Search

Many strategic industrial sectors prefer Directed Energy Deposition (DED) to other Additive Manufacturing (AM) technologies due to the high material deposition and build rates. However, the inadvertent formation of defects such as porosity, micro-cracks and microstructure anomalies hinders its adoption in industries that require specific mechanical and microstructural properties. These defects are caused by undesirable fluctuations in process conditions such as material flow rate, laser power, melt pool dynamics, environment gas composition, temperature gradients. This research proposes in situ quality monitoring of DED using images of process zone and contrastive learning-based Convolutional Neural Network (CNN). Experiments included deposition of titanium powder (Cp-Ti, grade 1) with the particle size ranging between 45 and 106 μm on the base plate (99.6 % Ti6Al4V grade 1), forming a cube geometry. The process parameters were tuned to achieve six quality grades. The video of the process zone was recorded co-axially to the laser beam during the entire manufacturing, which was eventually used as the input to train CNN's based on contrastive losses. An in situ monitoring strategy for classifying the different quality grades was demonstrated in a supervised and semi-supervised manner, with an accuracy ranging between 89 % and 97 %. The performance of the developed framework was compared to an alternative clustering technique, namely t-distributed stochastic neighbour embedding, justifying the efficiency of our approach. The developed methodology demonstrates the possibility to track workpiece manufacturing quality using simple CCD cameras with minimum interventions on the commercial machines.

Functional surfaces in relative contact and motion are prone to wear and tear, resulting in loss of efficiency and performance of the workpieces/machines. Wear occurs in the form of adhesion, abrasion, scuffing, galling, and scoring between contacts. However, the rate of the wear phenomenon depends primarily on the physical properties and the surrounding environment. Monitoring the integrity of surfaces by offline inspections leads to significant wasted machine time. A potential alternate option to offline inspection currently practiced in industries is the analysis of sensors signatures capable of capturing the wear state and correlating it with the wear phenomenon, followed by in situclassification using a state-of-the-art machine learning (ML) algorithm. Though this technique is better than offline inspection, it possesses inherent disadvantages for training the ML models. Ideally, supervised training of ML models requires the datasets considered for the classification to be of equal weightage to avoid biasing. The collection of such a dataset is very cumbersome and expensive in practice, as in real industrial applications, the malfunction period is minimal compared to normal operation. Furthermore, classification models would not classify new wear phenomena from the normal regimeif they are unfamiliar. As a promising alternative, in this work, we propose a methodology able to differentiate the abnormal regimes, i.e., wear phenomenon regimes, from the normal regime. This is carried out by familiarizing the ML algorithms only with the distribution of the acoustic emission (AE) signals captured using a microphone related to the normal regime. As a result, the ML algorithms would be able to detect whether some overlaps exist with the learnt distributions when a new, unseen signal arrives. To achieve this goal, a generative convolutional neural network (CNN) architecture based on variational auto encoder (VAE) is built and trained. During the validation procedure of the proposed CNN architectures, we were capable of identifying acoustics signals corresponding to the normal and abnormal wear regime with an accuracy of 97% and 80%. Hence, our approach shows very promising results for in situand real-time condition monitoring or even wear prediction in tribological applications.

Recent approaches in online monitoring of Laser Powder Bed Fusion (LPBF) processes rely on multimodal analysis of several sensors. The reported defect detection accuracies are generally very high for laboratory conditions but hardly generalize to real-life situations with arbitrary geometries. Indeed, under such circumstances, the acquired signals turn out to be very complex — with many uninformative portions due to non-constant laser emissions. This issue makes the introduction of a segmentation technique a requirement to eliminate the sections of the signals acquired when the laser is “off.” To close this gap, we present a novel segmentation algorithm based on optical emission data and mathematical morphology operations to identify precisely the portions of the signals recorded only during laser emissions, regardless of the adopted scanning strategy. Experimental results with a commercially available LPBF machine using different scanning strategies and geometries show an average Intersection-over-Union of 87% compared to 63% obtained by thresholding.

The present work demonstrates an adaptive closed-loop control for laser welding processes. Based on feedback signals from a sensing system, the controller interacts with the laser to adapt the processing parameters to achieve or maintain the target welding quality. The controller is constructed based on a model-free reinforcement learning approach, namely Q-learning. This algorithm allows autonomous learning of the control law independently from the starting conditions as well as any prior knowledge of the process dynamics. The controller's performance is demonstrated in both a well-controlled lab environment and more unpredictable industrial situations. For the demonstration, the control system is allowed to vary the laser power, and the feedback signal is given by an industrial laser process control unit (Coherent SmartSense+) using an optical sensor. The time needed to train the control system is approximately five and twenty minutes for the well-controlled and the industrial situations, respectively.

In this paper, the effects of the deposition speed and thermal cycles in gas-metal arc-welding (GMAW) additive manufacturing on the quality of as-built 308L stainless steel thin walls were investigated. The results exhibit that the deposition speed and thermal cycles play a crucial role in the quality of produced parts. An increase in deposition speed results in an improvement in the surface waviness. The surface waviness (Sa) decreases from 286 to 138 µm as the deposition speed increases from 0.2 to 0.4 m/min. On the other hand, the growth of microstructures in the walls fabricated with different deposition speeds shows a similar trend. The microstructure of as-built 308L-stainless-steel walls consists of dominant columnar/equiaxed dendrites of austenite and small amount of ferrite remaining in grain boundaries. The deposition speed mainly influences the grain size in microstructures. In the middle part of the walls, an augmentation in the deposition speed leads to a decrease in the secondary dendrite arm spacing, which results in an enhancement in mechanical properties of the walls. The microhardness and ultimate tensile strength increase from 153 ± 7.16 to 164 ± 8.96 HV0.1 and from 483 ± 4.24 to 518 ± 2.83 MPa, respectively, when the deposition speed increases from 0.2 to 0.4 m/min.

The existing knowledge regarding the interfacial forces, lubrication, and wear of bearings in real-world operation has significantly improved their designs over time, allowing for prolonged service life. As a result, self-lubricating bearings have become a viable alternative to traditional bearing designs in industrial machines. However, wear mechanisms are still inevitable and occur progressively in self-lubricating bearings, as characterized by the loss of the lubrication film and seizure. Therefore, monitoring the stages of the wear states in these components will help to impart the necessary countermeasures to reduce the machine maintenance downtime. This article proposes a methodology for using a long short-term memory (LSTM)-based encoder—decoder architecture on interfacial force signatures to detect abnormal regimes, aiming to provide early predictions of failure in self-lubricating sliding contacts even before they occur. Reciprocating sliding experiments were performed using a self-lubricating bronze bushing and steel shaft journal in a custom-built transversally oscillating tribometer setup. The force signatures corresponding to each cycle of the reciprocating sliding motion in the normal regime were used as inputs to train the encoder—decoder architecture, so as to reconstruct any new signal of the normal regime with the minimum error. With this semi-supervised training exercise, the force signatures corresponding to the abnormal regime could be differentiated from the normal regime, as their reconstruction errors would be very high. During the validation procedure for the proposed LSTM-based encoder—decoder model, the model predicted the force signals corresponding to the normal and abnormal regimes with an accuracy of 97%. In addition, a visualization of the reconstruction error across the entire force signature showed noticeable patterns in the reconstruction error when temporally decoded before the actual critical failure point, making it possible to be used for early predictions of failure.

Abrasive finishing processes such as grinding, lapping or disc polishing are one of the most practical means for processing materials to manufacture products with fine surface finish, surface quality and dimensional accuracy. However, they are one of the most difficult and least-understood processes for two main reasons. Firstly, the abrasive grains present in the tool surface are randomly oriented. Secondly, they undergo complex interactions in the machining zone. Given the advances in sensor technologies, the finishing processes can now be sensorized, and the vast amount of data produced can be exploited to model and monitor the processes using Artificial Intelligence techniques. Data-driven models have turned into a hot focus in engineering with the rise of machine learning and deep learning algorithms, which have greatly spread all through the academic community. The scope of this paper is mainly to review the application of Artificial Intelligence as well as supporting sensing and signal processing techniques in modelling and monitoring on different types of abrasive processes in metal finishing. The paper gives a detailed background on the key mechanisms and defects in the different abrasive finishing process and lists the suitable sensing techniques for their monitoring. The paper reports that most of the Artificial Intelligence algorithms available are not fully exploited for monitoring and modelling in abrasive finishing and emphasizes on bridging this gap. The probable research tendency on data-driven monitoring and modelling for abrasive finishing is also forecasted.

Sensor integration for in situ monitoring during additive manufacturing promises to enhance control over the process and assures quality in the fabricated workpieces. Acoustic emissions from the process zone of the laser powder-bed fusion process carry information about the events and failure modes of the printed workpiece. Analysis of acoustic signals emitted during different laser regimes, such as conduction, keyhole, etc. in time, frequency and time-frequency domains could provide quantitative information about the underlying physical mechanisms. This article reports a statistical analysis of the features in acoustic signals to perceive the characteristics of failure modes occurring during layering of stainless steel 316L. The visualization of the feature space distribution that corresponds to different failure modes shows the potentials of applying machine learning for in situ classification. The paper also proposes strategies in terms of data acquisition and preprocessing for building a comprehensive monitoring system.

During laser polishing of tool steel surfaces produced by electric discharge machining with NIR continuous wave laser, small craters were observed on the polished surface. Their formation was observed in situ with high-speed camera. Post mortem investigation of the composition of the craters with energy dispersive X-ray spectroscopy (SEM) has shown that they were caused by copper inclusions on the steel surface. The inclusions are probably coming from the electrode employed during the electric discharge machining process. The low absorption of copper in NIR signifies that the bigger copper particles do not melt at high speed which leads to the formation of the craters.

One of the remaining challenges in Laser Powder Bed Fusion (LPBF) of metals is the control of the formation of keyhole pores, resulting from a local excessive energy input during processing. Such defects can lead to degraded mechanical properties and are typically detected and/or removed after the process through non-destructive quality-inspection procedures and porosity-removal treatments. Monitoring and controlling the formation of defects during the LPBF process can allow circumventing such time-consuming and costly post-process stages. This paper develops a new approach to perform in-situ healing of deep keyhole pores, using a positively defocused laser beam with finely tuned laser remelting process parameters. Synchrotron radiographic images of the process zone are acquired during laser remelting. The use of operandoimaging enables the visualization of pore removal during processing, and unveils the effect of various remelting conditions on the healing efficiency. The acoustic signals generated during laser remelting are recorded using a high-sensitivity optical microphone, and analyzed in parallel with the X-ray images, allowing the acoustic signature of defect healing to be identified. The present paper demonstrates for the first time that an airborne acoustic sensor can be used to monitor the healing of keyhole pores during LPBF.

The Laser Power Bed Fusion (LPBF) process is of high interest to many industries, such as motors and vehicles, robotics, biomedical applications, aerospace, and others. LPBF workpieces can indeed achieve near full density and high resistance. However, a large amount of pore formation, in conjunction with the probabilistic nature of defect formation, results in a lack of process repeatability and reproducibility. This limits the range of industrial applications requiring high quality and defect free workpieces. To overcome this issue, we developed an acoustic monitoring system able to classify with high confidence three processing regimes (lack of fusion pores, conduction mode, keyhole pores)using a Convolution Neural Network (CNN).For the first time, we infer the processing regime based on AE waves produced during the LPBF process for conditions that are new and not part of the training database (>96%). The choice of processing conditions used in the database (training sets) is discussed in details, looking at the influence of their number, relative normalized distance, and position in the processing map on the classification accuracy. We found that the higher the number of processing conditions in the database, the higher the classification accuracies. Moreover, the higher the relative normalized “distance” between training and testing sets (measured in terms of laser speed and power), the lower the classification accuracies. Finally, the threshold defining the minimum number of training processing conditions is identified as eight to obtain a robust model able to identify the processing regimes for new laser parameters within the processing map. This number can be lowered to six if the training sets are in the surrounding region of the testing set. When one process parameter (speed, power, or normalized enthalpy) is constant between all the training and the testing sets, only four parameter sets allow a high classification accuracy (>88%). These results demonstrate the potential of in situ acoustic emission for monitoring the additive manufacturing process, in particular when the process conditions may deviate from the conduction mode. Finally, for a well-chosen set of training conditions, the model is able to construct a full processing map without additional experiments.

Harnessing the full potential of the metal-based Laser Powder Bed Fusion process (LPBF) relies heavily on how effectively the overall reliability and stability of the manufactured part can be ensured. To this aim, the recent advances in sensorization and processing of the associated signals using Machine Learning (ML) techniques have made in situ monitoring a viable alternative to post-mortem techniques such as X-ray tomography or ultrasounds for the assessment of parts. Indeed, the primary advantage of in situ monitoring over post-mortem analysis is that the process can be stopped in case of discrepancies, saving resources. Additionally, mitigations to repair the discrepancies can also be performed. However, the in situ monitoring strategies based on classifying processing regimes reported in the literature so far operate on signals of fixed length in time, constraining the generalization of the trained ML model by not allowing monitoring processes with heterogeneous laser scanning strategies. As a part of this work, we try to bridge this gap by developing a hybrid Deep Learning (DL) model by combining Convolutional Neural Networks (CNNs) with Long-Short Term Memory (LSTM) that can operate over variable time-scales. The proposed hybrid DL model was trained on signals obtained from a heterogeneous time-synced sensing system consisting of four sensors, namely back reflection (BR), Visible, Infra-Red (IR), and structure-borne Acoustic Emission (AE). The signals captured different phenomena related to the LPBF process zone and were used to classify three regimes: Lack of Fusion (LoF), conduction modeand Keyhole.Specifically, these three regimes were induced by printing cubes out of austenitic Stainless steel (316 L) on a mini-LPBF device with operandohigh-speed synchrotron X-ray imaging and signal acquisition with the developed heterogeneous sensing system. The operandoX-ray imaging analysis ensured that the regimes correlated with the defined process parameters. During the validation procedure of the trained hybrid DL model, the model predicted three regimes with an accuracy of about 98% across various time scales, ranging from 0.5 ms to 4 ms. In addition to tracking the model performance, a sensitivity analysis of the trained hybrid model was conducted, which showed that the BR and AE sensors carried more relevant information to guide the decision-making process than the other two sensors used in this work.

Abstract

Abstract

Abstract