Method of predictive detection of defects in absorbent sanitary articles

A production line of absorbent sanitary articles, comprising:

• production devices configured to perform respective production operations on the absorbent sanitary articles being processed along the production line;
• sensors adapted to detect current values of operating parameters which are indicative of the current operation of at least one of the production devices;
• a control unit comprising a memory and a processor configured to perform a comparison between the current values of the operating parameters detected by the sensors and respective predetermined reference values stored in said memory and, in a presence among said operating parameters of at least one anomalous operating parameter whose current value does not comply with the respective predetermined reference value, to identify, among a plurality of possible malfunctioning causes of the line, at least one current malfunctioning cause responsible for said at least one anomalous operating parameter;
• wherein the processor is configured to identify said at least one current malfunctioning cause starting from said at least one anomalous operating parameter by means of decision algorithms based on at least one of the following information stored in said memory:
• known and/or self-learned data related to correlations between said at least one anomalous operating parameter and said plurality of possible malfunctioning causes of the line;
• current and/or historical information manually provided by an operator in association with said at least one anomalous operating parameter and/or with said plurality of possible malfunctioning causes of the line;
• known and/or self-learned logical rules;
• and
• wherein the processor is configured to perform, by means of said decision algorithms, a prediction of defects in the absorbent sanitary articles being processed along the line and/or in absorbent sanitary articles in output from the production line starting from said at least one current malfunctioning cause identified and based on at least one of the following information stored in said memory:
• known and/or self-learned data related to correlations between said at least one current malfunctioning cause identified and at least one known possible defect;
• current and/or historical information manually provided by an operator in association with said at least one current malfunctioning cause identified and/or with said at least one known possible defect;
• known and/or self-learned logic rules.