US5283418A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Automated rotor welding processes using neural networks
Assignee: Siemens Energy Inc
•
Filing: 1992-02-27
•
Granted: 1994-02-01
•
Expires: 2012-02-27
Abstract
Abstract Methods and apparatus for monitoring an arc welding process are disclosed. In a preferred embodiment, the present invention creates a digital representation of the arc created during welding and, using a neural network computer, determines if the arc is representative of normal or abnormal welding conditions. The neural network disclosed is trained to identify abnormal conditions and normal conditions and may be adaptively retrained to classify images that are not in the initial set of normal and abnormal images. In certain embodiments, other data, such as current, weld wire emission spectra, or shielding gas flow rate are also collected and the neural network is trained to monitor these data. Also, in certain embodiments, an audio signal is collected from the vicinity of the welding process and is used by the neural network computer to further classify the arc as normal or abnormal. The present invention is most preferably implemented in repetitive and continuous welding operations, such as those encountered in the manufacture and rebuilding of steam turbines.
Inventors
- 👤 James C. Bellows
- 👤 Serge P. Barton
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-10-29 09:56
CPC Classifications
B; B23; B23K; B23K9/00; B23K9/12; B23K9/095; B23K9/0956; B23K31/00; B23K31/006
Claims Analysis
Total Claims: 28
Claim 1
Independent Claim
86 words
A method of controlling arc welding apparatus comprising the steps of: recording one or more arc images formed during normal and abnormal conditions respectively during a welding operation; training a neural network computer to recognize arc images indicative of normal and abnormal conditions; observing an arc during an occurring welding operation, and generating arc data; determining if said arc data represents a normal or abnormal condition using the neural network computer; and generating a control signal if the occurring welding operation is in an abnormal condition.
Claim 2
Independent Claim
86 words
A method of controlling arc welding apparatus comprising the steps of: recording one or more arc images formed during normal and abnormal conditions respectively during a welding operation; training a neural network computer to recognize arc images indicative of normal and abnormal conditions; observing an arc during an occurring welding operation, and generating arc data; determining if said arc data represents a normal or abnormal condition using the neural network computer; and generating a control signal if the occurring welding operation is in an abnormal condition.
Claim 3
Dependent Claim
30 words
The method of claim 1 wherein the step of observing an arc comprises viewing the arc using a video camera to create a video image and digitizing the video image.
Claim 4
Dependent Claim
18 words
The method of claim 2, further comprising the step of transmitting the video image to a video monitor.
Claim 5
Dependent Claim
18 words
The method of claim 2, further comprising the step of transmitting the video image to a video recorder.
Claim 6
Dependent Claim
13 words
The method of claim 1 wherein the control signal activates an audible alarm.
Claim 7
Dependent Claim
31 words
The method of claim 1, further comprising the step of processing the control signal to characterize the abnormal condition; and displaying an error message identifying the nature of the abnormal condition.
Claim 8
Dependent Claim
62 words
The method of claim 1, further comprising the steps of collecting data indicative of the condition of one or more welding process parameters; training the neural network computer to recognize normal and abnormal values for said welding process parameters; and generating a control signal if the occurring welding condition exhibits an abnormal value for one or more of said welding process parameters.
Claim 9
Dependent Claim
20 words
The method of claim 7 wherein one of the welding process parameters for which data is collected is electrical current.
Claim 10
Dependent Claim
26 words
The method of claim 8 wherein the step of collecting data comprises determining changes in welding wire composition by measuring the emission spectrum of the weld.
Claim 11
Dependent Claim
43 words
The method of claim 1, further comprising the steps of recording an audio signal generated during a welding operation, digitizing the signal and processing the signal in conjunction with the arc images to determine if the welding operation is in an abnormal condition.
Claim 12
Dependent Claim
106 words
Automated arc welding apparatus comprising: a power supply and source of shielding gas in combination with and connected to an arc torch for welding weld wire to an object; filter means for providing a view of the arc created during welding; camera means for creating an image of the view; a neural network computer comprising image digitizing means for producing a digital representation of the image and means for classifying the arc using the image, comprising a neural network computer apparatus programmed to classify the digital representation as a function of its training; and error signal means for indicating the existence of an abnormal arc image.
Claim 13
Dependent Claim
106 words
Automated arc welding apparatus comprising: a power supply and source of shielding gas in combination with and connected to an arc torch for welding weld wire to an object; filter means for providing a view of the arc created during welding; camera means for creating an image of the view; a neural network computer comprising image digitizing means for producing a digital representation of the image and means for classifying the arc using the image, comprising a neural network computer apparatus programmed to classify the digital representation as a function of its training; and error signal means for indicating the existence of an abnormal arc image.
Claim 14
Dependent Claim
19 words
The apparatus of claim 11, further comprising a monitor means for viewing the image created by the camera means.
Claim 15
Dependent Claim
20 words
The apparatus of claim 11, further comprising a video recorder means for recording the image created by the camera means.
Claim 16
Dependent Claim
32 words
The apparatus of claim 11, further comprising means for collecting data indicative of the condition of one or more welding parameters and means for transmitting said data to the neural network computer.
Claim 17
Dependent Claim
22 words
The apparatus of claim 14 wherein said means for collecting data comprises means for measuring the current drawn by the power supply.
Claim 18
Dependent Claim
31 words
The apparatus of claim 14, wherein said means for collecting data comprises emission spectrometer means for measuring the emission spectrum of the arc to determine composition changes in the welding wire.
Claim 19
Dependent Claim
26 words
The apparatus of claim 14, wherein said means for collecting data comprises flowmeter means for collecting data indicative of the flow rate of the shielding gas.
Claim 20
Dependent Claim
18 words
The apparatus of claim 11, wherein said error signal means comprises means for creating an audible error signal.
Claim 21
Dependent Claim
26 words
The apparatus of claim 11, wherein said error signal means comprises display means for creating a visual indication of the occurrence of an abnormal arc image.
Claim 22
Dependent Claim
44 words
The apparatus of claim 11 further comprising at least a second arc torch for welding weld wire to an object, spaced apart from the first arc torch, and a camera means for creating an image of the arc created by said second arc torch.
Claim 23
Dependent Claim
40 words
The apparatus of claim 11, further comprising a microphone for collecting an audio signal created during welding that is connected to the neural network computer, whereby the audio signal is used by the neural network computer to classify the arc.
Claim 24
Independent Claim
119 words
A method of training a neural network computer for monitoring an arc welding process comprising the steps of: creating a plurality of digital representations of normal arc images; transmitting the normal image representations to the neural network computer along with information identifying the images as normal; creating a plurality of digital representations of abnormal arc images; transmitting the abnormal image representations to the neural network computer along with information identifying the images as abnormal; recording video images of new arc images that differ from the normal and abnormal images; creating a digital representation of the new arc images; transmitting the new image representations to the neural network computer along with information classifying the new image as normal or abnormal.
Claim 25
Independent Claim
119 words
A method of training a neural network computer for monitoring an arc welding process comprising the steps of: creating a plurality of digital representations of normal arc images; transmitting the normal image representations to the neural network computer along with information identifying the images as normal; creating a plurality of digital representations of abnormal arc images; transmitting the abnormal image representations to the neural network computer along with information identifying the images as abnormal; recording video images of new arc images that differ from the normal and abnormal images; creating a digital representation of the new arc images; transmitting the new image representations to the neural network computer along with information classifying the new image as normal or abnormal.
Claim 26
Dependent Claim
15 words
The method of claim 22, wherein the arc welding process comprises welding steam turbine components.
Claim 27
Dependent Claim
33 words
The method of claim 22, further comprising the steps of creating an audio signal representative of a welding operation; and training a neural network computer to distinguish between normal and abnormal audio signals.
Claim 28
Dependent Claim
60 words
The method of claim 22 further comprising the steps of: transmitting data indicative of normal and abnormal arcs to a second neural network computer; concurrently inputting data indicative of one or more welding process parameters into the second neural network computer; and identifying combinations of welding process parameters that produce the normal arcs and those that produce the abnormal arcs.
References Cited (Backward Citations)
US4739404A
Cited
US4561059A
Cited
US4589140A
Cited
US4679077A
Cited
US4660166A
Cited
US4724524A
Cited
US4633554A
Cited
US4675730A
Cited
US4951218A
Cited
US4869109A
Cited
US4868649A
Cited
US4877940A
Cited
US4846928A
Cited
US4853680A
Cited
US4897519A
Cited
US4965725A
Cited
US4965725B1
Cited
US4940390A
Cited
US4903888A
Cited
US4916625A
Cited
US4989159A
Cited
US4954964A
Cited
US4893388A
Cited
US4912654A
Cited
US4933871A
Cited
US4918517A
Cited
US4918427A
Cited
US4995087A
Cited
US4998005A
Cited
US5001714A
Cited
US5159660A
Cited
US5204872A
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 1997-01-31 | 1997-08-02 | 1998-02-01 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2001-01-31 | 2001-08-02 | 2002-02-01 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2005-01-31 | 2005-08-02 | 2006-02-01 | $7700.00 | Window Closed / Lapsed |