Search

This paper demonstrates how the application of machine learning techniques can be used in Magnetron Sputtering (MS) processes, to detect anomalies and reduce their failure rate. Magnetron Sputtering is a widely used technique in materials science and engineering to deposit thin films of various materials for a range of applications. However, the process is complex and can be prone to various anomalies that can lead to defects in the deposited films, resulting in a non-negligible waste of coated objects. In this paper, we focus on the use of machine learning algorithms for both online and offline anomaly detection, which can help identify and diagnose process anomalies in real-time or post-process. Our results demonstrate that machine learning techniques can be used to develop anomaly detection systems, to limit failure in magnetron sputtering processes. • Simple machine learning tools can be applied to magnetron sputtering processes to detect anomalies, offline or online. • This innovative approach can limit failure in magnetron sputtering processes • Although machine learning is not common for surface scientist, this approach is reliable and based on accessible algorithms • The algorithms assist the operator in the decision making without seeking to replace him. [ABSTRACT FROM AUTHOR]

Detecting the anomalies in a large amounts of high-dimensional data has been a challenging task. In the Industry 4.0 environment, large-scale high-dimensional monitoring data features the complex pattern of high level semantics. In order to provide enterprise-wide monitoring solutions, it is necessary to identify the high-level semantic patterns of the anomalies in these data without splitting them. Existing end-to-end deep neural networks for time series are capable of recognizing the high-level semantics in natural language or speech signals, but they are barely applied in real-time anomaly detection of industrial data because of the large time costs. In this paper, we leverage the self-supervised contrastive learning methodology and propose a Composite Semantic Augmentation Encoder (CSAE) to provide an appropriate representation of industrial data and implement quick detection of anomalies in industrial application environments. CSAE is a non-sequential deep neural network with two augmentation layers and a mandatory layer. The two layers of data-augmentation are built to expand the size of samples of both low-level semantic anomalies and high-level semantic anomalies, which enables CSAE to discover diverse anomalies and improves its accuracy of high-level semantic pattern recognition. The mandatory layer is built to compress and reserve the temporal information in the industrial data to accelerate the anomaly detection. Therefore, as a non-sequential contrastive learning model, CSAE has faster training convergence than the usual sequence models. The experiment results have verified that CSAE can achieve higher prediction accuracy with less time consumption than existing machine learning models in the tasks of high dimensional anomaly pattern detection. [ABSTRACT FROM AUTHOR]

The evolution of journal bearings, from fixed profiles to adaptive configurations, signifies a captivating progression in journal bearing technology. A notable innovation in adaptive bearings is Multi-pad Bidirectional Adjustable Bearing (MBAB), which can control film thickness in radial and circumferential directions. This paper introduces a novel data-driven approach, utilizing machine learning to model and optimize the static performance of MBABs with asymmetric bearing element adjustments. The study uses a parameter-independent Jaya algorithm coupled with machine learning models to identify optimal combinations of adjustments. Results highlight the significance of negative radial adjustments and asymmetric profiles for optimal performance. This research contributes to Industry 4.0 by bridging the physical-digital gap, offering a data-driven solution to enhance the performance of these advanced bearings. [ABSTRACT FROM AUTHOR]

Localization models, particularly for indoor applications like autonomous vehicles and smart manufacturing in Industry 4.0, have emerged as a crucial research area because of the growing need for precise positioning. In this context, distributed cell-free MIMO have demonstrated capabilities for a higher fine-grained resolution and accuracy. This paper explores different approaches to address challenges related to positioning and corresponding data transmission and model compression problems in cell-free massive MIMO systems. We present a novel deep-learning powered CSI compression model for reducing the signalization data associated with the localization task, minimizing the fronthaul capacity requirement by at least a factor of 2, without a significant loss of accuracy. In our research, we also focus on the challenges of deploying deep end-to-end neural networks on resource-constrained embedded platforms. Two approaches are proposed: model modification and magnitude-based pruning. Our results show that the pruned end-to-end localization model has a smaller size by 36% at the cost of an increase on the localization error by a factor of 1.5 (precision is reduced from approximately 5 mm to 7.8 mm). [ABSTRACT FROM AUTHOR]

Abstract With the arrival of Industry 4.0, the overall transformation using digital integration and intelligent engineering has taken a giant leap towards futuristic technology. All devices today are equipped with machine learning, automation has become a priority and thus another industrial revolution is in the making. In this state-of-the-art paper, we have performed bibliometric analysis and an extensive survey on recent developments in the field of "Industry 4.0". In bibliometric analysis, different performance metrics are extracted, such as: total papers, total citations, and citation per paper. Further, top 10 of the most productive and highly cited authors, major subject areas, sources or journals, countries, and institutions are evaluated. A list of highly influential papers is also assessed. Later on, a detailed discussion of the most cited papers is analysed and a sectional classification is provided. This paper summarizes the growth structure of Industry 4.0 during the last 5 years and provides the concise background overview of Industry 4.0 related works and various application areas. [ABSTRACT FROM AUTHOR]

In the Industry 4.0 era, a myriad of sensors of diverse nature (temperature, pressure, etc.) is spreading throughout the entire value chain of industries, being potentially exploitable for multiple purposes, such as Predictive Maintenance (PdM): the just-in-time maintenance of industrial assets, which results in reduced operating costs, increased operator safety, etc. Nowadays, industrial processes require to be highly configurable, in order to proactively adapt their operation to diverse factors such as user needs, product updates or supply chain uncertainties. This limits current Industry 4.0-PdM solutions, typically consisting of ad-hoc developments intended for specific scenarios, i.e. they are designed to operate under certain conditions (configurations, employed sensors, etc.), being unable to manage changes in their setup. This paper presents a general Deep Learning (DL) architecture, MachNet, which deals with such heterogeneity and is able to address PdM problems of a diverse nature. The modularity of the proposed architecture enables it to deal with an arbitrary number of sensors of different types, also allowing the integration of prior information (age of assets, material type, etc.), which clearly affects performance and is often neglected. In practice, our architecture effortlessly adapts to the assets' specifications and to different PdM problems. That is, MachNet becomes an architectural template that can be instantiated for a given scenario. We tested our proposal in two different PdM-related problems: Health State (HS) and Remaining-useful-Life (RuL) estimation, achieving in both cases comparable or superior performance to other state-of-the-art approaches, with the additional advantage of the generality that MachNet offers. [ABSTRACT FROM AUTHOR]

The wider availability of low-cost sensing and data acquisition technologies means that real-time sensing of tribological parameters is becoming increasingly viable. Consequently, the potential to use these technologies to monitor in-service tribological components has increased significantly. This paper presents a review of a number of state-of-the-art in sensors for measuring friction, wear and lubricant properties. It also elaborates on the use of sensor coatings as an emerging area for directly probing the tribological interface. It is concluded that sensors will find ever increasing uses in condition monitoring" applications. However, sensing and tribology is beginning to evolve towards "Tribotronics" where combining the sensing of machine elements that have conventionally been passive with computational capability, or even embedded intelligence, along with actuation can create active machine elements, optimised to operate with say minimum power loss in all situations of duty. Additionally, it is noted that by incorporating sensing and responsive capabilities, functional surfaces can also become part of a bigger connected systems particularly in association with Industry 4.0. Increased use of sensors in tribological components alongside machine learning and artificial intelligence, will also support the shift in industrial tribological analytics. [ABSTRACT FROM AUTHOR]

The evaluation of process improvements measures in offsite construction shop floors often relies on experts' opinion, with limited use of empirical data gathered by sensors in real-time. To address this issue, there is a need for methods that integrate expert's tacit knowledge with robust data analysis techniques. This paper describes the application of exploratory data analysis techniques to evaluate improvement suggestions proposed by expert's, supported by data collected by sensors on the shop floor and building information models. The presented method involves a quantitative and qualitative digitalization-based approach where improvement suggestions are modelled and validated though machine learning algorithms and hypothesis testing. The contribution of this study is a method that combines real-time data, building information models, and knowledge modeling from experts to evaluate process improvement on offsite construction shop floors. • A method to assess improvements based on experts input and real-time data. • Machine learning is applied to analyze data from RFID sensors and BIM models. • The automation in workstations is rated based on production balance and efficiency. • Strategies to increase production flexibility are rated using statistical analysis. [ABSTRACT FROM AUTHOR]

Highlights • We overview the state of art in data fusion and analysis for industrial prognosis. • Descriptive, predictive and prescriptive prognostic models are reviewed. • Future trends and challenges are provided to stimulate research on this topic. Abstract The so-called "smartization" of manufacturing industries has been conceived as the fourth industrial revolution or Industry 4.0, a paradigm shift propelled by the upsurge and progressive maturity of new Information and Communication Technologies (ICT) applied to industrial processes and products. From a data science perspective, this paradigm shift allows extracting relevant knowledge from monitored assets through the adoption of intelligent monitoring and data fusion strategies, as well as by the application of machine learning and optimization methods. One of the main goals of data science in this context is to effectively predict abnormal behaviors in industrial machinery, tools and processes so as to anticipate critical events and damage, eventually causing important economical losses and safety issues. In this context, data-driven prognosis is gradually gaining attention in different industrial sectors. This paper provides a comprehensive survey of the recent developments in data fusion and machine learning for industrial prognosis, placing an emphasis on the identification of research trends, niches of opportunity and unexplored challenges. To this end, a principled categorization of the utilized feature extraction techniques and machine learning methods will be provided on the basis of its intended purpose: analyze what caused the failure (descriptive), determine when the monitored asset will fail (predictive) or decide what to do so as to minimize its impact on the industry at hand (prescriptive). This threefold analysis, along with a discussion on its hardware and software implications, intends to serve as a stepping stone for future researchers and practitioners to join the community investigating on this vibrant field. [ABSTRACT FROM AUTHOR]

In this paper we introduce the current thrust of development in machine learning and artificial intelligence, fueled by advances in statistical learning theory over the last 20 years and commercial successes by leading big data companies. Then we discuss the characteristics of process manufacturing systems and briefly review the data analytics research and development in the last three decades. We give three attributes for process data analytics to make machine learning techniques applicable in the process industries. Next we provide a perspective on the currently active topics in machine learning that could be opportunities for process data analytics research and development. Finally we address the importance of a data analytics culture. Issues discussed range from technology development to workforce education and from government initiatives to curriculum enhancement. [ABSTRACT FROM AUTHOR]

In the railway domain, rolling stock maintenance affects service operation time and efficiency. Minimizing train unavailability is essential for reducing capital loss and operational costs. To this aim, prediction of failures of rolling stock equipment is crucial to proactively trigger proper maintenance activities. Indeed, predictive maintenance is a golden example of the digital transformation within Industry 4.0, which affects several engineering processes in the railway domain. Nowadays, it may leverage artificial intelligence and machine learning algorithms to forecast failures and schedule the optimal time for maintenance actions. Generally, rail systems deteriorate gradually over time or fail directly, leading to data that vary extremely slowly. Indeed, ML approaches for predictive maintenance should consider this type of data to accurately predict and forecast failures. This paper proposes a methodology based on Long Short-Term Memory deep learning algorithms for predictive maintenance of railway rolling stock equipment. The methodology allows us to properly learn long-term dependencies for gradually changing data, and both predicting and forecasting failures of rail equipment. In the framework of an academic-industrial partnership, the methodology is experimented on a train traction converter cooling system, demonstrating its applicability and benefits. The results show that it outperforms state-of-the-art methods, reaching a failure prediction and forecasting accuracy over 99%, with a false alarm rate of ∼ 0. 4 % and a mean absolute error in the order of 1 0 − 4 , respectively. [ABSTRACT FROM AUTHOR]

The fourth industrial revolution creates ubiquitous sensor data in production plants. To generate maximum value out of these data, reliable and precise time series-based machine learning methods like temporal neural networks are needed. This paper proposes a novel sequence-to-sequence deep learning architecture for time series segmentation called PrecTime which tries to combine the concepts and advantages of sliding window and dense labeling approaches. The general-purpose architecture is evaluated on a real-world industry dataset containing the End-of-Line testing sensor data of hydraulic pumps. We are able to show that PrecTime outperforms five implemented state-of-the-art baseline networks based on multiple metrics. The achieved segmentation accuracy of around 96% shows that PrecTime can achieve results close to human intelligence in operational state segmentation within a testing cycle. [ABSTRACT FROM AUTHOR]

Normalisation is a preprocessing technique widely employed in Machine Learning (ML)-based solutions for industry to equalise the features' contribution. However, few researchers have analysed the normalisation effect and its implications on the ML algorithm performance, especially on Euclidean distance-based algorithms, such as the well-known K-Nearest Neighbours and K-means. In this sense, this paper formally analyses the effect of normalisation yielding results significantly far from the state-of-the-art traditional claims. In particular, this paper shows that normalisation does not equalise the contribution of the features, with the consequent impact on the performance of the learning process for a particular problem. More concretely, this demonstration is made on K-Nearest Neighbours and K-means Euclidean distance-based ML algorithms. This paper concludes that normalisation can be viewed as an unsupervised Feature Weighting method. In this context, a new metric (Normalisation weight) for measuring the impact of normalisation on the features is presented. Likewise, an analysis of the normalisation effect on the Euclidean distance is conducted and a new metric referred to as Proportional influence that measures the features influence on the Euclidean distance is proposed. Both metrics enable the automatic selection of the most appropriate normalisation method for a particular engineering problem, which can significantly improve both the computational cost and classification performance of K-Nearest Neighbours and K-means algorithms. The analytical conclusions are validated on well-known datasets from the UCI repository and a real-life application from the refinery industry. [ABSTRACT FROM AUTHOR]

Abstract Modern production and logistics systems, supply chains, and Industry 4.0 networks are challenged by increased uncertainty and risks, multiple feedback cycles, and dynamics. Control theory is an interesting research avenue which contributes to further insights concerning the management of the given challenges in operations and supply chain management. In this paper, the applicability of control theory to engineering and management problems in supply chain operations is investigated. Our analysis bridges the fundamentals of control and systems theory to supply chain and operations management. This study extends our previous survey in the Annual Reviews in Control (Ivanov et al. 2012) by including new literature published in 2012–2018, identifying two new directions of control theory applications (i.e., ripple effect analysis in the supply chains and scheduling in Industry 4.0) and analysis towards the digital technology use in control theoretic models. It describes important issues and perspectives that delineate dynamics in supply chains, operations, and Industry 4.0 networks and identifies and systemizes different streams in the application of control theory to operations and supply chain management and engineering in the period from 1960–2018. It updates the existing applications and classifications, performs a critical analysis, and discusses further research avenues. Further development of interdisciplinary approaches to supply chain optimization is argued. An extended cooperation between control engineers and supply chain experts may have the potential to introduce more realism to dynamic planning and models, and improve performance in production and logistics systems, supply chains, and Industry 4.0 networks. Finally, we analyze the trends towards the intellectualization of control and its development towards supply chain control analytics. [ABSTRACT FROM AUTHOR]

Currently, in manufacturing, massive useful data about health condition and maintenance is often available thanks to Industry 4.0 technologies. However, how to take advantage of historical data to optimize maintenance policies for multi-component systems has still been a challenging problem. This is especially true when maintenance cost models at component level are not available and/or maintenance actions are imperfect. In order to cope with this issue, we propose in this paper an artificial-intelligence-based maintenance approach which first constructs a predictor based on artificial neural network (ANN) for estimating maintenance cost at system level and then employs a customized multi-agent deep reinforcement learning algorithm to optimize maintenance decisions that can be applied for large-scale systems. To evaluate the performance and scalability of the proposed maintenance approach, numerical studies are conducted on a small 4-component system with different configurations and a large system composed of 15 components considering both deterministic and random maintenance quality. The simulation results show that ANN-based predictor is efficient for maintenance cost forecasting and multi-agent deep reinforcement learning is a promising solution for maintenance decision-making and optimization. • An artificial-intelligence-based maintenance approach is proposed. • Both deterministic and random quality for imperfect maintenance are investigated. • An ANN-based predictor is developed for estimating maintenance cost at system level. • A multi-agent deep RL is customized for maintenance optimization. • The performance of the proposed approach is numerically investigated. [ABSTRACT FROM AUTHOR]

Under the vision of industry 4.0, the integration of drones in last-mile delivery can transform traditional delivery practices and provide competitive advantages. However, the combinatorial nature of the routing problem and the technical limitations of the drones present a real challenge for adopting a stand-alone drone delivery as an alternative to truck delivery. This study introduces the Parking Location and Traveling Salesman Problem with Homogeneous Drones (PLTSPHD). This problem considers a scenario in which a single truck carries identical drones along with parcels from the depot to preassigned launching/parking sites, from where the drones complete the last-mile deliveries. In contrast to previous studies that tackle truck-drone delivery using conventional optimization approaches, this paper proposes a two-phase machine learning (ML) approach for the PLTSPHD, which minimizes the total operational cost of the last-mile problem. The proposed ML approach for PLTSPHD consists of clustering and routing phases. In the first phase, a constrained k -means clustering algorithm is proposed to cluster delivery locations based on the maximum flight range and number of available drones per truck. A deep reinforcement learning (DRL) model is then developed in the second stage to find an optimal route among all constrained clusters. Experimental results show that solving the presented truck-drone problem using the ML framework can significantly reduce the operational cost compared to standard truck delivery. The constrained clustering reduces the complexity of the routing problem while adhering to the constraints. In addition, the trained DRL model outperforms the state-of-art Google's OR-tools solver and other types of well-known heuristics in terms of both solution quality and computation time. Moreover, a sensitivity analysis of different key parameters is conducted to highlight some key trade-offs in using multiple drones and their dependence on operating costs and problem sizes. [ABSTRACT FROM AUTHOR]

• Process intensification 4.0 (PI4.0) framework allows system-level transformations. • PI4.0 incorporates a data approach based on the design principles of industry 4.0. • An iterative design strategy is presented as the backbone of PI4.0. • Machine Learning/advanced computing accelerate knowledge discovery, driving PI4.0. • Different case studies are reviewed that support the philosophy behind PI4.0. This paper reviews system-level transformations converging into the next generation of Process Intensification strategies defined as PI4.0. Process Intensification 4.0 uses data-driven algorithms to understand other physical and chemical processes that improve equipment design, predictive control, and optimization. Following this, an overview of the use of Artificial Intelligence techniques, particularly Machine Learning for the acceleration of equipment design, process optimization, and streamlining, is presented. This work will highlight and discuss the emerging framework of the integration between Circular Chemistry, Industry 4.0, and Process Intensification and how the data obtained from this integration is at the core of the next generation of Process Intensification strategies. This is supported by a discussion of different cases that apply data-driven models enabled by Machine Learning as a mean to enhance an intensified system (product synthesis, equipment or methods). [Display omitted] [ABSTRACT FROM AUTHOR]

• Recognize operator's activities through a machine learning classifier. • Assess Vibration risk in real time to enhance operator's safety. • Develop a lightweight wearable device for real time monitoring. • Validate the process in real conditions through experimental tests. While being a fundamental driver of competitiveness in agroindustry, technological innovation has also introduced new critical elements related, for example, to the sustainability of the production processes as well as to the safety of workers. In such regard, the advent of the 4th industrial revolution (Agriculture 4.0) based on digitalization, is an unprecedented opportunity of rethinking the role of innovation in a new human-centric perspective. In particular, the establishment of an interconnected work environment and the augmentation of the operator's physical, sensorial, and cognitive capabilities, are two technologies which can be effectively employed for substantially improving the ergonomics and safety conditions on the workplace. This paper approaches such topic referring to the vibration risk, which is a well-known cause of work-related pathologies, and proposes an original methodology for mapping the risk exposure of the operators to the activities performed. A miniaturized wearable device is employed to collect vibration data, and the signals obtained are segmented in time windows and processed in order to extract the significant features. Finally, a machine learning classifier has been developed to recognize the worker's activity and to evaluate the related exposure to vibration risks. To validate the methodology proposed, an experimental analysis in real operating conditions has been finally carried out by monitoring the activities performed by a team of workers during harvesting operations. The results obtained demonstrate the feasibility and the effectiveness of the methodology proposed. [ABSTRACT FROM AUTHOR]

This paper presents a comparative assessment of two-dimensional convolutional neural networks (2DCNN) and boosting methods for regression-based textile whiteness estimation, applied to high resolution images of textiles of an industrial cotton textiles producer, labeled with whiteness values, thus enabling supervised learning. The images were taken under various lighting conditions. Concerning the machine learning methods, Random Forest and XGBoost were the selected and tested boosting techniques on which model hyper-parameter tuning was applied, whereas regarding the 2DCNN architectures, the known from literature ColorNet architecture was selected and a more shallow one, called WERegNet, was introduced. Data augmentation was applied during pre-processing, due to the limited amount of available data. Based on the simulation results, the WERegNet architecture surpasses ColorNet and XGBoost in terms of performance, while it is comparable with Random Forest on test set, based on model selection measure Normalized Root Mean Squared Error. [ABSTRACT FROM AUTHOR]

Recent trends in intelligent manufacturing are transforming shop floor environments into digital factories, thanks to a pervasive integration of information and communication technologies in production lines. Industrial processes become the source of high-volume heterogeneous data, paving the way to create manufacturing intelligence by means of machine learning and data-driven methodologies. In such settings, predictive diagnostics play a crucial role, as they promise to predict future critical conditions in the production process. Unfortunately, the diffusion of data-driven predictive maintenance methodologies is limited by (i) the absence of timely ground-truth knowledge (i.e., class labels), required in the learning phase of data-driven supervised approaches, and (ii) the limited availability of data-mining expertise among application-domain experts, required to harness the power of machine learning techniques. Innovative data-driven services are needed to support domain experts in (i) applying powerful self-learning intelligent techniques with limited technical expertise and (ii) easily understanding results and choices operated by such intelligent techniques, to increase trust by means of transparency. To this aim, this paper presents UDaMP, an integrated platform to support manufacturing intelligence by providing a transparent, self-tuning, unsupervised discovery and assisted data labelling service for predictive maintenance, specifically targeted at cyclic industrial processes. UDaMP includes (i) production-cycle-aware feature engineering, (ii) unsupervised discovery of production-cycle categories, (iii) self-tuning of the optimal number of categories, (iv) human-readable characterisation of production-cycle categories, and (v) assisted data labelling for domain experts. Scalable clustering algorithms automatically discover groups of production cycles sharing common time-independent properties. A self-tuning strategy is integrated to automatically configure the specific input parameter and select the best approach for the data under analysis. Each cluster is then locally characterised through the data distribution of the top 10 most relevant features to support domain experts in uncovering its meaning. Experimental evaluation of UDaMP has been performed on real-world data collected in two different industrial settings. • Semi-supervised labelling of production cycles for predictive maintenance. • Wide applicability to any cycle-based production process without ground truth labels. • Exploitable by domain experts thanks to transparent self-tuning techniques. [ABSTRACT FROM AUTHOR]

In the fourth Industrial Revolution, cost-effective planning and rational management were the key to the success of the revolution. This paper mainly studies the development and application of models in machine learning technology. The abnormal activities monitored in real time are rectified so that the customer's electronic orders can be displayed through the support of big data, thus laying the foundation for the development of intelligent logistics. Under the data system, an exception model is created and classified and regressed. In this model, the security and stability of customer orders in the network can be automatically detected, and the abnormal data can be analyzed and evaluated. Unusual circumstances of this kind need to be in an intelligent logistics environment, and delivery tasks must be called intuitive for special care. Early detection of abnormal order events is expected to improve the accuracy of delivery planning. To enable new technical solutions, the logistics industry and economic decision-makers often lack the IT background and expertise needed to start developing new systems and technical solutions. Evaluate the benefits of using. Implementation and integration complexity is seen as one of the three major obstacles to the success of the IoT above. This is by hindering long-term investment in new technologies from slowing down digitization. [ABSTRACT FROM AUTHOR]

Industry 4.0 has brought forth many advantages and challenges for the industry players. Many organizations are strategizing to take advantage of this industrial paradigm shift, thus improving the sustainability of the enterprise. However, there are many factors such as talent development, machinery advancement and infrastructure development which involve huge investment that need to be considered. This paper presents an enhanced adaptive model for the implementation of the lean and green (L&G) strategy in processing sectors to solve dynamic industry problems associated with Industry 4.0. A feature of this enhanced adaptive model is that it combines experts' experience and operational data as input in dealing with real industry application. A lean and green index is coupled in the model to serve as a benchmark and process improvement tracking indicator. This allows the industrialists to set a lean and green index (LGI) target for effective process improvement. From this integrated model, an ensemble of backpropagation optimizers is then used to identify the best-optimized strategy. This ensemble optimizer is formulated to perform operation improvement and update the targeted LGI automatically when a higher index is achieved for continuous improvement. A case study on a combined heat and power plant is performed and reflects an improvement of 18.25% on the LGI. This work serves as a practical transition strategy for the industrialist desiring to improve the sustainability of the facility with Industry 4.0 elements at minimum investment cost. • Application of enhanced adaptive lean and green model in energy industry. • Combined qualitative and quantitative data input for process optimization. • Use of ensemble solver to analyze the best optimum pathway for process improvement. • The model improved the process performance by 18.25% in 8 months. • This model is user friendly and pragmatic in addressing real industry problems. [ABSTRACT FROM AUTHOR]

• Data-driven models (DDMs) will become widespread across manufacturing. • Paramount to DDMs is the collection of an accurate set of model development data. • Process manufacturers face unique considerations and challenges in collecting data. • These points are presented in the context of the CRISP-DM framework. • This supports the development of DDMs to meet manufacturers' requirements. The increasing availability of data, due to the adoption of low-cost industrial internet of things technologies, coupled with increasing processing power from cloud computing, is fuelling increase use of data-driven models in manufacturing. Utilising case studies from the food and drink industry and waste management industry, the considerations and challenges faced when developing data-driven models for manufacturing systems are explored. Ensuring a high-quality set of model development data that accurately represents the manufacturing system is key to the successful development of a data-driven model. The cross-industry standard process for data mining (CRISP-DM) framework is used to provide a reference at to what stage process manufacturers will face unique considerations and challenges when developing a data-driven model. This paper then explores how data-driven models can be utilised to characterise process streams and support the implementation of the circular economy principals, process resilience and waste valorisation. [ABSTRACT FROM AUTHOR]

The Architecture, Engineering and Construction (AEC) industry is fraught with complex and difficult problems. Artificial intelligence (AI) represents a powerful tool to assist in addressing these problems. Therefore, over the years, researchers have been conducting research on AI in the AEC industry (AI-in-the-AECI). In this paper, the first comprehensive scientometric study appraising the state-of-the-art of research on AI-in-the-AECI is presented. The science mapping method was used to systematically and quantitatively analyze 41,827 related bibliographic records retrieved from Scopus. The results indicated that genetic algorithms, neural networks, fuzzy logic, fuzzy sets, and machine learning have been the most widely used AI methods in AEC. Optimization, simulation, uncertainty, project management, and bridges have been the most commonly addressed topics/issues using AI methods/concepts. The primary value and uniqueness of this study lies in it being the first in providing an up-to-date inclusive, big picture of the literature on AI-in-the-AECI. This study adds value to the AEC literature through visualizing and understanding trends and patterns, identifying main research interests, journals, institutions, and countries, and how these are linked within now-available studies on AI-in-the-AECI. The findings bring to light the deficiencies in the current research and provide paths for future research, where they indicated that future research opportunities lie in applying robotic automation and convolutional neural networks to AEC problems. For the world of practice, the study offers a readily-available point of reference for practitioners, policy makers, and research and development (R&D) bodies. This study therefore raises the level of awareness of AI and facilitates building the intellectual wealth of the AI area in the AEC industry. • Genetic algorithms, neural networks, fuzzy logic, fuzzy sets, and machine learning are the most widely used AI methods in AEC. • Optimization, simulation, uncertainty, project management, and bridges are the most commonly addressed topics/issues using AI methods/concepts. • Future research opportunities lie in applying robotic automation to AEC problems. • There are future research opportunities also in using convolutional neural networks (CNNs) for solving AEC problems. [ABSTRACT FROM AUTHOR]

• Development of the innovative machine learning based design support system (DesSS). • Comparison with other state-of-the-art machine learning and deep learning model. • Validation of the DesSS on two real use case datasets. • High computational speed and accuracy compared to simulation tools. • Trade-off between the model interpretability, computation effort and accuracy. In the engineering practice, it frequently occurs that designers, final or intermediate users have to roughly estimate some basic performance or specification data on the basis of input data available at the moment, which can be time-consuming. There is the need for a tool that will fill the missing gap in the optimization problems in engineering design processes, by making use of the advances in the artificial intelligence field. This paper aims to fill this gap by introducing an innovative Design Support System (DesSS), originated from the Decision Support System, for the prediction and estimation of machine specification data such as machine geometry and machine design on the basis of heterogeneous input parameters. As the main core of the developed DesSS, we introduced different machine learning (ML) approaches based on Decision/Regression Tree, k-Nearest Neighbors, and Neighborhood Component Features Selection. Experimental results obtained on a real use case and using two different real datasets demonstrated the reliability and the effectiveness of the proposed approach. The innovative machine learning-based DesSS meant for supporting the designing choice, can bring various benefits such as the easier decision-making, conservation of the company's knowledge, savings in man-hours, higher computational speed and accuracy. [ABSTRACT FROM AUTHOR]

Polyacrylonitrile (PAN) is a versatile man-made polymer and has been used in a large array of products since its first mass production in the mid 40s. Among all applications of PAN the widely used application is in manufacture of precursor fiber for fabrication of carbon fibers. The process of PAN-based carbon fiber production comprises fiber spinning, thermal stabilization and carbonization stages. Carbon fiber properties are significantly dependent on the quality of PAN precursor fiber and in particular the process parameters involved in thermal stabilization. This paper is the first comprehensive review that provides a general understanding of the links between PAN fiber structure, properties, and its stabilization process along with the use of mathematical modelling as a powerful tool in prediction and optimization of the processes involved. Since the promise of the mathematical modelling is to predict the future behaviour of the system and the value of the variables for the unseen or unmeasured domain of variables; and in the era of industry 4.0 rise, this review will be valuable in further understanding of the intricate processes of carbon fiber manufacture and utilising the advanced mathematical modelling using machine learning techniques to predict and optimize a range of critical factors that control the quality of PAN and resultant carbon fibers. [ABSTRACT FROM AUTHOR]

The Industry 4.0 paradigm is being increasingly adopted in the production, distribution and commercialization chains worldwide. The integration of the cutting-edge techniques behind it entails a deep and complex revolution – changing from scheduled-based processes to smart, reactive ones – that has to be thoroughly applied at different levels. Aiming to shed some light on the path towards such evolution, this work presents an Industry 4.0 based approach for facing a key aspect within factories: the health assessment of critical assets. This work is framed in the context of the innovative project SiMoDiM , which pursues the design and integration of a predictive maintenance system for the stainless steel industry. As a case of study, it focuses on the machinery involved in the production of high-quality steel sheets, i.e. the Hot Rolling Process , and concretely on predicting the degradation of the drums within the heating coilers of Steckel mills (parts with an expensive replacement that work under severe mechanical and thermal stresses). This paper describes a predictive model based on a Bayesian Filter , a tool from the Machine Learning field, to estimate and predict the gradual degradation of such machinery, permitting the operators to make informed decisions regarding maintenance operations. For achieving that, the proposed model iteratively fuses expert knowledge with real time information coming from the hot rolling processes carried out in the factory. The predictive model has been fitted and evaluated with real data from ∼ 118k processes, proving its virtues for promoting the Industry 4.0 era. • A Bayesian Filter is proposed for predicting the degradation state of machinery. • Such a model iteratively fuses expert knowledge with real time data from the factory. • A set of recipes for analyzing and processing plant and processes data are provided. • As a case of study, it is applied to the machinery within the Hot Rolling process. • The proposal has been assessed with data from +118k processes from a real factory. [ABSTRACT FROM AUTHOR]