Real Industrial Environment Research Articles

Workflow management system with smart procedures

Supervision of repair and diagnostic works aimed at improving the safety of maintenance crews is one of the key objectives of the distributed INRED system. Working in a real industrial environment, the INRED system includes, among others, the so-called INRED-Workflow, which provides an infrastructure for process automation. Participants of the service processes, managed by the INRED-Workflow, are controlled at each stage of the performed service procedures, both by the system and other process participants, such as quality managers and technologists. All data collected from the service processes is stored in the System Knowledge Repository (SKR) for further processing by using advanced algorithms, and the so-called Smart Procedures merge services supplied by other INRED system modules. The applicability of workflow management systems in conjunction with image recognition and machine learning methods has not yet been thoroughly explored. The presented paper shows the innovative usage of such systems in the supervision of the repair and diagnostic works.

A Machine Vision-Based Pipe Leakage Detection System for Automated Power Plant Maintenance.

Industrial pipework maintenance inspection can be automated through machine vision-based effusion monitoring. However, colorless effusions such as water can be difficult to detect in a complex industrial environment due to weak illumination and poor visibility of the background. This paper deploys the reflective characteristics of effusion and its lower temperature compared to the environment in order to develop an automatic inspection system for power plant pipeworks’ maintenance. Such a system is aimed at detecting the colorless fluid effusion based on dual source images and a contour features algorithm. In this respect, a visible light source unit highlights the reflective features of the effusion edge. Meanwhile, high-definition images of the potential effusion are acquired under both visible and infrared lights. A customized image processing procedure extracts the potential effusion features from the infrared image to retrieve the region of interest for segmentation purposes and transfer such information to the visible light image to determine the effusion contour. Finally, a decision-making support tool based on the image contour closure is enabled for classification purposes. The implementation of the proposed system is tested on a real industrial environment. Experimental results show a classification accuracy up to 99%, demonstrating excellent suitability in meeting industrial requirements.

Single RGB Image 6D Object Grasping System Using Pixel-Wise Voting Network.

A robotic system that can autonomously recognize object and grasp it in a real scene with heavy occlusion would be desirable. In this paper, we integrate the techniques of object detection, pose estimation and grasping plan on Kinova Gen3 (KG3), a 7 degrees of freedom (DOF) robotic arm with a low-performance native camera sensor, to implement an autonomous real-time 6 dimensional (6D) robotic grasping system. To estimate the object 6D pose, the pixel-wise voting network (PV-net), is applied in the grasping system. However, the PV-net method can not distinguish the object from its photo through only RGB image input. To meet the demands of a real industrial environment, a rapid analytical method on a point cloud is developed to judge whether the detected object is real or not. In addition, our system shows a stable and robust performance in different installation positions with heavily cluttered scenes.

Intelligent Robust Cross-Domain Fault Diagnostic Method for Rotating Machines Using Noisy Condition Labels

Cross-domain fault diagnosis methods have been successfully and widely developed in the past years, which focus on practical industrial scenarios with training and testing data from numerous machinery working regimes. Due to the remarkable effectiveness in such problems, deep learning-based domain adaptation approaches have been attracting increasing attention. However, the existing methods in the literature are generally lower compared to environmental noise and data availability, and it is difficult to achieve promising performance under harsh practical conditions. This paper proposes a new cross-domain fault diagnosis method with enhanced robustness. Noisy labels are introduced to significantly increase the generalization ability of the data-driven model. Promising diagnosis performance can be obtained with strong noise interference in testing, as well as in practical cases with low-quality data. Experiments on two rotating machinery datasets are carried out for validation. The results indicate that the proposed algorithm is well suited to be applied in real industrial environments to achieve promising performance with variations of working conditions.

RFID Tag Classification for False-Positive Reads Identification via Deep Learning With Multidimensional Signal Fusion

Radio-frequency identification has been widely used in logistics, significantly improving logistics management’s informatization level and efficiency. However, false-positive detection is discussed as one of the major results leading to the recording error. Particularly in industrial scenarios, the measurement noise and high accuracy requirements pose challenges to the tag classification methods. In this paper, a novel two-stage deep learning model consisting of a multi-dimension deep residual network for feature learning and a single-dimensional convolutional neural network for feature fusion and classification is proposed. In the first stage, the multi-dimension deep residual networks are trained in each signal dimension respectively, to extract the optimal single-dimensional features for classification. In the second stage, the single-dimensional convolutional neural network is used for feature fusion and deeper feature learning to obtain a robust classifier. Real data obtained from a laboratory warehouse environment and a real logistics workshop in an automobile factory respectively are used to evaluate the performance of the proposed model. Experimental results show that the proposed method has satisfactory accuracies both in the laboratory and real industrial environment and has portability, specifically, which provides the classification accuracy of 100% in the laboratory and higher than 99.91% in the real industrial environment.

AI-based vision system for collision detection in HRC applications

Human-Robot Collaboration (HRC) enabling mechanisms require real-time detection of potential collisions among human and robots. Taking under consideration the already existing standards and the literature, most of collision detection techniques require the integration of sensorial systems on the robot aiming to identify the contact events. This paper deals with a novel approach for the identification of human and robot collision based on vision systems. Moreover, Artificial Intelligent (AI) algorithms are required to classify the captured data near real-time and to provide a score about the collision status (contact or non-contact) between a human and the robot. Accordingly, the AI models should be trained using the appropriate image data enabling an accurate classification. The proposed system has been developed in a lab environment. A detailed presentation of the system implementation, its performance and the potential integration in a real industrial environment are discussed in this paper.

Integrated planning framework for preventive maintenance grouping: A case study for a conveyor system in the Chilean mining industry

Increasingly complex production environments pose new challenges and require the development of new methodological approaches in the search for effective and cost-efficient maintenance planning, taking into account their applicability in real industrial environments. In this context, this article proposes an extension of the integrated framework for opportunistic preventive maintenance (PM) planning, proposed originally in Viveros et al. integrating non-negligible execution times of PM activities and time-window tolerances criterion for the generation of opportunistic grouping schemes. This work offers the implementation of the proposed framework in a practical case within the Chilean mining industry, expanding the analysis on scarce resource availability scenarios for PM tasks to be performed on conveyor belts in the grinding process. The proposed planning optimization model is formulated under the mixed-integer linear programming (MILP) paradigm, and searches minimizing the asset unavailability under several tolerance levels for the case study analyzed. The results show a 35% downtime reduction for a maximum tolerance factor of 10% considering an unconstrained maintenance resource scenario, which confirms the ability of the model to increase asset availability, minimizing system interruptions, improving its cost efficiency, and enhancing productivity levels in the mining industry.

Map-less long-term localization in complex industrial environments

PurposeThis paper aims to study the localization problem for autonomous industrial vehicles in the complex industrial environments. Aiming for practical applications, the pursuit is to build a map-less localization system which can be used in the presence of dynamic obstacles, short-term and long-term environment changes.Design/methodology/approachThe proposed system contains four main modules, including long-term place graph updating, global localization and re-localization, location tracking and pose registration. The first two modules fully exploit the deep-learning based three-dimensional point cloud learning techniques to achieve the map-less global localization task in large-scale environment. The location tracking module implements the particle filter framework with a newly designed perception model to track the vehicle location during movements. Finally, the pose registration module uses visual information to exclude the influence of dynamic obstacles and short-term changes and further introduces point cloud registration network to estimate the accurate vehicle pose.FindingsComprehensive experiments in real industrial environments demonstrate the effectiveness, robustness and practical applicability of the map-less localization approach.Practical implicationsThis paper provides comprehensive experiments in real industrial environments.Originality/valueThe system can be used in the practical automated industrial vehicles for long-term localization tasks. The dynamic objects, short-/long-term environment changes and hardware limitations of industrial vehicles are all considered in the system design. Thus, this work moves a big step toward achieving real implementations of the autonomous localization in practical industrial scenarios.

2D CNN-Based Multi-Output Diagnosis for Compound Bearing Faults under Variable Rotational Speeds

Bearings prevent damage caused by frictional forces between parts supporting the rotation and they keep rotating shafts in their correct position. However, the continuity of work under harsh conditions leads to inevitable bearing failure. Thus, methods for bearing fault diagnosis (FD) that can predict and categorize fault type, as well as the level of degradation, are increasingly necessary for factories. Owing to the advent of deep neural networks, especially convolutional neural networks (CNNs), intelligent FD methods have achieved significantly higher performance in terms of accuracy. However, in addition to accuracy, the efficiency issue still needs to be weathered in complicated diagnosis scenarios to adapt to real industrial environments. Here, we introduce a method based on multi-output classification, which utilizes the correlated features extracted for bearing compound fault type classification and crack-size classification to serve both aims. Additionally, the synergy of a time–frequency signal processing method and the proposed two-dimensional CNN helped the method perform well under the condition of variable rotational speeds. Monitoring signals of acoustic emission also had advantages for incipient FD. The experimental results indicated that utilizing correlated features in multi-output classification improved both the accuracy and efficiency of multi-task diagnosis compared to conventional CNN-based multiclass classification.

Development of an accelerated test for pitting corrosion of CUSTOM 450 by rDHM

Abstract A key application of CUSTOM 450 alloy is in the construction of gas turbine compressor blades. The study of pitting corrosion can prevent the failure of many gas turbine compressor blades. In this study, a reflective digital holography microscopy method was employed to investigate the growth of pitting corrosion in depth. To this end, a constant potential of 350 mVSCE in a 3.5 wt.% NaCl solution was applied to the specimen. The generated pits were simulated in three dimensions, and it was indicated that pitting corrosion rate was decreased as time passed. Comparing the obtained experimental data with the data gathered from the real industrial environment surrounding a compressor installation, an accelerated test was proposed. By the proposed accelerated test, it is possible to produce a pit similar to the one that will be initiated and propagated at any time in the future in real conditions.

Empirical validation of a quality framework for evaluating modelling languages in MDE environments

In previous research, we proposed the multiple modelling quality evaluation framework (MMQEF), which is a method and tool for evaluating modelling languages in model-driven engineering (MDE) environments. Rather than being exclusive, MMQEF attempts to complement other methods of evaluation of quality such as SEQUAL. However, to date, MMQEF has not been validated beyond some concept proofs. This paper evaluates the applicability of the MMQEF method in comparison with other existing methods. We performed an evaluation in which the subjects had to detect quality issues in modelling languages. A group of expert professionals and two experimental objects (i.e. two combinations of different modelling languages based on real industrial practices) were used. To analyse the results, we applied quantitative approaches, i.e. statistical tests on the results of the performance measures and the perception of subjects. We ran four replications of the experiment in Colombia between 2016 and 2019, with a total of 50 professionals. The results of the quantitative analysis show a low performance for all of the methods, but a positive perception of MMQEF.Conclusions: The application of modelling language quality evaluation methods within MDE settings is indeed tricky, and subjects did not succeed in identifying all quality problems. This experiment paves the way for additional investigation on the trade-offs between the methods and potential situational guidelines (i.e. circumstances under which each method is convenient). We encourage further inquiries on industrial applications to incrementally improve the method and tailor it to the needs of professionals working in real industrial environments.

Numerical study of the square cup stamping process: a stochastic analysis

The industrial demand for products with better quality and lower production costs have encouraged the widespread application of the finite element analysis (FEA) in the development and optimization of sheet metal forming processes. To ensure that the FEA solutions are reliable and robust it is important to take into account the uncertainties that inevitably arise in a real industrial environment. In this context, a numerical study on the influence of the material and process uncertainty in the stamping results of a square cup is presented. In this analysis, it is assumed uncertainty in the elasticity properties, hardening law parameters, anisotropy coefficients, blank thickness, friction coefficient and in the blank holder force. The effect of the uncertainty in these input parameters is evaluated in the punch force, equivalent plastic strain, thickness and cup geometry. Firstly, quasi-Monte Carlo method was used to evaluate the variability in the simulation outputs, considering the uncertainty of the input parameters. This analysis shows that the geometry is the output most sensitive to the uncertainty of the input parameters. Afterwards, a variance-based sensitivity analysis was carried out to identify the input parameters that most influence the output variability. It was concluded that the hardening law parameters and the anisotropy coefficients have the most influence in the stamping results variability of a square cup.

Optimisation of mixed-model assembly line balancing problem under uncertain demand

Assembly lines play a crucial role in determining the profitability of a company. Market conditions have increased the importance of mixed-model assembly lines. Variations in the demand are frequent in real industrial environments and often leads to failure of the mixed-model assembly line balancing scheme. Decision makers have to take into account this uncertainty. In an assembly line balancing problem, there is a massive amount of research in the literature assuming deterministic environment, and many other works consider uncertain task times. This research utilises the uncertainty theory to model uncertain demand and introduces complexity theory to measure the uncertainty of assembly lines. Scenario probability and triangular fuzzy number are used to describe the uncertain demand. The station complexity was measured based on information entropy and fuzzy entropy to assist in balancing systems with robust performances, considering the influence of multi-model products in the station on the assembly line. Taking minimum station complexity, minimum workload difference within station, maximum productivity as objective functions, a new optimization model for mixed-model assembly line balancing under uncertain demand was established. Then an improved genetic algorithm was applied to solve the model. Finally, the effectiveness of the model was verified by several instances of mixed-model assembly line for automobile engine.

The smart SME technology readiness assessment methodology in the context of industry 4.0

PurposeThis study proposes the Smart SME Technology Readiness Assessment (SSTRA) methodology which aims to enable practitioners to assess the SMEs Industry 4.0 technology readiness throughout the end-to-end engineering across the entire value chain; the smart product design phase is the focus in this paper.Design/methodology/approachThe proposed SSTRA utilises the analytic hierarchy process to prioritise smart SME requirements, a graphical interface which tracks technologies' benchmarks under Industry 4.0 Technology Readiness Levels (TRLs); a mathematical model used to determine the technology readiness and visual representation to understand the relative readiness of each smart main area. The validity of the SSTRA is confirmed by testing it in a real industrial environment. In addition, the conceptual model for Smart product design development is proposed and validated.FindingsThe proposed SSTRA offers decision-makers the facility to identify requirements and rank them to reflect the current priorities of the enterprise. It allows SMEs to assess their current capabilities in a range of technologies of high relevance to the Industry 4.0 area. The SSTRA assembles a readiness profile allowing decision-makers to not only perceive the overall score of technology readiness but also the distribution of technology readiness across the main smart areas. It helps to visualise strengths and weaknesses; whilst emphasising the fundamental gaps that require serious action to assist the program with a well-balanced effort towards a successful transition to Industry 4.0.Originality/valueThe SSTRA provides a step-by-step approach for decision-making based on data collection, analysis, visualisation and documentation. Hence, it greatly mitigates the risk of further Industry 4.0 technology investment and implementation.

An integrated approach for the HCl and metals recovery from waste pickling solutions: pilot plant and design operations

Continuous regeneration of industrial pickling solutions and recovery of valuable materials are implemented in a pilot-scale plant including diffusion dialysis (DD), where HCl is recovered, membrane distillation (MD), where HCl is concentrated, and reactive precipitation (CSTR), where metal ions are recovered in different forms. The integration of the three processes allows to minimize waste streams generation and to accomplish a closed-loop process, thus increasing the environmental sustainability and economic impact of the galvanizing industry. Process reliability was proved through the operation of a demonstrator in the real industrial environment of the Tecnozinco SrL hot-dip galvanizing plant (Carini, Italy), assessing the actual performance in fully reducing spent pickling solution disposal and recovering useful compounds. Tests were conducted firstly with synthetic solutions and then with real waste liquors from the pickling plant. A high acid recovery (80%) can be achieved in the diffusion dialysis unit and quantitative metals separation was accomplished, with iron hydroxide produced at 99% purity. The membrane distillation performance suffers when metal salts are present in large quantities, due to the “salting out” effect, resulting in reduced water vapor pressure, though the use of available low grade waste heat allows energy-sustainable operation of the MD.

Additive Manufacturing Processing of Plastics for Mass Production of Composites Tooling: Technical and Economic Analysis

AbstractThe present paper focuses on the use of several additive manufacturing (AM) techniques for tooling manufacturing for composite parts production. A geometry that prevents mandrel removal is selected and it is analyzed in a real industrial environment for composites production using autoclave molding. The use of two different AM mandrels is evaluated: fused deposition modeling and liquid crystal display. The results are analyzed in terms of composites part quality, processing time and final part cost.

Iron Carbon Masteralloy as Sintering Activator for Low Alloy Powder Metallurgy Steels

In real industrial environment there is always a difference between ideal theoretical condition and real production condition which bears the risk of producing defective or low quality parts. Getting closer to this ideal situation requires more effort and investment which tends to increase the production cost. In the P/M production lines, the sintering stage is one of the most critical processes. Maintaining an open continuous sintering furnace in an ideal condition is a challenge, and this issue gets more pronounced when using alloy powder containing oxygen-sensitive elements such as Cr or Mn which provide good hardenability at low cost but on the other hand form stable oxides that weaken the sintering contacts if they are not reduced properly. In the present study, using a carbon master alloy as a sintering enhancer in the sintering process of Cr-Mo alloyed powder compacts has been investigated. For clearly depicting the effect of carbon master alloy addition on carbon dissolution and deoxidation, sintering was done in argon as inert atmosphere to avoid other reducing agents such as H2. The physical and mechanical properties of the sintered specimens were investigated, and thermal chemical analysis by DIL/MS and carbon/oxygen measurements were performed. The experiments showed that adding iron-carbon masteralloys promote the sintering processes such as reduction of oxides and carbon dissolution in the early stages of sintering, resulting in better properties after final sintering.

A Systematic Review of Deep Learning for Silicon Wafer Defect Recognition

Advancements in technology have made deep learning a hot research area, and we see its applications in various fields. Its widespread use in silicon wafer defect recognition is replacing traditional machine learning and image processing methods of defect monitoring. This article presents a review of the deep learning methods employed for wafer map defect recognition. A systematic literature review (SLR) has been conducted to determine how the semiconductor industry is leveraged by deep learning research advancements for wafer defects recognition and analysis. Forty-four articles from well-known databases have been selected for this review. The articles’ detailed study identified the prominent deep learning algorithms and network architectures for wafer map defect classification, clustering, feature extraction, and data synthesis. The identified learning algorithms are grouped as supervised learning, unsupervised learning, and hybrid learning. The network architectures include different forms of Convolutional Neural Network (CNN), Generative Adversarial Network (GAN), and Auto-encoder (AE). Various issues of multi-class and multi-label defects have been addressed, solving data unavailability, class imbalance, instance labeling, and unknown defects. For future directions, it is recommended to invest more efforts in the accuracy of the data generation procedures and the defect pattern recognition frameworks for defect monitoring in real industrial environments.

A micro-service-based machinery monitoring solution towards realizing the Industry 4.0 vision in a real environment

Abstract This work presents a modular Smart Maintenance Platform (SMP), which offers a toolbox of condition monitoring tools for the day-to-day inspection of the machinery equipment in a real industrial environment. SMP complements the manual inspection performed by the maintenance engineers with advanced data-analytics solutions that utilize both reactive and proactive techniques based on state-of-the-art machine learning proposals. Appropriate data acquisition mechanisms are built to acquire the sensorial data and transform them to suitable forms for data processing. Higher-level monitoring policies are enabled, through the fusion of multiple monitoring tasks increasing the precision of the results. The generated results are persisted for the continuous enhancement of the solution through the retraining of the proactive approach or the re-configuration of the reactive ones. The generated results are visualized in a user-friendly format, to help the maintenance engineers to assess the severity of the situation and to intervene in the machinery equipment if deemed necessary.

How to Survive Identity Management in the Industry 4.0 Era

Industry 4.0 heavily builds on massive deployment of Industrial Internet of Things (IIoT) devices to monitor every aspect of the manufacturing processes. Since the data gathered by these devices impact the output of critical processes, identity management and communications security are critical aspects, which commonly rely on the deployment of X.509 certificates. Nevertheless, the provisioning and management of individual certificates for a high number of IIoT devices involves important challenges. In this paper, we present a solution to improve the management of digital certificates in IIoT environments, which relies on partially delegating the certificate enrolment process to an edge server. However, in order to preserve end-to-end security, private keys are never delegated. Additionally, for the protection of the communications between the edge server and the IIoT devices, an approach based on Identity Based Cryptography is deployed. The proposed solution considers also the issuance of very short-lived certificates, which reduces the risk of using expired or compromised certificates, and avoids the necessity of implementing performance expensive protocols such as Online Certificate Status Protocol (OCSP). The proposed solution has been successfully tested as an efficient identity management solution for IIoT environments in a real industrial environment.