Self-adaptive ARIMA modeling method and system for predictive maintenance of semiconductor manufacturing equipment

Abstract

The invention relates to the technical field of semiconductor manufacturing equipment predictive maintenance, and particularly discloses a self-adaptive ARIMA modeling method and system for semiconductor manufacturing equipment predictive maintenance. The method comprises the following steps: collecting process parameter data in real time, such as etching rate, temperature or pressure; the size of an ARIMA model training window is dynamically adjusted based on data volatility, and the window is automatically enlarged or shrunk to optimize model fitting by calculating a rolling standard deviation and comparing the rolling standard deviation with a self-adaptive threshold value based on a historical stability standard deviation; fitting an ARIMA model by using the adjusted window data; and predicting by using the model and generating an equipment state early warning signal. According to the method, the problems of prediction lag and false alarm caused by the fact that a fixed parameter ARIMA model cannot adapt to dynamic degradation of equipment are solved, early fault early warning is achieved, and the equipment operation efficiency and the product yield are improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of semiconductor manufacturing, and particularly relates to a self-adaptive ARIMA modeling method and system for predictive maintenance of semiconductor manufacturing equipment. BACKGROUND

[0002] In the field of semiconductor manufacturing, the process chambers of key equipment such as etching, deposition, and lithography will degrade in performance as the running time accumulates, such as radio frequency power attenuation, gas jet blockage, or accumulation of chamber wall deposits, resulting in drift or mutation of process parameters (such as etching rate, temperature, or pressure). In the prior art, equipment maintenance usually relies on regular maintenance or statistical process control (SPC) threshold alarms, but these methods have a lag and can only trigger an alarm when the parameter is significantly outside the control limit (such as ±3σ), at which point a large number of unqualified wafers may have already been produced. At the same time, the false positive rate is high and cannot effectively distinguish between normal slow drift and sudden abnormalities, resulting in unnecessary equipment downtime. In order to further improve the prediction accuracy, the prior art also uses a time series model with fixed parameters (such as the ARIMA model) for prediction control, but the training window size and parameters of the ARIMA model are often preset to fixed values and cannot adapt to the time series characteristics of dynamic changes in process parameters during equipment degradation. When the equipment state changes, the ARIMA model with a fixed window will gradually decrease in prediction accuracy due to outdated data or noise interference, making it difficult to achieve early fault warning, which is a major technical problem in the predictive maintenance of semiconductor manufacturing equipment. SUMMARY

[0003] The present application aims to solve the problem of inaccurate prediction caused by the inability of fixed parameter models to adapt to dynamic changes by adaptively adjusting the training window of the ARIMA model. In view of the above deficiencies in the prior art, the present application provides a self-adaptive ARIMA modeling method, system, device, and medium for predictive maintenance of semiconductor manufacturing equipment, which significantly improves the adaptability and accuracy of the ARIMA model in the predictive maintenance of semiconductor manufacturing equipment by introducing a dynamic window adjustment mechanism based on the volatility of process parameter data.

[0004] The first aspect of the present application discloses a self-adaptive ARIMA modeling method for predictive maintenance of semiconductor manufacturing equipment, comprising the following steps: Step S101: Real-time acquisition of process parameter data of semiconductor manufacturing equipment, the process parameter data including at least one of etching rate, temperature, or pressure; Step S102: Dynamically adjusting the training window size of the ARIMA model based on the volatility of the process parameter data, wherein the dynamic adjustment includes: calculating the rolling standard deviation σ of recent data roll where the rolling standard deviation σ rollis a standard deviation of the latest m data points, m is a preset positive integer, and m ranges from 5 to 20; σ roll is compared with an upper threshold Th high and a lower threshold Th low , where Th high =k1×σ process , Th low =k2×σ process , k1 and k2 are preset coefficients greater than 0, k1 ranges from 0.8 to 1.5, k2 ranges from 0.1 to 0.5, σ process is a historical stable standard deviation of the process parameter in a stable running stage of the equipment; and The size of the training window is adjusted according to the comparison result of the rolling standard deviation σ roll with the upper threshold Th high and the lower threshold Th low , including: If σ roll >Th high , the size of the training window is expanded to min(W old ×a, W max ), where W old is a current window size, a is a preset coefficient greater than 1, a ranges from 1.05 to 1.2, W max is a preset maximum window size, and W max ranges from 200 to 500; If σ roll <Th low , the size of the training window is reduced to max(W old ×b, W min ), where b is a preset coefficient less than 1, b ranges from 0.8 to 0.95, W min is a preset minimum window size, and W min ranges from 20 to 50; If Th low ≤σ roll ≤Th high , the size of the training window is kept unchanged; Step S103: using the process parameter data in the adjusted training window, fitting an ARIMA model, the ARIMA model including an autoregressive order p, a difference order d, and a moving average order q; Step S104: using the fitted ARIMA model to predict the next process parameter data, and generating an equipment state early warning signal based on the prediction result.

[0005] The adaptive ARIMA modeling method further comprises a step S105 of monitoring the change of the ARIMA model parameters after the step S103, wherein the monitoring comprises: calculating the difference order d of the currently fitted ARIMA model current of the last fitted model previous , and the difference Δd = |d current -d previous | is calculated. Comparing the Δd with a control upper limit UCL, wherein the UCL = c x σ delta , c is a preset coefficient greater than 0, the value range of c is 2 to 4, and σ delta is the standard deviation of the historical d value change sequence of the device health stable stage, and the device health stable stage is determined by the time period in which the process parameter fluctuation in the historical data is less than a preset threshold. If the Δd > UCL, it is determined that the parameter mutates, and a parameter mutation signal is generated.

[0006] The adaptive ARIMA modeling method further comprises that, in the step S105, if the Δd changes in the same direction in three consecutive calculations, and each Δd is greater than k x σ delta , wherein k is a preset coefficient, the value range of k is 1 to 2, it is determined that the parameter mutates, and a parameter mutation signal is generated.

[0007] The adaptive ARIMA modeling method, wherein the parameter mutation signal is used to trigger fault diagnosis, and the fault diagnosis comprises: analyzing the device operation data corresponding to the parameter mutation signal, and checking whether other process parameters have abnormal fluctuations in the same period; based on the analysis result, associating the device degradation type to determine the root cause.

[0008] The adaptive ARIMA modeling method further comprises a step S106 of analyzing the statistical characteristics of the prediction residual and matching with a fault mode library after the step S104, wherein the analysis comprises: calculating the residual of the prediction value and the actual value, and calculating the statistical characteristics of the residual, including at least one of the variance, the mean, the skewness or the kurtosis; comparing the statistical characteristics with a mapping rule in a preset fault mode library, the fault mode library is trained based on historical fault data, and the mapping rule includes the corresponding relationship between the fault type and the residual statistical characteristics; determining the fault type and the alarm level according to the matching result.

[0009] The adaptive ARIMA modeling method, wherein the fitting of the ARIMA model in the step S103 comprises: Smoothness test is performed on the process parameter data in the training window, an extended Dickey-Fuller test is used, if the p value is greater than 0.05, difference operation is performed until the sequence is smooth, and the difference order is denoted as d; On the smooth sequence, the optimal autoregressive order p and moving average order q are determined through autocorrelation function and partial autocorrelation function analysis, or using the auto_arima algorithm for grid search.

[0010] The adaptive ARIMA modeling method, the equipment state early warning signal includes a warning level and a severity level, after the equipment state early warning signal is generated, is sent to the equipment maintenance system or the engineer terminal through the interface.

[0011] The second aspect of the application discloses an adaptive ARIMA modeling system for predictive maintenance of a semiconductor manufacturing equipment, comprising: The data acquisition module is configured to acquire process parameter data of the semiconductor manufacturing equipment in real time, and the process parameter data comprises at least one of etching rate, temperature or pressure; The window adjustment module is configured to dynamically adjust the training window size of the ARIMA model based on the volatility of the process parameter data, wherein the window adjustment module comprises: The volatility calculation unit is configured to calculate the rolling standard deviation σ roll of the recent data roll , wherein the rolling standard deviation σ roll is the standard deviation of the last m data points, and m is a preset positive integer; The threshold comparison unit is configured to compare σ high with the upper threshold Th low and the lower threshold Th high , wherein Th process =k1×σ low , Th process =k2×σ process , k1 and k2 are preset coefficients greater than 0, and σ roll is the historical stable standard deviation of the process parameter in the stable running stage of the equipment; The window control unit is configured to adjust the training window size according to the comparison results of the rolling standard deviation σ high and the upper threshold Th low and the lower threshold Th roll , comprising: If σ high >Th old , the training window size is expanded to min(W max ×a, W old ), wherein W maxpreset maximum window size; if σ roll <Th low , then the training window size is reduced to max(W old ×b, W min ), where b is a preset coefficient less than 1, and W min is a preset minimum window size; if Th low ≤ σ roll ≤ Th high , then the training window size is kept unchanged; a model fitting module configured to fit an ARIMA model using the process parameter data in the adjusted training window, the ARIMA model including an autoregressive order p, a difference order d, and a moving average order q; a prediction and early warning module configured to predict next process parameter data using the fitted ARIMA model, and generate a device state early warning signal based on the prediction result.

[0012] A third aspect of the present application discloses an electronic device, comprising a memory and a processor, the processor and the memory are connected; the memory is used for storing programs; the processor calls the programs stored in the memory to execute the method provided in the first aspect and / or any possible embodiment in combination with the first aspect.

[0013] A fourth aspect of the present application discloses a computer readable storage medium, which stores a computer program, and the computer program is executed by a computer to execute the method provided in the first aspect and / or any possible embodiment in combination with the first aspect.

[0014] Compared with the prior art, the present application has the following advantages: by calculating the rolling standard deviation of recent data and comparing it with the adaptive threshold based on the historical stable standard deviation, the training window size can be optimized in real time: when the process fluctuation is large, the window is automatically enlarged to smooth the noise and enhance the model stability; when the fluctuation is small, the window is reduced to quickly respond to small changes and improve the model sensitivity. This adaptive adjustment effectively overcomes the problem of decreased prediction accuracy of the fixed parameter ARIMA model due to the inability to adapt to the dynamic timing characteristics in the device degradation process, thereby reducing the alarm lag and false alarm rate in the traditional method, achieving early warning of device failure, reducing unplanned downtime and yield loss.

[0015] The technical solutions of the present application will be further described in detail below with reference to the drawings and examples. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 Method flowchart for Example 1.

[0017] Figure 2 System module chart for Example 2. DETAILED DESCRIPTION

[0018] Example 1 As shown in the figure, an adaptive ARIMA modeling method for predictive maintenance of semiconductor manufacturing equipment includes the following steps: Figure 1 Step S101: Real-time collection of process parameter data of the semiconductor manufacturing equipment, the process parameter data including at least one of etching rate, temperature or pressure; In implementation, the etching rate data is continuously collected by a sensor integrated in the plasma etching machine chamber, with each wafer processing completion as a trigger point, in units of angstroms per minute. The collected data is transmitted to a central database in real time to ensure the integrity of the time series. For example, when the etching machine processes the 100th wafer, the sensor measures the etching rate as 500 angstroms per minute, and the system immediately records and stores the value. By collecting process parameter data in real time, continuous monitoring of the equipment running state is realized, providing an accurate data basis for subsequent analysis. Step S102: Dynamic adjustment of the training window size of the ARIMA model based on the volatility of the process parameter data, wherein the dynamic adjustment includes:

[0019] Calculation of the rolling standard deviation σ roll of recent data roll , where σ roll is the standard deviation of the last m data points, m is a preset positive integer, and the value range of m is 5 to 20; Comparison of σ high with an upper threshold Th low and a lower threshold Th high , where Th process =k1×σ low , Th process =k2×σ process , k1 and k2 are preset coefficients greater than 0, the value range of k1 is 0.8 to 1.5, the value range of k2 is 0.1 to 0.5, and σ roll is the historical stable standard deviation of the process parameter in the stable running stage of the equipment; Adjustment of the training window size according to the comparison results of the rolling standard deviation σ high and the upper threshold Th low and the lower threshold Th , including: roll If σ high >Th old , then expand the training window size to min(Wold ×a,W max ), wherein W old is a preset maximum window size, a is a preset coefficient greater than 1, a has a value range of 1.05 to 1.2, W max is a preset minimum window size, and W max has a value range of 20 to 50; If σ roll < Th low , the training window size is reduced to max(W old ×b,W min ), wherein b is a preset coefficient less than 1, b has a value range of 0.8 to 0.95, W min is a preset maximum window size, and W min has a value range of 200 to 500; If Th low ≤ σ roll ≤ Th high , the training window size is kept unchanged; The implementation includes three sub-steps: first, the rolling standard deviation σ roll of recent data is calculated, wherein the rolling window size m is set to 10, i.e., the standard deviation is calculated based on the last 10 etching rate data points to reflect short-term volatility. For example, the system extracts the etching rate values of the 91st to 100th wafer from the database, and calculates σ roll = 0.6 angstrom per minute. Second, σ roll is compared with the upper threshold Th high and the lower threshold Th low ; Th high is calculated by multiplying the historical stable standard deviation σ process by a coefficient k1 = 1.0, and Th low is calculated by multiplying σ process by a coefficient k2 = 0.3, wherein σ process is determined to be 0.5 angstrom per minute based on historical data of the stable running stage of the device. Therefore, Th high = 0.5, and Th low = 0.15. Finally, the training window size is adjusted according to the comparison result: if σ roll is greater than Th high , the window is expanded to a value obtained by multiplying the current window size W old by a coefficient a = 1.1, and does not exceed the maximum window size W max = 300; if σ roll is less than Th low , the window is reduced to W oldmultiplied by a coefficient b = 0.9, and is not less than a minimum window size Wmin = 30; if σroll is between the thresholds, the window is kept unchanged. For example, when σ roll = 0.6 is greater than Th high = 0.5, the system expands the training window from 100 wafers to 110 wafers. By dynamically adjusting the training window size, the adaptive matching of model training data volume and current process volatility is realized, so as to smooth the noise by expanding the window when the data fluctuation is large, and to enhance the sensitivity to small changes by reducing the window when the data is stable.

[0020] Step S103: using the process parameter data in the adjusted training window, fitting an ARIMA model, the ARIMA model including an autoregressive order p, a difference order d, and a moving average order q; In implementation, the etching rate data sequence in the current training window, for example 110 data points, is extracted from the database, and first, stationarity test is performed: the p value is calculated by applying the extended Dickey-Fuller test, if the p value is greater than 0.05, the sequence is first-order differentiated, and the test is repeated until the p value is less than 0.05, and the difference order is recorded as the difference order d. For example, the p value of the initial sequence is 0.8, which is non-stationary; after first-order differentiation, the p value is 0.01, which is stationary, so d = 1. Then, on the stationary sequence, the optimal autoregressive order p and moving average order q are determined by analyzing the truncation characteristics of the autocorrelation function and the partial autocorrelation function, or calling the Auto_Arim algorithm for grid search, with the minimum AIC as the target. For example, after analysis, p = 1 and q = 1 are determined, so that an ARIMA(1,1,1) model is fitted. By fitting the ARIMA model, the dynamic accurate capture of the process parameter time sequence is realized, providing a reliable base model for prediction.

[0021] Step S104: using the fitted ARIMA model to predict the next process parameter data, and generating a device state warning signal based on the prediction result.

[0022] In implementation, the fitted model is used to predict the etching rate of the next wafer, for example, the predicted value of the 111th wafer is 505 angstroms per minute; after the actual processing is completed, the actual value is recorded and the prediction residual is calculated. If the residual exceeds the threshold set based on the historical residual standard deviation (such as ± 2 times the standard deviation), a device state warning signal is generated, for example, a warning level signal, and is sent to the device maintenance system through the application programming interface. For example, the actual etching rate is 510 angstroms per minute, and the residual is 5 angstroms per minute, which exceeds the threshold, and the system triggers a warning signal. Through real-time prediction and residual monitoring, early warning of potential device failure is realized, thereby reducing unplanned downtime and improving production efficiency.

[0023] In one embodiment, after step S103, further comprising step S105: monitoring the change of the ARIMA model parameters, wherein the monitoring comprises: calculating the difference order d of the current fitted ARIMA model current of the last fitted model previous ; current = |d previous - d delta |; and comparing Ad with a control upper limit UCL, wherein UCL = c x s delta , c is a preset coefficient greater than 0, c ranges from 2 to 4, and s current is the standard deviation of the historical d value change sequence in the device health stable stage, which is determined by the time period in the historical data in which the process parameter fluctuation is less than a preset threshold; if Ad > UCL, determining that the parameter mutates and generating a parameter mutation signal.

[0024] In implementation, after fitting a new ARIMA model each time, the current difference order d previous is extracted from the model parameters and compared with the difference order d current of the last model, and the absolute value difference Ad = |d previous - d previous | is calculated. For example, the last model d current = 1 and the current model d delta = 2, then Ad = 1. The control upper limit UCL is calculated by multiplying the coefficient c = 3 by the standard deviation s delta of the historical d value change; s delta is determined based on the historical data in the device health stable stage, for example, the system analyzes the d value change sequence during the processing of the last 1000 wafers, and calculates s delta = 0.2, so UCL = 0.6. If Ad > UCL, it is determined that the parameter mutates and a parameter mutation signal is generated. By monitoring the mutation of the difference order d, early identification of step degradation of the device performance is achieved, so that early warning is triggered in the case of accelerated chamber wall deposition and the like, avoiding batch wafer loss.

[0025] In one embodiment, the step S105 further comprises: if Ad changes in the same direction in three consecutive calculations, and each Ad is greater than k x s delta , wherein k is a preset coefficient, k ranges from 1 to 2, it is determined that the parameter mutates and a parameter mutation signal is generated.

[0026] During implementation, record the Δd values ​​for three consecutive model fits and check if they all change in the same direction. For example, if the three consecutive Δd values ​​are 0.3, 0.4, and 0.5, they all increase positively. Simultaneously, each Δd must be greater than k×σ. delta Where k is set to 1.5, σ delta =0.2, therefore the threshold is 0.3; since 0.3, 0.4, and 0.5 are all greater than 0.3 and change in the same direction, the system determines it as a parameter mutation and generates a signal. Through continuous trend analysis, reliable detection of gradual degradation is achieved, false alarms caused by single data fluctuations are reduced, and the stability of mutation determination is improved.

[0027] In one embodiment, the parameter mutation signal is used to trigger fault diagnosis, wherein the fault diagnosis includes: Analyze the equipment operation data corresponding to the parameter mutation signal and check whether there are abnormal fluctuations in other process parameters during the same period; Based on the analysis results, the type of equipment degradation is correlated to determine the root cause.

[0028] During implementation, upon generating a parameter mutation signal, the system automatically correlates it with equipment operating data from the same period, checking for abnormal fluctuations in other process parameters such as RF power and gas flow. The system extracts data from 10 wafers before and after the mutation point from the database and calculates the rolling standard deviation of these parameters. If the standard deviation of RF power increases from 0.1 to 0.5, it correlates this with a pre-defined fault type, such as RF power degradation, based on the d-value mutation. Through multi-parameter collaborative analysis, the system enables rapid localization of the root cause of the fault, guiding maintenance personnel to clean or replace components in a targeted manner, thereby improving maintenance efficiency.

[0029] In one embodiment, after step S104, a step S106 is further included: analyzing the statistical characteristics of the predicted residuals and matching them with a failure mode library, wherein the analysis includes: Calculate the residual between the predicted value and the actual value, and calculate the statistical characteristics of the residual, including at least one of variance, mean, skewness or kurtosis; The statistical features are compared with the mapping rules in the preset fault mode library, which is trained based on historical fault data, and the mapping rules include the correspondence between fault types and residual statistical features. Based on the matching results, determine the fault type and alarm level.

[0030] During implementation, the residual between the predicted and actual values ​​is calculated. For example, for the 111th wafer, the predicted etching rate is 505 Å per minute, the actual rate is 510 Å per minute, and the residual is 5 Å per minute. Then, the statistical characteristics of the residual sequence are calculated, including variance, mean, skewness, and kurtosis. For example, the variance of the last 20 residuals increases from 0.1 to 0.5, and the mean shows a positive shift of 0.3. The fault mode library is trained based on historical fault data and includes mapping rules such as increased variance corresponding to "gas nozzle blockage" and a positive shift in the mean corresponding to "chamber wall deposition." The system compares the calculated features with the rules in the library. If a match is found, the fault type is determined to be gas nozzle blockage, and a warning level alarm is triggered. Through residual feature analysis, refined fault type identification is achieved, thereby providing specific maintenance recommendations and reducing unnecessary equipment downtime.

[0031] In one embodiment, fitting the ARIMA model in step S103 includes: The stationarity of the process parameter data within the training window is tested using the extended Dickey-Fuller test. If the p-value is greater than 0.05, the difference operation is performed until the sequence is stationary. The difference order is denoted as d. For stationary sequences, the optimal autoregression order p and moving average order q can be determined through autocorrelation function and partial autocorrelation function analysis, or by using the Auto_Arim algorithm for grid search.

[0032] In implementation, the stationarity of the etching rate data within the training window is tested using the extended Dickey-Fuller test to calculate the p-value. If the p-value is greater than 0.05, it indicates that the sequence is non-stationary, so first-order differencing is performed, and the test is repeated until the p-value is less than 0.05. The differencing order is denoted as d. For example, if the initial sequence p-value is 0.8, the p-value after first-order differencing is 0.01, therefore d=1. For stationary sequences, truncation features are analyzed using autocorrelation and partial autocorrelation functions, or a grid search is performed using the auto_arima algorithm, with the goal of minimizing the Akaike information criterion, to determine the optimal p and q values, such as p=1 and q=1. The standardized fitting process ensures the model's adaptability to non-stationary data, improving the accuracy and robustness of predictions.

[0033] In one embodiment, the device status warning signal includes a warning level and a severity level. After the device status warning signal is generated, it is sent to the device maintenance system or engineer terminal through an interface.

[0034] During implementation, alarm levels are categorized based on the degree of abrupt changes in predicted residuals or parameters; for example, residuals exceeding twice the historical residual standard deviation are set to a warning level, and those exceeding three times are set to a critical level. Signals are sent in real-time to the equipment maintenance system or engineer's terminal via a REST API, with the JSON format including timestamp, device ID, fault type, and level. Through tiered alarms and real-time transmission, priority management of maintenance responses is achieved, thereby optimizing resource allocation and shortening fault handling time.

[0035] Example 2 like Figure 2 As shown, an adaptive ARIMA modeling system for predictive maintenance of semiconductor manufacturing equipment includes: The data acquisition module is configured to acquire process parameter data of semiconductor manufacturing equipment in real time, wherein the process parameter data includes at least one of etching rate, temperature or pressure. A window adjustment module is configured to dynamically adjust the training window size of the ARIMA model based on the volatility of the process parameter data. The window adjustment module includes: The volatility calculation unit is configured to calculate the rolling standard deviation σ of recent data. roll The rolling standard deviation σ roll It is the standard deviation of the most recent m data points, where m is a preset positive integer; The threshold comparison unit is configured to compare σ roll With upper limit threshold Th high and lower limit threshold Th low Comparison, where Th high =k1×σ process ,Th low =k2×σ process k1 and k2 are preset coefficients greater than 0, σ process The historical stability standard deviation of the process parameters during the stable operation phase of the equipment; Window control unit, configured to operate based on rolling standard deviation σ roll With upper limit threshold Th high and lower limit threshold Th low Adjusting the training window size based on the comparison results includes: If σ roll >Th high Then increase the training window size to min(W) old ×a,W max ), where W old Where W is the current window size, 'a' is a preset coefficient greater than 1, and W is the current window size. max Set the maximum window size to the preset value; If σ roll <Th low Then reduce the training window size to max(W)old ×b,W min ), where b is a preset coefficient less than 1, W min Set the minimum window size; If Th low ≤σ roll ≤Th high If so, the training window size remains unchanged; The model fitting module is configured to use the process parameter data within the adjusted training window to fit an ARIMA model, wherein the ARIMA model includes an autoregressive order p, a difference order d, and a moving average order q. The prediction and early warning module is configured to use the fitted ARIMA model to predict the next process parameter data and generate equipment status early warning signals based on the prediction results.

[0036] The adaptive ARIMA modeling system for predictive maintenance of semiconductor manufacturing equipment provided in this embodiment has the same implementation principle and technical effect as the method embodiment in Embodiment 1. For the sake of brevity, any parts not mentioned in the system embodiment can be referred to the corresponding content in Embodiment 1.

[0037] Example 3 A computer-readable storage medium having a computer program stored thereon, the computer program being executed by a computer to perform the adaptive ARIMA modeling method for predictive maintenance of semiconductor manufacturing equipment as described in Embodiment 1 above.

[0038] Example 4 An electronic device includes: a memory and a processor, wherein the processor and the memory are connected; The memory is used to store programs; The processor invokes a program stored in the memory to execute an adaptive ARIMA modeling method for predictive maintenance of semiconductor manufacturing equipment as described in Embodiment 1.

[0039] It should be noted that the electronic device mentioned may be, but is not limited to, personal computers (PCs), tablet computers, mobile internet devices (MIDs), etc.

[0040] It should be noted that processors, memory, and other components that may be present in electronic devices are electrically connected to each other, directly or indirectly, to enable data transmission or interaction. For example, processors, memory, and other components may be electrically connected to each other via one or more communication buses or signal lines.

[0041] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0042] In the several embodiments provided in this application, it should be understood that the disclosed systems and methods can also be implemented in other ways. The system embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0043] In addition, the functional modules in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0044] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, laptop, server, mobile phone, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0045] The above description is merely a preferred embodiment of the present invention and does not constitute any limitation on the present invention. Any simple modifications, alterations, or equivalent structural changes made to the above embodiments based on the technical essence of the present invention shall still fall within the protection scope of the present invention.

Claims

1. An adaptive ARIMA modeling method for predictive maintenance of semiconductor manufacturing equipment, characterized in that, Includes the following steps: Step S101: Real-time acquisition of process parameter data of semiconductor manufacturing equipment, wherein the process parameter data includes at least one of etching rate, temperature or pressure; Step S102: Based on the volatility of the process parameter data, dynamically adjust the training window size of the ARIMA model, wherein the dynamic adjustment includes: Calculate the rolling standard deviation σ of recent data roll The rolling standard deviation σ roll It is the standard deviation of the most recent m data points, where m is a preset positive integer; σ roll With upper limit threshold Th high and lower limit threshold Th low Comparison, where Th high =k1×σ process ,Th low =k2×σ process k1 and k2 are preset coefficients greater than 0, σ process The historical stability standard deviation of the process parameters during the stable operation phase of the equipment; Based on the rolling standard deviation σ roll With upper limit threshold Th high and lower limit threshold Th low Adjusting the training window size based on the comparison results includes: If σ roll >Th high Then increase the training window size to min(W) old ×a,W max ), where W old Where W is the current window size, 'a' is a preset coefficient greater than 1, and W is the current window size. max Set the maximum window size to the preset value; If σ roll <Th low Then reduce the training window size to max(W) old ×b,W min ), where b is a preset coefficient less than 1, W min Set the minimum window size; If Th low ≤σ roll ≤Th high If so, the training window size remains unchanged; Step S103: Using the process parameter data within the adjusted training window, fit the ARIMA model, wherein the ARIMA model includes the autoregressive order p, the difference order d, and the moving average order q; Step S104: Use the fitted ARIMA model to predict the next process parameter data, and generate equipment status warning signals based on the prediction results.

2. The adaptive ARIMA modeling method as described in claim 1, characterized in that, Following step S103, step S105 is also included: monitoring changes in ARIMA model parameters, wherein monitoring includes: Calculate the difference order d of the currently fitted ARIMA model. current The difference order d with the previous fitted model previous The difference Δd = |d current -d previous |; Compare Δd with the control upper limit UCL, where UCL = c × σ delta c is a preset coefficient greater than 0, σ delta The standard deviation of the historical d-value change sequence during the equipment health and stability phase is defined by the time period in historical data when the fluctuation of process parameters is less than a preset threshold. If Δd > UCL, it is determined to be a parameter mutation, and a parameter mutation signal is generated.

3. The adaptive ARIMA modeling method as described in claim 2, characterized in that, Step S105 further includes: if in three consecutive calculations, Δd changes in the same direction, and each Δd is greater than k×σ delta If k is a preset coefficient, then it is determined to be a parameter mutation, and a parameter mutation signal is generated.

4. The adaptive ARIMA modeling method as described in claim 2, characterized in that, The parameter mutation signal is used to trigger fault diagnosis, wherein the fault diagnosis includes: Analyze the equipment operation data corresponding to the parameter mutation signal and check whether there are abnormal fluctuations in other process parameters during the same period; Based on the analysis results, the type of equipment degradation is correlated to determine the root cause.

5. The adaptive ARIMA modeling method as described in claim 1, characterized in that, Following step S104, step S106 is also included: analyzing the statistical characteristics of the predicted residuals and matching them with a failure mode library, wherein the analysis includes: Calculate the residual between the predicted value and the actual value, and calculate the statistical characteristics of the residual, including at least one of variance, mean, skewness or kurtosis; The statistical features are compared with the mapping rules in the preset fault mode library, which is trained based on historical fault data, and the mapping rules include the correspondence between fault types and residual statistical features. Based on the matching results, determine the fault type and alarm level.

6. The adaptive ARIMA modeling method as described in claim 1, characterized in that, In step S103, fitting the ARIMA model includes: The stationarity of the process parameter data within the training window is tested using the extended Dickey-Fuller test. If the p-value is greater than 0.05, the difference operation is performed until the sequence is stationary. The difference order is denoted as d. For stationary sequences, the optimal autoregression order p and moving average order q can be determined through autocorrelation function and partial autocorrelation function analysis, or by using the auto_arima algorithm for grid search.

7. The adaptive ARIMA modeling method as described in claim 1, characterized in that, The equipment status warning signal includes a warning level and a severity level. After the equipment status warning signal is generated, it is sent to the equipment maintenance system or engineer terminal through an interface.

8. An adaptive ARIMA modeling system for predictive maintenance of semiconductor manufacturing equipment, characterized in that, include: The data acquisition module is configured to acquire process parameter data of semiconductor manufacturing equipment in real time, wherein the process parameter data includes at least one of etching rate, temperature or pressure. A window adjustment module is configured to dynamically adjust the training window size of the ARIMA model based on the volatility of the process parameter data. The window adjustment module includes: The volatility calculation unit is configured to calculate the rolling standard deviation σ of recent data. roll The rolling standard deviation σ roll It is the standard deviation of the most recent m data points, where m is a preset positive integer; The threshold comparison unit is configured to compare σ roll With upper limit threshold Th high and lower limit threshold Th low Comparison, where Th high =k1×σ process ,Th low =k2×σ process k1 and k2 are preset coefficients greater than 0, σ process The historical stability standard deviation of the process parameters during the stable operation phase of the equipment; Window control unit, configured to operate based on rolling standard deviation σ roll With upper limit threshold Th high and lower limit threshold Th low Adjusting the training window size based on the comparison results includes: If σ roll >Th high Then increase the training window size to min(W) old ×a,W max ), where W old Where W is the current window size, 'a' is a preset coefficient greater than 1, and W is the current window size. max Set the maximum window size to the preset value; If σ roll <Th low Then reduce the training window size to max(W) old ×b,W min ), where b is a preset coefficient less than 1, W min Set the minimum window size; If Th low ≤σ roll ≤Th high If so, the training window size remains unchanged; The model fitting module is configured to use the process parameter data within the adjusted training window to fit an ARIMA model, wherein the ARIMA model includes an autoregressive order p, a difference order d, and a moving average order q. The prediction and early warning module is configured to use the fitted ARIMA model to predict the next process parameter data and generate equipment status early warning signals based on the prediction results.

9. An electronic device, characterized in that, include: A memory and a processor, wherein the processor and the memory are connected; The memory is used to store programs; The processor invokes a program stored in the memory to execute the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, It stores a computer program, which is executed by a computer to perform the method as described in any one of claims 1-7.

Images