Method and system for sink mark suppression of an integral die cast

Abstract

The application discloses a shrinkage mark inhibition method and system of an integrated die casting, and relates to the technical field of die casting. The method comprises the following steps: collecting position information of a solidification front to determine first state information; collecting vibration signals of shrinkage stress waves to reconstruct a shrinkage wave field distribution map as second state information; fusing the first state information and the second state information to generate a shrinkage mark probability cloud map of the whole solidification process; if the shrinkage mark probability cloud map does not satisfy a shrinkage mark probability threshold, activating a shrinkage mark inhibition module, executing compensation wave parameter decision through a first decision layer, executing local shrinkage compensation pressure pulse parameter decision through a second decision layer, outputting mapping and optimization iteration of a layer, and determining a die casting regulation parameter combination; and executing die casting process regulation according to the die casting regulation parameter combination. The technical problem that it is difficult to dynamically inhibit shrinkage marks in the die casting process in the prior art, leading to unstable die casting quality, is solved, dynamic inhibition of shrinkage defects is achieved, and the technical effect of improving the forming quality of die castings is achieved.

Description

Technical Field

[0001] This invention relates to the field of die casting technology, and specifically to a method and system for suppressing shrinkage marks in integral die castings. Background Technology

[0002] Die casting is widely used in the manufacturing of automotive structural parts and large integrated die-cast components due to its advantages such as high forming efficiency and good dimensional accuracy. However, during the solidification process of die-cast parts, shrinkage defects are prone to occur in thick-walled areas or hot spots due to the volume shrinkage of the molten metal and uneven solidification, which seriously affects the appearance quality, structural strength, and service reliability of the product. In existing technologies, shrinkage defects are usually addressed by optimizing the gating system, adjusting process parameters (such as injection speed, holding pressure, mold temperature, etc.), or making predictions in advance through numerical simulation. However, these methods mostly rely on experience or offline analysis and are difficult to accurately reflect the complex multi-physics coupling behavior in the actual production process. On the other hand, some technologies monitor the die casting process through temperature or pressure sensors, or identify shrinkage defects after forming through visual inspection, X-ray inspection, etc. However, these methods are mostly single physical quantity monitoring or post-detection methods, lacking real-time perception and evolution mechanism analysis of the shrinkage formation process, making it difficult to effectively intervene at the defect formation stage. Furthermore, existing technologies have not yet formed a quantitative assessment system for shrinkage risk based on the fusion of multi-source sensing information, nor do they have a means of coordinated control over the propagation of shrinkage stress waves and the dynamic changes in the solidification front. This makes it difficult to dynamically suppress shrinkage defects, thereby affecting the stability of the forming quality of die-cast parts. Summary of the Invention

[0003] This application provides a method and system for suppressing shrinkage marks in integral die castings, which solves the technical problem in the prior art that it is difficult to dynamically suppress shrinkage marks in the die casting process, resulting in unstable quality of die castings.

[0004] The first aspect of this application provides a method for suppressing shrinkage marks in an integral die-cast part, the method comprising:

[0005] The position information of the solidification front is collected by ultrasonic probes distributed within the die-casting mold. The first state information is determined by inverting the shrinkage risk based on the topology sensing network. The vibration signal of the shrinkage stress wave is collected by the piezoelectric sensing array. The shrinkage wave field distribution map is reconstructed by the waveform analysis unit as the second state information. The first state information and the second state information are fused to generate a shrinkage probability cloud map of the entire solidification process. If the shrinkage probability cloud map does not meet the shrinkage probability threshold, the shrinkage suppression module is activated. The first decision layer performs the compensation wave parameter decision of the shrinkage stress wave phase conjugate, and the second decision layer performs the local shrinkage pressure pulse parameter decision. The execution layer output mapping and optimization iteration determine the die-casting control parameter combination. The die-casting process is controlled according to the die-casting control parameter combination.

[0006] A second aspect of this application provides a shrinkage mark suppression system for integral die castings, the system comprising: The first information acquisition module collects the position information of the solidification front using ultrasonic probes distributed within the die-casting mold, and determines the first state information by performing shrinkage risk inversion based on a topological sensing network. The second information acquisition module collects the vibration signal of the shrinkage stress wave using a piezoelectric sensor array, and reconstructs the shrinkage wave field distribution map as the second state information based on the waveform analysis unit. The information fusion module fuses the first and second state information to generate a shrinkage probability cloud map of the entire solidification process. The decision execution module activates the shrinkage suppression module if the shrinkage probability cloud map does not meet the shrinkage probability threshold. It then performs compensation wave parameter decisions based on the phase conjugate of the shrinkage stress wave through the first decision layer, and local shrinkage pressure pulse parameter decisions through the second decision layer. The execution layer outputs mapping and optimization iterations to determine the die-casting control parameter combination. The process control module executes die-casting process control based on the die-casting control parameter combination.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: First, ultrasonic probes distributed within the die-casting mold collect the position information of the solidification front. A topological sensing network is used to perform shrinkage risk inversion to determine the first state information. A piezoelectric sensor array collects the vibration signal of the shrinkage stress wave, and a shrinkage wave field distribution map is reconstructed using waveform analysis units as the second state information. Next, the first and second state information are fused to generate a shrinkage probability cloud map of the entire solidification process. If the shrinkage probability cloud map does not meet the shrinkage probability threshold, a shrinkage suppression module is activated. The first decision layer executes the compensation wave parameter decision for the shrinkage stress wave phase conjugate, and the second decision layer executes the local shrinkage pressure pulse parameter decision. The execution layer outputs mapping and optimization iterations to determine the die-casting control parameter combination. Finally, die-casting process control is performed based on the die-casting control parameter combination. This solves the technical problem in existing technologies where it is difficult to dynamically suppress shrinkage during the die-casting process, leading to unstable die-casting quality. It achieves dynamic suppression of shrinkage defects, thereby improving the forming quality of die-casting parts. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 A schematic flowchart of a method for suppressing shrinkage marks in an integral die-cast part provided in an embodiment of this application; Figure 2This is a schematic diagram of the shrinkage mark suppression system for an integral die-cast part provided in an embodiment of this application.

[0010] Explanation of reference numerals in the attached drawings: First information acquisition module 11, Second information acquisition module 12, Information fusion module 13, Decision execution module 14, Process control module 15. Detailed Implementation

[0011] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0012] Example 1, as Figure 1 As shown, this application provides a method for suppressing shrinkage marks in integral die-cast parts, wherein the method includes: The position information of the solidification front is collected by ultrasonic probes distributed inside the die-casting mold, and the first state information is determined by shrinkage risk inversion based on the topology sensing network.

[0013] Multiple ultrasonic probe arrays are deployed around the cavity wall and hot spot area of ​​the die-casting mold according to a preset spatial resolution. The ultrasonic probes periodically emit high-frequency ultrasonic signals using a pulse-echo method and receive echo signals reflected from the solid-liquid interface. The echo signals are subjected to time-domain gating and bandpass filtering to extract the echo flight time, echo amplitude, and phase information. Temperature compensation is performed on the flight time according to the sound velocity-temperature coupling model to calculate the instantaneous spatial coordinates of the solidification front at each measuring point. Spatial interpolation and reconstruction are performed based on the multi-probe data to generate three-dimensional topographic point cloud data of the solidification front. The three-dimensional topographic point cloud data is meshed to construct a solidified shell topology diagram, and local thickness gradient, curvature distribution, and connectivity of closed regions are extracted as topological feature parameters. The topological feature parameters are input into a pre-trained topology sensing network, which adopts a graph neural network structure to propagate solidification state information through the adjacency relationship between nodes, outputting the solidified shell thickness distribution field and the spatial location distribution of the isolated liquid phase region, and outputting it as the first state information.

[0014] Furthermore, the first state information is determined by performing shrinkage risk inversion based on the topology-sensing network, including: Based on the ultrasonic probes distributed within the die-casting mold cavity, position information of the solidification front is collected as the die-casting solidification process progresses; the position information is input into the topological sensing network of the solidification front to invert the current thickness distribution of the solidified shell and the spatial position of the liquid-phase isolated region, which serves as the first state information, wherein the liquid-phase isolated region characterizes the shrinkage mark risk area.

[0015] Preferably, ultrasonic probes with a preset spatial distribution within the die-casting mold cavity are used to transmit pulsed ultrasonic signals at a set sampling period during the die-casting solidification process and receive echo signals reflected from the solid-liquid interface. The echo signals are then subjected to time-domain gating and bandpass filtering to extract the echo flight time, amplitude, and phase information. A preset sound velocity-temperature correction model is used to compensate for the flight time, obtaining the instantaneous spatial coordinates of the solidification front at each measuring point. Based on the spatial coordinates of multiple measuring points, interpolation reconstruction is performed to generate three-dimensional morphological data of the solidification front. This three-dimensional morphological data is then meshed to construct a solidification topology diagram, and local thickness gradients and curvature distributions are extracted. The distribution and connectivity features are used as topological feature parameters. These topological feature parameters are input into the topological sensing network at the solidification front for state inversion. The topological sensing network performs feature propagation and fusion of the adjacency relationships of each grid node based on a graph structure, outputting the thickness distribution field of the current solidified shell and the spatial distribution of the liquid phase region. Based on the connectivity analysis of the liquid phase region, closed liquid phase regions that are no longer connected to the main feeding channel are identified and determined as liquid phase isolated regions. Their volume fraction, spatial dispersion, and location distribution are used as shrinkage risk characterization parameters and output as first state information. The liquid phase isolated regions are used to characterize shrinkage risk areas.

[0016] Vibration signals of contraction stress waves are acquired by a piezoelectric sensing array, and the contraction wave field distribution map is reconstructed based on the waveform analysis unit as the second state information.

[0017] Furthermore, the contracted wave field distribution map reconstructed from the waveform analysis unit is used as the second state information, including: A piezoelectric sensor array is embedded at a key location in the die-casting mold to collect vibration signals during the die-casting solidification process. These vibration signals cover the frequency, amplitude, phase, and propagation direction of the contraction wave, characterizing the particle displacement caused by the propagation of contraction stress within the target casting. The vibration signals are amplified and filtered, and a waveform analysis unit is activated to perform coherent superposition and reverse-time propagation analysis based on the elastic wave equation. This reconstructs the contraction wave field distribution map within the target casting, serving as the second state information. The contraction wave field distribution map identifies the location of standing wave nodes and energy concentration areas.

[0018] A piezoelectric sensor array is embedded around the hot spot area, abrupt cross-section area, and feeding channel of the die-casting mold according to a preset spatial resolution. The piezoelectric sensor array is numbered and arranged according to the array topology, and vibration signals during the die-casting solidification process are synchronously acquired at a set sampling frequency. The vibration signals include displacement response signals caused by contraction waves, which are converted into electrical signals by the piezoelectric effect and then cover the frequency, amplitude, phase, and propagation direction information of the contraction waves. The vibration signals are pre-amplified, bandpass filtered, and multi-channel synchronously aligned to construct a spatiotemporally consistent vibration dataset. The waveform analysis unit is activated, and the vibration dataset is input into a wave field propagation model based on the elastic wave equation. The model is established by discretizing the elastic wave equation. A wave propagation mesh model of the internal medium of the target casting is used, and the propagation velocity field is determined by combining material density, elastic modulus, and temperature-related parameters. Based on the propagation mesh model, coherent superposition calculations are performed on multi-channel vibration signals to enhance the co-source wave field signals and suppress noise interference. At the same time, a reverse-time propagation algorithm is used to reconstruct the wave field in time, locating the source distribution and propagation path of the contraction stress wave. The contraction wave field distribution map inside the target casting is reconstructed according to the time inversion results, and the standing wave node positions, wave field propagation directions, and energy concentration areas are marked in the contraction wave field distribution map. The energy concentration areas are calculated by the wave field energy density function, and the contraction wave field distribution map is output as the second state information.

[0019] By integrating the first state information and the second state information, a shrinkage probability cloud map of the entire solidification process is generated.

[0020] Specifically, the solidified shell thickness distribution field, spatial location and volume fraction of the isolated liquid phase region, spatial dispersion and connectivity indices in the first state information are spatiotemporally aligned with the contraction wave field distribution map, wave field energy density and propagation direction information in the second state information to construct a multi-source state feature set under a unified spatiotemporal coordinate system. Based on the multi-source state feature set, the die-casting solidification process is divided into stages according to a preset time step to form a solidification stage sequence, and the corresponding multi-dimensional feature vector is extracted in each stage. The multi-dimensional feature vector is input into a preset shrinkage probability prediction model, wherein the shrinkage... The shrinkage probability prediction model is a lightweight generator structure. It models the coupling relationship between the solidified shell thickness gradient, the distribution of isolated liquid phase regions, and the energy concentration of shrinkage waves, and outputs the shrinkage probability value of each spatial grid node. Based on the shrinkage probability value of each spatial grid node, a three-dimensional probability field is constructed, and the three-dimensional probability field is continuously accumulated and updated in the time dimension to generate a shrinkage probability cloud map covering the entire solidification process. The shrinkage probability cloud map uses spatial location as the coordinate axis and shrinkage probability as the mapping value to characterize the spatial distribution and evolution trend of potential shrinkage defects at different solidification stages.

[0021] Furthermore, generating a probability cloud map of the shrinkage marks throughout the entire solidification process includes: By fusing the first state information and the second state information, the solidification state of the casting is determined. Specifically, the solidification state of the casting is updated by dividing the die-casting solidification process into stages. Based on the solidification state of the casting, a shrinkage probability prediction based on a lightweight generator is performed to generate a shrinkage probability cloud map throughout the solidification process. The topology sensing network, waveform analysis unit, and lightweight generator are embedded in the central control of the die-casting machine tool and are activated during the die-casting operation.

[0022] Preferably, by fusing first-state information and second-state information, a multi-source state feature set under a unified spatiotemporal coordinate system is constructed. This multi-source state feature set includes at least the solidified shell thickness distribution, the spatial location and volume fraction of isolated liquid phase regions, spatial dispersion and connectivity indices, as well as the energy density, propagation direction, and standing wave node distribution of the contraction wave field. Based on this multi-source state feature set, the die-casting solidification process is divided into stages according to the time step and the degree of solidification. The degree of solidification is characterized by the growth rate of the average solidified shell thickness or the change rate of the liquid phase volume fraction. The die-casting process is divided into an initial solidification stage, a rapid shell growth stage, and a final feeding stage. The multi-source state feature set is dynamically updated within each stage. The corresponding solidification state of the casting is determined; based on the solidification state, a multi-dimensional feature vector for the corresponding stage is constructed and input into a shrinkage probability prediction model based on a lightweight generator. The lightweight generator models the coupling relationship between the solidification shell thickness gradient, the distribution of the isolated liquid phase region, and the energy concentration of the shrinkage wave, and outputs the shrinkage probability value of each spatial grid node. The shrinkage probability values ​​output at each stage are spliced ​​temporally and reconstructed spatially to form a three-dimensional shrinkage probability cloud map covering the entire solidification process. The topology sensing network, waveform analysis unit, and lightweight generator are integrated into the central control system of the die-casting machine tool, deployed in an embedded manner, and activated synchronously when the die-casting operation starts, realizing online real-time prediction and dynamic updating of the solidification process.

[0023] If the shrinkage probability cloud map does not meet the shrinkage probability threshold, the shrinkage suppression module is activated. The first decision layer performs the compensation wave parameter decision of the shrinkage stress wave phase conjugate, and the second decision layer performs the local shrinkage pressure pulse parameter decision. The execution layer output mapping and optimization iteration determine the die casting control parameter combination.

[0024] Specifically, the shrinkage probability values ​​of each spatial grid node in the shrinkage probability cloud map are compared point by point with a preset shrinkage probability threshold. Regions exceeding the threshold are extracted as target control regions, and a risk intensity index for the target control region is calculated. This risk intensity index is weighted based on the probability mean, probability variance, and spatial clustering. The target control region and risk intensity index are input into a first decision layer and a second decision layer. The first decision layer, based on the shrinkage wave field distribution in the second state information, extracts the dominant frequency, phase distribution, and propagation direction information within the target control region, constructs phase conjugate constraints, and generates a set of compensation wave parameters, including compensation wave frequency, phase reversal amount, and wavefront modulation parameters, ensuring that the compensation wave satisfies the phase conjugate conditions of having the same frequency, opposite phase, and matching propagation path as the original shrinkage wave. The second decision layer, based on the distribution of the liquid-phase isolated region and the thickness of the solidified shell in the first state information... The gradient determines the application location, timing, and amplitude of the local feeding pressure pulse, generating a pressure pulse parameter set. The optimization objectives of these pressure pulse parameters are minimizing the volume fraction of the isolated liquid phase region and improving the connectivity of the feeding channel. The compensation wave parameter set and the pressure pulse parameter set are input to the execution layer for parameter mapping. Consistency constraints are applied to the parts of the two sets of parameters that have the same application area and time window, serving as quantitative parameters. Inconsistent parts are treated as variable parameters to construct a joint optimization space. Within this joint optimization space, multiple rounds of iterative optimization calculations are performed with the objective functions of minimizing the shrinkage probability and process stability constraints. The process stability constraints include the injection speed fluctuation range, the upper limit of mold force, and equipment response time constraints. Iteration stops when the rate of change of the objective function for two consecutive iterations is less than a preset convergence threshold or the maximum number of iterations is reached, and the optimal die-casting control parameter combination is output.

[0025] Furthermore, the process information of the target casting is read, and a shrinkage probability threshold based on the casting quality is set; if the shrinkage probability cloud map meets the shrinkage probability threshold, the die-casting machine continues to execute PLC automatic die-casting control; if the shrinkage probability cloud map does not meet the shrinkage probability threshold, a shrinkage suppression control command is generated.

[0026] The process information of the target casting is read from the central control system or process database of the die-casting machine tool. This process information includes at least the material type, casting wall thickness distribution, hot spot location, injection speed curve, pressurization sequence, and historical defect rate data. A quality constraint model is constructed based on this process information, normalizing the allowable defect level, critical area quality requirements, and historical shrinkage rate to generate quality weight coefficients. Based on these quality weight coefficients and the importance distribution of each spatial region, the probability values ​​of each grid node in the shrinkage probability cloud map are weighted and evaluated to calculate the global risk index and critical area risk index. Based on the global risk index and critical area risk index, a set of... A shrinkage probability threshold is defined, wherein the shrinkage probability threshold adopts a hierarchical threshold structure, including a global control threshold and a local sensitivity threshold. The shrinkage probability cloud map is compared with the shrinkage probability threshold. When both the global risk indicator and the key area risk indicator are lower than the corresponding threshold, it is determined that the shrinkage probability threshold is met, and the die-casting machine maintains the current PLC automatic die-casting control mode. When any risk indicator exceeds the corresponding threshold, it is determined that the shrinkage probability threshold is not met, and a shrinkage suppression control command is generated. The shrinkage suppression control command includes at least a target control area identifier, a risk level identifier, and a corresponding control trigger signal, which is used to drive the subsequent shrinkage suppression module to perform control.

[0027] Furthermore, the construction of the shrinkage suppression module before activation includes: Using active phase interferometry, a multi-segment velocity curve and pressurization sequence at the end of the injection phase are defined as a compensation wave parameter sequence. Based on the compensation wave parameter sequence, a first decision layer is deployed with the compensation wave excitation parameters based on phase conjugation as the target. Phase conjugation analysis is performed based on the same frequency, opposite phase, and consistent wavefront geometry of the contraction wave characteristics. To address the risk of shrinkage marks at the solidification front, a second decision layer is deployed with the location orientation, quantification, and timing of the local feeding pressure pulse as the target. Lateral data interaction between the first and second decision layers is established to generate a shrinkage mark suppression module, wherein the shrinkage mark suppression module outputs the process parameter combination that minimizes the shrinkage mark probability.

[0028] Preferably, based on active phase interference as a constraint, the historical injection curves of the die-casting machine and the current process setting parameters are read. The final stage of injection is divided into multiple stages, constructing a multi-segment injection speed curve and a pressurization timing model. By discretizing the injection speed change rate and pressurization response time, a compensation wave parameter sequence is generated. This sequence includes at least frequency mapping parameters, phase modulation parameters, and wavefront modulation parameters. Based on the compensation wave parameter sequence, a first decision layer is deployed with phase conjugation-based compensation wave excitation parameters as the target. The contraction wave characteristic parameters from the second state information are used as input to extract the dominant frequency, phase distribution, and propagation path. A phase conjugation constraint model is constructed to ensure the compensation wave satisfies the conditions of identical frequency, phase reversal, and wavefront geometric matching, outputting a set of compensation wave excitation parameters. To address the risk of shrinkage marks at the solidification front, the distribution of the isolated liquid phase region in the first state information and the solidification... Using the shell thickness gradient as input, a local feeding pressure pulse control model is established, and a second decision layer is deployed. This model determines the pressure application location by discretizing the spatial positioning of the target area into a grid, and calculates the pressure amplitude and application timing based on the liquid phase feeding requirements, outputting a pressure pulse parameter set. A lateral data interaction mechanism is established between the first and second decision layers. This mechanism shares the target control area, time window, and risk intensity index, and collaboratively constrains and dynamically corrects the compensation wave parameter set and the pressure pulse parameter set to construct a joint parameter space. Within this joint parameter space, minimizing the shrinkage probability is used as the objective function, with constraints including injection stability, equipment response capability, and mold bearing capacity. Multiple rounds of parameter search and optimization calculations are performed until convergence conditions are met, generating the shrinkage suppression module. The shrinkage suppression module outputs the optimal combination of process parameters to drive subsequent die-casting control.

[0029] Furthermore, determining the combination of die-casting control parameters includes: According to the shrinkage mark suppression control command, the shrinkage mark suppression module is activated, and the solidification state of the casting and the shrinkage mark probability cloud map are input; the first decision layer and the second decision layer make decisions in parallel to determine the first control group and the second control group; by mapping the first control group and the second control group, taking the consistent part of the mapping as the quantitative part and the inconsistent part of the mapping as the variable, multiple rounds of iterative optimization are performed to output the die casting control parameter combination.

[0030] Preferably, according to the shrinkage mark suppression control command, the shrinkage mark suppression module is activated, and the solidification state of the casting and the shrinkage mark probability cloud map are converted into a unified feature input vector. The feature input vector includes at least the spatial coordinates of the target control area, risk intensity index, solidified shell thickness gradient, and shrinkage wave energy density distribution. The first decision layer and the second decision layer perform parameter decisions in parallel based on the feature input vector. The first decision layer outputs a first parameter control group, which includes the compensation wave frequency, phase reversal amount, wavefront modulation parameters, and action time window. The second decision layer outputs a second parameter control group, which includes the feeding pressure application location, pressure amplitude, application sequence, and duration parameters. A parameter mapping relationship is constructed based on the first and second parameter control groups. In this process, the consistent portions of two sets of parameters in the action space and time window are matched to form a quantitative constraint parameter set, while the inconsistent portions in the action space or time window are defined as variable parameter sets. Using the quantitative constraint parameter set as boundary conditions, a joint optimization model is established within the parameter space formed by the variable parameter sets, with the minimization of shrinkage probability as the objective function and the constraints of injection stability, equipment response, and mold structure safety as limiting conditions. Multi-round iterative optimization calculations are performed in the joint optimization model. In each round of iteration, the shrinkage probability cloud map is updated according to the current parameter combination and the objective function value is calculated. The iteration is terminated when the rate of change of the objective function between two adjacent rounds is less than a preset convergence threshold or the maximum number of iterations is reached, and the optimal die-casting control parameter combination is output.

[0031] Die casting process control is performed according to the combination of die casting control parameters.

[0032] Furthermore, performing die-casting process control based on the aforementioned combination of die-casting control parameters includes: Based on the shrinkage probability threshold, the expected die-casting state is determined, wherein the expected die-casting state is determined based on the residual energy of the shrinkage wave field; the die-casting process is controlled and managed using the die-casting control parameter combination, and when the expected die-casting state is met, the die-casting machine tool is switched to the pre-configured PLC automatic die-casting control.

[0033] Preferably, the expected die-casting state is determined based on the shrinkage probability threshold, wherein the expected die-casting state is determined based on the residual energy of the shrinkage wave field. Specifically, based on the shrinkage wave field distribution map in the second state information, the wave field energy density of each spatial grid node inside the target casting is calculated, and the wave field energy density is spatially integrated and time-weighted to obtain a residual energy index; the residual energy index is compared with a preset residual energy threshold, wherein the residual energy threshold is calibrated based on the casting material characteristics, wall thickness distribution, and historical process data. When the residual energy index is lower than a certain threshold, the expected die-casting state is determined. When the residual energy threshold is reached, the expected die-casting state is determined to be reached. The die-casting process is controlled and managed using the die-casting control parameter combination. During execution, the wave field residual energy index is updated in real time. When the wave field residual energy index is continuously lower than the residual energy threshold and meets the preset stable time window, the process state is determined to be stable. When the expected die-casting state is met, the die-casting machine tool switches from the shrinkage suppression module control mode to the pre-configured PLC automatic die-casting control mode. The switching process includes mapping the current die-casting control parameter combination to PLC control parameters and completing the handover of control and state synchronization.

[0034] Furthermore, it also includes: Before the target casting is opened, a stiffness distribution cloud map of the solidified shell is obtained; based on the stiffness distribution cloud map, the pre-pressure application position and application vector are located, and the pre-pressure is applied to the target casting by driving the mold ejector pin.

[0035] Preferably, based on the solidified shell thickness distribution and temperature field information in the first state information, and combined with the material elastic modulus and temperature coupling relationship model, the equivalent stiffness parameters of each spatial grid node are calculated to construct a solidified shell stiffness distribution cloud map, wherein the equivalent stiffness parameters are a quantitative representation of the local deformation resistance of the solidified shell; gradient analysis and extreme value region identification are performed on the stiffness distribution cloud map to extract stiffness abrupt change regions and low stiffness weak regions as candidate control regions; combined with the shrinkage probability cloud map, risk-weighted screening is performed on the candidate control regions to determine the target pre-stress application location; based on Based on the stiffness distribution and local structural orientation at the target pre-stress application location, a pre-stress application vector is calculated. The application vector includes an application direction and an application amplitude. The application direction is determined along the opposite direction of the stiffness gradient or the direction of the shrinkage compensation path. The application amplitude is calculated based on constraints of local stiffness and allowable deformation range. The pre-stress application location and application vector are input into the mold ejector drive control unit. By controlling the stroke, speed, and timing of the ejector, a directional pre-stress is applied to the target casting before mold opening to adjust the local stress distribution and suppress the formation of potential shrinkage defects.

[0036] In summary, the embodiments of this application have at least the following technical effects: First, ultrasonic probes distributed within the die-casting mold collect the position information of the solidification front. A topological sensing network is used to perform shrinkage risk inversion to determine the first state information. A piezoelectric sensor array collects the vibration signal of the shrinkage stress wave, and a shrinkage wave field distribution map is reconstructed using waveform analysis units as the second state information. Next, the first and second state information are fused to generate a shrinkage probability cloud map of the entire solidification process. If the shrinkage probability cloud map does not meet the shrinkage probability threshold, a shrinkage suppression module is activated. The first decision layer executes the compensation wave parameter decision for the shrinkage stress wave phase conjugate, and the second decision layer executes the local shrinkage pressure pulse parameter decision. The execution layer outputs mapping and optimization iterations to determine the die-casting control parameter combination. Finally, die-casting process control is performed based on the die-casting control parameter combination. This solves the technical problem in existing technologies where it is difficult to dynamically suppress shrinkage during the die-casting process, leading to unstable die-casting quality. It achieves dynamic suppression of shrinkage defects, thereby improving the forming quality of die-casting parts.

[0037] Example 2, based on the same inventive concept as the method for suppressing shrinkage marks in integral die-cast parts in the foregoing examples, such as... Figure 2 As shown, this application provides a shrinkage mark suppression system for integral die castings, wherein the system includes: First information acquisition module 11: Acquires the position information of the solidification front through ultrasonic probes distributed within the die-casting mold, and determines the first state information by performing shrinkage risk inversion based on the topology sensing network; Second information acquisition module 12: Acquires the vibration signal of the shrinkage stress wave through the piezoelectric sensor array, and reconstructs the shrinkage wave field distribution map as the second state information based on the waveform analysis unit; Information fusion module 13: Fuses the first state information and the second state information to generate a shrinkage probability cloud map of the entire solidification process; Decision execution module 14: If the shrinkage probability cloud map does not meet the shrinkage probability threshold, activates the shrinkage suppression module, executes the compensation wave parameter decision of the shrinkage stress wave phase conjugate through the first decision layer, executes the local shrinkage pressure pulse parameter decision through the second decision layer, and outputs mapping and optimization iteration of the execution layer to determine the die-casting control parameter combination; Process control module 15: Executes die-casting process control according to the die-casting control parameter combination.

[0038] Furthermore, the first information acquisition module 11 is used to perform the following methods: Based on the ultrasonic probes distributed within the die-casting mold cavity, position information of the solidification front is collected as the die-casting solidification process progresses; the position information is input into the topological sensing network of the solidification front to invert the current thickness distribution of the solidified shell and the spatial position of the liquid-phase isolated region, which serves as the first state information, wherein the liquid-phase isolated region characterizes the shrinkage mark risk area.

[0039] Furthermore, the second information acquisition module 12 is used to perform the following methods: A piezoelectric sensor array is embedded at a key location in the die-casting mold to collect vibration signals during the die-casting solidification process. These vibration signals cover the frequency, amplitude, phase, and propagation direction of the contraction wave, characterizing the particle displacement caused by the propagation of contraction stress within the target casting. The vibration signals are amplified and filtered, and a waveform analysis unit is activated to perform coherent superposition and reverse-time propagation analysis based on the elastic wave equation. This reconstructs the contraction wave field distribution map within the target casting, serving as the second state information. The contraction wave field distribution map identifies the location of standing wave nodes and energy concentration areas.

[0040] Furthermore, the information fusion module 13 is used to perform the following method: By fusing the first state information and the second state information, the solidification state of the casting is determined. Specifically, the solidification state of the casting is updated by dividing the die-casting solidification process into stages. Based on the solidification state of the casting, a shrinkage probability prediction based on a lightweight generator is performed to generate a shrinkage probability cloud map throughout the solidification process. The topology sensing network, waveform analysis unit, and lightweight generator are embedded in the central control of the die-casting machine tool and are activated during the die-casting operation.

[0041] Furthermore, the decision execution module 14 is used to execute the following methods: Read the process information of the target casting and set a shrinkage probability threshold based on the casting quality; if the shrinkage probability cloud map meets the shrinkage probability threshold, the die casting machine continues to execute PLC automatic die casting control; if the shrinkage probability cloud map does not meet the shrinkage probability threshold, a shrinkage suppression control command is generated.

[0042] Furthermore, the decision execution module 14 is used to execute the following methods: Using active phase interferometry, a multi-segment velocity curve and pressurization sequence at the end of the injection phase are defined as a compensation wave parameter sequence. Based on the compensation wave parameter sequence, a first decision layer is deployed with the compensation wave excitation parameters based on phase conjugation as the target. Phase conjugation analysis is performed based on the same frequency, opposite phase, and consistent wavefront geometry of the contraction wave characteristics. To address the risk of shrinkage marks at the solidification front, a second decision layer is deployed with the location orientation, quantification, and timing of the local feeding pressure pulse as the target. Lateral data interaction between the first and second decision layers is established to generate a shrinkage mark suppression module, wherein the shrinkage mark suppression module outputs the process parameter combination that minimizes the shrinkage mark probability.

[0043] Furthermore, the decision execution module 14 is used to execute the following methods: According to the shrinkage mark suppression control command, the shrinkage mark suppression module is activated, and the solidification state of the casting and the shrinkage mark probability cloud map are input; the first decision layer and the second decision layer make decisions in parallel to determine the first control group and the second control group; by mapping the first control group and the second control group, taking the consistent part of the mapping as the quantitative part and the inconsistent part of the mapping as the variable, multiple rounds of iterative optimization are performed to output the die casting control parameter combination.

[0044] Furthermore, the process control module 15 is used to perform the following methods: Based on the shrinkage probability threshold, the expected die-casting state is determined, wherein the expected die-casting state is determined based on the residual energy of the shrinkage wave field; the die-casting process is controlled and managed using the die-casting control parameter combination, and when the expected die-casting state is met, the die-casting machine tool is switched to the pre-configured PLC automatic die-casting control.

[0045] Furthermore, the process control module 15 is used to perform the following methods: Before the target casting is opened, a stiffness distribution cloud map of the solidified shell is obtained; based on the stiffness distribution cloud map, the pre-pressure application position and application vector are located, and the pre-pressure is applied to the target casting by driving the mold ejector pin.

[0046] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for suppressing shrinkage marks in integral die-cast parts, characterized in that, The method includes: The position information of the solidification front is collected by ultrasonic probes distributed in the die-casting mold, and the first state information is determined by the shrinkage risk inversion based on the topology sensing network. Vibration signals of contraction stress waves are acquired by a piezoelectric sensing array, and the contraction wave field distribution map is reconstructed based on the waveform analysis unit as the second state information. By integrating the first state information and the second state information, a shrinkage probability cloud map of the entire solidification process is generated; If the shrinkage probability cloud map does not meet the shrinkage probability threshold, the shrinkage suppression module is activated. The compensation wave parameter decision of the shrinkage stress wave phase conjugate is executed through the first decision layer, and the local shrinkage pressure pulse parameter decision is executed through the second decision layer. The layer output mapping and optimization iteration are performed to determine the die casting control parameter combination. Die casting process control is performed according to the combination of die casting control parameters.

2. The method for suppressing shrinkage marks in an integral die-cast part as described in claim 1, characterized in that, The location information of the solidification front is collected, and the first state information is determined by inverting the shrinkage risk based on the topology sensing network, including: Based on the ultrasonic probes distributed within the die-casting mold cavity, position information of the solidification front is collected as the die-casting solidification process progresses; The location information is input into the topological sensing network of the solidification front to retrieve the thickness distribution of the current solidified shell and the spatial location of the isolated liquid phase region, which serves as the first state information. The isolated liquid phase region represents the shrinkage mark risk area.

3. The method for suppressing shrinkage marks in an integral die-cast part as described in claim 1, characterized in that, The vibration signal of the contraction stress wave is acquired, and the contraction wave field distribution map is reconstructed based on the waveform analysis unit as the second state information, including: A piezoelectric sensor array is embedded in a key position of the die-casting mold to collect vibration signals during the die-casting solidification process. The vibration signals cover the frequency, amplitude, phase and propagation direction of the shrinkage wave, which characterizes the particle displacement caused by the shrinkage stress propagating inside the target casting. The vibration signal is amplified and filtered, and the waveform analysis unit is activated to perform coherent superposition and reverse time propagation analysis on the vibration signal based on the elastic wave equation. The contraction wave field distribution map inside the target casting is reconstructed as the second state information. The contraction wave field distribution map is marked with the position of the standing wave node and the energy concentration area.

4. The method for suppressing shrinkage marks in an integral die-cast part as described in claim 1, characterized in that, Generate a probability cloud map of the shrinkage mark throughout the solidification process, including: The solidification state of the casting is determined by integrating the first state information and the second state information, wherein the solidification state of the casting is updated by dividing the die casting solidification process into stages. Based on the solidification state of the casting, a shrinkage probability prediction based on a lightweight generator is performed to generate a shrinkage probability cloud map throughout the solidification process. The topology sensing network, waveform analysis unit, and lightweight generator are embedded in the central control unit of the die-casting machine tool and are activated during the die-casting operation.

5. The method for suppressing shrinkage marks in an integral die-cast part as described in claim 4, characterized in that, Read the process information of the target casting and set a shrinkage probability threshold based on the casting quality; If the shrinkage probability cloud map meets the shrinkage probability threshold, the die-casting machine tool continues to execute PLC automatic die-casting control; If the shrinkage probability cloud map does not meet the shrinkage probability threshold, a shrinkage suppression control command is generated.

6. The method for suppressing shrinkage marks in an integral die-cast part as described in claim 5, characterized in that, Before activating the shrinkage suppression module, the construction of the shrinkage suppression module includes: Using active phase interferometry, we define multi-segment velocity curves and pressurization timing at the end of the injection phase as a sequence of compensation wave parameters; Based on the compensation wave parameter sequence, a first decision layer is deployed with the compensation wave excitation parameters based on phase conjugation as the target, wherein phase conjugation analysis is performed based on the characteristics of contraction waves with the same frequency, opposite phase, and consistent wavefront geometry. To address the risk of shrinkage marks at the solidification front, a second decision-making layer is deployed with the goal of determining the location, quantity, and timing of local feeding pressure pulses. A lateral data interaction is established between the first decision layer and the second decision layer to generate a shrinkage suppression module, wherein the shrinkage suppression module outputs a combination of process parameters that minimizes the shrinkage probability.

7. The method for suppressing shrinkage marks in an integral die-cast part as described in claim 6, characterized in that, Determine the combination of die-casting control parameters, including: According to the shrinkage mark suppression control command, activate the shrinkage mark suppression module and input the casting solidification state and shrinkage mark probability cloud map; The first and second decision-making levels make decisions in parallel to determine the first and second control groups. By mapping the first and second parameter control groups, taking the consistent portion as the quantitative part and the inconsistent portion as the variable, multiple rounds of iterative optimization are performed to output the die-casting control parameter combination.

8. The method for suppressing shrinkage marks in an integral die-cast part as described in claim 1, characterized in that, Die casting process control is performed according to the aforementioned die casting control parameter combination, including: The expected die-casting state is determined based on the shrinkage probability threshold, wherein the expected die-casting state is determined based on the residual energy of the shrinkage wave's wave field. The die-casting process is controlled and managed using the aforementioned combination of die-casting control parameters. When the expected die-casting state is met, the die-casting machine tool switches to the pre-configured PLC automatic die-casting control.

9. The method for suppressing shrinkage marks in an integral die-cast part as described in claim 1, characterized in that, The method further includes: Before the target casting is molded, obtain the stiffness distribution cloud map of the solidified shell; Based on the stiffness distribution cloud map, the location and vector of the pre-pressure application are determined, and the pre-pressure is applied to the target casting by driving the mold ejector pin.

10. A shrinkage mark suppression system for integral die-cast parts, characterized in that, A method for suppressing shrinkage marks in an integral die-cast part according to any one of claims 1-9, the system comprising: First information acquisition module: The position information of the solidification front is collected by ultrasonic probes distributed in the die-casting mold, and the first state information is determined by the shrinkage risk inversion based on the topology sensing network. The second information acquisition module: acquires the vibration signal of the contraction stress wave through the piezoelectric sensor array, and reconstructs the contraction wave field distribution map based on the waveform analysis unit as the second state information; Information fusion module: fuses the first state information and the second state information to generate a shrinkage probability cloud map of the entire solidification process; Decision execution module: If the shrinkage probability cloud map does not meet the shrinkage probability threshold, the shrinkage suppression module is activated. The first decision layer executes the compensation wave parameter decision of the shrinkage stress wave phase conjugate, and the second decision layer executes the local shrinkage pressure pulse parameter decision. The execution layer outputs mapping and optimization iteration to determine the die casting control parameter combination. Process control module: Performs die-casting process control according to the combination of die-casting control parameters.

Images