Diagnostic device, diagnostic system, and diagnostic method for process processing equipment.
The diagnostic device addresses false reports in process processing equipment by integrating sensor data and setting common thresholds, ensuring accurate diagnosis and efficient maintenance.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- HITACHI HIGH TECH CORP
- Filing Date
- 2023-09-15
- Publication Date
- 2026-07-01
Smart Images
Figure 0007883604000001 
Figure 0007883604000002 
Figure 0007883604000003
Abstract
Description
Technical Field
[0001] The present invention relates to a diagnostic device, a diagnostic system, and a diagnostic method for a process processing device that processes semiconductor wafers.
Background Art
[0002] A process processing device is a device that performs process processing for forming a fine shape on a semiconductor wafer. In a process processing device, maintenance work such as replacement or cleaning of components that are regularly subject to maintenance is usually performed based on the number of processed wafers, etc. However, due to component deterioration caused by aging changes and accumulation of reaction by-products depending on the usage method, unplanned maintenance work may occur. In order to reduce the downtime due to unplanned maintenance, it is required to sequentially monitor the deterioration state of components and take early measures such as cleaning and component replacement according to the deterioration state.
[0003] In order to realize such early measures, in a diagnostic device for a process processing device, first, using sensor waveform data, which is a time-series signal composed of a plurality of sensor items sequentially acquired for each process processing from a plurality of state sensors attached to the components of the process processing device (the data items of the sensor waveform data are called sensor items), a feature amount indicating the characteristics of the sensor waveform data is calculated. Further, based on the feature amount, the degree of deterioration indicating the deterioration state of the component is estimated from the deviation from the normal state, compared with a preset alarm threshold value to determine whether maintenance is required, and an alarm is issued.
[0004] Regarding diagnostic processing for process processing equipment, for example, International Publication No. 2018 / 061842 (Patent Document 1) states that "the anomaly detection device applies statistical modeling to a summary value compiled from observed values to estimate a state with noise removed from the summary value, and generates a predicted value that predicts the summary value one period ahead based on this estimation. Based on the predicted value, the anomaly detection device detects whether or not there is an anomaly in the monitored equipment." Furthermore, Japanese Patent Application Publication No. 2012-9064 (Patent Document 2) states that "a learning-type process anomaly diagnostic device that achieves accurate anomaly detection as well as anomaly diagnostic performance suitable for practical process monitoring." These prior arts are, so to speak, individual approaches that statistically process each sensor waveform data individually and estimate degradation by comparing the predicted value with the current value. [Prior art documents] [Patent Documents]
[0005] [Patent Document 1] International Publication No. 2018 / 061842 [Patent Document 2] Japanese Patent Publication No. 2012-9064 [Overview of the project] [Problems that the invention aims to solve]
[0006] However, in the diagnosis of process equipment, the prior art is thought to generate false reports regarding the diagnosis of degradation due to the following factors, making it difficult to take effective countermeasures based on the diagnostic results. Firstly, in process processing equipment, the state of the equipment's components changes according to the processing history, which can lead to false reports due to initial state differences between components in a group of process processing equipment. Individual diagnostic approaches, such as those described in Patent Documents 1 and 2, do not address the issue of variability among the components of a group of process equipment.
[0007] Secondly, false alarms can occur due to the setting of alarm thresholds. When thresholds are set based on the operator's past knowledge, the occurrence of false alarms depends on the operator's experience. In addition, thresholds may be set by establishing control standards such as 3σ for the variation in the normal state of the feature quantities of each sensor in each process processing device. In this case, since process processing devices are equipped with various sensors, they interfere with each other due to control and are affected by the device state, so it is rare for the output to be stable for all sensor items. Therefore, when diagnosing deviations from the normal state for various sensor items, false alarms may occur frequently. Furthermore, when thresholds are set based on sensor values at the time of degradation events, the frequency of degradation events is relatively low, making the setting itself difficult. Individual approaches to diagnosis, such as those described in Patent Documents 1 and 2, do not recognize the challenges related to setting degradation state thresholds that correspond to differences in devices and various sensor items.
[0008] Thirdly, because multiple sensors attached to the components of a process processing apparatus interfere with each other, diagnostic methods based on feature quantities calculated for each sensor, as in the individual approaches described in Patent Documents 1 and 2, may not necessarily directly reflect the deterioration state of the component, potentially leading to false information. Therefore, the present invention aims to provide a technology that enables efficient countermeasures by suppressing false reports generated for components whose status is measured by multiple sensors, or for each component of a group of process processing devices, in the diagnosis of deterioration of a process processing device. [Means for solving the problem]
[0009] To solve the above problems, one representative diagnostic device for process processing equipment of the present invention is a diagnostic device for process processing equipment that targets the components of the process processing equipment for maintenance, and comprises a feature change calculation unit that calculates a degree of degradation corresponding to a change index from the feature quantities of the sensor waveform data of each sensor item for a plurality of sensors (state sensor group) that measure the state of the components, and a component degradation calculation unit that calculates a component degradation degree that corresponds one-to-one with the component using a plurality of the said degradation degrees. [Effects of the Invention]
[0010] According to the present invention, in diagnosing the deterioration of a process processing apparatus, it is possible to provide a technology that suppresses false reports generated for components whose status is measured by multiple sensors, or for each component of a group of process processing apparatuses, thereby enabling efficient countermeasures. Other issues, configurations, and effects not mentioned above will be clarified by the following description of the embodiments. [Brief explanation of the drawing]
[0011] [Figure 1] Figure 1 is an overall configuration diagram of the process processing apparatus and the diagnostic apparatus for the process processing apparatus. [Figure 2] Figure 2 is a diagram showing the relationship between component degradation levels and common thresholds for the equipment. [Figure 3] Figure 3 is a flowchart showing an example of the processing flow on computer 30. [Figure 4] Figure 4 is a flowchart showing an example of the processing flow on server 50. [Figure 5] Figure 5 is a diagram illustrating an example of a method for setting common thresholds for the device. [Figure 6] Figure 6 shows an example of the display screen in the diagnostic result display unit 512. [Modes for carrying out the invention]
[0012] Embodiments of the present invention will be described below with reference to the drawings. In all drawings used to illustrate the embodiments, the same parts will be denoted by the same reference numerals, and repeated descriptions of them will be omitted.
[0013] (Processing equipment) Figure 1 is an overall configuration diagram of a process processing apparatus and a diagnostic apparatus for the process processing apparatus. As shown in Figure 1, the process processing apparatus A(10), process processing apparatus B(11), etc., which constitute the process processing apparatus group 1 in this embodiment, process a wafer (sample 101) according to preset process processing conditions. In addition, each process processing apparatus is equipped with the same type of component, and each component is attached to a state sensor group 102 consisting of multiple sensors that measure the state of the component, so that measured values of the sensor (e.g., temperature and pressure) during processing or idle can be acquired as sensor waveform data. Examples of process processing apparatuses, components, and state sensors include a plasma processing apparatus, a microwave generation unit, and current / voltage sensors, respectively.
[0014] (Diagnostic device) As shown in Figure 1, the diagnostic device 2 comprises a group of computers (computers 30, 40, ...) consisting of an execution unit 31 that acquires sensor waveform data corresponding to each process processing device of the process processing device group 1 and performs calculation processing, and a storage unit 32 that stores information necessary for the execution unit's processing. Furthermore, the diagnostic device 2 comprises a server 50 consisting of an analysis unit 51 that sets calculation conditions and common thresholds for the execution unit 31 and displays diagnostic results, and a storage unit 52 that stores information necessary for the analysis unit 51's processing. The process processing unit group 1 is connected to the computer group (computers 30, 40, ...) either directly or via a network. The computer group (computers 30, 40, ...) and the server 50 are also connected via a network. This allows each computer to perform calculations at high speed using sensor waveform data acquired from each process processing unit in the execution unit 31. The server 50 can also set calculation conditions and thresholds across the process processing unit group 1, and analyze and display diagnostic results. However, the connection between the computer group and the server 50 is not necessarily limited to a network; direct connections may also be made at the processing units.
[0015] (calculator) The execution unit 31 is composed of a processor such as a CPU or a GPU, and includes a preprocessing unit 310, a feature quantity change calculation unit 311, a diagnosis and reporting unit 312, a feature quantity calculation unit 313, and a component degradation degree calculation unit 314 as described later. The storage unit 32 is composed of a memory, a hard disk, etc., and includes a sensor waveform storage unit 320 and a calculation result storage unit 321 as described later.
[0016] (Server) The analysis unit 51 is composed of a processor such as a CPU or a GPU, and includes an operation condition setting unit 510, a device common threshold setting unit 511, and a diagnosis result display unit 512 as described later. The storage unit 52 is composed of a memory, a hard disk, etc., and includes a diagnosis result storage unit 520 and a maintenance timing information storage unit 521 as described later.
[0017] Figure 2 is a relational diagram showing the positioning of the component degradation degree and the device common threshold, etc. The process processing device has one or more components, and a plurality of sensors (state sensor groups) are added to each component. Then, one or more feature quantities are calculated for each sensor item of each sensor, and one or more degradation degrees are calculated for each feature quantity. In this way, the component degradation degree corresponding one-to-one to the component is obtained by calculation from the plurality of degradation degrees obtained from one component. Also, a device common threshold serving as a reference for degradation detection is set from the component degradation degrees of the same type of components provided in a plurality of process processing devices (process processing device group).
[0018] (Processing on a computer) An example of the diagnosis process of the degradation state of each component of the process processing device 10 performed by the computer 30 will be described. Figure 3 is a flowchart showing an example of the processing flow on the computer 30. Note that Figure 3 shows the processing flow for each diagnosis cycle. The diagnosis cycle may be, for example, for each process or at regular time intervals.
[0019] First, the computer 30 acquires sensor waveform data for the diagnostic cycle from the sensor waveform storage unit 320. The sensor waveform data is configured so that for each of the multiple sensor items, the sensor value is stored in a time series such as processing time or time. At this point, the pre-processing unit 310 selects one sensor item from the sensor waveform data (S1). Next, the preprocessing unit 310 performs preprocessing on the sensor waveform data of the selected sensor item (S2). The method of preprocessing is arbitrary, but for example, a process ID, which is information for identifying process conditions, is used to specify the process ID to be used for diagnosis, and the sensor waveform data associated with it is extracted. In addition, processing time intervals to be used for diagnosis are extracted, and missing values are removed and the data is standardized to make it easier to handle.
[0020] Next, the feature calculation unit 313 calculates feature quantities that represent the characteristics of the sensor waveform data using the pre-processed sensor waveform data (S3). As shown in Figure 2, one or more feature quantities are calculated for each sensor item. Statistical quantities such as the mean value and standard deviation of the sensor waveform data can be used as feature quantities. For example, in a plasma processing apparatus, the mean value over a specific time interval can be used as a feature quantity for the waveform data of the current value. Feature quantities with a high probability of being related to degradation may be calculated by comparing them with domain knowledge of the component, or multiple feature quantities may be calculated. Next, the feature change calculation unit 311 determines whether the initial state period has been exceeded (S4). The initial state period refers to the time interval from immediately after maintenance work on the component to a certain period of time. If the initial state interval has not been exceeded in the diagnostic cycle at the time of determination, the preprocessing unit 310 determines (S6) whether the feature calculation has been completed for all sensor items targeted for feature calculation. If it has been completed, the processing for the current diagnostic cycle ends. If the calculation has not been completed in S6, the process moves on to selecting another sensor item and continues.
[0021] Regarding S4, if the initial state interval has been exceeded, the feature change calculation unit 311 calculates the degree of degradation for each calculated feature by taking the change from its initial state (S5). For each feature, the degree of degradation is calculated using multiple change indicators. Examples of change indicators include the difference or ratio between the average value of the feature in the initial state interval and the feature in the current diagnostic cycle, or the Mahalanobis distance between the feature in the initial state interval and the feature in the current diagnostic cycle. In this way, each process processing unit of the process processing unit group 1 defines the degree of degradation based on the change from the initial state, making it possible to compare the degree of degradation between process processing units. Furthermore, since the change indicator to be compared depends on the sensor item, multiple change indicators are defined and multiple degrees of degradation are calculated based on them. For example, if the change from the initial state is a significant physical quantity, the difference is considered a valid change indicator, and if the rate of change is a significant physical quantity, the ratio is considered a valid change indicator. Also, if the feature has a lot of variation even in the initial state, the Mahalanobis distance that takes variation into account is considered a valid indicator. Next, the component degradation calculation unit 314 proceeds to determine (S7) whether the degradation calculation has been completed for all sensor items. If it has not been completed, it repeats the same process for other sensor items. If it has been completed, it proceeds to process S8.
[0022] In S8, the component degradation calculation unit 314 calculates an index that integrates the degradation levels calculated for each component into a single component degradation index. It also calculates the contribution of each degradation level to the component degradation index. The calculation conditions for the component degradation index are calculated in the calculation condition setting unit 510, and an example of the calculation method will be described later. It is assumed that the degradation of components in a process processing apparatus may be due to the overlapping of multiple degradation factors, or even a single degradation factor may affect the degradation state of multiple sensor items. A comprehensive approach using the component degradation index calculated from individual degradation levels enables the calculation and diagnosis of degradation levels that take into account the interactions between multiple sensor items. Next, the diagnostic and alarm unit 312 determines when the calculation of the component degradation degree is complete for all components to be diagnosed, and repeats this process until the calculation is complete (S9).
[0023] Finally, the diagnostic and alarm generation unit 312 compares the device common threshold value, which has been pre-set for each component by the device common threshold setting unit 511, with the value calculated for the degree of component deterioration during the current diagnostic cycle. If the device common threshold value is exceeded, the unit notifies the server 50 to generate an alarm (S10). The series of calculation results calculated by the computer 30 are stored in the calculation result storage unit 321.
[0024] (Server processing) This section describes an example of the process performed by server 50 to set calculation conditions and common thresholds for the computer 30. Figure 4 is a flowchart showing an example of the processing flow on server 50. Note that the processing flow for setting calculation conditions and common thresholds for the devices only needs to be performed when it becomes necessary to set or update new calculation conditions or common threshold values.
[0025] First, the calculation condition setting unit 510 obtains the calculation result of the degree of degradation in the process processing apparatus group 1 and the maintenance timing information from the calculation result storage unit 321 and the maintenance timing information storage unit 521, respectively, for the component of interest (T1). The maintenance timing information includes historical information such as the time and content of maintenance performed in response to the degradation of the components of each process processing apparatus.
[0026] Next, in the calculation condition setting unit 510, for the process processing device of interest, the degradation level in the interval of the maintenance interval is extracted from the time-series trend of the degradation level calculated by the computer 30 for each diagnostic cycle, based on the maintenance timing information (T2). As a result, the left end of the time-series trend of the degradation level represents the initial state, and the right end represents the degradation state. Since the computer 30 calculates the degradation level multiple times, there are also multiple time-series trends of the degradation level. For process processing devices that do not include maintenance timing information, i.e., new process processing devices or process processing devices that have been in use since maintenance (where no degradation events have occurred), the degradation level trend from the initial state to the latest calculation point is extracted. Since the occurrence of degradation events is relatively rare, utilizing data from process processing devices that do not include maintenance timing information is effective in suppressing false reports.
[0027] Next, the calculation condition setting unit 510 calculates the component degradation level using the time-series trends of multiple degradation levels (T3). Various methods can be considered for calculating the component degradation level, but for example, a model can be constructed that calculates the component degradation level by applying a machine learning method that includes the structure of a Decision Tree. When constructing the model, the model is made with each degradation level trend as an explanatory variable and the degradation timing as the dependent variable, and the model parameters are optimized according to the data. From the viewpoint of narrowing down the effective features for each sensor item, a machine learning method with a feature selection function may be used to select one degradation level for each sensor item and calculate the component degradation level, as shown in Figure 2. Furthermore, from the viewpoint of explainability of the diagnostic results, it is desirable to have a model that can also calculate the contribution of each degradation level to the component degradation level. In addition, multiple models may be constructed. In the device common threshold setting unit 511, it is determined that the calculation is complete for all process processing devices to be diagnosed, and the processes from T2 to T3 are repeated for each process processing device to calculate the degree of component degradation of the component of interest (T4).
[0028] Next, the common threshold setting unit 511 sets a common threshold for each component degradation level across the process processing unit group 1 (T5). Figure 5 is a diagram illustrating an example of the common threshold setting method for the devices. Figure 5 shows the state after calculating the time-series trend of component degradation levels across the process processing unit group 1. The set thresholds are scanned comprehensively, and a common threshold is set between devices that maximizes the prediction accuracy index for the desired degradation timing. To suppress false reports, for example, the recall rate under conditions where the precision rate is above a certain threshold is used as the prediction accuracy index. If multiple component degradation calculation models are constructed in T3, the component degradation calculation model to be used for diagnosis by the computer 30 is selected based on the prediction accuracy index. This comprehensive approach of setting the common threshold for the devices to optimize the prediction accuracy index makes it possible to suppress false reports caused by individual device factors.
[0029] Finally, in the calculation condition setting unit 510, the selected component degradation calculation model (calculation conditions) and the common threshold between devices are set as conditions to be used for diagnosis in the computer 30 (T6). Next, the calculation condition setting unit 510 determines that the settings are complete for all components to be diagnosed, repeats the processes from T1 to T6 (T7), and then terminates the process.
[0030] Figure 6 shows an example of the display screen in the diagnostic result display unit 512. As shown in Figure 6(a), the time-series trend of component degradation and the common threshold values between devices are displayed for all components targeted for diagnosis in the process processing apparatus group 1. This allows for a quick understanding of which components of which process processing apparatus are close to a degraded state. Furthermore, if an error occurs in a process processing apparatus with an unknown cause, efficient countermeasures can be taken by prioritizing maintenance work on components that are close to a degraded state. Furthermore, as shown in Figure 6(b), specifying a component ID allows you to check the contribution of each degradation level (degradation calculation condition) to the component degradation level. This provides supplementary information for factor estimation by matching with the domain knowledge of the component, enabling more efficient countermeasures.
[0031] Although embodiments have been described above, the present invention is not limited to the embodiments described above, and various modifications are possible without departing from the spirit of the invention. For example, if it is difficult to install the server 50, the computer 30 may be configured to perform part of the processing of the server 50.
[0032] The following describes, but is not limited to, embodiments that may constitute the present invention. (Aspect 1) A diagnostic device for a process processing apparatus, which is intended to maintain the components of the process processing apparatus, A feature change calculation unit calculates the degree of degradation corresponding to a change index from the feature quantities of the waveform data of each sensor item corresponding to multiple sensors (state sensor group) that measure the state of a component, A component degradation degree calculation unit that calculates a component degradation degree that corresponds one-to-one with the component using a plurality of degradation degrees, A diagnostic device for process processing equipment equipped with the following features. (Aspect 2) A diagnostic device for a process processing apparatus according to embodiment 1, comprising a diagnostic and alarm unit that diagnoses and alarms by comparing a common threshold set among the components of a group of process processing apparatuses with the degree of degradation of the components. (Aspect 3) The diagnostic device for a process processing apparatus according to embodiment 1 or 2, characterized in that the feature change calculation unit uses at least one of the following as the change index: the difference between the feature in the diagnostic cycle and the average value of the feature in the initial state interval of the component (hereinafter referred to as "initial feature"); the ratio to the average value of the initial feature; or the Mahalanobis distance to the average value of the initial feature. (Aspect 4) A diagnostic device for a process processing apparatus according to any one of embodiments 1 to 3, characterized in that the component degradation calculation unit calculates the component degradation as a continuous probability value using a method that applies a regression model including the structure of a Decision Tree. (Aspect 5) A diagnostic device for a process processing apparatus according to embodiment 2, comprising a common threshold setting unit that sets a common threshold for the apparatus by comprehensively searching for an accuracy index for predicting the timing of deterioration from the time-series trend of the degree of deterioration of each component in the group of process processing apparatuses. (Aspect 6) A diagnostic device for a process processing apparatus according to embodiment 5, characterized in that the device common threshold setting unit uses precision and recall as prediction accuracy indicators to set the device common threshold such that the precision is equal to or greater than a standard value and the recall is maximized. (Aspect 7) A diagnostic device for a process processing apparatus according to any one of embodiments 2, 5, or 6, comprising a diagnostic result display unit that displays together the time-series trend of the degree of component degradation in each component of the group of process processing apparatuses and the contributions of a plurality of degradation levels used in calculating the degree of component degradation. (Pattern 8) A diagnostic system for a process processing apparatus that maintains the components of the process processing apparatus, A feature change calculation unit calculates the degree of degradation corresponding to a change index from the feature quantities of the waveform data of each sensor item for multiple sensors (state sensor group) that measure the state of a component, A component degradation degree calculation unit that calculates a component degradation degree that corresponds one-to-one with the component using a plurality of degradation degrees, A diagnostic system for process processing equipment equipped with the following features. (Aspect 9) A diagnostic system for a process processing apparatus according to embodiment 8, comprising a diagnostic and alarm unit that diagnoses and alarms by comparing a common threshold set among the components of a group of process processing apparatuses with the degree of degradation of the component. (Aspect 10) A diagnostic method for a process processing apparatus, in which the components of the process processing apparatus are the target of maintenance, In the feature change calculation unit, the steps include: calculating the degree of degradation corresponding to the change index from the feature quantities of the sensor waveform data of each sensor item for multiple sensors (state sensor group) that measure the state of a component; The component degradation calculation unit includes the steps of calculating a component degradation degree that corresponds one-to-one with the component using a plurality of degradation degrees, A diagnostic method for a process processing apparatus having [a certain characteristic]. (Aspect 11) A diagnostic method for a process processing apparatus according to embodiment 10, comprising the step of diagnosing and issuing an alert by comparing a common threshold set among the components of the process processing apparatus group with the degree of degradation of the component in the diagnostic and alerting unit. [Explanation of Symbols]
[0033] 1…Process processing equipment group, 2…Diagnostic equipment, 10, 11…Process processing equipment 30, 40...Computer, 31...Execution unit, 32, 52...Storage unit, 50...Server 51…Analysis unit, 310…Preprocessing unit, 311…Feature quantity change calculation unit, 312…Diagnosis / Reporting unit 313...Feature extraction unit, 314...Component degradation calculation unit 510...Calculation condition setting unit, 511...Device common threshold setting unit, 512...Diagnostic result display unit 520...Diagnostic result storage unit, 521...Maintenance timing information storage unit
Claims
1. A diagnostic device for a process processing apparatus, wherein the components of the process processing apparatus are the target of maintenance, A feature change calculation unit calculates the degree of degradation corresponding to a change index from the feature quantities of the waveform data of each sensor item for multiple sensors (state sensor group) that measure the state of a component, The system includes a component degradation degree calculation unit that calculates a component degradation degree that corresponds one-to-one with the component using a plurality of degradation degrees, A diagnostic device for a process processing apparatus, characterized in that the feature change calculation unit uses at least one of the following as the change index: the difference between the feature in the diagnostic cycle and the average value of the feature in the initial state interval of the component (hereinafter referred to as "initial feature"), the ratio to the average value of the initial feature, or the Mahalanobis distance to the average value of the initial feature.
2. A diagnostic device for a process processing apparatus, wherein the components of the process processing apparatus are the target of maintenance, A feature change calculation unit calculates the degree of degradation corresponding to a change index from the feature quantities of the waveform data of each sensor item for multiple sensors (state sensor group) that measure the state of a component, The system includes a component degradation degree calculation unit that calculates a component degradation degree that corresponds one-to-one with the component using a plurality of degradation degrees, A diagnostic device for a process processing apparatus, characterized in that the component degradation degree calculation unit calculates the component degradation degree as a continuous probability value using a method that applies a regression model including the structure of a Decision Tree.
3. A diagnostic device for a process processing apparatus, wherein the components of the process processing apparatus are the target of maintenance, A feature change calculation unit calculates the degree of degradation corresponding to a change index from the feature quantities of the waveform data of each sensor item for multiple sensors (state sensor group) that measure the state of a component, A component degradation degree calculation unit that calculates a component degradation degree that corresponds one-to-one with the component using a plurality of degradation degrees, A diagnostic device for a process processing apparatus, comprising: a common threshold setting unit that sets a common threshold for the apparatus by comprehensively searching for an accuracy index for predicting the timing of degradation from the time-series trend of the degree of component degradation in each component of a group of process processing apparatuses.
4. A diagnostic device for a process processing apparatus, wherein the components of the process processing apparatus are the target of maintenance, A feature change calculation unit calculates the degree of degradation corresponding to a change index from the feature quantities of the waveform data of each sensor item for multiple sensors (state sensor group) that measure the state of a component, A component degradation degree calculation unit that calculates a component degradation degree that corresponds one-to-one with the component using a plurality of degradation degrees, A diagnostic device for a process processing apparatus, comprising: a diagnostic result display unit that displays the time-series trend of the degree of component degradation in each component of a group of process processing apparatuses, and the contributions of multiple degrees of degradation used in the calculation of the degree of component degradation.
5. A diagnostic device for a process processing apparatus according to claim 1 or 2, comprising a diagnostic and alarm unit that diagnoses and alarms by comparing a common threshold set among the components of a group of process processing apparatuses with the degree of degradation of the component.
6. The diagnostic device for a process processing apparatus according to claim 3, characterized in that the common threshold setting unit for the apparatus uses precision and recall as prediction accuracy indicators to set the common threshold for the apparatus such that the precision is equal to or greater than a standard value and the recall is maximized.
7. A diagnostic system for a process processing apparatus, wherein the components of the process processing apparatus are the target of maintenance, A feature change calculation unit calculates the degree of degradation corresponding to a change index from the feature quantities of the waveform data of each sensor item for multiple sensors (state sensor group) that measure the state of a component, The system includes a component degradation degree calculation unit that calculates a component degradation degree that corresponds one-to-one with the component using a plurality of degradation degrees, A diagnostic system for a process processing apparatus, characterized in that the feature change calculation unit uses at least one of the following as the change index: the difference between the feature in the diagnostic cycle and the average value of the feature in the initial state interval of the component (hereinafter referred to as "initial feature"), the ratio to the average value of the initial feature, or the Mahalanobis distance to the average value of the initial feature.
8. A diagnostic system for a process processing apparatus, wherein the components of the process processing apparatus are the target of maintenance, A feature change calculation unit calculates the degree of degradation corresponding to a change index from the feature quantities of the waveform data of each sensor item for multiple sensors (state sensor group) that measure the state of a component, The system includes a component degradation degree calculation unit that calculates a component degradation degree that corresponds one-to-one with the component using a plurality of degradation degrees, A diagnostic system for a process processing apparatus, characterized in that the component degradation degree calculation unit calculates the component degradation degree as a continuous probability value using a method that applies a regression model including the structure of a Decision Tree.
9. A diagnostic system for a process processing apparatus, wherein the components of the process processing apparatus are the target of maintenance, A feature change calculation unit calculates the degree of degradation corresponding to a change index from the feature quantities of the waveform data of each sensor item for multiple sensors (state sensor group) that measure the state of a component, A component degradation degree calculation unit that calculates a component degradation degree that corresponds one-to-one with the component using a plurality of degradation degrees, A diagnostic system for a process processing apparatus, comprising: a common threshold setting unit that sets a common threshold for the apparatus by comprehensively searching for an accuracy index for predicting the timing of degradation from the time-series trend of the degree of degradation of each component in a group of process processing apparatuses.
10. A diagnostic system for a process processing apparatus, wherein the components of the process processing apparatus are the target of maintenance, A feature change calculation unit calculates the degree of degradation corresponding to a change index from the feature quantities of the waveform data of each sensor item for multiple sensors (state sensor group) that measure the state of a component, A component degradation degree calculation unit that calculates a component degradation degree that corresponds one-to-one with the component using a plurality of degradation degrees, A diagnostic system for a process processing apparatus, comprising: a diagnostic result display unit that displays the time-series trend of the degree of component degradation in each component of a group of process processing apparatuses, and the contributions of multiple degrees of degradation used in calculating the degree of component degradation.
11. A method for diagnosing a process processing apparatus, wherein the components of the process processing apparatus are the target of maintenance, The first step involves calculating the degree of degradation corresponding to the change index from the characteristic quantities of the waveform data of each sensor item for multiple sensors (a group of state sensors) that measure the state of a component, and The second step involves calculating a component degradation degree that corresponds one-to-one with the component using a plurality of degradation degrees, A method for diagnosing a process processing apparatus, characterized in that the first step uses at least one of the following as the change indicator: the difference between the feature quantity in the diagnostic cycle and the mean value of the feature quantity in the initial state interval of the component (hereinafter referred to as "initial feature quantity"), the ratio to the mean value of the initial feature quantity, or the Mahalanobis distance to the mean value of the initial feature quantity.
12. A method for diagnosing a process processing apparatus, wherein the components of the process processing apparatus are the target of maintenance, The first step involves calculating the degree of degradation corresponding to the change index from the characteristic quantities of the waveform data of each sensor item for multiple sensors (a group of state sensors) that measure the state of a component, and The second step involves calculating a component degradation degree that corresponds one-to-one with the component using a plurality of degradation degrees, A method for diagnosing a process processing apparatus, characterized in that, in the second step, the degree of component degradation is calculated as a continuous probability value using a method that applies a regression model including the structure of a Decision Tree.
13. A method for diagnosing a process processing apparatus, wherein the components of the process processing apparatus are the target of maintenance, A step of calculating the degree of degradation corresponding to the change index from the characteristic quantities of the waveform data of each sensor item for multiple sensors (a group of state sensors) that measure the state of a component, A step of calculating a component degradation degree that corresponds one-to-one with the component using a plurality of degradation degrees, A diagnostic method for a process processing apparatus, comprising the steps of setting a common threshold for the apparatus by comprehensively searching for an indicator of the accuracy of predicting the timing of degradation from the time-series trend of the degree of component degradation in each component of a group of process processing apparatuses.
14. A method for diagnosing a process processing apparatus, wherein the components of the process processing apparatus are the target of maintenance, A step of calculating the degree of degradation corresponding to the change index from the characteristic quantities of the waveform data of each sensor item for multiple sensors (a group of state sensors) that measure the state of a component, A step of calculating a component degradation degree that corresponds one-to-one with the component using a plurality of degradation degrees, A method for diagnosing a process processing apparatus, comprising the steps of displaying a time-series trend of the degree of component degradation in each component of a group of process processing apparatuses, and a plurality of contributions of the degree of degradation used in calculating the degree of component degradation.