Production equipment monitoring system and production equipment monitoring method
View Patent ↗A production equipment monitoring system 20 including: an anomaly index determination unit 23 determining an anomaly index a of a production equipment 10 , based on a feature quantity obtained from an equipment information on the production equipment 10 ; a relevance determination unit 24 determining a degree of relevance D between each of anomalous conditions predicable to occur in the production equipment 10 and an observed condition of the production equipment 10 ; a detection threshold determination unit 25 determining a single detection threshold th for detecting a degree of anomaly A of the production equipment 10 , based on anomaly thresholds at which are thresholds of the anomaly index a corresponding respectively to the anomalous conditions, and on the degree of relevance D; and an anomaly degree detection unit 26 detecting the degree of anomaly A of the production equipment 10 , based on the anomaly index a and the detection threshold th.
1. A production equipment monitoring system, comprising:
a computer configured to:
determines an anomaly index of a production equipment, based on a feature quantity obtained from an equipment information on the production equipment;
determines a degree of relevance between each of a plurality of anomalous conditions predicable to occur in the production equipment and an observed condition of the production equipment;
determines a single detection threshold for detecting a degree of anomaly of the production equipment, based on a plurality of anomaly thresholds which are thresholds of the anomaly index corresponding respectively to the plurality of anomalous conditions, and on the degree of relevance;
detects the degree of anomaly of the production equipment, based on the anomaly index and the detection threshold; and
notify a degree of anomaly of the production equipment, wherein
when the anomalous condition whose degree of relevance with the observed condition is the highest is referred to as a first anomalous condition, and the anomalous condition whose degree of relevance with the observed condition is the second highest is referred to as a second anomalous condition,
the computer is configured to determine the detection threshold by interpolating between the anomaly threshold of the first anomalous condition and the anomaly threshold of the second anomalous condition, based on a Mahalanobis distance between the first anomalous condition and the observed condition, and on a Mahalanobis distance between the second anomalous condition and the observed condition, and
the computer is configured to notify the degree of anomaly of the production equipment so as to support a person who manages the production equipment to make a determination, before failures occur, on the maintenance timing and the maintenance target of the production equipment.
2. The production equipment monitoring system according to claim 1 , wherein
the feature quantity includes a first feature quantity and a second feature quantity,
the production equipment monitoring system further includes a normal model that simulates a normal condition of the production equipment, with the first feature quantity obtained from the equipment information on the production equipment in the normal condition used as an input, and the second feature quantity obtained from the equipment information on the production equipment in the normal condition used as an output,
the computer is configured to determines the anomaly index of the production equipment, based on the second feature quantity that is output from the normal model in response to an input of the first feature quantity of the equipment information observed, and on the second feature quantity of the equipment information observed.
3. The production equipment monitoring system according to claim 1 , wherein in notifying the degree of anomaly of the production equipment, the computer notifies the anomalous condition whose degree of relevance with the observed condition is the highest.
4. The production equipment monitoring system according to claim 3 , wherein in notifying the degree of anomaly of the production equipment, the notification unit the computer notifies also the anomalous condition whose degree of relevance with the observed condition is the second highest.
5. The production equipment monitoring system according to claim 1 , wherein
the production equipment is a plasma processing apparatus including a processing chamber in which plasma processing is performed, and
the degree of anomaly of the production equipment includes a degree of anomaly of an element placed in the processing chamber.
6. A production equipment monitoring method, comprising:
an anomaly index determination step of determining an anomaly index of a production equipment, based on a feature quantity obtained from an equipment information on the production equipment;
a relevance determination step of determining a degree of relevance between each of a plurality of anomalous conditions predicable to occur in the production equipment and an observed condition of the production equipment;
a detection threshold determination step of determining a single detection threshold for detecting an anomaly index of the production equipment, based on a plurality of anomaly thresholds which are thresholds of the anomaly index corresponding respectively to the plurality of anomalous conditions, and on the degree of relevance;
an anomaly degree detection step of detecting a degree of anomaly of the production equipment, based on the anomaly index and the detection threshold; and
a notification step of notifying the degree of anomaly of the production equipment, wherein
when the anomalous condition whose degree of relevance with the observed condition is the highest is referred to as a first anomalous condition, and the anomalous condition whose degree of relevance with the observed condition is the second highest is referred to as a second anomalous condition,
in the detection threshold determination step, the detection threshold is determined by interpolating between the anomaly threshold of the first anomalous condition and the anomaly threshold of the second anomalous condition, based on a Mahalanobis distance between the first anomalous condition and the observed condition, and on a Mahalanobis distance between the second anomalous condition and the observed condition, and
in the notification step, the degree of anomaly of the production equipment is notified so as to support a person who manages the production equipment to make a determination, before failures occur, on the maintenance timing and the maintenance target of the production equipment.
7. The production equipment monitoring method according to claim 6 , wherein
the feature quantity includes a first feature quantity and a second feature quantity, and
in the anomaly index determination step, a normal model is used, the normal model being configured to simulate a normal condition of the production equipment, with the first feature quantity obtained from the equipment information of the production equipment in the normal condition used as an input, and the second feature quantity obtained from the equipment information of the production equipment in the normal condition used as an output, and the anomaly index of the production equipment is determined based on the second feature quantity that is output in response to an input into the normal model of the first feature quantity of the equipment information observed, and on the second feature quantity of the equipment information observed.
8. The production equipment monitoring method according to claim 6 , wherein
the production equipment is a plasma processing apparatus including a processing chamber in which plasma processing is performed, and
the degree of anomaly of the production equipment includes a degree of anomaly of an element placed in the processing chamber.