Equipment health prediction methods and electronic devices

Abstract

This application discloses a method and electronic device for predicting equipment health. The method involves acquiring equipment data of the device under test (DUT); determining the operating parameters of each subsystem of the DUT, and determining the aging degree of each subsystem based on these operating parameters; calculating the initial health index of each subsystem based on sensor state parameters at multiple times within a first time period and the aging degree of each subsystem; inputting the initial health index of each subsystem into a preset process quality prediction model to obtain process quality prediction indicators; updating the initial health index based on the difference between the process quality prediction indicators and process quality parameters of each subsystem to obtain the target health index of each subsystem; and performing health prediction on the DUT based on the coupling strength between each subsystem and other subsystems, as well as the target health index, to obtain the health prediction result of the DUT.

Description

Technical Field

[0001] This application relates to the field of equipment health monitoring technology, and in particular to equipment health prediction methods and electronic devices. Background Technology

[0002] Surface Mount Technology (SMT) production line equipment requires health monitoring during operation. This can be achieved by collecting multi-source data during equipment operation and combining it with data analysis and modeling methods to assess and predict the performance degradation, potential failure risks, and remaining service life of key SMT equipment in real time. Currently, common health monitoring methods treat the entire machine as a single entity for health assessment. However, the impact of different subsystems on health varies, making it difficult to accurately identify anomalies in any particular subsystem. This can lead to localized subsystem failures escalating into global equipment failures. Therefore, how to conduct more accurate and comprehensive health assessments and predictions for SMT production line equipment has become a pressing issue. Summary of the Invention

[0003] This application provides a method and electronic equipment for predicting equipment health, in order to at least address the problem of how to conduct more accurate and comprehensive health status assessment and prediction of SMT production line equipment.

[0004] This application provides a method for predicting equipment health, including: Acquire equipment data of the device under test, including equipment operating parameters, sensor status parameters, and process quality parameters; Determine the equipment operating parameters corresponding to each subsystem of the device under test, and determine the aging degree of each subsystem based on the equipment operating parameters corresponding to each subsystem. Based on the sensor state parameters at multiple moments within the first time period and the aging degree of each subsystem, the initial health index of each subsystem is calculated. The initial health index of each subsystem is input into a preset process quality prediction model to obtain process quality prediction indicators. The initial health index is updated based on the difference between the process quality prediction index and the process quality parameter to obtain the target health index of each subsystem. Based on the coupling strength between each subsystem and other subsystems and the target health index, a health prediction is made for the device under test, and the health prediction result of the device under test is obtained.

[0005] This application also provides a device for predicting device health, comprising: The acquisition module is used to acquire equipment data of the device under test, including equipment operating parameters, sensor status parameters, and process quality parameters. The processing module is used to determine the equipment operating parameters corresponding to each subsystem of the device under test, and to determine the aging degree of each subsystem based on the equipment operating parameters corresponding to each subsystem. The processing module is also used to calculate the initial health index of each subsystem based on the sensor state parameters at multiple times within the first time period and the aging degree of each subsystem. The processing module is also used to input the initial health index of each subsystem into a preset process quality prediction model to obtain process quality prediction indicators. The processing module is further configured to update the initial health index based on the difference between the process quality prediction index and the process quality parameter, so as to obtain the target health index of each subsystem. The processing module is further configured to perform health prediction on the device under test based on the coupling strength between each subsystem and other subsystems and the target health index, and obtain the health prediction result of the device under test.

[0006] This application also provides an electronic device, including: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of any of the above-described device health prediction methods.

[0007] This application also provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the steps of any of the above-described device health prediction methods.

[0008] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of any of the above-described device health prediction methods.

[0009] This application acquires equipment data for the device under test (DUT), including equipment operating parameters, sensor status parameters, and process quality parameters. It determines the operating parameters for each subsystem of the DUT and, based on these parameters, determines the aging degree of each subsystem. Based on the sensor status parameters at multiple points within a first time period and the aging degree of each subsystem, it calculates the initial health index for each subsystem. The initial health index of each subsystem is input into a preset process quality prediction model to obtain a process quality prediction index. The initial health index is updated based on the difference between the process quality prediction index and the process quality parameters to obtain the target health index for each subsystem. Finally, based on the coupling strength between each subsystem and other subsystems, and the target health index, a health prediction is performed on the DUT to obtain the health prediction result. This approach divides the device under test (DUT) into subsystems with different functional characteristics and calculates health indices for each. This ensures that the health index reflects the physical characteristics and failure mechanisms of each subsystem as closely as possible, preventing local anomalies from being overlooked and achieving fine-grained health assessment. Furthermore, by combining multi-dimensional parameters to calculate the health index, closed-loop correction is achieved through process quality prediction. The coupling relationships between various subsystems are also considered, which effectively improves the authenticity and accuracy of the health status assessment, ensuring fault warnings for the DUT and guaranteeing the stability of the device mounting process. Attached Figure Description

[0010] To more clearly illustrate the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 A flowchart of a device health prediction method provided in this application embodiment Figure 1 ; Figure 2 A flowchart of a device health prediction method provided in this application embodiment Figure 2 ; Figure 3 A structural diagram of a device health prediction device provided in an embodiment of this application; Figure 4 This is a structural diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0012] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of this application.

[0013] It should be noted that, in the description of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. The terms "first," "second," etc., in this application are used to distinguish similar objects and are not used to describe a specific order or sequence.

[0014] It should be noted that in the embodiments of this application, the words "exemplary" or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of the words "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0015] SMT production line equipment health status monitoring technology refers to a technical system that collects multi-source data during equipment operation and combines data analysis and modeling methods to conduct real-time assessment and prediction of the performance degradation, potential failure risks, and remaining service life of key SMT equipment. Its core objective is to shift from reactive or periodic maintenance to predictive maintenance, ensuring the stability and reliability of high-precision, high-efficiency electronic manufacturing processes. Therefore, how to utilize advanced technologies to improve the intelligence level and safety of SMT production line equipment health status monitoring has become one of the most pressing issues to be addressed.

[0016] In the field of SMT production line equipment health status monitoring, traditional solutions treat the entire equipment as a single object for health assessment, which fails to identify early anomalies in a certain subsystem, causing local failures to develop into global shutdowns, missing the opportunity for precise maintenance, and lacking functional coupling modeling between subsystems, resulting in a lack of system-level risk transmission. SMT equipment modules are highly coordinated, but existing health assessments do not consider functional dependencies and cannot reflect the real risk of a chain reaction caused by the deterioration of a subsystem, making the overall machine health status assessment one-sided.

[0017] In summary, to address all or part of the aforementioned technical problems, this application provides a method and electronic device for predicting equipment health. The method involves acquiring equipment data of the device under test (DUT), including equipment operating parameters, sensor status parameters, and process quality parameters; determining the equipment operating parameters corresponding to each subsystem of the DUT, and determining the aging degree of each subsystem based on these parameters; calculating the initial health index of each subsystem based on sensor status parameters at multiple times within a first time period and the aging degree of each subsystem; inputting the initial health index of each subsystem into a preset process quality prediction model to obtain a process quality prediction index; updating the initial health index based on the difference between the process quality prediction index and the process quality parameters to obtain the target health index of each subsystem; and performing a health prediction of the DUT based on the coupling strength between each subsystem and other subsystems, as well as the target health index, to obtain the health prediction result of the DUT. This approach divides the device under test (DUT) into subsystems with different functional characteristics and calculates health indices for each. This ensures that the health index reflects the physical characteristics and failure mechanisms of each subsystem as closely as possible, preventing local anomalies from being overlooked and achieving fine-grained health assessment. Furthermore, by combining multi-dimensional parameters to calculate the health index, closed-loop correction is achieved through process quality prediction. The coupling relationships between various subsystems are also considered, which effectively improves the authenticity and accuracy of the health status assessment, ensuring fault warnings for the DUT and guaranteeing the stability of the device mounting process.

[0018] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0019] like Figure 1 As shown, Figure 1 A flowchart of a device health prediction method provided for embodiments of this application, the method may include the following steps: 101. Obtain the device data of the device under test.

[0020] In this embodiment of the application, the device under test is the device that needs to be health predicted. The device under test can be an SMT production line device, and the device data can include device operating parameters, sensor status parameters, and process quality parameters.

[0021] It should be noted that the equipment operating parameters can include data directly generated during equipment operation, specifically including: the number of placement actions performed by the placement head per minute, the number of batch switching by the feeder per unit time, and the real-time current value of the drive servo motor, etc.

[0022] It should be noted that the sensing status parameters can include data that the device under test needs to collect through sensors during operation. Specifically, these can include: the effective value of vibration acceleration at the mounting base of the mounting head, the temperature sensor reading near the lens of the vision system, and the pressure fluctuation amplitude in the pneumatic circuit, etc.

[0023] It should be noted that process quality parameters can include data used by the equipment under test to measure the quality of the products produced during operation. Specifically, these can include: component placement offset output by automatic optical inspection equipment, the proportion of component rejections to total material input as counted by the material tracking system, and soldering yield rate fed back by the online testing system.

[0024] It should be noted that the device data may include real-time device data collected at the current moment, as well as historical device data from a past period.

[0025] 102. Determine the equipment operating parameters corresponding to each subsystem of the device under test, and determine the aging degree of each subsystem based on the equipment operating parameters corresponding to each subsystem.

[0026] In the embodiments of this application, due to the mechanical and electrical functions of the device under test, the entire device under test can be divided into multiple subsystems with independent degradation characteristics, which may specifically include: placement head assembly, vision alignment system, feeder array and product conveying mechanism, etc.

[0027] The placement head assembly may include a Z-axis lifting mechanism, a nozzle array, and a drive motor; the vision alignment system may include an industrial camera, a light source module, and an image processing unit; the feeder array may include multiple tape or tray-type feeders and their switching mechanisms; and the product conveying mechanism may include a track, a clamping device, and a conveyor motor.

[0028] It should be noted that the above describes four subsystems, which can be considered to affect the health of the entire device under test, and these four subsystems have different functional characteristics.

[0029] In this embodiment of the application, the device operating parameters of the device under test collected above can correspond to each subsystem. That is, each subsystem has its corresponding device operating parameters. Therefore, the device operating parameters corresponding to each subsystem of the device under test can be determined, and then the aging degree of the subsystem can be determined based on the device operating parameters of each subsystem.

[0030] It should be noted that, for the placement head assembly, the corresponding equipment operating parameters can be the placement action frequency and the motion acceleration of the placement head; for the vision alignment system, the corresponding equipment operating parameters can be the image acquisition frame rate; for the feeder array, the corresponding equipment operating parameters can be the number of feeder switching times per unit time; and for the product conveying mechanism, the corresponding equipment operating parameters can be the product throughput frequency.

[0031] It should be noted that the equipment operating parameters of the above subsystems can be used to indicate the current workload of the subsystem. It can be understood that the higher the aging level of a subsystem, the lower its working efficiency will be. Therefore, the aging level of the corresponding subsystem can be determined based on the above equipment operating parameters. For example, the lower the placement operation frequency, the higher the aging level of the placement head assembly; the lower the image acquisition frame rate, the higher the aging level of the vision alignment system; the lower the number of feeder switching times per unit time, the higher the aging level of the feeder array; and the lower the product throughput frequency, the higher the aging level of the product conveying mechanism.

[0032] 103. Based on the sensor state parameters at multiple moments within the first time period and the aging degree of each subsystem, the initial health index of each subsystem is calculated.

[0033] In this embodiment of the application, after determining the aging degree of each subsystem, the current health index of each subsystem can be determined by combining the changes in the sensor status parameters over a period of time. This means that the initial health index is determined by the device operating parameters and sensor status parameters. The first duration can refer to a period of time before the current moment. Within the first duration, the sensor status parameters of the device under test can be acquired at various moments. Then, the sensor status parameters at multiple moments within the first duration, including the current moment, are summarized. The calculated initial health index of each subsystem can be determined by combining the stability of the sensor status parameters of each subsystem within the first duration.

[0034] It should be noted that each subsystem also has its corresponding sensing state parameters. In other words, for a subsystem, the initial health index of the subsystem can be calculated based on the sensing state parameters of the subsystem at multiple times within the first time period and the aging degree of the subsystem.

[0035] 104. Input the initial health index of each subsystem into the preset process quality prediction model to obtain the process quality prediction index.

[0036] In this embodiment, since the device under test is to output a product, and each subsystem contributes to the output of the product, the health index of each subsystem of the device under test can also be measured by the process quality of the product. At this time, the deviation between the process quality parameters and the process quality predicted based on the initial health index can be used to perform closed-loop correction on the initial health index of the subsystem. The process quality predicted based on the initial health index can be obtained by inputting the initial health index into a preset process quality prediction model. Specifically, after determining the initial health index of each subsystem, the initial health index of each subsystem can be input into the preset process quality prediction model to predict the process quality of each subsystem and obtain the process quality prediction index.

[0037] In some embodiments, since the SMT equipment can be a placement device, the process quality of the device under test mainly depends on the process quality of the placement head assembly. The preset process quality prediction model can be obtained through pre-training. The health index can be used to characterize the current health status of the subsystem. After processing the initial health index using the preset process quality prediction model, the process quality of the output product can be predicted after the subsystem continues to operate based on the initial health index.

[0038] 105. Based on the difference between the process quality prediction index and the process quality parameter, update the initial health index to obtain the target health index for each subsystem.

[0039] In this embodiment of the application, after obtaining the process quality prediction index through the preset process quality prediction model, it can be compared with the current process quality prediction index. Since the process quality prediction index output by the preset process quality prediction model can be the process quality of the product output by the device under test continuing to work according to the current initial health index, while the process quality parameter is the process quality of the actual output product, the initial health index can be updated by the difference between the process quality prediction index and the process quality parameter, thereby obtaining the target health index of each subsystem. The target health index can be considered as an index that can more accurately measure the current health status of the subsystem.

[0040] It should be noted that each of the above steps can be performed separately for each subsystem. That is, for a given subsystem, first determine the corresponding equipment operating parameters, sensor status parameters, and process quality parameters. Then, based on the equipment operating parameters, determine the aging degree of the subsystem. Next, based on the sensor status parameters and aging degree, determine the initial health index of the subsystem. Then, input the initial health index into a preset process quality prediction model to obtain the process quality prediction index for the subsystem. Finally, update the initial health index based on the difference between the process quality prediction index and the process quality parameters to obtain the target health index for the subsystem. Performing the above operations for each subsystem ensures that the subsystems do not affect each other, ultimately yielding the target health index for each subsystem.

[0041] 106. Based on the coupling strength between each subsystem and other subsystems and the target health index, perform health prediction on the device under test to obtain the health prediction result of the device under test.

[0042] In this embodiment of the application, although the various subsystems in the device under test have different functional characteristics, there is a certain coupling relationship between the various subsystems during the operation of the device. That is to say, there is a functional dependency relationship between some subsystems. When a subsystem is abnormal, it may affect other subsystems that depend on the abnormal subsystem and cause them to fail. Therefore, when making a health prediction of the device under test, the current target health index of each subsystem can be combined with the coupling strength between each subsystem and other subsystems to make a health prediction, thereby obtaining the health prediction result of the entire device under test.

[0043] It should be noted that the coupling strength between each subsystem and other subsystems can reflect the degree of functional dependence between the two subsystems. For example, the placement head assembly and the vision alignment system need to work together to complete the alignment and placement work, so the coupling strength between these two subsystems can be set to a higher value. On the other hand, the interaction between the feeder array and the product conveying mechanism is relatively small, and an abnormality in the feeder array will not affect the operation of the product conveying mechanism, so the coupling strength between these two subsystems can be set to a lower value.

[0044] It should be noted that when making health predictions for the device under test, the target health index of each subsystem can be used as a basis. This is because the health status of each subsystem affects the overall operating status of the device under test during operation. Therefore, the target health parameters of each subsystem are combined to predict the health status at future moments.

[0045] In this embodiment, equipment data of the device under test (DUT) is acquired, including equipment operating parameters, sensor status parameters, and process quality parameters. The operating parameters corresponding to each subsystem of the DUT are determined, and the aging degree of each subsystem is determined based on these parameters. An initial health index for each subsystem is calculated based on sensor status parameters at multiple times within a first time period and the aging degree of each subsystem. The initial health index of each subsystem is input into a preset process quality prediction model to obtain a process quality prediction index. The initial health index is updated based on the difference between the process quality prediction index and the process quality parameters to obtain a target health index for each subsystem. Finally, a health prediction is performed on the DUT based on the coupling strength between each subsystem and other subsystems, as well as the target health index, to obtain the health prediction result of the DUT. This approach divides the device under test (DUT) into subsystems with different functional characteristics and calculates health indices for each. This ensures that the health index reflects the physical characteristics and failure mechanisms of each subsystem as closely as possible, preventing local anomalies from being overlooked and achieving fine-grained health assessment. Furthermore, by combining multi-dimensional parameters to calculate the health index, closed-loop correction is achieved through process quality prediction. The coupling relationships between various subsystems are also considered, which effectively improves the authenticity and accuracy of the health status assessment, ensuring fault warnings for the DUT and guaranteeing the stability of the device mounting process.

[0046] like Figure 2 As shown, Figure 2 Another flowchart of a device health prediction method provided for embodiments of this application, the method may include the following steps: 201. Obtain the device data of the device under test.

[0047] 202. Determine the equipment operating parameters corresponding to each subsystem of the device under test.

[0048] In the embodiments of this application, the description of steps 201 to 202 is the same as the detailed description of steps 101 to 102 in the above embodiments, and will not be repeated in the embodiments of this application.

[0049] 203. Determine the load intensity of each subsystem based on the equipment operating parameters corresponding to each subsystem.

[0050] In this embodiment of the application, after determining the equipment operating parameters corresponding to each subsystem, the load intensity of the subsystem can be determined based on the equipment operating parameters. The specific method for determining the load intensity can be determined based on the equipment operating parameters.

[0051] For example, for a placement head assembly, the load strength can be determined by both the placement frequency and the motion acceleration. The load strength of the placement head assembly can be a weighted sum of the placement frequency and the peak acceleration. The weights can be set by the user or determined based on historical data. For feeder arrays, the load intensity can be directly characterized by the number of feeder switching times per unit time. For visual alignment systems, the load intensity can be directly represented by the image acquisition frame rate; For product conveying mechanisms, load strength can be directly characterized by the frequency of product passage.

[0052] 204. Integrate the load intensity corresponding to multiple moments within the second time period to obtain the cumulative aging amount of each subsystem.

[0053] In this embodiment of the application, after determining the load intensity of each subsystem, the load intensity of the subsystem over a continuous operating period can be integrated over time to obtain the aging accumulation amount reflecting the actual usage intensity. The second duration can be a period of time prior to the current moment, and the plurality of moments includes the current moment and historical moments within the second duration prior to the current moment.

[0054] In some embodiments, integrating the load intensity can be expressed as:

[0055] Where L represents the cumulative aging amount. Let represent the load intensity, and t represent the specific time intervals. For each subsystem, the load intensity at each time interval can be integrated using the above formula to obtain the cumulative aging amount of each subsystem.

[0056] 205. Determine the aging degree of each subsystem based on the exponential decay function and the cumulative aging amount of each subsystem.

[0057] In this embodiment of the application, after determining the cumulative aging amount of each subsystem, the cumulative aging amount of each subsystem can be substituted into the exponential decay function to obtain the aging degree of each subsystem. The larger the cumulative aging amount, the higher the corresponding aging degree.

[0058] In some embodiments, the exponential decay function can be expressed as:

[0059] Where A represents the degree of aging, and L represents the cumulative amount of aging. This is the aging rate coefficient corresponding to the subsystem, which is a property parameter of each subsystem itself.

[0060] In this embodiment, the aging degree of the subsystem is characterized by time integral based on actual load intensity rather than calendar time. This can truly reflect the cumulative usage effect of each subsystem in the SMT equipment under typical working conditions such as intermittent production and variable cycle time. Especially for highly dynamic components such as the placement head, their wear and fatigue mainly depend on the frequency of action and acceleration, rather than simple running time. Therefore, the working condition-aware aging calculation method improves the physical rationality and engineering applicability of health status assessment and remaining life prediction.

[0061] 206. Normalize the sensor state parameters at multiple times within the first time period to obtain multiple normalized sensor state parameters.

[0062] In this embodiment of the application, since the value range of the sensing state parameters corresponding to each subsystem is different, in order to facilitate calculation, the sensing state parameters can be normalized and converted into the same interval value. Specifically, each sensing state parameter has multiple values ​​at different times. Therefore, for each sensing state parameter, multiple values ​​can be normalized to obtain multiple normalized sensing state parameters.

[0063] 207. Divide the sensor into multiple preset intervals based on multiple normalized sensor state parameters.

[0064] In this embodiment of the application, for a sensing state parameter, there are multiple normalized sensing state parameters. Then, the changes between these multiple normalized sensing state parameters can be determined, and the distribution of each normalized sensing state parameter in each interval can be determined. Therefore, multiple preset intervals can be divided according to the multiple normalized sensing state parameters, and the range of each preset interval is equal.

[0065] It should be noted that the number of preset intervals can be determined according to the needs. The more preset intervals there are, the more accurately the normalized sensor state parameters will be divided, and the more accurate the subsequent calculation of weight factors will be.

[0066] In some embodiments, the range of the preset interval can be determined based on the specific values ​​of the normalized sensing state parameters. For example, for a sensing state parameter, there are 10 normalized sensing state parameters at different times, namely: 0.10, 0.13, 0.15, 0.09, 0.185, 0.20, 0.24, 0.21, 0.06, and 0.22. It can be seen that these normalized sensing state parameters are concentrated in the range of 0.06 to 0.24. Then, the range of 0.06 to 0.24 can be divided into multiple intervals. Suppose there are 6 intervals: 0.06 to 0.09, 0.09 to 0.12, 0.12 to 0.15, 0.15 to 0.18, 0.18 to 0.21, and 0.21 to 0.24. Then, it is possible to count which interval each of the above 10 normalized sensing state parameters falls into.

[0067] 208. Determine the probability distribution corresponding to each preset interval based on the preset intervals in which each normalized sensing state parameter is located.

[0068] In this embodiment of the application, after dividing multiple preset intervals, the preset interval in which each normalized sensing state parameter is located can be determined. Then, the distribution probability corresponding to each preset interval is calculated by dividing the number of normalized sensing state parameters included in each preset interval by the total number of normalized sensing state parameters.

[0069] For example, the normalized sensor state parameters at 10 time points are: 0.10, 0.13, 0.15, 0.09, 0.185, 0.20, 0.24, 0.21, 0.06, and 0.22. These are divided into six intervals: 0.06~0.09, 0.09~0.12, 0.12~0.15, 0.15~0.18, 0.18~0.21, and 0.21~0.24. It can be seen that there is a value of 0.006 in the interval 0.06~0.09, with a probability distribution... The probability distribution is 0.1; in the interval 0.09~0.12, there are two values, 0.09 and 0.10, with a probability distribution of 0.2; in the interval 0.12~0.15, there is one value, 0.13, with a probability distribution of 0.1; in the interval 0.15~0.18, there is one value, 0.15, with a probability distribution of 0.1; in the interval 0.18~0.21, there are two values, 0.185 and 0.20, with a probability distribution of 0.2; in the interval 0.21~0.24, there are three values, 0.21, 0.22 and 0.24, with a probability distribution of 0.3.

[0070] 209. Based on the probability distribution, determine the dynamic weighting factors corresponding to each normalized sensing state parameter.

[0071] In this embodiment of the application, after determining the distribution probability of each preset interval, the changes in the sensing state parameters over a period of time can be known. Therefore, by combining the distribution probability of each sensing state parameter corresponding to each preset interval, the dynamic weight factor corresponding to each normalized sensing state parameter can be determined. This dynamic weight factor is used to reflect the fluctuation of the sensing state parameters.

[0072] In some embodiments, the dynamic weighting factor corresponding to each normalized sensing state parameter is determined according to the probability distribution. Specifically, this may include: calculating the information entropy of each normalized sensing state parameter according to the probability distribution; and determining the dynamic weighting factor of each normalized sensing state parameter based on the exponential decay function and the information entropy of each subsystem.

[0073] It should be noted that information entropy can be used to visually represent the degree of dispersion of normalized sensor state parameters through numerical values, and can be expressed as:

[0074] in, This refers to the information entropy of the normalized sensor state parameters. This refers to the distribution probability of the normalized sensor state parameters falling into each preset interval. There are a total of b intervals, where i represents each normalized sensor state parameter.

[0075] It should be noted that the higher the information entropy, the more drastic the fluctuation of the normalized sensing state parameter within the first time period, and the lower the data reliability. Therefore, the corresponding dynamic weighting factor should be set to a lower value. In other words, the dynamic weighting factor is negatively correlated with the information entropy. This dynamic weighting factor can be determined using an exponential decay function.

[0076] in, This refers to the dynamic weighting factor for the normalized sensor state parameters. This refers to the information entropy of the normalized sensor state parameters, where i represents each normalized sensor state parameter. The preset sensitivity adjustment parameter can be set by the user based on the attribute information of the sensing state parameter.

[0077] 210. Based on the dynamic weighting factors corresponding to each normalized sensing state parameter and the aging degree of each subsystem, the initial health index of each subsystem is calculated.

[0078] In this embodiment of the application, after determining the dynamic weighting factors corresponding to each normalized sensing state parameter, the initial health index of each subsystem can be determined based on the dynamic weighting factors of the normalized sensing state parameters corresponding to each subsystem and the aging degree corresponding to each subsystem.

[0079] In some embodiments, the initial health index of each subsystem is calculated based on the dynamic weighting factors corresponding to each normalized sensing state parameter and the aging degree of each subsystem. Specifically, this may include: determining at least one normalized sensing state parameter corresponding to each subsystem; and for the first subsystem, calculating the initial health index of the first subsystem based on at least one normalized sensing state parameter corresponding to the first subsystem, the dynamic weighting factor of at least one normalized sensing state parameter, and the aging degree of the first subsystem, wherein the first subsystem is any one of the subsystems.

[0080] It should be noted that each subsystem can correspond to at least one normalized sensor state parameter; the health index can be calculated based on the at least one normalized sensor state parameter corresponding to the subsystem, the dynamic weighting factor of each normalized sensor state parameter, and the aging degree of the subsystem, which can be specifically expressed as follows:

[0081] in, This is the initial health index. This refers to the dynamic weighting factor for the normalized sensor state parameters. That is, the normalized sensor state parameters. This refers to the degree of aging. The contribution coefficient to the degree of aging is denoted by i, which represents each normalized sensor state parameter. There are a total of m normalized sensor state parameters.

[0082] In this embodiment, the initial health index achieves a unified representation of the current physical state of the equipment and the cumulative effects of historical use by fusing normalized sensor state parameters with the aging degree perceived by operating conditions. The aging degree, as an irreversible degradation factor, is explicitly introduced into the health index, making up for the shortcomings of traditional methods that rely solely on instantaneous sensor data and ignore long-term performance degradation. In addition, a dynamic weight adjustment mechanism based on information entropy is introduced to effectively solve the problem of feature reliability fluctuation caused by noise interference, sensor drift, or occasional anomalies in complex industrial environments. By automatically reducing the weight of features with high information entropy and increasing the contribution of features with low information entropy, the health index calculation process has adaptive anti-interference capabilities, enhancing its robustness and stability in long-term operation on actual production lines.

[0083] 211. Input the initial health index of each subsystem into the preset process quality prediction model to obtain the process quality prediction index.

[0084] In this embodiment, the description of step 211 is the same as the detailed description of step 104 in the above embodiments, and will not be repeated in this embodiment.

[0085] 212. The difference between the process quality prediction index and the process quality parameter is allocated according to the ratio between the sensitivity coefficients of each subsystem to obtain the quality deviation of each subsystem.

[0086] In this embodiment, the process quality prediction index output by the preset process quality prediction model can be the process quality prediction index of the device under test. This process quality prediction index can be a weighted combination of placement offset, rejection rate, and yield. The difference between the process quality prediction index and the process quality parameters can be allocated to each subsystem according to the ratio between the sensitivity coefficients of each subsystem. The sensitivity coefficient can be used to indicate the degree of influence of the subsystem on the process quality. Generally speaking, for SMT equipment, the placement head assembly has a greater influence on the process quality, so the sensitivity coefficient of the placement head assembly can be set to a larger value, while the sensitivity coefficients of other subsystems can be set to smaller values.

[0087] 213. The initial health index is superimposed by the quality deviations of each subsystem to obtain the target health index of each subsystem.

[0088] In this embodiment of the application, after determining the quality deviation of each subsystem, the initial health index can be superimposed and corrected to obtain the target health index of each subsystem, which can be specifically expressed as:

[0089]

[0090] in, This is the target health index of the subsystem. This is the initial health index of the subsystem. This is the sensitivity coefficient of the subsystem. This is the difference between the predicted process quality indicators and the process quality parameters. These are the process quality parameters. This refers to the indicators for predicting process quality.

[0091] In this embodiment, the health index of the subsystem is corrected by process quality deviation in a closed loop, and a feedback loop between equipment status and process output is constructed. This allows the health status estimate to be indirectly inferred and corrected even if some internal degradation is not directly detected by sensors, through its impact on process results such as mounting accuracy and rejection rate. This mechanism breaks through the dependence of traditional prediction methods on the completeness of sensor coverage and improves the detection capability of latent faults and early degradation.

[0092] 214. Determine the global health index of the device under test based on the coupling strength between each subsystem and other subsystems and the target health index.

[0093] In this embodiment of the application, after determining the target health index of each subsystem, it is necessary to fuse the target health indices of each subsystem. During the fusion, the coupling strength between each subsystem and other subsystems needs to be considered to obtain the global health index of the device under test. The global health index can be used to indicate the overall health status of the entire device under test.

[0094] In some embodiments, the global health index of the device under test is determined based on the coupling strength between each subsystem and other subsystems and the target health index. Specifically, this may include: obtaining the functional coupling relationships between the various subsystems of the device under test and generating a coupling relationship matrix; for the second subsystem, calculating the coupling enhancement factor of the second subsystem based on the coupling strength between the second subsystem and other subsystems and the target health index, where the second subsystem is any one of the subsystems; and determining the global health index of the device under test based on the coupling enhancement factors of each subsystem and the target health index.

[0095] It should be noted that the functional coupling relationship between the various subsystems can be a relationship that has been set during the production process of the device under test. This functional coupling relationship can be obtained directly from the device information. By summarizing the coupling strength of the functional coupling relationship between each subsystem and other subsystems, a coupling relationship matrix can be obtained. Each element in the coupling relationship matrix represents the coupling strength between each subsystem.

[0096] Then, for each subsystem, based on the coupling strength between that subsystem and each other subsystem, the coupling enhancement factor of that subsystem can be calculated, which can be expressed as:

[0097] in, This is the coupling enhancement factor. This refers to the coupling strength between the k-th subsystem and the j-th subsystem. This is the target health index.

[0098] Then, after calculating the coupling enhancement factor of each subsystem, it can be multiplied by the target health index. After summing the products of each subsystem and normalizing them, the global health index of the device under test can be obtained.

[0099] 215. Generate a global health time series based on the global health index corresponding to multiple consecutive time points.

[0100] In this embodiment of the application, the global health index can be the one at the current moment. Then, the global health index at multiple consecutive moments in history can be determined and summarized to obtain the global health time series. The global health time series is based on consecutive moments.

[0101] 216. Input the global health time series into the preset health prediction model to obtain the health prediction results of the device under test.

[0102] In this embodiment of the application, after obtaining the global health time series, the global health time series can be input into the preset health prediction model to predict the health status of the device under test at future times, and obtain the health prediction result of the device under test.

[0103] In some embodiments, the preset health prediction model may be a trained long short-term memory neural network model.

[0104] In this embodiment, a coupling relationship matrix of the subsystem is introduced and a coupling enhancement factor is calculated, which realizes the leap from isolated subsystem evaluation to system-level risk transmission modeling. When the health status of a certain subsystem deteriorates, the risk impact on other strongly coupled subsystems is automatically amplified, thereby more realistically reflecting the collaborative degradation behavior of the overall system health status.

[0105] 217. Based on the health prediction results of the device under test, determine the health prediction index of the device under test at each time after the current time.

[0106] 218. When the health prediction index at the target time is less than or equal to the failure threshold and the health prediction index at the time before the target time is greater than the failure threshold, the remaining service life of the device under test is determined based on the time interval between the target time and the current time.

[0107] In this embodiment, the health prediction result can include the health prediction index of the device under test (DUT) at each time after the current time. A failure threshold can be set for the health prediction index. If the health prediction index at the target time is less than or equal to the failure threshold, it indicates that the DUT may fail at the target time, which is a time after the current time. Furthermore, if the health prediction index at the time before the target time is greater than the failure threshold, it indicates that the DUT may start to fail from the target time. In other words, the target time is the time when the health prediction index is greater than or equal to the failure threshold for the first time after the current time. Therefore, the time interval between the target time and the current time can be determined as the remaining service life of the DUT.

[0108] In some embodiments, the slope of the current global health index change can also be calculated. If the slope exceeds a preset warning value, it is determined to be a high-risk state and a high-level maintenance warning is triggered.

[0109] In the embodiments of this application, by combining the time-series prediction model with threshold judgment, not only can the remaining service life be accurately predicted, but also the gradual degradation and sudden deterioration can be distinguished, providing a basis for decision-making for differentiated maintenance strategies.

[0110] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method.

[0111] like Figure 3 As shown, embodiments of this application also provide a device health prediction apparatus, which may include: The acquisition module 301 is used to acquire the equipment data of the device under test, including equipment operating parameters, sensor status parameters and process quality parameters. The processing module 302 is used to determine the equipment operating parameters corresponding to each subsystem of the device under test, and to determine the aging degree of each subsystem based on the equipment operating parameters corresponding to each subsystem. The processing module 302 is also used to calculate the initial health index of each subsystem based on the sensing state parameters at multiple times within the first time period and the aging degree of each subsystem. The processing module 302 is also used to input the initial health index of each subsystem into the preset process quality prediction model to obtain the process quality prediction index. The processing module 302 is also used to update the initial health index based on the difference between the process quality prediction index and the process quality parameter, so as to obtain the target health index of each subsystem. The processing module 302 is also used to perform health prediction on the device under test based on the coupling strength between each subsystem and other subsystems and the target health index, and obtain the health prediction result of the device under test.

[0112] In some embodiments, the processing module 302 is specifically used to determine the load intensity of each subsystem based on the device operating parameters corresponding to each subsystem. Processing module 302 is specifically used to integrate the load intensity corresponding to multiple moments within the second time period to obtain the aging accumulation of each subsystem. The processing module 302 is specifically used to determine the aging degree of each subsystem based on the exponential decay function and the aging accumulation of each subsystem.

[0113] In some embodiments, the processing module 302 is specifically used to normalize the sensing state parameters at multiple times within a first time period to obtain multiple normalized sensing state parameters. The processing module 302 is specifically used to divide multiple preset intervals based on multiple normalized sensing state parameters, with each preset interval having an equal range; The processing module 302 is specifically used to determine the distribution probability corresponding to each preset interval based on the preset interval in which each normalized sensing state parameter is located. The processing module 302 is specifically used to determine the dynamic weighting factor corresponding to each normalized sensing state parameter according to the probability distribution. The dynamic weighting factor is used to reflect the fluctuation of the sensing state parameter. The processing module 302 is specifically used to calculate the initial health index of each subsystem based on the dynamic weighting factors corresponding to each normalized sensing state parameter and the aging degree of each subsystem.

[0114] In some embodiments, the processing module 302 is specifically used to calculate the information entropy of each normalized sensing state parameter based on the distribution probability. The processing module 302 is specifically used to determine the dynamic weighting factor of each normalized sensing state parameter based on the exponential decay function and the information entropy of each subsystem. The dynamic weighting factor is negatively correlated with the information entropy.

[0115] In some embodiments, the processing module 302 is specifically used to determine at least one normalized sensing state parameter corresponding to each subsystem. The processing module 302 is specifically used to calculate the initial health index of the first subsystem based on at least one normalized sensing state parameter corresponding to the first subsystem, the dynamic weighting factor of at least one normalized sensing state parameter, and the aging degree of the first subsystem. The first subsystem is any one of the various subsystems.

[0116] In some embodiments, the processing module 302 is specifically used to allocate the difference between the process quality prediction index and the process quality parameter according to the ratio between the sensitivity coefficients of each subsystem to obtain the quality deviation of each subsystem. The sensitivity coefficient is used to indicate the degree of influence of the subsystem on the process quality. The processing module 302 is specifically used to superimpose the initial health index based on the quality deviation of each subsystem to obtain the target health index of each subsystem.

[0117] In some embodiments, the processing module 302 is specifically used to determine the global health index of the device under test based on the coupling strength between each subsystem and other subsystems and the target health index. Processing module 302 is specifically used to generate a global health time series based on the global health index corresponding to multiple consecutive time points; The processing module 302 is specifically used to input the global health time series into the preset health prediction model to obtain the health prediction result of the device under test.

[0118] In some embodiments, the acquisition module 301 is specifically used to acquire the functional coupling relationship between the various subsystems of the device under test and generate a coupling relationship matrix, wherein each element in the coupling relationship matrix represents the coupling strength between the various subsystems. The processing module 302 is specifically used to calculate the coupling enhancement factor of the second subsystem based on the coupling strength between the second subsystem and other subsystems and the target health index. The second subsystem is any one of the subsystems. The processing module 302 is specifically used to determine the global health index of the device under test based on the coupling enhancement factor of each subsystem and the target health index.

[0119] In some embodiments, the processing module 302 is further configured to determine the health prediction index of the device under test at various times after the current time based on the health prediction result of the device under test. The processing module 302 is further configured to determine the remaining service life of the device under test based on the time interval between the target time and the current time when the health prediction index at the target time is less than or equal to the failure threshold and the health prediction index at the time before the target time is greater than the failure threshold.

[0120] In the embodiments of this application, the description of the features corresponding to the device health prediction device in the embodiments can be found in the relevant description of the device health prediction method in the embodiments, and will not be repeated here.

[0121] like Figure 4As shown, embodiments of this application also provide an electronic device, including a memory 401 and a processor 402, wherein the memory 401 stores a computer program and the processor 402 is configured to run the computer program to perform the steps in any of the above-described embodiments of the device health prediction method.

[0122] Embodiments of this application also provide a computer-readable storage medium storing a computer program, wherein the computer program is configured to execute the steps in any of the above-described embodiments of the device health prediction method when it is run.

[0123] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard disk, magnetic disk, or optical disk.

[0124] Embodiments of this application also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the above-described embodiments of the device health prediction method.

[0125] Embodiments of this application also provide another computer program product, including a non-volatile computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps in any of the above-described embodiments of the device health prediction method.

[0126] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0127] The foregoing has provided a detailed description of the process monitoring of a storage system provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only intended to help understand the method and core ideas of this application. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of the claims of this application.

Claims

1. A device health prediction method, comprising: The method includes: Acquire equipment data of the device under test, including equipment operating parameters, sensor status parameters, and process quality parameters; Determine the equipment operating parameters corresponding to each subsystem of the device under test, and determine the aging degree of each subsystem based on the equipment operating parameters corresponding to each subsystem. Based on the sensor state parameters at multiple moments within the first time period and the aging degree of each subsystem, the initial health index of each subsystem is calculated. The initial health index of each subsystem is input into a preset process quality prediction model to obtain process quality prediction indicators. The initial health index is updated based on the difference between the process quality prediction index and the process quality parameter to obtain the target health index of each subsystem. The functional coupling relationship between the various subsystems of the device under test is obtained, and a coupling relationship matrix is ​​generated. Each element in the coupling relationship matrix represents the coupling strength between the various subsystems. For the second subsystem, a coupling enhancement factor is calculated based on the coupling strength between the second subsystem and other subsystems and the target health index, wherein the second subsystem is any one of the subsystems. The global health index of the device under test is determined based on the coupling enhancement factor of each subsystem and the target health index. Generate a global health time series based on the global health index corresponding to multiple consecutive time points; The global health time series is input into a preset health prediction model to obtain the health prediction result of the device under test.

2. The method of claim 1, wherein, The step of determining the aging degree of each subsystem based on the equipment operating parameters corresponding to each subsystem includes: The load intensity of each subsystem is determined based on the equipment operating parameters corresponding to each subsystem. The cumulative aging amount of each subsystem is obtained by integrating the load intensity corresponding to multiple moments within the second time period. The degree of aging of each subsystem is determined based on the exponential decay function and the cumulative aging amount of each subsystem.

3. The method according to claim 1, characterized in that, The calculation of the initial health index of each subsystem based on the sensor state parameters at multiple moments within the first time period and the aging degree of each subsystem includes: The sensor state parameters at multiple times within the first time period are normalized to obtain multiple normalized sensor state parameters. Based on the multiple normalized sensing state parameters, multiple preset intervals are divided, and the range of each preset interval is equal. Based on the preset intervals in which each normalized sensing state parameter is located, determine the distribution probability corresponding to each preset interval; Based on the distribution probability, a dynamic weighting factor is determined for each normalized sensing state parameter, and the dynamic weighting factor is used to reflect the fluctuation of the sensing state parameter. The initial health index of each subsystem is calculated based on the dynamic weighting factors corresponding to each normalized sensing state parameter and the aging degree of each subsystem.

4. The method according to claim 3, characterized in that, The step of determining the dynamic weighting factor corresponding to each normalized sensing state parameter based on the distribution probability includes: Based on the distribution probability, the information entropy of each normalized sensing state parameter is calculated; Based on the exponential decay function and the information entropy of each subsystem, a dynamic weighting factor is determined for each normalized sensing state parameter, and the dynamic weighting factor is negatively correlated with the information entropy.

5. The method according to claim 3, characterized in that, The initial health index of each subsystem is calculated based on the dynamic weighting factors corresponding to each normalized sensing state parameter and the aging degree of each subsystem, including: Determine at least one normalized sensing state parameter corresponding to each of the subsystems; For the first subsystem, an initial health index is calculated based on at least one normalized sensor state parameter corresponding to the first subsystem, the dynamic weighting factor of the at least one normalized sensor state parameter, and the aging degree of the first subsystem. The first subsystem is any one of the various subsystems.

6. The method according to claim 1, characterized in that, The initial health index is updated based on the difference between the process quality prediction index and the process quality parameter to obtain the target health index of each subsystem, including: The difference between the process quality prediction index and the process quality parameter is allocated according to the ratio between the sensitivity coefficients of each subsystem to obtain the quality deviation of each subsystem. The sensitivity coefficient is used to indicate the degree of influence of the subsystem on the process quality. The initial health index is obtained by superimposing the quality deviations of each subsystem onto the initial health index.

7. The method according to claim 1, characterized in that, After performing a health prediction on the device under test based on the coupling strength between each subsystem and other subsystems and the target health index, and obtaining the health prediction result of the device under test, the method further includes: Based on the health prediction results of the device under test, the health prediction index of the device under test at each time after the current time is determined. When the health prediction index at the target time is less than or equal to the failure threshold and the health prediction index at the time preceding the target time is greater than the failure threshold, the remaining service life of the device under test is determined based on the time interval between the target time and the current time.

8. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of the device health prediction method as described in any one of claims 1 to 7 when executing the computer program.

Images