A machine monitoring parameter extraction method and device based on large language model

Abstract

The present invention provides a method and device for selecting monitoring parameters of machine equipment, the method comprising: first, extracting the full text information, pictures and corresponding picture names in the technical manual of the machine equipment; then, using LLM to analyze the picture names to obtain partial picture names describing the overall architecture and components of the machine equipment; then, using a multimodal model to obtain machine architecture information and component composition information from the pictures corresponding to these partial picture names; then, inputting these information and the full text information of the manual into LLM to obtain a summary corresponding to the partial picture, and inputting it into LLM to obtain component classification information; then, inputting the summary, component classification information, association information between parameters and components, and multiple parameter information into LLM to obtain reasoning information of each parameter therein; finally, inputting the reasoning information of each parameter and information describing the basis for judging the importance of the parameter into LLM to obtain the final monitoring parameters.

Description

Technical Field

[0001] The present invention relates to the field of equipment monitoring, and in particular to a method and device for extracting machine monitoring parameters based on a large language model. Background Art

[0002] In the semiconductor manufacturing process, equipment monitoring and control are the key to ensuring product quality and production efficiency. Parameter monitoring of machine equipment can be applied to multiple production links such as fault detection classification, yield prediction, process optimization, equipment health management, process control, and environmental monitoring. Among them, fault detection classification and yield prediction play a vital role in ensuring the stability of the production process and the quality of the product.

[0003] The number and types of parameters of semiconductor equipment vary depending on the equipment model. Complex equipment parameters can reach tens of thousands. In actual production, due to the limitations of data storage efficiency and development efficiency, only some key parameters can be selectively collected and monitored. These key parameters can provide data support for fault detection classification and yield prediction, which are called monitoring parameters. Monitoring parameters include two main data types: data variables (DV) and state variables (SV).

[0004] Among them, DV is a data item collected when a specific event occurs. For example, in semiconductor production, when a recipe changes, the GEM interface may trigger an event called PPChanged and provide DVs related to it, such as "changedrecipe". These data are only valid in the context of a specific event and may produce invalid or inaccurate information outside of that context. Due to the context dependency of DV, accurate event recording and data collection are critical in failure analysis and process optimization. For real-time monitoring and failure prediction, how to effectively capture and utilize these instantaneous data has become a major challenge.

[0005] SV is a persistent data item containing device information, which is guaranteed to be valid at any time. For example, the reading of the temperature sensor ModuleTemperature can be requested through the GEM interface, and it always reflects the current state of the device. SV provides stable and repeatable information, and its accuracy directly affects the evaluation of device performance and the effectiveness of fault prediction, making it of great significance in equipment status monitoring, process control and fault detection.

[0006] In traditional semiconductor manufacturing processes, the selection of equipment monitoring parameters relies on experienced process engineers and implementation engineers. Although this approach can work in the short term, it has obvious limitations when faced with the following challenges:

[0007] 1. High dependence and low efficiency: Process engineers and implementation engineers usually rely on personal experience and intuition when selecting monitoring parameters. When the amount of parameter data is huge, this dependence leads to a time-consuming selection process, which seriously affects work efficiency.

[0008] 2. Lack of migration flexibility: When new equipment is put into use, engineers often lack sufficient experience in the characteristics and parameter selection of the equipment. This leads to the cold start problem, that is, in the absence of historical data and experience support, it is impossible to quickly make effective parameter selections.

[0009] 3. Lack of accuracy: The current keyword or specified rule screening methods lack flexibility and are difficult to adapt to complex and changing production environments. These methods usually ignore contextual information and fail to fully utilize the relationship between device events and parameters, resulting in inaccurate screening results.

[0010] 4. Information islands and knowledge transfer barriers: The inheritance of experience and the transfer of knowledge often rely on individual engineers, resulting in information islands and easily leading to inaccurate screening results. This also means that new engineers find it difficult to effectively acquire the necessary knowledge and experience in the absence of guidance, further exacerbating the cold start problem.

[0011] 5. Unadaptable to dynamically changing environments: The semiconductor manufacturing process is highly complex and changes rapidly. Traditional experience-driven and rule-based screening methods cannot be quickly adjusted to adapt to emerging process changes and equipment characteristics, thus limiting production flexibility and adaptability.

[0012] In order to solve the above problems, a more flexible, accurate and efficient method for screening machine equipment monitoring parameters is urgently needed to adapt to the emergence of new processes and equipment. Summary of the invention

[0013] In order to solve the above problems, the present invention proposes a machine monitoring parameter extraction method and device based on a large language model.

[0014] According to a first aspect, the present invention provides a method for extracting machine monitoring parameters based on a large language model, comprising the following steps:

[0015] Parse the technical manuals of machine equipment to extract the manual text, pictures and corresponding figure names.

[0016] LLM was used to analyze all the extracted graphic names to obtain some graphic names that describe the overall architecture and components of the machine equipment.

[0017] Part of the images corresponding to the partial image names among all the extracted images are input into the multimodal model to obtain machine architecture information and component composition information.

[0018] The machine architecture information and component composition information and the manual text are input into the LLM to obtain a summary of the text information corresponding to the partial picture, and the summary is input into the LLM to obtain component classification information.

[0019] The summary, component classification information, association information of pre-configured parameters and components, and multiple parameter information corresponding to multiple parameters of the machine equipment are input into LLM to obtain reasoning information of each parameter, wherein the reasoning information includes parameter ID, component to which the parameter belongs, and reasoning basis.

[0020] The reasoning information of the various parameters and the information describing the basis for judging the importance of the parameters are input into the LLM to obtain the important parameters among the multiple parameters as the final monitoring parameters.

[0021] In some embodiments, the document parsing specifically includes:

[0022] Through OCR technology, the text information and all pictures in the technical manual are extracted.

[0023] Based on the text information near each image, a large language model is used to determine the image name corresponding to the image.

[0024] In some embodiments, the component classification information includes component categories, and the component categories include MainFrame, Factory Interface, and Load Lock A.

[0025] In some embodiments, the parameter information includes a data type, a field description, and an identifier.

[0026] In some embodiments, the machine equipment is a semiconductor machine equipment.

[0027] According to a second aspect, the present invention provides a machine monitoring parameter extraction device based on a large language model, comprising:

[0028] The document parsing module is used to parse the technical manual of the machine equipment and extract the manual text, pictures and corresponding picture names.

[0029] The image name screening module is used to analyze all the image names extracted by the document parsing module through LLM to obtain partial image names that describe the overall architecture and components of the machine equipment.

[0030] The component identification module is used to input some of the images corresponding to the partial image names among all the images extracted by the document parsing module into the multimodal model to obtain the machine architecture information and component composition information.

[0031] The component classification module is used to input the machine architecture information and component composition information and the manual text extracted by the document parsing module into the LLM to obtain a summary of the text information corresponding to the partial picture, and to input the summary into the LLM to obtain component classification information.

[0032] The primary screening module is used to input the summary, component classification information, association information of pre-configured parameters and components, and multiple parameter information corresponding to multiple parameters of the machine equipment into the LLM to obtain reasoning information of each parameter, which includes parameter ID, component to which the parameter belongs, and reasoning basis.

[0033] The secondary screening module is used to input the reasoning information of each parameter and the information describing the basis for judging the importance of the parameters into the LLM, so as to obtain the important parameters among the multiple parameters as the final monitoring parameters.

[0034] In some embodiments, the document parsing module is specifically configured to: extract text information and all pictures in the technical manual through OCR technology; and determine the picture name corresponding to the picture using a large language model based on the text information near each picture.

[0035] In some embodiments, the device is applied to semiconductor equipment.

[0036] The monitoring parameter selection method provided by the present invention performs text analysis through LLM. Even in the face of a huge amount of parameter data, LLM can fully associate and utilize the context information and image information in the technical manual, and combine manual experience to efficiently and accurately complete the monitoring parameter selection. In addition, the method can fully adapt to the characteristics of the device itself and has high migration flexibility. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0038] Figure 1 A flow chart of the machine monitoring parameter extraction method provided by the present invention;

[0039] Figure 2 This is a diagram of the machine monitoring parameter extraction method provided by the present invention;

[0040] Figure 3 This is a schematic diagram of the machine monitoring parameter extraction device provided by the present invention. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0042] In order to overcome the limitations of the traditional machine equipment monitoring parameter screening method, the present invention introduces a more intelligent large language model into the machine equipment monitoring parameter screening method, and performs automatic monitoring parameter screening based on the large language model.

[0043] Large Language Models (LLM) are artificial intelligence models that learn the universal characteristics of language by pre-training on large amounts of text data and are able to demonstrate strong understanding and generation capabilities in a variety of natural language processing tasks.

[0044] The powerful understanding and generation capabilities of the large language model can be used to parse the equipment technical manual, identify the association mechanism between equipment, and efficiently infer the parameters related to the equipment. Even in the face of a huge amount of equipment technical manual text and parameter data, LLM can fully associate contextual information for reasoning to achieve more efficient, flexible and accurate monitoring parameter selection. However, when LLM is directly applied to the selection of monitoring parameters, due to the lack of relevant experience, LLM cannot effectively distinguish which parameters are important and give inferences. In view of this, the present invention provides a monitoring parameter screening method that integrates LLM, multimodal LLM, manual experience, equipment manual, and parameter information, effectively refines and summarizes manual experience, equipment information, and parameter information, and combines empirical knowledge for secondary screening to generate a list of monitoring parameters suitable for a specific process, thereby improving the accuracy and efficiency of fault prediction.

[0045] The following is a detailed description of the monitoring parameter screening method provided by the present invention in conjunction with the accompanying drawings. The flowchart of the method is as follows: Figure 1 As shown, the following steps are included:

[0046] S1. Document parsing: Parse the technical manual of the machine equipment to extract the manual text, pictures and corresponding figure names.

[0047] In some embodiments, the machine equipment is a semiconductor machine equipment.

[0048] The technical manual is provided by the equipment manufacturer and usually contains equipment overview, installation guide, operating instructions, maintenance, troubleshooting, parts list and drawings, etc. It is intended to guide users to install, operate, maintain and repair equipment safely and effectively.

[0049] In some embodiments, step S1 specifically includes:

[0050] S11. Extract the manual text and all the images in the technical manual by using optical character recognition (OCR) technology.

[0051] OCR technology can convert printed or handwritten text information into computer-processable information.

[0052] Specifically, OCR technology can be used to analyze the layout of the document, extract the text information in the technical manual page by page, and combine the text with the location information to form a readable text segment. While extracting the image, the image is linked to the nearby text based on the location information, which facilitates the subsequent determination of the image name corresponding to each image.

[0053] S12. Determine the name of the picture corresponding to the picture using a large language model based on the text information near each picture.

[0054] Specifically, the text information near the image can be sent to LLM, and the name of the image can be found using the prompt method. The image names in general technical manuals have fixed formats, and LLM can recognize the image names based on these formats.

[0055] For example, a prompt for finding the figure name could be: "Please output all the figure names in the technical manual, with the figure number before the figure name, e.g. Figure 1 .1 Equipment structure diagram. "

[0056] In this way, the document parsing of the technical manual of the machine equipment is completed, and the text information, pictures and corresponding picture names are successfully extracted.

[0057] S2. Graphic name screening: Use LLM to analyze all extracted graphic names and obtain some graphic names that describe the overall architecture and components of the machine equipment.

[0058] Equipment manuals often use some specific descriptions such as top view, overhead view, side view, etc. as the names of the equipment structure.

[0059] Specifically, you can use the prompt "Please find the picture names related to 'top view' or 'top view' or 'side view' in the above picture names" to find the picture names that describe the overall structure and components of the machine equipment.

[0060] You can use the prompt "Please find the picture name related to 'overall architecture' in the above picture names" to find the picture name that describes the overall architecture of the machine equipment.

[0061] You can also use the prompt "Please find the picture name related to 'component' or 'interface' or 'part' in the above picture names" to find the picture name describing the machine equipment component.

[0062] S3. Component identification: inputting some of the extracted images corresponding to the partial image names into the multimodal model to obtain machine architecture information and component composition information.

[0063] In this step, images describing the overall architecture and components of the equipment are fed into a multimodal model that can process and understand data in different modes or forms, including images. Using a multimodal model, it is possible to extract component composition information contained in the image, supplementing the lack of text information in the equipment manual.

[0064] The machine architecture information includes the overall structure of the machine architecture, the connection between the various components, etc. The component composition information includes the component structure, component shape, size, etc.

[0065] S4. Component classification: input the machine architecture information, component composition information and the manual text information into LLM to obtain a summary of the text information corresponding to the partial picture, and input the summary into LLM to obtain component classification information.

[0066] This step specifically includes:

[0067] S41. The machine architecture information and component composition information extracted from the machine architecture diagram and the equipment component diagram are combined with the full-text information or full-text segmented information extracted from the technical manual, and a text summary corresponding to the machine architecture diagram and the equipment component diagram is generated through LLM analysis and integration.

[0068] Among them, using LLM to generate a text summary corresponding to the device component diagram can be achieved through the following prompt:

[0069] “Please combine the technical manual and component composition information to generate a text summary for each component. The text summary needs to include: component number, component name, component structure, connection method with other components, function in the device, operation method, and other information you think is important.”

[0070] S42. Input the text summary obtained in S41 into LLM to obtain component classification information.

[0071] The component classification information includes component description and component category.

[0072] The component description is a specific description of the architecture or component, including the component structure, connection method with other components, role in the device, operation method, etc., wherein the information comes from the text summary of each component.

[0073] The component categories may be MainFrame, Factory Interface, Load Lock A, Load LockB, Chamber A, Chamber B, Chamber C, Chamber D, etc. In addition to the text summary, human experience on component classification may also be supplemented as input.

[0074] Specifically, after entering the text summary into LLM, you can get the component classification information through the following prompt: "Please classify all components based on the above text summary information. Please input: component name - component category - component description"

[0075] The above describes how to obtain a summary of the text information of the partial image, and how to input the summary into LLM to obtain component classification information. The above steps can fully utilize the information in the technical manual and give full play to the excellent text reasoning ability of LLM to achieve more efficient and accurate information extraction and component classification.

[0076] S5. Initial screening: Input the summary, component classification information, parameter-component association information, and multiple parameter information corresponding to multiple parameters of the machine equipment into LLM to obtain reasoning information of each parameter, wherein the reasoning information includes parameter ID, component to which the parameter belongs, and reasoning basis.

[0077] The multiple parameters include all possible monitoring parameters involved in the corresponding field of the machine equipment, which come from the monitoring parameter database pre-built in the corresponding field. Parameter information may include data type, position field description, field description, general window definition and identifier (i.e. parameter ID), etc. The association information between parameters and components comes from manual experience input, including which parameter corresponds to which component of the equipment, for example, temperature corresponds to a heating furnace.

[0078] Specifically, after the above information is input into LLM, the inference information of each parameter can be obtained through the following prompt: "Please infer all the monitoring parameters involved in this set of equipment based on the above information. The monitoring parameters involved in this set of equipment need to be associated with one or several components in the equipment. The output format is required to be: inference basis | parameter ID | component to which the parameter belongs."

[0079] This step can filter out monitoring parameters associated with each component in the machine equipment, that is, monitoring parameters that may be applied to the equipment.

[0080] S6. Secondary screening: input the reasoning information of each parameter and the information describing the basis for judging the importance of the parameter into the LLM to obtain the important parameters among the multiple parameters as the final monitoring parameters.

[0081] The information describing the importance of parameters comes from manual experience input, including which parameters under which components are important, which data types of parameters are important, which parameters with which descriptive characteristics are unimportant, etc. For example, the parameters under the process chamber component are important. Similar manual experience information can be directly input by experienced engineers, or extracted by LLM from technical documents written by engineers.

[0082] After inputting the information obtained from the initial screening and the information describing the component categories and the basis for judging the importance of the parameters into the LLM, the final monitoring parameters can be obtained through the following prompt: "Please divide the parameters obtained from the last reasoning into two categories: important and others based on the above information, and output the parameter ID of the important parameters."

[0083] In addition, the number of monitoring parameters finally selected can be limited in the prompt as needed.

[0084] The above is the machine equipment monitoring parameter selection method provided by the present invention. The overall structure of the method and the input and output of each step can also be seen in Figure 2 It can be seen that this method has the following beneficial effects:

[0085] 1. High selection efficiency: The monitoring parameter selection method provided by the present invention performs text analysis through LLM, and can quickly complete the monitoring parameter selection even if the amount of parameter data is huge.

[0086] 2. High accuracy: The monitoring parameter selection method provided by the present invention can make full use of the context information and picture information in the technical manual, and combined with manual experience, can make full use of the relationship between equipment events and parameters and accurately judge the importance of parameters, thereby obtaining accurate monitoring parameter screening results.

[0087] 3. High migration flexibility: The monitoring parameter selection method provided by the present invention selects monitoring parameters based on the equipment manual, can be applied to various machines and equipment and fully adapt to the characteristics of the equipment itself, and has high migration flexibility.

[0088] 4. Full automation: The monitoring parameter selection method provided by the present invention uses LLM to achieve fully automated monitoring parameter selection, which basically does not require the human involvement of engineers, can save labor costs and improve factory efficiency.

[0089] On the other hand, the present invention also provides a machine monitoring parameter extraction device 300 based on a large language model, the schematic diagram of which is shown in FIG. Figure 3 As shown, including:

[0090] The document parsing module 301 is used to parse the technical manual of the machine equipment and extract the manual text, pictures and corresponding picture names therein.

[0091] The image name screening module 302 is used to analyze all the image names extracted by the document parsing module through LLM to obtain partial image names that describe the overall architecture and components of the machine equipment.

[0092] The component identification module 303 is used to input some of the pictures corresponding to the partial picture names among all the pictures extracted by the document parsing module into the multimodal model to obtain the machine architecture information and component composition information.

[0093] The component classification module 304 is used to input the machine architecture information and component composition information and the manual text extracted by the document parsing module into the LLM to obtain a summary of the text information corresponding to the partial picture, and to input the summary into the LLM to obtain component classification information.

[0094] The initial screening module 305 is used to input the summary, component classification information and the association information of pre-configured parameters and components, as well as multiple parameter information corresponding to multiple parameters of the machine equipment into the LLM to obtain the reasoning information of each parameter, wherein the reasoning information includes the parameter ID, the component to which the parameter belongs and the reasoning basis.

[0095] The secondary screening module 306 is used to input the reasoning information of each parameter and the information describing the basis for judging the importance of the parameter into the LLM, so as to obtain the important parameters among the multiple parameters as the final monitoring parameters.

[0096] In some embodiments, the document parsing module is specifically configured to: extract text information and all pictures in the technical manual through OCR technology; and determine the picture name corresponding to the picture using a large language model based on the text information near each picture.

[0097] In some embodiments, the device is applied to semiconductor equipment.

[0098] It should be noted that the above-mentioned device can execute the above-mentioned machine monitoring parameter extraction method. For the function of each module, please refer to the above-mentioned introduction to the method and will not be repeated here.

[0099] In the description of the embodiments of the present application, words such as "exemplary", "for example" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary", "for example" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary", "for example" or "for example" is intended to present related concepts in a concrete way.

[0100] In the description of the embodiments of the present application, the term "and / or" is merely a description of the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B may represent: A exists alone, B exists alone, and A and B exist at the same time. In addition, unless otherwise specified, the term "plurality" means two or more.

[0101] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. The terms "include", "comprises", "has" and their variations all mean "including but not limited to", unless otherwise specifically emphasized.

[0102] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made on the basis of the technical solution of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for selecting monitoring parameters of a machine equipment, characterized in that: The following steps are involved: Parse the technical manuals of machine equipment to extract the manual text, pictures and corresponding picture names; Use the large language model (LLM) to analyze all the extracted graphic names and obtain some graphic names that describe the overall architecture and components of the equipment. Inputting some of the extracted pictures corresponding to the part of the picture names into the multimodal model to obtain machine architecture information and component composition information; the machine architecture information includes the overall structure of the machine architecture and the connection mode between the various components; the component composition information includes the component structure, component shape, and size; Input the machine architecture information and component composition information and the manual text into the LLM to obtain a summary of the text information corresponding to the partial picture, and input the summary into the LLM to obtain component classification information; Input the summary, component classification information, association information of pre-configured parameters and components, and multiple pieces of parameter information corresponding to multiple parameters of the machine equipment into the LLM to obtain inference information of each parameter, wherein the inference information includes parameter ID, component to which the parameter belongs, and inference basis; Input the reasoning information of each parameter and the information describing the basis for judging the importance of the parameter into the LLM to obtain the important parameters among the multiple parameters as the final monitoring parameters; The document parsing specifically includes: Extract the manual text and all images in the technical manual through OCR technology; Based on the text information near each image, a large language model is used to determine the corresponding image name; The machine architecture information and component composition information and the manual text are input into the LLM to obtain a summary of the text information corresponding to the partial picture, and the summary is input into the LLM to obtain component classification information, specifically including: The machine architecture information and component composition information extracted from the machine architecture diagram and the equipment component diagram are combined with the full text information or full text segment information extracted from the technical manual, and a text summary corresponding to the machine architecture diagram and the equipment component diagram is generated through LLM analysis and integration; the summary is input into LLM to obtain component classification information; The multiple parameters include all monitoring parameters involved in the corresponding field of the machine equipment, which come from a monitoring parameter database pre-built in the corresponding field; the parameter information includes data type, position field description, field description, general window definition and identifier; The information describing the basis for judging the importance of the parameters is input from human experience.

2. The method according to claim 1, characterized in that The component classification information includes component description and component category, and the component category includes MainFrame, Factory Interface and Load Lock A.

3. The method according to claim 1, characterized in that The parameter information includes data type, field description and identifier.

4. The method according to claim 1, characterized in that The machine equipment is a semiconductor machine equipment.

5. A device for selecting monitoring parameters of a machine equipment, characterized in that: include: The document parsing module is used to parse the technical manuals of the equipment and extract the manual text, pictures and corresponding picture names; The image name screening module is used to analyze all the image names extracted by the document parsing module through the large language model LLM to obtain some image names that describe the overall architecture and components of the machine equipment; The component identification module is used to input the part of the pictures corresponding to the part of the picture names among all the pictures extracted by the document parsing module into the multimodal model to obtain the machine architecture information and component composition information; the machine architecture information includes the overall structure of the machine architecture and the connection mode between the various components; the component composition information includes the component structure, component shape and size; A component classification module, used for inputting the machine architecture information and component composition information and the manual text extracted by the document parsing module into the LLM to obtain a summary of the text information corresponding to the partial picture, and inputting the summary into the LLM to obtain component classification information; A primary screening module, used to input the summary, component classification information, association information of pre-configured parameters and components, and multiple parameter information corresponding to multiple parameters of the machine equipment into the LLM, and obtain reasoning information of each parameter, wherein the reasoning information includes parameter ID, component to which the parameter belongs, and reasoning basis; A secondary screening module is used to input the reasoning information of each parameter and the information describing the basis for judging the importance of the parameter into the LLM to obtain the important parameters among the multiple parameters as the final monitoring parameters; The document parsing module is specifically configured as follows: Extract the manual text and all images in the technical manual through OCR technology; Based on the text information near each image, a large language model is used to determine the corresponding image name; The component classification module is specifically configured to: combine the machine architecture information and component composition information extracted from the machine architecture diagram and the equipment component diagram with the full text information or full text segment information extracted from the technical manual, and generate a text summary corresponding to the machine architecture diagram and the equipment component diagram through LLM analysis and integration; input the summary into the LLM to obtain component classification information; The multiple parameters include all monitoring parameters involved in the corresponding field of the machine equipment, which come from a monitoring parameter database pre-built in the corresponding field; the parameter information includes data type, position field description, field description, general window definition and identifier; The information describing the basis for judging the importance of the parameters is input from human experience.

6. The device according to claim 5, characterized in that The component classification information includes component categories, and the component categories include MainFrame, Factory Interface, and Load Lock A.

7. The device according to claim 5, characterized in that The parameter information includes data type, field description and identifier.

8. The device according to claim 5, characterized in that The device is applied to semiconductor equipment.

Images