CN113762366A - Additive manufacturing forming state prediction control method and system - Google Patents
Additive manufacturing forming state prediction control method and system Download PDFInfo
- Publication number
- CN113762366A CN113762366A CN202110973673.5A CN202110973673A CN113762366A CN 113762366 A CN113762366 A CN 113762366A CN 202110973673 A CN202110973673 A CN 202110973673A CN 113762366 A CN113762366 A CN 113762366A
- Authority
- CN
- China
- Prior art keywords
- forming state
- additive manufacturing
- prediction
- forecast
- state label
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000004519 manufacturing process Methods 0.000 title claims abstract description 78
- 239000000654 additive Substances 0.000 title claims abstract description 76
- 230000000996 additive effect Effects 0.000 title claims abstract description 76
- 238000000034 method Methods 0.000 title claims abstract description 54
- 238000012545 processing Methods 0.000 claims abstract description 13
- 238000000605 extraction Methods 0.000 claims abstract description 9
- 230000015654 memory Effects 0.000 claims abstract description 8
- 230000008569 process Effects 0.000 claims description 26
- 230000004927 fusion Effects 0.000 claims description 22
- 230000006403 short-term memory Effects 0.000 claims description 13
- 230000007787 long-term memory Effects 0.000 claims description 12
- 238000013528 artificial neural network Methods 0.000 claims description 11
- 238000012549 training Methods 0.000 claims description 11
- 230000007547 defect Effects 0.000 claims description 9
- 239000002356 single layer Substances 0.000 claims description 9
- 238000013527 convolutional neural network Methods 0.000 claims description 8
- 125000004122 cyclic group Chemical group 0.000 claims description 8
- 239000010410 layer Substances 0.000 claims description 7
- 239000011324 bead Substances 0.000 claims description 6
- 238000002474 experimental method Methods 0.000 claims description 6
- 230000006870 function Effects 0.000 claims description 6
- 239000002893 slag Substances 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 5
- 238000004422 calculation algorithm Methods 0.000 claims description 4
- 238000002372 labelling Methods 0.000 claims description 4
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 238000003466 welding Methods 0.000 description 8
- 238000011217 control strategy Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 239000002699 waste material Substances 0.000 description 2
- 238000010146 3D printing Methods 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 230000001934 delay Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 230000000877 morphologic effect Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 229910052755 nonmetal Inorganic materials 0.000 description 1
- 238000007639 printing Methods 0.000 description 1
- 238000003908 quality control method Methods 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B29—WORKING OF PLASTICS; WORKING OF SUBSTANCES IN A PLASTIC STATE IN GENERAL
- B29C—SHAPING OR JOINING OF PLASTICS; SHAPING OF MATERIAL IN A PLASTIC STATE, NOT OTHERWISE PROVIDED FOR; AFTER-TREATMENT OF THE SHAPED PRODUCTS, e.g. REPAIRING
- B29C64/00—Additive manufacturing, i.e. manufacturing of three-dimensional [3D] objects by additive deposition, additive agglomeration or additive layering, e.g. by 3D printing, stereolithography or selective laser sintering
- B29C64/30—Auxiliary operations or equipment
- B29C64/386—Data acquisition or data processing for additive manufacturing
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B29—WORKING OF PLASTICS; WORKING OF SUBSTANCES IN A PLASTIC STATE IN GENERAL
- B29C—SHAPING OR JOINING OF PLASTICS; SHAPING OF MATERIAL IN A PLASTIC STATE, NOT OTHERWISE PROVIDED FOR; AFTER-TREATMENT OF THE SHAPED PRODUCTS, e.g. REPAIRING
- B29C64/00—Additive manufacturing, i.e. manufacturing of three-dimensional [3D] objects by additive deposition, additive agglomeration or additive layering, e.g. by 3D printing, stereolithography or selective laser sintering
- B29C64/30—Auxiliary operations or equipment
- B29C64/386—Data acquisition or data processing for additive manufacturing
- B29C64/393—Data acquisition or data processing for additive manufacturing for controlling or regulating additive manufacturing processes
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B33—ADDITIVE MANUFACTURING TECHNOLOGY
- B33Y—ADDITIVE MANUFACTURING, i.e. MANUFACTURING OF THREE-DIMENSIONAL [3-D] OBJECTS BY ADDITIVE DEPOSITION, ADDITIVE AGGLOMERATION OR ADDITIVE LAYERING, e.g. BY 3-D PRINTING, STEREOLITHOGRAPHY OR SELECTIVE LASER SINTERING
- B33Y50/00—Data acquisition or data processing for additive manufacturing
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B33—ADDITIVE MANUFACTURING TECHNOLOGY
- B33Y—ADDITIVE MANUFACTURING, i.e. MANUFACTURING OF THREE-DIMENSIONAL [3-D] OBJECTS BY ADDITIVE DEPOSITION, ADDITIVE AGGLOMERATION OR ADDITIVE LAYERING, e.g. BY 3-D PRINTING, STEREOLITHOGRAPHY OR SELECTIVE LASER SINTERING
- B33Y50/00—Data acquisition or data processing for additive manufacturing
- B33Y50/02—Data acquisition or data processing for additive manufacturing for controlling or regulating additive manufacturing processes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Materials Engineering (AREA)
- Chemical & Material Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Manufacturing & Machinery (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Mechanical Engineering (AREA)
- Optics & Photonics (AREA)
- Probability & Statistics with Applications (AREA)
- Image Analysis (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a predictive control method and a predictive control system for additive manufacturing forming states, which belong to the field of additive manufacturing control, and comprise the following steps: collecting video data of an additive manufacturing forming area and inputting the video data into a forecasting model frame by frame; performing feature extraction on each frame of image input into the prediction model by using an encoder, and outputting the image, and performing sequence prediction on the output of the encoder by using a long-short term memory network module to obtain prediction feature information in the next time period; classifying the predicted characteristic information by using a classifier to obtain a first predicted forming state label, reconstructing the predicted characteristic information by using a decoder to obtain a molten pool image in the next time period, and performing image processing on the molten pool image to obtain a second predicted forming state label; and fusing the first forecast forming state label and the second forecast forming state label to obtain a final forecast forming state label so as to control the additive manufacturing strategy in the next time period. Various quality problems in additive manufacturing are accurately predicted, and accurate control is achieved.
Description
Technical Field
The invention belongs to the field of additive manufacturing control, and particularly relates to a predictive control method and system for additive manufacturing forming states.
Background
Additive manufacturing is a technology for planning the stacking of raw materials to obtain a solid part, and is mainly based on the existing three-dimensional model to slice and stack, and the printing process is completed layer by layer, so the technology is also called 3D printing. In recent years, the additive manufacturing attracts more and more researchers' attention due to the great advantages of low cost and high efficiency, and various additive manufacturing technologies are generated and gradually expanded from non-metal parts to metal parts. However, most of the additive manufacturing techniques are still in the research stage, and the parts directly formed by the additive manufacturing techniques are easy to have internal defects or morphological defects such as cracks, air holes, incomplete fusion, abnormal accumulation and the like, so that the yield is low, and the parts cannot be manufactured efficiently at low cost. Therefore, there is a need for a rapid diagnosis of the quality of the forming and a corresponding control in the additive manufacturing of the formed part, so as to ensure a defect-free manufacturing of the part.
The existing quality control method for the additive manufacturing part generally processes images, and continuous video data in a forming process is not fully utilized, so that deviation of forming prediction is caused. In addition, the output of the existing method is mostly quality grade when the forming quality is diagnosed, whether the forming quality has a problem can be roughly judged, the type of the quality problem cannot be accurately described, subsequent process adjustment is inconvenient, and high-quality additive manufacturing is not facilitated.
Disclosure of Invention
Aiming at the defects and improvement requirements of the prior art, the invention provides a predictive control method and a predictive control system for additive manufacturing forming states, and aims to accurately predict various quality problems in additive manufacturing and realize accurate control.
To achieve the above object, according to an aspect of the present invention, there is provided an additive manufacturing forming state prediction control method including: s1, collecting video data of the additive manufacturing forming area and inputting the video data into a forecasting model frame by frame, wherein the forecasting model comprises an encoder, a long-term and short-term memory network module, a classifier and a decoder; s2, extracting the features of each frame of image input into the forecasting model by the encoder, outputting the extracted features, and performing sequence prediction on the output of the encoder by the long-short term memory network module to obtain prediction feature information in the next time period; s3, classifying the predicted characteristic information by the classifier to obtain a first predicted forming state label, reconstructing the predicted characteristic information by the decoder to obtain a molten pool image in the next time period, and performing image processing on the molten pool image to obtain a second predicted forming state label; and S4, fusing the first forecast forming state label and the second forecast forming state label to obtain a final forecast forming state label, and controlling an additive manufacturing strategy in the next time period according to the final forecast forming state label.
Furthermore, the first and second predicted forming state labels each include a plurality of sub-labels and the occurrence probability of each sub-label, and the plurality of sub-labels include normal, wrong path, unstable welding machine process, wire shaking, unsmooth wire feeding, bead flowing, excessive slag and shutdown.
Further, the fusing the first forecast forming state label and the second forecast forming state label in S4 includes: and respectively fusing the same sub-labels in the first forecast forming state label and the second forecast forming state label, and setting the sub-label with the highest occurrence probability after fusion as the final forecast forming state label.
Further, the fusion in S4 is an average fusion, a maximum fusion, a weight fusion or a generalized average fusion.
Further, S1 is preceded by: performing additive manufacturing experiments for multiple times by adopting different process parameters to obtain various defects in various forming states, and collecting video data of additive manufacturing forming areas in the experiment process; sampling the video data to obtain an image set, dividing the image set into a plurality of groups of subsets by taking a preset length N and a preset step length as a sliding window, and labeling the later x images in each subset to obtain a labeled data set, wherein x is less than N/3; and training the forecasting model by taking the first N-x images in each subset as input and the labeled data set as output.
Still further, the training the predictive model includes: and training the forecasting model by using Adam optimization algorithm and taking the minimum loss function of the forecasting model as a training target, wherein the loss function adopts average absolute error.
Further, in S2, the encoder performs feature extraction on each frame of image input into the prediction model through a single-layer convolutional neural network, a multi-layer convolutional neural network, a single-layer cyclic neural network, or a multi-layer cyclic neural network, and outputs the extracted features.
Further, the decoder reconstructs the predicted feature information through a reverse network corresponding to the encoder in S3.
According to another aspect of the invention, there is provided an additive manufacturing forming state prediction control system comprising: the system comprises an acquisition module, a prediction model and a prediction module, wherein the acquisition module is used for acquiring video data of an additive manufacturing forming area and inputting the video data into the prediction model frame by frame, and the prediction model comprises an encoder, a long-term and short-term memory network module, a classifier and a decoder; the encoding module is used for performing feature extraction on each frame of image input into the forecasting model by using the encoder and outputting the frame of image; the prediction module is used for performing sequence prediction on the output of the encoder by utilizing the long-short term memory network module to obtain prediction characteristic information in the next time period; the classification module is used for classifying the prediction characteristic information by using the classifier to obtain a first forecast forming state label; the decoding module is used for reconstructing the predicted characteristic information by using the decoder to obtain a molten pool image in the next time period; the image processing module is used for carrying out image processing on the molten pool image to obtain a second forecast forming state label; the fusion module is used for fusing the first forecast forming state label and the second forecast forming state label to obtain a final forecast forming state label; and the control module is used for controlling an additive manufacturing strategy in the next time interval according to the final forecast forming state label.
Generally, by the above technical solution conceived by the present invention, the following beneficial effects can be obtained: the method provides a new model suitable for prediction of the additive manufacturing forming state, utilizes a long-term and short-term memory network module to carry out sequence prediction on the output of an encoder, and improves the reliability of subsequent classification and reconstructed images, thereby improving the accuracy of prediction control; the long-term and short-term memory network is used for sequence prediction, and the other classifier is independently used, and the classifier can set the output label categories at will, so that the label categories output by the classifier can be set according to all possible situation types in the additive manufacturing forming process, various quality problems in additive manufacturing can be accurately predicted, and accurate control can be realized; in addition, video data in the additive manufacturing forming process is fully utilized, the forming state of additive manufacturing can be predicted more accurately based on continuous sampling images, and the accuracy of prediction control is further improved.
Drawings
Fig. 1 is a flowchart of a predictive control method for an additive manufacturing forming state according to an embodiment of the present invention;
fig. 2 is a block diagram of an additive manufacturing forming state prediction control system according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
In the present application, the terms "first," "second," and the like (if any) in the description and the drawings are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order.
Fig. 1 is a flowchart of a predictive control method for an additive manufacturing forming state according to an embodiment of the present invention. Referring to fig. 1, the additive manufacturing forming state prediction control method of the embodiment includes operations S1-S4.
In this embodiment, a neural network model of an encoder-long and short term memory-classifier-decoder structure is built as a prediction model by using an encoder-decoder neural network and a long and short term memory neural network as a backbone. The input of the forecasting model is multi-frame continuous image data; the output of the forecasting model comprises two parts, namely the forming state labels and the probabilities output by the classifier and the images of the forecasting forming areas output by the encoder.
Before operation S1 is performed, the constructed forecasting model needs to be trained, specifically including operation S0 '-operation S0'.
In operation S0', additive manufacturing experiments are performed multiple times using different process parameters to obtain various defects in each forming state, and video data of additive manufacturing forming regions are collected during the experiments.
And operation S0', sampling the video data to obtain an image set, dividing the image set into a plurality of groups of subsets by taking a preset length N and a preset step length as a sliding window, and labeling the last x images in each subset to obtain a labeled data set, wherein x is less than N/3.
Specifically, for example, a certain time duration between 10ms and 200ms is selected as a sampling time to sample video data, an ordered image set is obtained, all images in the image set are subjected to size transformation, and are uniformly converted into an input size of an encoder, and then segmentation and labeling operations are performed. And (4) marking, for example, manually marking according to the forming state of the last x images in each subset to obtain a marked data set, wherein the manually marked label marks comprise normal marks, path errors, unstable welding machine process, welding wire shaking, unsmooth wire feeding, bead flowing, excessive slag, machine halt and the like.
Operation S0' "trains the predictive model with the first N-x images in each subset as input and the annotated dataset as output.
Further, the subsets obtained in operation S0 ″ are divided into training sets and testing sets, and the dividing method includes, but is not limited to, leave method, cross-validation method, self-service method, and the like. In this embodiment, the subsets obtained in operation S0 ″ may be further expanded, and the expansion manner includes, but is not limited to, random cropping, horizontal flipping, vertical flipping, mirroring, rotating by different angles, adjusting brightness of the picture, adjusting contrast of the picture, adjusting chromaticity, changing proportions of RGB color components, adjusting saturation of the image, performing gaussian blur on the image, sharpening, adding noise, converting into a grayscale image, and the like. The expansion mode can be one or more.
Preferably, the prediction model is trained by using Adam optimization algorithm and taking the loss function minimum of the prediction model as a training target, and the loss function adopts average absolute error. After the training is finished, the structure and the weight of the forecasting model are frozen and stored as an externally-callable mode.
Operation S1, video data of the additive manufacturing forming area is collected and input frame by frame to a predictive model, which includes an encoder, a long-short term memory network module, a classifier, and a decoder.
In the actual additive manufacturing control process, the additive manufacturing of the part is carried out according to the initial process parameters, and the set industrial camera can be used for collecting video data of a forming area and the vicinity of the forming area in the additive manufacturing process in real time and inputting the video data into a trained forecasting model frame by frame.
In operation S2, the encoder performs feature extraction on each frame of image input into the prediction model, and outputs the frame of image, and the long-term and short-term memory network module performs sequence prediction on the output of the encoder to obtain predicted feature information in the next time period.
And the encoder performs feature extraction on each frame of image input into the forecasting model through a single-layer convolutional neural network, a multilayer convolutional neural network, a single-layer cyclic neural network or a multilayer cyclic neural network and outputs the frame of image after the feature extraction.
The Long Short-Term Memory network (LSTM) is a time-cycle neural network, has the characteristic of time Memory, and is suitable for processing and predicting important events with very Long intervals and delays in a time sequence. The LSTM module comprises a single-layer or multi-layer LSTM and is used for realizing the forecasting of the time sequence characteristic information. In the embodiment, the long-term and short-term memory network module is used for carrying out sequence prediction on the output of the encoder to obtain the prediction characteristic information in the next time period, so that the reliability of subsequent classification and image reconstruction is improved, and the accuracy of prediction control is improved.
And operation S3, classifying the predicted feature information by using the classifier to obtain a first predicted forming state label, reconstructing the predicted feature information by using the decoder to obtain a molten pool image in the next time period, and performing image processing on the molten pool image to obtain a second predicted forming state label.
In the embodiment, the long-term and short-term memory network is used as sequence prediction, the other classifier is independently used, and the output label categories can be set at will, so that the label categories output by the classifier can be set according to all possible situation types in the additive manufacturing forming process, various quality problems in additive manufacturing can be accurately predicted, and accurate control is realized.
According to actual requirements, in this embodiment, the label categories output by the classifier include normal, wrong path, unstable welding machine process, welding wire shaking, unsmooth wire feeding, bead flowing, excessive slag and shutdown, and the classifier also outputs occurrence probabilities of the label categories, and the sum of the occurrence probabilities of the label categories is 1.
Based on this, in this embodiment, each of the first and second forecasted forming state tags includes a plurality of sub-tags and an occurrence probability of each sub-tag. The plurality of sub-labels include normal, misrouted, unstable welder process, wire chatter, unsmooth wire feed, bead flow, excessive slag, and downtime.
The decoder reconstructs the predicted feature information through a reverse network corresponding to the encoder. The decoder reconstructs the predicted characteristic information output by the long-term and short-term memory network module through a reverse network corresponding to the single-layer convolutional neural network, the multilayer convolutional neural network, the single-layer cyclic neural network or the multilayer cyclic neural network to obtain a molten pool image in the next time period.
And carrying out image classification processing on the molten pool image by adopting a machine learning algorithm to obtain a second forecast forming state label. Preferably, a k-Nearest Neighbor classification algorithm is selected, and a corresponding forming state probability vector is calculated according to the number of the adjacent label tickets.
And operation S4, fusing the first and second predicted forming state labels to obtain a final predicted forming state label, and controlling an additive manufacturing strategy in a next time interval according to the final predicted forming state label.
And performing fusion judgment on the first forecast forming state label and the second forecast forming state label by using a decision fusion rule, respectively fusing the same sub-labels in the first forecast forming state label and the second forecast forming state label, and setting the sub-label with the highest occurrence probability after fusion as the final forecast forming state label. Specifically, average fusion, maximum fusion, weight fusion or generalized average fusion is performed on the first forecast forming state label and the second forecast forming state label to obtain a final forecast forming state label. Preferably, the decision fusion rule adopts an average fusion plus maximum support rule, the probability vector U1 in the first forecast forming state label and the probability vector U2 in the second forecast forming state label are subjected to fusion calculation to obtain a probability vector U3, U3 is (U1+ U2)/2, and the sub label corresponding to the maximum probability value in U3 is selected as the final forecast forming state label according to the maximum support rule.
The additive manufacturing control strategy for the next time period is, for example, a process parameter modification, a path modification, issuing an alarm or shutdown, etc. And if the final forecast forming state label is normal, keeping the original strategy of the additive manufacturing control strategy in the next time interval unchanged so as to continuously work according to the established process parameters and the process path. And if the final forecast forming state label is a path error, modifying the additive manufacturing control strategy in the next time interval to correspondingly adjust the process path in the additive manufacturing device, and continuing the subsequent manufacturing of the additive manufacturing device after the adjustment is finished. And if the final forecast forming state label is shutdown, immediately stopping the additive manufacturing device in the additive manufacturing control strategy in the next time period so as to avoid material waste, part failure and serious defects. If the final forecast forming state label is that the welding machine is unstable in process, the welding wire shakes, the wire feeding is not smooth, the welding bead flows or the molten slag is too much, the process parameters are timely adjusted according to a specific defect form, and an additive manufacturing control strategy in the next time period is formed. Therefore, the method can forecast in advance when the defects do not appear, respond quickly, adjust process parameters or stop the machine in time, improve the stability and the part quality in the additive manufacturing process and reduce the appearance of waste products.
Fig. 2 is a block diagram of an additive manufacturing forming state prediction control system according to an embodiment of the present invention. Referring to fig. 2, the additive manufacturing forming state prediction control system includes an acquisition module, an encoding module, a prediction module, a classification module, a decoding module, an image processing module, a fusion module, and a control module.
The acquisition module is used for acquiring video data of the additive manufacturing forming area and inputting the video data into the forecasting model frame by frame, and the forecasting model comprises an encoder, a long-term and short-term memory network module, a classifier and a decoder. And the coding module is used for performing feature extraction on each frame of image input into the forecasting model by using the coder and outputting the frame of image. And the prediction module is used for performing sequence prediction on the output of the encoder by using the long-short term memory network module to obtain prediction characteristic information in the next time period. The classification module is used for classifying the prediction characteristic information by using the classifier to obtain a first forecast forming state label. And the decoding module is used for reconstructing the predicted characteristic information by using a decoder to obtain a molten pool image in the next time period. And the image processing module is used for carrying out image processing on the molten pool image to obtain a second forecast forming state label. The fusion module is used for fusing the first forecast forming state label and the second forecast forming state label to obtain a final forecast forming state label. And the control module is used for controlling the additive manufacturing strategy in the next time period according to the final forecast forming state label.
The additive manufacturing forming state prediction control system is used for executing the additive manufacturing forming state prediction control method in the embodiment shown in fig. 1. For details that are not described in the present embodiment, please refer to the additive manufacturing forming state prediction control method in the embodiment shown in fig. 1, which is not described herein again.
It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and that any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (9)
1. An additive manufacturing forming state prediction control method, comprising:
s1, collecting video data of the additive manufacturing forming area and inputting the video data into a forecasting model frame by frame, wherein the forecasting model comprises an encoder, a long-term and short-term memory network module, a classifier and a decoder;
s2, extracting the features of each frame of image input into the forecasting model by the encoder, outputting the extracted features, and performing sequence prediction on the output of the encoder by the long-short term memory network module to obtain prediction feature information in the next time period;
s3, classifying the predicted characteristic information by the classifier to obtain a first predicted forming state label, reconstructing the predicted characteristic information by the decoder to obtain a molten pool image in the next time period, and performing image processing on the molten pool image to obtain a second predicted forming state label;
and S4, fusing the first forecast forming state label and the second forecast forming state label to obtain a final forecast forming state label, and controlling an additive manufacturing strategy in the next time period according to the final forecast forming state label.
2. The additive manufacturing forming state predictive control method of claim 1, wherein the first and second predicted forming state labels each include a plurality of sub-labels and a probability of occurrence of each of the sub-labels, the plurality of sub-labels including normal, misrouted, unstable welder process, wire chatter, poor wire feed, bead flow, excessive slag, and downtime.
3. The additive manufacturing forming state prediction control method of claim 2, wherein fusing the first predicted forming state label and the second predicted forming state label in S4 comprises: and respectively fusing the same sub-labels in the first forecast forming state label and the second forecast forming state label, and setting the sub-label with the highest occurrence probability after fusion as the final forecast forming state label.
4. The additive manufacturing forming state predictive control method of any one of claims 1-3, wherein the fusing in S4 is an average fusing, a maximum fusing, a weight fusing, or a generalized average fusing.
5. The additive manufacturing forming state prediction control method of claim 1, wherein the S1 is preceded by:
performing additive manufacturing experiments for multiple times by adopting different process parameters to obtain various defects in various forming states, and collecting video data of additive manufacturing forming areas in the experiment process;
sampling the video data to obtain an image set, dividing the image set into a plurality of groups of subsets by taking a preset length N and a preset step length as a sliding window, and labeling the later x images in each subset to obtain a labeled data set, wherein x is less than N/3;
and training the forecasting model by taking the first N-x images in each subset as input and the labeled data set as output.
6. The additive manufacturing forming state prediction control method of claim 5, wherein the training the predictive model comprises: and training the forecasting model by using Adam optimization algorithm and taking the minimum loss function of the forecasting model as a training target, wherein the loss function adopts average absolute error.
7. The additive manufacturing forming state prediction control method according to claim 1, wherein the encoder performs feature extraction on each frame image input into the prediction model by a single-layer convolutional neural network, a multi-layer convolutional neural network, a single-layer cyclic neural network, or a multi-layer cyclic neural network, and outputs the feature extracted image at S2.
8. The additive manufacturing forming state prediction control method according to claim 1 or 7, wherein the decoder reconstructs the prediction feature information through a reverse network corresponding to the encoder in S3.
9. An additive manufacturing forming state predictive control system, comprising:
the system comprises an acquisition module, a prediction model and a prediction module, wherein the acquisition module is used for acquiring video data of an additive manufacturing forming area and inputting the video data into the prediction model frame by frame, and the prediction model comprises an encoder, a long-term and short-term memory network module, a classifier and a decoder;
the encoding module is used for performing feature extraction on each frame of image input into the forecasting model by using the encoder and outputting the frame of image;
the prediction module is used for performing sequence prediction on the output of the encoder by utilizing the long-short term memory network module to obtain prediction characteristic information in the next time period;
the classification module is used for classifying the prediction characteristic information by using the classifier to obtain a first forecast forming state label;
the decoding module is used for reconstructing the predicted characteristic information by using the decoder to obtain a molten pool image in the next time period;
the image processing module is used for carrying out image processing on the molten pool image to obtain a second forecast forming state label;
the fusion module is used for fusing the first forecast forming state label and the second forecast forming state label to obtain a final forecast forming state label;
and the control module is used for controlling an additive manufacturing strategy in the next time interval according to the final forecast forming state label.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110973673.5A CN113762366B (en) | 2021-08-24 | 2021-08-24 | Method and system for predictive control of forming state of additive manufacturing |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110973673.5A CN113762366B (en) | 2021-08-24 | 2021-08-24 | Method and system for predictive control of forming state of additive manufacturing |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113762366A true CN113762366A (en) | 2021-12-07 |
CN113762366B CN113762366B (en) | 2024-02-06 |
Family
ID=78790962
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110973673.5A Active CN113762366B (en) | 2021-08-24 | 2021-08-24 | Method and system for predictive control of forming state of additive manufacturing |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113762366B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114619119A (en) * | 2022-03-29 | 2022-06-14 | 北京理工大学 | Electric arc vibration material disk makes on-line monitoring and real-time control system |
CN117984006A (en) * | 2024-04-03 | 2024-05-07 | 国网山东省电力公司潍坊供电公司 | Welding quality prediction method, device and medium based on welding infrared video generation |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20170144378A1 (en) * | 2015-11-25 | 2017-05-25 | Lawrence Livermore National Security, Llc | Rapid closed-loop control based on machine learning |
EP3579148A1 (en) * | 2018-06-08 | 2019-12-11 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Method and apparatus for training, classification model, mobile terminal, and readable storage medium |
CN112801091A (en) * | 2021-01-26 | 2021-05-14 | 汕头大学 | Additive manufacturing part forming quality monitoring and predicting method based on deep learning |
CN112989703A (en) * | 2021-04-02 | 2021-06-18 | 唐剑军 | Feature fusion metal additive manufacturing fusion state real-time prediction method based on deep learning |
-
2021
- 2021-08-24 CN CN202110973673.5A patent/CN113762366B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20170144378A1 (en) * | 2015-11-25 | 2017-05-25 | Lawrence Livermore National Security, Llc | Rapid closed-loop control based on machine learning |
EP3579148A1 (en) * | 2018-06-08 | 2019-12-11 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Method and apparatus for training, classification model, mobile terminal, and readable storage medium |
CN112801091A (en) * | 2021-01-26 | 2021-05-14 | 汕头大学 | Additive manufacturing part forming quality monitoring and predicting method based on deep learning |
CN112989703A (en) * | 2021-04-02 | 2021-06-18 | 唐剑军 | Feature fusion metal additive manufacturing fusion state real-time prediction method based on deep learning |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114619119A (en) * | 2022-03-29 | 2022-06-14 | 北京理工大学 | Electric arc vibration material disk makes on-line monitoring and real-time control system |
CN114619119B (en) * | 2022-03-29 | 2023-01-13 | 北京理工大学 | Electric arc vibration material disk makes on-line monitoring and real-time control system |
CN117984006A (en) * | 2024-04-03 | 2024-05-07 | 国网山东省电力公司潍坊供电公司 | Welding quality prediction method, device and medium based on welding infrared video generation |
Also Published As
Publication number | Publication date |
---|---|
CN113762366B (en) | 2024-02-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113762366A (en) | Additive manufacturing forming state prediction control method and system | |
KR102631031B1 (en) | Method for detecting defects in semiconductor device | |
CN112381788B (en) | Part surface defect increment detection method based on double-branch matching network | |
Valizadeh et al. | Convolutional Neural Network applications in additive manufacturing: A review | |
CN110648310B (en) | Weak supervision casting defect identification method based on attention mechanism | |
KR102265459B1 (en) | Predictive maintenance system for defects in plant facilities | |
CN110245227B (en) | Training method and device for text classification fusion classifier | |
KR20230147636A (en) | Manufacturing quality control system and method using automated visual inspection | |
CN111833313B (en) | Industrial product surface defect detection method and system based on deep active learning | |
CN114332008B (en) | Unsupervised defect detection and positioning method based on multi-level feature reconstruction | |
CN112801091A (en) | Additive manufacturing part forming quality monitoring and predicting method based on deep learning | |
CN114529894A (en) | Rapid scene text detection method fusing hole convolution | |
JP2021033564A (en) | Image classification method, image classification device and image classification program | |
CN117540779A (en) | Lightweight metal surface defect detection method based on double-source knowledge distillation | |
CN111753653A (en) | High-speed rail contact net fastener identification and positioning method based on attention mechanism | |
CN114743099A (en) | Method and device for identifying faults of train key components | |
CN114529821A (en) | Offshore wind power safety monitoring and early warning method based on machine vision | |
CN110378916A (en) | A kind of TBM image based on multitask deep learning is slagged tap dividing method | |
CN115082923B (en) | Milk packing box production date identification method based on machine vision | |
CN115797326A (en) | Metal casting surface defect detection method based on improved YOLOX-S algorithm | |
CN116152606A (en) | Sample generation method, target detection model training, target detection method and system | |
CN114037960B (en) | Method and system for identifying state of flap valve based on machine vision | |
Gong et al. | Research on surface defects detection method and system in manufacturing processes based on the fusion of multi-scale features and semantic segmentation for intelligent manufacturing | |
CN114387211A (en) | Training method of vaccine defect detection model, vaccine defect detection method and device | |
Zhang et al. | A Transformer-Based Approach for Metal 3d Printing Quality Recognition |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |