A method and system for quality control of a plastic article production process

Abstract

This invention discloses a quality control method and system for the production process of plastic products, relating to the field of plastic production supervision technology. The invention generates a three-dimensional melt front advancing surface during the injection molding process and detects whether local nodes experience sudden pressure drops or temperature breaks based on the arrival time difference matrix and filling temperature matrix. This determines whether filling imbalance occurs at the corresponding location and identifies it as stagnation or overfilling anomalies. The invention outputs the location information of the abnormal area and its offset relative to the normal front surface. Based on the offset of the abnormal area relative to the normal front surface, the interface contour line is extracted on the three-dimensional melt front advancing surface, and the corresponding fractal dimension is calculated. The fractal dimension is used as a quantitative indicator of melt merging. When the fractal dimension of the weld line area is determined to be higher than a preset threshold, the coordinates and risk level of the abnormal area are output. The parameters of the injection molding machine nozzle are optimized based on the coordinates and risk level of the abnormal area.

Description

Technical Field

[0001] This invention relates to the field of plastic production supervision technology, specifically to a method and system for quality control in the production process of plastic products. Background Technology

[0002] Injection molding is one of the most important production methods for plastic products. The quality of these products is influenced by a combination of factors, including melt filling behavior, pressure and temperature field distribution, molecular orientation, and weld line strength. Traditional injection molding quality control methods rely heavily on mold design experience, offline process parameter trial and error, and finished product sampling, making it difficult to perceive the real-time advancement of the melt front within the mold cavity. Existing online monitoring methods, such as cavity pressure sensors or cavity temperature sensors, are typically discretely arranged at single points, only acquiring local signals and failing to reconstruct the dynamic evolution of the global melt flow front. Furthermore, they struggle to quantify the interface morphology and molecular chain entanglement when two melt streams converge.

[0003] When abnormalities such as filling imbalance, stagnation, or overfill occur, operators often only passively adjust parameters after a large number of defective products have been produced, resulting in a significant lag in response. Although there has been research on mold flow analysis based on simulation software in recent years, factors such as raw material batch fluctuations, mold temperature drift, and environmental interference in actual production cause significant deviations between offline simulations and real-world conditions. Therefore, there is an urgent need for a quality control method that can construct a three-dimensional melt front advancement surface in real time, automatically detect filling abnormalities, and quantify weak areas of weld lines to achieve online diagnosis and closed-loop adaptive optimization of the injection molding process. To this end, a quality control method and system for the plastic product manufacturing process is provided. Summary of the Invention

[0004] The purpose of this invention is to provide a quality control method and system for the production process of plastic products, so as to solve the problems in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A quality control method for the production process of plastic products includes the following steps: Step S1: Install a thin-film pressure-temperature composite sensor on the surface of the mold cavity, and then collect the arrival time of the melt front, pressure distribution and temperature change of each sensor node in real time during the injection molding process to form an arrival time difference matrix and a filling temperature matrix, as well as the corresponding three-dimensional melt front advancement surface. Step S2: Based on the arrival time difference matrix and filling temperature matrix, detect whether there is a sudden drop in pressure or temperature fault at local nodes, and then determine whether there is filling imbalance at the corresponding location and identify it as stagnation or overfilling abnormality, and output the location information of the abnormal area and the offset relative to the normal front. Step S3: Extract the interface contour line on the three-dimensional melt front advancement surface based on the offset of the abnormal area relative to the normal front surface, calculate the corresponding fractal dimension, and use the fractal dimension as a quantitative indicator when the melt merges. When the fractal dimension of the weld line area is higher than the preset threshold, mark it as a potential weak point in strength, and output the coordinates and risk level of the abnormal area in combination with the location information of the abnormal area. Step S4: Optimize the parameters of the injection molding machine nozzle based on the coordinates of the abnormal area and the risk level, and continuously monitor the offset of the abnormal area relative to the normal front face until no pressure drop or temperature break occurs.

[0006] Furthermore, the process of mounting a thin-film pressure-temperature composite sensor on the surface of the mold cavity includes: Obtain the three-dimensional geometric model of the injection mold for the target plastic product, and lay out the sensor on the surface of the mold cavity. With the gate position as the origin, along the expected main direction of melt flow, arrange the sensor nodes in an equidistant or non-equidistant grid form from the area near the gate to the farthest end of the cavity. A thin-film pressure-temperature composite sensor was selected as the sensing unit for the sensor node, and each sensor node was assigned a number. 1，1 c 1，2 c i，j , where i represents the row of the sensor node, j represents the column of the sensor node, and i and j are natural numbers greater than 0.

[0007] Furthermore, the process of real-time acquisition of melt front arrival time, pressure distribution, and temperature changes at each sensor node during injection molding to form an arrival time difference matrix, a filling temperature matrix, and the corresponding three-dimensional melt front advancement surface includes: The injection molding machine is started to begin a complete injection cycle. After the molten plastic is injected into the mold cavity from the gate, the melt front begins to move forward. Each sensor node continuously records the pressure signal and temperature signal at a preset sampling frequency. Both the pressure signal and the temperature signal are curves that change with time. For both pressure and temperature signals, when the sensor node determines that both signals have arrived simultaneously, the arrival time t of the melt front at the sensor node is confirmed. i，j ; From arrival time t i，j The pressure and temperature change curves during the period from the start of the filling stage to the end of the filling stage are obtained, and the stable filling pressure and average filling temperature of each sensor node during the filling stage are extracted. Based on the arrival time t of the melt front at all sensor nodes i，j Construct an arrival time matrix and a filling temperature matrix based on the average filling temperature of each sensor node; By integrating a pressure-temperature sensor array on the surface of the mold cavity, the moment when the melt arrives, the instantaneous pressure and temperature are recorded by each sensor node. The sensed data is then spatiotemporally interpolated and surface reconstructed on the three-dimensional geometric model of the mold to generate a dynamic surface that advances forward over time, which is the three-dimensional melt front advancing surface.

[0008] Furthermore, the process of detecting whether a local node experiences a sudden drop in pressure or a temperature fault based on the arrival time difference matrix and the filling temperature matrix includes: Obtain the arrival time difference matrix for the current injection cycle, and obtain the global mean and standard deviation of all elements in the arrival time difference matrix. For each sensor node, compare its arrival time difference with the arrival time difference of its neighboring nodes. Define the neighboring nodes as the set of sensor nodes that have a four-neighbor or eight-neighbor relationship with the sensor node on the surface of the mold cavity. Obtain the local arrival time difference index of the sensor node relative to its neighboring nodes, and determine whether a local fill time anomaly occurs at the corresponding sensor node based on the relationship between the local arrival time difference index and a preset threshold. Obtain the temperature difference matrix for the current injection molding cycle. Similar to detecting a sudden drop in pressure, obtain the global average and standard deviation of all elements in the temperature difference matrix. For each sensor node, obtain the local outlier degree of its temperature difference and classify temperature fault anomalies based on the local outlier degree.

[0009] Furthermore, the process of determining whether a filling imbalance has occurred at the corresponding location and identifying it as a stagnation or overfill anomaly, and outputting the location information of the abnormal area and its offset relative to the normal front surface, includes: For each sensor node, an abnormal state vector is defined. Based on the abnormal state vector, it is determined whether a stagnation abnormality has occurred at each sensor node. The spatial locations of all sensor nodes that are determined to have stagnation abnormality or overcharge abnormality are clustered. Adjacent abnormal sensor nodes in the space are grouped into the same abnormal region; otherwise, they are not grouped. The abnormal region is characterized by a set of sensor node coordinates, and the geometric center of the abnormal region is used as the abnormal location information. The offset of the abnormal region relative to the normal front is calculated. The normal front is defined as: at the current filling time, the normal melt front propulsion surface is generated by interpolation based on the arrival time of the global sensor nodes. The normal melt front propulsion surface assumes that the melt is filled in an ideal radial flow mode and does not consider local defects. For each sensor node within the abnormal region, its actual arrival time is obtained. The difference between the theoretical arrival time of the corresponding sensor node when the melt front propagates in the ideal mode is used as the time offset. The time offset is converted into a spatial offset. The sign of the spatial offset indicates whether the abnormal region is lagging or leading the normal front. Then, the spatial offsets of each sensor node in the abnormal region are accumulated to obtain the total offset.

[0010] Furthermore, the process of extracting the interface contour line on the three-dimensional melt front advancement surface based on the offset of the abnormal region relative to the normal front surface and calculating the corresponding fractal dimension includes: Based on the abnormal location information and overall offset of the abnormal area, it is divided into attention areas. Within the attention areas, the final state of the three-dimensional melt front advancement surface at the end of the current injection cycle is obtained. In the final state, the two melts have completely merged, and their boundary is the physical location of the weld line. On the three-dimensional melt front advancing surface within the attention area, when two melts merge, the surface tension and cooling contraction are inconsistent, and the surface presents a valley line with a local minimum or curvature abrupt change. The second derivative of the surface is obtained, and the eigenvalues ​​of the second derivatives that are greater than a preset threshold are selected. Then, the gradient is diffused outward from the center of the abnormal region to track the point with the largest gradient change and connect them to form a spatial curve. The spatial curve is the three-dimensional spatial representation of the weld line on the three-dimensional melt front advancing surface. Projecting it onto a two-dimensional plane perpendicular to the mold opening direction yields a two-dimensional curve, which is defined as the interface contour line. At the same time, the fractal dimension of the interface contour line is obtained.

[0011] Furthermore, using fractal dimension as a quantitative indicator for melt fusion, when the fractal dimension of the weld line region is determined to be higher than a preset threshold, it is marked as a potential weak point. The process of outputting the coordinates and risk level of the abnormal region, combined with the location information of the abnormal region, includes: Based on the negative correlation between fractal dimension and the bonding strength of the weld line region, fractal dimension is used as a quantitative indicator for melt fusion. A state threshold range is set, and the coordinates and risk level of the abnormal area are output according to the relationship between the fractal dimension and the state threshold range. The coordinates of the abnormal area are defined by the abnormal location information of the abnormal area.

[0012] Furthermore, the process of optimizing the injection molding machine nozzle parameters based on the coordinates of the abnormal area and the risk level includes: Based on the coordinates of the abnormal area, the specific location of the abnormal area in the mold cavity can be determined. If the abnormal area is a stagnant flow abnormality and the risk level is high or medium, the parameter adjustment includes increasing the melt temperature, adjusting the injection speed, and increasing the holding pressure and time. If the abnormal area is a filling abnormality and the risk level is high or medium risk, parameter adjustments include reducing the injection speed, reducing the holding pressure, and adjusting the stroke of the injection and holding pressure switching point. The optimized injection molding process parameters are downloaded to the injection molding machine controller and will take effect in the next injection molding cycle.

[0013] Furthermore, the process of continuously monitoring the offset of the abnormal region relative to the normal front until no pressure drop or temperature fault occurs includes: During subsequent injection molding production with adjusted injection molding process parameters, the current spatial offset of the abnormal area is reacquired. When the injection cycle is completed, the rate of change of the offset relative to the previous injection cycle is obtained. If the absolute value of the rate of change of the offset gradually decreases, it indicates that the optimization measures are effective and the size of the abnormal area has converged. Then, in the next five consecutive injection cycles, if no new pressure drop or temperature fault is detected near the abnormal area and the fractal dimension is less than the state threshold range, it is determined that the abnormal area control target has been achieved and the injection process has stabilized in a qualified state. Otherwise, the parameters are readjusted until no new pressure drop or temperature fault is detected near the abnormal area and the fractal dimension is less than the state threshold range.

[0014] A quality control system for the production process of plastic products includes a real-time injection molding monitoring module, an anomaly risk analysis module, and a status control module; The injection molding real-time monitoring module is used to install a thin-film pressure-temperature composite sensor on the surface of the mold cavity, and then collect the arrival time of the melt front, pressure distribution and temperature change of each sensor node in real time during the injection molding process, forming an arrival time difference matrix and a filling temperature matrix, as well as a corresponding three-dimensional melt front advancement surface. Based on the arrival time difference matrix and the filling temperature matrix, it detects whether there is a sudden drop in pressure or temperature discontinuity at local nodes, and then determines whether there is a filling imbalance at the corresponding position and identifies it as stagnation or overfilling abnormality, and outputs the positioning information of the abnormal area and the offset relative to the normal front surface. The abnormal risk analysis module is used to extract the interface contour line on the three-dimensional melt front advancement surface based on the offset of the abnormal area relative to the normal front surface, calculate the corresponding fractal dimension, and use the fractal dimension as a quantitative indicator when the melt merges. When the fractal dimension of the weld line area is higher than the preset threshold, it is marked as a potential weak point. Combined with the location information of the abnormal area, the coordinates and risk level of the abnormal area are output. The state control module is used to optimize the parameter adjustment of the injection molding machine nozzle according to the coordinates of the abnormal area and the risk level, and continuously monitor the offset of the abnormal area relative to the normal front surface until no pressure drop or temperature break occurs.

[0015] The technical effects and advantages provided by the present invention in the above technical solution are as follows: 1. This invention integrates a thin-film pressure-temperature composite sensor array to construct a three-dimensional melt front advancement surface in real time, which visualizes the previously invisible melt flow behavior in the cavity. It can accurately locate abnormal areas of filling imbalance and their offset relative to the normal front surface, providing a direct basis for closed-loop control. 2. This invention extracts the interface contour line of the weld line region and calculates its fractal dimension. The fractal dimension is used as a quantitative indicator of the degree of molecular chain entanglement, which realizes the automatic identification and risk classification of potential weak points. Combined with the adaptive optimization of injection molding process parameters, it effectively reduces the scrap rate of products. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0017] Figure 1 This is a flowchart of the method of the present invention.

[0018] Figure 2 This is a system block diagram of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see Figure 1 As shown, a quality control method for the production process of plastic products includes the following steps: Step S1: Install a thin-film pressure-temperature composite sensor on the surface of the mold cavity, and then collect the arrival time of the melt front, pressure distribution and temperature change of each sensor node in real time during the injection molding process to form an arrival time difference matrix and a filling temperature matrix, as well as the corresponding three-dimensional melt front advancement surface. Step S2: Based on the arrival time difference matrix and filling temperature matrix, detect whether there is a sudden drop in pressure or temperature fault at local nodes, and then determine whether there is filling imbalance at the corresponding location and identify it as stagnation or overfilling abnormality, and output the location information of the abnormal area and the offset relative to the normal front. Step S3: Extract the interface contour line on the three-dimensional melt front advancement surface based on the offset of the abnormal area relative to the normal front surface, calculate the corresponding fractal dimension, and use the fractal dimension as a quantitative indicator when the melt merges. When the fractal dimension of the weld line area is higher than the preset threshold, mark it as a potential weak point in strength, and output the coordinates and risk level of the abnormal area in combination with the location information of the abnormal area. Step S4: Optimize the parameters of the injection molding machine nozzle based on the coordinates of the abnormal area and the risk level, and continuously monitor the offset of the abnormal area relative to the normal front face until no pressure drop or temperature break occurs.

[0021] Furthermore, step S1 is implemented through the following process: Step S101: Install a thin-film pressure-temperature composite sensor on the surface of the mold cavity. The specific process includes: A three-dimensional geometric model of the injection mold for the target plastic product is obtained. The shape of the mold cavity is consistent with the shape of the target plastic product. Sensors are arranged on the surface of the mold cavity. With the gate position as the origin, sensor nodes are arranged in an equidistant or unequal-spacing grid along the expected main direction of melt flow, from the area near the gate to the farthest end of the cavity. A thin-film pressure-temperature composite sensor is selected as the sensing unit of the sensor node. The thin-film pressure-temperature composite sensor has a multi-layer structure. The bottom layer is a flexible polyimide substrate, the middle layer contains a piezoresistive pressure sensing unit and a thermocouple temperature sensing unit, and the top layer is a protective layer. The thin-film pressure-temperature composite sensor is embedded and integrated into the inner surface of the mold cavity according to a predetermined layout, ensuring that the sensor sensing surface is flush with or slightly lower than the inner wall of the cavity (the recess is less than 0.05 mm) to avoid forming raised marks on the surface of the injection molded part. Assign a number to each sensor node, c 1，1 c 1，2 c i，j , where i represents the row of the sensor node (along the main direction of melt flow), j represents the column of the sensor node (perpendicular to the main direction of melt flow), and i and j are natural numbers greater than 0.

[0022] Step S102: During the injection molding process, the arrival time of the melt front, pressure distribution, and temperature changes of each sensor node are collected in real time to form an arrival time difference matrix and a filling temperature matrix, as well as the corresponding three-dimensional melt front advancement surface. The specific process includes: The injection molding machine is started to begin a complete injection cycle. After the molten plastic is injected into the mold cavity from the gate, the melt front begins to move forward. Each sensor node continuously records the pressure signal and temperature signal at a preset sampling frequency. At each sensor node, the pressure signal and temperature signal are curves that change with time. For pressure signals, when the melt front has not yet reached a certain sensor node, the pressure value measured by that sensor node is zero (or close to the ambient atmospheric pressure). When the melt front just covers the sensor node, the pressure value will experience a step increase, rapidly jumping from zero to a positive pressure value. The time point at which the pressure signal at each sensor node experiences its first positive change from zero is defined as the arrival time of the melt front at that sensor node, denoted as t. i，j The criteria for determining the positive mutation are: the rate of increase of the pressure value within a continuous time window of Δτ (usually Δτ is set to 0.5 milliseconds to 1 millisecond) exceeds a preset slope threshold. For temperature signals, when the melt front has not yet arrived, the sensor node measures the temperature of the mold cavity wall (usually maintained within a certain range by the mold cooling system, such as 40°C to 80°C). Once the melt front arrives, because the temperature of the molten plastic is much higher than the mold temperature (e.g., 200°C to 300°C), the temperature value of the sensor node will rise rapidly. The rising edge of the temperature signal can be used simultaneously as an auxiliary criterion for the arrival of the melt front. Therefore, when both the pressure signal and the temperature signal are simultaneously determined to have arrived, the arrival time t of the melt front at that sensor node can be confirmed. i，j ; Record the arrival time t for each sensor node after the melt front arrives. i，j The pressure change curve P during the period from the start of the injection molding process to the end of the filling stage (usually marked by the cavity being filled and the injection molding machine switching from speed control to pressure control). i，j (t) and temperature change curve T i，j (t), and extract the steady-state filling pressure of each sensor node during the filling phase. and average filling temperature ; Based on the arrival time t of the melt front at all sensor nodes i，j Construct the arrival time matrix T arr Its dimension is consistent with the number of rows and columns of the sensor nodes, and the elements in the matrix [T] arr ] i，j =t i，j ; The arrival time t of the gate position (usually the reference origin of the sensor array) 0，0 Using the reference point as the baseline, calculate the time difference of arrival of each sensor node relative to the reference point: This forms the arrival time difference matrix ΔT arr , where element [ΔT arr ] i，j =Δt i，j ; Similarly, based on the average fill temperature of each sensor node , constructing a filling temperature matrix T temp , selecting the average value of filling temperatures of all sensor nodes as the reference temperature T ref , calculating the temperature difference of each sensor node relative to the reference temperature: , thus obtaining the temperature difference matrix ΔT temp , wherein the element [ΔT temp i，j =ΔTemp i，j ; Further, a three-dimensional melt front propagation surface is constructed, wherein the three-dimensional melt front propagation surface is a dynamic surface that propagates forward with time during the injection filling stage: when molten plastic enters a mold cavity through a gate, the foremost end of the flow of the molten plastic is called a melt front; through a pressure-temperature sensor array integrated on the surface of the mold cavity, the time when the melt arrives, the instantaneous pressure and the temperature sensed by each sensor node are recorded, and the sensed data are subjected to spatio-temporal interpolation and surface reconstruction on the three-dimensional geometric model of the mold, so as to generate the dynamic surface that propagates forward with time, which is the three-dimensional melt front propagation surface; The specific construction process is as follows: taking the three-dimensional geometric model of a mold as a spatial carrier, the spatial coordinates (x i，j , y i，j , z i，j ) of all sensor nodes are used as discrete data points, and the arrival time of melt at each sensor node is added as t i，j information in the time dimension; for any given filling time t*, (t* is between the earliest arrival time and the latest arrival time), all sensor nodes satisfying t i，j <t* are selected as sensor nodes already covered by the melt front, the spatial coordinates of the selected sensor nodes are used as input, a closed contour line is generated on the surface of the mold cavity, the closed contour line is the melt front line at the time t*, and further the normal height (distance relative to the bottom surface of the mold cavity) of each point on the melt front line is associated with the time t*, so as to construct a three-dimensional surface that changes with time, the equation of the three-dimensional surface can be expressed as S(x, y, t*)=z front (x, y, t*), wherein z front represents the coordinate in the cavity height direction where the melt front arrives at the time t* at the position (x, y), and further the melt front surfaces of continuous time series are superimposed, so as to obtain the dynamic three-dimensional melt front propagation surface that propagates forward with time.

[0023] Further, said step S2 is implemented by the following process: Step S201, detecting whether a pressure sudden drop or a temperature fault occurs at a local node based on an arrival time difference matrix and a filling temperature matrix, the specific process comprising: obtaining an arrival time difference matrix ΔT of the current injection molding cycle arr , calculating ΔT​arr The global mean μΔt and standard deviation σΔt of all elements in the dataset are given, and for each sensor node (i, j), its arrival time difference Δt is calculated. i，j By comparing the arrival time difference with that of neighboring nodes, a neighboring node is defined as the set N(i,j) of sensor nodes that have a four-neighbor or eight-neighbor relationship with sensor node (i,j) on the surface of the mold cavity. Calculate the local arrival time dissimilarity index of sensor node (i, j) with respect to its neighboring nodes: Extract the stable filling pressure during the filling phase of sensor node (i, j) exhibiting localized filling time anomalies. Obtain the global average value μ of the sensor node at the filling pressure. P With standard deviation σ P Then, the Z-score of the filling pressure of the corresponding sensor node is calculated: , when If the pressure drop is less than -2 and continues to decrease by more than 15% over 2 to 3 consecutive injection cycles, it is determined that a pressure drop has occurred at the corresponding sensor node. The pressure drop indicates that the melt filling in the corresponding area is insufficient. Otherwise, it is determined that no pressure drop has occurred. Obtain the temperature difference matrix ΔT for the current injection molding cycle. temp Similar to detecting a sudden drop in pressure, calculate ΔT. tem The global average value μΔT of all elements in tem With standard deviation σΔT tem For each sensor node (i, j), calculate the degree of local outlier in its temperature difference; The temperature tomography is defined as a temperature difference exceeding a preset physical threshold between two or more spatially adjacent sensor nodes, forming a discontinuous temperature field distribution. The detection process includes: For any given sensor node, compare the temperature difference with all neighboring sensor nodes. If there is any neighboring sensor node whose temperature difference is greater than a preset difference threshold (set to 5℃ to 10℃), then mark the connection edge between the corresponding sensor node and the corresponding neighboring sensor node as a temperature fault edge; otherwise, do not mark it. All sensor nodes marked as temperature fault edges are aggregated to form a temperature fault region. If multiple temperature fault edges exist within a continuous region, and the average filling temperature of the region is significantly lower than (or higher than) the surrounding non-fault regions, then the region is determined to have a temperature fault anomaly.

[0024] Step S202: Determine whether a filling imbalance has occurred at the corresponding location and identify it as a stagnation or overfilling anomaly, and output the location information of the abnormal area and its offset relative to the normal front surface. The specific process includes: For each sensor node, define an anomaly state vector. i，j = [A time A press A temp Each component has a value of 0 (no anomaly) or 1 (anomaly exists). When A time When =1, A press =1 (sudden pressure drop) or A temp =1 (temperature fault), then it is determined that a flow stagnation anomaly has occurred at sensor node (i, j). The flow stagnation anomaly indicates that the melt filling speed in the corresponding area is too slow, which may lead to local underfilling or a decrease in mechanical properties of the product. When A time When =1, A press =0 (no sudden drop in pressure, or even high pressure), and A temp =1 (no temperature fault), then it is determined that an overcharge anomaly has occurred at sensor node (i, j). The overcharge anomaly indicates that the melt filling speed in the corresponding area is too fast, which may cause over-holding pressure, flash, or internal residual stress concentration. Cluster the spatial locations of all sensor nodes that are determined to be stagnation or overcharge abnormalities. Group adjacent abnormal sensor nodes (with Euclidean distance less than the preset cluster radius, which is usually 2 to 3 times the sensor spacing) into the same abnormal region; otherwise, do not group them. The abnormal region is defined by a set of sensor node coordinates. Characterize and calculate the geometric center of the anomalous region. As abnormal location information; Calculate the offset of the abnormal region relative to the normal front surface. The normal front surface is defined as: at the current filling time t*, the normal melt front advancement surface is generated by interpolation based on the arrival time of the global sensor nodes. The normal melt front advancement surface assumes that the melt is filled in an ideal radial flow mode and does not consider local defects. Specifically, for each sensor node within the abnormal region, its actual arrival time t is calculated. actual =t i，j When the melt front propagates in the ideal mode, the theoretical arrival time t of the corresponding sensor node is... ideal (x k y k The difference between ) is used as the time offset δ tk =t actual -t ideal According to the average velocity v of the melt front front (via v) front =L flow / max(t i，j ) obtained, for Lflow (Length of the cavity in the main direction), converting the time offset into a spatial offset δ dk =v front *δ tk The sign of the spatial offset indicates whether the abnormal region is lagging (stagnant) or leading (overcharged) relative to the normal front. The spatial offsets of each sensor node in the abnormal region are then summed to obtain the total offset.

[0025] Furthermore, step S3 is implemented through the following process: Step S301: Extract the interface contour line on the three-dimensional melt front advancement surface based on the offset of the abnormal region relative to the normal front surface, and calculate the corresponding fractal dimension. The specific process includes: When it is determined that there is stagnation or overfilling abnormality, the abnormal area is often a high-risk area for weld line formation. Weld line refers to the physical interface formed when two or more melts from different directions meet in the mold cavity and fail to fuse completely. Near the abnormal area, due to the unbalanced filling, the leading edges of the two melts often meet at a non-ideal angle, forming a significant weld line. Based on the anomaly location information and overall offset of the abnormal area, the spatial range in which the weld line may appear is determined. Using the anomaly location information as the center and R as the reference point... search =2*|Total Offset|+L margin The region of interest is divided into regions with radius L. margin Take 5 mm to 10 mm; Within the area of ​​attention, obtain the final state of the three-dimensional melt front advancement surface at the end of the current injection cycle (i.e., all sensor nodes have detected the melt front and the filling stage has just ended). In this final state, the two melts have completely merged, and their junction is the physical location of the weld line. On the three-dimensional melt front advancing surface within the attention area, due to the inconsistency of surface tension and cooling contraction when the two melts merge, the surface will present a valley line with a local minimum or abrupt curvature change. The weld line usually corresponds to the region where the gradient direction changes abruptly, that is, the first derivative of the surface is discontinuous. Further calculation of the second derivative of the surface reveals that the weld line position often corresponds to the characteristic value of the second derivative that is greater than a preset threshold. Then, the gradient changes outward from the center of the abnormal region, and the point with the largest gradient change is traced and connected to form a spatial curve. The spatial curve is the three-dimensional spatial representation of the weld line on the three-dimensional melt front advancing surface. Projecting it onto a two-dimensional plane perpendicular to the mold opening direction yields a two-dimensional curve. This two-dimensional curve is defined as the interface contour line. The interface outline is formed when two streams of melt from different directions meet in the mold cavity, creating a physical bonding interface. The two-dimensional projection of the physical bonding interface on the weld line cross-sectional image is a meandering line. The fractal dimension of the interface contour line is calculated using the box-counting dimension method. The fractal dimension is used to quantitatively describe the complexity of irregular curves or surfaces. The value of the fractal dimension is usually between 1 and 2: the closer the value is to 1, the closer the curve is to a smooth straight line; the closer the value is to 2, the more meandering the curve is and the stronger the space-filling ability is.

[0026] Step S302: Using the fractal dimension as a quantitative indicator for melt fusion, when the fractal dimension of the weld line region is determined to be higher than a preset threshold, it is marked as a potential weak point. Combined with the location information of the abnormal region, the coordinates and risk level of the abnormal region are output. The specific process includes: When melts merge, if the front-end temperatures of the two melts are high, the pressure is sufficient, and the meeting angle is appropriate, the molecular chains have enough time to diffuse and entangle, resulting in a relatively smooth interface profile with a fractal dimension close to 1. Conversely, if the melt temperature is low, the pressure is insufficient, or the meeting angle is large, the molecular chains do not have enough time to entangle, and the interface profile will exhibit a highly meandering and complex branching shape with a fractal dimension close to 2. Therefore, the fractal dimension is negatively correlated with the bonding strength of the weld line region. The fractal dimension is used as a quantitative indicator for melt fusion, and a state threshold range is set (usually [1.25, 1.65]). Based on the relationship between the fractal dimension and the state threshold range, the coordinates and risk level of the abnormal region are output, where the coordinates of the abnormal region are based on the geometric center of the abnormal region. definition; If the fractal dimension is less than the state threshold range, it is determined that the molecular chain entanglement in the weld line region is good, and the mechanical strength can reach more than 90% of the strength of the matrix material. The risk level is low, indicating that it is currently within the allowable range of production errors. If the fractal dimension is within the state threshold range, the molecular chain entanglement is judged to be moderate, the mechanical strength is 60%-90% of the matrix material, and the risk level is medium risk. If the fractal dimension is greater than the state threshold range, the molecular chain entanglement is poor, the mechanical strength is less than 60% of the matrix material, the risk level is high, and the weld line area should be marked as a potential weak point.

[0027] Furthermore, step S4 is implemented through the following process: Step S401: Optimize the parameter adjustment of the injection molding machine nozzle based on the coordinates of the abnormal area and the risk level. The specific process includes: Based on the coordinates of the abnormal area, the specific location of the abnormal area in the mold cavity is located (such as near the gate, far from the gate, near the thin-walled area, etc.). If the abnormal area is a stagnant flow abnormality and the risk level is high or medium risk, the parameter adjustment includes increasing the melt temperature, adjusting the injection speed, and increasing the holding pressure and time. If the abnormal area is a filling abnormality and the risk level is high or medium risk, parameter adjustments include reducing the injection speed, reducing the holding pressure, and adjusting the stroke of the injection and holding pressure switching point. The optimized injection molding process parameters are downloaded to the injection molding machine controller and will take effect in the next injection molding cycle.

[0028] Step S402: Continuously monitor the offset of the abnormal area relative to the normal front until no pressure drop or temperature fault occurs. The specific process includes: During the subsequent injection molding production process with adjusted injection molding process parameters, steps S1, S2, and S3 are repeated in the next injection molding cycle to re-acquire the current spatial offset of the abnormal area. When the injection molding cycle is completed, the offset change rate relative to the previous injection molding cycle is obtained, where the offset change rate = (current spatial offset - previous spatial offset) / previous spatial offset. When the previous spatial offset is 0, the offset change rate is 1 by default. If the absolute value of the rate of change of the offset gradually decreases, it indicates that the optimization measures are effective and the size of the abnormal area has converged. Then, in the next five consecutive injection cycles, if no new pressure drop or temperature fault is detected near the abnormal area and the fractal dimension is less than the state threshold range, it is determined that the abnormal area control target has been achieved and the injection process has stabilized in a qualified state. Otherwise, the parameters are readjusted until no new pressure drop or temperature fault is detected near the abnormal area and the fractal dimension is less than the state threshold range.

[0029] Please see Figure 2 As shown, a quality control system for the production process of plastic products includes a real-time injection molding monitoring module, an anomaly risk analysis module, and a status control module. The injection molding real-time monitoring module is used to install a thin-film pressure-temperature composite sensor on the surface of the mold cavity, and then collect the arrival time of the melt front, pressure distribution and temperature change of each sensor node in real time during the injection molding process, forming an arrival time difference matrix and a filling temperature matrix, as well as a corresponding three-dimensional melt front advancement surface. Based on the arrival time difference matrix and the filling temperature matrix, it detects whether there is a sudden drop in pressure or temperature discontinuity at local nodes, and then determines whether there is a filling imbalance at the corresponding position and identifies it as stagnation or overfilling abnormality, and outputs the positioning information of the abnormal area and the offset relative to the normal front surface. The abnormal risk analysis module is used to extract the interface contour line on the three-dimensional melt front advancement surface based on the offset of the abnormal area relative to the normal front surface, calculate the corresponding fractal dimension, and use the fractal dimension as a quantitative indicator when the melt merges. When the fractal dimension of the weld line area is higher than the preset threshold, it is marked as a potential weak point. Combined with the location information of the abnormal area, the coordinates and risk level of the abnormal area are output. The state control module is used to optimize the parameter adjustment of the injection molding machine nozzle according to the coordinates of the abnormal area and the risk level, and continuously monitor the offset of the abnormal area relative to the normal front surface until no pressure drop or temperature break occurs.

[0030] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A quality control method for the production process of plastic products, characterized in that, Includes the following steps: Step S1: Install a thin-film pressure-temperature composite sensor on the surface of the mold cavity, and then collect the arrival time of the melt front, pressure distribution and temperature change of each sensor node in real time during the injection molding process to form an arrival time difference matrix and a filling temperature matrix, as well as the corresponding three-dimensional melt front advancement surface. Step S2: Based on the arrival time difference matrix and filling temperature matrix, detect whether there is a sudden drop in pressure or temperature fault at local nodes, and then determine whether there is filling imbalance at the corresponding location and identify it as stagnation or overfilling abnormality, and output the location information of the abnormal area and the offset relative to the normal front. Step S3: Extract the interface contour line on the three-dimensional melt front advancement surface based on the offset of the abnormal area relative to the normal front surface, calculate the corresponding fractal dimension, and use the fractal dimension as a quantitative indicator when the melt merges. When the fractal dimension of the weld line area is higher than the preset threshold, mark it as a potential weak point in strength, and output the coordinates and risk level of the abnormal area in combination with the location information of the abnormal area. Step S4: Optimize the parameters of the injection molding machine nozzle based on the coordinates and risk level of the abnormal area, and continuously monitor the offset of the abnormal area relative to the normal front face until no pressure drop or temperature break occurs.

2. The quality control method for the production process of plastic products according to claim 1, characterized in that, The process of mounting a thin-film pressure-temperature composite sensor on the surface of a mold cavity includes: Obtain the three-dimensional geometric model of the injection mold for the target plastic product, and lay out the sensor on the surface of the mold cavity. With the gate position as the origin, along the expected main direction of melt flow, arrange the sensor nodes in an equidistant or non-equidistant grid form from the area near the gate to the farthest end of the cavity. A thin-film pressure-temperature composite sensor was selected as the sensing unit for the sensor node, and each sensor node was assigned a number. 1，1 c 1，2 c i，j , where i represents the row of the sensor node, j represents the column of the sensor node, and i and j are natural numbers greater than 0.

3. The quality control method for the production process of plastic products according to claim 2, characterized in that, The process of acquiring real-time data on the arrival time of the melt front, pressure distribution, and temperature changes at each sensor node during injection molding, forming an arrival time difference matrix and a filling temperature matrix, as well as the corresponding three-dimensional melt front advancement surface, includes: The injection molding machine is started to begin a complete injection cycle. After the molten plastic is injected into the mold cavity from the gate, the melt front begins to move forward. Each sensor node continuously records the pressure signal and temperature signal at a preset sampling frequency. Both the pressure signal and the temperature signal are curves that change with time. For both pressure and temperature signals, when the sensor node determines that both signals have arrived simultaneously, the arrival time t of the melt front at the sensor node is confirmed. i，j ; From arrival time t i，j The pressure and temperature change curves during the period from the start of the filling stage to the end of the filling stage are obtained, and the stable filling pressure and average filling temperature of each sensor node during the filling stage are extracted. The arrival time t of the melt front based on all sensor nodes i，j Construct an arrival time matrix and a filling temperature matrix based on the average filling temperature of each sensor node; By integrating a pressure-temperature sensor array on the surface of the mold cavity, the moment when the melt arrives, the instantaneous pressure and temperature are recorded by each sensor node. The sensed data is then spatiotemporally interpolated and surface reconstructed on the three-dimensional geometric model of the mold to generate a dynamic surface that advances forward over time, which is the three-dimensional melt front advancing surface.

4. The quality control method for the production process of plastic products according to claim 3, characterized in that, The process of detecting whether a local node experiences a sudden drop in pressure or a temperature fault based on the arrival time difference matrix and the filling temperature matrix includes: Obtain the arrival time difference matrix for the current injection cycle, and obtain the global mean and standard deviation of all elements in the arrival time difference matrix. For each sensor node, compare its arrival time difference with the arrival time difference of its neighboring nodes. Define the neighboring nodes as the set of sensor nodes that have a four-neighbor or eight-neighbor relationship with the sensor node on the surface of the mold cavity. Obtain the local arrival time difference index of the sensor node relative to its neighboring nodes, and determine whether a local fill time anomaly occurs at the corresponding sensor node based on the relationship between the local arrival time difference index and a preset threshold. Obtain the temperature difference matrix for the current injection molding cycle. Similar to detecting a sudden drop in pressure, obtain the global average and standard deviation of all elements in the temperature difference matrix. For each sensor node, obtain the local outlier degree of its temperature difference and classify temperature fault anomalies based on the local outlier degree.

5. The quality control method for the production process of plastic products according to claim 4, characterized in that, The process of determining whether a filling imbalance has occurred at the corresponding location and identifying it as a stagnant or overfilling anomaly, and outputting the location information of the abnormal area and its offset relative to the normal front surface, includes: For each sensor node, an abnormal state vector is defined. Based on the abnormal state vector, it is determined whether a stagnation abnormality has occurred at each sensor node. The spatial locations of all sensor nodes that are determined to have stagnation abnormality or overcharge abnormality are clustered. Adjacent abnormal sensor nodes in the space are grouped into the same abnormal region; otherwise, they are not grouped. The abnormal region is characterized by a set of sensor node coordinates, and the geometric center of the abnormal region is used as the abnormal location information. The offset of the abnormal region relative to the normal front is calculated. The normal front is defined as: at the current filling time, the normal melt front propulsion surface is generated by interpolation based on the arrival time of the global sensor nodes. The normal melt front propulsion surface assumes that the melt is filled in an ideal radial flow mode and does not consider local defects. For each sensor node within the abnormal region, its actual arrival time is obtained. The difference between the theoretical arrival time of the corresponding sensor node when the melt front propagates in the ideal mode is used as the time offset. The time offset is converted into a spatial offset. The sign of the spatial offset indicates whether the abnormal region is lagging or leading the normal front. Then, the spatial offsets of each sensor node in the abnormal region are accumulated to obtain the total offset.

6. The quality control method for the production process of plastic products according to claim 5, characterized in that, The process of extracting the interface contour line on the 3D melt front advancement surface based on the offset of the abnormal region relative to the normal front surface and calculating the corresponding fractal dimension includes: Based on the abnormal location information and overall offset of the abnormal area, it is divided into attention areas. Within the attention areas, the final state of the three-dimensional melt front advancement surface at the end of the current injection cycle is obtained. In the final state, the two melts have completely merged, and their boundary is the physical location of the weld line. On the three-dimensional melt front advancing surface within the attention area, when two melts merge, the surface tension and cooling contraction are inconsistent, and the surface presents a valley line with a local minimum or curvature abrupt change. The second derivative of the surface is obtained, and the eigenvalues ​​of the second derivatives that are greater than a preset threshold are selected. Then, the gradient is diffused outward from the center of the abnormal region to track the point with the largest gradient change and connect them to form a spatial curve. The spatial curve is the three-dimensional spatial representation of the weld line on the three-dimensional melt front advancing surface. Projecting it onto a two-dimensional plane perpendicular to the mold opening direction yields a two-dimensional curve, which is defined as the interface contour line. At the same time, the fractal dimension of the interface contour line is obtained.

7. A quality control method for the production process of plastic products according to claim 6, characterized in that, Using fractal dimension as a quantitative indicator of melt fusion, when the fractal dimension of the weld line region is determined to be higher than a preset threshold, it is marked as a potential weak point. The process of combining the location information of the abnormal region to output the coordinates and risk level of the abnormal region includes: Based on the negative correlation between fractal dimension and the bonding strength of the weld line region, fractal dimension is used as a quantitative indicator for melt fusion. A state threshold range is set, and the coordinates and risk level of the abnormal area are output according to the relationship between the fractal dimension and the state threshold range. The coordinates of the abnormal area are defined by the abnormal location information of the abnormal area.

8. The quality control method for the production process of plastic products according to claim 7, characterized in that, The process of optimizing the parameters of the injection molding machine nozzle based on the coordinates and risk level of the abnormal area includes: Based on the coordinates of the abnormal area, the specific location of the abnormal area in the mold cavity can be determined. If the abnormal area is a stagnant flow abnormality and the risk level is high or medium, the parameter adjustment includes increasing the melt temperature, adjusting the injection speed, and increasing the holding pressure and time. If the abnormal area is a filling abnormality and the risk level is high or medium risk, parameter adjustments include reducing the injection speed, reducing the holding pressure, and adjusting the stroke of the injection and holding pressure switching point. The optimized injection molding process parameters are downloaded to the injection molding machine controller and will take effect in the next injection molding cycle.

9. A quality control method for the production process of plastic products according to claim 8, characterized in that, The process of continuously monitoring the offset of the abnormal region relative to the normal front until no pressure drop or temperature fault occurs includes: During subsequent injection molding production with adjusted injection molding process parameters, the current spatial offset of the abnormal area is reacquired. When the injection cycle is completed, the rate of change of the offset relative to the previous injection cycle is obtained. If the absolute value of the rate of change of the offset gradually decreases and the size of the abnormal region converges, and then in the subsequent five consecutive injection cycles, no new pressure drop or temperature fault is detected near the abnormal region and the fractal dimension is less than the state threshold range, then it is determined that the abnormal region control target has been achieved and the injection process has stabilized in a qualified state. Otherwise, the parameters are readjusted until no new pressure drop or temperature fault is detected near the abnormal region and the fractal dimension is less than the state threshold range.

10. A quality control system for a plastic product manufacturing process, used to implement the quality control method for a plastic product manufacturing process as described in any one of claims 1-9, characterized in that, It includes a real-time injection molding monitoring module, an anomaly risk analysis module, and a status control module; The injection molding real-time monitoring module is used to install a thin-film pressure-temperature composite sensor on the surface of the mold cavity, and then collect the arrival time of the melt front, pressure distribution and temperature change of each sensor node in real time during the injection molding process, forming an arrival time difference matrix and a filling temperature matrix, as well as a corresponding three-dimensional melt front advancement surface. Based on the arrival time difference matrix and the filling temperature matrix, it detects whether there is a sudden drop in pressure or temperature discontinuity at local nodes, and then determines whether there is a filling imbalance at the corresponding position and identifies it as stagnation or overfilling abnormality, and outputs the positioning information of the abnormal area and the offset relative to the normal front surface. The abnormal risk analysis module is used to extract the interface contour line on the three-dimensional melt front advancement surface based on the offset of the abnormal area relative to the normal front surface, calculate the corresponding fractal dimension, and use the fractal dimension as a quantitative indicator when the melt merges. When the fractal dimension of the weld line area is higher than the preset threshold, it is marked as a potential weak point. Combined with the location information of the abnormal area, the coordinates and risk level of the abnormal area are output. The state control module is used to optimize the parameter adjustment of the injection molding machine nozzle according to the coordinates of the abnormal area and the risk level, and continuously monitor the offset of the abnormal area relative to the normal front surface until no pressure drop or temperature break occurs.

Images