Advanced Process Monitoring Research Articles

Advanced Process Monitoring Through Fault Detection and Classification for the Process Development of Tantalum Nitride Thin-Film Resistors

This paper discusses the optimization of an advanced process monitoring scheme with interdicting fault detection and classification (FDC) capabilities that improved the control over the process development of Tantalum Nitride thin film resistors (TaN TFR). Its implementation in a high-volume manufacturing environment resulted in a reduction of misprocessed wafers, shorter equipment downtime, higher throughput, enhanced process visibility, and yield improvement. Along with optimizing FDC capabilities, implementing a short loop sampling plan reduced the turnaround time for early electrical characterization by a factor of four; this allowed for timely inline adjustments within the fabrication process to tighten the statistical process control (SPC) over the distribution of the process control monitor (PCM) TaN resistor value and improve thin film uniformity.

In Situ Inclusion Detection and Material Characterization in an Electron Beam Powder Bed Fusion Process Using Electron Optical Imaging.

Electron Beam Powder Bed Fusion (PBF-EB) is an Additive Manufacturing (AM) method that utilizes an electron beam to melt and consolidate metal powder. The beam, combined with a backscattered electron detector, enables advanced process monitoring, a method termed Electron Optical Imaging (ELO). ELO is already known to provide great topographical information, but its capabilities regarding material contrast are less studied. In this article the extents of material contrast using ELO are investigated, focusing mainly on identifying powder contamination. It will be shown that an ELO detector is capable of distinguishing a single 100 μm foreign powder particle, during an PBF-EB process, if the backscattering coefficient of the inclusion is sufficiently higher than its surroundings. Additionally, it is investigated how the material contrast can be used for material characterization. A mathematical framework is provided to describe the relationship between the signal intensity in the detector and the effective atomic number Zeff of the imaged alloy. The approach is verified with empirical data from twelve different materials, demonstrating that the effective atomic number of an alloy can be predicted to within one atomic number from its ELO intensity.

Research on advanced manufacturing process monitoring and fault prediction method based on machine learning

Research on advanced manufacturing process monitoring and fault prediction method based on machine learning

Large Scale Microalgae Biofuel Technology—Development Perspectives in Light of the Barriers and Limitations

Microalgal biomass can be used to derive many different types of biofuels. In order to widely commercialize this technology, its limitations and weaknesses must be eliminated. Many technical and economic issues also need to be clarified and unknowns answered. Microalgae-based technologies have been shown to be versatile, efficient and suitable for practical and commercial use. However, the current technological readiness level (TRL) of most microalgae-based bioenergy production systems precludes their quick and widespread implementation. Their development is limited by a combination of factors that must be precisely identified, after which their negative impact on scale-up prospects can be eliminated or minimized. The present study identifies the main barriers to the development of industrial microalgae-production and microalgae-to-biofuel systems. In addition, it proposes measures and efforts necessary to achieve a higher TRL, which will provide investors with sought-after performance and cost-efficiency data for the given project. The main barriers to the development of microalgae cultivation and processing systems have been identified to include: the complex nature of the cultivation process with multiple variables involved; lack of sufficient data from pilot-scale and near-full-scale plants, which often precludes reliable life cycle assessment (LCA); and insufficient legal assistance, advisory assistance, subsidies and funding for innovative projects. Potential ways of improving performance and competitiveness of microalgae-based systems include: process optimization techniques, genetic engineering, yield improvement through advanced process control and monitoring, use of waste as feedstock and dedicated support programs. The presented summary of the current stage of microalgal biofuel production technology development indicates the directions for further research and implementation work, which are necessary for the final verification of the application potential of these solutions.

Zernike model for overlay control and tool monitor for lithography and etch process

In advanced lithography and etch processes, the after-etch overlay is rendered unequal to the after-development overlay, and an etch process–induced overlay plays an extremely important role in determining the total overlay. This makes it difficult to achieve feedback control and also throws the overlay residue out of the control specifications. As the layer number of three dimensional integrated circuits increases or the overlay residue is only several nanometers in extreme ultraviolet technology, differentiations in etch tools lead to lot-to-lot overlay differences. Tool difference should be considered to compensate the overlay, irrespective of which method, interfield, intrafield, or correction per exposure (CPE), is used. However, it is not recommended to compensate the overlay for every lot as it will increase the cost of metrology. One of the widely used methods is to separate the lots by groups. In this paper, we propose a Zernike-CPE method to conduct a systematic study on how to reduce the overlay residue during the process of providing CPE feedback and to monitor tool performance. The research results demonstrate that the proposed method has potential roles to play in achieving overlay control and is highly suitable for advanced technology process monitoring.

An interpretable unsupervised Bayesian network model for fault detection and diagnosis

Process monitoring is a critical activity in manufacturing industries. A wide variety of data-driven approaches have been developed and employed for fault detection and fault diagnosis. Analyzing the existing process monitoring schemes, prediction accuracy of the process status is usually the primary focus while the explanation (diagnosis) of a detected fault is relegated to a secondary role. In this paper, an interpretable unsupervised machine learning model based on Bayesian Networks (BN) is proposed to be the fundamental model supporting the process monitoring scheme. The proposed methodology is aligned with the recent efforts of eXplanatory Artificial Intelligence (XAI) for knowledge induction and decision making, now brought to the scope of advanced process monitoring. A BN is capable of combining data-driven induction with existing domain knowledge about the process and to display the underlying causal interactions of a process system in an easily interpretable graphical form. The proposed fault detection scheme consists of two levels of monitoring. In the first level, a global index is computed and monitored to detect any deviation from normal operation conditions. In the second level, two local indices are proposed to examine the fine structure of the fault, once it is signaled at the first level. These local indices support the diagnosis of the fault, and are based on the individual unconditional and conditional distributions of the monitored variables. A new labeling procedure is also proposed to narrow down the search and identify the fault type. Unlike many existing diagnosis methods that require access to faulty data (supervised diagnosis methods), the proposed diagnosis methodology belongs to the class that only requires data under normal conditions (unsupervised diagnosis methods). The effectiveness of the proposed monitoring scheme is demonstrated and validated through simulated datasets and an industrial dataset from semiconductor manufacturing.

Parameter and uncertainty estimation in stochastic differential equation models with multi-rate data and nonstationary disturbances

Updated methods are proposed for estimating model parameters and error-covariance matrices in stochastic differential equation (SDE) models. New expressions, based on the LAMLE (Laplace Approximation Maximum Likelihood Estimation) and LAB (Laplace Approximation Bayesian) SDE estimation methods are derived so that LAMLE and LAB can be used for systems with multi-rate data and nonstationary disturbances. The updated LAMLE method is tested using simulated data from a two-state continuous stirred tank reactor (CSTR) model. Comparisons are made with the results obtained using the continuous-time stochastic method (CTSM), confirming that the updated LAMLE algorithm provides reliable estimates for unknown model parameters, measurement noise variances, and process-error variances. In all situations tested, the updated LAMLE algorithm converged to reliable parameter and uncertainty estimates, while avoiding convergence difficulties encountered by CTSM in some situations. The updated LAMLE and LAB algorithms will be useful for parameter estimation in online models used in advanced process monitoring and control applications where multi-rate data and nonstationary disturbances are often encountered. Estimates of measurement noise variances and process-error variances will be helpful for state-estimator tuning.

A Novel Approach for Non-Invasive Continuous In-Line Control of Perfusion Cell Cultivations by Raman Spectroscopy.

Continuous manufacturing is becoming more important in the biopharmaceutical industry. This processing strategy is favorable, as it is more efficient, flexible, and has the potential to produce higher and more consistent product quality. At the same time, it faces some challenges, especially in cell culture. As a steady state has to be maintained over a prolonged time, it is unavoidable to implement advanced process analytical technologies to control the relevant process parameters in a fast and precise manner. One such analytical technology is Raman spectroscopy, which has proven its advantages for process monitoring and control mostly in (fed-) batch cultivations. In this study, an in-line flow cell for Raman spectroscopy is included in the cell-free harvest stream of a perfusion process. Quantitative models for glucose and lactate were generated based on five cultivations originating from varying bioreactor scales. After successfully validating the glucose model (Root Mean Square Error of Prediction (RMSEP) of ∼0.2 g/L), it was employed for control of an external glucose feed in cultivation with a glucose-free perfusion medium. The generated model was successfully applied to perform process control at 4 g/L and 1.5 g/L glucose over several days, respectively, with variability of ±0.4 g/L. The results demonstrate the high potential of Raman spectroscopy for advanced process monitoring and control of a perfusion process with a bioreactor and scale-independent measurement method.

Defining the Challenges—Identifying the Key Poisoning Elements to Be Separated in a Future Integrated Molten Salt Fast Reactor Clean-Up System for iMAGINE

Nuclear fission technologies have the potential to play a significant role in the energy mix of a net-zero and sustainable society. However, to achieve the sustainability goal two significant challenges remain: efficient and sustainable fuel usage and the minimization of long-term nuclear waste. Civil nuclear molten salt systems and technologies offer the opportunity to address both, delivering future reactors at scale for efficient and effective power production and nuclear waste burnup. Potentially, both objectives could be fulfilled in one reactor system, which could significantly improve sustainability indices. The key to this innovation is demand driven development of a significantly reduced fuel cycle with enhanced proliferation resistance which offers further potential for improvement. To achieve these goals, a transformative approach for salt clean-up during molten salt reactor operation is proposed, by concentrating on the detection and removal of key neutron poisoning elements which prevent the reactor from long-term operation. To enable this highly innovative development work, a novel analysis of the evolving elementary fuel composition, their concentrations, and their criticality influence is now provided in this work. This, combined with consideration of the oxidation states of each of these elements then provides the basis for the selection of these key poisons and the development of advanced separation processes and process monitoring. This work also discusses the importance of the effective integration of physics and chemistry when systems modelling in achieving these system development goals.

Special Issue “Advanced Process Monitoring for Industry 4.0”

Industry 4 [...]

Clustered Discriminant Regression for High-Dimensional Data Feature Extraction and Its Applications in Healthcare and Additive Manufacturing

The recent increase in applications of high-dimensional data poses a severe challenge to data analytics, such as supervised classification, particularly for online applications. To tackle this challenge, efficient and effective methods for feature extraction are critical to the performance of classification analysis. The objective of this work is to develop a new supervised feature extraction method for high-dimensional data. It is achieved by developing a clustered discriminant regression (CDR) to extract informative and discriminant features for high-dimensional data. In CDR, the variables are clustered into different groups or subspaces, within which feature extraction is performed separately. The CDR algorithm, which is a greedy approach, is implemented to obtain the solution toward optimal feature extraction. One numerical study is performed to demonstrate the performance of the proposed method for variable selection. Three case studies using healthcare and additive manufacturing data sets are accomplished to demonstrate the classification performance of the proposed methods for real-world applications. The results clearly show that the proposed method is superior over the existing method for high-dimensional data feature extraction. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —This article forwards a new supervised feature extraction method termed clustered discriminant regression. This method is highly effective for classification analysis of high-dimensional data, such as images or videos, where the number of variables is much larger than the number of samples. In our case studies on healthcare and additive manufacturing, the performance of classification analysis based on our method is superior over the existing feature extraction methods, which is confirmed by using various popular classification algorithms. For image classification, our method with elaborately selected classification algorithms can outperform a convolutional neural network. In addition, the computation efficiency of the proposed method is also promising, which enables its online applications, such as advanced manufacturing process monitoring and control.

Advanced Process Monitoring in Additive Manufacturing

AbstractLaser‐based additive manufacturing processes like laser metal deposition (LMD) and laser powder bed fusion (LPBF) are becoming more and more popular in industry. In demanding applications, however, every part has to be tested non‐destructively in a subsequent step. Online process monitoring can provide detailed documentation for each part and has the potential to reduce the post‐process inspection costs. BIAS has developed an advanced monitoring approach for in‐process temperature field measurement using a coaxial two‐channel pyrometric camera.

Brewery wastewater treatment plant key-component estimation using moving-window recurrent neural networks

This work proposes an experimental validation of software sensors for advanced on-line anaerobic digester process monitoring. The considered strategy is based on cheap available measurements (conductivity, temperature, pH, redox potential, etc) to reconstruct key component trajectories such as volatile fatty acid, carbonate and alkalinity concentrations, as well as biogas composition (methane, carbon dioxide, etc). The proposed solution considers a radial basis function artificial neural network (RBF -ANN) structure, using data processing (principal component analysis) and an efficient and fast sequential learning algorithm. In order to better reproduce unknown and complex process dynamics, the combination of a moving-window technique with a simple Jordan recurrent ANN structure (MW - RBF - RNN) is proposed. Comparative results based on real industrial data illustrate the estimation improvements provided by the MW - RBF - RNN with respect to the classical RBF - ANN structure.

A Structure Data-Driven Framework for Virtual Metrology Modeling

Virtual metrology (VM) has been widely studied in the semiconductor industry with the purpose of decreasing the cycle time and reducing the expensive metrology measurements. Ideally, a VM model should not only be able to provide accurate predictions but also present an interpretable and rational structure to accommodate fundamental restrictions and relationships that are known to be present in the process. The last aspects have been missing in the VM models proposed hitherto. Therefore, in this article, we propose a novel framework by combining in a single VM model the capability to learn from data with the ability to incorporate the domain knowledge on the process. Thus, the new methodology can use the best of both information sources: data and the subject-matter expert (SME) knowledge. The framework consists of two phases. In the first phase, a Gaussian Bayesian network (GBN) is used to extract the implicit relationships between the metrology and production/process variables. In the second phase, the target response variable is defined, the predictors are selected through the associated Markov blanket, and finally, an empirical model is estimated to accurately predict the response. The proposed framework was tested and its effectiveness was confirmed through a real industrial data from a chemical–mechanical polishing (CMP) process in semiconductor fabrication. The physical meaning of the model obtained was also scrutinized by an SME. Note to Practitioners— Unlike the conventional virtual metrology (VM) techniques that model the x–y relationship based only on process data, the proposed method aims at consolidating the process knowledge from experts into an extensive relational structure. Not only the x–y model is inferred but also the relationships among the predictors are revealed. The structure is illustrated in a form of a connected graph so that the correlation between parameters can be expressed explicitly, which is also compatible with the physical laws. With the clear visualization of the correlation structure of variables, the practitioners are able to utilize the result in different applications. The learned structure can be further integrated into the plant advanced process control system, including VM, process monitoring and diagnosis, and run-to-run (R2R) control.

Friction-Induced Vibrations during Tightening of Bolted Joints—Analytical and Experimental Results

Bolted joints are one of the most used machine elements. Holding together structures of all sizes, the integrity rests on their shoulders. Thus, an accurate and reliable assembly of the joint is crucial. While it is the aim to not experience friction-induced vibrations at all, at some situations, it is unavoidable. These cases, however, have yet been out of the focus of control algorithms due to the volatile nature of the process. This contribution delivers analytical and experimental results for the occurrence of friction-induced vibrations during tightening of bolted joints. Previous findings of system characteristic constants could be validated, which can be used to monitor the tightening process even while strong vibrations during the tightening process occur. Additionally, a real-time algorithm is presented which allows for an advanced process monitoring and control by identifying process characteristics based on which predictions of the process can be made. These measures significantly improve the capabilities to reach the target values of the tightening process.

Advanced process control and monitoring of a continuous flow micro-reactor

This paper presents a real-time advanced process control and monitoring scheme for a continuous flow micro-reactor producing an ether compound. A PLS calibration model is designed to predict the ether product yield using inline spectral data. This yield is then controlled to the desired setpoint by the proposed MPC scheme. Through real-time results, it is shown that the MPC controller is able to deliver accurate setpoint tracking even in the face of substantial plant-model mismatch caused by diluted catalyst. Furthermore, it is shown that the designed PCA monitor can effectively detect process/reaction faults, such as irregular reaction chemistry.

Quality by control: Towards model predictive control of mammalian cell culture bioprocesses.

The industrial production of complex biopharmaceuticals using recombinant mammalian cell lines is still mainly built on a quality by testing approach, which is represented by fixed process conditions and extensive testing of the end-product. In 2004 the FDA launched the process analytical technology initiative, aiming to guide the industry towards advanced process monitoring and better understanding of how critical process parameters affect the critical quality attributes. Implementation of process analytical technology into the bio-production process enables moving from the quality by testing to a more flexible quality by design approach. The application of advanced sensor systems in combination with mathematical modelling techniques offers enhanced process understanding, allows on-line prediction of critical quality attributes and subsequently real-time product quality control. In this review opportunities and unsolved issues on the road to a successful quality by design and dynamic control implementation are discussed. A major focus is directed on the preconditions for the application of model predictive control for mammalian cell culture bioprocesses. Design of experiments providing information about the process dynamics upon parameter change, dynamic process models, on-line process state predictions and powerful software environments seem to be a prerequisite for quality by control realization.

The effect of process and geometric parameters on longitudinal edge strain and product defects in cold roll forming

Roll forming has been used traditionally in the construction and housing industry for the production of longitudinal components but is now increasingly applied in the automotive industry for the manufacture of structural and crash components from ultra high-strength steel (UHSS); the incremental nature of this process allows these hard-to-form materials to be shaped with higher efficiency and less shape defects than observed in common sheet forming processes such as stamping. Tight dimensional tolerances are imposed on automotive components, and this can lead to problems when roll forming UHSS where the high material strength results in shape defect and forming problems. Recent work has therefore increasingly focused on developing process monitoring and control routines for roll forming to improve process robustness and part quality. In roll forming, the longitudinal edge strain is considered to be related to product defects such as bow, twist and end flare. Process and part shape parameters have been shown to significantly influence peak longitudinal edge strain, and the link between process and product parameters, longitudinal edge strain and shape defects needs to be understood for the roll forming of UHSS if routines for process monitoring and control are to be established. Previous studies were mainly focused on traditional roll forming materials used for building products and the like. In this paper, the effect of process and part shape parameters on the peak longitudinal edge strain, longitudinal bow and springback is experimentally and statistically investigated for three different advanced high-strength steel (AHSS) and UHSS commonly used in automotive manufacturing. The results show that there are significant differences in behaviour when forming UHSS and that forming trends differ from those reported for softer steel grades. The experimental data presented in this paper should contribute to the further development of advanced process monitoring and part shape quality control routines in the roll forming AHSS and UHSS.

Data mining and clustering in chemical process databases for monitoring and knowledge discovery

Modern chemical plants maintain large historical databases recording past sensor measurements which advanced process monitoring techniques analyze to help plant operators and engineers interpret the meaning of live trends in databases. However, many of the best process monitoring methods require data organized into groups before training is possible. In practice, such organization rarely exists and the time required to create classified training data is an obstacle to the use of advanced process monitoring strategies. Data mining and knowledge discovery techniques drawn from computer science literature can help engineers find fault states in historical databases and group them together with little detailed knowledge of the process. This study evaluates how several data clustering and feature extraction techniques work together to reveal useful trends in industrial chemical process data. Two studies on an industrial scale separation tower and the Tennessee Eastman process simulation demonstrate data clustering and feature extraction effectively revealing significant process trends from high dimensional, multivariate data. Process knowledge and supervised clustering metrics compare the cluster results against true labels in the data to compare performance of different combinations of dimensionality reduction and data clustering approaches.

Advanced data analysis and process monitoring across manufacturing plants using multilevel methods

Three‐level versions of Multilevel Simultaneous Component Analysis (MLSCA) and Multilevel Partial Least Squares (MLPLS) were developed, which are capable of separating between‐plant, between‐run and within‐run process variation, and modeling these three levels in a multivariate way. In comparison to the two‐level versions they allow to discriminate between overall differences between plants and the variation between runs within a plant. It was shown that the three‐level version of MLSCA has clear added value for the analysis of process runs from different plants. In MLPLS other projections of the multivariate data onto latent variables and different views of the data are obtained when relevant Y information is available. This has clear added value for obtaining insight into the relation between process data and Y. A special use of MLPLS is to diagnose aberrations in first principles models.In batch process monitoring MLSCA at three levels allows simultaneous multivariate modelling of batch data from different manufacturing plants. By filtering out the between‐plant and between‐run sources of variation, and using only within‐run variation, monitoring models can be improved. Using within‐run data, it is possible to build monitoring models across manufacturing units and reduce the number of nuisance alarms, while improving abnormal situation detection and diagnosis. Model transfer is only possible if static between‐plant differences exist, but not if there are dynamic differences.