A method for inspecting the quality of a molded article of a tail node of a batch of insert injection molding of an automobile

Abstract

The application discloses a kind of automobile insert injection molding batch tail node forming quality inspection method, its sampling data to tail node forming quality qualified automobile insert injection molding batch is preprocessed, and corresponding batch feature matrix is obtained;All batch feature matrix is converted into a data matrix, the average value and standard deviation of all elements in each column vector in the data matrix are calculated to implement standardization processing, the data matrix after standardization processing is implemented singular value decomposition, and then tail node forming quality inspection threshold value is obtained;Combining the average value and standard deviation of all elements in each column vector in the above data matrix, the unitary matrix and diagonal matrix obtained by singular value decomposition, tail node forming quality inspection threshold value, the sampling data of latest automobile insert injection molding batch collected by injection molding machine is used to inspect tail node forming quality in time;Advantages are that the relationship between measurement variables and the data relationship on sampling time sequence can be analyzed, and redundancy in characteristics can be eliminated.

Description

Technical Field

[0001] This invention relates to a method for inspecting the molding quality of automotive inserts, and more particularly to a method for inspecting the molding quality of the last node of an injection molding batch of automotive inserts. Background Technology

[0002] Insert injection molding is a process in which inserts are pre-fixed in appropriate positions within an injection mold, and then plastic is injected to form the insert. After the mold opens, the insert is encased and embedded within the cooled and solidified plastic, resulting in a product with inserts such as threads or electrodes. Insert injection molding is widely used in various industries, including automotive, electronics, and connectors. In the production process of automotive inserts, granular or powdered raw materials are typically added to the hopper of an injection molding machine. The raw materials are heated and melted into a flowing state. Driven by the screw or piston of the injection molding machine, they enter the mold cavity through the nozzle and the mold's gating system, where they harden and solidify. The entire molding cycle is short and production efficiency is high, but it is also highly susceptible to the effects of injection pressure, injection time, and injection temperature, which can lead to substandard molding quality. Since automotive inserts are molded as a single piece, a corresponding molded product is obtained at the end of the entire injection batch. Therefore, real-time inspection of the molding quality at the end of the batch is of great significance for the quality control of automotive inserts.

[0003] Although the injection molding batches of automotive inserts are short-lived, the injection molding machines still measure and report a large amount of status data, such as temperature, pressure, and displacement, through sensors. These status data are not only interconnected but also sequentially related, both of which directly or indirectly affect the inserts formed at the end of the batch. Therefore, by monitoring the changes in the status data of automotive insert injection molding batches, it is possible to indirectly inspect whether the quality of the final stage of molding meets requirements. In other words, real-time inspection of the final stage molding quality can be achieved using batch data. However, slight differences in the time between automotive insert injection molding batches lead to inconsistencies in the amount of status data obtained from different batches. This batch length variation problem presents new technical challenges.

[0004] Because the entire batch production process of automotive inserts affects the molding quality of the final stage, and the short production time of each batch means that the temporal sequence of state data influences the molding quality of the final stage, the unequal length of state data for each batch presents a technical obstacle to analyzing the differences in normal variations between batches. Therefore, analyzing the data for each batch of automotive inserts requires considering not only the temporal relationship of state data but also the differences in normal variations between different production batches. In existing technical methods, feature analysis of the collected data is the basic approach. However, how to design a corresponding feature analysis modeling process to address the complexity of automotive insert batch data, thereby considering not only the relationships between state data and their temporal sequence but also the differences in normal variations between different batches, remains an unsolved problem. Summary of the Invention

[0005] The main technical problem this invention aims to solve is how to analyze batch data of automotive inserts to represent the relationships between data points, their order, and batches, thereby defining a normal range of variation and enabling real-time inspection of the molding quality of the tail node. Specifically, the method first performs feature transformation on the data of each normal batch of automotive inserts, converting batch data of unequal length into feature vectors of the same dimension. Then, feature analysis is performed on multiple feature vectors to extract the normal variation differences between batches. Finally, by applying the same transformation and processing to new batch data, corresponding features reflecting the molding quality of the tail node are obtained. By monitoring changes in these features, the quality of the tail node molding can be verified to ensure it meets quality requirements.

[0006] The technical solution adopted by this invention to solve the above-mentioned technical problems is: a method for inspecting the molding quality of the last node of an injection molding batch of automotive inserts, characterized by specifically including the following steps:

[0007] Step 1: Preprocess the sampling data of the automotive insert injection molding batches with qualified tail node molding quality, specifically including steps 1.1 to 1.3 as shown below;

[0008] Step 1.1: From the historical database corresponding to the injection molding machine, select the sampling data of B batches of automotive inserts with qualified tail-end molding quality, and represent them as B batch data matrices, denoted as X1, X2, ..., X... B ; where the data matrix of the b-th batch is X b X b The data in columns 1 to 13 are generated from 13 measured variables of the injection molding machine in N. bThe data collected at each sampling time point consists of 13 measured variables arranged in the following order: cylinder pressure, nozzle pressure, plasticizing pressure, mold cavity pressure, screw stroke, screw speed, barrel section 1 temperature, barrel section 2 temperature, barrel section 3 temperature, barrel section 4 temperature, oil temperature, nozzle temperature, and mold temperature. b The dimension is N b ×13, numbered b=1,2,…,B;

[0009] Step 1.2: For X1, X2, ..., X B After standardization, a standardized batch data matrix is ​​obtained, which is denoted as follows: The specific implementation process includes steps A1 to D1 as shown below;

[0010] Step A1: Set b = 1, then set X = X b ;

[0011] Step B1: Represent the first, second, through thirteenth column vectors of X as x1, x2, ..., x 13 Next, calculate the mean and standard deviation of all elements in each column vector of X, and denot the mean of all elements in each column vector of X as μ1, μ2, ..., μ 13 Let the standard deviations of all elements in each column vector of X be denoted as δ1, δ2, ..., δ1, respectively. 13 Where μ1 and δ1 represent the mean and standard deviation of all elements in x1, respectively, and μ2 and δ2 represent the mean and standard deviation of all elements in x2, respectively. 13 and δ 13 They respectively represent x 13 The mean and standard deviation of all elements in the dataset;

[0012] Step C1: Set m = 1, 2, ..., 13 sequentially, and simultaneously apply the formula... For x m Standardization is performed to obtain the standardized column vector. Then Combined into a single data matrix, denoted as Where, x m Let m be the m-th column vector in X. x represents m The column vector obtained after standardization, μ m x represents m The average value of all elements in the δ m x represents m The standard deviation of all elements in the sample;

[0013] Step D1: Set the standardized batch b data matrix as follows make Next, check if b is less than B. If so, set b = b + 1, and then set X = X. b If not, return to step B1 to continue execution; otherwise, obtain the standardized batch data matrix.

[0014] Step 1.3: For each After performing feature length equalization, B 13×13 dimensional batch feature matrices are obtained, denoted as Φ1, Φ2, ..., Φ3. B Specifically, it includes steps A2 to C2 as shown below;

[0015] Step A2: Set b=1, then set...

[0016] Step B2: Iterative calculation The corresponding left transformation matrix and right transformation matrix Thus the matrix The sum of squares of the elements on the middle diagonal is the largest;

[0017] Step C2: Set Φ b =Φ; then check if b is less than B. If so, set b = b + 1, then set... Then return to step B2 to continue execution; otherwise, obtain B batch feature matrices Φ1,Φ2,…,Φ B The dimension of the feature matrix for each batch is 13×13.

[0018] Step 2: Establish an inspection mechanism to determine whether the tail node forming quality is up to standard, specifically including steps 2.1 to 2.5 as shown below;

[0019] Step 2.1: According to the formula The batch feature matrices Φ1, Φ2, ..., Φ are respectively... B Expand into their respective row vectors, and denote them as φ1, φ2, ..., φ B Then, φ1, φ2, ..., φ B The data are combined into a B×169 dimensional data matrix, denoted as Y; where φ b Φ b The corresponding row vector, They respectively represent Φ b The first, second, ..., thirteenth row vectors in Y, and the first, second, ..., Bth row vectors in Y are respectively equal to φ1, φ2, ..., φ B Numbered b = 1, 2, ..., B;

[0020] Step 2.2: Represent the first column vector, the second column vector, up to the 169th column vector in Y as y1, y2, ..., y 169 Next, calculate the mean and standard deviation of all elements in each column vector of Y, and denot the mean of all elements in each column vector of Y as ζ1, ζ2, ..., ζ. 169 Let the standard deviations of all elements in each column vector of Y be denoted as θ1, θ2, ..., θ 169 Then according to the formula For y1, y2, ..., y 169 After standardization, we obtain a standardized column vector, which is denoted as follows: Among them, y j Let j represent the j-th column vector in Y. Indicates y j The standardized column vectors are ζ1 and θ1, which represent the mean and standard deviation of all elements in y1, respectively, and ζ2 and θ2, which represent the mean and standard deviation of all elements in y2, respectively. 169 and θ 169 Each corresponds to y 169 The mean and standard deviation of all elements in the ζ j and θ j These respectively represent the j-th column vector y in Y. j The mean and standard deviation of all elements in the column, column number j = 1, 2, ..., 169;

[0021] Step 2.3: [The text appears to be incomplete and contains several grammatical errors. A more accurate translation would require Combined into a B×169 dimensional data matrix, denoted as Again Perform singular value decomposition, that is: Where U and V represent two unitary matrices of the singular value decomposition, S represents a diagonal matrix composed of non-zero singular values, and the superscript "T" indicates the transpose of a matrix or vector.

[0022] Step 2.4: According to the formula Θ=UU T Calculate matrix Θ; then calculate the mean p and standard deviation q of the B elements on the diagonal of Θ; finally, set the tail node molding quality inspection threshold. It equals p + 3.5q;

[0023] Step 2.5: Retain the parameters used for the tail node forming quality inspection, specifically including the average values ​​ζ1, ζ2, ..., ζ from Step 2.2. 169 Sum of standard deviations θ1, θ2, ..., θ 169 The unitary matrix V and diagonal matrix S in step 2.3, and the tail node forming quality inspection threshold in step 2.4.

[0024] Step 3: Instantly inspect the molding quality of the tail node using the latest sampling data of the automotive insert injection batch collected by the injection molding machine, specifically including steps 3.1 to 3.5 as shown below;

[0025] Step 3.1: After the latest automotive insert injection molding batch is completed, immediately acquire the corresponding sampling data and represent it as a batch data matrix, denoted as X. new ; where X new The data in columns 1 through 13 correspond to the 13 measurement variables in step 1.1, respectively.

[0026] Step 3.2: Following the procedures in steps B1 and C1, process X in the same way. new Standardization is performed to obtain a standardized batch data matrix, denoted as... ;

[0027] Step 3.3: Settings Then, following the process in step B2, iteratively calculate Z in the same manner. new The corresponding left transformation matrix U new and right transformation matrix V new Then calculate matrix Φ new =U new Z new V new ;

[0028] Step 3.4: According to the formula Φ new Expand into the corresponding row vector, denoted as y; then according to the formula Standardize each element in y individually to obtain the standardized row vector, denoted as . in, Corresponding to Φ new The first row vector, the second row vector, ..., the thirteenth row vector in the vector, y (j) This represents the j-th element in y. This represents the j-th element in y, with column numbers j = 1, 2, ..., 169;

[0029] Step 3.5: According to the formula Calculate test indicators Re-evaluation Is it greater than If not, the tail node molding quality is considered qualified; if yes, the tail node molding quality is considered unqualified. After the inspection is completed, return to step 3.1 to continue to inspect the tail node molding quality of the latest automotive insert injection molding batch.

[0030] The specific implementation process of step B2 includes the following steps B2.1 to B2.5:

[0031] Step B2.1: Randomly generate a 13×13 dimensional right transformation matrix from the interval (-1, 1).

[0032] Step B2.2: Solve for the symmetric matrix The 13 eigenvalues ​​correspond to the eigenvectors u1, u2, ..., u 13 Then according to the formula For u1, u2, ..., u respectively 13 Normalization is performed; where the superscript "T" denotes the transpose of a matrix or vector, and m = 1, 2, ..., 13. The "=" sign in the equation is the assignment operator, and the u on the right side of the "=" sign... m express The eigenvector corresponding to the m-th eigenvalue, u on the left side of the "=" m This represents the vector obtained after normalization.

[0033] Step B2.3: Normalize u1, u2, ..., u 13 Combined into a left transformation matrix Solve the symmetric matrix again The 13 eigenvalues ​​correspond to the eigenvectors v1, v2, ..., v 13 ;

[0034] Step B2.4: According to the formula For v1, v2, ..., v respectively 13 Perform normalization; then normalize the v1, v2, ..., v 13 Combined into a right transformation matrix in, The "=" sign in the equation is the assignment operator, and the value v on the right side of the "=" sign is... m express The eigenvector corresponding to the m-th eigenvalue, with v on the left side of the "=" sign. m This represents the vector obtained after normalization.

[0035] Step B2.5: Determine If convergence has occurred, return to step B2.2 to continue execution; if convergence has occurred, utilize the obtained left transformation matrix. and right transformation matrix Calculate matrix

[0036] In step B2.5, The criteria for convergence are The elements in the middle no longer change.

[0037] Compared with the prior art, the advantages of the present invention are:

[0038] First, the method of the present invention performs two transformations simultaneously on the sampling data of each batch of automotive insert injection molding. This not only converts batch data matrices of unequal length into batch feature matrices of the same dimension, but also allows for the simultaneous analysis and extraction of the relationships between measured variables and the data relationships in terms of sampling time through the left and right transformation matrices. Second, the method of the present invention performs singular value decomposition on the expanded batch feature matrix after the left and right transformation to eliminate redundancy in the features, and achieves the purpose of verifying whether the molding quality of the tail node is qualified by examining the changes in the features. Attached Figure Description

[0039] Figure 1 This is a block diagram illustrating the overall implementation of the method of the present invention;

[0040] Figure 2 This is a quality inspection diagram for the tail node forming process. Detailed Implementation

[0041] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0042] The present invention proposes a batch tail node molding quality inspection method for automotive insert injection molding, the overall implementation block diagram of which is shown below. Figure 1 As shown, it specifically includes the following steps:

[0043] Step 1: Preprocess the sampling data of the automotive insert injection molding batches with qualified tail node molding quality, specifically including steps 1.1 to 1.3 as shown below.

[0044] Step 1.1: From the historical database corresponding to the injection molding machine, select the sampling data of B batches of automotive inserts with qualified tail-end molding quality, and represent them as B batch data matrices, denoted as X1, X2, ..., X... B Where B represents the number of batch data matrices selected, X1 represents the first batch data matrix, X2 represents the second batch data matrix, and X... B Let X represent the data matrix of the Bth batch and X represent the data matrix of the bth batch. b X b The data in columns 1 to 13 are generated from 13 measured variables of the injection molding machine in N. b The data collected at each sampling time point consists of 13 measured variables arranged in the following order: cylinder pressure, nozzle pressure, plasticizing pressure, mold cavity pressure, screw stroke, screw speed, barrel section 1 temperature, barrel section 2 temperature, barrel section 3 temperature, barrel section 4 temperature, oil temperature, nozzle temperature, and mold temperature. b The dimension is Nb ×13, numbered b=1,2,…,B.

[0045] Step 1.2: For X1, X2, ..., X B After standardization, a standardized batch data matrix is ​​obtained, which is denoted as follows: The specific implementation process includes steps A1 to D1 as shown below.

[0046] Step A1: Set b = 1, then set X = X b .

[0047] Step B1: Represent the first, second, through thirteenth column vectors of X as x1, x2, ..., x 13 Next, calculate the mean and standard deviation of all elements in each column vector of X, and denot the mean of all elements in each column vector of X as μ1, μ2, ..., μ 13 Let the standard deviations of all elements in each column vector of X be denoted as δ1, δ2, ..., δ1, respectively. 13 Where x1 represents the first column vector in X, x2 represents the second column vector in X, and x... 13 Let X represent the 13th column vector. The calculation of the mean and standard deviation of multiple data is in the prior art. μ1 and δ1 respectively represent the mean and standard deviation of all elements in x1, and μ2 and δ2 respectively represent the mean and standard deviation of all elements in x2. 13 and δ 13 They respectively represent x 13 The mean and standard deviation of all elements in the dataset.

[0048] Step C1: Set m = 1, 2, ..., 13 sequentially, and simultaneously apply the formula... For x m Standardization is performed to obtain the standardized column vector. Then Combined into a single data matrix, denoted as Where, x m Let m be the m-th column vector in X. x represents m The column vector obtained after standardization, μ m x represents m The average value of all elements in the δ m x represents m The standard deviation of all elements in the sample. This represents the column vector obtained after standardizing x1. This represents the column vector obtained after standardizing x2. x represents13 The column vector obtained after standardization.

[0049] Step D1: Set the standardized batch b data matrix as follows make Next, check if b is less than B. If so, set b = b + 1, and then set X = X. b If not, return to step B1 to continue execution; otherwise, obtain the standardized batch data matrix. In b = b + 1, the "=" sign is the assignment operator. This represents the batch data matrix obtained after X1 has been standardized. This represents the batch data matrix obtained after X2 has been standardized. X represents B The batch data matrix obtained after standardization.

[0050] Step 1.3: For each After performing feature length equalization, B 13×13 dimensional batch feature matrices are obtained, denoted as Φ1, Φ2, ..., Φ3. B Specifically, it includes steps A2 to C2 as shown below.

[0051] Step A2: Set b=1, then set...

[0052] Step B2: Iterative calculation The corresponding left transformation matrix and right transformation matrix Thus the matrix The sum of squares of the elements on the diagonal is the largest.

[0053] In this specific embodiment, step B2 is implemented through steps B2.1 to B2.5 as shown below:

[0054] Step B2.1: Randomly generate a 13×13 dimensional right transformation matrix from the interval (-1, 1).

[0055] Step B2.2: Solve for the symmetric matrix The 13 eigenvalues ​​correspond to the eigenvectors u1, u2, ..., u 13 Then according to the formula For u1, u2, ..., u respectively 13 Normalization is performed; where the superscript "T" denotes the transpose of a matrix or vector, and m = 1, 2, ..., 13. The "=" sign in the equation is the assignment operator, and the u on the right side of the "=" sign... m express The eigenvector corresponding to the m-th eigenvalue, u on the left side of the "=" m This represents the vector obtained after normalization.

[0056] Step B2.3: Normalize u1, u2, ..., u 13 Combined into a left transformation matrix Solve the symmetric matrix again The 13 eigenvalues ​​correspond to the eigenvectors v1, v2, ..., v 13 .

[0057] Step B2.4: According to the formula For v1, v2, ..., v respectively 13 Perform normalization; then normalize the v1, v2, ..., v 13 Combined into a right transformation matrix in, The "=" sign in the equation is the assignment operator, and the value v on the right side of the "=" sign is... m express The eigenvector corresponding to the m-th eigenvalue, with v on the left side of the "=" sign. m This represents the vector obtained after normalization.

[0058] Step B2.5: Determine If convergence has occurred, return to step B2.2 to continue execution; if convergence has occurred, utilize the obtained left transformation matrix. and right transformation matrix Calculate matrix Here, The criteria for convergence are The elements in the middle no longer change.

[0059] In step B2, obtain The corresponding left transformation matrix and right transformation matrix To maximize the sum of squares of the elements on the diagonal of Φ, we need to: in, This represents calculating the sum of squares of all elements, where `max` is the function to find the maximum value. This is achieved through a left transformation matrix. It can analyze and extract The data relationship at different sampling time points is reflected in the right transformation matrix. Then it can be analyzed and extracted The relationship between different measurement variables.

[0060] because The `trace()` function calculates the sum of the elements on the diagonal, thus the Lagrange multiplier method easily yields two typical eigenvalue problems: and Here, λ and η represent diagonal matrices composed of their respective 13 eigenvalues.

[0061] Due to the problem of solving eigenvalues Require Given, and solving the eigenvalue problem Require As is known, the method of this invention is designed to obtain the left transformation matrix iteratively. and right transformation matrix The implementation process is shown in steps B2.1 to B2.5.

[0062] Step C2: Set Φ b =Φ; then check if b is less than B. If so, set b = b + 1, then set... Then return to step B2 to continue execution; otherwise, obtain B batch feature matrices Φ1,Φ2,…,Φ B In b = b + 1, the "=" sign is an assignment operator; Φ1 represents the feature matrix of the first batch; Φ2 represents the feature matrix of the second batch; Φ B Let B represent the feature matrix of the Bth batch. The dimension of each batch feature matrix is ​​13×13.

[0063] Step 2: Establish an inspection mechanism to determine whether the tail node forming quality is up to standard, specifically including steps 2.1 to 2.5 as shown below.

[0064] Step 2.1: According to the formula The batch feature matrices Φ1, Φ2, ..., Φ are respectively... B Expand into their respective row vectors, and denote them as φ1, φ2, ..., φ B Then, φ1, φ2, ..., φ B The data are combined into a B×169 dimensional data matrix, denoted as Y; where φ b Φ b The corresponding row vector, They respectively represent Φ b In the vectors, the first row vector, the second row vector, ..., the thirteenth row vector, φ1 represents the row vector corresponding to Φ1, φ2 represents the row vector corresponding to Φ2, and φ B Φ B The corresponding row vectors, the first row vector, the second row vector, up to the Bth row vector in Y, are equal to φ1, φ2, ..., φ1 respectively. B Numbered b = 1, 2, ..., B.

[0065] Step 2.2: Represent the first column vector, the second column vector, up to the 169th column vector in Y as y1, y2, ..., y 169Next, calculate the mean and standard deviation of all elements in each column vector of Y, and denot the mean of all elements in each column vector of Y as ζ1, ζ2, ..., ζ. 169 Let the standard deviations of all elements in each column vector of Y be denoted as θ1, θ2, ..., θ 169 Then according to the formula For y1, y2, ..., y 169 After standardization, we obtain a standardized column vector, which is denoted as follows: Among them, y j Let j represent the j-th column vector in Y. Indicates y j The standardized column vectors are ζ1 and θ1, which represent the mean and standard deviation of all elements in y1, respectively, and ζ2 and θ2, which represent the mean and standard deviation of all elements in y2, respectively. 169 and θ 169 Each corresponds to y 169 The mean and standard deviation of all elements in the ζ j and θ j These respectively represent the j-th column vector y in Y. j The mean and standard deviation of all elements in the matrix, column numbers j = 1, 2, ..., 169. This represents the column vector obtained after standardization of y1. This represents the column vector obtained after standardization of y2. Indicates y 169 The column vector obtained after standardization.

[0066] Step 2.3: [The text appears to be incomplete and contains several grammatical errors. A more accurate translation would require Combined into a B×169 dimensional data matrix, denoted as Again Perform singular value decomposition, that is: Where U and V represent two unitary matrices of the singular value decomposition, S represents a diagonal matrix composed of non-zero singular values, and the superscript "T" indicates the transpose of a matrix or vector.

[0067] Step 2.4: According to the formula Θ=UU T Calculate matrix Θ; then calculate the mean p and standard deviation q of the B elements on the diagonal of Θ; finally, set the tail node molding quality inspection threshold. It equals p + 3.5q.

[0068] Step 2.5: Retain the parameters used for the tail node forming quality inspection, specifically including the average values ​​ζ1, ζ2, ..., ζ from Step 2.2. 169 Sum of standard deviations θ1, θ2, ..., θ 169The unitary matrix V and diagonal matrix S in step 2.3, and the tail node forming quality inspection threshold in step 2.4.

[0069] Step 3: Instantly inspect the molding quality of the tail node using the latest sampling data of the automotive insert injection batch collected by the injection molding machine, specifically including steps 3.1 to 3.5 as shown below.

[0070] Step 3.1: After the latest automotive insert injection molding batch is completed, immediately acquire the corresponding sampling data and represent it as a batch data matrix, denoted as X. new ; where X new The data in columns 1 through 13 correspond to the 13 measurement variables in step 1.1, respectively.

[0071] Step 3.2: Following the procedures in steps B1 and C1, process X in the same way. new Standardization is performed to obtain a standardized batch data matrix, denoted as...

[0072] Step 3.3: Settings Then, following the process in step B2 (see steps B2.1 to B2.5 for details), iterate to obtain Z in the same way. new The corresponding left transformation matrix U new and right transformation matrix V new Then calculate matrix Φ new =U new Z new V new .

[0073] Step 3.4: According to the formula Φ new Expand into the corresponding row vector, denoted as y; then according to the formula Standardize each element in y individually to obtain the standardized row vector, denoted as . in, Corresponding to Φ new The first row vector, the second row vector, ..., the thirteenth row vector in the vector, y (j) This represents the j-th element in y. Let y represent the j-th element in y. (1) ,y (2) , ..., y (16 9) represent the 1st element, 2nd element, ..., 169th element in y, respectively. They represent The first element, the second element, ..., the 169th element in the array, with column numbers j = 1, 2, ..., 169; here, ζ1, ζ2, ..., ζ 169 and θ1,θ2,…,θ 169 These are the parameters retained in step 2.5.

[0074] Step 3.5: According to the formula Calculate test indicators Re-evaluation Is it greater than If not, the tail node molding quality is considered acceptable; if yes, the tail node molding quality is considered unacceptable. After inspection, return to step 3.1 to continue the tail node molding quality inspection for the latest automotive insert injection molding batch. Here, V, S, These are the parameters retained in step 2.5.

[0075] Figure 2 The paper shows the inspection results of the tail node molding quality of multiple automotive insert injection molding batches using the method of the present invention. The injection molding batches corresponding to the dots above the dotted line have unqualified tail node molding quality.

Claims

1. A method for inspecting the molding quality of the last node of an injection-molded batch of automotive inserts, characterized in that... Specifically, the following steps are included: Step 1: Preprocess the sampling data of the automotive insert injection molding batches with qualified tail node molding quality, specifically including steps 1.1 to 1.3 as shown below; Step 1.1: From the historical database corresponding to the injection molding machine, select the sampling data of B batches of automotive inserts with qualified tail-end molding quality, and represent them as B batch data matrices, denoted as X1, X2, ..., X... B ; Among them, the data matrix of the b-th batch is X b X b The data in columns 1 to 13 are generated from 13 measured variables of the injection molding machine in N. b The data collected at each sampling time point consists of 13 measured variables arranged in the following order: cylinder pressure, nozzle pressure, plasticizing pressure, mold cavity pressure, screw stroke, screw speed, barrel section 1 temperature, barrel section 2 temperature, barrel section 3 temperature, barrel section 4 temperature, oil temperature, nozzle temperature, and mold temperature. b The dimension is N b ×13, numbered b=1,2,…,B; Step 1.2: For X1, X2, ..., X B After standardization, a standardized batch data matrix is ​​obtained, which is denoted as follows: The specific implementation process includes steps A1 to D1 as shown below; Step A1: Set b = 1, then set X = X b ; Step B1: Represent the first, second, through thirteenth column vectors of X as x1, x2, ..., x 13 Next, calculate the mean and standard deviation of all elements in each column vector of X, and denot the mean of all elements in each column vector of X as μ1, μ2, ..., μ 13 Let the standard deviations of all elements in each column vector of X be denoted as δ1, δ2, ..., δ1, respectively. 13 Where μ1 and δ1 represent the mean and standard deviation of all elements in x1, respectively, and μ2 and δ2 represent the mean and standard deviation of all elements in x2, respectively. 13 and δ 13 They respectively represent x 13 The mean and standard deviation of all elements in the dataset; Step C1: Set m = 1, 2, ..., 13 sequentially, and simultaneously apply the formula... For x m Standardization is performed to obtain the standardized column vector. Then Combined into a single data matrix, denoted as Where, x m Let m be the m-th column vector in X. x represents m The column vector obtained after standardization, μ m x represents m The average value of all elements in the δ m x represents m The standard deviation of all elements in the sample; Step D1: Set the standardized batch b data matrix as follows make Next, check if b is less than B. If so, set b = b + 1, and then set X = X. b If not, return to step B1 to continue execution; otherwise, obtain the standardized batch data matrix. Step 1.3: For each After performing feature length equalization, B 13×13 dimensional batch feature matrices are obtained, denoted as Φ1, Φ2, ..., Φ3. B Specifically, it includes steps A2 to C2 as shown below; Step A2: Set b=1, then set... Step B2: Iterative calculation The corresponding left transformation matrix and right transformation matrix Thus the matrix The sum of squares of the elements on the middle diagonal is the largest; Step C2: Set Φ b =Φ; then check if b is less than B. If so, set b = b + 1, then set... Then return to step B2 to continue execution; otherwise, obtain B batch feature matrices Φ1,Φ2,…,Φ B The dimension of the feature matrix for each batch is 13×13. Step 2: Establish an inspection mechanism to determine whether the tail node forming quality is up to standard, specifically including steps 2.1 to 2.5 as shown below; Step 2.1: According to the formula The batch feature matrices Φ1, Φ2, ..., Φ are respectively... B Expand into their respective row vectors, and denote them as φ1, φ2, ..., φ B Then, φ1, φ2, ..., φ B The data are combined into a B×169 dimensional data matrix, denoted as Y; where φ b Φ b The corresponding row vector, They respectively represent Φ b The first, second, ..., thirteenth row vectors in Y, and the first, second, ..., Bth row vectors in Y are respectively equal to φ1, φ2, ..., φ B Numbered b = 1, 2, ..., B; Step 2.2: Represent the first column vector, the second column vector, up to the 169th column vector in Y as y1, y2, ..., y 169 Next, calculate the mean and standard deviation of all elements in each column vector of Y, and denot the mean of all elements in each column vector of Y as ζ1, ζ2, ..., ζ. 169 Let the standard deviations of all elements in each column vector of Y be denoted as θ1, θ2, ..., θ 169 Then according to the formula For y1, y2, ..., y 169 After standardization, we obtain a standardized column vector, which is denoted as follows: Among them, y j Let j represent the j-th column vector in Y. Indicates y j The standardized column vectors are ζ1 and θ1, which represent the mean and standard deviation of all elements in y1, respectively, and ζ2 and θ2, which represent the mean and standard deviation of all elements in y2, respectively. 169 and θ 169 Each corresponds to y 169 The mean and standard deviation of all elements in the ζ j and θ j These respectively represent the j-th column vector y in Y. j The mean and standard deviation of all elements in the column, with column numbers j = 1, 2, ..., 169; Step 2.3: [The text appears to be incomplete and contains several grammatical errors. A more accurate translation would require Combined into a B×169 dimensional data matrix, denoted as Again Perform singular value decomposition, that is: Where U and V represent two unitary matrices of the singular value decomposition, S represents a diagonal matrix composed of non-zero singular values, and the superscript "T" indicates the transpose of a matrix or vector; Step 2.4: According to the formula Θ=UU T Calculate matrix Θ; then calculate the mean p and standard deviation q of the B elements on the diagonal of Θ; finally, set the tail node molding quality inspection threshold. It equals p + 3.5q; Step 2.5: Retain the parameters used for the tail node forming quality inspection, specifically including the average values ​​ζ1, ζ2, ..., ζ from Step 2.

2. 169 Sum of standard deviations θ1, θ2, ..., θ 169 The unitary matrix V and diagonal matrix S in step 2.3, and the tail node forming quality inspection threshold in step 2.

4. Step 3: Instantly inspect the molding quality of the tail node using the latest sampling data of the automotive insert injection batch collected by the injection molding machine, specifically including steps 3.1 to 3.5 as shown below; Step 3.1: After the latest automotive insert injection molding batch is completed, immediately acquire the corresponding sampling data and represent it as a batch data matrix, denoted as X. new ; where X new The data in columns 1 through 13 correspond to the 13 measurement variables in step 1.1, respectively. Step 3.2: Following the procedures in steps B1 and C1, process X in the same way. new Standardization is performed to obtain a standardized batch data matrix, denoted as... Step 3.3: Settings Then, following the process in step B2, iteratively calculate Z in the same manner. new The corresponding left transformation matrix U new and right transformation matrix V new Then calculate matrix Φ new =U new Z new V new ; Step 3.4: According to the formula Φ new Expand into the corresponding row vector, denoted as y; then according to the formula Standardize each element in y individually to obtain the standardized row vector, denoted as . in, Corresponding to Φ new The first row vector, the second row vector, ..., the thirteenth row vector in the vector, y (j) This represents the j-th element in y. This represents the j-th element in y, with column numbers j = 1, 2, ..., 169; Step 3.5: According to the formula Calculate test indicators Re-evaluation Is it greater than If not, the tail node molding quality is considered qualified; if yes, the tail node molding quality is considered unqualified. After the inspection is completed, return to step 3.1 to continue to inspect the tail node molding quality of the latest automotive insert injection molding batch.

2. The method for inspecting the molding quality of the last node of an injection-molded automotive insert as described in claim 1, characterized in that... The specific implementation process of step B2 includes the following steps B2.1 to B2.5: Step B2.1: Randomly generate a 13×13 dimensional right transformation matrix from the interval (-1, 1). Step B2.2: Solve for the symmetric matrix The 13 eigenvalues ​​correspond to the eigenvectors u1, u2, ..., u 13 Then according to the formula For u1, u2, ..., u respectively 13 Normalization is performed; where the superscript "T" denotes the transpose of a matrix or vector, and m = 1, 2, ..., 13. The "=" sign is the assignment operator, and the u on the right side of "=" is... m express The eigenvector corresponding to the m-th eigenvalue, u on the left side of "=" m This represents the vector obtained after normalization. Step B2.3: Normalize u1, u2, ..., u 13 Combined into a left transformation matrix Solve the symmetric matrix again The 13 eigenvalues ​​correspond to the eigenvectors v1, v2, ..., v 13 ; Step B2.4: According to the formula For v1, v2, ..., v respectively 13 Perform normalization; then normalize the v1, v2, ..., v 13 Combined into a right transformation matrix in, The "=" sign in the equation is the assignment operator, and the value v on the right side of the "=" sign is... m express The eigenvector corresponding to the m-th eigenvalue, and v on the left side of "=" m This represents the vector obtained after normalization. Step B2.5: Determine If convergence has occurred, return to step B2.2 to continue execution; if convergence has occurred, utilize the obtained left transformation matrix. and right transformation matrix Calculate matrix 3. The method for inspecting the molding quality of the last node of an injection-molded automotive insert as described in claim 2, characterized in that... In step B2.5, The criteria for convergence are The elements in the middle no longer change.

Images