Method and system for avoiding anomaly stop of production device

Abstract

A method for avoiding irregular shutoff of production equipment, includes: measuring regularly time-series data of characteristics of a rotary machine used in the production equipment running for the production; obtaining first failure diagnosis data subjecting the time-series data to a first real-time analysis; obtaining second failure diagnosis data subjecting the first failure diagnosis data to a second real-time analysis; predicting a status of the production equipment several minutes later using the second failure diagnosis data; and shutting off during a production process if the result of the prediction determines that the production equipment will shut off irregularly, and switching to a purge sequence for conducting a gas purge of the production equipment.

Description

technical field [0001] The present invention relates to an abnormal stop avoidance method and an abnormal stop avoidance system of a production device having a vacuum pump. Background technique [0002] In the failure prediction technology of vacuum pumps attached to semiconductor manufacturing equipment, the lifetime of pumps is predicted for the purpose of planning maintenance from a long-term standpoint. However, since a plurality of processing conditions with different pump loads are used to form a film in normal device operation, when a film is formed under a processing condition with a high pump load, the pump stops suddenly during film formation. [0003] When this phenomenon occurs, it is often the case that a batch of products during film formation has a lotout. In addition, since the pump stops abnormally during film formation, highly reactive and toxic gases are released into the atmosphere when the pump is replaced, deteriorating the working environment, and may...

AI Technical Summary

Problems solved by technology

However, depending on the gas used, even normal values ​​for these parameters can cause the pump to fail

Therefore, this parameter alone cannot predict failures, either in a single process or in a composite process

In the case of analyzing only the data of the pump and executing multiple processes, the monitored parameters fluctuate greatly because the pump load varies depending on the conditions, and life prediction cannot be performed.

[0005] As mentioned above, in the case of the production of multiple types of industrial products, when the same production equipment is used to produce multiple types of industrial products, various process conditions are required, and there is a threshold that cannot be uniquely determined as a criterion for predicting failures of production equipment. The problem

Images