Information processing method, computer program, and information processing device

Abstract

The invention provides an information processing method, a computer program, and an information processing device which can expect to estimate the operation state of a semiconductor manufacturing device. An information processing method according to an embodiment includes: acquiring operation data of a first semiconductor manufacturing apparatus and sensor data output by a plurality of sensors when the first semiconductor manufacturing apparatus is operated, the first semiconductor manufacturing apparatus being provided with the plurality of sensors; generating, by machine learning, a learning model for outputting the sensor data in accordance with the input of the operation data on the basis of the acquired operation data and the sensor data; and disposing the generated learning model in a second semiconductor manufacturing apparatus different from the first semiconductor manufacturing apparatus.

Description

Technical Field

[0001] The present disclosure relates to an information processing method, a computer program, and an information processing device. Background Art

[0002] Patent document 1 proposes an integrated management system that accumulates various information collected from substrate processing devices and uses the accumulated data to display information required for energy saving of each substrate processing device installed in a semiconductor manufacturing plant. The integrated management system accumulates various information including power consumption information, gas consumption information, or operation information of the substrate processing device, obtains information that satisfies a predetermined condition from the accumulated information, and calculates and displays at least one of the power consumption consumed by the substrate processing device, the inactive gas consumption, and the device operation rate of the substrate processing device.

[0003] Prior art literature

[0004] Patent Literature

[0005] Patent Document 1: International Publication No. 2014 / 050808 Summary of the invention

[0006] Problem that the invention aims to solve

[0007] The present disclosure provides an information processing method, a computer program, and an information processing device that can be expected to estimate the operating state of a semiconductor manufacturing device.

[0008] Solutions for solving problems

[0009] An information processing method involved in one embodiment includes the following processing: when a first semiconductor manufacturing device equipped with multiple sensors is operated, operating data of the first semiconductor manufacturing device and sensor data output by the sensors are obtained; based on the acquired operating data and sensor data, a learning model that outputs the sensor data according to the input of the operating data is generated through machine learning; and the generated learning model is configured in a second semiconductor manufacturing device that is different from the first semiconductor manufacturing device.

[0010] Effects of the Invention

[0011] According to the present disclosure, it is expected that the operating state of a semiconductor manufacturing apparatus and the like can be estimated. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 It is a schematic diagram showing a configuration example of the information processing system involved in this embodiment.

[0013] Figure 2This is a schematic diagram for explaining an example of a sensor for a semiconductor manufacturing device.

[0014] Figure 3 This is a block diagram showing a configuration example of an information processing device according to the present embodiment.

[0015] Figure 4 It is a schematic diagram showing a structural example of a process DB.

[0016] Figure 5 This is a flowchart showing an example of the procedure of the prediction model generation process performed by the information processing device according to the present embodiment.

[0017] Figure 6 This is a schematic diagram showing an example of information display related to the generation of a prediction model.

[0018] Figure 7 This is a schematic diagram showing an example of information display related to the generation of a prediction model.

[0019] Figure 8 This is a schematic diagram for explaining an example of using a prediction model in a simulation.

[0020] Fig. 9 This is a flowchart showing an example of a procedure of a contribution degree calculation process performed by the information processing device according to the present embodiment.

[0021] Fig.10 It is a schematic diagram showing an example of information display related to contribution degree.

[0022] Fig.11 This is a flowchart showing an example of a procedure of a unit stop process performed by the information processing device according to the present embodiment.

[0023] Fig.12 This is a schematic diagram for explaining the effect of automatic stop of the standby state unit.

[0024] Fig.13 This is a schematic diagram showing a display example related to processing timing verification.

[0025] Fig.14 This is a schematic diagram showing a display example related to processing timing verification. DETAILED DESCRIPTION

[0026] Hereinafter, specific examples of the information processing system according to the embodiments of the present disclosure will be described with reference to the accompanying drawings. In addition, the present disclosure is not limited to these examples, and is intended to include all changes within the meaning and scope of the claims and equivalents to the claims.

[0027] <System Structure>

[0028] Figure 1 : is a schematic diagram showing a structural example of the information processing system involved in the present embodiment. The information processing system involved in the present embodiment is configured to include an information processing device 1 and a semiconductor manufacturing device 2. The semiconductor manufacturing device 2 is, for example, a device that performs various processes for semiconductor manufacturing such as CVD (Chemical Vapor Deposition), sputtering, or etching. The information processing device 1 is, for example, a device that acquires and collects operating data output by the semiconductor manufacturing device 2, and monitors and controls the operation of the semiconductor manufacturing device 2. There may be multiple semiconductor manufacturing devices 2 that are the objects of monitoring and control of the information processing device 1. The operating data acquired by the information processing device 1 can include various data such as process log data, process data, wafer transport history data, and error data.

[0029] In addition, in the information processing system according to the present embodiment, a plurality of sensors 3 for measuring detailed operation status, etc. are provided for at least one semiconductor manufacturing device 2 that is the object of monitoring and control of the information processing device 1. In the case where there are a plurality of semiconductor manufacturing devices 2, the sensors 3 may be provided not for all semiconductor manufacturing devices 2 but for some (at least one) of the plurality of semiconductor manufacturing devices 2. Figure 1 In the following description, the semiconductor manufacturing apparatus 2 provided with the sensor 3 is described as the semiconductor manufacturing apparatus 2A, and the semiconductor manufacturing apparatus 2 not provided with the sensor 3 is described as the semiconductor manufacturing apparatus 2B, and when there is no need to distinguish them, it is simply described as the semiconductor manufacturing apparatus 2. Similarly, the information processing apparatus 1 that monitors and controls the semiconductor manufacturing apparatus 2A provided with the sensor 3 is described as the information processing apparatus 1A, and the information processing apparatus 1 that monitors and controls the semiconductor manufacturing apparatus 2B not provided with the sensor 3 is described as the information processing apparatus 1B, and when there is no need to distinguish them, it is simply described as the information processing apparatus 1.

[0030] In the present embodiment, the plurality of sensors 3 provided in the semiconductor manufacturing device 2A are sensors for performing measurements related to the environment, and include, for example, sensors for measuring power consumption, sensors for measuring water discharge, or sensors for measuring gas discharge. The gas whose discharge is measured by the sensor may include, for example, gases such as CO2 (carbon dioxide) or NOx (nitrogen oxides). In addition, in the present embodiment, the semiconductor manufacturing device 2 is composed of a combination of various units such as an upper cooling unit, an RF (Radio Frequency) power supply unit, a DC (Direct Current) power supply unit, a vacuum pump unit, a chamber heater unit, an ESC (Electric Static Chuck) heater unit, and a lower cooling unit. The plurality of sensors 3 may include sensors for measuring the power consumption, water discharge, or gas discharge of each unit of the semiconductor manufacturing device 2A.

[0031] In the information processing system according to the present embodiment, when a semiconductor manufacturing process is implemented using a semiconductor manufacturing apparatus 2A provided with sensors 3, the information processing apparatus 1A acquires the operation data output by the semiconductor manufacturing apparatus 2A and the measurement results of the plurality of sensors 3, i.e., sensor data, and stores and accumulates the acquired operation data and the sensor data in a database in correspondence. The information processing apparatus 1A performs so-called supervised machine learning processing based on the operation data and sensor data accumulated in the database, and generates a learning model as a prediction model 5 that accepts the operation data as input and outputs a prediction value of the sensor data. The information processing apparatus 1A deploys (configures) the generated prediction model 5 to an information processing apparatus 1B of a semiconductor manufacturing apparatus 2B that is not provided with sensors 3. The information processing apparatus 1B can acquire the prediction value of the sensor data from the operation data output by the semiconductor manufacturing apparatus 2B using the prediction model 5 generated by the information processing apparatus 1, and perform processing such as control or monitoring of the semiconductor manufacturing apparatus 2B based on the prediction value. In the present embodiment, configuring the prediction model 5 for the semiconductor manufacturing apparatus 2B or configuring the prediction model 5 for the information processing device 1B of the semiconductor manufacturing apparatus 2B means that the prediction model 5 can predict the state of the sensor data based on the operation data of the semiconductor manufacturing apparatus 2B.

[0032] For example, in the information processing system according to the present embodiment, the semiconductor manufacturing apparatus 2A is installed at a base such as a company that develops and sells the semiconductor manufacturing apparatus 2. The semiconductor manufacturing apparatus 2A is operated experimentally, and various measurements such as power consumption, water discharge, and gas discharge are performed using the sensor 3, and the sensor data output by the sensor 3 is accumulated together with the operation data. After sufficient operation data and sensor data are accumulated, the prediction model 5 is generated using the information processing device 1A.

[0033] Thereafter, in the information processing system according to the present embodiment, when one or more semiconductor manufacturing apparatuses 2B are installed at one or more bases, for example, sensors 3 are not installed for each semiconductor manufacturing apparatus 2B (or sensors 3 may be installed in a smaller number than the number of sensors 3 installed for the semiconductor manufacturing apparatus 2A). Instead of installing sensors 3 for each semiconductor manufacturing apparatus 2B, the prediction model 5 generated by the information processing apparatus 1A is installed (imported) for the information processing apparatus 1B that monitors and controls each semiconductor manufacturing apparatus 2B, and the prediction model 5 is deployed (configured) for each semiconductor manufacturing apparatus 2B. For example, the information processing apparatus 1B acquires operation data accompanying the initial operation of the semiconductor manufacturing apparatus 2B, and inputs the acquired operation data to the prediction model 5, thereby being able to acquire sensor data output by the prediction model 5 without using the sensors 3.

[0034] Figure 2 This is a schematic diagram for explaining an example of the sensor 3 for the semiconductor manufacturing apparatus 2A. Figure 2 The semiconductor manufacturing apparatus 2A shown is configured to include six units, namely, units 1 to 6, and a control unit for controlling these units. The semiconductor manufacturing apparatus 2A is supplied with power from three power supplies 1 to 3. Power supply 1 supplies power to the control unit, unit 1, and unit 2. Power supply 2 supplies power to units 3 and 4. Power supply 3 supplies power to units 5 and 6. Power supplies 1 to 3, for example, supply power from a commercial AC power supply to each unit of the semiconductor manufacturing apparatus 2A.

[0035] exist Figure 2 In FIG. 1 , a single power supply path from a power source to each unit is shown by a thick solid line. In the information processing system according to the present embodiment, for example, a sensor for measuring power consumption is provided for each of these power supply paths. Figure 2In the figure, the circle marked in a manner overlapping with the thick solid line is the position where the sensor 3 is set. In the case of this example, there are seven power supply paths from three power sources 1 to 3 to seven units, and seven sensors 3 are set. In this way, in order to measure the power consumption of each unit of the semiconductor manufacturing device 2, many sensors 3 are required, but when a large number of semiconductor manufacturing devices 2 are set up in, for example, a semiconductor manufacturing factory, setting sensors 3 for all semiconductor manufacturing devices 2 may lead to increased costs. Therefore, in the information processing system involved in this embodiment, the semiconductor manufacturing device 2A provided with the sensor 3 is used to obtain operation data and sensor data in the test operation stage, and the prediction model 5 is generated based on the acquired data. In the actual operation stage, the sensor 3 is not set in the semiconductor manufacturing device 2B (or the number of sensors 3 less than the number in the test stage may be set), and the prediction model 5 is used to obtain sensor data instead of using the sensor 3.

[0036] Figure 3 1 is a block diagram showing a configuration example of an information processing device 1 according to the present embodiment. The information processing device 1 according to the present embodiment is configured to include a processing unit 11, a storage unit (storage device) 12, a communication unit (transceiver) 13, an input / output unit 14, a display unit 15, and an operation unit 16. In addition, in the present embodiment, the processing is performed by one information processing device 1, but the processing of the information processing device 1 may be performed in a distributed manner by a plurality of devices.

[0037] The processing unit 11 is composed of a CPU (Central Processing Unit), an MPU (Micro-Processing Unit), a GPU (Graphics Processing Unit), a quantum processor, or other computing devices, a ROM (Read Only Memory), a RAM (Random Access Memory), etc. The processing unit 11 reads and executes a program 12a stored in the storage unit 12 to perform various processes such as a process of acquiring operation data of the semiconductor manufacturing device 2, a process of acquiring sensor data measured by the sensor 3, and a process of generating a prediction model 5 based on these data.

[0038] The storage unit 12 is constituted by a large-capacity storage device such as a hard disk. The storage unit 12 stores various programs executed by the processing unit 11 and various data required for the processing of the processing unit 11. In the present embodiment, the storage unit 12 stores the program 12a executed by the processing unit 11. In addition, the storage unit 12 is provided with a process DB (database) 12b and a prediction model storage unit 12c. The process DB 12b stores the operation data acquired from the semiconductor manufacturing device 2 and the sensor data acquired from the sensor 3 in correspondence, and the prediction model storage unit 12c stores information related to the prediction model 5 generated by machine learning based on these data.

[0039] In the present embodiment, the program (computer program, program product) 12a is provided in the form of being recorded on a recording medium 99 such as a memory card or an optical disk, and the information processing device 1 reads the program 12a from the recording medium 99 and stores it in the storage unit 12. However, the program 12a may be written into the storage unit 12, for example, during the manufacturing stage of the information processing device 1. In addition, for example, the program 12a may be published by a remote server device and acquired by the information processing device 1 through communication. For example, the program 12a may be recorded in the recording medium 99 and read by a writing device and written into the storage unit 12 of the information processing device 1. The program 12a may be provided in the form of being published via a network, or may be provided in the form of being recorded in the recording medium 99.

[0040] The process DB 12 b is a database that stores the operation data acquired by the information processing apparatus 1A from the semiconductor manufacturing apparatus 2A and the sensor data acquired from the sensor 3 in association with each other. Figure 4 1 is a schematic diagram showing a structural example of the process DB 12b. The illustrated process DB 12b stores data such as "time stamp", "device ID", "process ID", "operation data" and "sensor data" in correspondence. "Time stamp" is information on the date and time when the information processing device 1A acquires the operation data and the sensor data. "Device ID" is identification information uniquely assigned to the semiconductor manufacturing device 2, and can be predetermined by, for example, a designer or manager of the information processing system involved in this embodiment. "Process ID" is identification information uniquely assigned to information such as the process or setting of the semiconductor manufacturing process performed using the semiconductor manufacturing device 2, i.e., the so-called process, and can be predetermined by, for example, a user such as a manager of the information processing system or an operator performing semiconductor manufacturing operations.

[0041] “Operation data” may include, for example, various information such as “high frequency RF power [W]”, “high frequency RF power pulse duty cycle [%]”, “high frequency RF power pulse frequency [kHz]”, “low frequency RF power [W]”, “low frequency RF power pulse duty cycle [%]”, “low frequency RF power pulse frequency [kHz]”, “pressure [mTorr]”, “total gas flow rate [sccm]”, “ESC temperature [°C]”, and “cooler temperature [°C]”. In addition, the information included in the operation data illustrated in this figure is information included in the data of the so-called process log output by the semiconductor manufacturing device 2 accompanying the implementation of the semiconductor manufacturing process. However, the information included in the operation data is not limited to the information in the process log data, and may include, for example, various information such as process data, wafer transport history data, and error data. Process data is data that collects information such as the process or settings of the semiconductor manufacturing process. Wafer transport history data is data that collects information related to the transport of the wafer, such as the ID of the wafer being transported and the date and time of the transport. The error data is data that contains information such as the date and time when an error (abnormality or failure, etc.) is detected in the semiconductor manufacturing apparatus 2 and the type of error.

[0042] The “sensor data” may include information such as “power consumption 1 [W]”, “power consumption 2 [W]”, and “power consumption N [W]”. The sensor data illustrated in this figure is sensor data assuming that the sensor 3 measures the power consumption of N units of the semiconductor manufacturing device 2 separately. The information included in the sensor data is not limited to the power consumption, and may include various information that the sensor 3 can detect, such as the amount of water discharged or the amount of gas discharged. In addition, the information included in the sensor data may not be information about each unit of the semiconductor manufacturing device 2.

[0043] The information processing device 1A, for example, repeatedly acquires operation data during the period when the semiconductor manufacturing device 2A implements the semiconductor manufacturing process, and stores the acquired operation data in the process DB 12b. As a result, the operation data in time series are gradually accumulated in the process DB 12b, for example, by implementing the semiconductor manufacturing process on a wafer, a plurality of operation data are accumulated in the process DB 12b. In addition, the information processing device 1B monitored and controlled by the semiconductor manufacturing device 2B that is not provided with the sensor 3 may not have the process DB 12b. However, in the present embodiment, the information processing device 1B also has the process DB 12b, which stores data other than the sensor data, such as "time stamp", "device ID", "process ID" and "operation data".

[0044] The prediction model storage unit 12c of the storage unit 12 stores information related to the prediction model 5 generated by the information processing device 1A based on the operation data and sensor data stored in the process DB 12b. The information related to the prediction model 5 may include, for example, structure information indicating what the structure of the prediction model 5 is, and information such as the value of the internal parameter of the prediction model 5. In the case of the information processing device 1A that monitors and controls the semiconductor manufacturing device 2A provided with the sensor 3, the information related to the prediction model 5 generated by itself is stored in the prediction model storage unit 12c. In the case of the information processing device 1B that monitors and controls the semiconductor manufacturing device 2B not provided with the sensor 3, the information related to the prediction model 5 generated by other information processing devices 1A is stored in the prediction model storage unit 12c.

[0045] The communication unit 13 of the information processing device 1 communicates with various devices via a wired or wireless network N including a LAN (Local Area Network), the Internet, or a mobile phone communication network. The communication unit 13 sends data provided by the processing unit 11 to other devices, and provides data received from other devices to the processing unit 11. In this embodiment, the communication unit 13 communicates with other information processing devices 1 via the network N to send and receive information related to the prediction model 5.

[0046] The input / output unit 14 is connected to the semiconductor manufacturing apparatus 2 via a communication line or a signal line, and transmits and receives information to and from the semiconductor manufacturing apparatus 2. In the present embodiment, the input / output unit 14 receives the operation data output by the semiconductor manufacturing apparatus 2, acquires the operation data, and provides it to the processing unit 11. The input / output unit 14 outputs control instructions, etc., provided by the processing unit 11 to the semiconductor manufacturing apparatus 2. In addition, in the information processing device 1A that monitors and controls the semiconductor manufacturing apparatus 2A provided with the sensor 3, the input / output unit 14 is connected to one or more sensors 3. The input / output unit 14 receives the sensor data output by the sensor 3, acquires the sensor data, and provides it to the processing unit 11.

[0047] In addition, in this embodiment, the information processing device 1A directly obtains the sensor data of the sensor 3, but it is not limited to this. The semiconductor manufacturing device 2A may obtain the sensor data output by the sensor 3 and the semiconductor manufacturing device 2A sends the operation data and the sensor data to the information processing device 1A.

[0048] The display unit 15 is configured using a liquid crystal display or the like, and displays various images and characters based on the processing of the processing unit 11. The display unit 15 displays various information such as operation data acquired from the semiconductor manufacturing device 2, sensor data acquired from the sensor 3, and information related to the prediction model generated based on these data.

[0049] The operation unit 16 receives the user's operation and notifies the processing unit 11 of the received operation. For example, the operation unit 16 receives the user's operation through an input device such as a mechanical button or a touch panel provided on the surface of the display unit 15. In addition, for example, the operation unit 16 may be an input device such as a mouse and a keyboard, and these input devices may also be detachable from the information processing device 1.

[0050] In addition, the storage unit 12 may be an external storage device connected to the information processing device 1. In addition, the information processing device 1 may be a multi-computer configured in a manner including a plurality of computers, or may be a virtual machine virtually constructed by software. In addition, the information processing device 1 is not limited to the above-mentioned structure, and for example, it may not include the display unit 15 and the operation unit 16.

[0051] In the information processing device 1 according to the present embodiment, the processing unit 11 reads and executes the program 12a stored in the storage unit 12, whereby the data acquisition unit 11a, the prediction model generation unit 11b, the prediction model configuration unit 11c, the control processing unit 11d, and the display processing unit 11e are implemented as software functional units by the processing unit 11. In addition, the information processing device 1B that monitors and controls the semiconductor manufacturing device 2B that is not provided with the sensor 3 may not include the data acquisition unit 11a, the prediction model generation unit 11b, and the prediction model configuration unit 11c.

[0052] The data acquisition unit 11a acquires the operation data output by the semiconductor manufacturing device 2A, for example, when or after the semiconductor manufacturing device 2A performs a semiconductor manufacturing process. In addition, the data acquisition unit 11a acquires the sensor data output by one or more sensors 3, for example, at the same time as the operation data is acquired from the semiconductor manufacturing device 2A. The data acquisition unit 11a adds information such as "time stamp", "device ID" and "process ID" to the acquired operation data and sensor data, and stores the information in the process DB 12b in correspondence.

[0053] The prediction model generation unit 11b performs machine learning using the operation data and sensor data acquired by the data acquisition unit 11a and accumulated in the process DB 12b, thereby generating a prediction model 5. In the present embodiment, the prediction model 5 is a learning model that accepts the operation data as input and outputs the predicted value of the sensor data. The prediction model generation unit 11b generates learning data (so-called training data) obtained by setting the operation data stored in the process DB 12b as input data (explanatory variables) and the sensor data corresponding to the operation data as the correct answer value (target variable), and generates the prediction model 5 by performing machine learning (so-called supervised learning) using the learning data. The prediction model generation unit 11b stores information related to the generated prediction model 5 in the prediction model storage unit 12c. In addition, the prediction model generation unit 11b may also repeat the generation (update) of the prediction model 5 at a predetermined period such as once a week or once a month.

[0054] In the present embodiment, the prediction model 5 can adopt, for example, a PLS (Partial Least Squares) regression model. However, the prediction model 5 can also adopt, for example, various learning models such as an ARX (Auto-Regressive with eXogenous) model, an SVM (Support Vector Machine), a random forest, or a neural network. In addition, for example, the prediction model 5 can also be a structure that accepts operation data in a time series as input and outputs sensor data in a time series. In this case, the prediction model 5 can adopt, for example, a learning model such as RNN (Recurrent Neural Network), LSTM (Long Short Term Memory), or Transfomer. In addition, the structure of these learning models and the method of generating learning models based on machine learning are existing technologies, so detailed descriptions are omitted in this embodiment.

[0055] The prediction model configuration unit 11c performs a process of configuring (deploying) the prediction model 5 generated by the prediction model generation unit 11b to the information processing device 1B that monitors and controls the semiconductor manufacturing device 2B that is not provided with the sensor 3. For example, the prediction model configuration unit 11c can receive input of information such as the device ID of the information processing device 1B as the configuration destination from the user, read out information related to the prediction model 5 from the prediction model storage unit 12c, and send the information related to the prediction model 5 to the information processing device 1B as the configuration destination that received the input, thereby configuring the prediction model 5 to the information processing device 1B. The information processing device 1B that receives the information related to the prediction model 5 from the information processing device 1A can store the received information in the prediction model storage unit 12c of its own storage unit 12, and use the prediction model 5 in subsequent processing. In addition, when the prediction model generation unit 11b repeatedly generates (updates) the prediction model 5, the prediction model configuration unit 11c can send (reconfigure) the information related to the updated prediction model 5 to the information processing device 1B that previously configured the prediction model 5 each time the prediction model 5 is updated.

[0056] The control processing unit 11d performs control processing of the semiconductor manufacturing device 2 based on the operation data acquired from the semiconductor manufacturing device 2, the sensor data acquired from the sensor 3, or the sensor data predicted by the prediction model 5. For example, the control processing unit 11d can monitor the change of the power consumption (power consumption amount) of each unit of the semiconductor manufacturing device 2 based on the sensor data, judge the operation state of each unit based on the power consumption, and stop the operation of the unit in the standby state (the state where no processing is performed) (stop the power supply), thereby performing control to reduce the power consumption of the entire semiconductor manufacturing device 2. In addition, for example, the control processing unit 11d can judge whether there is an abnormality in each unit based on the operation data or the sensor data, and when it is judged that there is an abnormality, control is performed to stop the operation of the unit or the semiconductor manufacturing device 2. In addition, the control processing of the semiconductor manufacturing device 2 performed by the control processing unit 11d can be any processing.

[0057] The display processing unit 11e performs processing to display various information on the display unit 15. For example, the display processing unit 11e can display the changes over time of various information included in the operation data and sensor data acquired by the data acquisition unit 11a in the form of a graph. In addition, for example, the display processing unit 11e can display various information such as information related to the prediction model 5 generated by the prediction model generation unit 11b, such as comparison information between the measured value measured by the sensor 3 and the predicted value obtained from the prediction model 5, and the prediction accuracy of the prediction model 5. In addition, for example, the display processing unit 11e can display the predicted values ​​of the power consumption, water discharge, or gas discharge of the semiconductor manufacturing device 2 on the display unit 15 based on the sensor data output by the prediction model 5 based on the operation data acquired by the data acquisition unit 11a. In addition, at this time, the display processing unit 11e can also perform a predetermined calculation on the predicted value of the power consumption of the semiconductor manufacturing device 2, thereby converting the power consumption into information such as the discharge amount of carbon dioxide and displaying it.

[0058] <Prediction Model Generation Process>

[0059] In the information processing system according to the present embodiment, for example, a user such as a designer or manager of the present system appropriately selects one or more semiconductor manufacturing apparatuses 2A from a plurality of semiconductor manufacturing apparatuses 2, and sets a sensor 3 for the selected semiconductor manufacturing apparatus 2A. The user implements a semiconductor manufacturing process in the semiconductor manufacturing apparatus 2A, and at this time, the information processing apparatus 1A acquires the operation data of the semiconductor manufacturing apparatus 2A and the sensor data of the sensor 3, and stores and accumulates them in the process DB 12b. After sufficient data is accumulated in the process DB 12b, the information processing apparatus 1A, for example, repeatedly performs machine learning using the accumulated operation data and sensor data according to the user's operation or at a predetermined cycle, thereby generating a prediction model 5. The information processing apparatus 1A sends information related to the generated prediction model 5 to the information processing apparatus 1B, thereby setting the prediction model 5 to a state in which the prediction model 5 can be used for monitoring and controlling the semiconductor manufacturing apparatus 2B that is not provided with the sensor 3.

[0060] Figure 5 1 is a flowchart showing an example of a process of prediction model generation processing performed by the information processing device 1A according to the present embodiment. When the semiconductor manufacturing process of the semiconductor manufacturing device 2A is implemented, the data acquisition unit 11a of the processing unit 11 of the information processing device 1A according to the present embodiment acquires the operation data output by the semiconductor manufacturing device 2A and the sensor data output by the sensor 3 at a predetermined period (step S1). The data acquisition unit 11a stores the operation data and sensor data acquired in step S1 in the process DB 12b of the storage unit 12 (step S2).

[0061] The processing unit 11 determines whether it is time to generate the prediction model 5 (step S3). At this time, the processing unit 11 can determine that it is time to generate the prediction model 5, for example, when it is time to update the prediction model 5 that is repeated at a predetermined period, when an instruction to generate the prediction model 5 is received from the user, or when a predetermined amount of operation data and sensor data is accumulated in the process DB 12b. If it is not time to generate the prediction model 5 (S3: "No"), the processing unit 11 returns the process to step S1 and continues to collect data through steps S1 and S2.

[0062] When the time has come to generate the prediction model 5 (S3: "Yes"), the prediction model generation unit 11b of the processing unit 11 reads the operating data and sensor data stored in the process DB 12b (step S4). The prediction model generation unit 11b generates learning data corresponding to the obtained operation data read in step S4 as input (explanatory variable) and the sensor data as the correct answer value (target variable) (step S5). In addition, at this time, the prediction model generation unit 11b can appropriately perform normalization or regularization of the numerical values ​​contained in the operation data and sensor data. The prediction model generation unit 11b uses the learning data generated in step S5 to perform supervised machine learning processing on the learning model of a predetermined structure (step S6) to determine the internal parameters of the learning model, thereby generating the prediction model 5.

[0063] The prediction model generation unit 11b uses the verification data generated separately from the learning data to make predictions for the prediction model 5 generated in step S6, thereby calculating the prediction accuracy (step S7). In addition, the verification data is data in the same format as the learning data. For example, a portion of all the data stored in the process DB 12b can be used as verification data, and the remaining data can be used as learning data. The prediction model generation unit 11b determines whether the prediction accuracy calculated in step S7 exceeds a threshold value (step S8). In addition, the threshold value is the prediction accuracy required for the prediction model 5, and is pre-determined by users such as designers or managers of the system. When the prediction accuracy does not exceed the threshold value (S8: "No"), the prediction model generation unit 11b returns the process to step S6 and repeats the machine learning process to improve the prediction accuracy.

[0064] When the prediction accuracy exceeds the threshold value (S8: Yes), the prediction model generation unit 11b stores information related to the generated prediction model in the prediction model storage unit 12c of the storage unit 12 (step S9). The display processing unit 11e of the processing unit 11 displays information such as the prediction accuracy of the generated prediction model 5 on the display unit 15 (step S10). The prediction model configuration unit 11c of the processing unit 11 sends information related to the generated prediction model 5 to the information processing device 1B that monitors and controls the semiconductor manufacturing device 2B without the sensor 3 (step S11), thereby configuring the prediction model 5 for the semiconductor manufacturing device 2B, and the processing ends.

[0065] Figure 6 and Figure 7 This is a schematic diagram showing an example of information display related to the generation of the prediction model 5. Figure 5 An example of information displayed in step S10 of the flowchart. Figure 6 In the information display example shown, the information processing device 1A displays a table representing the predicted errors of the power consumption of each unit of the semiconductor manufacturing device 2A and a graph representing the changes in the measured value and predicted value of the total power consumption over time side by side on the screen of the display unit 15. In the table representing the predicted errors, for example, the unit names of the units of the semiconductor manufacturing device 2A, the range of the measured values ​​of the power consumption, and the value [kW] and the ratio [%] of the maximum predicted error are shown in correspondence. In addition, in the graph of the total power consumption, the horizontal axis is set to time, the vertical axis is set to power consumption, the measured value of the total power consumption is represented by a light gray broken line, and the predicted value is represented by a dark gray broken line. According to the graph, it can be seen that the measured value oscillates up and down, and the predicted value represents a value close to the central value or average value of the measured value.

[0066] The information processing device 1A can calculate the illustrated prediction error and the like by verifying the generated prediction model 5 using the verification data prepared in advance. The verification data is data having the same structure as the learning data, which is generated using the operation data and sensor data stored in the process DB 12b. The information processing device 1A inputs the operation data included in the verification data to the prediction model 5, and obtains the predicted value of the sensor data output by the prediction model 5. The information processing device 1A can calculate the prediction error by calculating the difference between the actual measured value of the sensor data included in the verification data and the predicted value of the sensor data obtained from the prediction model 5.

[0067] In addition, Figure 7In the information display example shown, the information processing device 1A displays information for comparing the total power consumption consumed in each process for each process condition (process recipe) when the semiconductor manufacturing process is implemented by the semiconductor manufacturing device 2A. The information processing device 1A displays a table representing the prediction error of each process condition and a bar graph representing the total power consumption of each process condition side by side on the screen of the display unit 15. In the table representing the prediction error of each process condition, for example, for the three process conditions A to C, the measured value [kWh] and the predicted value [kWh] of the total power consumption are represented in correspondence with the value [kWh] and the ratio [%] of the prediction error. In addition, in the bar graph of the total power consumption, for example, for the three process conditions A to C, the measured value of the total power consumption is represented by light gray, and the predicted value is represented by dark gray.

[0068] <Use of prediction model 5>

[0069] In the information processing system according to the present embodiment, the prediction model 5 generated by the information processing device 1A can be used for various processes. Several examples of using the prediction model 5 will be described below.

[0070] (1) Use in simulation

[0071] Figure 8 This is a schematic diagram for explaining an example of using the prediction model 5 in a simulation. Figure 8 The upper part of shows the process of predicting the sensor data based on the operation data output by the semiconductor manufacturing device 2 by the prediction model 5. For example, a user such as a designer or manager of the present system generates process data of a semiconductor manufacturing process and inputs it to the semiconductor manufacturing device 2B. The semiconductor manufacturing device 2B implements the semiconductor manufacturing process based on the input process data and outputs operation data such as process log data. The information processing device 1B inputs the operation data obtained from the semiconductor manufacturing device 2B to the prediction model 5 generated in advance, and obtains the predicted value of the sensor data output by the prediction model 5.

[0072] exist Figure 8The lower half of the diagram shows a flow of using the prediction model 5 in the simulator 7 of the semiconductor manufacturing apparatus 2. The simulator 7 is software for simulating the semiconductor manufacturing process performed by the semiconductor manufacturing apparatus 2. The simulator 7, for example, receives process data including information such as settings of the semiconductor manufacturing process performed by the semiconductor manufacturing apparatus 2 as input, simulates the processing performed by the semiconductor manufacturing apparatus 2 according to the process data, and generates operation data output by the semiconductor manufacturing apparatus 2 as a simulation result. For example, the user can input the process data generated by the user to the simulator 7 for simulation, and input the operation data output by the simulator 7 to the prediction model 5, thereby obtaining the predicted value of the sensor data output by the prediction model 5. Based on the sensor data output by the prediction model 5, the user can verify the power consumption, water discharge, or gas discharge of the semiconductor manufacturing apparatus 2 when the processing is performed based on the process data generated by the user. For example, the user can repeatedly perform the simulation while applying corrections to the process data to improve the power consumption, water discharge, or gas discharge, etc.

[0073] In this way, by using the prediction model 5 generated by the information processing device 1A for simulation, the user can verify the power consumption, water discharge, gas discharge, etc. of the semiconductor manufacturing device 2 under various conditions without actually operating the semiconductor manufacturing device 2A. Figure 8 In the figure, the simulator 7 and the prediction model 5 are shown as different blocks, but the prediction model 5 can also be used by incorporating it into the simulator 7. In this case, the simulator 7 can output the predicted value of the sensor data together with the operation data. In addition, the simulation performed by the simulator 7 and the acquisition of the predicted value of the sensor data using the prediction model 5 can be performed by a device different from the information processing device 1 that monitors and controls the semiconductor manufacturing device 2, such as an information processing device for simulation.

[0074] (2) Contribution verification

[0075] The learning model generated by machine learning can calculate the contribution (importance, influence, etc.) of the input data (explanatory variables) to the output data (target variable) using XAI (eXplainable Artificial Intelligence) technologies such as SHAP (SHapley Additive Explanations) or LIME (Local Interpretable Model-agnostic Explanations). In the information processing system involved in this embodiment, the information processing device 1 can calculate the contribution of various items included in the operating data input to the prediction model 5 to the sensor data output by the prediction model 5, and display the calculation results on the display unit 15 to present to the user. The information processing device 1 can, for example, display information related to the contribution based on the user's operation, and can also display information related to the contribution when, for example, an abnormality is detected in the sensor data output by the prediction model 5.

[0076] Fig. 9 1 is a flowchart showing an example of the process of calculating the contribution degree by the information processing device 1 according to the present embodiment. The processing unit 11 of the information processing device 1 according to the present embodiment receives, for example, a designation of a range of operating data to be verified for the contribution degree from a user, and acquires the operating data to be the object from the operating data stored in the process DB 12b (step S21). The processing unit 11 inputs the operating data acquired in step S21 to the prediction model 5 in which the learned information such as the structure and parameters is stored in the prediction model storage unit 12c (step S22). The processing unit 11 acquires the sensor data output by the prediction model 5 for the operating data input in step S22 (step S23).

[0077] In addition, the processing unit 11 calculates the contribution of each item included in the operating data input to the prediction model 5 (step S24). At this time, the processing unit 11 can, for example, calculate the SHAP value (SHapely value) of each item included in each operating data, and use the average value or total value of the calculated SHAP value of each item as the contribution of each item in the operating data. In addition, the SHAP value is a numerical value indicating the extent to which each item of the input data of the learning model affects the output data. The calculation of the SHAP value is a prior art, so a detailed description of the calculation method is omitted. The processing unit 11 displays the contribution calculated in step S24 on the display unit 15 (step S25) and ends the processing.

[0078] In addition, in this example, SHAP values ​​are used to calculate the contribution of each item of the operating data, but the contribution can be calculated by methods other than the method using SHAP values. For example, when a learning model such as RNN (Recurrent Neural Network), LSTM (Long Short Term Memory) or Transformer is used as the prediction model 5, the contribution can also be calculated using the attention mechanism. In addition, for example, when a tree-structured learning model such as a decision tree, random forest, XGBoost (eXtreme Gradient Boosting) or LightGBM (Light Gradient Boosting Machine) is used as the prediction model 5, the contribution can also be calculated based on the structure of the generated learning model.

[0079] In addition, when the PLS regression model is used as the prediction model 5, the contribution can also be calculated based on the following calculation formula, for example. Regarding the item Ai (i = 1, 2 ... N) of the input data, the information processing device 1 can, for example, standardize the regression coefficient in such a way that the sum of the absolute values ​​of the standardized regression coefficients is 100, and the value obtained by converting the regression coefficient into a ratio (%) is the contribution.

[0080] [Number 1]

[0081]

[0082] Fig.10 is a schematic diagram showing an example of information display related to contribution. Fig. 9 An example of information displayed in step S25 of the flowchart. Fig.10 In the information display example shown, the information processing device 1 displays a table representing the contribution of each item included in the operating data and a pie chart representing the ratio of the contribution of multiple items included in the operating data side by side on the screen of the display unit 15. In the table representing the contribution of each item, for example, multiple items included in the operating data are arranged from top to bottom in order of contribution from high to low, and the ratio [%] of the contribution of each item to the whole is shown for each item. In addition, in this example, the top three items with the highest contribution among the multiple items included in the operating data are shown, and the total value of the contribution of the other items is shown as "others". In addition, the pie chart is a pie chart obtained by graphing the contribution shown in the above table and the area of ​​the sector corresponding to the ratio of the contribution of the multiple items included in the operating data to the whole after color differentiation.

[0083] The information processing device 1 calculates and displays the contribution of the plurality of items included in the operation data, so that the user can determine, for example, which item is effective to improve when the power consumption is high. In addition, the calculation and display of the contribution of the plurality of items included in the operation data may be performed by an information processing device different from the information processing device 1 that monitors and controls the semiconductor manufacturing device 2.

[0084] (3) Automatic stop of the unit in standby mode

[0085] The information processing device 1B according to the present embodiment acquires operation data when the semiconductor manufacturing device 2B implements the semiconductor manufacturing process, and acquires the predicted value of the sensor data based on the acquired operation data using the prediction model 5. For example, the information processing device 1B acquires the predicted value of the power consumption of each unit obtained from the prediction model 5, and when the predicted value of the power consumption of each unit is less than a threshold value, it is determined that the unit is in the standby state. The information processing device 1B can reduce the power consumption by stopping the operation of the unit in the standby state and stopping the power supply to the unit.

[0086] Fig.11 1 is a flowchart showing an example of a process of stopping a unit performed by the information processing device 1B according to the present embodiment. The processing unit 11 of the information processing device 1B according to the present embodiment acquires the operation data output by the semiconductor manufacturing device 2B when the semiconductor manufacturing device 2B performs the semiconductor manufacturing process (step S41). The processing unit 11 inputs the operation data acquired in step S41 to the prediction model 5 in which the learned information such as structure and parameters is stored in the prediction model storage unit 12c (step S42). The processing unit 11 acquires the sensor data output by the prediction model 5 in response to the operation data input in step S42 (step S43).

[0087] The processing unit 11 selects a unit as a processing object from a plurality of units provided in the semiconductor manufacturing apparatus 2B (step S44). The processing unit 11 obtains a predicted value of the power consumption related to the unit selected in step S44 from the sensor data obtained in step S43, and determines whether the power consumption is less than a predetermined threshold value (step S45). In addition, the threshold value is, for example, a value predetermined by a user such as a designer or manager of the information processing system involved in the present embodiment, and a different threshold value may be used for each unit. When the power consumption is less than the threshold value (S45: "Yes"), the processing unit 11 stops the operation of the unit as the object (step S46), and proceeds to step S47. When the power consumption is greater than the threshold value (S45: "No"), the processing unit 11 proceeds to step S47.

[0088] The processing unit 11 determines whether the power consumption determination has been completed for all units of the semiconductor manufacturing device 2B (step S47). When the power consumption determination has not been completed for all units (S47: "No"), the processing unit 11 returns the processing to step S44, selects other units and repeats the above processing. When the power consumption determination has been completed for all units (S47: "Yes"), the processing unit 11 determines whether the semiconductor manufacturing process of the semiconductor manufacturing device 2B has been completed (step S48). When the semiconductor manufacturing process has not been completed (S48: "No"), the processing unit 11 returns the processing to step S41, obtains new operating data from the semiconductor manufacturing device 2B and repeats the above processing. When the semiconductor manufacturing process has been completed (S48: "Yes"), the processing unit 11 ends the processing.

[0089] Fig.12 This is a schematic diagram for explaining the effect of automatic stop of the standby state unit. In this figure, the changes in power consumption of three units, namely, the RF power supply unit, the dry pump unit, and the heater, among the multiple units provided in the semiconductor manufacturing apparatus 2B are shown in graphs. Each graph is a line graph with time as the horizontal axis and power consumption of the unit as the vertical axis. Fig.12 The left side of is a graph showing a case where the automatic stop of the standby state unit is not performed, and the right side of is a graph showing a case where the automatic stop of the standby state unit is performed.

[0090] If the automatic stop is not performed, Fig.12 As shown in the graph on the left side of , each unit continues to consume standby power even in a standby state where no processing is actually performed (the area indicated by light gray in this figure). For example, a value with a margin relative to the standby power is set as a threshold in advance, and the information processing device 1B determines whether each unit is in a standby state based on the threshold, and automatically stops the operation of the unit whose power consumption exceeds the threshold. In addition, in the present embodiment, the automatic stop of the unit performed by the information processing device 1B is to switch the state of the unit to a state where standby power is not consumed, such as cutting off the power supply to the unit. Thus, as Fig.12 As shown in the graph on the right side of , the RF power supply unit and the dry pump unit determined to be in the standby state are automatically stopped by the information processing device 1B so that the power consumption of these units is reduced to a value less than the standby power amount, for example, 0.

[0091] In addition, in this example, the case where the information processing device 1B that monitors and controls the semiconductor manufacturing device 2B that is not provided with the sensor 3 uses the prediction model 5 to automatically stop the unit is described, but the information processing device 1A that monitors and controls the semiconductor manufacturing device 2A that is provided with the sensor 3 can also perform the same processing. However, the information processing device 1A can also perform the same automatic stop using the sensor data output by the sensor 3 instead of using the prediction model 5.

[0092] (4) Verification of processing timing

[0093] In the information processing system according to the present embodiment, the user can verify the processing timing of the plurality of units provided in the semiconductor manufacturing apparatus 2 based on the sensor data output by the prediction model 5. The information processing apparatus 1B reads out the operation data from the start to the end of the semiconductor manufacturing process for one wafer, for example, from the operation data acquired and stored in the process DB 12b when the semiconductor manufacturing apparatus 2B performs the semiconductor manufacturing process. The information processing apparatus 1B inputs the operation data read out from the process DB 12b to the prediction model 5, and acquires the sensor data output by the prediction model 5.

[0094] The semiconductor manufacturing device 2B is configured to include a plurality of units, and the sensor data includes, for example, predicted values ​​of power consumption, water discharge, or gas discharge of each of these units. In addition, regarding the water discharge or gas discharge, the sensor data only needs to include the value of the water discharge or gas discharge of the unit that discharges water or gas. Regarding power consumption, it is preferred that the sensor data include the value of the power consumption of all units that consume power. The information processing device 1B generates a graph showing the change over time of the power consumption, water discharge, or gas discharge of each unit based on the sensor data output by the prediction model 5, and displays the plurality of graphs generated for the plurality of units in parallel on the display unit 15.

[0095] Fig.13 This is a schematic diagram showing a display example related to processing timing verification. In the display example shown in the figure, the information processing device 1B generates a graph showing the change in power consumption over time for each of the three units 1 to 3 provided in the semiconductor manufacturing device 2B, and displays the three graphs side by side in the vertical direction. In addition, the information processing device 1B displays a graph showing the total power consumption of the three units 1 to 3 below the three graphs. Each graph of units 1 to 3 and the graph of total power consumption are graphs in which the horizontal axis is set to time and the vertical axis is set to power consumption, and preferably the vertical and horizontal axes of the four graphs have the same scale.

[0096] In addition, the information processing device 1B sets the chart based on the sensor data output by the prediction model 5 as the chart of the current power consumption, generates a chart of the power consumption when the processing timing of each unit is changed, and displays the chart of the current power consumption and the chart of the power consumption after the timing is changed side by side on the display unit 15. In the illustrated display example, according to the four charts displayed on the left side of the display unit 15, it can be seen that the current situation is that the three units 1 to 3 process at the same time, and the timing of the peak power consumption overlaps, so the maximum value of the total power consumption reaches about 15kW. In contrast, in the four charts displayed on the right side of the display unit 15, the timing of the processing of the three units 1 to 3 is different, and the maximum value of the total power consumption is suppressed to less than 12kW.

[0097] For example, the information processing device 1B receives input from the user regarding the degree to which the processing of each unit is to be changed, for a graph of current power consumption, and moves the graph of the current power consumption of each unit 1 to 3 in the direction of the time axis based on the received information, thereby generating a graph after the timing is changed. The information processing device 1B generates a graph of total power consumption based on the graph after the timing of the processing of each unit 1 to 3 is changed, and displays these graphs side by side on the display unit 15. In addition, the information processing device 1B may search for the timing of the processing when the maximum value of the total power consumption becomes the minimum, and generate and display the graph at this timing without receiving input from the user. In addition, the graph after the timing is changed may not be generated by moving the current graph in the direction of the time axis, but by using, for example Figure 8 The simulator 7 shown is generated.

[0098] Fig.14 FIG. 1 is a schematic diagram showing a display example related to processing timing verification. Fig.14 In the display example, the information processing device 1B displays a graph showing the change of the set temperature of the semiconductor manufacturing process over time and a graph showing the change of the power consumption of the semiconductor manufacturing device 2B over time side by side. The information processing device 1B extracts the information of the set temperature from the process data included in the operation data obtained from the semiconductor manufacturing device 2B, and generates a graph showing the change of the set temperature over time based on the extracted information. In addition, the information processing device 1B inputs the operation data to the prediction model 5, obtains the sensor data output by the prediction model 5, extracts the information of the power consumption included in the obtained sensor data, and generates a graph showing the change of the power consumption of the semiconductor manufacturing device 2B over time based on the extracted information.

[0099] In addition, the information processing device 1 generates a chart of the set temperature and power consumption when the timing of changing the set temperature of the semiconductor manufacturing device 2B is changed, for example, for the chart of the current set temperature and power consumption, and displays the current chart and the chart after the timing is changed side by side on the display unit 15. In the example shown in the figure, according to the chart displayed on the left, the current situation is that low-temperature treatment and high-temperature treatment are alternately performed for the set temperature, and when switching from low-temperature treatment to high-temperature treatment, power consumption increases. In contrast, in the chart displayed on the right, the timing of changing the temperature setting is changed so that low-temperature treatment is performed continuously first, and then switching from low-temperature treatment to high-temperature treatment is performed and high-temperature treatment is performed continuously. By performing such a timing change, the number of switching from low-temperature treatment to high-temperature treatment is reduced, and the number of times power consumption increases is reduced, so that the overall power consumption can be reduced.

[0100] The information processing device 1B receives input of a new setting temperature change pattern from the user, for example, with respect to the current setting temperature change pattern, and performs simulation based on the received content, for example, using the simulator 7. The information processing device 1B acquires the operation data output as a simulation result by the simulator 7, inputs the acquired operation data to the prediction model 5 to acquire sensor data, and generates a graph of power consumption.

[0101] The process related to the verification of the processing timing may be performed by a device different from the information processing device 1 that monitors and controls the semiconductor manufacturing apparatus 2 , and preferably by an information processing device that operates the simulator 7 .

[0102] <Conclusion>

[0103] In the information processing system according to the present embodiment having the above structure, when the semiconductor manufacturing apparatus 2A provided with a plurality of sensors 3 is operated, the information processing apparatus 1A acquires the operation data of the semiconductor manufacturing apparatus 2A and the sensor data output by the sensors 3, and generates a prediction model 5 that outputs a predicted value of the sensor data according to the input of the operation data by machine learning based on the acquired operation data and sensor data. The information processing apparatus 1A configures the prediction model 5 for the semiconductor manufacturing apparatus 2A by sending the generated prediction model 5 to the information processing apparatus 1B that monitors and controls the semiconductor manufacturing apparatus 2B not provided with the sensors 3. Thus, in the information processing system according to the present embodiment, for example, when the semiconductor manufacturing apparatus 2B provided with the sensors 3 is operated, the predicted value of the sensor data based on the operation data can be acquired and used for various controls or verifications.

[0104] In addition, in the information processing system according to the present embodiment, the information processing device 1B acquires the operation data of the semiconductor manufacturing device 2B, inputs the acquired operation data to the prediction model 5 that has been machine-learned in such a manner as to output sensor data based on the input of the operation data, acquires the sensor data output by the prediction model 5, and outputs information related to the sensor data. Thus, the information processing device 1B can predict the sensor data using the prediction model 5 generated in advance based on the operation data obtained from the semiconductor manufacturing device 2B that is not equipped with the sensor 3, so as to monitor and control the semiconductor manufacturing device 2B.

[0105] In the information processing system according to the present embodiment, it is preferred that the plurality of sensors 3 include a sensor for measuring power consumption, a sensor for measuring water discharge, or a sensor for measuring gas discharge. In addition, it is preferred that the operation data include process log data output by the semiconductor manufacturing device 2, process data set for the semiconductor manufacturing device 2, wafer transport history data of the semiconductor manufacturing device 2, or error data output by the semiconductor manufacturing device 2. Thus, a prediction model 5 for predicting power consumption, water discharge, or gas discharge of the semiconductor manufacturing device 2 can be generated based on data acquired from the semiconductor manufacturing device 2A and the sensor 3, and the power consumption, water discharge, or gas discharge of the semiconductor manufacturing device 2B without the sensor 3 can be predicted using the pre-generated prediction model 5.

[0106] In addition, in the information processing system according to the present embodiment, the information processing device 1B calculates the contribution of each item of the operating data to the sensor data output by the prediction model 5, and outputs information related to the calculated contribution. As a result, the user can expect to understand which item has a high contribution to the sensor data among the multiple items included in the operating data, and conduct research on the process of semiconductor manufacturing process, for example.

[0107] In the information processing system according to the present embodiment, the information processing device 1B determines the standby state of a plurality of units of the semiconductor manufacturing device 2B based on the sensor data output by the prediction model 5 and stops the operation of the units in the standby state. This can further reduce the power consumption of the semiconductor manufacturing device 2B.

[0108] In addition, in the information processing system according to the present embodiment, information on the temporal change of power consumption of a plurality of units provided in the semiconductor manufacturing apparatus 2B is outputted based on the sensor data outputted by the prediction model 5. Thus, for example, the user can expect to grasp the peak value of power consumption of each unit, and set the semiconductor manufacturing process of the semiconductor manufacturing apparatus 2B so as to avoid the peak values ​​of power consumption of a plurality of units from overlapping in time, thereby further reducing the peak value of power consumption of the semiconductor manufacturing apparatus 2B.

[0109] In addition, in the present embodiment, the information processing device 1A is configured to send the generated prediction model 5 to the information processing device 1B via the network N, but the present invention is not limited thereto. For example, the prediction model 5 may be sent and received between the information processing device 1A and the information processing device 1B via a recording medium such as a memory card or an optical disk, or may be sent and received by a method other than these methods. In addition, instead of the information processing device 1A that monitors and controls the semiconductor manufacturing device 2A, for example, an information processing device such as a server device that is different from the information processing devices 1A and 1B may acquire operation data and sensor data from the information processing device 1A to generate the prediction model 5, and the information processing device 5 such as the server device may send the prediction model to the information processing device 1B.

[0110] The embodiments disclosed herein are to be considered in all respects as illustrative rather than restrictive. The scope of the present disclosure is indicated by the claims rather than the above meaning, and is intended to include all modifications within the meaning and scope equivalent to the claims.

[0111] The matters described in each embodiment can be combined with each other. In addition, the independent claims and dependent claims described in the claims can be combined with each other in all arbitrary combinations regardless of the reference form. Moreover, the form of recording a claim that references two or more other claims (multiple claim form) is used in the claims, but it is not limited to this. It can also be recorded in the form of recording multiple claims (multiple references to multiple claims) that reference at least one multiple claim. For example, the present disclosure includes the following embodiments.

[0112] (Note 1)

[0113] An information processing method, comprising the following processing:

[0114] When a first semiconductor manufacturing device provided with a plurality of sensors is operated, operating data of the first semiconductor manufacturing device and sensor data output by the sensors are acquired;

[0115] Based on the acquired operation data and the sensor data, generating, by machine learning, a learning model that outputs the sensor data according to an input of the operation data; and

[0116] The generated learning model is disposed in a second semiconductor manufacturing apparatus different from the first semiconductor manufacturing apparatus.

[0117] (Note 2)

[0118] According to the information processing method described in Appendix 1,

[0119] The plurality of sensors include a sensor for measuring power consumption, a sensor for measuring a discharge amount, or a sensor for measuring a discharge amount of gas.

[0120] (Note 3)

[0121] According to the information processing method described in Appendix 1 or 2,

[0122] The operation data includes process log data output by the first semiconductor manufacturing device, processing process data set for the first semiconductor manufacturing device, wafer transport history data of the first semiconductor manufacturing device, or error data output by the first semiconductor manufacturing device.

[0123] (Note 4)

[0124] An information processing method according to any one of Appendixes 1 to 3, wherein:

[0125] The number of sensors provided in the second semiconductor manufacturing equipment is smaller than the number of sensors provided in the first semiconductor manufacturing equipment.

[0126] (Note 5)

[0127] An information processing method according to any one of Appendixes 1 to 4, wherein:

[0128] The learning model outputs power consumption, water discharge, or gas discharge of the semiconductor manufacturing apparatus when process log data outputted from the semiconductor manufacturing apparatus is inputted.

[0129] (Note 6)

[0130] An information processing method according to any one of Appendixes 1 to 5, wherein:

[0131] The first semiconductor manufacturing device used experimentally is provided with a predetermined number of sensors for measuring sensor data related to the environment, the plurality of sensors including a sensor for measuring power consumption, a sensor for measuring a discharge amount, and a sensor for measuring a discharge amount of gas.

[0132] When the second semiconductor manufacturing device is newly installed at one or more locations, the number of sensors less than the specified number or no sensors are installed in the second semiconductor manufacturing device, and the generated learning model is configured in the second semiconductor manufacturing device,

[0133] Operation data acquired in conjunction with initial operation of the second semiconductor manufacturing apparatus is input into the learning model, and sensor data is acquired without using the sensor.

[0134] (Note 7)

[0135] A computer program that causes a computer to:

[0136] When a first semiconductor manufacturing device provided with a plurality of sensors is operated, operating data of the first semiconductor manufacturing device and sensor data output by the sensors are acquired;

[0137] Based on the acquired operation data and the sensor data, generating, by machine learning, a learning model that outputs the sensor data according to an input of the operation data; and

[0138] Information related to the generated learning model is output as information for use in a second semiconductor manufacturing apparatus different from the first semiconductor manufacturing apparatus.

[0139] (Note 8)

[0140] A computer program that causes a computer to:

[0141] acquiring operation data of a second semiconductor manufacturing device;

[0142] Inputting the acquired operating data to a learning model obtained by machine learning in a manner of outputting sensor data based on input operating data, and acquiring sensor data output by the learning model; and

[0143] The acquired sensor data is outputted.

[0144] (Note 9)

[0145] A computer program according to Appendix 7 or 8, wherein:

[0146] The learning model is a learning model generated by performing machine learning in a manner that outputs the sensor data based on input of the operating data, based on operating data of the first semiconductor manufacturing device acquired when the first semiconductor manufacturing device equipped with multiple sensors is operated, and sensor data output by the sensors.

[0147] (Note 10)

[0148] A computer program according to any one of Appendixes 7 to 9, wherein:

[0149] The plurality of sensors include a sensor for measuring power consumption, a sensor for measuring a discharge amount, or a sensor for measuring a discharge amount of gas.

[0150] (Note 11)

[0151] A computer program according to any one of Appendixes 7 to 10, wherein:

[0152] The operation data includes process log data output by the first semiconductor manufacturing device, processing process data set for the first semiconductor manufacturing device, wafer transport history data of the first semiconductor manufacturing device, or error data output by the first semiconductor manufacturing device.

[0153] (Note 12)

[0154] A computer program according to any one of Appendixes 7 to 11, wherein:

[0155] The learning model outputs power consumption, water discharge, or gas discharge of the semiconductor manufacturing apparatus when process log data outputted from the semiconductor manufacturing apparatus is inputted.

[0156] (Note 13)

[0157] A computer program according to any one of Appendixes 7 to 12, wherein:

[0158] The operation data includes multiple items.

[0159] calculating the contribution of each item of the operating data to the sensor data output by the learning model,

[0160] The calculated contribution is outputted.

[0161] (Note 14)

[0162] A computer program according to any one of Appendixes 7 to 13, wherein:

[0163] determining the standby states of a plurality of units included in the second semiconductor manufacturing apparatus based on the sensor data output by the learning model,

[0164] Stops the operation of the unit in the standby state.

[0165] (Note 15)

[0166] A computer program according to any one of Appendixes 7 to 14, wherein:

[0167] The sensor data includes data on power consumption of each of a plurality of units included in the second semiconductor manufacturing apparatus.

[0168] Based on the sensor data output by the learning model, temporal changes in power consumption of a plurality of units are output.

[0169] (Note 16)

[0170] An information processing device comprising:

[0171] an acquisition unit that acquires operation data of the first semiconductor manufacturing apparatus and sensor data output by the sensors when a first semiconductor manufacturing apparatus provided with a plurality of sensors is operated;

[0172] a generating unit that generates, by machine learning, a learning model that outputs the sensor data according to an input of the operating data, based on the acquired operating data and the sensor data; and

[0173] An output unit outputs information related to the generated learning model as information for use in a second semiconductor manufacturing apparatus different from the first semiconductor manufacturing apparatus.

[0174] (Note 17)

[0175] An information processing device comprising:

[0176] a first acquisition unit that acquires operation data of a second semiconductor manufacturing device;

[0177] a second acquisition unit that inputs the acquired operating data into a learning model obtained by machine learning in a manner of outputting sensor data based on input operating data, and acquires the sensor data output by the learning model; and

[0178] An output unit outputs the acquired sensor data.

[0179] Description of Reference Numerals

[0180] 1, 1A, 1B: information processing device (computer); 2: semiconductor manufacturing device; 2A: semiconductor manufacturing device (first semiconductor manufacturing device); 2B: semiconductor manufacturing device (second semiconductor manufacturing device); 3: sensor; 5: prediction model; 11: processing unit; 11a: data acquisition unit (acquisition unit, first acquisition unit); 11b: prediction model generation unit (generation unit); 11c: prediction model configuration unit (configuration unit, output unit); 11d: control processing unit (second acquisition unit); 11e: display processing unit (output unit); 12: storage unit; 12a: program (computer program); 12b: process DB; 12c: prediction model storage unit; 13: communication unit; 14: input and output unit; 15: display unit; 16: operation unit; 99: recording medium; N: network.

Claims

1. An information processing method, comprising the following processing: When a first semiconductor manufacturing device provided with a plurality of sensors is operated, operating data of the first semiconductor manufacturing device and sensor data output by the sensors are acquired; Based on the acquired operation data and the sensor data, generating, by machine learning, a learning model that outputs the sensor data according to an input of the operation data; as well as The generated learning model is disposed in a second semiconductor manufacturing apparatus different from the first semiconductor manufacturing apparatus.

2. The information processing method according to claim 1, wherein: The plurality of sensors include a sensor for measuring power consumption, a sensor for measuring a discharge amount, or a sensor for measuring a discharge amount of gas.

3. The information processing method according to claim 1, wherein: The operation data includes process log data output by the first semiconductor manufacturing device, processing process data set for the first semiconductor manufacturing device, wafer transport history data of the first semiconductor manufacturing device, or error data output by the first semiconductor manufacturing device.

4. The information processing method according to claim 1, wherein: The number of sensors provided in the second semiconductor manufacturing equipment is smaller than the number of sensors provided in the first semiconductor manufacturing equipment.

5. The information processing method according to claim 1, wherein: The learning model outputs power consumption, water discharge, or gas discharge of the semiconductor manufacturing apparatus when process log data outputted from the semiconductor manufacturing apparatus is inputted.

6. The information processing method according to claim 1, wherein: The first semiconductor manufacturing device used experimentally is provided with a predetermined number of sensors for measuring sensor data related to the environment, the plurality of sensors including a sensor for measuring power consumption, a sensor for measuring a discharge amount, and a sensor for measuring a discharge amount of gas. When the second semiconductor manufacturing device is newly installed at one or more locations, the number of sensors less than the specified number or no sensors are installed in the second semiconductor manufacturing device, and the generated learning model is configured in the second semiconductor manufacturing device, Operation data acquired in conjunction with initial operation of the second semiconductor manufacturing apparatus is input into the learning model, and sensor data is acquired without using the sensor.

7. A computer program that causes a computer to execute the following processing: When a first semiconductor manufacturing device provided with a plurality of sensors is operated, operating data of the first semiconductor manufacturing device and sensor data output by the sensors are acquired; Based on the acquired operation data and the sensor data, generating, by machine learning, a learning model that outputs the sensor data according to an input of the operation data; as well as Information related to the generated learning model is output as information for use in a second semiconductor manufacturing apparatus different from the first semiconductor manufacturing apparatus.

8. A computer program that causes a computer to execute the following processing: acquiring operation data of a second semiconductor manufacturing device; Inputting the acquired operating data to a learning model obtained by machine learning in a manner of outputting sensor data based on input operating data, and acquiring sensor data output by the learning model; and The acquired sensor data is outputted.

9. The computer program according to claim 7, wherein: The learning model is a learning model generated by performing machine learning in a manner that outputs the sensor data based on input of the operating data, based on operating data of the first semiconductor manufacturing device acquired when the first semiconductor manufacturing device equipped with multiple sensors is operated, and sensor data output by the sensors.

10. The computer program according to claim 9, wherein: The plurality of sensors include a sensor for measuring power consumption, a sensor for measuring a discharge amount, or a sensor for measuring a discharge amount of gas.

11. The computer program according to claim 9, wherein: The operation data includes process log data output by the first semiconductor manufacturing device, processing process data set for the first semiconductor manufacturing device, wafer transport history data of the first semiconductor manufacturing device, or error data output by the first semiconductor manufacturing device.

12. The computer program according to claim 8, wherein: The learning model outputs power consumption, water discharge, or gas discharge of the semiconductor manufacturing apparatus when process log data outputted from the semiconductor manufacturing apparatus is inputted.

13. The computer program according to claim 8, wherein: The operation data includes multiple items. calculating the contribution of each item of the operating data to the sensor data output by the learning model, The calculated contribution is outputted.

14. The computer program according to claim 8, wherein: determining the standby states of a plurality of units included in the second semiconductor manufacturing apparatus based on the sensor data output by the learning model, Stops the operation of the unit in the standby state.

15. The computer program according to claim 8, wherein: The sensor data includes data on power consumption of each of a plurality of units included in the second semiconductor manufacturing apparatus. Based on the sensor data output by the learning model, temporal changes in power consumption of a plurality of units are output.

16. An information processing device comprising: an acquisition unit that acquires operation data of the first semiconductor manufacturing apparatus and sensor data output by the sensors when a first semiconductor manufacturing apparatus provided with a plurality of sensors is operated; a generating unit that generates, by machine learning, a learning model that outputs the sensor data according to an input of the operating data, based on the acquired operating data and the sensor data; as well as An output unit outputs information related to the generated learning model as information for use in a second semiconductor manufacturing apparatus different from the first semiconductor manufacturing apparatus.

17. An information processing device comprising: a first acquisition unit that acquires operation data of a second semiconductor manufacturing device; a second acquisition unit that inputs the acquired operating data into a learning model obtained by machine learning in a manner of outputting sensor data based on input operating data, and acquires the sensor data output by the learning model; and An output unit outputs the acquired sensor data.

Images