CN-122019986-A - Machine data management method, equipment and storage medium

CN122019986ACN 122019986 ACN122019986 ACN 122019986ACN-122019986-A

Abstract

The present application relates to the field of semiconductors, and in particular, to a method and apparatus for managing machine data, and a storage medium. The method comprises the steps of obtaining target process steps executed by a target machine and machine data generated during the execution of the process steps, determining time effect information of the machine data according to process categories and/or process sequence positions of the target process steps, dividing a future time axis into a plurality of continuous time intervals which are not overlapped with each other based on start and stop time of windows in the time effect information, calculating comprehensive access probability of each time interval based on predicted access probability corresponding to different time effect information, distributing corresponding compression strategies for each time interval based on the comprehensive access probability, and compressing and storing the machine data in different time intervals by applying the corresponding compression strategies. Therefore, dynamic and on-demand allocation of storage resources is realized, and the balance between the data storage cost and the response efficiency is achieved.

Inventors

LI DAN
Fu Huichu
GUO YUXIANG
LI JIE
LIU BIN
Guo Dingjingjun

Assignees

埃克斯控股(北京)有限公司

Dates

Publication Date: 20260512
Application Date: 20260414

Claims (10)

1. The machine data management method is characterized by comprising the following steps: s101, acquiring target process steps executed by a target machine and machine data generated during execution of the target process steps; S102, determining ageing information of machine station data according to the procedure category and/or the procedure sequence position of the target process step, wherein the ageing information comprises at least one of a life cycle stage window, an analysis evaluation window and a preset calling window; S103, dividing a future time axis into a plurality of continuous time intervals which are not overlapped with each other based on the starting and ending time of the window in the aging information; s104, calculating comprehensive access probability of each time interval based on predicted access probability corresponding to different time efficiency information, and distributing corresponding compression strategies for each time interval based on the comprehensive access probability, wherein the compression degree of the compression strategies on machine data is inversely related to the comprehensive access probability; s105, applying corresponding compression strategies to compress and store the machine data in different time intervals; Wherein, before the compression storage, the method further comprises updating the compression strategy, which comprises the following steps: predicting the accumulated memory occupation amounts at different time points in a future time axis according to compression strategies of machine data of a plurality of batches in a target machine; when the accumulated memory occupation amount at any time point is larger than a preset memory threshold value, updating the mapping relation between the comprehensive access probability and the compression strategy at the corresponding abnormal time point, wherein the method comprises the steps of reducing the grading number in the mapping relation; and updating the compression strategy of the time interval in which the abnormal time point is based on the updated mapping relation.
2. The method according to claim 1, wherein S104 comprises: The method comprises the steps of inquiring a preset data life cycle stage table to determine a first prediction probability of a time interval, and/or acquiring a second prediction probability of the time interval positioned in the analysis evaluation window, and/or acquiring a third prediction probability of the time interval positioned in the preset calling window; When at least two probability components in the first prediction probability, the second prediction probability and the third prediction probability are obtained, weighting and fusing the at least two probability components to obtain comprehensive access probability; when any one probability component of the first prediction probability, the second prediction probability and the third prediction probability is acquired, the corresponding probability component is used as the comprehensive access probability.
3. The method according to claim 1, wherein S104 comprises: a mapping relation between the comprehensive access probability and the compression strategy is established in advance; The mapping relation comprises a preset grading number, probability threshold intervals corresponding to all grades and compression strategies associated with each probability threshold interval.
4. A method according to claim 3, wherein S104 comprises: When the comprehensive access probability of the time interval is higher than a first probability threshold, a first compression strategy is adopted; when the comprehensive access probability of the time interval is lower than or equal to the first probability threshold and higher than a second probability threshold, adopting a second compression strategy; when the comprehensive access probability of the time interval is lower than or equal to the second probability threshold, a third compression strategy is adopted; The compression degree of the first compression strategy is lower than that of the second compression strategy, and the compression degree of the second compression strategy is lower than that of the third compression strategy.
5. The method according to claim 1, wherein the method further comprises: Acquiring a target machine station group, wherein the target machine station group comprises a target machine station and a plurality of other machine stations which are the same as the target machine station, and the other machine stations comprise preceding machine stations with service life limit higher than that of the target machine station; And determining a peak time period of the target machine data accessed in the future according to the data access log of the historical operation of the prior machine to the current same use year of the target machine, and taking the peak time period as a preset call window.
6. The method of claim 5, wherein the method further comprises: acquiring actually measured access frequencies of a plurality of machines in the target machine group in a preset time period in real time; And when abnormal fluctuation of the group access frequency is monitored, updating the compression strategy of the current time interval based on the group access frequency.
7. The method of claim 5, wherein the compression policy for the current time interval is updated based on the measured access frequency of the target station.
8. The method according to claim 6 or 7, wherein the updated mapping relation comprises: When the comprehensive access probability of the time interval is higher than a first probability threshold, a first compression strategy is adopted; And when the integrated access probability of the time interval is lower than or equal to the first probability threshold, adopting a third compression strategy.
9. A computer device, the device comprising: A memory for storing a computer program; a processor for executing the computer program and for implementing the method of managing machine data according to any one of claims 1 to 8 when the computer program is executed.
10. A computer-readable storage medium, wherein the computer-readable storage medium stores a computer program which, when executed by a processor, causes the processor to implement the method of managing machine data according to any one of claims 1 to 8.

Description

Machine data management method, equipment and storage medium Technical Field The present application relates to the field of semiconductors, and in particular, to a method and apparatus for managing machine data, and a storage medium. Background In the field of semiconductor fabrication, as process nodes continue to evolve toward advanced processes, the complexity of chip production increases exponentially. The production process of modern wafer factories comprises hundreds or even thousands of process steps, and comprises a plurality of key links such as photoetching, etching, film deposition, ion implantation, chemical mechanical polishing and the like. These processes are typically performed in concert by a wide variety of specialized machines with varying functions. In order to ensure yield and realize precise process control, each machine is integrated with a large number of high-precision sensors for monitoring hundreds or thousands of process parameters such as temperature, pressure, gas flow, radio frequency power, vibration frequency and the like in real time. In order to realize fine monitoring, fault prediction and root cause analysis of the production process, the industry tends to adopt high-frequency (millisecond or microsecond sampling) and large-dynamic-range data acquisition strategies. This strategy, while capable of capturing transient process fluctuations and subtle anomaly signals, also results in explosive growth of data volume. When a single advanced machine operates at full load, the time sequence data generated every day can reach tens of GB or even TB level, and the total data generated every day by the whole wafer factory is higher than PB level, so that efficient storage and management optimization of the data is required. For example, patent application publication number CN119937937A discloses an intelligent storage method for data collected by a semiconductor device, which comprises the steps of carrying out sectional processing on a time sequence data sequence, determining initial abnormality degree of data in each time period according to trend abnormality characteristics of data change trend in each time period in the time sequence data sequence and abnormality characteristics of data period in each time period, determining accurate abnormality degree of data in each time period according to possibility of real abnormality data in each time period and initial abnormality degree of data in each time period, determining dead zone threshold value and dead zone range corresponding to each time period according to accurate abnormality degree of data in each time period and average value of adjacent data difference absolute values in each time period, and carrying out efficient compression processing on the time sequence data according to dead zone range corresponding to each time period. For example, the patent application with the publication number of CN120811397A discloses a method, a device, electronic equipment and a medium for compressing operation and maintenance data in the general semiconductor industry, and relates to the technical field of data compression. The method comprises the steps of obtaining the semiconductor industry operation and maintenance data to be compressed, identifying the time sequence of the semiconductor industry operation and maintenance data to be compressed, judging whether the semiconductor industry operation and maintenance data to be compressed are time sequence data according to the time sequence, splitting the semiconductor industry operation and maintenance data to be compressed into a time data set and a characteristic data set if the semiconductor industry operation and maintenance data to be compressed are the time sequence data, and compressing the time data set and the characteristic data set respectively. However, the existing data storage and compression technology reduces the data volume to a certain extent, but it is difficult to satisfy the stringent requirements of data fidelity and response speed in the semiconductor manufacturing scene. Disclosure of Invention The application mainly aims to provide a machine data management method, equipment and a storage medium. In order to solve the technical problems, the application adopts the following technical scheme: The first aspect of the present application provides a method for managing machine data, where the method includes: S101, acquiring a target process step executed by a target machine and machine data generated when the process step is executed; S102, determining ageing information of machine station data according to the procedure category and/or the procedure sequence position of the target process step, wherein the ageing information comprises at least one of a life cycle stage window, an analysis evaluation window and a preset calling window; S103, dividing a future time axis into a plurality of continuous time intervals which are not overlapped with each other based o