One variation of a method for predicting manufacturing defects includes: accessing a set of inspection images of a set of assembly units recorded by an optical inspection station; for each inspection image in the set of inspection images, detecting a set of features in the inspection image and generating a vector representing the set of features in a multi-dimensional feature space; grouping neighboring vectors in the multi-dimensional feature space into a set of vector groups; and, in response to receipt of a first inspection result indicting a defect in a first assembly unit, in the set of assembly units, associated with a first vector in a first vector group, in the set of vector groups, labeling the first vector group with the defect and flagging a second assembly unit associated with a second vector, in the first vector group, as exhibiting characteristics of the defect.

According to one embodiment, there is provided a quality controlling device including: a predictor, a frequency calculator, and an implementing signal creator. The predictor employs a prediction model that associates an inspection result value of a first inspection with a predicted value being a value relating to a possibility of pass or failure in a second inspection and calculates the predicted value from an inspection result value that is obtained for an inspection target in the first inspection. The frequency calculator calculates, for the inspection target, an implementation frequency to implement the second inspection in accordance with the predicted value calculated by the predictor. The implementing signal creator creates a signal that indicates, for the inspection target, necessity of implementing the second inspection in accordance with the implementation frequency.

A quality control apparatus (20) includes: a regression analyzer (33) for calculating a regression formula on the basis of measurement values acquired from an upstream step and comparative measurement values acquired from a downstream step; a margin determination unit (34) for calculating a prediction value by substituting a determination reference value defining a determination reference range in the upstream step for an explanatory variable of the regression formula, comparing the prediction value with a comparative determination reference range in a downstream step, and determining whether the measurement values are accepted; and a reference value calculator (35) for calculating a new determination reference value to replace the determination reference value in accordance with the determination result.

This invention teaches a multi-sensor quality inference system for additive manufacturing. This invention still further teaches a quality system that is capable of discerning and addressing three quality issues: i) process anomalies, or extreme unpredictable events uncorrelated to process inputs; ii) process variations, or difference between desired process parameters and actual operating conditions; and iii) material structure and properties, or the quality of the resultant material created by the Additive Manufacturing process. This invention further teaches experimental observations of the Additive Manufacturing process made only in a Lagrangian frame of reference. This invention even further teaches the use of the gathered sensor data to evaluate and control additive manufacturing operations in real time.

Aspects of the disclosed technology encompass the use of a deep learning controller for monitoring and improving a manufacturing process. In some aspects, a method of the disclosed technology includes steps for: receiving a plurality of control values from two or more stations, at a deep learning controller, wherein the control values are generated at the two or more stations deployed in a manufacturing process, predicting an expected value for an intermediate or final output of an article of manufacture, based on the control values, and determining if the predicted expected value for the article of manufacture is in-specification. In some aspects, the process can further include steps for generating control inputs if the predicted expected value for the article of manufacture is not in-specification. Systems and computer-readable media are also provided.

Aspects of the disclosed technology encompass the use of a deep learning controller for monitoring and improving a manufacturing process. In some aspects, a method of the disclosed technology includes steps for: receiving a plurality of control values from two or more stations, at a deep learning controller, wherein the control values are generated at the two or more stations deployed in a manufacturing process, predicting an expected value for an intermediate or final output of an article of manufacture, based on the control values, and determining if the predicted expected value for the article of manufacture is in-specification. In some aspects, the process can further include steps for generating control inputs if the predicted expected value for the article of manufacture is not in-specification. Systems and computer-readable media are also provided.

A quality abnormality analysis method for a product manufactured by a manufacturing process includes: predicting quality of the product by inputting manufacturing conditions to a quality prediction model generated by using a plurality of manufacturing conditions of the manufacturing process as input variables and using the quality of the product as an output variable; calculating a quality evaluation value of an actual product manufactured by the manufacturing process; calculating, as a quality prediction error, a difference between a quality prediction value obtained as an output and the quality evaluation value; a quality contribution calculation step of quality contribution degrees of the input manufacturing conditions when predicting the quality of the product using the quality prediction model; and presenting, based on the quality prediction error and the quality contribution degree, a manufacturing condition causing a quality abnormality of the product.

A machine learning device of a finish-machining amount prediction apparatus observes, as state variables expressing a current state of an environment, finish-machining amount data indicating finish-machining amounts of the respective parts of a component and accuracy data indicating the accuracy of the respective parts of a machine, to which the component is attached. Then, the machine learning device acquires determination data indicating propriety determination results of the accuracy of the respective parts of the machine, to which the component after being subjected to finish machining is attached. After that, the machine learning device learns the finish-machining amounts of the respective parts of the component in association with the accuracy data by using the state variables and the determination data.

A system, method, and computer-readable medium are disclosed for analysis and characterization of manufacturing information such as process trees or genealogies using graph theory. More specifically, using graph theory to analyze manufacturing information of a manufacturing operation allows for deep analysis of relationships between batches or units in a process tree and their closeness or distance, to identify clusters associated with specific quality characteristics or problems, to identify common antecedents of specifically labeled batches (e.g., problem batches), and/or to detect overall desirable or undesirable characteristics of the process tree (e.g., centrality, etc.).

Provided is a management system for performing quality management on production equipment. A management system that includes: an acquisition component that acquires status information for production equipment that is subject to management; a detection component that, on the basis of the acquired status information, detects the occurrence of some event; and a display component that displays, separated according to inclusion in the four perspectives Machine, Man, Material, and Method, a plurality of factors that could be presumed to have caused the detected event in a manner in which the contents thereof and a probability of having caused the event can be compared.