Predictive maintenance aims to find out when the target device (TD) is in the sick state and almost entering the dead state before its actual occurrence to conduct just-in-time maintenance, so as to avoid unexpected TD down time. In this way, not only tool availability and manufacturing quality are improved, but the additional cost of excessive maintenance in preventive maintenance strategy can also be reduced. Among the predictive maintenance technologies proposed by many scholars, exponential model was commonly applied to predict the remaining useful life (RUL) of TD. However, due to the algorithm limitations, when TD is about to die, whether the TD's aging feature suddenly rises or becomes smooth, the exponential model may not be able to keep up with the real-time prediction or even falsely predicts long RUL. To solve the problem of inaccurate RUL prediction, the authors propose the time series prediction (TSP) algorithm. TSP applies the time series analysis model built by information criterion to adapt to the complicated future trend of solving TD fault prediction. Also, the Pre-Alarm Module (PreAM) to make alert of immediate maintenance when a TD is likely to shut down shortly as well as the Death Correlation Index (DCI) to reveal the possibility of entering the dead state are proposed in this work. How to select the most effective predictors and adjust the predictor weights to construct high-performance prediction model are also illustrated in this letter with the tools in various industries (such as solar-cell manufacturing and machine tool industry) being the examples of the TSP algorithm.

The capability of measuring melt pool variation is the key evaluating metal additive manufacturing quality. To measure the variation, a metrology architecture with in-situ melt pool measurement and an estimation module is required. However, it is a challenge to effectively extract significant features from the huge data collected by the in-situ metrology for quality estimation requirement. The purpose of this paper is to propose an intelligent metrology architecture with an in-situ metrology (ISM) module and an enhanced automatic virtual metrology (AVM) system. The ISM module can extract the melt pool features with a coaxial camera and a pyrometer. On the other hand, the AVM system is improved with a feature selection method to solve the issue of limited samples in the component modeling quality. The examples with different metals are adopted to illustrate how the system works for estimating surface roughness and density of components, and, in the future, the system can even serve as the feedback signal for adaptive control of the process parameters by layering in an additive manufacturing (AM) system.

Quality control is the key for the widespread adoption of metal additive manufacturing (AM). However, online quality estimation is challenging because high-frequency stream data derived from in situ metrology have to be processed in a timely manner to figure out the complicated interactions among material, machine, and part. To tackle such issue, this article proposes an intelligent AM metrology (IAMM) architecture to decouple and evaluate quality variations caused by material properties, machine issues, and process parameters when building an AM part. The IAMM architecture can also estimate the online layer-to-layer quality (e.g., roughness and density of an AM part) by applying the robust parameters derived from the uniform design (UD) method via the enhanced automatic virtual metrology technology in a parallel computing environment, as soon as the microfeatures of melt-pools and the macrofeatures of each layer are extracted. In addition, the associated indices and features in the IAMM architecture can also be used to evaluate the defects caused by machine issues in the given process parameters. The results of case studies validate the applicability of the IAMM architecture and show that the proposed estimation models are prospective for future closed-loop control in an AM process.

The abilities to both monitor and control additive manufacturing (AM) processes in real-time are necessary before the routine production of quality AM parts will be possible. Currently, neither ability exist! The major reason is that AM processes are different from traditional manufacturing processes in many ways and so are the sensors and the monitoring data collected from them. In traditional manufacturing, that data is mostly numeric in nature. To that numeric data, AM monitoring data add large volumes of a variety of in situ, high-speed, image data. Collecting, fusing, and analyzing all that AM data and making the necessary control decisions is not possible using traditional, rigid, hierarchical-control architectures. Therefore, researchers are proposing to use real-time, machine-learning algorithms to analyze the data and to execute the other control functions. This paper identifies those control functions and proposes a new architecture to integrate them. This paper also shows an example of using that architecture to analyze the melt-pool, shape analysis using a clustering method.

Determining thresholds of the primary control loops (System 1) of an additive manufacturing (AM) process is challenging when realizing System 1 with its fast and intuitive capability for adapting to different metal powers, machine configurations, and process parameters. Based on the convolution neural network and long short-term memory models, this letter presents a secondary tuning loop (System 2) to classify the types of melt-pool images (MPIs) from a coaxial camera online, suggest polishing parameters, and determine the control thresholds of System 1 offline. Case studies indicate that the thresholds and parameters of System 1 including smoke discharging, powder coating, and laser polishing of control loops of a laser powder bed fusion (LPBF) machine can be more deliberatively and logically decided by the proposed MPI-based System 2.

For the current wafer sawing process, the wafers in the same lot are inspected at the end of the entire process. Therefore, a defect, such as chipping, occurs during processing will only be detected until the end of the process, which is too late and may cause massive defects. If Automatic Virtual Metrology (AVM) is implemented in the wafer sawing process, when chippings occur and are detected, its chipping amount can be predicted by AVM on-line and in real time. Also, AVM's individual similarity index (ISI) analysis can be applied to identify the root cause of chipping. As a result, this root cause can be fixed to avoid generating defects in the subsequent wafers. However, chipping won't happen to all wafers. Since the AVM system deals mainly with the regression problem, it cannot classify whether a wafer is chipped or not. Hence, there is a need to predict wafer-chipping occurrence before applying AVM to the wafer sawing process. To solve the above mentioned problem, the wafer sawing quality monitoring is divided into two stages. An Automated Classification Scheme (ACS) based on ensemble learning is developed in Stage I to pre-determine whether a wafer is chipped. If chipping is detected, then proceed to Stage II for the AVM system to predict the chipping amount and identify the root cause that results in this chipping. With the so-called ACS-plus-AVM scheme, the AVM application in the wafer sawing process can be realized.