AI extensions and intelligent model validation for an industrial digital twin

A method, comprising:

• defining, on a system comprising a processor based on configuration input data, a digital twin that defines an industrial asset in terms of hierarchical elements, wherein the digital twin comprises data tags that correspond to data items maintained on devices of the industrial asset;
• performing, by the system, artificial intelligence (AI) analysis on the digital twin;
• identifying, by the system based on a result of the AI analysis, a key variable indicative of a quality of operation of the industrial asset, related variables that determine a value of the key variable, and a functional relationship between the key variable and the related variables;
• adding, by the system, AI fields to a data tag, of the data tags, that corresponds to the key variable, wherein the AI fields identify the related variables and the functional relationship between the key variable and the related variables;
• collecting, by the system, process data from the data items of the industrial asset;
• augmenting, by the system, the process data with calculated asset behavior data generated by applying the process data to the digital twin;
• and
• determining, by the system, whether the digital twin models an aspect of the industrial asset within a defined degree of accuracy based on a result of an AI validation analysis performed on the digital twin, the process data, and the calculated asset behavior data.

A non-transitory computer-readable medium having stored thereon instructions that, in response to execution, cause a system comprising a processor to perform operations, the operations comprising:

• defining, based on configuration input data, a digital twin that defines an industrial asset in terms of hierarchical elements, wherein the digital twin comprises data tags that correspond to data items maintained on devices of the industrial asset;
• performing artificial intelligence (AI) analysis on the digital twin;
• identifying, based on a result of the AI analysis, a key variable indicative of a performance quality of the industrial asset, related variables that determine a value of the key variable, and a functional relationship between the key variable and the related variables;
• appending AI fields to a data tag, of the data tags, that corresponds to the key variable, wherein the AI fields identify the related variables and the functional relationship between the key variable and the related variables;
• receiving process data from the data tags of the industrial asset;
• augmenting the process data with calculated asset behavior data generated based on application of the process data to the digital twin;
• and
• determining whether the digital twin models an aspect of the industrial asset within a defined degree of accuracy based on a result of an AI validation analysis performed on the digital twin, the process data, and the calculated asset behavior data.