Acoustic emission for in situ quality monitoring in additive manufacturing using spectral convolutional neural networks

TLDR: In this paper, the authors investigated the feasibility of using acoustic emission for quality monitoring and combined a sensitive acoustic emission sensor with machine learning, where the acoustic signals were recorded using a fiber Bragg grating sensor during the powder bed additive manufacturing process in a commercially available selective laser melting machine.

Abstract: Additive manufacturing, also known as 3D printing, is a new technology that obliterates the geometrical limits of the produced workpieces and promises low running costs as compared to traditional manufacturing methods. Hence, additive manufacturing technology has high expectations in industry. Unfortunately, the lack of a proper quality monitoring prohibits the penetration of this technology into an extensive practice. This work investigates the feasibility of using acoustic emission for quality monitoring and combines a sensitive acoustic emission sensor with machine learning. The acoustic signals were recorded using a fiber Bragg grating sensor during the powder bed additive manufacturing process in a commercially available selective laser melting machine. The process parameters were intentionally tuned to invoke different processing regimes that lead to the formation of different types and concentrations of pores (1.42 ± 0.85 %, 0.3 ± 0.18 % and 0.07 ± 0.02 %) inside the workpiece. According to this poor, medium and high part qualities were defined. The acoustic signals collected during processing were grouped accordingly and divided into two separate datasets; one for the training and one for the testing. The acoustic features were the relative energies of the narrow frequency bands of the wavelet packet transform, extracted from all the signals. The classifier, based on spectral convolutional neural network, was trained to differentiate the acoustic features of dissimilar quality. The confidence in classifications varies between 83 and 89 %. In view of the narrow range of porosity, the results can be considered as promising and they showed the feasibility of the quality monitoring using acoustic emission with the sub-layer spatial resolution.

Figures

Figure 2. a) SLM test workpiece produced with three energy densities where 50 J/mm3 are bright regions, 79

Figure 1: a) (left) view of the FBG location inside the AM chamber with the optical feedthrough on the chamber panel and (right) the FBG read out system and (b) scheme of the FBG read out system.

Frequently asked questions

Q1. What are the contributions mentioned in the paper "Acoustic emission for in situ quality monitoring in additive manufacturing using spectral convolutional neural networks" ?

This is a PDF fil of an unedited manuscript that has been accepted for publication. As a service to their customers the authors are providing this early version of the manuscript. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. 

Q2. What have the authors stated for future works in "Acoustic emission for in situ quality monitoring in additive manufacturing using spectral convolutional neural networks" ?

The possible transfer of the results presented to other technologies is the subject of the future investigations. 

Q3. What are the main techniques used in additive manufacturing?

at present, additive manufacturing processes incorporate several independently developing techniques, such as stereo-lithography, laser sintering, multi-jet printing, powder-bed fusion and others. 

Q4. What is the recent extent of conventional CNN?

Spectral convolutional neural networks (SCNN) are a recent extent of conventional CNN with improved efficiency in classification/regression tasks [27,28,29]. 

Q5. What is the advantage of using acoustic gratings for AE?

At present, Fiber Bragg gratings (FBG) are often used for detecting AE signals because of its high sensitivity in a wide acoustic spectral range [22]. 

Q6. What is the definition of the classification accuracies in the table?

Theclassification accuracies in the table are defined as the number of true positives divided by the total number of tests for each category. 

Q7. What was the connection to the read out system?

Its connection to the read out system was provided via an optical feedthrough mounted on a plate that hermetically closed the working chamber. 

Q8. Why were the signals not included in the training and tests data sets?

Those signals were used to evaluate the existence of changes in the AE content and noise parameters but they were not included into the training and tests data sets. 

Q9. What is the spectrogram of the AE signal produced?

An example of a fragment of an AE signal that corresponds to a high quality layerproduced with the optimal process condition that is with an energy density (79 J/mm3, 500 mm/s) and b) a spectrogram corresponding to the relative energies of the narrow frequency bands, localized in time-frequency domain. 

Q10. What is the classification error structure of the acoustic signal recorded by the FBG?

the authors can conclude that the acoustic signal recorded by the FBG and its processing with the SCNN could be a solution for in situ and real-time quality monitoring in AM. 

Q11. How many narrow frequency bands were extracted?

Each RW was decomposed by WPT at ten decomposition levels, resulting in the extraction of 2,046 narrow frequency bands with their corresponding values of the relative energies (See Section 3.1). 

Q12. How is the dimension of the feature space reduced?

The dimensions of the feature space are reduced via the principle component analysis (PCA) projection [41] into a lower 3D dimensional space.