Method and system for detection of tool performance degradation and mismatch

A method, comprising:

• measuring a value of a reference variable associated with a first tool that manufactures at least one asset and a second tool that manufactures the at least one asset;
• measuring a value for each variable in a set of influencing variables that influences the reference variable;
• predicting a value for the reference variable based on a learned formal relationship among the reference variable and the set of influencing variables associated with the first tool that manufactures the at least one asset and the second tool that manufactures the at least one asset;
• removing an influencing variable from the set of influencing variables in response to a determination that a weight associated with the influencing variable satisfies a function of a threshold value and a previous weight associated with the influencing variable does not satisfy the function of the threshold value;
• adjusting the learned formal relationship among the reference variable and the set of influencing variables in response to the removing of the influencing variable from the set of influencing variables in order to account for the removing of the influencing variable from the set of influencing variables;
• and
• reporting at least one of performance degradation of at least one tool or performance mismatch amongst the first tool that manufactures the at least one asset and the second tool that manufactures the at least one asset in response to a difference among the predicted value for the reference variable and the measured value for the reference variable being above a performance metric threshold.

A non-transitory computer-readable storage medium comprising a set of instructions that, when executed by a computer, cause the computer to perform the following acts:

• measuring a set of values of a reference variable associated with a first tool for manufacturing an asset and a second tool for manufacturing the asset;
• measuring a set of values for each variable in a set of influencing variables that influences the reference variable;
• learning autonomously a formal relationship among the reference variable and the set of influencing variables associated with the first tool for manufacturing the asset and the second tool for manufacturing the asset;
• predicting a set of values for the reference variable based on the learned formal relationship;
• removing an influencing variable from the set of influencing variables in response to a determination that a first weight associated with the influencing variable is above a threshold value and a second weight associated with the influencing variable is below the threshold value;
• modifying the formal relationship among the reference variable and the set of influencing variables in response to the removing of the influencing variable from the set of influencing variables in order to account for the removing of the influencing variable from the set of influencing variables;
• and
• reporting performance degradation of at least a tool or performance mismatch of the first tool for manufacturing the asset and the second tool for manufacturing the asset when a difference among a predicted value of the reference variable and a measured value of the reference variable is above a key performance indicator threshold.

An apparatus, comprising:

• means for identifying a set of variables and data associated with a first unit of equipment in a tool system that produces an asset and a second unit of equipment in the tool system that produces the asset, wherein the set of variables includes a reference variable and a set of influencing variables that influences the reference variable;
• means for autonomously learning a functional relationship among the reference variable and the set of influencing variables associated with the first unit of equipment and the second unit of equipment;
• means for removing a selected influencing variable from the set of influencing variables in response to a determination that a weight associated with the selected influencing variable is below a set value and a previous weight associated with the selected influencing variable is above the set value;
• means for autonomously updating the learned functional relationship among the reference variable and the set of influencing variables in response to the removing of the selected influencing variable from the set of influencing variables in order to account for the removing of the selected influencing variable from the set of influencing variables;
• and
• means for assessing at least one of performance degradation or performance mismatch of the first unit of equipment and the second unit of equipment in the tool system based at least in part on values of the reference variable predicted through the learned functional relationship.