Injection mold punching precision self-adaptive control method and system

Abstract

This invention discloses an adaptive control method and system for blanking accuracy of injection molds, relating to the field of injection mold processing technology. The method includes the following steps: acquiring initial blanking parameters, blanking process parameters, and blanking result parameters of the injection mold to be blanked; establishing the correspondence between the blanking edge state and blanking accuracy of the injection mold; determining the characteristic changes in blanking accuracy under the current blanking gap; determining the accuracy deviation feature set; generating candidate control periods and key source factors through abnormal moment determination and abnormal source determination; identifying the current blanking accuracy; determining the control classification range associated with the key source factors; and determining the adjustment direction and adjustment range of the control strategy. This invention adjusts based on the blanking edge state and blanking gap, improving the accuracy and specificity of blanking accuracy determination.

Description

Technical Field

[0001] This invention relates to the field of injection mold processing technology, specifically to an adaptive control method and system for injection mold punching accuracy. Background Technology

[0002] Blanking is a stamping process that uses injection molds on a press to separate sheet metal along closed or open contour lines. Currently, this technology is rapidly evolving towards adapting to multiple materials and integrating digital simulation, strongly supporting the development needs of high-end manufacturing fields such as automobiles, electronics, aerospace, and instrumentation. Blanking accuracy directly affects the dimensional consistency, edge quality, and subsequent assembly performance of injection molds.

[0003] However, existing injection mold blanking control technology still has the following shortcomings: On the one hand, traditional blanking control methods typically employ uniform blanking parameters for the entire blanking station, such as fixed blanking gaps, blanking speeds, and fixed edge pressing methods, and complete the blanking of injection molded parts according to preset process parameters. In actual scenarios, such as in the processing of complex injection molds with thin-walled structures, reinforcing rib structures, curved surfaces, and different wall thickness combinations, the injection mold typically has different local thicknesses, curvatures, structural stiffness, and material states in different blanking station areas. This results in significant differences in the stress and deformation characteristics of different areas during the blanking process. Using uniform initial blanking parameters makes it difficult to balance the processing accuracy of each blanking station area, which can easily lead to dimensional deviations in some areas and excessive constraints in others. On the other hand, when there is a precision deviation, existing blanking control technology is usually set once based on predetermined process parameters. There is a lack of continuous feedback between the blanking result after control and the historical control strategy. When the blanking precision is improved or further deteriorated after the control strategy is adjusted, the existing system usually does not update the correlation between the corresponding blanking state, blanking gap and blanking precision in a timely manner. This makes it difficult for historical control experience to continuously adapt to changes in material condition, production batch and mold usage condition. Especially during the continuous operation of injection molds, mold wear may gradually manifest as changes in blanking force response, changes in blanking edge quality and continuous increase in dimensional deviation. If the control strategy cannot be dynamically updated by combining the blanking process characteristics of multiple historical batches and the actual blanking precision, the blanking precision judgment will be inaccurate and the efficiency will be reduced. Summary of the Invention

[0004] (a) Technical problems to be solved To address the shortcomings of existing technologies, this invention provides an adaptive control method and system for injection mold blanking accuracy. The method acquires the initial blanking parameters, blanking process parameters, and blanking result parameters of the injection mold to be blanked; establishes the correspondence between the mold's edge-pressing state and blanking accuracy; determines the characteristic changes in blanking accuracy under the current blanking gap; and identifies the accuracy deviation feature set. Through anomaly moment and anomaly source determination, candidate control periods and key source factors are generated. The current blanking accuracy is identified, the control classification range associated with key source factors is determined, and the adjustment direction and magnitude of the control strategy are determined, thus solving the problems raised in the background technology.

[0005] (II) Technical Solution To achieve the above objectives, the present invention provides the following technical solution: In a first aspect, this application provides an adaptive control method for the blanking accuracy of injection molds, the method comprising: Obtain the initial blanking parameters, blanking process parameters, and blanking result parameters of the injection mold to be blanked; Based on the initial blanking parameters, blanking process parameters, and blanking result parameters, establish the correspondence between the blanking state of the injection mold and the blanking accuracy, determine the characteristic changes of the blanking accuracy under the current blanking gap, and determine the set of accuracy deviation characteristics. Anomaly detection is performed based on the set of accuracy deviation features, including the detection of abnormal moments and sources of anomalies in blanking accuracy. By sorting and grouping abnormal moments and configuring dimensional constraints in conjunction with the edge pressing state, adjustment processing is performed on abnormal moments that do not meet the dimensional constraints to generate candidate control periods. For candidate control periods, the key source factors are determined by detecting the abnormal sources of the set of accuracy deviation features. Identify the current blanking accuracy, determine the control classification range associated with key source factors, and determine the adjustment direction and magnitude of the control strategy.

[0006] Furthermore, the steps for obtaining the initial blanking parameters, blanking process parameters, and blanking result parameters include: Before the punching process, obtain the thickness, temperature, position of the injection mold, and the material of each batch of injection mold; The degree of thickness variation in the blanking station area is determined based on the thickness. Determine the degree of temperature change in the punching station area based on the temperature; Determine the initial state of the injection mold relative to the punching tool based on its position; The thickness variation, temperature variation, initial state, and each batch of injection mold material are correlated according to the blanking station area to form the initial blanking parameters; During the blanking process, the punch displacement, punch speed, blanking pressure, and blanking force are obtained; Based on the punch displacement and punching force, a first response curve is established, and the first slope characteristic is identified; Based on the blanking pressure and punching speed, a second response curve is established, and a second slope characteristic is identified; The first slope feature and the second slope feature are correlated in chronological order to form the punching process parameters; After the blanking process is completed, obtain the blanking edge dimensions and deformation data; Determine the dimensional deviation based on the blanking dimensions; Determine the deformation deviation based on the deformation data; The blanking result parameters are generated by associating dimensional deviation and deformation deviation according to different edge clamping states; The pressing state includes no pressing, single-sided pressing, and double-sided pressing.

[0007] Furthermore, establish the correspondence between the blanking state of the injection mold and the blanking accuracy, including: Divide the work area into multiple blanking stations, determine the initial blanking parameters for each blanking station area, and obtain status data for several blanking station areas. Associate the status data of each punching station area with the punching process parameters and punching result parameters corresponding to different pressing states of the same punching station area respectively; Identify the changes in initial blanking parameters, blanking process parameters, and blanking result parameters of the same blanking station area in different batches; wherein the changes include at least one representation of the magnitude and direction of change; The influence relationship on the status data of the punching station area is determined based on the changes in initial punching parameters, changes in punching process parameters, and changes in punching result parameters. Mark the blanking station areas that have the same influence relationship but correspond to different blanking accuracies.

[0008] Furthermore, the determination of abnormal moments includes: The first-order difference of the precision deviation feature set is used to obtain the difference value sequence; Multiple initial time periods are obtained by segmenting the differential value sequence based on the positive and negative change times. At any initial time period, based on the known punching gap and feature changes, a multi-dimensional feature vector is constructed, which includes at least: feature parameters, punching gap and timestamp; Among them, the blanking clearance includes small clearance, medium clearance and large clearance, and the characteristic parameters include the change of initial blanking parameters, the change of blanking process parameters and the change of blanking result parameters; Extract the peak value and variance of the characteristic parameters of any initial time period, and obtain the punching gap of the initial time period. Establish several control vectors according to different combinations, and generate control deviation index based on the control vectors. For the control vectors at the same time, select at least one corresponding time when the control deviation index is greater than the standard abnormal threshold, and mark it as an abnormal time.

[0009] Furthermore, candidate control periods are generated, including: Based on the control deviation index corresponding to one quantity time before and one quantity time after the abnormal time, determine whether the abnormal time meets the continuous abnormality condition. If the consecutive anomaly condition is met, select one time point before and after the anomaly to enter the sorting and grouping process; if the consecutive anomaly condition is not met, only select the anomaly time point to enter the sorting and grouping process.

[0010] Furthermore, the determination of the source of the anomaly includes: Based on the precision deviation feature set, analyze the linear correlation between feature changes and source factors, and configure feature processors; if the correlation is linear, analyze the source factors whose deviations in the precision deviation feature set are synchronous or have consistent trends, and mark them as key source factors; if the correlation is nonlinear, perform feature processing, determine the feature contribution degree of the source factors, and determine the key source factors by sorting them in descending order based on the feature contribution degree. Among them, the key source factors include at least one of the following: changes in material condition, changes in blanking clearance, changes in positioning condition, and changes in the wear condition of injection mold.

[0011] Furthermore, the scope of control classifications associated with key source factors is determined, including: Extract the similarity between the control vector and the feature contribution level, filter the precision deviation feature set corresponding to the similarity greater than or equal to the standard similarity threshold, determine whether the feature parameters analyzed in real time fall into the control classification range corresponding to the fluctuation of normal operating conditions, and perform circular clustering analysis to determine the circular judgment radius; Identify at least one normal type cluster, and use the cluster center of each normal type cluster as the normal control center. Perform a circle drawing operation with a circular decision radius. Mark those falling inside the circle as normal deformation and match the control strategy; mark those falling outside the circle as abnormal cutting.

[0012] Furthermore, determine the direction and magnitude of the control strategy adjustment, including: Obtain the punching accuracy before and after the control strategy change in multiple historical batches; Based on the correspondence between control strategy and blanking accuracy, determine the degree of influence of control strategy on the current blanking accuracy deviation; Based on the degree of influence, the control strategy with the highest degree of influence on the current punching accuracy deviation is selected from multiple candidate control parameters as the priority adjustment strategy. Based on the current deviation direction of the punching accuracy, determine the adjustment direction of the priority adjustment strategy; Based on the blanking accuracy corresponding to the same control strategy in historical blanking processes, determine the adjustment range of the priority adjustment strategy.

[0013] Furthermore, after determining the direction and magnitude of the control strategy adjustment, this includes: The injection mold is subjected to blanking control according to the adjustment direction and adjustment range, and the blanking result data after control is obtained; Based on the deviation between the controlled blanking result data and the standard blanking accuracy threshold, update the blanking accuracy corresponding to the current edge pressing state.

[0014] Secondly, this application provides an adaptive control system for the blanking accuracy of injection molds, the system comprising: The acquisition module acquires the initial blanking parameters, blanking process parameters, and blanking result parameters of the injection mold to be blanked; The feature determination module establishes the correspondence between the blanking die pressing state and the blanking accuracy based on the initial blanking parameters, blanking process parameters, and blanking result parameters. It determines the feature changes corresponding to the blanking accuracy under the current blanking gap and determines the set of accuracy deviation features. The anomaly detection module performs anomaly detection based on the set of precision deviation features, including the detection of abnormal moments and sources of anomalies in blanking precision. By sorting and grouping abnormal moments and configuring dimensional constraints based on the edge pressing state, adjustment processing is performed on abnormal moments that do not meet the dimensional constraints to generate candidate control periods. For candidate control periods, the key source factors are determined by using the anomaly source detection method on the set of precision deviation features. The adaptive control module identifies the current blanking accuracy, determines the control classification range associated with key source factors, and determines the adjustment direction and magnitude of the control strategy.

[0015] (III) Beneficial Effects This invention provides an adaptive control method and system for the blanking accuracy of injection molds, which has the following beneficial effects: By acquiring the initial blanking parameters, blanking process parameters, and blanking result parameters, and establishing the correspondence between the blanking process and blanking accuracy according to the blanking station area and the edge pressing state, and determining the characteristic changes corresponding to the accuracy based on the current blanking gap, a set of accuracy deviation features is constructed, so that the initial blanking state, blanking process, and final blanking result are associated, providing a regionally specific state benchmark for subsequent anomaly identification and control parameter adjustment, thereby improving the accuracy and specificity of blanking accuracy judgment; Anomaly detection is performed on the precision deviation feature set. The determination of the anomaly moment in blanking precision is combined with the determination of the anomaly source. By sorting and grouping the anomaly moments and configuring dimensional constraints in conjunction with the current blanking state, the anomaly moments that do not meet the dimensional constraints are adjusted to form candidate control periods. The changing patterns of precision deviation features within the candidate control periods are used to identify key source factors, which improves the accuracy and stability of blanking anomaly location. A correspondence between key source factors and control classification ranges is established, and the corresponding control classification range is matched according to the current blanking precision. Priority adjustment strategies are selected from multiple candidate control parameters, which improves the targeting of adaptive control and reduces control coupling and over-adjustment caused by simultaneous adjustment of multiple parameters. Based on the current blanking accuracy as feedback, the blanking accuracy corresponding to the current blanking state is updated according to the deviation between the controlled blanking result and the standard blanking accuracy range. The updated blanking process data and blanking result data are used for subsequent control strategy selection, so that the blanking state of no blanking, single-sided blanking, and double-sided blanking can be continuously corrected according to the actual blanking results. To a certain extent, this improves the blanking accuracy stability of complex injection molded parts under different material states, blanking gaps, and partitioned area structures. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the present invention. Detailed Implementation

[0017] 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, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0018] Research and Development Concept: The initial blanking parameters, blanking process parameters, and blanking result parameters of the injection mold to be blanked are established. Corresponding characteristic changes are established for blanking without a blank holder, with one side blank holder, and with both sides blank holder. By extracting the peak value and variance of characteristic changes within different initial time periods, a control vector associated with the blanking clearance is constructed to identify moments of abnormal accuracy. For abnormal states, key source factors such as material state, blanking clearance, positioning state, and mold wear are further analyzed. Based on the actual accuracy improvement effect of historical control strategies, priority adjustment parameters, adjustment direction, and adjustment magnitude are determined. The correspondence between blank holder state and blanking accuracy is updated according to the blanking results after control, achieving intelligent closed-loop adaptive control.

[0019] Example 1: This embodiment of the invention provides an adaptive control method for the blanking accuracy of injection molds; Figure 1This is a flowchart illustrating the present invention; please refer to [link / reference]. Figure 1 The method includes the following steps: S1: Obtain the initial blanking parameters, blanking process parameters, and blanking result parameters of the injection mold to be blanked; Initial parameters for blanking: Before the injection mold to be blanked enters the blanking station area, the thickness, temperature, position of the injection mold and the material of each batch of injection mold are obtained; among them, the thickness is used to characterize the material geometry of the blanking station area, the temperature is used to characterize the thermal state of the blanking station area, the position is used to characterize the positioning state of the injection mold to be blanked relative to the blanking tool, and the material of each batch of injection mold is used to characterize the material source and batch differences. Since different batches, different thermal states, and different wall thicknesses may have different punching responses, determining the above data before punching begins can avoid misjudging changes in punching force and dimensional deviations caused by differences in initial parameters as changes in punching clearance or abnormal edge clamping conditions. Specifically, for each batch of injection mold material, the raw material batch number, injection production batch number, or batch identifier associated with the production management system can be used as the data source. For position, the relative position between the positioning reference point, reference edge, or reference hole on the outer contour of the injection mold to be punched and the center line of the punching tool can be used to characterize the position. The thickness variation of the blanking station area is determined based on the thickness. The steps are as follows: the thickness deviation is equal to the absolute value of the difference between the actual thickness and the standard thickness, divided by the standard thickness. Here, the actual thickness represents the thickness detected in the current blanking station area, and the standard thickness represents the preset reference thickness corresponding to this area. It is set based on historical statistics. By obtaining historical thickness data, its mean and standard deviation are calculated, and the preset reference thickness is set to the mean plus a multiple of the standard deviation. It should be noted that the value of the multiple is only an example. The specific value should be set according to the actual situation, which will not be elaborated here. The temperature variation in the blanking station area is determined by the following steps: the temperature variation is equal to the absolute value of the difference between the actual temperature and the standard temperature, divided by the standard temperature. Here, the actual temperature represents the temperature detected in the current blanking station area, and the standard temperature represents the preset reference temperature corresponding to that area. This preset reference temperature is set based on historical statistics. By collecting historical temperature data, calculating its mean and standard deviation, the preset reference temperature is set to the mean plus a multiple of the standard deviation. It should be noted that the multiple is only an example; the specific value should be set according to the actual situation, and will not be elaborated further here. The initial state of the injection mold relative to the blanking tool is determined based on its position; for injection molds with planar contours, two or more reference points can be obtained to simultaneously determine translational and angular deviations; for injection molds with complex curved surfaces, the overall positional state can be obtained through three-dimensional contour fitting. Based on the degree of thickness change, temperature change, initial change, and the material batch of each injection mold, the blanking initial parameters are formed by associating them with the blanking station area. This regional association can avoid data overlap between different areas. The thickness change, temperature change, initial position state, and material batch information corresponding to the same blanking area are bound together, so that each blanking area has an independent initial state description. The blanking initial parameters are used to characterize the initial parameters of the injection mold to be blanked before the blanking action is performed.

[0020] Punching process parameters: During the punching process, the punch displacement, punch speed, blanking pressure, and punching force are acquired. Punch displacement characterizes the degree to which the punch enters the injection mold material; it can be obtained through feedback values ​​from a linear displacement sensor, encoder, or servo drive. Punch speed characterizes the motion state during the punching process and is directly obtained by the servo controller. Blanking pressure characterizes the constraint state of the blanking mechanism on the injection mold; it can be obtained by a pressure sensor installed in the blanking mechanism. Punching force characterizes the mechanical response of the injection mold material to the punching action; it can be obtained from a force sensor or the force feedback signal corresponding to the drive mechanism. It should be noted that the sensors or related mechanisms mentioned above are not shown in the figures; they are adapted for installation and will not be elaborated upon here. Based on the punch displacement and punching force, with the punch displacement as the horizontal axis and the punching force as the vertical axis, a first response curve is established to show the change of punching force with punch displacement during the punching process. The first slope feature of the first response curve is identified. The first slope feature is used to characterize the degree of change of punching force corresponding to a unit punch displacement. Therefore, compared with using the peak value of punching force alone, the first slope feature can more accurately describe the change of material state from contact, deformation, fracture to completion of punching. Based on the blanking pressure and the blanking speed, a second response curve is established with the blanking speed as the horizontal axis and the blanking pressure as the vertical axis, to reflect the change in blanking pressure with the blanking speed during the blanking process. The second slope feature of the second response curve is then identified. The second slope feature is used to characterize the correspondence between the change in blanking pressure and the change in blanking speed. The first slope feature and the second slope feature are correlated in time sequence to form the blanking process parameters.

[0021] Blanking result parameters: After the blanking process is completed, the blanking edge dimensions and deformation data after blanking are obtained. Among them, the blanking edge dimensions reflect whether the blanking contour has reached the design dimensions. The blanking edge dimensions can be obtained by visual measurement, laser displacement measurement, contact dimension measurement, or three-dimensional contour measurement. The deformation data reflects whether the injection molded part has warped beyond the allowable range during the blanking process. The deformation data can be obtained by comparing the contours before and after blanking, three-dimensional scanning, or datum point displacement. For mature dimension inspection technologies, existing industrial inspection equipment can be selected according to the product accuracy requirements, but is not limited to a single sensor type. It should be noted that the sensors or related mechanisms mentioned above are not shown in the figure. They are adapted for installation and will not be described in detail here. The dimensional deviation is determined based on the blanking dimensions. The steps for obtaining this deviation include: the dimensional deviation is represented as the difference between the actual blanking dimension and the standard blanking dimension; the standard blanking dimension is set based on historical statistics. By collecting historical actual blanking dimension data, its mean and standard deviation are calculated, and the standard blanking dimension is set as the mean plus a multiple of 2 to 3 times the standard deviation; it should be noted that the value of the multiple is only an example, and the specific value should be set according to the actual situation, which will not be elaborated here. The deformation deviation is determined based on the deformation data. The steps for obtaining the deformation deviation include: the deformation represents the maximum distance between the blanked contour and the reference contour; the reference contour is a predefined setting of the system; the blanking result parameters are formed by associating the dimensional deviation and the deformation deviation according to different blanking states, which are used to reflect the actual blanking accuracy of the current injection mold. The edge-pressing states include no edge-pressing, single-sided edge-pressing, and double-sided edge-pressing. Specifically, the edge-pressing state can be identified based on whether effective pressing constraints are generated on both sides: no edge-pressing when neither side meets the effective edge-pressing condition, single-sided edge-pressing when only one side meets the effective edge-pressing condition, and double-sided edge-pressing when both sides meet the effective edge-pressing condition. The threshold for the effective edge-pressing condition can be determined through no-load testing, normal punching tests, and the product's allowable deformation range, and calibrated according to the mold structure and product material. If the threshold is too low, mechanical contact or noise may be misjudged as effective edge-pressing, while if the threshold is too high, the actual edge-pressing state may be missed. Therefore, the goal should be to be able to stably distinguish between the three edge-pressing states.

[0022] S2: Based on the initial blanking parameters, blanking process parameters, and blanking result parameters, establish the correspondence between the blanking state of the injection mold and the blanking accuracy, determine the characteristic changes of the blanking accuracy under the current blanking gap, and determine the set of accuracy deviation characteristics. Establishing the correspondence between the blanking state of the injection mold and the blanking accuracy includes: dividing multiple blanking station areas, determining the initial blanking parameters of each blanking station area, and obtaining several blanking station area state data; for dividing the areas, alternative technical means can be used; for example, fixed mesh division based on the blanking contour can be used, adaptive division based on the location of local thickness abrupt changes can be used, or the area can be divided according to structural boundaries such as reinforcing ribs, corners, and holes; for storing the state data of the divided areas, a structured data table corresponding to the area number and parameter fields can be used, or a timestamp plus area index can be used for recording; The status data of each blanking station area is associated with the blanking process parameters and blanking result parameters corresponding to different edge pressing states of the same blanking station area. That is, for the same blanking station area, when blanking is performed with no edge pressing, single-sided edge pressing, and double-sided edge pressing respectively, the corresponding blanking process parameters and blanking result parameters are obtained respectively. These data are associated with the status data of the blanking station area of ​​that area, so that for any blanking operation, a complete record similar to: area, edge pressing state, blanking process parameters, and blanking result parameters can be formed. For example, in the same first blanking station area, the first set of blanking process parameters and the first set of blanking result parameters are obtained under the no-edge-holding state; the second set of blanking process parameters and the second set of blanking result parameters are obtained under the single-sided edge-holding state; and the third set of blanking process parameters and the third set of blanking result parameters are obtained under the double-sided edge-holding state. If the state data of the first blanking station area is always kept within the range of thin wall, stable temperature, and material batch A, then these three sets of data can be directly used to compare the impact of different edge-holding states on the accuracy of this area. This method identifies variations in initial blanking parameters, blanking process parameters, and blanking result parameters for the same blanking station area across different batches. The variations include at least one representation of the magnitude and direction of change. Specifically, variations in initial blanking parameters may include: changes in thickness, temperature, initial positioning, and material batch variation. Variations in blanking process parameters may include: changes in peak blanking force, changes in blanking force variance, changes in the slope of the first response curve, changes in the slope of the second response curve, changes in the duration of the blanking stage, and changes in the edge-pressing state response. Variations in blanking result parameters may include: changes in dimensional deviation and changes in deformation deviation. Based on the changes in initial blanking parameters, blanking process parameters, and blanking result parameters, the influence relationship on the status data of the blanking station area is determined; for each blanking station area, the correlation between different initial blanking parameter changes and blanking process and blanking result changes is established; to facilitate calculation, the parameters involved are normalized and dimensionless to make the subsequent analysis physically meaningful. To determine the influencing relationship, mature technical means such as historical data statistics, correlation analysis, and sensitivity analysis can be used; the blanking station areas with the same influencing relationship but corresponding to different blanking accuracies are marked to form a correspondence between the blanking state of the injection mold and the blanking accuracy; Based on the status data of the current blanking station area and the current edge pressing status, the normal blanking reference features corresponding to the current blanking station area are determined from the correspondence between the edge pressing status of the injection mold and the blanking accuracy. Obtain the punching process parameters corresponding to each punching moment in the current punching process, and extract the punching process features corresponding to each punching moment according to the preset time window; compare the punching process features corresponding to each punching moment with the normal punching reference features, and determine the punching force deviation features, punch displacement deviation features, first slope deviation features and second slope deviation features respectively. Obtain the blanking result parameters corresponding to the current blanking process, and determine the dimensional deviation characteristics and deformation deviation characteristics; according to the blanking station area, the edge pressing state and the blanking time, associate the blanking force deviation characteristics, punch displacement deviation characteristics, the first slope deviation characteristics, the second slope deviation characteristics, the dimensional deviation characteristics and the deformation deviation characteristics. Based on the deviation amplitude, deviation direction, and continuous change trend of each deviation characteristic, determine the set of precision deviation characteristics corresponding to the current blanking station area; among them, the continuous change trend can be described by the difference between adjacent batches or the average change rate of the sliding window; in order to avoid the abnormal establishment of the influence relationship of a single batch, the statistical median, quantile, or robust average of multiple batches can be used as a reference.

[0023] By acquiring initial blanking parameters, blanking process parameters, and blanking result parameters, and establishing a correspondence between the blanking process and blanking accuracy according to the blanking station area and the edge pressing state, the characteristic changes corresponding to the accuracy are determined based on the current blanking gap, and a set of accuracy deviation features is constructed. This establishes a correlation between the initial blanking state, the blanking process, and the final blanking result, providing a regionally specific state benchmark for subsequent anomaly identification and control parameter adjustment, thereby improving the accuracy and specificity of blanking accuracy determination.

[0024] S3: Perform anomaly determination based on the precision deviation feature set, including the determination of the abnormal moment and the abnormal source of the blanking precision; sort and group the abnormal moments, combine the edge pressing state to configure the dimensional constraints, perform adjustment processing on the abnormal moments that do not meet the dimensional constraints, and generate candidate control periods; for the candidate control periods, determine the key source factors by judging the abnormal source of the precision deviation feature set. The abnormal moment determination includes: first-order difference of the precision deviation feature set to obtain a difference value sequence; segmentation of the positive and negative change moments in the difference value sequence to obtain multiple initial time periods; under any initial time period, based on the known punching gap and feature changes, constructing a multi-dimensional feature vector, which includes at least: feature parameters, punching gap, and timestamp; wherein, the punching gap includes small gap, medium gap, and large gap, and the small gap and large gap are usually determined based on the punching speed. Under high-speed punching, such as greater than 600 times / min, a small punching gap is used; under low-speed punching, such as less than 200 times / min, a larger punching gap is used; if greater than or equal to 200 times / min and less than or equal to 600 times / min, a conventional punching gap is used; the feature parameters include changes in initial punching parameters, changes in punching process parameters, and changes in punching result parameters; The peak value and variance of the feature parameters for any initial time period are extracted, and the punching gap of the initial time period is obtained. The peak value is summed and the variance is subtracted respectively. The peak value is associated with the corresponding punching gap according to different combinations. Several control vectors are established according to different combinations. The form of the vector is [peak value ± variance, punching gap]. It should be noted that the peak value plus variance is used to represent the reference state of the feature parameter after it fluctuates in the direction of high response in the current initial time period, and the peak value minus variance is used to represent the reference state of the feature parameter after it fluctuates in the direction of low response in the current initial time period. By associating the peak value plus variance and the peak value minus variance with the corresponding punching gap respectively, the feature parameters under the same punching gap form a control reference range with upper and lower boundaries, so as to obtain the high-side control boundary and low-side control boundary corresponding to the initial time period. The specific steps for generating a control deviation index based on a control vector include: obtaining the characteristic parameters and the current punching gap at the current punching moment; determining the control vector corresponding to the current punching condition from multiple control vectors based on the current punching gap; and comparing the characteristic parameters at the current punching moment with the high-side control boundary and the low-side control boundary in the control vector to determine whether the characteristic parameters at the current punching moment are within the corresponding control range. If the characteristic parameter at the current punching moment is between the low-side control boundary and the high-side control boundary, that is, the low-side control boundary < characteristic parameter < high-side control boundary, the control deviation index at the current punching moment is set to zero, indicating that the current characteristic parameter is within the control reference range under the corresponding punching gap; If the characteristic parameter at the current punching moment is higher than the high-side control boundary (i.e., characteristic parameter ≥ high-side control boundary), the excess amount of the current characteristic parameter exceeding the high-side control boundary is determined, and the ratio between the excess amount and the variance of the corresponding initial time period is used as the control deviation index at the current punching moment; wherein, the excess amount is used to characterize the numerical difference of the current characteristic parameter being higher than the high-side control boundary, and the variance is used to scale the excess amount to characterize the degree to which the current characteristic parameter exceeds the normal fluctuation range; If the characteristic parameter at the current punching moment is lower than the lower-side control boundary (i.e., characteristic parameter ≤ lower-side control boundary), the difference between the lower-side control boundary and the current characteristic parameter is determined, and the ratio of the difference to the variance of the corresponding initial time period is used as the control deviation index at the current punching moment; wherein, the difference represents the degree to which the current characteristic parameter is lower than the lower-side control boundary, and the variance is used to normalize the difference. For the control vector at the same moment, at least one moment with a control deviation index greater than the standard anomaly threshold is selected and marked as an anomaly moment. The steps for obtaining the standard anomaly threshold include: taking multiple normal blanking periods under the same blanking station area, the same material state, the same edge-pressing state, and the same blanking gap, and obtaining the control deviation index for each normal blanking period in the manner described above; sorting the obtained multiple normal control deviation indices according to their numerical values, and determining the standard anomaly threshold based on the 95th quantile of the normal control deviation index, so that the standard anomaly threshold can reflect the maximum allowable deviation under normal blanking conditions; it should be noted that the quantile value is only an example, and the specific value should be set according to the actual situation, which will not be elaborated here. The process of generating candidate control periods includes: determining whether an abnormal time meets the continuous abnormality condition by taking the control deviation index corresponding to one number of time points before and after the abnormal time; the criteria for meeting the continuous abnormality condition are: the control deviation index corresponding to one number of time points before and after the abnormal time point is greater than the standard abnormality threshold; if the continuous abnormality condition is met, one number of time points before and after the abnormal time point are selected for sorting and grouping; if the continuous abnormality condition is not met, only the abnormal time point is selected for sorting and grouping.

[0025] Anomaly source determination includes: analyzing the linear correlation between changes in initial blanking parameters, changes in blanking process parameters, and changes in blanking result parameters and the source factors based on the accuracy deviation feature set, and configuring a feature analyzer; it should be noted that changes in initial blanking parameters include thickness changes, temperature changes, changes in initial positioning state, and material batch changes; changes in blanking process parameters include changes in blanking force, punch displacement, changes in the first slope characteristic, changes in blanking pressure, and changes in the second slope characteristic; changes in blanking result parameters include dimensional deviations and deformation deviations; Linear correlation can be characterized by the Pearson correlation coefficient, which represents the degree of linear synchronization between two data sequences. The calculation logic is as follows: First, calculate the average value of the change sequence and the precision deviation feature set corresponding to the source factors respectively. Then, calculate the deviation between the data at each corresponding time point and their respective average values. Multiply the two deviations at each time point and sum them. Then, calculate the square root of the sum of squares of the deviations of the two sequences respectively. Finally, divide the sum of the products of the two deviations mentioned above by the product of the square roots of the two sums of squares mentioned above to obtain the linear correlation coefficient. It should be noted that the closer the correlation coefficient is to positive one, the more the two sequences tend to change in the same direction; the closer it is to negative one, the more the two sequences tend to change in opposite directions; and the closer it is to zero, the weaker the linear relationship between the two. The feature analyzer can be implemented using a sliding time window, difference processing, peak extraction, variance extraction, trend extraction, or a combination of multiple features. First, the data is segmented according to the aforementioned initial time period. Then, the peak value and variance of the source factors and precision deviation feature set are extracted in each initial time period, so that subsequent comparisons can correspond to the same punching stage. If the correlation is linear, the source factors whose deviations in the feature set are synchronous or have consistent trends are marked as key source factors. If the correlation is nonlinear, feature standardization is performed, feature processing is carried out, and the feature contribution of the source factors is determined. The specific calculation methods include: Determine the magnitude of change of the source factor during the current abnormal period; determine the degree of consistency between the changes in the process characteristics and accuracy deviation characteristics corresponding to the source factor; determine the degree of temporal consistency between the time when the change in the source factor occurs and the time when the accuracy deviation changes significantly; normalize the above three degrees, and then synthesize them according to preset weights to obtain the characteristic contribution degree of the source factor: The contributions of the magnitude of change, the consistency of change direction, and the chronological order between the changes in the source factor and the changes in the accuracy deviation are converted into contribution scores under the same dimension. The three contribution scores are multiplied by their corresponding weights and then summed to obtain the basic contribution level of the source factor. It should be noted that the larger the contribution of the magnitude of change, the more obvious the change in the relevant accuracy deviation is when the source factor changes; the higher the contribution of the consistency of change direction, the more consistent the change direction of the source factor and the accuracy deviation; and the higher the contribution of the chronological order, the more stably the change of the source factor precedes the occurrence of the accuracy deviation. The steps for determining the weights include: for each contribution score, calculating the difference between the average level of the data in the corresponding sample at the confirmed abnormal time and the average level in the corresponding sample at the normal time, and combining the fluctuation of the data in the abnormal and normal samples, where the fluctuation is defined as half of the absolute value of the difference between the maximum and minimum values, to determine the discriminative power of the data; normalizing the discriminative power of each contribution score to obtain the corresponding original weights, and the sum of the three weights is 1; Key source factors are identified by sorting them in descending order of feature contribution; among them, key source factors include at least one of the following: material state change, blanking clearance change, positioning state change, and injection mold wear state change.

[0026] Anomaly detection is performed on the set of precision deviation features. The determination of the time of anomaly in punching precision is combined with the determination of the source of anomaly. By sorting and grouping the anomaly times and configuring dimensional constraints in combination with the current edge pressing state, the anomaly times that do not meet the dimensional constraints are adjusted and processed to form candidate control periods. By converting discrete anomaly times into candidate control periods with temporal continuity, the misjudgment caused by instantaneous vibration, sampling noise and local material disturbance during the punching process can be reduced. Furthermore, the change law of precision deviation features within the candidate control periods is used to identify key source factors, thereby improving the accuracy and stability of punching anomaly positioning. For the identified key source factors, a correspondence between key source factors and control classification ranges is established, and the corresponding control classification range is matched according to the current punching accuracy. The influence of the control strategy on the current punching accuracy deviation is determined according to the accuracy improvement degree corresponding to the historical control strategy. Priority adjustment strategies are selected from multiple candidate control parameters, and the adjustment direction and adjustment magnitude are determined in combination with the current accuracy deviation direction. This enables the control system to select different control parameters for different sources of anomalies, rather than uniformly performing fixed parameter adjustments for all anomalies. This improves the targeting of adaptive control and reduces control coupling and over-adjustment caused by simultaneous adjustment of multiple parameters.

[0027] S4: Identify the current blanking accuracy, determine the control classification range associated with key source factors, and determine the adjustment direction and magnitude of the control strategy; Based on key source factors, source features corresponding to the key source factors are determined, and these features are ranked according to their contribution level. Source features whose contribution level reaches a preset contribution threshold are selected, and corresponding precision deviation feature data are extracted from historically confirmed normal time samples based on these source features. Then, the control classification range associated with the key source factors is determined, including: Establish the correlation between key source factors and blanking accuracy; extract the similarity between control vectors and feature contribution levels, filter the accuracy deviation feature set corresponding to similarity greater than or equal to the standard similarity threshold, determine whether the feature parameters analyzed in real time fall within the control classification range corresponding to normal operating condition fluctuations, and perform circular clustering analysis to determine the circular judgment radius; determine at least one normal type classification cluster, and use the cluster center of each normal type classification cluster as the normal control center, perform a circle drawing operation with the circular judgment radius, mark those falling inside the circle as normal deformation, and match the control strategy; mark those falling outside the circle as abnormal blanking; The injection mold is subjected to blanking control according to the adjustment direction and adjustment range, and the blanking result data after control is obtained; Based on the deviation between the controlled blanking result data and the standard blanking accuracy threshold, the blanking accuracy corresponding to the current edge pressing state is updated, so that the subsequent control strategy of the injection mold can be adaptively adjusted according to the updated edge pressing state. Determine the adjustment direction and magnitude of the control strategy, including: obtaining the blanking accuracy before and after the change of the control strategy in multiple historical batches; determining the degree of influence of the control strategy on the current blanking accuracy deviation based on the correspondence between the control strategy and the blanking accuracy; wherein, the degree of influence is defined as: the ratio between the adjustment amount of each control strategy and the allowable adjustment range of the corresponding control strategy, and normalized to make it range between 0 and 1. Based on the degree of influence, the control strategy with the highest degree of influence on the current blanking accuracy deviation is selected from multiple candidate control parameters as the priority adjustment strategy; based on the deviation direction of the current blanking accuracy, the adjustment direction of the priority adjustment strategy is determined, specifically: based on the deviation direction of the current blanking accuracy deviation and the correspondence between the change direction of the priority adjustment strategy and the change direction of the blanking accuracy deviation in historical batches, the adjustment direction of the priority adjustment strategy is determined. Based on the blanking accuracy corresponding to the same control strategy in the historical blanking process, the adjustment range of the priority adjustment strategy is determined. Specifically, the similarity between the current blanking and the historical batch blanking is used to screen similar historical batches. Based on the adjustment range corresponding to the priority adjustment strategy in the similar historical batches, the initial adjustment range of the priority adjustment strategy is determined. If the initial adjustment range is greater than the preset maximum adjustment range, the initial adjustment range is limited to the preset maximum adjustment range. After the adjustment is completed, the blanking accuracy is re-acquired so as to determine the next adjustment based on the adjusted blanking accuracy. After determining the adjustment direction and magnitude of the control strategy, the process includes: performing blanking control on the injection mold according to the adjustment direction and magnitude, and obtaining the blanking result data after control; updating the blanking accuracy corresponding to the current blanking state based on the deviation between the blanking result data after control and the standard blanking accuracy threshold. If the blanking result data after control is less than the minimum value of the standard blanking accuracy threshold or greater than the maximum value of the standard blanking accuracy threshold, and the blanking accuracy deviation is reduced relative to the blanking accuracy deviation before control, the adjustment direction and adjustment range are re-determined based on the remaining blanking accuracy deviation after control, and the next blanking control is executed. If the deviation of the blanking result data after control from the blanking accuracy before control increases, the impact of the current control strategy is reduced, and the abnormality source determination is re-executed. If the blanking result data after performing blanking control for a preset number of consecutive times is still less than the minimum value of the standard blanking accuracy threshold or greater than the maximum value of the standard blanking accuracy threshold, the current blanking operation will be marked as abnormal control, and the adjustment range will be stopped from being further expanded. Based on the current controlled blanking result data and the updated blanking accuracy corresponding to different edge pressing states, the edge pressing state of the current blanking station area is re-determined. The standard blanking accuracy threshold is used to characterize the allowable accuracy range of the injection molded part to meet product quality requirements after blanking. Its sources include at least one of the following: product design dimensions and corresponding dimensional tolerances, product functional requirements, assembly requirements, applicable product standards or process quality standards, and quality requirements determined through process verification. In this embodiment, it specifically refers to the allowable accuracy range of dimensional deviation and deformation deviation. For dimensional deviation, the standard dimensional range is determined based on the product design dimensions and corresponding dimensional tolerances. For deformation deviation, the standard deformation range is determined based on product functional requirements, assembly requirements, or the allowable deformation range determined through process verification. If it falls into the deviation anomaly type category cluster, it will be marked as an abnormal punching, and an alarm will be issued directly.

[0028] Example 2: This embodiment of the invention provides an adaptive control system for the blanking accuracy of injection molds. The system includes: an acquisition module, a feature determination module, an anomaly determination module, and an adaptive control module, and the acquisition module, feature determination module, anomaly determination module, and adaptive control module are communicatively connected to each other. Acquisition module: used to acquire the initial blanking parameters, blanking process parameters, and blanking result parameters of the injection mold to be blanked; Feature Determination Module: Based on the initial blanking parameters, blanking process parameters, and blanking result parameters, this module establishes the correspondence between the blanking die pressing state and the blanking accuracy, determines the feature changes corresponding to the blanking accuracy under the current blanking gap, and determines the set of accuracy deviation features. Anomaly Detection Module: This module is used to perform anomaly detection based on the set of precision deviation features, including the detection of abnormal moments and sources of anomalies in blanking precision. By sorting and grouping abnormal moments and configuring dimensional constraints based on edge pressing status, adjustment processing is performed on abnormal moments that do not meet the dimensional constraints to generate candidate control periods. For candidate control periods, key source factors are determined by using the set of precision deviation features to detect anomalies. Adaptive control module: used to identify the current blanking accuracy, determine the control classification range associated with key source factors, and determine the adjustment direction and adjustment range of the control strategy.

[0029] Based on the current blanking accuracy as feedback, the blanking accuracy corresponding to the current blanking state is updated according to the deviation between the controlled blanking result and the standard blanking accuracy range. The updated blanking process data and blanking result data are used for subsequent control strategy selection, so that the blanking state of no blanking, single-sided blanking, and double-sided blanking can be continuously corrected according to the actual blanking results. This realizes a closed-loop correlation between blanking accuracy judgment, anomaly source identification, control parameter adjustment, and control result feedback, thereby improving the blanking accuracy stability of complex injection molded parts under different material states, blanking gaps, and partitioned area structures.

[0030] In the application, the various formulas mentioned are all calculated by removing dimensions and taking their numerical values. The formulas are derived from the most recent real-world situation by collecting a large amount of data and conducting software simulations. The formulas are set by those skilled in the art according to the actual situation.

[0031] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0032] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0033] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes 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.

Claims

1. An adaptive control method for the blanking accuracy of injection molds, characterized in that, The method includes: Obtain the initial blanking parameters, blanking process parameters, and blanking result parameters of the injection mold to be blanked; Based on the initial blanking parameters, blanking process parameters, and blanking result parameters, establish the correspondence between the blanking state of the injection mold and the blanking accuracy, determine the characteristic changes of the blanking accuracy under the current blanking gap, and determine the set of accuracy deviation characteristics. Anomaly detection is performed based on the set of accuracy deviation features, including the detection of abnormal moments and sources of anomalies in blanking accuracy. By sorting and grouping abnormal moments and configuring dimensional constraints in conjunction with the edge pressing state, adjustment processing is performed on abnormal moments that do not meet the dimensional constraints to generate candidate control periods. For candidate control periods, the key source factors are determined by detecting the abnormal sources of the set of accuracy deviation features. Identify the current blanking accuracy, determine the control classification range associated with key source factors, and determine the adjustment direction and magnitude of the control strategy.

2. The adaptive control method for punching accuracy of injection molds according to claim 1, characterized in that, The steps for obtaining the initial blanking parameters, blanking process parameters, and blanking result parameters include: Before the punching process, obtain the thickness, temperature, position of the injection mold, and the material of each batch of injection mold; The degree of thickness variation in the blanking station area is determined based on the thickness. Determine the degree of temperature change in the punching station area based on the temperature; Determine the initial state of the injection mold relative to the punching tool based on its position; The thickness variation, temperature variation, initial state, and each batch of injection mold material are correlated according to the blanking station area to form the initial blanking parameters; During the blanking process, the punch displacement, punch speed, blanking pressure, and blanking force are obtained; Based on the punch displacement and punching force, a first response curve is established, and the first slope characteristic is identified; Based on the blanking pressure and punching speed, a second response curve is established, and a second slope characteristic is identified; The first slope feature and the second slope feature are correlated in chronological order to form the punching process parameters; After the blanking process is completed, obtain the blanking edge dimensions and deformation data; Determine the dimensional deviation based on the blanking dimensions; Determine the deformation deviation based on the deformation data; The blanking result parameters are generated by associating dimensional deviation and deformation deviation according to different edge clamping states; The pressing state includes no pressing, single-sided pressing, and double-sided pressing.

3. The adaptive control method for punching accuracy of injection molds according to claim 1, characterized in that, Establish the correspondence between the blanking state of the injection mold and the blanking accuracy, including: Divide the work area into multiple blanking stations, determine the initial blanking parameters for each blanking station area, and obtain status data for several blanking station areas. Associate the status data of each punching station area with the punching process parameters and punching result parameters corresponding to different pressing states of the same punching station area respectively; Identify the changes in initial blanking parameters, blanking process parameters, and blanking result parameters of the same blanking station area in different batches; wherein the changes include at least one representation of the magnitude and direction of change; The influence relationship on the status data of the punching station area is determined based on the changes in initial punching parameters, changes in punching process parameters, and changes in punching result parameters. Mark the blanking station areas that have the same influence relationship but correspond to different blanking accuracies.

4. The adaptive control method for punching accuracy of injection molds according to claim 3, characterized in that, Abnormal moment determination includes: The first-order difference of the precision deviation feature set is used to obtain the difference value sequence; Multiple initial time periods are obtained by segmenting the differential value sequence based on the positive and negative change times. At any initial time period, based on the known punching gap and feature changes, a multi-dimensional feature vector is constructed, which includes at least: feature parameters, punching gap and timestamp; Among them, the blanking clearance includes small clearance, medium clearance and large clearance, and the characteristic parameters include the change of initial blanking parameters, the change of blanking process parameters and the change of blanking result parameters; Extract the peak value and variance of the characteristic parameters of any initial time period, and obtain the punching gap of the initial time period. Establish several control vectors according to different combinations, and generate control deviation index based on the control vectors. For the control vectors at the same time, select at least one corresponding time when the control deviation index is greater than the standard abnormal threshold, and mark it as an abnormal time.

5. The adaptive control method for punching accuracy of injection molds according to claim 4, characterized in that, Generate candidate control periods, including: Based on the control deviation index corresponding to one quantity time before and one quantity time after the abnormal time, determine whether the abnormal time meets the continuous abnormality condition. If the consecutive anomaly condition is met, select one time point before and after the anomaly to enter the sorting and grouping process; if the consecutive anomaly condition is not met, only select the anomaly time point to enter the sorting and grouping process.

6. The adaptive control method for punching accuracy of injection molds according to claim 5, characterized in that, Anomaly source determination includes: Based on the precision deviation feature set, analyze the linear correlation between feature changes and source factors, and configure feature processors; if the correlation is linear, analyze the source factors whose deviations in the precision deviation feature set are synchronous or have consistent trends, and mark them as key source factors; if the correlation is nonlinear, perform feature processing, determine the feature contribution degree of the source factors, and determine the key source factors by sorting them in descending order based on the feature contribution degree. Among them, the key source factors include at least one of the following: changes in material condition, changes in blanking clearance, changes in positioning condition, and changes in the wear condition of injection mold.

7. The adaptive control method for punching accuracy of injection molds according to claim 1, characterized in that, Determine the scope of control classifications associated with key source factors, including: Extract the similarity between the control vector and the feature contribution level, filter the precision deviation feature set corresponding to the similarity greater than or equal to the standard similarity threshold, determine whether the feature parameters analyzed in real time fall into the control classification range corresponding to the fluctuation of normal operating conditions, and perform circular clustering analysis to determine the circular judgment radius; Identify at least one normal type cluster, and use the cluster center of each normal type cluster as the normal control center. Perform a circle drawing operation with a circular decision radius. Mark those falling inside the circle as normal deformation and match the control strategy; mark those falling outside the circle as abnormal cutting.

8. The adaptive control method for punching accuracy of injection molds according to claim 7, characterized in that, Determine the direction and magnitude of the control strategy adjustment, including: Obtain the punching accuracy before and after the control strategy change in multiple historical batches; Based on the correspondence between control strategy and blanking accuracy, determine the degree of influence of control strategy on the current blanking accuracy deviation; Based on the degree of influence, the control strategy with the highest degree of influence on the current punching accuracy deviation is selected from multiple candidate control parameters as the priority adjustment strategy. Based on the current deviation direction of the punching accuracy, determine the adjustment direction of the priority adjustment strategy; Based on the blanking accuracy corresponding to the same control strategy in historical blanking processes, determine the adjustment range of the priority adjustment strategy.

9. The adaptive control method for blanking accuracy of injection molds according to claim 8, characterized in that, After determining the direction and magnitude of the control strategy adjustment, the following steps are included: The injection mold is subjected to blanking control according to the adjustment direction and adjustment range, and the blanking result data after control is obtained; Based on the deviation between the controlled blanking result data and the standard blanking accuracy threshold, update the blanking accuracy corresponding to the current edge pressing state.

10. A system applied to the adaptive control method for blanking accuracy of injection molds according to any one of claims 1-9, characterized in that, The system includes: The acquisition module acquires the initial blanking parameters, blanking process parameters, and blanking result parameters of the injection mold to be blanked; The feature determination module establishes the correspondence between the blanking die pressing state and the blanking accuracy based on the initial blanking parameters, blanking process parameters, and blanking result parameters. It determines the feature changes corresponding to the blanking accuracy under the current blanking gap and determines the set of accuracy deviation features. The anomaly detection module performs anomaly detection based on the set of precision deviation features, including the detection of abnormal moments and sources of anomalies in blanking precision. By sorting and grouping abnormal moments and configuring dimensional constraints based on the edge pressing state, adjustment processing is performed on abnormal moments that do not meet the dimensional constraints to generate candidate control periods. For candidate control periods, the key source factors are determined by using the anomaly source detection method on the set of precision deviation features. The adaptive control module identifies the current blanking accuracy, determines the control classification range associated with key source factors, and determines the adjustment direction and magnitude of the control strategy.

Images