Information processing device, monitoring method, program, and article manufacturing method

Abstract

The information processing device includes: an acquisition unit configured to acquire a plurality of time series data indicating changes in output values ​​of a plurality of sensors from a monitoring target including a plurality of sensors; a model generation unit configured to generate a model indicating a relationship between the plurality of time series data from each of a plurality of time periods in the plurality of time series data, thereby generating a plurality of models corresponding to the plurality of time periods respectively; and a detection unit configured to detect an abnormality of the monitoring target based on the plurality of models and the plurality of time series data. The plurality of time periods may include two time periods partially overlapping each other.

Description

Technical Field

[0001] The present invention relates to an information processing device, a monitoring method, a program, and an article manufacturing method. Background Art

[0002] Japanese Patent Laid-Open No. 2017-021702 describes a method for monitoring the signs of failure of equipment installed in a factory. The method uses a failure sign monitoring device that monitors the signs of failure of each device installed in the factory based on measurement data measured by a sensor for measuring the behavior of the equipment. The failure sign monitoring device divides the startup process of the factory into multiple time periods based on the operation content of the processing, and displays the failure sign part of each device by checking the validity of the behavior state based on the measurement data at the end of each time period. In addition, when it is confirmed that the behavior state in each time period is valid, the failure sign monitoring device advances to the next time period.

[0003] In order to monitor the signs of failure, a model indicating the relationship between multiple time series data can be generated based on multiple time series data indicating changes in the output values ​​of multiple sensors within a given period. If the difference between the output value of the model and the output value of the sensor corresponding to the model exceeds a threshold, then this indicates that an abnormality (sign of failure or failure) has occurred.

[0004] In the case where the model is inappropriate for some reason as described above, if the output value of the inappropriate model is close to the output value of the sensor even when the abnormality has actually occurred, the occurrence of the abnormality cannot be detected. However, it is difficult to verify the inappropriateness of the model. In addition, even if the abnormality is detected, it cannot be guaranteed that the model is appropriate. Summary of the invention

[0005] The present invention provides a technology that is useful for detecting the occurrence of abnormality in a monitoring target with higher accuracy.

[0006] One aspect of the present invention relates to an information processing device, and the information processing device includes: an acquisition unit configured to acquire a plurality of time series data indicating changes in output values ​​of a plurality of sensors from a monitoring target including a plurality of sensors; a model generation unit configured to generate a model indicating a relationship between the plurality of time series data from each of a plurality of time periods in the plurality of time series data, thereby generating a plurality of models corresponding to the plurality of time periods respectively; and a detection unit configured to detect an abnormality of the monitoring target based on the plurality of models and the plurality of time series data. The plurality of time periods may include two time periods that partially overlap each other. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 is a diagram showing an arrangement example of an information processing apparatus of the embodiment.

[0008] Figure 2A is a diagram showing an example of a relationship between a monitoring target and an information processing device.

[0009] Figure 2B is a diagram showing an example of a relationship between a monitoring target and an information processing device.

[0010] Figure 3 is a diagram showing an arrangement example of an information processing apparatus of the embodiment.

[0011] Figure 4 is a view showing the arrangement of a substrate processing apparatus as a monitoring target.

[0012] Figure 5 is a diagram showing an example of a temperature adjustment system as a monitoring target.

[0013] Figure 6 is a diagram showing an example of the operation of the information processing apparatus.

[0014] Fig. 7A is a diagram showing an example of an output value of a temperature sensor.

[0015] Figure 7B is a diagram showing an example of an output value of a temperature sensor.

[0016] Figure 8 is a diagram showing an example of the operation of the information processing apparatus.

[0017] Fig. 9 is a diagram showing an example of the operation of the information processing apparatus.

[0018] Fig.10 2 is a diagram showing an example of abnormality detection of the embodiment.

[0019] Fig.11 is a view showing an example of abnormality detection of a comparative example. DETAILED DESCRIPTION

[0020] Hereinafter, the embodiments will be described in detail with reference to the accompanying drawings. It is noted that the following embodiments are not intended to limit the scope of the claimed invention. A plurality of features are described in the embodiments, but the invention requiring all of these features is not limited, and a plurality of such features may be appropriately combined. In addition, in the accompanying drawings, the same reference numerals are assigned to the same or similar configurations, and repeated descriptions thereof are omitted.

[0021] Figure 1An example of the configuration of the information processing device 100 of the embodiment is shown. The information processing device 100 can be operated as a monitoring device for monitoring the state of the monitoring target 150, or as a device for executing a monitoring method for monitoring the state of the monitoring target 150. The information processing device 100 can be composed of a computer including a memory for storing a program and a processor operating based on the program. The program can be carried by being stored in a storage medium. Alternatively, the program can be transmitted via a communication path such as a network. The information processing device 100 can be composed of a single computer, or a plurality of computers connected to each other via a communication path such as a network, and can also be composed of other forms. In one example, the information processing device 100 may include a CPU 101, a ROM 102, a RAM 103, an auxiliary storage device 104, an input device 105, a display device 106, a communication device 107, and a bus 108.

[0022] The CPU 101 can operate based on the program stored in the ROM 102, and can define the functions of the information processing apparatus 100. The CPU 101 can communicate with the ROM 102, the RAM 103, the auxiliary storage device 104, the input device 105, the display device 106, and the communication device 107 via the bus 108, and can control these devices. The RAM 103 can be used to temporarily store data. The auxiliary storage device 104 can be a non-volatile storage device composed of one or more storage devices. The program for controlling the CPU 101 can also be stored in the auxiliary storage device 104. The storage device can include, for example, an HDD (hard disk drive), a storage disk, and / or a memory card.

[0023] The input device 105 may include a keyboard, a pointing device, a scanner, etc. The display device 106 may include a liquid crystal monitor, an organic EL monitor, etc. The communication device 107 has a function of communicating with an external device. The communication device 107 can communicate with the monitoring target 150 via a communication path such as a network. For example, the communication device 107 can obtain or receive a plurality of time series data indicating changes in output values ​​of a plurality of sensors 151 of the monitoring target 150 from the monitoring target 150. The monitoring target 150 may be a manufacturing device for manufacturing an article, such as a substrate processing device. This substrate processing device may be, for example, an exposure device, a pattern forming device such as an imprint device, or a photolithography device.

[0024] like Figure 2A As shown in FIG. 1 , an information processing device 100 for monitoring a monitoring target 150 may be installed. Alternatively, as Figure 2BAs shown in FIG. 1 , one information processing apparatus 100 may be installed to monitor a plurality of monitoring targets 150. In addition, although not shown, the information processing apparatus 100 may also be incorporated into the monitoring target 150.

[0025] Figure 3 An example of a configuration of the information processing device 100 according to another perspective is shown. The information processing device 100 may include an acquisition unit 121, a model generation unit 122, and a detection unit 123. The acquisition unit 121 may be configured to acquire a plurality of time series data indicating changes in output values ​​of the plurality of sensors 151 from a monitoring target 150 including a plurality of sensors 151. The model generation unit 122 may generate a model indicating a relationship between the plurality of time series data from each of a plurality of time periods in the plurality of time series data, thereby generating a plurality of models corresponding to the plurality of time periods. The detection unit 123 may detect an abnormality of the monitoring target 150 based on the plurality of models and the plurality of time series data. The information processing device 100 may further include an exclusion unit 124 for excluding a model that is not used to detect an abnormality of the monitoring target 150. The information processing device 100 may further include a notification unit 125 for notifying the detection unit 123 that an abnormality of the monitoring target 150 has been detected.

[0026] The acquisition unit 121 may acquire error information indicating that an error has occurred from the monitoring target 150, and the model generation unit 122 may determine a plurality of time periods during a period in which no error has occurred. The monitoring target 150 may have an error detection function for detecting the occurrence of an error, and the error information may be provided by the error detection function. By determining a plurality of time periods during a period in which no error has occurred, generation of a model based on inappropriate time series data may be prevented.

[0027] The model generation unit 122 may also determine the multiple time periods so that they partially overlap each other, which is beneficial for shortening the time period required to generate the multiple models.

[0028] The exclusion unit 124 may be configured to exclude a model that is not used to detect anomalies of the monitoring target 150 from the plurality of models. The exclusion unit 124 may determine the model to be excluded based on, for example, the time that has passed since the model generation unit 122 generated the model. For example, the exclusion unit 124 may determine a model that has exceeded a predetermined time since the model generation unit 122 generated the model as a model to be excluded. If there is no exclusion unit 124, the number of models continues to increase. In addition, if there is no exclusion unit 124, the old model has an impact on the anomaly detection, and this may reduce the sensitivity of the monitoring target 150 and / or the sensor to changes that occur over time.

[0029] The detection unit 123 may calculate the difference between the output value of each of the plurality of sensors and the output value given by the model. If the value obtained by processing the difference exceeds a predetermined value, the detection unit 123 may detect the occurrence of an abnormality of the monitoring target 150.

[0030] For example, the exclusion unit 124 may determine a model in which the frequency of the output value of the abnormal occurrence detected exceeds a predetermined frequency as a model to be excluded. For example, the exclusion unit 124 may determine a model to be excluded from at least three models based on the frequency of the abnormal occurrence detected by using at least three models, and the three models correspond to at least three time periods arranged in a time order in a plurality of time periods. As an example, the exclusion unit 124 may determine a threshold value based on the frequency of the abnormal occurrence detected by using at least three models. Then, the exclusion unit 124 may determine a model that generates a frequency exceeding the threshold value among the frequencies of the abnormal occurrence detected by using at least three models as a model to be excluded.

[0031] The exclusion unit 124 may also determine whether to exclude a model generated thereafter by the model generation unit 122 based on a threshold value determined by using an outlier value calculated for each of the already existing models.

[0032] The detection unit 123 may calculate an evaluation value while weighting a plurality of differences calculated by using a plurality of models (i.e., the difference between the sensor output value and the output value generated by the model), and may detect the state of the monitoring target 150 based on the evaluation value. The detection unit 123 may also perform weighting based on the time that has passed since the model generation unit 122 generated the model. The detection unit 123 may reduce the weight as the elapsed time of the model becomes longer. Alternatively, the detection unit 123 may reduce the weight as the elapsed time of the model becomes shorter.

[0033] The notification unit 125 may be configured to notify the detection unit 123 that an abnormality of the monitoring target 150 has been detected. Such notification is conducive to rapid maintenance of the monitoring target 150.

[0034] The information processing device 100 can be understood as a device for executing a method, which includes an acquisition step, a model generation step, and a detection step corresponding to the acquisition unit 121, the model generation unit 122, and the detection unit 123. The method may also include an exclusion method corresponding to the exclusion unit 124 and / or a notification step corresponding to the notification unit 125.

[0035] Figure 4The substrate processing device, more specifically, an example of a monitoring target 150 configured as an exposure device 10 is shown. The exposure device 10 may include a light source unit 1 including a light source. The light source may be, for example, a high-pressure mercury lamp or an excimer laser. When the light source is an excimer laser, the light source unit 1 is installed outside the chamber of the exposure device 10 in some cases, but may also be installed inside the chamber.

[0036] The exposure device 10 may include an illumination system 2, a reticle stage 3, a projection optical system 5, and a substrate stage 6. The illumination system 2 may illuminate a reticle R held by the reticle stage 3 by using light from a light source unit 1. The illumination optical system 2 may project a pattern of the illuminated reticle R onto a substrate S. The substrate stage 6 may include a substrate chuck 7 for holding the substrate S. The substrate stage 6 may position the substrate S by moving in a state where the substrate S is held by the substrate chuck 7. The exposure device 10 may be configured to expose the substrate S by a step-and-repeat method or a step-and-scan method.

[0037] The exposure device 10 may provide, for example, a substrate box 12 that holds a plurality of substrates S. The exposure device 10 may include a pre-aligner 9. The exposure device 10 may include a robot (not shown), and the robot may extract the substrate S from the substrate box 12 and transport the substrate to the pre-aligner 9. The pre-aligner 9 may perform pre-alignment (alignment of direction and position) of the substrate S. The substrate S pre-aligned by the pre-aligner 9 may be transported to the substrate chuck 7 of the substrate stage 6 by the robot.

[0038] Other devices may also be connected to the exposure device 10. The other device may be a coating / developing device. The substrate S coated with a resist by the coating / developing device may be loaded into the exposure device 10, and the substrate S exposed by the exposure device 10 may be unloaded to the coating / developing device and developed by the coating / developing device.

[0039] The exposure device 10 may include a controller 11 for controlling the operation of the exposure device 10. The controller 11 may control, for example, the light source unit, the illumination system 2, the original plate stage 3, the projection optical system 5, the substrate stage 6, and the pre-aligner 9. The controller 11 may communicate with (the communication device 107 of) the information processing device 100. Figure 4 Although not shown, the exposure device 10 may include a plurality of sensors, and the controller 11 may transmit or provide a plurality of time series data indicating changes in the output values ​​of the plurality of sensors to (the communication device 107 of) the information processing device 100. The plurality of time series data may also be transmitted or provided to (the communication device 107 of) the information processing device 100 via other devices.

[0040] Figure 5 Shown is incorporated into Figure 41 is a configuration example of a temperature adjustment system 300 in the exposure device 10 shown in FIG. The temperature adjustment system 300 is an example of the monitoring target 150. Figure 5 , the thick arrow indicates the flow of the coolant, and the thin arrow indicates the flow of information. The temperature adjustment system 300 may include, for example, a first module 301 and a second module 302. A plurality of temperature adjustment targets 416 to 419 may be installed in the second module 302. The plurality of temperature adjustment targets 416 to 419 may include, for example, a light source 1, an illumination optical system 2, a master stage 3, a projection optical system 5, and a substrate stage 6. The first module 301 may adjust the temperature of the coolant and supply the temperature-adjusted coolant to the second module 302. The coolant whose temperature is adjusted in the first module 301 may perform temperature adjustment on one or more units in the second module 302 while absorbing heat from the one or more units, and may thereafter return to the first module 301.

[0041] The first module 301 may include, for example, a temperature adjustment unit 401, a temperature adjustment unit 402, a temperature sensor 401S, a temperature sensor 402S, a control unit 401C, and a control unit 402C. The temperature adjustment unit 401 may reduce the temperature of the coolant to a target temperature and supply the coolant to the temperature adjustment unit 402. The control unit 401C may determine a control amount according to the temperature measured by the temperature sensor 401S so that the temperature of the coolant matches the target temperature, and may operate the temperature adjustment unit 401 according to the control amount. The temperature adjustment unit 402 may adjust the temperature of the coolant within a temperature range allowed by the second module 302, and supply the coolant to the second module 302. The control unit 402C may determine a control amount according to the temperature measured by the temperature sensor 402S so that the temperature of the coolant falls within a temperature range allowed by the second module 302, and may operate the temperature adjustment unit 402 according to the control amount.

[0042] In the second module 302, the temperature adjustment units 412 to 415 may adjust the temperature of the coolant so that the temperature adjustment targets 416 to 419 fall within the target temperature range. The control unit 412C may determine the control amount according to the temperature measured by the temperature sensors 412S1 and 412S2 so that the temperature adjustment target 416 falls within the target temperature range, and may operate the temperature adjustment unit 412 according to the control amount. The control unit 413C may determine the control amount according to the temperature measured by the temperature sensors 413S1 and 413S2 so that the temperature adjustment target 417 falls within the target temperature range, and may operate the temperature adjustment unit 413 according to the control amount.

[0043] The control unit 411C may determine a control amount according to the temperature measured by the temperature sensor 411S and the information from the control units 414C and 415C so that the temperature of the coolant falls within the target temperature range, and may operate the temperature adjustment unit 411 according to the control amount. The control unit 414C may determine a control amount according to the temperature measured by the temperature sensors 414S1 and 414S2 so that the temperature adjustment target 418 falls within the target temperature range, and may operate the temperature adjustment unit 414 according to the control amount. The control unit 415C may determine a control amount according to the temperature measured by the temperature sensors 415S1 and 415S2 so that the temperature adjustment target 419 falls within the target temperature range, and may operate the temperature adjustment unit 415 according to the control amount.

[0044] The temperature regulating units 401, 402 and 412 to 415 may also be heating / cooling units using heat exchange. In another perspective, the temperature regulating unit 401 may be a cooling unit, and the temperature regulating units 402 and 412 to 415 may be heating units. The coolant may be a liquid or a gas.

[0045] exist Figure 5 In the temperature adjustment system 300 shown in FIG. 1 , the temperature of the temperature adjustment target is controlled, and temperature sensors 401S, 402S, 411S to 415S2 are installed. However, the temperature adjustment system 300 may also include sensors for measuring information other than temperature (for example, flow sensors, pressure sensors, acceleration sensors, and position sensors). In addition, the exposure device 10 may include a control system for controlling a control target related to parameters other than temperature.

[0046] An example of a model to be generated by the model generation unit 122 will be described below. To simplify the description, it is assumed that Figure 5 The output values ​​of the two temperature sensors (e.g., temperature sensors 401S and 402S) in the temperature control system shown in FIG. at time t are a t and b t The output values ​​of the two temperature sensors are t and b t The relationship between can be defined by the model (function) given by the following equation (1):

[0047] b t =f(a t )...(1)

[0048] For example, the model f may be a regression equation determined by the least square method or the like based on the measurement data measured by the two temperature sensors. Alternatively, the model f may also be a learning model generated, for example, by using machine learning. For example, the model f may be a model including a neural network. The neural network is a model having a multi-layer network structure including, for example, an input layer, an intermediate layer, and an output layer. By using a indicating a as input data t and b as training data t The learning model of the relationship between the model f can be obtained by optimizing the combined weight coefficients and the like inside the neural network according to an algorithm such as the error back propagation method. The error back propagation method is a method of adjusting the combined weight coefficients and the like between the nodes of each neural network so that the difference between the output data and the training data is reduced. In addition, the model f may also be a learning model including an SVM (support vector machine) instead of a model including a neural network.

[0049] Given a sensor S i The output value (hereinafter referred to as the predicted output value) x ij Model f ij (x j ) can be used as the sensor S by the following equation (2) j The output value (hereinafter referred to as the predicted output value) x j In equation (2), i is an integer from 1 to N, N is the number of sensors, and j is an integer from 1 to N other than i.

[0050] x ij =f ij (x j )...(2)

[0051] Equation (2) can be expressed as follows:

[0052]

[0053] Figure 6An operation example of the information processing apparatus 100 regarding model generation is shown. In step S301, the information processing apparatus 100 causes the acquisition unit 121 to acquire multiple data indicating output values ​​of multiple sensors from the monitoring target 150 including multiple sensors. In step S302, the information processing apparatus 100 saves the multiple data acquired in step S301 in the storage area of ​​the auxiliary storage device 104. When the processing unit including steps S301 and S302 is executed multiple times, multiple time series data indicating changes in the output values ​​of multiple sensors are saved in the storage area of ​​the auxiliary storage device 104. Each time series data contains data obtained by arranging the output value of one sensor in a time series manner. In step S303, the information processing apparatus 100 determines whether the time series data saved in the storage area of ​​the auxiliary storage device 104 has reached the amount required to form the model. If the time series data saved in the storage area of ​​the auxiliary storage device 104 has reached the amount required to form the model, the information processing apparatus 100 proceeds to step S304, otherwise returns to step S301.

[0054] In step S304, the information processing apparatus 100 causes the model generation unit 122 to generate a plurality of models based on the time series data stored in the storage area of ​​the auxiliary storage device 104. ij (x j In this step, the model generation unit 122 generates a model indicating the relationship between the plurality of time series data from each of the plurality of time periods in the plurality of time series data, thereby generating a plurality of models f corresponding to the plurality of time periods, respectively. ij (x j In step S305, the information processing device 100 generates the multiple models f generated by the model generation unit 122 in step S304. ij (x j ) is stored in the storage area of ​​the auxiliary storage device 104.

[0055] Fig. 7A An example of changes in the output value of the temperature sensor 401S is shown. Figure 7B An example of a change in the output value of the temperature sensor 402S is shown. Fig. 7A and Figure 7B , the abscissa indicates time, and the ordinate indicates output values ​​(values ​​indicating temperature) of the temperature sensors 401S and 402S. 1 To P 5 is the period used to generate multiple models. 1 is the period used to generate a model, period P 2 is the period used to generate a model, and period P 3is the time period used to generate a model. In addition, the time period P 4 is the period used to generate a model, and period P 5 is the period used to generate a model.

[0056] In one example, the output values ​​(data) required to generate a model are 5 days of output values ​​measured at 5 second intervals. 1 To P 5 Each of the periods is 5 days. 1 To P 5 It can also be determined to have overlapping periods. For example, assuming that the output value (data) required to generate each model is the output value of 5 days, if the period P 1 To P 5 If there is no overlap, then 25 days of measurement data are required, but if 1 day of output values ​​overlap, then only 21 days of output values ​​need to be obtained.

[0057] Figure 8 FIG. 1 shows an example of the operation of the information processing device 100 regarding prediction using a model. The detection unit 123 may perform Figure 8 In step S401, the information processing apparatus 100 determines whether there is a Fig. 7A and Figure 7B If one or more models exist, the information processing apparatus 100 proceeds to step S402. If not, the information processing apparatus 100 terminates. Figure 8 The processing shown in .

[0058] In step S402, the information processing apparatus 100 determines whether there is a model (unapplied model) that is not applied to the sensor output value newly acquired by the acquisition unit 121. If the model exists, the information processing apparatus 100 proceeds to step S403, and if not, terminates. Figure 8 In this step, the sensor output value newly acquired by the acquisition unit 121 is the sensor output value acquired by the acquisition unit 121 after the model is generated. If there is no model that is not applied to the sensor output value newly acquired by the acquisition unit 121, this means that all models applicable to the output value have been applied to the output value.

[0059] In step S404, the information processing apparatus 100 reads out the unapplied model from the storage area of ​​the auxiliary storage device 104, and in step S404, determines whether the unapplied model is valid. If the unapplied model is valid, the information processing apparatus 100 proceeds to step S405, otherwise terminates. Figure 8The process shown in . This determination may be performed based on the time that has elapsed since the model was generated. For example, if the time that has elapsed since the model was generated has exceeded a predetermined time, the information processing apparatus 100 may determine that the model is invalid.

[0060] In step S405, the information processing apparatus 100 performs the following operations by using the model f read out in step S403. ij (x j ) to calculate the predicted output value x ij Then, in step S406, the information processing device 100 determines the predicted output value x calculated in step S405. ij More specifically, if the predicted output value x ij With the measured output value x i If the difference or ratio between exceeds the threshold, the information processing device 100 can determine the predicted output value x ij Invalid. If the predicted output value x calculated in step S405 is ij If the predicted output value x calculated in step S405 is invalid, the information processing device 100 returns to step S402. ij If the information processing device 100 is effective, then in step S407 the predicted output value x calculated in step S405 is ij The data is stored in the storage area of ​​the auxiliary storage device 104 .

[0061] Fig. 9 An example of the operation of the information processing apparatus 100 related to the process of detecting the occurrence of an abnormality by using a model is shown. The detection unit 123 performs Fig. 9 In step S501, the information processing apparatus 100 determines whether there are one or more predicted output values ​​stored in the storage area of ​​the auxiliary storage device 104. If there are one or more predicted output values, the information processing apparatus 100 proceeds to step S502, and if not, terminates. Fig. 9 The processing shown in .

[0062] In step S502, the information processing apparatus 100 reads the predicted output value stored in the storage area of ​​the auxiliary storage device 104. In step S503, the information processing apparatus 100 calculates an evaluation value based on the predicted output value read in step S502. The evaluation value may be obtained by comparing the predicted output values ​​x ij Each sum in and the predicted output value x ij The corresponding measured output value x iThe information processing apparatus 100 may be used to process the difference between the values ​​of the evaluation values ​​and the values ​​obtained by processing ...

[0063] Fig.10 An example of changes in the evaluation values ​​calculated in this embodiment is shown. Fig.11 An example of changes in evaluation values ​​in the comparative example is shown. Fig.10 and Fig.11 In , the abscissa indicates time, and the ordinate indicates the evaluation value. Fig.10 In this embodiment shown in , a plurality of models corresponding to a plurality of time periods in a plurality of time series data are generated by generating a model indicating a relationship between a plurality of time series data from each of the plurality of time periods, and an evaluation value is calculated by using the plurality of models. Evaluation using a plurality of models like this can play a role in reducing uncertainty as to whether each individual model is appropriate. On the other hand, in Fig.11 In the comparative example shown in , a model indicating the relationship between a plurality of time series data is generated from a period in a plurality of time series data, and an evaluation value is calculated using the model.

[0064] exist Fig.10 In the embodiment shown in FIG. 1 , the occurrence of an abnormality is detected at time t1. Fig.11 In the comparative example shown in , the abnormality is not detected until time t2 after time t1. That is, this embodiment can detect the occurrence of abnormality earlier.

[0065] In addition, even when an abnormality actually occurs in the comparative example when the model generated from one period is inappropriate for some reason, if the output value of the inappropriate model is close to the output value of the sensor, the occurrence of this abnormality cannot be detected. However, it is not easy to verify whether the model is appropriate. In addition, even when an abnormality is detected, it cannot be guaranteed that the model is appropriate.

[0066] The article manufacturing method of this embodiment may include: an exposure step of exposing a substrate by using an exposure device including a plurality of sensors as a monitoring target, a development step of developing the substrate exposed in the exposure step, and a processing step of obtaining an article by processing the substrate developed in the development step. The article manufacturing method may also include the above-mentioned acquisition step and the above-mentioned model generation step. The article manufacturing method may also include a detection step of detecting the state of the exposure device as a monitoring target based on a plurality of models and a plurality of time series data, and a maintenance step of maintaining the exposure device based on the state of the exposure device detected in the detection step.

[0067] The present invention is not limited to the above-described embodiments, and various changes and modifications can be made within the spirit and scope of the present invention. Therefore, in order to make the public aware of the scope of the present invention, the following claims are made.

[0068] This application claims priority from Japanese Patent Application No. 2020-037158, filed on March 4, 2020, which is hereby incorporated by reference herein.

Claims

1. An information processing device, comprising: an acquisition unit configured to acquire, from a monitoring target including a plurality of sensors, a plurality of time series data indicating changes in output values ​​of the plurality of sensors; a model generating unit configured to generate a plurality of models corresponding to different time periods among a plurality of time periods, each model indicating a relationship between the plurality of time series data within a corresponding time period among the plurality of time periods; a detection unit configured to detect anomalies of a monitoring target based on the plurality of models and the plurality of time series data; as well as An exclusion unit is configured to exclude a model that is not used to detect an abnormality of the monitoring target from the plurality of models.

2. The information processing device according to claim 1, wherein: an acquisition unit acquires error information indicating that an error has occurred from a monitoring target, and The model generation unit determines the plurality of time periods based on the error information during a time period in which no error occurs.

3. The information processing device according to claim 1, wherein: The plurality of time periods include two time periods partially overlapping each other.

4. The information processing device according to claim 1, wherein: The exclusion unit determines a model to be excluded based on an elapsed time since the model was generated by the model generation unit.

5. The information processing device according to claim 1, wherein: The exclusion unit determines a model to be excluded based on a change over time in an output value of each of the plurality of sensors.

6. The information processing device according to claim 5, wherein: The detection unit calculates a difference between an output value of each of the plurality of sensors and an output value generated by a model.

7. The information processing device according to claim 6, wherein: If a value obtained by processing the difference exceeds a predetermined value, the detection unit detects that an abnormality has occurred in the monitoring target.

8. The information processing device according to claim 7, wherein: The exclusion unit determines a model in which a generation frequency of an output value in which an abnormality occurrence is detected exceeds a predetermined frequency as a model to be excluded.

9. The information processing device according to claim 7, wherein: The exclusion unit determines a model to be excluded from the at least three models based on a frequency of abnormality occurrence detected by using the at least three models, the at least three models respectively corresponding to at least three time periods arranged in chronological order among the plurality of time periods.

10. The information processing device according to claim 6, wherein: The exclusion unit determines whether to exclude a model to be generated by the model generation unit thereafter, based on a threshold value determined according to a value obtained by processing a difference calculated using each of the already existing models.

11. The information processing device according to claim 6, wherein: The detection unit calculates an evaluation value while performing weighting on a plurality of differences respectively calculated by using the plurality of models, and detects a state of a monitoring target based on the evaluation value.

12. The information processing device according to claim 11, wherein: The detection unit performs weighting based on an elapsed time from when the model is generated by the model generation unit.

13. The information processing device according to claim 12, wherein: The detection unit reduces the weight as the model becomes older.

14. The information processing device according to claim 12, wherein: The detection unit reduces the weight as the elapsed time of the model becomes shorter.

15. The information processing device according to any one of claims 1 to 14, further comprising: A notification unit is configured to notify the detection unit that an abnormality of the monitoring target has been detected.

16. A manufacturing device comprising: A substrate processing apparatus including a plurality of sensors; as well as An information processing device according to any one of claims 1 to 15 and configured to detect the state of a monitoring target.

17. A monitoring method, comprising: an acquisition step of acquiring, from a substrate processing apparatus including a plurality of sensors, a plurality of time series data indicating changes in output values ​​of the plurality of sensors; A model generating step, generating a plurality of models corresponding to different time periods among the plurality of time periods, each model indicating a relationship between the plurality of time series data within a corresponding time period among the plurality of time periods; a detection step of detecting an abnormality of a substrate processing device based on the plurality of models and the plurality of time series data; as well as The step of excluding a model that is not used to detect abnormality of the substrate processing apparatus from the plurality of models.

18. A storage medium storing a program for causing a computer to execute the monitoring method according to claim 17.

19. A method for manufacturing an article, comprising: an exposure step of exposing the substrate by using an exposure device including a plurality of sensors; a developing step of developing the substrate exposed in the exposing step; a processing step of obtaining an article by processing the substrate developed in the developing step; an acquisition step of acquiring, from an exposure device, a plurality of time series data indicating changes in output values ​​of the plurality of sensors; A model generating step, generating a plurality of models corresponding to a plurality of time periods respectively, each model indicating a relationship between the plurality of time series data in a corresponding time period among the plurality of time periods; a detection step of detecting a state of an exposure device based on the plurality of models and the plurality of time series data; a maintenance step of maintaining the exposure device based on the state of the exposure device detected in the detection step; as well as The step of excluding a model that is not used to detect abnormality of the exposure device from the plurality of models.

Images