A soft failure prediction method based on ICT and FCT

Abstract

A soft failure prediction method based on ICT and FCT includes the following steps: step S1, selecting key test items from the ICT and FCT test items of a product; collecting the ICT and FCT test data of products that have been rejected by customers due to problems related to the key test items, thereby forming an abnormal data set; step S2, obtaining the ICT and FCT test data of products of the same batch that have not been rejected by customers according to the production time of the products rejected by customers, thereby forming a normal data set; step S3, based on the abnormal data set and the normal data set, constructing a soft failure prediction model for predicting whether a product has a soft failure according to the ICT and FCT test data of the product; step S4, analyzing the ICT and FCT test data of the product in real time according to the soft failure prediction model, thereby predicting whether the product has a soft failure. The soft failure prediction method of the present invention is novel in design and highly practical.

Description

Technical Field

[0001] The present invention relates to the field of industrial big data mining, and in particular to a soft failure prediction method based on ICT and FCT. Background Art

[0002] In the current electronic product manufacturing process, the ICT (in-circuit test) / FCT (functional circuit test) test results are determined based on the electrical performance, functions and electronic device specifications that must be achieved during the product design phase. The test results do not take into account the deviation of the test results within the acceptable range caused by external forces, which can lead to reduced product reliability, easily causing product rejection, increased repair costs and loss of reputation. Summary of the invention

[0003] In view of the above problems, the present invention proposes a soft failure prediction method based on ICT and FCT.

[0004] The technical solution proposed by the present invention is as follows:

[0005] The present invention proposes a soft failure prediction method based on ICT and FCT, comprising the following steps:

[0006] Step S1, selecting key test items from the ICT and FCT test items of the product; collecting ICT and FCT test data of products rejected by customers due to problems related to the key test items, thereby forming an abnormal data set;

[0007] Step S2: according to the production time of the product rejected by the customer, obtain the ICT and FCT test data of the products of the same batch but not rejected by the customer, so as to form a normal data set;

[0008] Step S3: Based on the abnormal data set and the normal data set, a soft failure prediction model is constructed for predicting whether a product has a soft failure according to the ICT and FCT test data of the product;

[0009] Step S4: Analyze the ICT and FCT test data of the product in real time according to the soft failure prediction model, so as to predict whether the product has soft failure.

[0010] In the above-mentioned soft failure prediction method of the present invention, step S3 is performed by adopting a machine learning algorithm.

[0011] In the above-mentioned soft failure prediction method of the present invention, step S3 comprises the following steps:

[0012] Step S3.1: normalize all ICT and FCT test data in the union of the normal data set and the abnormal data set. The normalization algorithm is:

[0013] kij =|(T ij -μ j ) / σ j |;

[0014] Among them, T ij It represents the jth ICT and FCT test data of the i-th product in the union of the normal data set and the abnormal data set;

[0015] μ j represents the mean of the j-th ICT and FCT test data of all products in the normal data set;

[0016] σ j represents the standard deviation of the j-th ICT and FCT test data of all products in the normal data set;

[0017] The total number of products in the union of the normal data set and the abnormal data set is recorded as n, and the total number of ICT and FCT test data items of the products is recorded as m. The 1st to mth ICT and FCT test data after normalization of the first product, the 1st to mth ICT and FCT test data after normalization of the second product, ..., the 1st to mth ICT and FCT test data after normalization of the nth product are used to construct A. nm ;in,

[0018]

[0019] Step S3.2, respectively calculate the K values ​​of all products in the union of the normal data set and the abnormal data set; the K value of the i-th product in the union of the normal data set and the abnormal data set is recorded as K i , where K i =max(k i1 ,k i2 ,…,k im ); With K value as the horizontal coordinate, and with the proportion of the number of returned products corresponding to K value in the total number of returned products and the proportion of the number of non-returned products corresponding to K value in the total number of non-returned products as the vertical coordinate, a coordinate system is established and corresponding curves are drawn, thereby obtaining a curve of the proportion of returned products-K value and a curve of the proportion of non-returned products-K value; obtain the K value at the position with the largest spacing between the curve of the proportion of returned products-K value and the curve of the proportion of non-returned products-K value in the vertical coordinate direction, recorded as k';

[0020] Step S3.3: define the products in the abnormal data set whose K value is greater than the absolute value of k' as soft failure products; extract the ICT and FCT test data of soft failure products to form a soft failure product data set; construct B n(m+1) ,in,

[0021]

[0022] Here, y1 represents the value of whether the first product in the union of the normal data set and the abnormal data set is a soft failure product; y2 represents the value of whether the second product in the union of the normal data set and the abnormal data set is a soft failure product; ...; y n Indicates whether the nth product in the union of the normal data set and the abnormal data set is a soft failure product;

[0023] When y1, y2, …, y n When any one of y1, y2, ..., y n When any one of them is 1, it means that the corresponding product is a soft failure product;

[0024] Step S3.4: n(m+1) The first m columns are used as independent variables, and the m+1th column is used as the dependent variable. The machine learning classification algorithm is used for analysis to construct a soft failure prediction model.

[0025] In the above-mentioned soft failure prediction method of the present invention, step S4 also includes: normalizing the real-time ICT and FCT test data of the product, and then inputting the normalized data into the soft failure prediction model as independent variables to obtain and output the prediction results.

[0026] The present invention constructs a soft failure prediction method based on ICT and FCT, and uses ICT / FCT test results combined with machine learning algorithms to achieve prejudgment of soft failures, reduce the impact of deviations within an acceptable range of test results caused by external forces, and reduce the customer return rate. The soft failure prediction method based on ICT and FCT of the present invention is novel in design and highly practical. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The present invention will be further described below with reference to the accompanying drawings and embodiments, in which:

[0028] Figure 1 A flowchart of a soft failure prediction method based on ICT and FCT according to a preferred embodiment of the present invention is shown;

[0029] Figure 2 A curve graph showing the proportion of products returned by customers versus k value and a curve graph showing the proportion of products not returned by customers versus k value. DETAILED DESCRIPTION

[0030] The technical problem to be solved by the present invention is: in the current electronic product manufacturing process, the test results of ICT (online test) / FCT (functional circuit test) do not take into account the deviation within the acceptable range of the test results caused by external forces, and these deviations will lead to reduced product reliability, easily leading to product rejection, increased rework costs and loss of reputation. The technical idea proposed by the present invention to solve this technical problem is: construct a soft failure prediction method based on ICT and FCT, use ICT / FCT test results combined with machine learning algorithms to achieve pre-judgment of soft failures, reduce the impact caused by the deviation within the acceptable range of test results caused by external forces, and reduce the customer rejection rate.

[0031] In order to make the technical purpose, technical scheme and technical effect of the present invention clearer and to facilitate those skilled in the art to understand and implement the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0032] like Figure 1 As shown, Figure 1 The flowchart of the soft failure prediction method based on ICT and FCT of the preferred embodiment of the present invention is shown. The present invention proposes a soft failure prediction method based on ICT and FCT, comprising the following steps:

[0033] Step S1, selecting key test items from the ICT and FCT test items of the product; collecting ICT and FCT test data of products rejected by customers due to problems related to the key test items, thereby forming an abnormal data set;

[0034] Step S2: according to the production time of the product rejected by the customer, obtain the ICT and FCT test data of the products of the same batch but not rejected by the customer, so as to form a normal data set;

[0035] Here, the same batch of products refers to products from the same batch of materials, the same production line, and the same shift.

[0036] Step S3: Based on the abnormal data set and the normal data set, a soft failure prediction model is constructed for predicting whether a product has a soft failure according to the ICT and FCT test data of the product;

[0037] Here, this step is performed by employing a machine learning algorithm.

[0038] Step S3 includes the following steps:

[0039] Step S3.1: normalize all ICT and FCT test data in the union of the normal data set and the abnormal data set. The normalization algorithm is:

[0040] k ij =|(T ij-μ j ) / σ j |;

[0041] Among them, T ij It represents the jth ICT and FCT test data of the i-th product in the union of the normal data set and the abnormal data set;

[0042] μ j represents the mean of the j-th ICT and FCT test data of all products in the normal data set;

[0043] σ j represents the standard deviation of the j-th ICT and FCT test data of all products in the normal data set;

[0044] The total number of products in the union of the normal data set and the abnormal data set is recorded as n, and the total number of ICT and FCT test data items of the products is recorded as m. The 1st to mth ICT and FCT test data after normalization of the first product, the 1st to mth ICT and FCT test data after normalization of the second product, ..., the 1st to mth ICT and FCT test data after normalization of the nth product are used to construct A. nm ;in,

[0045]

[0046] Step S3.2, respectively calculate the K values ​​of all products in the union of the normal data set and the abnormal data set; the K value of the i-th product in the union of the normal data set and the abnormal data set is recorded as K i , where K i =max(k i1 ,k i2 ,…,k im ); take K value as the horizontal coordinate, take the proportion of the number of returned products corresponding to K value in the total number of returned products and the proportion of the number of non-returned products corresponding to K value in the total number of non-returned products as the vertical coordinate to establish a coordinate system and draw corresponding curves, so as to obtain the proportion of returned products-K value curve and the proportion of non-returned products-K value curve; obtain the K value at the position with the largest spacing between the proportion of returned products-K value curve and the proportion of non-returned products-K value curve in the vertical coordinate direction, recorded as k', such as Figure 2 As shown;

[0047] Step S3.3: define the products in the abnormal data set whose K value is greater than the absolute value of k' as soft failure products; extract the ICT and FCT test data of soft failure products to form a soft failure product data set; construct B n(m+1) ,in,

[0048]

[0049] Here, y1 represents the value of whether the first product in the union of the normal data set and the abnormal data set is a soft failure product; y2 represents the value of whether the second product in the union of the normal data set and the abnormal data set is a soft failure product; ...; y n Indicates whether the nth product in the union of the normal data set and the abnormal data set is a soft failure product;

[0050] When y1, y2, …, y n When any one of y1, y2, ..., y n When any one of them is 1, it means that the corresponding product is a soft failure product;

[0051] Step S3.4: n(m+1) The first m columns are used as independent variables, and the m+1th column is used as the dependent variable. The machine learning classification algorithm is used for analysis to construct a soft failure prediction model.

[0052] Step S4: Analyze the ICT and FCT test data of the product in real time according to the soft failure prediction model, so as to predict whether the product has soft failure.

[0053] Step S4 also includes: normalizing the real-time ICT and FCT test data of the product, and then inputting the normalized data into the soft failure prediction model as independent variables to obtain and output the prediction results.

[0054] In the above technical solution, soft failure means that the product has external force damage (such as electrical damage, vibration damage, etc.) before shipment, but it does not affect its electrical performance and function. It can pass ICT and FCT tests and can be shipped normally.

[0055] Products with soft failures will expose electrical performance and functional defects and maintenance diagrams in advance during use after shipment, leading to customer returns.

[0056] Products with soft failures may have offsets or mutations in some ICT and FCT test data.

[0057] In the above technical solution, the key test items of the product are mainly the ICT and FCT test items of semiconductor devices, mainly including analog quantity tests of semiconductor devices, such as voltage test, current test, capacitive reactance test, etc.

[0058] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all these improvements and changes should fall within the scope of protection of the appended claims of the present invention.

Claims

1. A soft failure prediction method based on ICT and FCT, characterized in that: The following steps are involved: Step S1, selecting key test items from the ICT and FCT test items of the product; Collect ICT and FCT test data of products rejected by customers due to issues related to the test items, thus forming an abnormal data set; Step S2: according to the production time of the product rejected by the customer, obtain the ICT and FCT test data of the products of the same batch but not rejected by the customer, so as to form a normal data set; Step S3: Based on the abnormal data set and the normal data set, a soft failure prediction model is constructed for predicting whether a product has a soft failure according to the ICT and FCT test data of the product; Step S4: Analyze the ICT and FCT test data of the product in real time according to the soft failure prediction model, so as to predict whether the product has soft failure; Step S3 includes the following steps: Step S3.1: normalize all ICT and FCT test data in the union of the normal data set and the abnormal data set. The normalization algorithm is: k ij =|(T ij -m j ) / s j |; Among them, T ij It represents the jth ICT and FCT test data of the i-th product in the union of the normal data set and the abnormal data set; μ j represents the mean of the j-th ICT and FCT test data of all products in the normal data set; σ j represents the standard deviation of the j-th ICT and FCT test data of all products in the normal data set; The total number of products in the union of the normal data set and the abnormal data set is recorded as n, and the total number of ICT and FCT test data items of the products is recorded as m. The 1st to mth ICT and FCT test data after normalization of the first product, the 1st to mth ICT and FCT test data after normalization of the second product, ..., the 1st to mth ICT and FCT test data after normalization of the nth product are used to construct A. nm ;in, Step S3.2, respectively calculate the K values ​​of all products in the union of the normal data set and the abnormal data set; the K value of the i-th product in the union of the normal data set and the abnormal data set is recorded as K i , where K i =max(k i1 ,k i2 ,…,k im ); With K value as the horizontal coordinate, and with the proportion of the number of returned products corresponding to K value in the total number of returned products and the proportion of the number of non-returned products corresponding to K value in the total number of non-returned products as the vertical coordinate, a coordinate system is established and corresponding curves are drawn, thereby obtaining a curve of the proportion of returned products-K value and a curve of the proportion of non-returned products-K value; obtain the K value at the position with the largest spacing between the curve of the proportion of returned products-K value and the curve of the proportion of non-returned products-K value in the vertical coordinate direction, recorded as k'; Step S3.3: define the products in the abnormal data set whose K value is greater than the absolute value of k' as soft failure products; extract the ICT and FCT test data of soft failure products to form a soft failure product data set; construct B n(m+1) ,in, Here, y1 represents the value of whether the first product in the union of the normal data set and the abnormal data set is a soft failure product; y2 represents the value of whether the second product in the union of the normal data set and the abnormal data set is a soft failure product; ...; y n Indicates whether the nth product in the union of the normal data set and the abnormal data set is a soft failure product; When y1, y2, …, y n When any one of y1, y2, ..., y n When any one of them is 1, it means that the corresponding product is a soft failure product; Step S3.4: n(m+1) The first m columns are used as independent variables, and the m+1th column is used as the dependent variable. The machine learning classification algorithm is used for analysis to construct a soft failure prediction model.

2. The soft failure prediction method according to claim 1, characterized in that: Step S3 is performed by adopting a machine learning algorithm.

3. The soft failure prediction method according to claim 1, characterized in that: Step S4 also includes: normalizing the real-time ICT and FCT test data of the product, and then inputting the normalized data into the soft failure prediction model as independent variables to obtain and output the prediction results.

Images