Information acquisition monitoring system and method for semiconductor accessory equipment

Abstract

The invention relates to the technical field of semiconductor manufacturing, and discloses an information acquisition monitoring system and method for semiconductor accessory equipment. An information acquisition monitoring system of semiconductor accessory equipment comprises a data acquisition end which comprises built-in sensors and input ports of a plurality of target equipment, the built-in sensors are used for acquiring state parameters of the target equipment, and the input ports are used for acquiring attribute information data of the target equipment; the edge computing module comprises a plurality of edge computing nodes, and the edge computing nodes are deployed in a preset equipment cluster area; and the edge computing node is used for performing data fusion on the state parameters of the equipment to obtain health feature data of the equipment. According to the invention, by constructing a distributed monitoring architecture with edge computing and cloud collaboration, the problem of insufficient real-time performance in traditional semiconductor equipment management is effectively solved.

Description

Technical Field

[0001] This invention relates to the field of semiconductor manufacturing technology, and more specifically to an information acquisition and monitoring system and method for semiconductor auxiliary equipment. Background Technology

[0002] The semiconductor manufacturing industry heavily relies on the stable operation of precision equipment, and the health status of its auxiliary equipment directly impacts production efficiency and product yield. Traditional equipment management methods typically employ periodic inspections and reactive maintenance, which struggle to achieve real-time monitoring and predictive maintenance of equipment status. With the development of Industrial Internet of Things (IIoT) technology, achieving full lifecycle management of equipment through multi-source data fusion has become an important research direction in this field.

[0003] Existing monitoring systems generally suffer from high data processing latency and insufficient edge-side analysis capabilities, leading to delayed decision-making and response in the cloud. Furthermore, the lack of effective health feature extraction and maintenance prediction models makes it difficult to develop differentiated maintenance strategies for different equipment states, resulting in frequent downtime for maintenance, which wastes resources and fails to effectively prevent sudden failures. Summary of the Invention

[0004] The purpose of this invention is to provide an information acquisition and monitoring system and method for semiconductor auxiliary devices, thereby solving the above-mentioned technical problems:

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] An information acquisition and monitoring system for semiconductor accessories, comprising:

[0007] The data acquisition terminal includes built-in sensors and input ports of multiple target devices. The built-in sensors are used to collect the status parameters of the target devices, and the input ports are used to collect the attribute information data of the target devices.

[0008] The edge computing module includes several edge computing nodes, which are deployed in a preset device cluster area; the edge computing nodes are used to perform data fusion on the status parameters of the devices to obtain the health feature data of the devices.

[0009] The cloud layer includes a database and a maintenance management module. The database stores the status parameter data, attribute information data, and health characteristic data from all edge computing nodes. The maintenance management module includes a predictive analysis unit and a strategy execution unit. The predictive analysis unit predicts the required maintenance duration of the device by analyzing the device's attribute information data and health characteristic data. The strategy execution unit generates and executes a maintenance strategy based on the device's maintenance duration prediction and health characteristic data.

[0010] The communication module is used to establish communication connections between the cloud layer and each edge computing node.

[0011] As a further technical solution, the target equipment includes: a vacuum pump, a cooling water machine, a tail gas treatment device, a gas supply cabinet, and a temperature control device; the status parameters include: temperature, pressure, flow rate, vibration frequency, current, voltage, and power; the attribute information data includes: equipment age, historical maintenance records, maintenance operation complexity, and information on the maintenance operators corresponding to the equipment.

[0012] As a further technical solution, the process by which the edge computing node fuses the device's state parameters to obtain the device's health characteristic data includes:

[0013] The original state parameters are subjected to noise filtering, outlier correction and standardization to generate standardized time series data;

[0014] A sliding window mechanism is used to extract time-domain and frequency-domain features from standardized time-series data;

[0015] (1);

[0016] (2);

[0017] (3);

[0018] (4);

[0019] The health coefficient of the equipment over time is calculated using formulas (1) to (4). ;in, The number of sensors built into the target device; This represents the current measurement value of the i-th built-in sensor; The standard value of the parameter measured by the i-th built-in sensor; , These are the upper and lower limits of the parameter measured by the i-th built-in sensor, respectively; For sensing health status; For aging health; To maintain health; Equipment aging rate; This represents the interval between the last maintenance completion time and the current time. Standard maintenance cycle; , , Preset weights.

[0020] As a further technical solution, the trigger condition for the predictive analysis unit to predict the required maintenance time of the equipment is: determining that the equipment health does not meet the preset standard, and the determination process includes:

[0021] The change in the health coefficient of the equipment over time Health preset threshold function for corresponding devices Compare:

[0022] like If so, the current device health status is determined to meet the preset standard;

[0023] like If the current health status of the device does not meet the preset standard, the device maintenance will be triggered.

[0024] As a further technical solution, the process of predicting the required maintenance time of equipment through analysis of equipment attribute information data and health characteristic data includes:

[0025] (5);

[0026] (6);

[0027] (7);

[0028] in, Baseline forecast, This is the predicted value for the nonlinear part; , The preset regression coefficients; This reduces the operational complexity of equipment maintenance; The preset coupling coefficient; Equipment age reference factor; As an environmental correction factor, This represents the total number of environmental correction factors.

[0029] The predicted maintenance time required for equipment whose health does not meet the preset standard is calculated using formulas (5) to (7). .

[0030] As a further technical solution, the maintenance strategy generated by the strategy execution unit based on the equipment's maintenance duration prediction results and health characteristic data includes:

[0031] Preset health threshold range and maintenance time threshold

[0032] when At that time, no maintenance was performed on the target equipment;

[0033] when hour:

[0034] like or and If so, the equipment maintenance will be postponed until the next planned shutdown.

[0035] like and If so, equipment maintenance will be scheduled during off-peak hours;

[0036] like If the problem occurs, immediately stop the machine and perform maintenance on the equipment.

[0037] As a further technical solution, the communication module includes a link quality monitoring submodule, which is used to collect network performance parameters in real time and switch to a backup link and trigger a data retransmission mechanism when the performance of the main link deteriorates.

[0038] The communication between the cloud layer and the edge computing nodes adopts an end-to-end encryption and authentication mechanism. Each data packet is attached with a digital signature and a timestamp. The edge computing node verifies the signature and then performs data decryption.

[0039] A method for information acquisition and monitoring of semiconductor accessory devices, the method comprising the following steps:

[0040] S1. Collect the status parameters of the corresponding target device through the built-in sensor, and collect the attribute information data of the device through the input port;

[0041] S2. By using edge computing nodes deployed in the device cluster area to perform data fusion processing on the collected status parameters, health feature data of the devices is generated.

[0042] S3. Upload the status parameter data, attribute information data, and health characteristic data to the cloud database for storage. Analyze the attribute information data and health characteristic data of the device through the cloud-based predictive analysis unit to predict the maintenance time required for the device.

[0043] S4. Based on the predicted maintenance duration and health characteristic data, the corresponding equipment maintenance strategy is generated and executed through the strategy execution unit.

[0044] The beneficial effects of this invention are:

[0045] (1) This invention effectively solves the problem of insufficient real-time performance in traditional semiconductor equipment management by constructing a distributed monitoring architecture that combines edge computing and cloud collaboration. Edge computing nodes are deployed near the equipment cluster area, enabling real-time data fusion and feature extraction of raw status parameters, which significantly reduces data transmission latency. The cloud maintenance management module performs intelligent analysis based on multi-dimensional health feature data, realizing a shift from passive maintenance to predictive maintenance, which greatly improves the accuracy and timeliness of equipment management.

[0046] (2) The hierarchical processing mechanism of the present invention fully optimizes the system resource utilization efficiency. Data preprocessing and feature extraction are completed at the edge, reducing the computing load on the cloud. The cloud integrates data to establish a predictive model, and the generated maintenance strategy considers both the health status of individual equipment and the overall scheduling needs of the production line. This collaborative working mode ensures the system response speed while providing a reliable guarantee for the stable operation of the semiconductor manufacturing process. Attached Figure Description

[0047] The invention will now be further described with reference to the accompanying drawings.

[0048] Figure 1 This is a framework diagram of the information acquisition and monitoring system for semiconductor auxiliary devices in this invention;

[0049] Figure 2 This is a flowchart of the information acquisition and monitoring method for semiconductor auxiliary devices in this invention. Detailed Implementation

[0050] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] Please see Figure 1 As shown, an information acquisition and monitoring system for semiconductor accessories includes:

[0052] The data acquisition terminal includes built-in sensors and input ports for multiple target devices. The built-in sensors are used to collect the status parameters of the target devices, specifically employing multimodal sensing technology and integrating multi-dimensional sensor arrays for vibration, temperature, current, pressure, etc. The input ports are used to collect attribute information data of the target devices and support multiple industrial protocols such as Modbus and OPCUA, ensuring compatibility with both new and old devices.

[0053] The edge computing module includes several edge computing nodes, which are deployed in a preset device cluster area; the edge computing nodes are used to perform data fusion on the status parameters of the devices to obtain the health feature data of the devices.

[0054] The cloud layer includes a database and a maintenance management module. The database stores the status parameter data, attribute information data, and health characteristic data from all edge computing nodes. The maintenance management module includes a predictive analysis unit and a strategy execution unit. The predictive analysis unit predicts the required maintenance duration of the device by analyzing the device's attribute information data and health characteristic data. The strategy execution unit generates and executes a maintenance strategy based on the device's maintenance duration prediction and health characteristic data.

[0055] The communication module is used to establish communication connections between the cloud layer and each edge computing node.

[0056] Through the above technical solution, this embodiment provides an information acquisition and monitoring system for semiconductor auxiliary equipment. By constructing a distributed monitoring architecture that combines edge computing and cloud collaboration, the system effectively solves the problem of insufficient real-time performance in traditional semiconductor equipment management. Edge computing nodes are deployed near the equipment cluster area, enabling real-time data fusion and feature extraction of raw status parameters, significantly reducing data transmission latency. The cloud-based maintenance management module performs intelligent analysis based on multi-dimensional health feature data, realizing a shift from passive maintenance to predictive maintenance, greatly improving the accuracy and timeliness of equipment management.

[0057] The target equipment includes: a vacuum pump, a cooling water machine, an exhaust gas treatment device (e.g., a gaseous molecular pollutant filtration unit), a gas supply cabinet, and a temperature control device; the status parameters include: temperature, pressure, flow rate, vibration frequency, current, voltage, and power; the attribute information data includes: equipment age, historical maintenance records, maintenance operation complexity, and information on the maintenance personnel corresponding to the equipment.

[0058] Through the above technical solution, this embodiment establishes a monitoring data system for semiconductor auxiliary equipment. By comprehensively analyzing multi-dimensional parameters, potential fault modes of the equipment can be accurately identified.

[0059] The process by which the edge computing node fuses data on the device's state parameters to obtain the device's health characteristic data includes:

[0060] The original state parameters are subjected to noise filtering, outlier correction and standardization to generate standardized time series data;

[0061] A sliding window mechanism is used to extract time-domain and frequency-domain features from standardized time-series data;

[0062] (1);

[0063] (2);

[0064] (3);

[0065] (4);

[0066] The health coefficient of the equipment over time is calculated using formulas (1) to (4). ;in, The number of sensors built into the target device; This represents the current measurement value of the i-th built-in sensor; The standard value of the parameter measured by the i-th built-in sensor; , These are the upper and lower limits of the parameter measured by the i-th built-in sensor, respectively; For sensing health status; For aging health; To maintain health; The equipment aging rate is calculated by analyzing historical operating data, maintenance records, and performance degradation trends, combined with the equipment's usage time and load intensity. This represents the interval between the last maintenance completion time and the current time. Standard maintenance cycle; , , Preset weights.

[0067] Through the above technical solution, this embodiment provides a process for fusing data on the state parameters of a device to obtain the device's health characteristic data. Specifically, firstly, noise filtering, outlier correction, and standardization are performed on the original state parameters to generate standardized time-series data. Then, a sliding window mechanism is used to extract time-domain and frequency-domain features from the standardized time-series data. Finally, the change in the device's health coefficient over time is calculated using formulas (1) to (4). .

[0068] The predictive analysis unit predicts the required maintenance time for the equipment based on the following trigger condition: the equipment's health status is determined to be below a preset standard. The determination process includes:

[0069] The change in the health coefficient of the equipment over time Health preset threshold function for corresponding devices Compare:

[0070] like If so, the current device health status is determined to meet the preset standard;

[0071] like If the current health status of the device does not meet the preset standard, the device maintenance will be triggered.

[0072] Through the above technical solution, this embodiment provides the triggering conditions for the predictive analysis unit to predict the maintenance time required by the equipment. Specifically, when If the current equipment health status meets the preset standard, equipment maintenance will not be triggered. If the current health status of the device does not meet the preset standard, the device maintenance will be triggered.

[0073] The process of predicting the required maintenance time for equipment by analyzing its attribute information data and health characteristic data includes:

[0074] (5);

[0075] (6);

[0076] (7);

[0077] in, Baseline forecast, This is the predicted value for the nonlinear part; , The preset regression coefficients; This reduces the operational complexity of equipment maintenance; The preset coupling coefficient; Equipment age reference factor; As an environmental correction factor, This represents the total number of environmental correction factors, such as the coefficient corresponding to the spare parts readiness factor. ,in, The physical distance between the current location of the spare part and the target equipment. This represents the current inventory quantity of spare parts. This is the safety stock threshold for the spare part. This is the distance attenuation coefficient (related to spare parts transportation efficiency). This is the inventory sensitivity coefficient (reflecting supply chain responsiveness). The corresponding coefficient for the personnel skills factor is: ,in, This represents the operator's current skill level (e.g., level 1-5). This represents the minimum skill level required for this maintenance task. This represents the steepness coefficient of the skill threshold.

[0078] The predicted maintenance time required for equipment whose health does not meet the preset standard is calculated using formulas (5) to (7). .

[0079] Through the above technical solution, this embodiment provides a process for predicting the maintenance time required for a device by analyzing the device's attribute information data and health characteristic data.

[0080] Based on the equipment's maintenance duration prediction results and health characteristic data, the strategy execution unit generates the following maintenance strategies:

[0081] Preset health threshold range and maintenance time threshold

[0082] when At that time, no maintenance was performed on the target equipment;

[0083] when hour:

[0084] like or and If so, the equipment maintenance will be postponed until the next planned shutdown.

[0085] like and If so, equipment maintenance will be scheduled during off-peak hours;

[0086] like If the problem occurs, immediately stop the machine and perform maintenance on the equipment.

[0087] Through the above technical solution, this embodiment provides a maintenance strategy generated based on the equipment's maintenance duration prediction results and health characteristic data. Specifically, when When the equipment is healthy, maintenance is not required; when If this occurs, it indicates that the equipment is not healthy. Under this premise, if or and Then the equipment maintenance will be postponed until the next planned downtime; if and Then, equipment maintenance will be scheduled during off-peak hours; if If a problem occurs, the system is immediately shut down, and maintenance is performed on the equipment. By integrating data to build a predictive model, the generated maintenance strategy considers both the health status of individual devices and the overall scheduling needs of the production line. This collaborative working mode ensures system response speed while providing a reliable guarantee for the stable operation of the semiconductor manufacturing process.

[0088] The communication module includes a link quality monitoring submodule, which is used to collect network performance parameters in real time and switch to a backup link and trigger a data retransmission mechanism when the performance of the main link deteriorates.

[0089] The communication between the cloud layer and edge computing nodes employs an end-to-end encryption and authentication mechanism. Each data packet is appended with a digital signature and a timestamp, and the edge computing node verifies the signature before decrypting the data. This technical solution achieves high reliability and security for data transmission by deploying intelligent link monitoring and secure communication mechanisms. The link quality monitoring submodule senses the network status in real time and automatically performs link switching, ensuring uninterrupted transmission of critical monitoring data. The dual protection system, combining end-to-end encryption with digital signature verification, ensures both the confidentiality and integrity of data during transmission and effectively prevents security threats such as replay attacks. This collaborative design enables the system to simultaneously meet real-time requirements and information security standards in complex industrial environments, providing a robust communication guarantee for the stable monitoring of semiconductor manufacturing equipment.

[0090] Please see Figure 2 As shown, a method for information acquisition and monitoring of semiconductor peripheral devices includes the following steps:

[0091] S1. Collects status parameters of the target equipment through built-in sensors, such as: mechanical vibration (obtained via a 3-axis accelerometer); motor current (three-phase harmonic analysis); bearing temperature (infrared + contact dual measurement); and oil pressure (differential pressure sensor). It also collects equipment attribute information data through input ports, such as: equipment file data: reading the equipment's electronic history via OPC UA; and maintenance history data: obtaining work order records (including fault codes and replacement parts lists) from the CMMS system.

[0092] S2. By using edge computing nodes deployed in the device cluster area to perform data fusion processing on the collected status parameters, health feature data of the devices is generated.

[0093] S3. Upload the status parameter data, attribute information data, and health characteristic data to the cloud database for storage. Analyze the attribute information data and health characteristic data of the device through the cloud-based predictive analysis unit to predict the maintenance time required for the device.

[0094] S4. Based on the predicted maintenance duration and health characteristic data, the corresponding equipment maintenance strategy is generated and executed through the strategy execution unit.

[0095] By constructing an intelligent monitoring system that integrates edge computing and cloud collaboration, closed-loop management from data acquisition to maintenance decision-making is achieved. This method first completes real-time acquisition and feature extraction of device status parameters at the edge, effectively reducing data transmission latency. Then, multi-dimensional data analysis and predictive modeling are performed in the cloud to generate precise maintenance strategies. The distributed deployment of edge computing nodes significantly improves the real-time performance of data processing and local decision-making efficiency, ensuring timely monitoring of critical equipment status. Furthermore, the cloud-based intelligent analysis module integrates data from the entire equipment lifecycle, enabling a shift from passive to predictive maintenance, improving maintenance efficiency and reducing the risk of unplanned downtime.

[0096] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. An information acquisition and monitoring system for semiconductor auxiliary devices, characterized in that, include: The data acquisition terminal includes built-in sensors and input ports of multiple target devices. The built-in sensors are used to collect the status parameters of the target devices, and the input ports are used to collect the attribute information data of the target devices. The edge computing module includes several edge computing nodes, which are deployed in a preset device cluster area; the edge computing nodes are used to perform data fusion on the status parameters of the devices to obtain the health feature data of the devices. The cloud layer includes databases and maintenance management modules; The database is used to store the status parameter data, attribute information data, and health feature data from all edge computing nodes; The maintenance management module includes a predictive analysis unit and a strategy execution unit; the predictive analysis unit is used to predict the required maintenance time of the equipment by analyzing the equipment's attribute information data and health characteristic data. The strategy execution unit is used to generate and execute maintenance strategies based on the equipment's maintenance duration prediction results and health characteristic data. The communication module is used to establish communication connections between the cloud layer and each edge computing node.

2. The information acquisition and monitoring system for semiconductor auxiliary devices according to claim 1, characterized in that, The target equipment includes: vacuum pump, cooling water machine, exhaust gas treatment device, gas supply cabinet and temperature control equipment; the status parameters include: temperature, pressure, flow rate, vibration frequency, current, voltage and power; the attribute information data includes: equipment age, historical maintenance records, maintenance operation complexity and the corresponding maintenance operator information.

3. The information acquisition and monitoring system for semiconductor auxiliary equipment according to claim 2, characterized in that, The process by which the edge computing node fuses data on the device's state parameters to obtain the device's health characteristic data includes: The original state parameters are subjected to noise filtering, outlier correction and standardization to generate standardized time series data; A sliding window mechanism is used to extract time-domain and frequency-domain features from standardized time-series data; （1）； （2）； （3）； （4）； The health coefficient of the equipment over time is calculated using formulas (1) to (4). ;in, The number of sensors built into the target device; This represents the current measurement value of the i-th built-in sensor; The standard value of the parameter measured by the i-th built-in sensor; , These are the upper and lower limits of the parameter measured by the i-th built-in sensor, respectively; For sensing health status; For aging health; To maintain health; Equipment aging rate; This represents the interval between the last maintenance completion time and the current time. Standard maintenance cycle; , , Preset weights.

4. The information acquisition and monitoring system for semiconductor auxiliary devices according to claim 3, characterized in that, The predictive analysis unit predicts the required maintenance time for the equipment based on the following trigger condition: the equipment's health status is determined to be below a preset standard. The determination process includes: The change in the health coefficient of the equipment over time Health preset threshold function for corresponding devices Compare: like If so, the current device health status is determined to meet the preset standard; like If the current health status of the device does not meet the preset standard, the device maintenance will be triggered.

5. The information acquisition and monitoring system for semiconductor auxiliary devices according to claim 4, characterized in that, The process of predicting the required maintenance time for equipment by analyzing its attribute information data and health characteristic data includes: （5）； （6）； （7）； in, Baseline forecast value, This is the predicted value for the nonlinear part; , The preset regression coefficients; This reduces the operational complexity of equipment maintenance; The preset coupling coefficient; Equipment age reference factor; As an environmental correction factor, This represents the total number of environmental correction factors. The predicted maintenance time required for equipment whose health does not meet the preset standard is calculated using formulas (5) to (7). .

6. The information acquisition and monitoring system for semiconductor auxiliary devices according to claim 5, characterized in that, Based on the equipment's maintenance duration prediction results and health characteristic data, the strategy execution unit generates the following maintenance strategies: Preset health threshold range and maintenance time threshold when At that time, no maintenance was performed on the target equipment; when hour: like or and If so, the equipment maintenance will be postponed until the next planned shutdown. like and If so, equipment maintenance will be scheduled during off-peak hours; like If the problem occurs, immediately stop the machine and perform maintenance on the equipment.

7. The information acquisition and monitoring system for semiconductor auxiliary devices according to claim 6, characterized in that, The communication module includes a link quality monitoring submodule, which is used to collect network performance parameters in real time and switch to a backup link and trigger a data retransmission mechanism when the performance of the main link deteriorates. The communication between the cloud layer and the edge computing nodes adopts an end-to-end encryption and authentication mechanism. Each data packet is attached with a digital signature and a timestamp. The edge computing node verifies the signature and then performs data decryption.

8. A method for information acquisition and monitoring of semiconductor auxiliary devices, characterized in that, The method is applied to the information acquisition and monitoring system for semiconductor auxiliary devices as described in claim 7, and the method includes the following steps: S1. Collect the status parameters of the corresponding target device through the built-in sensor, and collect the attribute information data of the device through the input port; S2. By using edge computing nodes deployed in the device cluster area to perform data fusion processing on the collected status parameters, health feature data of the devices is generated. S3. Upload the status parameter data, attribute information data, and health characteristic data to the cloud database for storage. Analyze the attribute information data and health characteristic data of the device through the cloud-based predictive analysis unit to predict the maintenance time required for the device. S4. Based on the predicted maintenance duration and health characteristic data, the corresponding equipment maintenance strategy is generated and executed through the strategy execution unit.

Images