Learned model generation method

A method of generating a learned model for maintenance of a production apparatus, the production apparatus including:

• a first supply reel configured to supply a first electrode sheet;
• a second supply reel configured to supply a second electrode sheet;
• a first bonding roller provided on a first electrode sheet side;
• a second bonding roller provided on a second electrode sheet side and paired with the first bonding roller, the second bonding roller being configured to bond the first electrode sheet and the second electrode sheet together;
• a winding core around which the first electrode sheet and the second electrode sheet bonded together are wound;
• and
• a sensor configured to read a first end surface of the first electrode sheet and a second end surface of the second electrode sheet along a radial direction of a roll member of the first electrode sheet and the second electrode sheet wound around the winding core, the method comprising:
• acquiring, from the sensor, first data representing a position of the first end surface read along the radial direction and second data representing a position of the second end surface read along the radial direction;
• outputting information indicating that the roll member has a defect and that a cause of the defect is the first supply reel or the first bonding roller to a display apparatus for maintenance when it is determined that a sequence of the position of the first end surface represented by the first data intersects a sequence of the position of the second end surface represented by the second data on a basis of the first data and the second data;
• and
• generating the learned model such that on a basis of the first data and the second data before and after maintenance of the first supply reel, when it is determined that a fifth defect rate of the roll member after the maintenance of the first supply reel is determined to be equal to or greater than a predetermined value, the first data and the second data read before the maintenance of the first supply reel are not used for the generating of the learned model, whereas when it is determined that the fifth defect rate is smaller than the predetermined value, the first data and the second data read before the maintenance of the first supply reel are used for the generating of the learned model.

A method of generating a learned model for maintenance of a production apparatus, the production apparatus including:

• a first supply reel configured to supply a first electrode sheet;
• a second supply reel configured to supply a second electrode sheet;
• a first bonding roller provided on a first electrode sheet side;
• a second bonding roller provided on a second electrode sheet side and paired with the first bonding roller, the second bonding roller being configured to bond the first electrode sheet and the second electrode sheet together;
• a winding core around which the first electrode sheet and the second electrode sheet bonded together are wound;
• and
• a sensor configured to read a first end surface of the first electrode sheet and a second end surface of the second electrode sheet along a radial direction of a roll member of the first electrode sheet and the second electrode sheet wound around the winding core, the method comprising:
• acquiring, from the sensor, first data representing a position of the first end surface read along the radial direction and second data representing a position of the second end surface read along the radial direction;
• generating a learned model for determining that a cause of a defect of the roll member is the first supply reel or the first bonding roller when it is determined that a sequence of the position of the first end surface represented by the first data intersects a sequence of the position of the second end surface represented by the second data on a basis of the first data and the second data;
• outputting information indicating that the cause of the defect is the first supply reel or the first bonding roller to a display apparatus for maintenance on a basis of a determination using the learned model;
• and
• updating the learned model such that when it is determined that a first difference between a first probability that the defect of the roll member is improved and a second probability that the defect of the roll member is improved is smaller than a predetermined value, the first data and the second data read before the maintenance of the first supply reel are not used for the updating of the learned model, whereas when it is determined that the first difference is equal to or greater than the predetermined value, the first data and the second data before the maintenance of the first supply reel are used for the updating of the learned model, the first probability being obtained by inputting the first data and the second data before the maintenance of the first supply reel to the learned model, the second probability being obtained by inputting by the first data and the second data after the maintenance of the first supply reel to the learned model.