A machining step monitoring apparatus includes a data segmentation unit, a data length adjuster, and a determination processor. The data segmentation unit defines machining start time and machining end time of each single cycle, and segments, from a first physical quantity that temporally changes during the cycle machining in the machining machine, first measurement data indicating the first physical quantity from the machining start time to the machining end time for each single cycle. The data length adjuster generates the second measurement data by adjusting the data length of the first measurement data to match with the data length of the determination reference data indicating, as a reference, a change in the first physical quantity during the single cycle when the machining machine is normally operating. The determination processor compares the second measurement data with the determination reference data to determine whether the machining machine is normally operating.

A processor performs an experiment of machining a device to acquire first-type and second-type information each indicating conditions of the experiment of machining and third-type and fourth-type information each indicating a result of the experiment of machining (S401). The processor derives a first expression and a second expression, where the first expression receives first-type and second-type information as inputs and outputs third-type information as more than one solution, and the second expression receives first-type and second-type information as inputs and outputs fourth-type information. The processor derives more than one third expression from the first expression, where the more than one third expression each receives second-type and third-type information as inputs and outputs first-type information (S402). The processor receives second-type and third-type information each measured in machining as inputs and outputs fourth-type information indicating a result of machining using the second expression and the more than one third expression (S403).

A processor performs an experiment of machining a device to acquire first-type and second-type information each indicating conditions of the experiment of machining and third-type and fourth-type information each indicating a result of the experiment of machining (S401). The processor derives a first expression and a second expression, where the first expression receives first-type and second-type information as inputs and outputs third-type information as more than one solution, and the second expression receives first-type and second-type information as inputs and outputs fourth-type information. The processor derives more than one third expression from the first expression, where the more than one third expression each receives second-type and third-type information as inputs and outputs first-type information (S402). The processor receives second-type and third-type information each measured in machining as inputs and outputs fourth-type information indicating a result of machining using the second expression and the more than one third expression (S403).

A method and system for identifying parts manufactured by a workstation by measuring signals generated by machines in the workstation, extracting features from the signals, clustering the features into clusters, associating clusters with manufactured parts and recognizing the parts through the clusters.

An abnormality detector includes a signal output unit for detecting a sign of an abnormality based on a physical quantity acquired from a manufacturing machine and outputting a signal; and a machine learning device including state observation unit for observing, as a state variable representing a present state of the environment, physical quantity data indicating the physical quantity related to an operation of the manufacturing machine from the manufacturing machine; a label data acquisition unit for acquiring, as label data, operation state data indicating an operation state of the manufacturing machine; a learning unit for learning the operation state of the manufacturing machine with respect to the physical quantity, using the state variable and the label data; and an estimation result output unit for estimating the operation state of the manufacturing machine using a learning result by the learning unit and outputting an estimation result.

An abnormality detector includes a signal output unit for detecting a sign of an abnormality based on a physical quantity acquired from a manufacturing machine and outputting a signal; and a machine learning device including state observation unit for observing, as a state variable representing a present state of the environment, physical quantity data indicating the physical quantity related to an operation of the manufacturing machine from the manufacturing machine; a label data acquisition unit for acquiring, as label data, operation state data indicating an operation state of the manufacturing machine; a learning unit for learning the operation state of the manufacturing machine with respect to the physical quantity, using the state variable and the label data; and an estimation result output unit for estimating the operation state of the manufacturing machine using a learning result by the learning unit and outputting an estimation result.

A method includes obtaining data associated with operation of a model-based industrial process controller. The method also includes identifying at least one estimated impact of at least one operational problem of the industrial process controller, where each estimated impact is expressed in terms of a lost opportunity associated with operation of the industrial process controller. The method further includes presenting the at least one estimated impact to a user. The at least one estimated impact could include impacts associated with noise or variance in process variables used by the industrial process controller, misconfiguration of an optimizer in the industrial process controller, one or more limits on one or more process variables, a quality of at least one model used by the industrial process controller, a quality of one or more inferred properties used by the industrial process controller, or one or more process variables being dropped from use by the industrial process controller.

A method includes obtaining data associated with operation of a model-based industrial process controller. The method also includes identifying at least one estimated impact of at least one operational problem of the industrial process controller, where each estimated impact is expressed in terms of a lost opportunity associated with operation of the industrial process controller. The method further includes presenting the at least one estimated impact to a user. The at least one estimated impact could include impacts associated with noise or variance in process variables used by the industrial process controller, misconfiguration of an optimizer in the industrial process controller, one or more limits on one or more process variables, a quality of at least one model used by the industrial process controller, a quality of one or more inferred properties used by the industrial process controller, or one or more process variables being dropped from use by the industrial process controller.

A method of generating training spectra for training of a neural network includes generating a plurality of theoretically generated initial spectra from an optical model, sending the plurality of theoretically generated initial spectra to a feedforward neural network to generate a plurality of modified theoretically generated spectra, sending an output of the feedforward neural network and empirically collected spectra to a discriminatory convolutional neural network, determining that the discriminatory convolutional neural network does not discriminate between the modified theoretically generated spectra and empirically collected spectra, and thereafter, generating a plurality of training spectra from the feedforward neural network.

A method of generating training spectra for training of a neural network includes generating a plurality of theoretically generated initial spectra from an optical model, sending the plurality of theoretically generated initial spectra to a feedforward neural network to generate a plurality of modified theoretically generated spectra, sending an output of the feedforward neural network and empirically collected spectra to a discriminatory convolutional neural network, determining that the discriminatory convolutional neural network does not discriminate between the modified theoretically generated spectra and empirically collected spectra, and thereafter, generating a plurality of training spectra from the feedforward neural network.