A method for evaluating and improving the quality of injection molded products

Abstract

This invention relates to a method for evaluating and improving the quality of injection-molded products, relating to the field of injection molding machine production. The method includes: collecting injection molding process parameters and target product quality; conducting preliminary single-factor experiments on the injection molding process parameters to screen out the main influencing factors; constructing a product regression equation based on the main influencing factors and the target product quality, and simultaneously calculating correlation parameters based on the main influencing factors and the target product quality; obtaining the correlation strength based on the correlation parameters; generating improvement parameters by combining the product regression equation and the correlation strength; and applying the improvement parameters and a preset response surface methodology to verify the adjustment effect, thereby stabilizing the product quality within a preset threshold range. This application has the effect of efficiently and stably controlling product quality within a preset threshold.

Description

Technical Field

[0001] This invention relates to the field of injection molding machine production, and in particular to a method for evaluating and improving the quality of injection molded products. Background Technology

[0002] Injection molding product quality evaluation and improvement is a full-process technical approach in the field of injection molding machine production. It involves systematically collecting data, screening key influencing factors, quantifying the correlation between parameters and quality, generating optimization solutions, and verifying them, all based on the pre-set quality standards for injection molding products.

[0003] Currently, manufacturers first determine the specifications of polymer injection molding raw materials based on injection molding requirements, then screen key process parameters, and collect parameter feedback data in real time through a host computer to dynamically adjust the values ​​of core parameters such as barrel temperature, material back pressure, and material rotation speed. To verify the impact of parameters on quality, a fourth-order covariance matrix is ​​introduced to analyze the correlation between parameters and between parameters and performance indicators. At the same time, the influence coefficients of each parameter are calculated by combining a regression model. After repeated testing and verification to determine the final coefficients, a response surface model is constructed to control the quality of injection molded products within a preset threshold range.

[0004] The above operations failed to screen the main influencing factors through standardized single-factor preliminary experiments, and did not combine the product regression equation and correlation strength to generate targeted improvement parameters. As a result, the parameter optimization lacked accuracy and it was difficult to efficiently and stably control the product quality within the preset threshold. This needs to be improved. Summary of the Invention

[0005] In order to efficiently and stably control product quality within a preset threshold, this invention provides a method for evaluating and improving the quality of injection molded products.

[0006] This invention provides a method for evaluating and improving the quality of injection molded products, employing the following technical solution:

[0007] A method for evaluating and improving the quality of injection molded products, comprising:

[0008] Collect injection molding process parameters and target product quality;

[0009] Single-factor preliminary experiments were conducted on injection molding process parameters to screen out the main influencing factors.

[0010] A product regression equation is constructed based on the main influencing factors and the product target quality, and a correlation parameter is calculated based on the main influencing factors and the product target quality.

[0011] The correlation strength is obtained based on the correlation parameter;

[0012] Combine product regression equations and correlation strengths to generate improvement parameters;

[0013] The adjustment effect is verified by applying improved parameters and preset response surface methodology to stabilize product quality within the preset threshold range.

[0014] By adopting the above technical solution, the core parameters that have a significant impact on product quality are first accurately screened through standardized single-factor preliminary experiments, avoiding interference from irrelevant parameters in the optimization process. Then, based on the main influencing factors, a product regression equation including main effects, interaction effects, and quadratic terms is constructed. At the same time, correlation parameters are calculated and the correlation strength is divided, which can comprehensively and quantitatively reflect the complex relationship between parameters and quality. Finally, the regression equation and correlation strength are combined to generate targeted improvement parameters, and the adjustment effect is verified by the response surface methodology. This can achieve precise optimization of parameters, effectively solving the problem of lack of precision in parameter optimization in existing technologies, and ultimately efficiently and stably controlling product quality within the preset threshold.

[0015] Optionally, specific screening methods for key influencing factors may also be included:

[0016] Collect historical production data;

[0017] Based on historical production data, the initial baseline value and range of variation are retrieved;

[0018] The experimental gradient is determined based on the initial baseline value and the range of variation.

[0019] Injection molding production is carried out based on experimental gradients and injection molding process parameters, and the product quality of the produced products is collected.

[0020] The range coefficient and coefficient of variation were obtained by combining the experimental gradient, injection molding process parameters, and product manufacturing quality.

[0021] When the range coefficient exceeds the preset range threshold or the coefficient of variation exceeds the preset coefficient of variation threshold, the parameter is marked as a major influencing factor.

[0022] Optionally, methods for constructing product regression equations are also included:

[0023] Based on the actual values ​​of the main influencing factors, the actual values ​​of the target product quality, and the actual values ​​of the production quality of the product during injection molding production.

[0024] Based on the actual values ​​of the factors, the actual values ​​of the targets, and the actual values ​​of production, the main effect coefficient, the interaction coefficient, the quadratic term coefficient, and the constant term of the equation are calculated.

[0025] The product regression equation is generated by combining the main effect coefficient, interaction coefficient, quadratic term coefficient, constant term of the equation, and preset error term.

[0026] Optionally, specific product regression equations may also be included:

[0027] Y = β0 + β1F1 + β2F2 + β3F3 + β 12 F1F2+β 13 F1F3+β 23 F2F3+β 11 F1 2 +β 22 F2 2 +β 33 F3 2 +ϵ;

[0028] Where Y represents the target product quality, β0 is a constant term, and β1, β2, and β3 are the main effect coefficients of the first influencing factor F1, the second influencing factor F2, and the third influencing factor F3 on Y, respectively. 12 β 13 β 23 The interaction coefficients β and F1 are the interaction coefficients between F1 and F2, F1 and F3, and F2 and F3, respectively. 11 β 22 β 33 ϵ represents the coefficients of the quadratic terms F1, F2, and F3, respectively, and ϵ is the error term.

[0029] Optionally, the specific calculation method for the relevance parameter may also be included:

[0030] A fourth-order covariance matrix was constructed based on the main influencing factors and the target quality of the product.

[0031] The sample variance and sample covariance are calculated by combining the actual values ​​of the factors, the actual values ​​of the target, the actual values ​​of production, and the fourth-order covariance matrix.

[0032] The correlation coefficient matrix is ​​obtained based on the sample variance and sample covariance;

[0033] The correlation coefficient matrix is ​​used to obtain the correlation parameter.

[0034] Optionally, it also includes the specific construction of the fourth-order covariance matrix and the refined calculation methods for sample variance and sample covariance:

[0035] The expression for the fourth-order covariance matrix is:

[0036] ;

[0037] Among them, Fi is the main influencing factor, F1 is the first main influencing factor, F2 is the second main influencing factor, and F3 is the third main influencing factor.

[0038] The formula for calculating the sample variance is: ;

[0039] Where n is the number of samples, The mean of the observed sample of the main influencing factor Fi. The sample variances on the diagonal of the fourth-order covariance matrix;

[0040] The sample covariance includes the sample covariance between factors and the sample covariance between factors and quality. The formula for calculating the sample covariance between factors is: ;

[0041] The formula for calculating the sample covariance between factors and quality is: ;

[0042] in, Let yk be the mean of the observed samples of the main influencing factor Fj, and yk be the k-th value of the product target quality. This represents the mean of the observed samples for the product's target quality.

[0043] Optionally, a method for refining the correlation coefficient matrix may also be included:

[0044] ;

[0045] ;

[0046] in, The correlation coefficient between factor i and factor j Let be the correlation coefficient between factor i and index Y. Let Fi be the sample standard deviation. Let Fj be the sample standard deviation, and SY be the sample standard deviation of Y;

[0047] Combination and To generate the correlation coefficient matrix expression;

[0048] , where R is the expression for the correlation coefficient matrix.

[0049] In summary, this application includes at least one of the following beneficial technical effects:

[0050] 1. First, standardized single-factor preliminary experiments are used to accurately screen out the core parameters that have a significant impact on product quality, avoiding interference from irrelevant parameters in the optimization process. Then, based on the main influencing factors, a product regression equation including main effects, interaction effects, and quadratic terms is constructed. At the same time, correlation parameters are calculated and the correlation strength is divided, which can comprehensively and quantitatively reflect the complex relationship between parameters and quality. Finally, the regression equation and correlation strength are combined to generate targeted improvement parameters. The adjustment effect is verified by using response surface methodology, which can achieve precise optimization of parameters and effectively solve the problem of lack of precision in parameter optimization in existing technologies. Ultimately, product quality is controlled within the preset threshold efficiently and stably.

[0051] 2. By using a regression model and solving for the regression coefficients, the main effect coefficients are obtained, thereby quantifying the influence of the parameters and reducing the product loss rate;

[0052] 3. Determining the initial baseline value and variation range based on historical production data ensures that the experimental gradient division closely matches the actual production scenario. By conducting injection molding production according to the experimental gradient and collecting product production quality data, the impact data of different process parameter values ​​on quality can be systematically obtained. Then, the influence of parameters on quality is quantified by the range coefficient and the coefficient of variation. When either of them exceeds the preset threshold, it is marked as a major influencing factor. This avoids subjective bias caused by human experience judgment and accurately identifies the core parameters that have a significant impact on quality. Attached Figure Description

[0053] Figure 1 This is a flowchart illustrating a method for evaluating and improving the quality of injection molded products.

[0054] Figure 2 This is a flowchart illustrating the specific screening methods for the main influencing factors;

[0055] Figure 3 This is a flowchart illustrating the method for constructing the product regression equation;

[0056] Figure 4 A flowchart illustrating the specific calculation method for the relevance parameter. Detailed Implementation

[0057] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments.

[0058] Reference Figure 1 This application discloses a method for evaluating and improving the quality of injection molded products, including the following steps:

[0059] S1: Collect injection molding process parameters and target product quality.

[0060] Injection molding process parameters refer to the key, controllable technical parameters during the injection molding production process. Product target quality refers to the preset quality standards that the injection-molded products must meet. Both injection molding process parameters and product target quality are pre-input by on-site technical personnel and will not be elaborated upon here.

[0061] S2: Conduct preliminary single-factor experiments on injection molding process parameters to screen out the main influencing factors.

[0062] The main influencing factors refer to the injection molding process parameters that significantly affect the target quality of the product. Preliminary single-factor experiments can screen out the main influencing factors among the injection molding process parameters. Specific operating procedures are detailed in subsequent sections S20 to S25 and will not be repeated here.

[0063] S3: Construct a product regression equation based on the main influencing factors and the product target quality, and calculate the correlation parameter based on the main influencing factors and the product target quality.

[0064] The product regression equation is a mathematical model that quantifies the relationship between key influencing factors and the target quality of the product. The product regression equation is constructed by combining key influencing factors and the target quality of the product. The specific construction method will be explained in detail in sections S30 to S301, and will not be elaborated upon here.

[0065] The correlation parameter is a quantitative indicator used to measure the degree of linear correlation between major influencing factors and between major influencing factors and product target quality. The correlation parameter can be calculated by understanding the major influencing factors and product target quality. The specific calculation steps for the correlation parameter will be explained in detail in subsequent sections S31 to S312, and will not be repeated here.

[0066] S4: Obtain the correlation strength based on the correlation parameter.

[0067] Association strength refers to a graded description of the degree of association between major influencing factors or between major influencing factors and quality, based on the correlation parameter.

[0068] First, extract specific values ​​from the relevance parameters and classify the association strength level according to the absolute value of the value: an absolute value close to 1 (e.g., greater than 0.8) is "strong association", between 0.5 and 0.8 is "moderate association", between 0.3 and 0.5 is "weak association", and close to 0 (less than 0.3) is "no association". Then, determine the association direction according to the sign of the value: positive numbers indicate "positive association" (if one factor increases, the other factor or quality also tends to increase), and negative numbers indicate "negative association" (if one factor increases, the other factor or quality tends to decrease). Integrate the level and direction to obtain the association strength.

[0069] S5: Combine product regression equations and association strength to generate improvement parameters.

[0070] Improvement parameters refer to the specific values ​​of the main influencing factors after adjustment.

[0071] First, extract the main effect coefficients and interaction coefficients of each factor from the product regression equation. Sort the factors by absolute value to determine which factors have a more significant impact on quality (the larger the absolute value, the higher the priority). Then, based on the correlation strength results, prioritize adjusting factors with strong correlations and high priority. For example, if a factor is positively correlated with quality (e.g., the larger the factor, the better the quality), and the current quality is substandard, increase the value of that factor according to priority; if it is negatively correlated (e.g., the larger the factor, the worse the quality), decrease the value of that factor.

[0072] Finally, the machine automatically corrects the parameters (such as PID control or intelligent algorithms) based on the real-time production quality data, avoiding excessive adjustment (in this embodiment, it is no more than ±20% of the initial value). The final value obtained is the improved parameter.

[0073] S6: Apply improved parameters and preset response surface methodology to verify the adjustment effect, so that product quality is stabilized within the preset threshold range.

[0074] The response surface methodology is an experimental verification method that uses a three-dimensional response surface and contour plot of "key influencing factors - product quality" to visually demonstrate the impact of factor interactions on quality and verify whether improving parameters can stabilize quality within the target range.

[0075] The threshold range refers to the allowable fluctuation range of the product's target quality.

[0076] The response surface methodology and threshold range are both preset by those skilled in the art and will not be elaborated upon here.

[0077] The improved parameters were applied to the injection molding machine for actual production, and then the response surface methodology was used for verification to determine whether the quality of the products produced under the improved parameters fell within the threshold range.

[0078] If it falls within the range, it indicates that the product quality is qualified under the improved parameters.

[0079] Reference Figure 2 It also includes specific screening methods for the main influencing factors:

[0080] S20: Collect historical production data.

[0081] Historical production data refers to the correlation data between the values ​​of injection molding process parameters recorded in past injection molding production processes and the corresponding product production quality data. Historical production data is obtained by retrieving a pre-set production record database. This database contains data from each injection molding production process. The production record database is pre-set by those skilled in the art and will not be elaborated upon here.

[0082] S21: Retrieve initial baseline values ​​and ranges of variation based on historical production data.

[0083] The initial baseline value refers to the default value of the injection molding process parameters that ensures the product quality is basically acceptable. The variation range refers to the safe adjustment range of the injection molding process parameters. The initial baseline value and variation range can be obtained by reviewing historical production data. Historical production data contains both the initial baseline value and the variation range.

[0084] S22: Divide the test gradient based on the initial baseline value and the range of variation.

[0085] An experimental gradient refers to the gradient used in a single-factor experiment to test the influence of different parameters on quality one by one. It is obtained by dividing the parameter test points at equal intervals within the variation range, centered on the initial baseline value. The specific intervals are predetermined by those skilled in the art and will not be elaborated upon here.

[0086] S23: Injection molding production is carried out based on experimental gradients and injection molding process parameters, and the product quality of the produced products is collected.

[0087] Product manufacturing quality refers to the actual quality test value of the products produced under a specific process parameter gradient in a single-factor experiment. Product manufacturing quality is obtained through quality inspection of the produced products by on-site technical personnel. The methods for quality inspection are selected by the on-site technical personnel themselves and will not be elaborated here.

[0088] Injection molding production is carried out using the obtained experimental gradient and injection molding process parameters, and the product quality of the produced products is collected for subsequent steps.

[0089] S24: Combine experimental gradient, injection molding process parameters, and product manufacturing quality to obtain the range coefficient and coefficient of variation.

[0090] The range coefficient refers to the difference between the maximum and minimum values ​​of product production quality under different gradients of a certain parameter in a single-factor experiment.

[0091] The coefficient of variation refers to the ratio of the standard deviation to the mean of a product's production quality under different gradients of a certain parameter in a single-factor experiment.

[0092] First, record the product production quality data corresponding to each injection molding process parameter under all experimental gradients. Then, sort these data by numerical value, mark the maximum value, minimum value, and all original data, and finally calculate the range coefficient and coefficient of variation. Since the calculation formulas for the range coefficient and coefficient of variation are common knowledge in this field, they will not be elaborated here.

[0093] S25: When the range coefficient exceeds the preset range threshold or the coefficient of variation exceeds the preset coefficient of variation threshold, mark this parameter as the main influencing factor.

[0094] The range threshold is a critical value used to determine whether a parameter is a major influencing factor. When the range coefficient of a parameter exceeds this value, it indicates that its impact on product quality is significant and it should be marked as a major influencing factor.

[0095] The coefficient of variation threshold is a critical value used to determine whether a parameter is a major influencing factor. When the coefficient of variation of a parameter exceeds this value, it indicates that its impact on product quality stability is significant and it should be marked as a major influencing factor.

[0096] The range threshold and the coefficient of variation threshold are both set in advance by those skilled in the art, and will not be elaborated here.

[0097] Reference Figure 3 It also includes methods for constructing product regression equations:

[0098] S30: Based on the actual values ​​of the main influencing factors, the actual values ​​of the target quality of the product, and the actual values ​​of the production quality of the product during injection molding production.

[0099] The actual value of a factor refers to the actual measured value of the main influencing factor. The actual value of a factor is obtained in real time through the parameter monitoring system (such as pressure sensor and temperature sensor) built into the injection molding machine during the injection molding production process.

[0100] The actual target value refers to the preset standard value of the product's target quality. The actual target value is obtained through pre-input by those skilled in the art, and will not be elaborated here.

[0101] Actual production value refers to the actual test value of the product's production quality under the corresponding "actual factor value". Actual production value is obtained by testing 3 to 5 randomly selected samples of each batch of products according to preset testing standards (such as weighing with an electronic scale with an accuracy of 0.01g and measuring dimensions with a micrometer) in a constant temperature and humidity environment, and taking the average value as the actual production value of that batch.

[0102] During injection molding production, it is necessary to first collect the actual values ​​of the main influencing factors, the actual values ​​of the target product quality, and the actual values ​​of the production quality of the product, in order to facilitate subsequent steps.

[0103] S300: Based on the actual values ​​of the factors, the actual values ​​of the target, and the actual values ​​of production, calculate the main effect coefficient, interaction coefficient, quadratic term coefficient, and constant term of the equation.

[0104] The main effect coefficient is a quantitative value of the degree of independent influence of a single main influencing factor on the target quality of a product.

[0105] The interaction coefficient is a quantitative value of the additional impact on the target quality of a product when two main influencing factors work together.

[0106] The quadratic coefficient refers to the quantitative value of the nonlinear impact of the square term of a single major influencing factor on the target quality of the product.

[0107] The constant term in the equation refers to the theoretical basis value of the product target quality when all major influencing factors are zero.

[0108] The actual values ​​of the collected factors, the actual values ​​of the targets, and the actual values ​​of production are organized into a structured dataset. Outliers (such as data that deviate from the mean by 3 times the standard deviation) are removed, and the sample size is ensured to be no less than 30 groups (to meet the statistical requirements of regression analysis).

[0109] First, a simplified regression model is constructed (for solving the main effect coefficients), as follows:

[0110] ;

[0111] in, For the target product quality, β0 is the constant term in the equation, and β1, β2, and β3 are the main effect coefficients. This is the error term of the simplified regression model.

[0112] Secondly, the covariance matrix of the main influencing factors is extracted from the fourth-order covariance matrix, as follows:

[0113] ;

[0114] in, The covariance matrix of the main influencing factors. , , For sample variance, , , The sample covariance is between factors.

[0115] Simultaneously, construct the covariance vector between the main influencing factors and product quality:

[0116] ;

[0117] in, This represents the covariance vector between the main influencing factors and product quality. , , The sample covariance is between factors and quality.

[0118] Finally, the main effect coefficients are solved by matrix inversion, using the following formula:

[0119] ;

[0120] in, β1, β2, and β3 are the inverse of the covariance matrix, and β1, β2, and β3 are the main effect coefficients.

[0121] The interaction coefficients and quadratic coefficients can be solved simultaneously by those skilled in the art by expanding the independent variable matrix and using the least squares method, following the logic described above. The constant term of the equation is derived by substituting the mean into the regression equation, based on the known coefficients. Substituting the mean into the regression equation is common knowledge in this field and will not be elaborated upon here.

[0122] S301: Combine the main effect coefficient, interaction coefficient, quadratic coefficient, constant term of the equation, and preset error term to generate the product regression equation.

[0123] The error term is a random variable in the regression equation used to characterize the "quality fluctuations not explained by the main influencing factors". The error term is predetermined by those skilled in the art and will not be elaborated here.

[0124] The product regression equation can be obtained by combining the main effect coefficient, interaction coefficient, quadratic term coefficient, constant term, and error term, as follows:

[0125] Y = β0 + β1F1 + β2F2 + β3F3 + β 12 F1F2+β 13 F1F3+β 23 F2F3+β 11 F1 2 +β 22 F2 2 +β 33 F3 2 +ϵ.

[0126] Where Y represents the target product quality, β0 is a constant term, and β1, β2, and β3 are the main effect coefficients of the first influencing factor F1, the second influencing factor F2, and the third influencing factor F3 on Y, respectively. 12 β 13 β 23 The interaction coefficients β and F1 are the interaction coefficients between F1 and F2, F1 and F3, and F2 and F3, respectively. 11 β 22 β 33 ϵ represents the coefficients of the quadratic terms F1, F2, and F3, respectively, and ϵ is the error term.

[0127] In this embodiment, F1 is the back pressure of the storage material, F2 is the injection pressure, and F3 is the barrel temperature.

[0128] Reference Figure 4 It also includes the specific calculation method for the relevance parameter:

[0129] S31: Construct a fourth-order covariance matrix based on key influencing factors and product target quality.

[0130] The fourth-order covariance matrix is ​​a 4×4 matrix constructed using three main influencing factors F1, F2, and F3, plus a product target quality as its dimensions.

[0131] The fourth-order covariance matrix can be obtained by combining the main influencing factors and the product's target quality. The expression is as follows:

[0132] ;

[0133] Among them, F i F1 is the primary influencing factor, F2 is the primary influencing factor one, F3 is the primary influencing factor three, and Y is the target quality of the product.

[0134] S310: Combine the actual values ​​of factors, the actual values ​​of targets, the actual values ​​of production, and the fourth-order covariance matrix to calculate the sample variance and sample covariance.

[0135] Sample variance refers to the elements on the diagonal of the fourth-order covariance matrix. It is used to quantify the dispersion of the values ​​of a single variable.

[0136] The sample variance is calculated using the following formula: .

[0137] Where n is the number of samples, The main influencing factor F i The mean of the observed sample, The variance of the samples is the diagonal of the fourth-order covariance matrix.

[0138] Sample covariance refers to the off-diagonal elements of a fourth-order covariance matrix. It is used to quantify the degree of covariance between two variables.

[0139] Sample covariance includes the sample covariance between factors and the sample covariance between factors and quality. It can be calculated using the following formula:

[0140] The formula for calculating the sample covariance among factors is as follows: .

[0141] The formula for calculating the sample covariance between factors and quality is: .

[0142] in, The main influencing factor F j The mean of the observed sample, y k For the k-th value of the product's target quality, This represents the mean of the observed samples for the product's target quality. For the sample covariance among factors, The sample covariance is between factors and quality.

[0143] S311: Obtain the correlation coefficient matrix based on the sample variance and sample covariance.

[0144] The correlation coefficient matrix is ​​a 4×4 matrix formed by converting the "covariance / variance" in the fourth-order covariance matrix into "correlation coefficient".

[0145] The correlation coefficient is calculated using the following formula:

[0146] ;

[0147] ;

[0148] in, The correlation coefficient between factor i and factor j Let be the correlation coefficient between factor i and index Y. For F i The sample standard deviation For F j The sample standard deviation, S Y Let Y be the sample standard deviation;

[0149] Combination and To generate the correlation coefficient matrix expression;

[0150] , where R is the expression for the correlation coefficient matrix.

[0151] S312: Obtain the correlation parameter based on the correlation coefficient matrix.

[0152] The correlation parameter is all the off-diagonal elements in the correlation coefficient matrix. The correlation parameter can be found using the obtained correlation coefficient matrix.

[0153] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional modules is used as an example. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, device, and unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0154] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A method for evaluating and improving the quality of injection-molded articles, characterized by, The application relates to an injection molding process parameter and product target quality acquisition method, an injection molding process parameter single-factor preliminary test method for screening main influencing factors, a main influencing factor and product target quality-based product regression equation construction method and a correlation degree parameter calculation method, a correlation strength acquisition method based on the correlation degree parameter, a product regression equation and correlation strength-based improvement parameter generation method, a preset response surface method and improvement parameter application verification adjustment effect verification method for stabilizing product quality in a preset threshold range, a product regression equation construction method, a main influencing factor, product target quality and product production quality-based main effect coefficient, interaction coefficient, quadratic term coefficient and equation constant item calculation method, a product regression equation generation method based on the main effect coefficient, interaction coefficient, quadratic term coefficient, equation constant item and preset error term, a correlation degree parameter specific calculation method, a four-order covariance matrix construction method based on the main influencing factors and product target quality, a sample variance and sample covariance calculation method based on the factor actual value, target actual value, production actual value and four-order covariance matrix, a correlation coefficient matrix acquisition method based on the sample variance and sample covariance, and a correlation degree parameter acquisition method based on the correlation coefficient matrix, a main influencing factor specific screening method, historical production data acquisition, an initial benchmark value and variation interval acquisition method based on the historical production data, a test gradient division method based on the initial benchmark value and variation interval, an injection molding production method based on the test gradient and injection molding process parameters, a product production quality acquisition method, a range coefficient and variation coefficient acquisition method based on the test gradient, injection molding process parameters and product production quality, a main influencing factor marking method based on the range coefficient exceeding a preset range threshold or the variation coefficient exceeding a preset variation coefficient threshold, a specific product regression equation, a four-order covariance matrix specific construction method and sample variance and sample covariance refinement calculation method, and a correlation coefficient matrix refinement calculation method. ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 2. The method for evaluating and improving the quality of an injection-molded article according to claim 1, characterized in that, ​ ​ ​ ​ ​ ​ ​ 3. The method for evaluating and improving the quality of an injection-molded article according to claim 1, characterized in that, ​ Y = β0+ β1F1+ β2F2+ β3F3+ β 12 F1F2+ β 13 F1F3+ β 23 F2F3+ β 11 F1 2 + β 22 F2 2 + β 33 F3 2 + ε; Wherein, Y is the target quality of the product, β0 is a constant term, β1, β2, β3 are main effect coefficients of main influencing factor one F1, main influencing factor two F2, and main influencing factor three F3 on Y, respectively, β 12 , β 13 , β 23 are interaction coefficients of F1 and F2, F1 and F3, and F2 and F3, respectively, β 11 , β 22 , β 33 are quadratic coefficients of F1, F2, and F3, respectively, and ϵ is an error term.

4. The method for evaluating and improving the quality of an injection-molded article according to claim 1, characterized in that, ​ ​ ； Wherein, F i is the main influencing factor, F1 is the main influencing factor one, F2 is the main influencing factor two, and F3 is the main influencing factor three. The formula for calculating the sample variance is: ; where n is the number of samples, is the main influencing factor F i is the mean of the observed samples, is the sample variance on the diagonal of the fourth-order covariance matrix; The sample covariance includes a factor-to-factor sample covariance and a factor-to-quality sample covariance, and the factor-to-factor sample covariance is calculated according to the following formula: ; The sample covariance between the factors and the quality is calculated by the formula: ; wherein is the main influencing factor F j is the mean of the observed samples of y k is the kth value of the target quality of the product, is the mean of the observed samples of the target quality of the product.

5. The method for evaluating and improving the quality of an injection-molded article according to claim 4, characterized in that, ​ ； ； wherein, is the correlation coefficient of factor i and factor j, is the correlation coefficient of factor i and index Y, is F i the sample standard deviation, is F j the sample standard deviation, S Y is the sample standard deviation of Y; combining and to generate a correlation coefficient matrix expression; where R is the correlation coefficient matrix expression.

Images