Automobile part quality early warning method and device, computer equipment and readable storage medium

Abstract

This application relates to a quality early warning method, apparatus, computer device, and computer-readable storage medium for automotive parts. The method includes: performing early warning detection on the automotive parts under preset early warning dimensions based on historical quality defect data, and obtaining early warning detection results; the preset early warning dimensions include at least one of the following: defective quantity, defective quantity growth rate, and predicted defective quantity change rate; the historical quality defect data includes the number of defective automotive parts over multiple periods; fusing the early warning detection results under each early warning dimension to obtain a comprehensive early warning detection result for the automotive parts; and performing graded quality early warning for the automotive parts based on the comprehensive early warning detection result. This method can improve the accuracy of early warnings.

Description

Technical Field

[0001] This application relates to the field of vehicle quality management technology, and in particular to a quality early warning method, device, computer equipment, and computer-readable storage medium for automotive parts. Background Technology

[0002] With the rapid development of the automotive industry, the situation of defective products in the automotive market, such as the outbreak of quality problems, consumes a lot of labor costs in isolating and analyzing them; if defective parts flow into the market, it may cause economic losses to car manufacturers, dealers and related supply chains. Therefore, timely and effective early warning of defective products is needed so that quality management personnel can take corresponding measures in advance to reduce potential economic losses.

[0003] Traditional methods for analyzing the market quality of automotive parts often rely on human experience and simple statistical data. Warning results are obtained by comparing preset warning thresholds with statistical data, resulting in relatively simple warning results and warning benchmarks, thus leading to low warning accuracy. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, device, computer equipment, computer-readable storage medium, and computer program product for quality early warning of automotive parts that can improve the accuracy of early warning, in response to the above-mentioned technical problems.

[0005] Firstly, this application provides a quality early warning method for automotive parts, including:

[0006] Based on historical quality defect data of automotive parts, the automotive parts are subjected to early warning detection under a preset early warning dimension to obtain early warning detection results; the preset early warning dimension includes at least one of the number of defective products, the growth rate of defective products, and the predicted change rate of defective products; the historical quality defect data includes the number of defective products of the automotive parts in multiple periods.

[0007] The warning detection results under each warning dimension are fused to obtain the comprehensive warning detection results for the automotive parts;

[0008] Based on the comprehensive early warning detection results, the automotive parts are classified and given early warnings.

[0009] In one embodiment, the step of performing early warning detection on the automotive parts based on historical quality defect data of the automotive parts under a preset early warning dimension, and obtaining early warning detection results, includes:

[0010] Based on the historical quality defect data, determine the cumulative number of defective automotive parts;

[0011] Based on the cumulative number of defective products and the number of defective products in each cycle, the early warning detection result of the automotive parts under the dimension of the number of defective products is obtained.

[0012] In one embodiment, obtaining the early warning detection result of the automotive part in terms of the number of defective products based on the cumulative number of defective products and the number of defective products in each cycle includes:

[0013] Based on the number of defective items in each cycle, the increase in the number of defective items in the cycle closest to the current time relative to the next closest cycle is determined, as well as the trend of the number of defective items in the multiple cycles closest to the current time. If the cumulative number of defective items is greater than a quantity threshold, the increase in the number of defective items is greater than an increase threshold, or the trend of change is a continuous increase, the warning detection result of the automotive part in the dimension of the number of defective items is determined as triggering a warning. The increase threshold is determined based on the average number of defective items in each cycle.

[0014] In one embodiment, the step of performing early warning detection on the automotive parts based on historical quality defect data of the automotive parts under a preset early warning dimension, and obtaining early warning detection results, includes:

[0015] Based on the historical quality defect data, the defect growth rate for each period and the average defect growth rate for each period are determined; from each period, a target period in which the defect growth rate is greater than the average defect growth rate is identified; if the number of target periods is greater than a preset target period number threshold, the warning detection result of the automotive part in the defect growth rate dimension is determined as triggering a warning.

[0016] In one embodiment, the step of performing early warning detection on the automotive parts based on historical quality defect data of the automotive parts under a preset early warning dimension, and obtaining early warning detection results, includes:

[0017] The quality defect data of the period most recent to the current time is input into a pre-trained defect quantity prediction model to obtain the defect quantity prediction result for the next period most recent to the current time. The pre-trained defect quantity prediction model is obtained by training an initial long short-term memory neural network with historical quality defect data of various automotive parts. The defect quantity prediction result is subjected to secondary difference processing to obtain the defect change rate of the next period. If the defect change rate of the next period exceeds a preset warning value, the warning detection result of the automotive part in the dimension of the predicted defect change rate is determined as a trigger warning.

[0018] In one embodiment, the quality early warning method for automotive parts described in the above embodiments further includes:

[0019] Based on the historical quality defect data, the mean and variance of the defect growth rate are determined; based on the time period length of a single period, the mean and variance of the defect growth rate, and the value at risk algorithm, the preset warning value is determined.

[0020] In one embodiment, the step of providing graded quality warnings for each type of automotive part based on the comprehensive early warning detection results includes:

[0021] Based on the comprehensive early warning detection results and the preset graded early warning rules, graded quality early warning results are obtained; based on the graded quality early warning results, a heat map corresponding to the graded quality early warning results is generated.

[0022] The heatmap is shown.

[0023] Secondly, this application also provides a quality warning device for automotive parts, comprising:

[0024] The early warning determination module is used to perform early warning detection on the automotive parts under a preset early warning dimension based on the historical quality defect data of the automotive parts, and obtain the early warning detection result; the preset early warning dimension includes at least one of the number of defective products, the growth rate of defective products, and the predicted change rate of defective products; the historical quality defect data includes the number of defective products of the automotive parts in multiple periods.

[0025] The result fusion module is used to fuse the warning detection results under various warning dimensions to obtain the comprehensive warning detection results of the automotive parts.

[0026] The graded early warning module is used to provide graded quality early warnings for the automotive parts based on the comprehensive early warning detection results.

[0027] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0028] Based on historical quality defect data of automotive parts, the automotive parts are subjected to early warning detection under a preset early warning dimension to obtain early warning detection results; the preset early warning dimension includes at least one of the number of defective products, the growth rate of defective products, and the predicted change rate of defective products; the historical quality defect data includes the number of defective products of the automotive parts in multiple periods.

[0029] The warning detection results under each warning dimension are fused to obtain the comprehensive warning detection results for the automotive parts;

[0030] Based on the comprehensive early warning detection results, the automotive parts are classified and given early warnings.

[0031] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0032] Based on historical quality defect data of automotive parts, the automotive parts are subjected to early warning detection under a preset early warning dimension to obtain early warning detection results; the preset early warning dimension includes at least one of the number of defective products, the growth rate of defective products, and the predicted change rate of defective products; the historical quality defect data includes the number of defective products of the automotive parts in multiple periods.

[0033] The warning detection results under each warning dimension are fused to obtain the comprehensive warning detection results for the automotive parts;

[0034] Based on the comprehensive early warning detection results, the automotive parts are classified and given early warnings.

[0035] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0036] Based on historical quality defect data of automotive parts, the automotive parts are subjected to early warning detection under a preset early warning dimension to obtain early warning detection results; the preset early warning dimension includes at least one of the number of defective products, the growth rate of defective products, and the predicted change rate of defective products; the historical quality defect data includes the number of defective products of the automotive parts in multiple periods.

[0037] The warning detection results under each warning dimension are fused to obtain the comprehensive warning detection results for the automotive parts;

[0038] Based on the comprehensive early warning detection results, the automotive parts are classified and given early warnings.

[0039] The aforementioned quality early warning method, device, computer equipment, computer-readable storage medium, and computer program product for automotive parts, by using historical quality defect data of automotive parts, performs early warning detection on automotive parts under preset early warning dimensions, and obtains early warning detection results. This achieves multi-faceted early warning detection of historical quality defect data for automotive parts. The preset early warning dimensions include at least one of the following: the number of defective products, the defective product growth rate, and the predicted defective product change rate. Historical quality defect data includes the number of defective automotive parts over multiple periods. This approach considers not only the historical quality status and defective product changes of automotive parts but also the predicted changes in defects, enabling timely implementation of corresponding measures based on the prediction results. Then, the early warning detection results under each early warning dimension are fused to obtain a comprehensive early warning detection result for the automotive parts. The fusion of multiple early warning detection results to obtain the final comprehensive early warning detection result avoids the one-sidedness of early warnings. Furthermore, based on the comprehensive early warning detection result, graded quality early warnings are applied to the automotive parts. Graded early warnings improve the effectiveness of early warnings and facilitate targeted responses and appropriate measures. In summary, this method achieves multi-dimensional early warning detection for automotive parts, improving the accuracy of early warnings. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 This is a flowchart illustrating a quality early warning method for automotive parts in one embodiment;

[0042] Figure 2 This is a schematic diagram of the quantity threshold calculation process in one embodiment;

[0043] Figure 3 This is a schematic diagram of the change in the number of defective products showing a fluctuating increase in one embodiment;

[0044] Figure 4 This is a schematic diagram illustrating the relationship between cumulative defect rate (cumulative defect rate) and time in one embodiment;

[0045] Figure 5 This is a schematic diagram illustrating the test results obtained through training in one embodiment.

[0046] Figure 6 This is a schematic diagram of the preprocessing flow for historical quality defect data in one embodiment.

[0047] Figure 7 This is a flowchart illustrating the process of determining the defect rate of change for the next cycle in one embodiment.

[0048] Figure 8 This is a schematic diagram illustrating the result of a warning triggered by the rate of change of defective products in the next cycle in one embodiment.

[0049] Figure 9 This is a schematic diagram of the warning score heatmap in one embodiment;

[0050] Figure 10 This is a schematic diagram of a defect early warning method for automotive parts in another embodiment;

[0051] Figure 11 This is a schematic diagram of the structure of a pre-emptive warning model for a defective automotive parts market in one embodiment;

[0052] Figure 12 This is a schematic diagram illustrating the calculation process of the cumulative non-performing growth rate and the average value in one embodiment;

[0053] Figure 13 This is a schematic diagram illustrating the verification and early warning results of a market distress classification and early warning model in one embodiment;

[0054] Figure 14 This is a schematic diagram of the failure warning results of the 2020 EVCT transmission in one embodiment;

[0055] Figure 15 This is a structural block diagram of a quality warning device for automotive parts in one embodiment;

[0056] Figure 16 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0058] In one embodiment, such as Figure 1 As shown, a quality early warning method for automotive parts is provided. This embodiment illustrates the method's application to a terminal, but it is understood that the method can also be applied to a server, or to a system including both a terminal and a server, and implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0059] Step S102: Based on the historical quality defect data of automotive parts, perform early warning detection on the automotive parts under the preset early warning dimensions, and obtain the early warning detection results.

[0060] The preset early warning dimensions include at least one of the following: the number of defective products, the growth rate of defective products, and the predicted rate of change of defective products.

[0061] Historical quality defect data includes the number of defective automotive parts over multiple periods. The periods can be set according to actual statistical needs, such as one month, one week, or one day. Historical quality defect data can be relevant data on defective automotive parts from any historical point in time to the present. Data characteristics include production and purchase dates, failure dates, mileage, model year, VIN, engine and transmission numbers, diagnostic information, authorized dealer information, vehicle type, model number, part number, and sales volume.

[0062] The types of automotive parts can be varied, and each type of automotive part is applicable to the quality early warning method for automotive parts proposed in this embodiment.

[0063] Optionally, the system acquires historical quality defect data of automotive parts, performs data cleaning and other preprocessing on the acquired historical quality defect data, and performs early warning detection on automotive parts under preset dimensions based on the preprocessed historical quality defect data to obtain corresponding early warning detection results.

[0064] Step S104: The warning detection results under each warning dimension are fused to obtain the comprehensive warning detection results for the automotive parts.

[0065] Among them, the comprehensive early warning detection result can be the final early warning result for the quality of automotive parts, which integrates the early warning detection results under various early warning dimensions, and is more comprehensive.

[0066] Optionally, the system integrates the warning detection results under each warning dimension. For example, the warning detection results under each dimension are determined by a scoring system, with each warning detection result corresponding to its own score. The scores of all warning detection results are added together to obtain the comprehensive warning detection result for the automotive parts.

[0067] Step S106: Based on the comprehensive early warning detection results, classify and issue early warnings for the quality of automotive parts.

[0068] Among them, graded quality early warning can be issued at different levels for different quality grades.

[0069] Optionally, the system determines the quality level of the inspected automotive parts based on the comprehensive early warning detection results, determines the corresponding early warning method based on the quality level, and performs graded quality early warnings according to the corresponding early warning method. For example, if the comprehensive quality inspection result of the automotive parts is severe, a red warning is issued; if the comprehensive quality inspection result is mild, a yellow alert is issued, etc.

[0070] The aforementioned quality early warning method for automotive parts utilizes historical quality defect data to perform early warning detection on automotive parts under preset early warning dimensions, obtaining early warning detection results. This achieves multi-faceted early warning detection of historical quality defect data for automotive parts. The preset early warning dimensions include at least one of the following: the number of defective parts, the defective part growth rate, and the predicted defective part change rate. The historical quality defect data includes the number of defective parts within multiple periods, considering not only the historical quality status and defective part changes of the automotive parts but also the predicted defective part changes, enabling timely implementation of corresponding measures based on the prediction results. Then, the early warning detection results under each early warning dimension are fused to obtain a comprehensive early warning detection result for the automotive parts. This fusion of multiple early warning detection results avoids the one-sidedness of early warnings. Furthermore, based on the comprehensive early warning detection result, graded quality early warnings are applied to the automotive parts. Graded early warnings improve the effectiveness of early warnings and facilitate targeted responses and appropriate measures. In summary, this method achieves multi-dimensional early warning detection for automotive parts, improving the accuracy of early warnings.

[0071] In an exemplary embodiment, step S102 involves performing early warning detection on the automotive parts based on historical quality defect data of the automotive parts under a preset early warning dimension, and obtaining early warning detection results, including:

[0072] Based on historical quality defect data, the cumulative number of defective automotive parts is determined; based on the cumulative number of defective parts and the number of defective parts in each period, the early warning detection results of automotive parts in terms of the number of defective parts are obtained.

[0073] The cumulative number of defective products can be the sum of the number of defective products counted within the historical time period corresponding to the historical quality defect data.

[0074] Optionally, the system calculates the cumulative number of defective automotive parts based on the total number of defective parts in each period of historical quality defect data. Then, based on the cumulative number of defective parts and the number of defective parts in each period, and according to preset warning rules, it issues warnings for both the cumulative number of defective parts and the number of defective parts in each period, thus obtaining the warning detection results for the automotive parts in terms of defect quantity. For example, the warning rule could be to compare the cumulative number of defective parts and the number of defective parts in each period with preset defect thresholds, and determine the warning detection result based on the comparison result.

[0075] In this embodiment, when performing early warning detection on automobiles based on the quantity of defective products, the early warning is made by considering both the cumulative number of defective products and the number of defective products in each cycle, which further improves the accuracy of the early warning.

[0076] In an exemplary embodiment, the steps of the above embodiment, based on the cumulative number of defective products and the number of defective products in each cycle, obtain the early warning detection result of automotive parts in terms of the number of defective products, including:

[0077] Based on the number of defective products in each cycle, determine the increase in the number of defective products in the cycle closest to the current time relative to the next closest cycle, as well as the trend of the number of defective products in the multiple cycles closest to the current time; if the cumulative number of defective products is greater than the quantity threshold, the increase in the number of defective products is greater than the increase threshold, or the trend is a continuous increase, determine the warning detection result of the automotive parts in the dimension of the number of defective products as the trigger warning.

[0078] The increase in defective products can be any period in which the number of defective products in that period increases relative to the number of defective products in the previous period.

[0079] The trend of the number of defective products can be the change in the number of defective products with each cycle, such as an increasing trend over time or a fluctuating trend over time.

[0080] Both the quantity threshold and the growth threshold can be set and flexibly adjusted according to actual early warning needs. The method for setting the quantity threshold can be as follows: Figure 2 The diagram illustrates the quantity threshold calculation process. The cumulative number of defective items in the diagram is equivalent to the cumulative number of defective products in this embodiment, and the cumulative defective item warning value is equivalent to the quantity threshold in this embodiment. The cumulative defective product quantity threshold = 0.3% × (Year / Year Launch Date ÷ Warranty Period) × Cumulative Sales (refer to the market measures of competitors for judgment; if the incidence rate > 0.3%, proactive market measures should be considered). The date (year-month) is used as the reference. ,in It is the start date of the statistical data. This is the cutoff date for statistical data. It represents the number of defective parts for a specific component in the corresponding month (cycle). It's a model year; the corresponding release date and sales volume are respectively... and The warranty period is The cumulative defective product quantity threshold for a period of one month can be written as:

[0081]

[0082] in, These are adjustable parameters; industry-experienced parameters are selected here. When the cumulative number of defective products exceeds the quantity threshold, i.e. At that time, the auto part was flagged as a warning. The growth threshold is determined based on the average number of defective items across all periods. A warning is triggered when the growth in defective items in the period closest to the current time, relative to the second closest period, exceeds the growth threshold. The corresponding formula is:

[0083]

[0084] in, It is an adjustable parameter. This refers to the cumulative number of defective products, which is calculated when the number of defective products in a single cycle exceeds the average number of defective products across all cycles. When the value is doubled, an early warning is triggered.

[0085] Optionally, the system determines the difference between the number of defective products in the most recent period and the number of defective products in the second most recent period based on the number of defective products in each period, determines the increase in the number of defective products in the most recent period relative to the second most recent period, and further determines the trend of the number of defective products in multiple periods most recent to the current time. If the cumulative number of defective products is greater than the quantity threshold, the increase in the number of defective products is greater than the increase threshold, or the trend of change is continuous growth, the system determines the warning detection result of the automotive parts in the dimension of the number of defective products as the trigger warning. For example, the system is set to trigger the warning if the trend of the number of defective products in three or more consecutive periods shows an increase.

[0086] It should be noted that in this embodiment, comparing the cumulative number of defective products with a quantity threshold to trigger an early warning is for the overall quality level of each type of automotive part. If an early warning is triggered, it may be due to defects in the part's design itself, insufficient overall quality assurance capabilities of the component supplier, or other reasons. However, using the increase in defective products and the increase threshold for each period, as well as determining the trend of defective product quantity changes in the most recent several periods, to trigger an early warning is for defective products in recent production batches. If an early warning is triggered, it may be due to changes in the component supplier, or other reasons.

[0087] In this embodiment, the cumulative number of defective products is compared with a quantity threshold to trigger an early warning. This takes into account the overall quality level of automotive parts from the time they leave the factory. The early warning is triggered by comparing the growth amount of defective products in each cycle with the growth threshold, as well as by determining the trend of the number of defective products in the most recent cycles. This is to check the recent quality level of automotive parts, thereby enabling more accurate early warnings. At the same time, it can also quickly determine the cause of the early warning based on the warning results, so as to take corresponding measures.

[0088] In an exemplary embodiment, step S104 involves performing early warning detection on the automotive parts based on historical quality defect data of the automotive parts under a preset early warning dimension, and obtaining early warning detection results, including:

[0089] Based on historical quality defect data, determine the defect growth rate for each period and the average defect growth rate for each period; identify target periods from each period where the defect growth rate is greater than the average defect growth rate; and determine the early warning detection result for automotive parts in the defect growth rate dimension as the trigger for an early warning when the number of target periods exceeds the preset target period number threshold.

[0090] The defective product growth rate can refer to the growth ratio of the number of defective products in any period relative to the number of defective products in the previous period. It can be calculated as the difference between the number of defective products in the current period and the number of defective products in the previous period, and the ratio of this difference to the number of defective products in the previous period is taken as the defective product growth rate in the current period.

[0091] The target period number threshold can be set and adjusted according to actual early warning needs.

[0092] Optionally, the system calculates the defective product growth rate for each period and the average defective product growth rate for each period based on historical quality defect data. It then compares the defective product growth rate for each period with the average defective product growth rate to obtain a comparison result. Based on this comparison result, it identifies target periods from each period where the defective product growth rate exceeds the average defective product growth rate. If the number of target periods exceeds a preset threshold, the system determines the warning detection result for automotive parts in the defective product growth rate dimension as the trigger for an alert. For example, it selects historical quality defect data from the past six months, calculates the defective product growth rate and average defective product growth rate for the past five months, and if the defective product growth rate exceeds the average defective product growth rate for three or more of the five months, an alert is triggered.

[0093] It should be noted that this embodiment determines whether to trigger an early warning by determining the number of target periods in which the defective product growth rate exceeds the average defective product growth rate. The derivation of this early warning rule is based on the fact that the above embodiment provides early warnings for the cumulative number of defective products, effectively alerting for parts with excessively high cumulative defective product numbers, recent frequent increases, and sharp increases. However, it lacks early warning capabilities for parts with fluctuating defective product numbers, such as... Figure 3 The diagram illustrates a fluctuating upward trend in the number of defective products (number of defective parts). To address this situation, the defective product growth rate needs to be used for assessment. Since the defect rate of automotive parts generally follows a Weibull distribution, its cumulative defect rate over the first 12 months can be approximated as a linear function, such as... Figure 4 As shown, a curve relating cumulative defect rate (cumulative non-conforming rate) to time is provided. Differentiating the cumulative defect rate for the first 12 months yields a constant approximating the slope. By taking the cumulative defect rate over the past six months, and considering the intervals on the horizontal axis, the defect rate can be obtained by directly differencing the data. For components where the number of defective parts fluctuates, when the number of defective parts increases, the corresponding growth rate must have at least one [variable]. Greater than The argument is as follows: Assume It is the cumulative defect rate for the corresponding month (period). It is the average growth rate of defective products. If the corresponding month's defective product growth rate is given, then:

[0094]

[0095]

[0096]

[0097]

[0098] For components where the number of defective parts fluctuates, when the number of defective parts decreases, the growth rate... Since the number is negative, according to the pigeonhole principle, when the number of defective products increases, there must be at least one growth rate. ,otherwise:

[0099] With formula The contradiction is proven.

[0100] In this embodiment, the defect rate of defective products is used to provide early warning for changes in the number of defective automotive parts that are showing fluctuating increases. This compensates for the fact that early warning rules based on the number of defective products cannot detect and provide early warnings for automotive parts with fluctuating increases, and further improves the accuracy of quality early warnings for automotive parts.

[0101] In an exemplary embodiment, step S106 involves performing a warning detection on the automotive parts based on historical quality defect data of the automotive parts under a preset warning dimension, and obtaining the warning detection result, including:

[0102] The quality defect data of the period most recent to the current time is input into a pre-trained defect quantity prediction model to obtain the defect quantity prediction result of the next period most recent to the current time; the defect quantity prediction result is subjected to secondary difference processing to obtain the defect change rate of the next period; if the defect change rate of the next period exceeds the preset warning value, the warning detection result of the automotive parts in the dimension of predicted defect change rate is determined as the trigger warning.

[0103] The pre-trained defect prediction model is obtained by training an initial Long Short-Term Memory (LSTM) neural network using historical quality defect data from various automotive parts. The dataset is divided into two parts: 80% for training and 20% for testing. The LSTM is a special type of Recurrent Neural Network (RNN), designed for processing sequential data. Compared to general neural networks, RNNs can handle data with varying sequences. Since the horizontal axis of the automotive market defect data is time, it is time-series data, making RNNs suitable for analysis and prediction. Compared to RNNs, LSTMs have the advantage of remembering long-term information, making them suitable for processing and predicting important events with long intervals and delays in time series. Therefore, LSTMs are chosen to predict the number of defective automotive parts. The hyperparameter settings of the initial LSTM neural network in this embodiment are shown in Table 1.

[0104] Table 1

[0105]

[0106] A model for predicting the number of defective products for automotive part No. 12513 of the 2020 model year was trained, such as... Figure 5 The diagram shows the test results obtained from the training. The Actual curve represents the actual number of defects in each month, and the Predicted curve represents the predicted number of defects in each month, output by the defect quantity prediction model.

[0107] Difference is the calculation of the difference between adjacent elements in a sequence, which can help eliminate trends and seasonality in the data, making the data more suitable for linear modeling. First difference is the calculation of the difference between two adjacent time points; while second difference is a second difference based on first difference.

[0108] Optionally, after the system obtains historical quality defect data, it needs to preprocess the historical quality defect data, such as... Figure 6As shown, a preprocessing workflow for historical quality defect data is provided. The market's past defect data originates from the company's internal database. The data features include production and purchase dates, failure dates, mileage, model year, chassis number, engine and transmission numbers, diagnostic information, authorized dealer information, vehicle type, model number, part number, and sales volume. First, data cleaning is performed. Since the model provides warnings based on failure dates, and the output is the warning value for a specific component of the corresponding model year, only the failure date, model year, and main part number are retained after processing. First, data containing blank values ​​is removed, and then abnormal data such as incorrectly entered dates (mistakenly entered as mileage or part number) are removed. Next, the cleaned data is preprocessed. After extracting and integrating information such as model year, part number, and failure date, the final defect data table is obtained, as shown in Table 2 (taking part 16450 as an example). This table needs to be converted into a four-level data structure of "part number-model year-date-number of defective products" in Python (a computer language) to reduce computer memory usage and facilitate indexing during subsequent processing.

[0109] Table 2

[0110]

[0111] In addition, such as Figure 7 As shown, a flowchart illustrating the process of obtaining the defect rate of change (predicted defect rate of change) for the next period is provided. Preprocessed quality defect data from the period most recent to the current time is input into a pre-trained defect quantity prediction model to obtain the defect quantity prediction result for the next period from the period most recent to the current time. The defect quantity prediction result is then subjected to a second difference processing to obtain the defect rate of change for the next period, as shown below. Figure 8 As shown, a schematic diagram of the result of triggering an early warning based on the defect change rate (predicted defect change rate) of the next period is provided. When the defect change rate of the next period exceeds the preset early warning value (VaR warning line), the early warning detection result of the automotive parts in the dimension of predicted defect change rate is determined as the triggering of the early warning.

[0112] In this embodiment, a long short-term memory neural network is introduced to predict the number of defective products in the next cycle. By analyzing the prediction results, early warnings can be issued in advance, thereby improving the timeliness of the warnings and enabling timely implementation of corresponding measures.

[0113] In one exemplary embodiment, the quality early warning method for automotive parts described in the above embodiments further includes:

[0114] Based on historical quality defect data, determine the mean and variance of the defect growth rate; based on the length of a single period, the mean and variance of the defect growth rate, and the value at risk algorithm, determine the preset warning value.

[0115] The preset warning value is the VaR (Value at Risk) value. Confidence intervals are defined for the predicted results based on confidence levels. A warning is triggered when the predicted rate of change of defective goods exceeds the confidence interval. This represents the maximum possible loss of a financial asset or portfolio of securities within a specific future period, at a certain probability level (confidence level). Let be the confidence level, be the loss over the time period, and be the maximum possible loss. Then, the VaR model can be expressed as:

[0116]

[0117] In the automotive market's defective product sector, VaR refers to the maximum rate of increase in defective products for a given automotive part over a specific future period, at a given confidence level. (Note that this rate of increase differs from the cumulative rate of increase in defective products in the previous example; the latter is derived from the difference in cumulative defect rates and is a positive number; the former, derived from the difference in defect rates, can be positive or negative, reflecting the degree of change in defect rates. For ease of distinction, it will be referred to as the rate of change in defective products below.) For example, taking... , It is calculated to be 3 months. This indicates that in the next three months, there will be The probability guarantee of defective product variation rate Less than .

[0118] Optionally, based on historical quality defect data, the mean and variance of the defective product growth rate are determined; based on the time period length of a single cycle, the mean and variance of the defective product growth rate, and the Value at Risk (VaR) algorithm, a preset warning value VaR is determined, and the corresponding formulas include:

[0119]

[0120]

[0121]

[0122] in, It is under the standard normal distribution Quantiles These are the mean and variance of the defective product growth rate. This refers to the length of a single period. For the historical simulation method, the non-performing loan growth rates are sorted from largest to smallest, and the 5th percentile value is selected as the VaR value. The VaR values ​​of the parametric method and the historical simulation method are then weighted and averaged.

[0123] In this embodiment, the VaR model value is calculated using the Value at Risk algorithm as a preset warning value, which can reduce the error of the warning for the rate of change of defective products and further improve the accuracy of quality warning.

[0124] In an exemplary embodiment, step S106, based on the comprehensive early warning detection results, performs graded quality warnings for each type of automotive part, including:

[0125] Based on the comprehensive early warning detection results and the preset graded early warning rules, graded quality early warning results are obtained; based on the graded quality early warning results, a heat map corresponding to the graded quality early warning results is generated; and the heat map is displayed.

[0126] Among them, the tiered early warning rules can be the correspondence between comprehensive early warning detection results and early warning categories. As shown in Table 3, the specific content of the preset tiered rules is provided:

[0127] Table 3

[0128]

[0129] Here, the score P represents the score corresponding to the comprehensive early warning detection result, and the category represents the corresponding early warning category.

[0130] Heatmaps, also known as heat maps, are a data visualization technique that uses color changes to represent the numerical density or intensity of data. They are typically used to analyze and display the distribution of large amounts of data, helping people quickly identify patterns, trends, or outliers in the data.

[0131] Optionally, the system obtains graded quality warning results based on the comprehensive warning detection results and preset graded warning rules. For example, in the above embodiments, five warning triggering conditions are described. A scoring system is used to assign points to each triggered warning. Triggering a warning earns 2 points to the warning detection result. The comprehensive warning detection result is the sum of multiple warning detection results, with an upper limit of 10 points. As shown in Table 3, a comprehensive warning detection result of 0 to 2 points results in a warning; a comprehensive warning detection result of 2 to 6 points results in a warning; and a comprehensive warning detection result of 6 to 10 points results in a clarification. Further, the system generates a heatmap corresponding to the graded quality warning results for the automotive parts and displays the heatmap through the system's display module. Figure 9 As shown, a schematic diagram of the warning score heatmap is provided.

[0132] In this embodiment, a heatmap is introduced to display the graded quality early warning results of the comprehensive early warning detection. Technicians can intuitively determine the different levels of early warning from the heatmap and take corresponding measures, thus realizing the visualization of the comprehensive early warning detection results.

[0133] In one exemplary embodiment, such as Figure 10 As shown, another method for early warning of defects in automotive parts is provided, including:

[0134] Step 1, Data Preprocessing: Using Python's NumPy and Pandas libraries, market defect data is cleaned and preprocessed, stored in a data structure of "part number-model year-month-date-number of defects." A three-layer defect warning model is constructed based on deep learning algorithms, the Value at Risk (VaR) model, and differential difference methods from the field of artificial intelligence: a quantity layer, a defect rate layer, and an LSTM-VaR layer, to achieve graded warnings for automotive parts defects. The quantity layer executes three independent logical judgments: whether the cumulative number of defective parts exceeds a threshold, whether the number of defective parts is continuously increasing, and whether it is increasing rapidly. The growth rate layer mainly supplements the quantity layer by performing differential difference on the cumulative defect rate to determine if there is a fluctuating upward trend. The LSTM-VaR layer sets a warning line based on confidence level and judges whether there will be a significant risk of defects in the future based on the prediction results. All three layers use a scoring system, and the graded warning effect is displayed in the form of a heatmap based on the different scores. Overall, the implementation of this model is divided into four sub-parts: data preprocessing, defect analysis, graded warning, and warning visualization. Subsequent calculations require the launch dates and sales volumes of different model years. The model year, launch date, and sales volume data are directly retrieved from the database and integrated into a sub-table.

[0135] Among them, the market's past adverse data (historical quality defect data) comes from the company's internal database. The data features include production and purchase dates, failure dates, mileage, model year, chassis number, engine and transmission numbers, diagnostic information, authorized dealer information, vehicle type, model, part number, sales volume, etc.

[0136] Optionally, data cleaning is performed first. Since the model provides warnings based on the failure date, and the output is the warning value for a specific component of the corresponding model year, only the failure date, model year, and main part number are retained after processing. First, data containing blank values ​​are removed, and then abnormal data such as incorrectly entered dates (incorrectly entered as mileage or part number) are removed. Considering that some component defects are rare and sporadic (occasionally one or two defects occur), in order not to affect the subsequent growth rate layer and LSTM-VaR layer calculations, data with a monthly average number of defective components less than (an adjustable parameter, 0.5 is selected here) and no recent consecutive defects are removed, completing the data screening and cleaning. Then, the cleaned data is preprocessed. After extracting and integrating information such as model year, part number, and failure date, the final defect data table is obtained.

[0137] Step 2, Establishment of the Defect Early Warning Model: The architecture of the defect classification early warning model for the automotive parts market is as follows: Figure 11 As shown, the preprocessed data (the monthly defect count of a certain component of a certain model year) is input into the model and will sequentially enter the quantity layer, the defect rate layer, and the LSTM-VaR layer. Each layer calculates the data using different algorithms, judges and scores the calculation results according to the established rules, adds up the scores of each layer (the early warning detection results under each dimension), and finally outputs the early warning score of the data (the comprehensive early warning detection result).

[0138] The early warning functions of each layer in the model are complementary: the quantity layer is mainly responsible for judging the recent increase in the number of defects and providing early warnings for a recent outbreak of defects; the defect rate layer can determine whether the component has fluctuating growth. If it happens to be in a decline period after a peak, the quantity layer cannot provide an early warning, and the defect rate layer fills this gap; the LSTM-VaR layer uses the variance and quantiles of past data to give a warning line based on confidence level, and provides an early warning based on whether the prediction result reaches the line, supplementing the first two layers' single time dimension and making the early warning effect more timely. Finally, some adjustable parameters are involved in the calculation formulas of each layer, so that the early warning benchmark can be adjusted according to the actual situation of defect data and quality targets.

[0139] Optionally, regarding the functional principle of the quantity layer: In early warning methods, judging based on quantity is the most basic and intuitive method. For defective automotive parts, this means judging whether the number of defective parts is trending upward recently and whether the cumulative number of defective parts exceeds a threshold. Based on the above judgment, the following three early warning rules are established for this layer.

[0140] The first judgment scenario is based on the cumulative number of defective parts. The cumulative defective parts warning threshold is calculated as 0.3% × (model year launch date ÷ warranty period) × cumulative sales (referencing the market measures of competitors; if the incidence rate > 0.3%, proactive market measures should be considered). When the cumulative number of defective parts (cumulative number of defective products) exceeds the warning threshold, the component is immediately flagged for warning. The second judgment scenario is based on the increase in defective parts over multiple periods, considering the trend of defective parts over the most recent three months. If the number of defective parts continues to increase, the component is flagged for warning.

[0141] The third scenario is based on the increase in the number of defective parts in a single period. If the increase in the number of defective parts in a single period is greater than a multiple of the monthly average number of defective parts, then the component will be flagged as an alert. Here, is an adjustable parameter.

[0142] Furthermore, regarding the three judgment scenarios mentioned above, the reasons for triggering a warning in each scenario are as follows: For the first scenario, the overall quality level of the components is the primary concern. If a warning is triggered, it may be due to defects in the component design itself, insufficient overall quality assurance capabilities of the component supplier, or other reasons. For the second and third scenarios, recent production batch defects are the primary concerns. If a warning is triggered, it may be due to changes in the component supplier, or other reasons.

[0143] Optionally, regarding the function and principle of the defect rate layer, while the quantity layer's early warning system can effectively warn of parts with excessively high cumulative defective numbers, recent frequent increases, and sharp increases, it lacks early warning capabilities for parts with fluctuating defective numbers. To address this, it is necessary to use the cumulative defect rate growth rate for judgment, i.e., the second layer—the defect rate layer. Since the defect rate of automotive parts generally follows a Weibull distribution, its cumulative defect rate in the first 12 months can be approximated as a linear function. Taking the derivative of the cumulative defect rate for the first 12 months yields a constant approximating the slope; taking the cumulative defect rate for the past six months, due to the interval on the horizontal axis, directly differencing the data yields the defect rate growth rate.

[0144] This layer selects defect data from the past six months and calculates the cumulative defect growth rate over the past five periods. If the growth rate is greater than the average for three or more periods, the component is flagged for warning. Figure 12 The diagram illustrates the calculation process for the cumulative defect rate and average defect rate. In summary, the defect rate layer provides a good early warning for components with fluctuating increases in the number of defective parts, thus compensating for the shortcomings of the quantity layer.

[0145] Optionally, the function and principle of the LSTM-VaR layer are explained. The quantity layer and defect rate layer are already quite comprehensive in identifying various defect situations. However, the data used by these two layers is only updated to the statistical cutoff date, causing them to mostly issue warnings only when defects have already broken out on a large scale. In this situation, quality management personnel often can only take stopgap measures, which are essentially remedial. From a cost perspective, the cost of remedial measures is far higher than that of preventative measures. To improve this situation, this layer introduces a Long Short-Term Memory (LSTM) neural network to predict the number of defective items. By analyzing the prediction results, early warnings can be issued in advance, thus overcoming the shortcomings of the first two layers in terms of the timeliness of warnings. Prediction results are often uncertain, and directly issuing warnings based on prediction results will introduce large errors. Therefore, this layer additionally introduces a Value at Risk (VaR) model, which divides the prediction results into confidence intervals based on confidence levels. This layer only issues a warning when the predicted defect rate exceeds the confidence interval. By introducing the VaR model, the error of the warning results can be greatly reduced. First, the variance and mean are calculated using historical defect rates. The VaR value is then obtained using a formula. Next, LSTM is used to predict the number of defective items over three months. The predicted data is then subjected to a second difference to obtain the predicted defect change rate. Finally, the predicted defect change rate and VaR value are compared to determine whether an alert is necessary.

[0146] Step 3: Defect Warning and Visualization. Parts from each year undergo a total of 5 warning items across the quantity layer, growth rate layer, and LSTM-VaR layer: Cumulative number of defective parts > Warning value? Continuous increase in the number of defective parts? Significant increase in the number of defective parts? Cumulative defect growth rate > Average growth rate? Defect change rate > VaR value? Each item has a score of 2, with a maximum warning score of 10. A part receives a warning score for each item triggered. Clearly, the higher the warning score, the more severe the recent defects, requiring greater attention from quality management personnel and potentially necessitating appropriate countermeasures. Therefore, this model categorizes parts based on their scores and provides graded warnings. A warning heatmap is then generated using Python's seaborn library based on the part scores.

[0147] Step 4, Model Validation: Using the currently used early warning model as a benchmark, comparative validation is performed. The dataset selected is engine warranty failure data for models 2019, 2020, 2021, and 2022 of ACC10 vehicles from March 2019 to August 2022, with the early warning time point being August 2022. Figure 13 As shown, a schematic diagram of the verification and early warning results of this model is provided.

[0148] Of the 472 parts, 66 triggered warnings. Among these, parts with a score of 2 accounted for 77.55%, falling into the "reminder" category; parts with scores of 4 and 6 accounted for 22.45%, falling into the "warning" category. During the verification process, the warning data calculated by this model and the benchmark are shown in Table 4. In the table, ○ indicates that the part was only warned by this model, and the benchmark model did not issue a warning; ● indicates that the part was only warned by the benchmark model, and this model did not issue a warning; ▲ indicates that the part was warned by both the benchmark and this model.

[0149] Table 4

[0150]

[0151] As shown in Table 4, this model achieves full coverage of warning parts compared to the baseline model, and the number of warning parts increases from 33 to 66, significantly improving warning sensitivity. For parts like 12513-2020 (2020 cylinder head blind hole plug), the baseline model cannot provide warnings for this type of fluctuating and increasing defect, while this model can accurately identify and issue warnings through the defect rate layer. Similarly, for parts with a future trend of increasing defect numbers, the baseline model cannot provide warnings, but this model, through the LSTM-VaR layer, predicts the number of defective parts in the next three months, discovering that the number of defects in these parts may increase significantly in the future, thus providing a warning and intervening in advance. In the validation results, this model identified a total of 10 parts with a future growth trend through the LSTM-VaR layer, two of which were not identified by the baseline model and the quantity and defect rate layers of this model. Taking the 2020 EVCT transmission as an example, if... Figure 14 The image shows a schematic diagram illustrating the adverse warning results of the 2020 EVCT (Electric Variable Continuously Variable Transmission) transmission.

[0152] In this embodiment, the model's early warning sensitivity and timeliness are significantly improved compared to the baseline model. It can provide effective and timely early warnings for parts with fluctuating defective numbers and future growth trends, thus compensating for the shortcomings of the baseline model. The early warning results are visualized in the form of a heatmap, with tiered early warnings to facilitate analysis and judgment by quality management personnel. For cases of over-warning by this model, the early warning threshold can be adjusted by adjusting the model parameters to optimize the model performance to meet the desired quality objectives.

[0153] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0154] Based on the same inventive concept, this application also provides a quality warning device for automobile parts for implementing the quality warning method for automobile parts described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more embodiments of the quality warning device for automobile parts provided below can be found in the limitations of the quality warning method for automobile parts described above, and will not be repeated here.

[0155] In one exemplary embodiment, such as Figure 15 As shown, a quality early warning device 1500 for automotive parts is provided, comprising: an early warning determination module 1501, a result fusion module 1502, and a graded early warning module 1503, wherein:

[0156] The early warning determination module 1501 is used to perform early warning detection on automotive parts under preset early warning dimensions based on historical quality defect data of automotive parts, and obtain early warning detection results; the preset early warning dimensions include at least one of the following: number of defective products, defective product growth rate, and predicted defective product change rate; historical quality defect data includes the number of defective automotive parts in multiple periods.

[0157] The result fusion module 1502 is used to fuse the early warning detection results under various early warning dimensions to obtain the comprehensive early warning detection results of automotive parts.

[0158] The graded early warning module 1503 is used to provide graded quality early warnings for automotive parts based on comprehensive early warning detection results.

[0159] Furthermore, in one embodiment, the early warning determination module 1501 is also used to determine the cumulative number of defective automotive parts based on historical quality defect data; and to obtain the early warning detection result of automotive parts in the dimension of defect quantity based on the cumulative number of defective parts and the number of defective parts in each period.

[0160] Furthermore, in one embodiment, the early warning determination module 1501 is also used to determine the increase in the number of defective products in the most recent period relative to the second most recent period, and the trend of the number of defective products in the most recent multiple periods, based on the number of defective products in each period; if the cumulative number of defective products is greater than the quantity threshold, the increase in the number of defective products is greater than the increase threshold, or the trend of change is continuous growth, the early warning detection result of the automotive parts in the dimension of the number of defective products is determined as triggering an early warning; the increase threshold is determined based on the average number of defective products in each period.

[0161] Furthermore, in one embodiment, the early warning determination module 1501 is also used to determine the defective product growth rate of each period and the average defective product growth rate of each period based on historical quality defect data; determine the target periods from each period where the defective product growth rate is greater than the average defective product growth rate; and determine the early warning detection result of the automotive parts in the dimension of defective product growth rate as triggering an early warning when the number of target periods is greater than the preset target period number threshold.

[0162] Furthermore, in one embodiment, the early warning determination module 1501 is also used to input the quality defect data of the period most recent to the current time into a pre-trained defect quantity prediction model to obtain the defect quantity prediction result of the next period most recent to the current time; the pre-trained defect prediction model is obtained by training an initial long short-term memory neural network with historical quality defect data of various automotive parts; the defect quantity prediction result is subjected to secondary difference processing to obtain the defect change rate of the next period; if the defect change rate of the next period exceeds a preset early warning value, the early warning detection result of the automotive parts in the dimension of predicted defect change rate is determined as triggering an early warning.

[0163] Furthermore, in one embodiment, the early warning determination module 1501 is also used to determine the mean and variance of the defective product growth rate based on historical quality defect data; and to determine a preset early warning value based on the time period length of a single cycle, the mean and variance of the defective product growth rate, and the risk value algorithm.

[0164] Furthermore, in one embodiment, the graded early warning module 1503 is also used to obtain graded quality early warning results based on the comprehensive early warning detection results and preset graded early warning rules; generate a heat map corresponding to the graded quality early warning results based on the graded quality early warning results; and display the heat map.

[0165] The various modules in the aforementioned quality warning device for automotive parts can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0166] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 16 As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs in the non-volatile storage media to run. The database stores historical defect quality data, early warning detection results under various warning dimensions, and other data. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a quality early warning method for automotive parts.

[0167] Those skilled in the art will understand that Figure 16 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0168] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0169] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0170] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0171] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0172] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0173] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method of quality early warning of an automobile part, characterized by, The method comprises: According to the historical quality defect data of the automobile parts, the automobile parts are detected in the preset warning dimension to obtain a warning detection result; the preset warning dimension includes the number of defective products, the growth rate of defective products and the predicted change rate of defective products; the historical quality defect data includes the number of defective products of the automobile parts in multiple periods; The warning detection results in each warning dimension are fused to obtain a comprehensive warning detection result of the automobile parts; According to the comprehensive warning detection result, the automobile parts are given a graded quality warning; Wherein, the warning detection result in the number of defective products dimension is determined based on the cumulative number of defective products and the number threshold, to give a warning for the overall quality level of each automobile part, or based on the number of defective products in the period closest to the current time relative to the period next closest, the growth amount threshold, and the change trend of the number of defective products in multiple periods closest to the current time, to give a warning for the defective products in the recent production batch; the warning detection result in the growth rate of defective products dimension is determined based on the number of target periods and the preset target period number threshold, the target period being the period with a defective product growth rate greater than the average defective product growth rate, to give a warning for the automobile parts with fluctuating upward trend; the warning detection result in the predicted change rate of defective products dimension is determined based on the next period's defective product change rate and the preset warning value, the next period's defective product change rate being determined based on the quality defect data of the period closest to the current time and the pre-trained defective product quantity prediction model, the obtained defective product quantity prediction result, and the preset warning value being a risk value model value representing the maximum defective product growth rate of the automobile parts in a given confidence period in the future.

2. The method of claim 1, wherein, The method comprises: According to the historical quality defect data of the automobile parts, the automobile parts are detected in the preset warning dimension to obtain a warning detection result; the preset warning dimension includes the number of defective products, the growth rate of defective products and the predicted change rate of defective products; the historical quality defect data includes the number of defective products of the automobile parts in multiple periods; The warning detection results in each warning dimension are fused to obtain a comprehensive warning detection result of the automobile parts; 3. The method of claim 2, wherein, According to the comprehensive warning detection result, the automobile parts are given a graded quality warning; Wherein, the warning detection result in the number of defective products dimension is determined based on the cumulative number of defective products and the number threshold, to give a warning for the overall quality level of each automobile part, or based on the number of defective products in the period closest to the current time relative to the period next closest, the growth amount threshold, and the change trend of the number of defective products in multiple periods closest to the current time, to give a warning for the defective products in the recent production batch; the warning detection result in the growth rate of defective products dimension is determined based on the number of target periods and the preset target period number threshold, the target period being the period with a defective product growth rate greater than the average defective product growth rate, to give a warning for the automobile parts with fluctuating upward trend; the warning detection result in the predicted change rate of defective products dimension is determined based on the next period's defective product change rate and the preset warning value, the next period's defective product change rate being determined based on the quality defect data of the period closest to the current time and the pre-trained defective product quantity prediction model, the obtained defective product quantity prediction result, and the preset warning value being a risk value model value representing the maximum defective product growth rate of the automobile parts in a given confidence period in the future. The method comprises: According to the historical quality defect data of the automobile parts, the automobile parts are detected in the preset warning dimension to obtain a warning detection result; the preset warning dimension includes the number of defective products, the growth rate of defective products and the predicted change rate of defective products; the historical quality defect data includes the number of defective products of the automobile parts in multiple periods; The warning detection results in each warning dimension are fused to obtain a comprehensive warning detection result of the automobile parts; According to the comprehensive warning detection result, the automobile parts are given a graded quality warning; Wherein, the warning detection result in the number of defective products dimension is determined based on the cumulative number of defective products and the number threshold, to give a warning for the overall quality level of each automobile part, or based on the number of defective products in the period closest to the current time relative to the period next closest, the growth amount threshold, and the change trend of the number of defective products in multiple periods closest to the current time, to give a warning for the defective products in the recent production batch; the warning detection result in the growth rate of defective products dimension is determined based on the number of target periods and the preset target period number threshold, the target period being the period with a defective product growth rate greater than the average defective product growth rate, to give a warning for the automobile parts with fluctuating upward trend; the warning detection result in the predicted change rate of defective products dimension is determined based on the next period's defective product change rate and the preset warning value, the next period's defective product change rate being determined based on the quality defect data of the period closest to the current time and the pre-trained defective product quantity prediction model, the obtained defective product quantity prediction result, and the preset warning value being a risk value model value representing the maximum defective product growth rate of the automobile parts in a given confidence period in the future. The method comprises: According to the historical quality defect data of the automobile parts, the automobile parts are detected in the preset warning dimension to obtain a warning detection result; the preset warning dimension includes the number of defective products, the growth rate of defective products and the predicted change rate of defective products; the historical quality defect data includes the number of defective products of the automobile parts in multiple periods; The warning detection results in each warning dimension are fused to obtain a comprehensive warning detection result of the automobile parts; According to the comprehensive warning detection result, the automobile parts are given a graded quality warning; Wherein, the warning detection result in the number of defective products dimension is determined based on the cumulative number of defective products and the number threshold, to give a warning for the overall quality level of each automobile part, or based on the number of defective products in the period closest to the current time relative to the period next closest, the growth amount threshold, and the change trend of the number of defective products in multiple periods closest to the current time, to give a warning for the defective products in the recent production batch; the warning detection result in the growth rate of defective products dimension is determined based on the number of target periods and the preset target period number threshold, the target period being the period with a defective product growth rate greater than the average defective product growth rate, to give a warning for the automobile parts with fluctuating upward trend; the warning detection result in the predicted change rate of defective products dimension is determined based on the next period's defective product change rate and the preset warning value, the next period's defective product change rate being determined based on the quality defect data of the period closest to the current time and the pre-trained defective product quantity prediction model, the obtained defective product quantity prediction result, and the preset warning value being a risk value model value representing the maximum defective product growth rate of the automobile parts in a given confidence period in the future.

4. The method of claim 1, wherein, The method comprises the following steps: According to the historical quality defect data of the automobile parts, the automobile parts are detected in a preset warning dimension to obtain a warning detection result, which comprises the following steps: According to the historical quality defect data, the defect product growth rate of each period and the average defect product growth rate of each period are determined; The target period in which the defect product growth rate is greater than the average defect product growth rate is determined from each period; 5. The method of claim 1, wherein, If the number of target periods is greater than a preset target period threshold, it is determined that the warning detection result of the automobile parts in the defect product growth rate dimension is a triggered warning. The method comprises the following steps: The quality defect data of the period closest to the current time is input into the pre-trained defect product quantity prediction model to obtain the defect product quantity prediction result of the next period closest to the current time; the pre-trained defect product prediction model is obtained by training an initial long short-term memory neural network through historical quality defect data of multiple automobile parts; The defect product quantity prediction result is subjected to secondary difference processing to obtain the defect product change rate of the next period; 6. The method of claim 5, wherein, If the defect product change rate of the next period exceeds the preset warning value, it is determined that the warning detection result of the automobile parts in the predicted defect product change rate dimension is a triggered warning. The method further comprises the following steps: According to the historical quality defect data, the mean of the defect product growth rate and the variance of the defect product growth rate are determined; 7. The method of claim 1, wherein, According to the time length of a single period, the mean and variance of the defect product growth rate, and a risk value algorithm, the preset warning value is determined. The method further comprises the following steps: According to the comprehensive warning detection result, a graded quality warning is performed on each type of automobile parts, which comprises the following steps: According to the comprehensive warning detection result and a preset grading warning rule, a graded quality warning result is obtained; 8. A quality early warning device for automotive parts, characterized in that, According to the graded quality warning result, a heat map corresponding to the graded quality warning result is generated; The heat map is displayed. The device comprises the following modules: A warning determination module is configured to detect automobile parts in a preset warning dimension according to historical quality defect data of the automobile parts to obtain a warning detection result; the preset warning dimension comprises defect product quantity, defect product growth rate and predicted defect product change rate; the historical quality defect data comprises the defect product quantity of the automobile parts in multiple periods; A result fusion module is configured to fuse the warning detection results in each warning dimension to obtain a comprehensive warning detection result of the automobile parts; A graded warning module is configured to perform a graded quality warning on the automobile parts according to the comprehensive warning detection result. The early warning detection result in the defect product quantity dimension is determined based on an accumulated defect product quantity and a quantity threshold value, to early warn the overall quality level of each automobile part, or is determined based on a defect product growth amount of a period closest to the current time relative to a second closest period, a growth amount threshold value, and a change trend of the defect product quantity of multiple periods closest to the current time, to early warn a defect product of a recent production batch; the early warning detection result in the defect product growth rate dimension is determined based on a target period quantity and a preset target period quantity threshold value, the target period being a period in which the defect product growth rate is greater than an average defect product growth rate, to early warn an automobile part with fluctuating rising; and the early warning detection result in the predicted defect product change rate dimension is determined based on a next period defect product change rate and a preset early warning value, the next period defect product change rate being determined based on quality defect data of the period closest to the current time and a pre-trained defect product quantity prediction model, a resulting defect product quantity prediction result, and the preset early warning value being a risk value model value representing a maximum defect product growth rate of the automobile part in a future specific period under a given confidence level. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. The computer program, when executed by the processor, implements the steps of the method of any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program, when executed by the processor, implements the steps of the method of any one of claims 1 to 7.

Images