Diagnostic apparatus for generating verification data including at least one piece of abnormal data based on normal data

1. A diagnostic apparatus for diagnosing an operating state of an industrial machine, the diagnostic apparatus comprising:

a data acquisitor for acquiring normal data related to an operating state during a normal operation of the industrial machine;

an acquired data storage for storing the normal data acquired by the data acquisitor;

a learner for generating a learning model by learning based on the normal data stored in the acquired data storage;

an estimator for performing an estimation process for normality or abnormality of an operation of the industrial machine using the learning model;

a verification data generator for generating verification data including at least one piece of abnormal data based on the normal data stored in the acquired data storage; and

a verificator for verifying validity of the learning model on receiving a result of the estimation process performed by the estimator using the learning model based on the verification data, wherein

the learner generates a plurality of learning models,

the verificator causes the estimator to perform estimation based on the verification data using each of the plurality of learning models, and

the verificator selects an average learning model from estimation results obtained by the estimator as a valid learning model.

2. A diagnostic apparatus for diagnosing an operating state of an industrial machine, the diagnostic apparatus comprising:

a data acquisitor for acquiring normal data related to an operating state during a normal operation of the industrial machine;

an acquired data storage for storing the normal data acquired by the data acquisitor;

a learner for generating a learning model by learning based on the normal data stored in the acquired data storage;

an estimator for performing an estimation process for normality or abnormality of an operation of the industrial machine using the learning model;

a verification data generator for generating verification data including at least one piece of abnormal data based on the normal data stored in the acquired data storage; and

a verificator for verifying validity of the learning model on receiving a result of the estimation process performed by the estimator using the learning model based on the verification data, wherein

the verificator causes the estimator to perform estimation based on the verification data using the learning model, and

the verificator determines that the learning model is invalid to regenerate a learning model when an estimation result obtained by the estimator dissatisfies a predetermined condition.

3. A diagnostic apparatus for diagnosing an operating state of an industrial machine, the diagnostic apparatus comprising:

a data acquisitor for acquiring normal data related to an operating state during a normal operation of the industrial machine;

an acquired data storage for storing the normal data acquired by the data acquisitor;

a learner for generating a learning model by learning based on the normal data stored in the acquired data storage;

an estimator for performing an estimation process for normality or abnormality of an operation of the industrial machine using the learning model;

a verification data generator for generating verification data including at least one piece of abnormal data by applying a predetermined change to the normal data stored in the acquired data storage; and

a verificator for verifying validity of the learning model on receiving a result of the estimation process performed by the estimator using the learning model based on the verification data,

wherein the verification data generator is configured to generate abnormal data by adding to normal data at least one of an impulse, a fixed value component, an ax+b component, a frequency component, a data value defect or a sampling defect,

the diagnostic apparatus further comprises a display, wherein when a generation target of the abnormal data is an image,

the verification data generator is configured to

add a dot of a predetermined color and a predetermined size to a predetermined position on the display, when the impulse is added to normal data,

change a color component of entire image data by a predetermined amount when the fixed value component is added to normal data,

change color components of the entire image data to gradation when the ax+b component is added to normal data,

add a predetermined two-dimensional frequency component to the entire image data when the frequency component is added to normal data, and

add a black or white point of a predetermined size to a predetermined position on the display when the data value defect is added to normal data, and

the verificator is configured to output an estimation result obtained by the estimator to the display.

4. The diagnostic apparatus according to claim 3 , wherein the learning model is an auto-encoder.

5. The diagnostic apparatus according to claim 3 , wherein the verification data is a set of abnormal data, or includes normal data and abnormal data at a predetermined ratio.