Unlock instant, AI-driven research and patent intelligence for your innovation.

Automated stochastic methods for feature discovery and their use in repeated processes

A technology of signal characteristics and programming modules, applied to controllers with specific characteristics, program control, electric controllers, etc., can solve the problems of manufacturing unqualified workpieces and not always ensuring the quality of workpieces

Active Publication Date: 2020-03-20
GM GLOBAL TECH OPERATIONS LLC
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, using stable process control parameters can still produce parts of substandard quality, and therefore closed-loop threshold-based control methods do not always ensure stable part quality over time

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Automated stochastic methods for feature discovery and their use in repeated processes
  • Automated stochastic methods for feature discovery and their use in repeated processes
  • Automated stochastic methods for feature discovery and their use in repeated processes

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0015] Referring to the drawings, in which like reference numerals denote like parts throughout the several views, in figure 1 A repeated process 11 for producing a workpiece 30 is schematically shown in . The process 11 is repeated using respective first and second controllers in the form of a feature generation module (FGM) 55 and a process control module (PCM) 50, wherein the FGM 55 is used to generate predictive features in a stochastic manner and the PCM 50 uses the same predictive feature , for example in the overall control of the iterative process 11 and / or in the targeted product recall as explained herein. Subsequently, the FGM 55 temporally iteratively processes the generated features that can be automatically derived from the sensor data (arrows 28, 128), as the FGM 55 learns these features that are most predictive through evolutionary programming. as below reference figure 2 with 3 As stated, FGM 55 is programmed to perform method 100 .

[0016] As those of o...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

An automated method for feature discovery in an iterative process includes measuring raw time-series data during the process using sensors. Time series data describe multiple parameters of the process. The method includes receiving timing data from the sensor via a first controller, and randomly generating candidate features from the raw timing data using one or more logic blocks of the first controller. Candidate features predict the quality of workpieces manufactured through repeated processes. The method also includes determining, via the genetic or evolutionary programming module, which of the generated candidate features are most predictive of the quality of the workpiece, and using the most predictive candidate features, via the second controller, to perform control actions with respect to the iterative process. The system includes a controller, programming module and sensors.

Description

technical field [0001] The present disclosure relates to an automated stochastic method for feature discovery and its use in an iterative process. Background technique [0002] Various processes are repetitive and thus facilitate real-time process monitoring. An example of such an iterative process is ultrasonic welding, which involves the controlled application of high frequency vibrational energy to the interface surfaces of clamped workpieces. Surface friction generates heat, which eventually softens and bonds the interfacial surfaces. For a given workpiece, the formation of multiple identical welds typically occurs in a consistent, repeatable manner. Another exemplary procedure is cold testing of internal combustion engines, in which the performance of the engine is tested with the cylinders uncombusted (eg, by driving the engine via an electric motor), including static and dynamic leak tests. [0003] Conventional process control methods for repetitive processes incl...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): B23K20/10G01D21/02
CPCB23K20/10G01D21/02G05B19/41875G05B2219/32194Y02P90/02G05B11/42G05B13/0255G05B19/21G05B19/33G05B19/39G05B19/425G05B2219/36219
Inventor D·M·韦格纳J·A·阿贝尔M·A·温凯克
Owner GM GLOBAL TECH OPERATIONS LLC