parameters (e.g., acceleration 500 steps / s², maximum speed 1000 steps / s). The driver converts the electrical pulses into phase current (PC) of the motor windings, precisely controlling the rotor rotation. During rotation, the RVD's guide grooves gradually align with the inlet of the target material channel (e.g., channel B) and the main feed tube of the printhead, while simultaneously disconnecting the connection of the previous material channel (e.g., channel A). To ensure positioning accuracy, the system employs closed-loop feedback: a micro-magnet is embedded in the RVD, which, together with a Hall Effect Sensor (HES), detects the actual angle (AA) in real time. By comparing the deviation between AA and SM_TA (Angle Deviation, AD), the CMCM dynamically fine-tunes the pulse until AD approaches zero (e.g., within ±0.1°), thus completing the physical-level switching of the material feeding path.

[0192] The reliability and sealing of the switching action are crucial. A Teflon-Ceramic Composite Seal Ring (TCCSR) is used at the contact surface between the RVD and the fixed valve seat (VS) to ensure zero leakage under high pressure (e.g., 10-30 bar). During stepper motor drive, the system synchronously monitors the motor winding temperature (WT, acquired via a thermistor) and load current (LC). If an abnormal increase in LC is detected (e.g., exceeding the threshold of 2.5 amperes), indicating potential mechanical jamming, the system immediately triggers a Safety Interrupt (SI), stopping pulse transmission and initiating a Back-off Retraction (BOR) action to prevent damage to the motor or valve. After the switch is completed, the CMCM sends a PathSwitch Confirmation Signal (PSCS) to the main control system, which includes data such as the actual switching timestamp (TS), final angle (FA), and switching duration (SD), providing a timing reference for subsequent temperature control.

[0193] After the material switching is completed, the real-time temperature distribution map of the printhead is collected through the microchannel temperature sensor array;

[0194] After the material supply path switching operation is completed (marked by the receipt of the PSCS), the system immediately initiates the temperature monitoring phase. The core monitoring device is a micro-channel temperature sensor array (MTSA). This array consists of dozens (e.g., 48 or 96) miniature temperature sensing units (TSUs), arranged in a high-density grid pattern in key heat-sensitive areas inside the printhead: including the entry zone (EZ) of each material channel, the mixing chamber (MC), the nozzle throat (NT), and the heat spread zone (HSZ) of the printhead shell. Each TSU uses a thin-film platinum resistance temperature detector (Pt-RTD), whose resistance value (RV) changes linearly with temperature (e.g., PT100 type, 100 ohms at 0°C, rate of change 0.385Ω / °C). These Pt-RTDs are directly integrated onto the stainless steel or ceramic substrate of the printhead using micro-electro-mechanical systems (MEMS) technology, with a distance of only tens of micrometers (μm) from the flow channel, ensuring rapid response to changes in fluid temperature.

[0195] Data acquisition is performed by a dedicated multi-channel temperature acquisition circuit (MTAC). The MTAC powers each Pt-RTD with a constant current source (CCS, e.g., 1 mA) and measures the voltage drop (VD) across it using a 24-bit high-precision analog-to-digital converter (ADC). The resistance value (RV) of each sensor is calculated in real-time according to Ohm's law (Voltage = Current × Resistance). RV is then converted to a Celsius temperature value (TV) using a pre-stored platinum resistance-temperature lookup table (RT-LUT) or a polynomial fitting algorithm. The acquisition process synchronously scans all TSU channels at a high frequency (e.g., a sampling rate of 100 times per second). For example, an array containing 96 sensors can complete a full array scan every 10 milliseconds (ms), generating a raw dataset (RTD) containing 96 temperature points.

[0196] The raw temperature data is fed into the Temperature Distribution Map Generator (TDMG). This engine performs three processing steps:

[0197] Spatial Calibration (SC): Based on the preset three-dimensional coordinates (X, Y, Z) of each TSU, discrete points are mapped onto the spatial grid (SG) of the digital model inside the printhead.

[0198] Anomaly Filtering (AF): A moving median filter (MMF with a window size of 5 sampling points) is used to remove transient interference (such as spikes caused by electrical noise).

[0199] Interpolation Reconstruction (IR): The filtered data is interpolated on the spatial grid SG using the Bicubic Spline Interpolation Algorithm (BSIA) to generate a continuous and smooth Real-time Temperature Distribution Map (RTDM). This map is visualized as a heatmap (HM), with different colors (e.g., blue-green-yellow-red) representing different temperature ranges (e.g., 20°C-50°C-80°C-120°C), and the current temperature (CT) and temperature gradient vector (TGV) of key locations (e.g., MC, NT) are labeled. The RTDM is updated several times per second (e.g., 10 times), providing high-resolution thermal field information for cooling control.

[0200] The real-time temperature distribution map is input into the PID-fuzzy composite controller to calculate the cooling rate compensation parameters; the current of the semiconductor cooling chip is adjusted according to the cooling rate compensation parameters to maintain the printhead temperature within the set threshold.

[0201] The generated real-time temperature distribution map (RTDM) is transmitted in real time to a PID-FuzzyHybrid Controller (PID-FHC). This controller combines the stability of Proportional-Integral-Derivative (PID) control with the advantages of fuzzy logic (FL) in handling nonlinearity. The core input of PID-FHC is the actual temperature (AT) of key regions in the RTDM, such as the temperature T_MC of the mixing chamber (MC) and the temperature T_NT of the nozzle throat (NT). The system presets a target temperature threshold range (TTTR) for each region; for example, T_MC needs to be maintained at 75±2°C, and T_NT needs to be maintained at 85±3°C. The controller first calculates the temperature deviation (Temperature Error, TE = AT - TTTR_center) and its rate of change (Error Derivative, ED = d(TE) / dt) for each region.

[0202] The PID module (PIDM) employs the positional PID algorithm. Its output (PID Output, PID_OUT) is a weighted sum of three parts:

[0203] The proportional term (PT) is proportional to TE, and its coefficient is the proportional gain Kp. Kp determines the strength of the system's response to the current deviation.

[0204] Integral Term (IT): Accumulated historical data TE, with coefficients equal to the integral gain Ki. Ki eliminates steady-state error, but excessive amounts can lead to oscillations.

[0205] The derivative term (DT) is proportional to ED, and its coefficient is the derivative gain Kd (DerivativeGain). Kd predicts the trend and suppresses overshoot.

[0206] For example, if T_MC is currently 78°C (TE = +3°C) and continues to rise (ED > 0), PIDM may output a large negative PID_OUT (such as -12.5) to enhance cooling.

[0207] The Fuzzy Module (FM) handles nonlinear scenarios that PID controllers struggle with (such as sudden changes in material heat capacity and cooling hysteresis). Its inputs are the fuzzified values ​​of TE and ED (e.g., "negative large NB", "negative small NS", "zero ZO", "positive small PS", "positive large PB"), and its output is the fuzzy compensation (FC). The Fuzzy Rule Base (FRB) contains empirical rules, such as:

[0208] "If TE is positive and ED is positive, then FC is negative (powerful cooling)";

[0209] "If TE is negative and ED is negative, then FC is positive (weak heating)."

[0210] PID_OUT and FC are combined using a weighted fusion algorithm (WFA) to generate the final cooling rate compensation parameter (CRCP). CRCP is a dimensionless adjustment coefficient (ranging from -1.0 to +1.0), where a negative value indicates that cooling needs to be enhanced, and a positive value indicates that cooling or heating needs to be reduced.

[0211] The CRCP is fed into a thermoelectric cooler (TEC) drive system. The TEC is based on the Peltier effect, and its cooling / heating power is strictly dependent on the magnitude and direction of the input current. The drive system includes a programmable bidirectional current source (PBCS). The system maps the CRCP to a target current value (TCV):

[0212] When CRCP < 0 (cooling required): TCV = |CRCP| × Max_Cooling_Current (e.g., maximum cooling current 3.0 amperes A).

[0213] When CRCP>0 (heating required): TCV = CRCP × Max_Heating_Current (e.g., maximum heating current 2.0 amperes A).

[0214] When CRCP = 0: TCV = 0 (TEC is off).

[0215] The PBCS precisely outputs DC current based on the TCV, and its direction and magnitude are controllable in real time (e.g., continuously adjustable from -2.0A to +3.0A). The TEC is tightly attached to the printhead heat exchange baseplate (HEB). When the current is negative (e.g., -1.8A), the cold end of the TEC absorbs heat from the printhead, while the hot end dissipates heat through heat sink fins and a fan; when the current is positive (e.g., +0.5A), the TEC provides auxiliary heating to the printhead. The system maintains this effect through closed-loop control: after the TEC current is adjusted, new RTDM data is continuously fed back to the PID-FHC, dynamically calculating a new CRCP, forming a "sensing-decision-execution-re-sensing" closed loop to ensure that the printhead temperature (especially T_MC and T_NT) remains stable within the TTTR. Simultaneously, the system monitors the temperature difference (ΔT) between the cold and hot ends of the TEC. If ΔT exceeds the safety limit (e.g., 60°C), the current is automatically reduced to prevent overheating and damage to the TEC, ensuring long-term reliable operation.

[0216] While performing physical switching, a closed-loop temperature control system dynamically balances the thermal field distribution of the printhead. Based on a real-time temperature feedback-based intelligent cooling algorithm, it automatically adjusts the cooling power according to the material's thermal capacity differences, ensuring the printhead is always within the optimal operating temperature window. This solves the most critical thermal disturbance problem in multi-material printing and maintains process stability through adaptive cooling. This mechanism can control temperature fluctuations within, for example, ±1℃, effectively preventing the decrease in interlayer adhesion and printing defects caused by thermal stress.

[0217] As can be seen, the system receives print task instructions and generates a material switching demand vector based on these instructions. Based on this demand vector, a predictive preheating algorithm activates the preheating mechanism of the multi-channel feeding system, dynamically preheating the material channel to be used and generating a preheating ready state signal. Using this signal, a parallel switching control algorithm triggers the material switching action of the multi-channel feeding system, generating a switching instruction sequence. Responding to the switching instruction sequence, the system executes material supply switching and adjusts the channel temperature in real time through an integrated cooling mechanism to maintain a stable material switching state. This enables efficient collaborative switching of multiple material channels, improving the printhead's response speed and temperature stability.

[0218] Another embodiment of the present invention provides a multi-material switching system for a printhead, see [link to relevant documentation]. Figure 3 The system may include:

[0219] The receiving module 301 is used to receive a printing task instruction including a material type sequence, switching point coordinates and material property parameters, and generate a material switching requirement vector based on the printing task instruction. The material switching requirement vector encodes the material switching order, switching time point and material thermal property requirements.

[0220] The preheating module 302 is used to activate the preheating mechanism of the multi-channel feeding system based on the material switching demand vector and through a predictive preheating algorithm to dynamically preheat the material channel to be used and generate a preheating ready state signal. The predictive preheating algorithm optimizes the preheating temperature curve in real time to match the thermal characteristics of different materials.

[0221] The generation module 303 is used to trigger the material switching action of the multi-channel feeding system by using the preheating ready state signal and the parallel switching control algorithm to generate a switching instruction sequence. The parallel switching control algorithm adopts a multi-threaded scheduling mechanism to synchronously execute the operation of the material channel before cooling and the operation of activating the new material channel.

[0222] The switching module 304 is used to respond to the switching command sequence, perform material supply switching, and adjust the channel temperature in real time through an integrated cooling mechanism to maintain a stable material switching state. The integrated cooling mechanism combines temperature sensor feedback to adaptively adjust the cooling rate to maintain the printhead temperature within a set threshold.

[0223] This invention also provides a storage medium storing a computer program, wherein the computer program is configured to execute the steps in any of the above method embodiments when running.

[0224] Specifically, in this embodiment, the storage medium can be configured to store a computer program for performing the following steps:

[0225] S201, receive a printing task instruction including a material type sequence, switching point coordinates and material property parameters, and generate a material switching requirement vector based on the printing task instruction, wherein the material switching requirement vector encodes the material switching order, switching time point and material thermal property requirements;

[0226] S202, based on the material switching demand vector, the preheating mechanism of the multi-channel feeding system is activated by a predictive preheating algorithm to dynamically preheat the material channel to be used and generate a preheating ready state signal. The predictive preheating algorithm optimizes the preheating temperature curve in real time to match the thermal characteristics of different materials.

[0227] S203, using the preheating ready state signal, trigger the material switching action of the multi-channel feeding system through the parallel switching control algorithm to generate a switching instruction sequence, wherein the parallel switching control algorithm adopts a multi-thread scheduling mechanism to synchronously execute the operation of the material channel before cooling and the operation of activating the new material channel;

[0228] S204, in response to the switching command sequence, execute the material supply switching, and adjust the channel temperature in real time through an integrated cooling mechanism to maintain a stable material switching state. The integrated cooling mechanism, combined with temperature sensor feedback, adaptively adjusts the cooling rate to maintain the printhead temperature within a set threshold.

[0229] This invention also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0230] Specifically, the aforementioned electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the aforementioned processor, and the input / output device is connected to the aforementioned processor.

[0231] Specifically, in this embodiment, the processor can be configured to perform the following steps via a computer program:

[0232] S201, receive a printing task instruction including a material type sequence, switching point coordinates and material property parameters, and generate a material switching requirement vector based on the printing task instruction, wherein the material switching requirement vector encodes the material switching order, switching time point and material thermal property requirements;

[0233] S202, based on the material switching demand vector, the preheating mechanism of the multi-channel feeding system is activated by a predictive preheating algorithm to dynamically preheat the material channel to be used and generate a preheating ready state signal. The predictive preheating algorithm optimizes the preheating temperature curve in real time to match the thermal characteristics of different materials.

[0234] S203, using the preheating ready state signal, trigger the material switching action of the multi-channel feeding system through the parallel switching control algorithm to generate a switching instruction sequence, wherein the parallel switching control algorithm adopts a multi-thread scheduling mechanism to synchronously execute the operation of the material channel before cooling and the operation of activating the new material channel;

[0235] S204, in response to the switching command sequence, execute the material supply switching, and adjust the channel temperature in real time through an integrated cooling mechanism to maintain a stable material switching state. The integrated cooling mechanism, combined with temperature sensor feedback, adaptively adjusts the cooling rate to maintain the printhead temperature within a set threshold.

[0236] The above description, based on the embodiments shown in the figures, details the structure, features, and effects of the present invention. The above description is only a preferred embodiment of the present invention, but the present invention is not limited to the scope of implementation shown in the figures. Any changes made in accordance with the concept of the present invention, or equivalent embodiments modified to have equivalent changes, that do not exceed the spirit covered by the specification and figures, should be within the protection scope of the present invention.

Claims

1. A method for switching multiple materials in a printhead, characterized in that, The method includes: Receive a print task instruction including a material type sequence, switching point coordinates and material property parameters, and generate a material switching requirement vector based on the print task instruction. The material switching requirement vector encodes the material switching order, switching time point and material thermal property requirements. Based on the material switching demand vector, the preheating mechanism of the multi-channel feeding system is activated by a predictive preheating algorithm to dynamically preheat the material channels to be used and generate a preheating ready state signal. The predictive preheating algorithm optimizes the preheating temperature curve in real time to match the thermal characteristics of different materials. Using the preheating ready status signal, the material switching action of the multi-channel feeding system is triggered by the parallel switching control algorithm to generate a switching instruction sequence. The parallel switching control algorithm adopts a multi-threaded scheduling mechanism to simultaneously execute the operation of the material channel before cooling and the operation of activating the new material channel. In response to the switching command sequence, the material supply is switched, and the channel temperature is adjusted in real time through an integrated cooling mechanism to maintain a stable material switching state. The integrated cooling mechanism combines temperature sensor feedback to adaptively adjust the cooling rate to maintain the printhead temperature within a set threshold. Based on the material switching demand vector, the preheating mechanism of the multi-channel feeding system is activated through a predictive preheating algorithm to dynamically preheat the material channels to be used, generating a preheating ready state signal. The predictive preheating algorithm optimizes the preheating temperature curve in real time to match the thermal characteristics of different materials, including: Extract the next set of thermal property vector subsets of the materials to be switched from the material switching demand vector; Input a subset of thermal property vectors into a preheating prediction model built on LSTM, and output the initial preheating temperature curves for each material channel; By combining real-time temperature feedback from the printhead, the initial preheating temperature curve is dynamically optimized using a gradient descent algorithm to generate a dynamic preheating temperature curve. The PTC heating element of the multi-channel feeding system is controlled according to the dynamic preheating temperature curve to dynamically preheat the material channel corresponding to the material to be switched. When the difference between the preheated channel temperature and the target temperature enters the preset tolerance range, the preheating ready state signal is triggered.

2. The method according to claim 1, characterized in that, The receiving process includes a print task instruction comprising a material type sequence, switching point coordinates, and material property parameters. Based on the print task instruction, a material switching requirement vector is generated. This material switching requirement vector encodes the material switching sequence, switching time point, and material thermal property requirements, including: Parse the material type sequence in the print task instruction, extract the unique identifier for each material, and generate a material identification sequence; Based on the switching point coordinates and the printhead motion trajectory model, calculate the precise timestamp of each material switching action and output the set of switching time points; Query the material property database, quantify the material thermal property requirements into feature values, and generate a material thermal property vector. The material identification sequence, the set of switching time points, and the material thermal property vector are encoded into a fixed-dimensional tensor, and the material switching demand vector is output.

3. The method according to claim 2, characterized in that, The process involves using the preheating ready state signal to trigger material switching actions in the multi-channel feeding system via a parallel switching control algorithm, generating a switching instruction sequence. The parallel switching control algorithm employs a multi-threaded scheduling mechanism to simultaneously execute operations on the pre-cooling material channel and the activation of the new material channel, including: Receive the preheating-ready status signal and activate the multi-threaded scheduling controller; Create two threads in the controller, where thread 1 executes a cooling command sequence for the current material channel, and thread 2 executes an activation command sequence for the target new material channel to be switched to. Synchronize the dual-thread states using shared memory and generate channel synchronization status codes; When the channel synchronization status code reaches the preset threshold, the piezoelectric ceramic valve is triggered to open and close, and a material switching command sequence is output.

4. The method according to claim 3, characterized in that, The system responds to the switching command sequence by performing a material supply switch and adjusting the channel temperature in real time through an integrated cooling mechanism to maintain a stable material switching state. This integrated cooling mechanism, combined with temperature sensor feedback, adaptively adjusts the cooling rate to maintain the printhead temperature within a set threshold, including: The switching instruction sequence is analyzed, and the stepper motor is driven to perform the material supply path switching action. After the material switching is completed, the real-time temperature distribution map of the printhead is collected through the microchannel temperature sensor array; Input the real-time temperature distribution map into the PID-fuzzy composite controller to calculate the cooling rate compensation parameters; The current of the semiconductor cooling chip is adjusted according to the cooling rate compensation parameter to maintain the printhead temperature within the set threshold.

5. A multi-material switching system for a printhead, characterized in that, The system includes: The receiving module is used to receive a printing task instruction including a material type sequence, switching point coordinates and material property parameters, and generate a material switching requirement vector based on the printing task instruction. The material switching requirement vector encodes the material switching order, switching time point and material thermal property requirements. The preheating module is used to activate the preheating mechanism of the multi-channel feeding system based on the material switching demand vector and through a predictive preheating algorithm to dynamically preheat the material channel to be used and generate a preheating ready status signal. The predictive preheating algorithm optimizes the preheating temperature curve in real time to match the thermal characteristics of different materials. The generation module is used to trigger the material switching action of the multi-channel feeding system by means of the preheating ready state signal and the parallel switching control algorithm to generate a switching instruction sequence. The parallel switching control algorithm adopts a multi-threaded scheduling mechanism to synchronously execute the operation of the material channel before cooling and the operation of activating the new material channel. The switching module is used to respond to the switching command sequence, perform material supply switching, and adjust the channel temperature in real time through an integrated cooling mechanism to maintain a stable material switching state. The integrated cooling mechanism combines temperature sensor feedback to adaptively adjust the cooling rate to maintain the printhead temperature within a set threshold.

6. The system according to claim 5, characterized in that, The receiving module is specifically used for: Parse the material type sequence in the print task instruction, extract the unique identifier for each material, and generate a material identification sequence; Based on the switching point coordinates and the printhead motion trajectory model, calculate the precise timestamp of each material switching action and output the set of switching time points; Query the material property database, quantify the material thermal property requirements into feature values, and generate a material thermal property vector. The material identification sequence, the set of switching time points, and the material thermal property vector are encoded into a fixed-dimensional tensor, and the material switching demand vector is output.

7. The system according to claim 6, characterized in that, The preheating module is specifically used for: Extract the next set of thermal property vector subsets of the materials to be switched from the material switching demand vector; Input a subset of thermal property vectors into a preheating prediction model built on LSTM, and output the initial preheating temperature curves for each material channel; By combining real-time temperature feedback from the printhead, the initial preheating temperature curve is dynamically optimized using a gradient descent algorithm to generate a dynamic preheating temperature curve. The PTC heating element of the multi-channel feeding system is controlled according to the dynamic preheating temperature curve to dynamically preheat the material channel corresponding to the material to be switched. When the difference between the preheated channel temperature and the target temperature enters the preset tolerance range, the preheating ready state signal is triggered.

8. A storage medium, characterized in that, The storage medium stores a computer program, wherein the computer program is configured to execute the method of any one of claims 1-4 when it is run.

9. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to run the computer program to perform the method of any one of claims 1-4.

Images