Automatic modeling for monitoring, diagnostics, optimization and control
AssigneeRockwell Automation Technologies, Inc.
InventorsBijan SayyarRodsari, Alexander B. Smith, Kadir Liano
Life of this patent
Estimating the term…
Abstract
A modular modeling engine is provided for industrial automation applications. The module may be instantiated upon demand, such as upon receipt of annotated data for a system or process being monitored and/or controlled. The model is agnostic insomuch as little or no prior knowledge is required of the system or process. Variables, functions, and their combinations are selected and the model is refined automatically. A data structure is received for instantiation of the model, and following modeling, a similar data structure is produced. The module may be used together with other modules for caning out complex automation processing at the same or multiple levels in an automation setting.
Claims20
A non-transitory computer readable medium storing a program causing a computer to execute a method for a physical processor comprising:
- receiving, from a plurality of sensors, operational parameters of a controlled machine or process to produce sensed data;
- receiving annotated input data for respective input variables, the annotated input data being derived from the sensed data or from data generated by an automation controller;
- and
- generating a model, via an analytics engine executed by a processor that receives the annotated input data and employs a mixed-integer nonlinear programming solver to generate the model based upon an agnostic model stored in a memory and upon the annotated input data derived from the sensed data or from the data generated by an automation controller by naming of the input data, selection of the input data, selection of one or more parameters or variables of interest in the input data, or computations made on the input data, wherein the agnostic model comprises a set of candidate base functions that are used to build the model from variables of interest, and wherein the analytics engine selects input variables of interest of the system represented by the annotated input data for provision to the candidate base functions.
A non-transitory computer readable medium storing a program causing a computer to execute a method for a physical processor comprising:
- receiving, from a plurality of sensors, operational parameters of a controlled machine or process;
- generating a model, via an analytics engine executed by a processor that receives annotated input data derived from the sensed data or from data generated by an automation controller by naming of input data, selection of the input data, selection of one or more parameters or variables of interest in the input data, or computations made on the input data, based upon an agnostic model stored in a memory and upon the annotated input data, wherein the analytics engine selects input variables of interest of the system represented by the annotated input data for provision to candidate base functions of the agnostic model, and reduces an error between a prediction-based variable and a measurement-based variable based upon integer decision variables, a subset of the input variables, and a subset of the candidate functions.
A non-transitory computer readable medium storing a program causing a computer to execute a method for a physical processor comprising:
- receiving annotated input data derived from the sensed data or from data generated by an automation controller by naming of input data, selection of the input data, selection of one or more parameters or variables of interest in the input data, or computations made on the input data;
- and
- generating a model via an analytics engine executed by a processor based upon the annotated input data and an agnostic model stored in a memory;
- wherein the analytics engine selects input variables of interest of the system represented by the annotated input data for provision to candidate base functions of the agnostic model, and reduces an error between a prediction-based variable and a measurement-based variable based upon integer decision variables, a subset of the input variables, and a subset of the candidate functions.
Dependencies are inferred from the claim text.