Reinforcement learning (RL) is still rarely applied in industrial control, partly due to the difficulty of training reliable agents for real-world conditions. This work investigates how evolution strategies can support RL in such settings by introducing a continuous-control adaptation of an industrial sorting benchmark. The CMA-ES algorithm is used to generate high-quality demonstrations that warm-start RL agents. Results show that CMA-ES-guided initialization significantly improves stability and performance. Furthermore, the demonstration trajectories generated with the CMA-ES provide a strong oracle reference performance level, which is of interest in its own right. The study delivers a focused proof of concept for hybrid evolutionary-RL approaches and a basis for future, more complex industrial applications.
Lilian Dias Pereira, Irenilza de Alencar Nääs, Vando Aparecido Monteiro
et al.
This study presents a quantitative, cross-sectional analysis of work-related musculoskeletal disorders (WRMSDs) among sick leave recipients in Brazil’s meat production chain, using official surveillance data. A marked temporal shift was observed; women remained more affected by upper limb injuries, such as shoulder and wrist disorders. In 2022, male notifications surpassed female ones, marking a turning point linked to improved reporting and the inclusion of WRMSDs in Brazil’s compulsory notification list. Workers aged 20–49 were the most impacted group, with diagnoses including shoulder lesions, tenosynovitis, carpal tunnel syndrome, back pain, and occupational risk exposure. The findings highlight systemic barriers, including underreporting, inadequate protection, and weak return-to-work protocols. Implementing gender-differentiated ergonomic protocols is crucial, as it requires reducing repetitive strain for women in line-feeding/cutting roles, and mitigating environmental hazards (such as cold, vibration, and chemical exposure) for men in farming/slaughtering. These results underscore the urgent need for gender-sensitive preventive strategies and occupational health policies tailored to the meat processing industry.
Industrial safety. Industrial accident prevention, Medicine (General)
Digital Twins (DTs) are virtual representations of physical systems synchronized in real time through Internet of Things (IoT) sensors and computational models. In industrial applications, DTs enable predictive maintenance, fault diagnosis, and process optimization. This paper explores the mathematical foundations of DTs, hybrid modeling techniques, including Physics Informed Neural Networks (PINNs), and their implementation in industrial scenarios. We present key applications, computational tools, and future research directions.
Managing delay is one of the core requirements of industrial automation applications due to the high risk associated for equipment and human lives. Using efficient Media Access Control (MAC) schemes guarantees the timely transmission of critical data, particularly in the industrial environments where heterogeneous data is inherently expected. This paper compares the performance of Fragmentation based MAC (FROG-MAC) against Fuzzy Priority Scheduling based MAC (FPS-MAC), both of which have been designed to optimize the performance of heterogenous wireless networks. Contiki has been used as a simulation platform and a single hop star topology has been assumed to resemble the industrial environment. It has been shown that FROG-MAC has the potential to outperform FPS-MAC in terms of energy efficiency and delay both, due to its inherent feature of interrupting ongoing lower priority transmission on the channel.
This document describes the development and implementation of a technological solution based on IoT devices to modernize a machine known as the Cyclone. This equipment is used by a contractor collaborating with petrochemical companies in the state of Texas, performing specialized work in mechanics, engineering, catalytic material replacement, and rescue operations in refinery complexes. The Cyclone machine, with outdated relay logic technology, poses challenges in terms of operational efficiency, critical condition monitoring, and safety. The project was carried out with the collaboration of specialists in equipment handling, focusing on demonstrating the feasibility of integrating advanced Industry 4.0 technologies into legacy industrial equipment. The methodology included the incorporation of IoT sensors for real-time monitoring, an automated control system, and the digitization of key processes. Preliminary results indicate improvements in the precision of operational control and the ability for remote supervision, highlighting the potential for modernization in critical industrial applications. This work not only validates the use of IoT devices in obsolete equipment but also sets a precedent for the transition towards more sustainable and efficient technologies in the petrochemical sector.
Prodromos D. Chatzoglou, Athanasios E. Kotzakolios, Panagiotis K. Marhavilas
The main aim of this study is to investigate the association of an assortment of factors pertaining to the implementation of a Health and Safety Management System (HSMS) with firms’ Health and Safety (H&S) level and employees’ satisfaction and performance. The proposed research model incorporates six independent factors: (i) the development of a safety culture, (ii) the availability of H&S specific procedures/instructions/rules, (iii) the forethought of H&S-focused training, (iv) the availability of the essential resources to improve H&S equipment, (v) the augmentation of employee motivation for safe behavior, and (vi) the top management commitment to upgrade H&S at the workplace. The model was empirically tested using primary data from 230 employees of 10 manufacturing firms operating in Greece. It was found that H&S meliorates employees’ performance, but, on the other hand, firm’s management should be committed towards creating a high-level safety culture. To achieve this, proper resources should be invested, comprehensible procedures/instructions/rules should be established, and focused training should be provided. The acceptance of this policy would result in an enhanced safety culture, an augmented firm H&S level, amplified employee satisfaction and, accordingly, improved employee performance. In addition, this article suggests a new HSMS model, which, by relying on the principles of the Plan–Do–Check–Act cycle, incorporates the examined six H&S factors, which can upgrade other known standards (like OHSAS 18001 and ISO 45001).
Industrial safety. Industrial accident prevention, Medicine (General)
In this pilot study, the crossing behavior of elementary school students commuting on bicycles was investigated with the objective of enhancing safety around pedestrian crossings within school zones. With a noticeable increase in crashes involving young cyclists near schools, this research assessed the effectiveness of visual nudges in the form of red strips displaying “CYCLISTS DISMOUNT” instructions. Initial observations indicated a lack of compliance with dismounting regulations. After the initial observations, a specific elementary school was selected for the implementation of the nudging intervention and additional pre- (N = 91) and post-intervention (N = 71) observations. The pre-intervention observations again revealed poor adherence to the regulations requiring cyclists to dismount at specific points. Following our targeted intervention, the post-intervention observations marked an improvement in compliance. Indeed, the visual nudge effectively communicated the necessity of dismounting at a critical location, leading to a higher rate of adherence among cyclists (52.74% pre-intervention, 97.18% post-intervention). Although it also indirectly affected the behavior of the accompanying adult, who more often held hands with their children while crossing, this effect was weaker than the direct effect on dismounting behavior (20.88% pre-intervention, 39.44% post-intervention). The findings of the current pilot study underscore the possible impact of nudging on behavior and advocate for a combined approach utilizing physical nudges to bolster safety within school zones. Follow-up research, including, for instance, multiple sites, long-term effects, or children traveling alone, is called for.
Industrial safety. Industrial accident prevention, Medicine (General)
The study develops a mathematical system that encompasses the theory of reliability, risk management principles, and predictive analytics to assess the impact of digital twins on equipment reliability and personnel safety. The model incorporates key parameters, including the intensity of failures, recovery rates, and efficiency factors, which quantitatively determine the impact of a digital twin on hazard detection and elimination. The proposed model formalizes the interconnection between the implementation of digital twins and the outcomes in the sphere of safety through several interconnected components, which are equipment reliability functions considering both baseline failure rates and those enhanced by digital twins, integrated safety indices uniting the technical reliability with human factor assessment, and optimization functions for the comprehensive cost-benefit analysis of digital twin deployment. The key results have demonstrated that digital twins significantly reduce effective failure intensity due to the early detection of malfunctions and possible preventive maintenance. The improvements of the system’s reliability can be quantitatively assessed and show the measurable risk reduction in industrial operations. The model also considers the improvement of maintenance efficiency through the accelerated detection of failures, improved accuracy of diagnostics, and optimized repair processes. The study introduces the utility function that balances multiple factors, which are the expenses on accident prevention, implementation expenses, increasing operational efficiency, and general safety improvement. This helps organizations to optimize digital twin strategy deployment based on empirical data. Dynamic Bayesian approaches have been proposed for continuous parameter update in real-time mode based on the accumulated operational data. The practical application encompasses predictive maintenance software, hazard virtual modeling for training purposes, safety monitoring systems in real time, and risk assessment protocols. These applications are specifically valuable for high-risk industries, i.e., oil and gas, chemical, and heavy industries. The model provides top managers with quantitative tools to prioritize investments into digital twins based on the demonstrated impact on safety and cost efficiency, supporting the evidence-based safety management strategies in the Industry 4.0 environment.
Hans Aoyang Zhou, Dominik Wolfschläger, Constantinos Florides
et al.
Machine vision enhances automation, quality control, and operational efficiency in industrial applications by enabling machines to interpret and act on visual data. While traditional computer vision algorithms and approaches remain widely utilized, machine learning has become pivotal in current research activities. In particular, generative AI demonstrates promising potential by improving pattern recognition capabilities, through data augmentation, increasing image resolution, and identifying anomalies for quality control. However, the application of generative AI in machine vision is still in its early stages due to challenges in data diversity, computational requirements, and the necessity for robust validation methods. A comprehensive literature review is essential to understand the current state of generative AI in industrial machine vision, focusing on recent advancements, applications, and research trends. Thus, a literature review based on the PRISMA guidelines was conducted, analyzing over 1,200 papers on generative AI in industrial machine vision. Our findings reveal various patterns in current research, with the primary use of generative AI being data augmentation, for machine vision tasks such as classification and object detection. Furthermore, we gather a collection of application challenges together with data requirements to enable a successful application of generative AI in industrial machine vision. This overview aims to provide researchers with insights into the different areas and applications within current research, highlighting significant advancements and identifying opportunities for future work.
Thomas Rosenstatter, Christian Schäfer, Olaf Saßnick
et al.
As Industry 4.0 and the Industrial Internet of Things continue to advance, industrial control systems are increasingly adopting IT solutions, including communication standards and protocols. As these systems become more decentralized and interconnected, a critical need for enhanced security measures arises. Threat modeling is traditionally performed in structured brainstorming sessions involving domain and security experts. Such sessions, however, often fail to provide an exhaustive identification of assets and interfaces due to the lack of a systematic approach. This is a major issue, as it leads to poor threat modeling, resulting in insufficient mitigation strategies and, lastly, a flawed security architecture. We propose a method for the analysis of assets in industrial systems, with special focus on physical threats. Inspired by the ISO/OSI reference model, a systematic approach is introduced to help identify and classify asset interfaces. This results in an enriched system model of the asset, offering a comprehensive overview visually represented as an interface tree, thereby laying the foundation for subsequent threat modeling steps. To demonstrate the proposed method, the results of its application to a programmable logic controller (PLC) are presented. In support of this, a study involving a group of 12 security experts was conducted. Additionally, the study offers valuable insights into the experts' general perspectives and workflows on threat modeling.
Graph-based patterns are extensively employed and favored by practitioners within industrial companies due to their capacity to represent the behavioral attributes and topological relationships among users, thereby offering enhanced interpretability in comparison to black-box models commonly utilized for classification and recognition tasks. For instance, within the scenario of transaction risk management, a graph pattern that is characteristic of a particular risk category can be readily employed to discern transactions fraught with risk, delineate networks of criminal activity, or investigate the methodologies employed by fraudsters. Nonetheless, graph data in industrial settings is often characterized by its massive scale, encompassing data sets with millions or even billions of nodes, making the manual extraction of graph patterns not only labor-intensive but also necessitating specialized knowledge in particular domains of risk. Moreover, existing methodologies for mining graph patterns encounter significant obstacles when tasked with analyzing large-scale attributed graphs. In this work, we introduce GraphRPM, an industry-purpose parallel and distributed risk pattern mining framework on large attributed graphs. The framework incorporates a novel edge-involved graph isomorphism network alongside optimized operations for parallel graph computation, which collectively contribute to a considerable reduction in computational complexity and resource expenditure. Moreover, the intelligent filtration of efficacious risky graph patterns is facilitated by the proposed evaluation metrics. Comprehensive experimental evaluations conducted on real-world datasets of varying sizes substantiate the capability of GraphRPM to adeptly address the challenges inherent in mining patterns from large-scale industrial attributed graphs, thereby underscoring its substantial value for industrial deployment.
Carlos Rafael Silva de Oliveira, Catia Rosana Lange de Aguiar, Maria Elisa Philippsen Missner
et al.
Textile chemistry and textile processing laboratories are essential environments for textile product research and development, but they also pose hazards that require rigorous precautions. Among the most common risks is handling chemicals used in the textile industry, such as dyes, solvents, and finishing chemicals, which can be contaminants, corrosive, and flammable, presenting risks of poisoning and fire. Textile processing laboratories also require proper ventilation, as a lack of appropriate ventilation in these environments can accumulate toxic vapors in the air. The most relevant risks and hazards of using textile chemistry laboratories include using equipment such as dyeing autoclaves under pressure and high temperature; drying ovens like furnaces/lab stenters; cylinders of squeezing, calenders, and others, capable of causing severe accidents. These laboratories also generate or handle solid waste and effluents containing, heavy metals to pathogens (e.g., from industrial sludge). It is essential to adopt rigorous safety measures in textile chemistry laboratories, including using personal protective equipment (PPE), proper training of workers, effective ventilation systems, and safe waste disposal protocols. Good laboratory work practices not only reduce risk but also promote better research; more accurate results; and better data. Therefore, this study aimed to map the risks and hazards of textile processing laboratories with a view to accident prevention and formalizing a protocol for good practices.
Accident detection and traffic analysis is a critical component of smart city and autonomous transportation systems that can reduce accident frequency, severity and improve overall traffic management. This paper presents a comprehensive analysis of traffic accidents in different regions across the United States using data from the National Highway Traffic Safety Administration (NHTSA) Crash Report Sampling System (CRSS). To address the challenges of accident detection and traffic analysis, this paper proposes a framework that uses traffic surveillance cameras and action recognition systems to detect and respond to traffic accidents spontaneously. Integrating the proposed framework with emergency services will harness the power of traffic cameras and machine learning algorithms to create an efficient solution for responding to traffic accidents and reducing human errors. Advanced intelligence technologies, such as the proposed accident detection systems in smart cities, will improve traffic management and traffic accident severity. Overall, this study provides valuable insights into traffic accidents in the US and presents a practical solution to enhance the safety and efficiency of transportation systems.
Industrial Cyber-Physical Systems (ICPS) integrate the disciplines of computer science, communication technology, and engineering, and have emerged as integral components of contemporary manufacturing and industries. However, ICPS encounters various challenges in long-term operation, including equipment failures, performance degradation, and security threats. To achieve efficient maintenance and management, prognostics and health management (PHM) finds widespread application in ICPS for critical tasks, including failure prediction, health monitoring, and maintenance decision-making. The emergence of large-scale foundation models (LFMs) like BERT and GPT signifies a significant advancement in AI technology, and ChatGPT stands as a remarkable accomplishment within this research paradigm, harboring potential for General Artificial Intelligence. Considering the ongoing enhancement in data acquisition technology and data processing capability, LFMs are anticipated to assume a crucial role in the PHM domain of ICPS. However, at present, a consensus is lacking regarding the application of LFMs to PHM in ICPS, necessitating systematic reviews and roadmaps to elucidate future directions. To bridge this gap, this paper elucidates the key components and recent advances in the underlying model.A comprehensive examination and comprehension of the latest advances in grand modeling for PHM in ICPS can offer valuable references for decision makers and researchers in the industrial field while facilitating further enhancements in the reliability, availability, and safety of ICPS.
Guido Alfaro Degan, Andrea Antonucci, Dario Lippiello
The ISO Standard 10819:2013 defines the method for evaluating the performances of antivibration (AV) gloves, but when used in real fields, the protection can be dissimilar to that labeled. This paper investigates the transmissibility, at the palm level, of three different types of AV gloves (air, gel, neoprene) and an ordinary leather glove, during the use of four similar electric hammers (average weight of 10 kg, and average impact energy of 18 J), in a limestone quarry plant. As the average triaxial transmissibility for all the hammers, results show very limited benefits in reducing the vibration (6%), with no significative differences among the different gloves. The working leather glove, instead, shows a transmissibility quite equal to the unit. Anyway, results can be different for the same glove when used among the different hammers, providing in some cases 19% of protection. Some differences can be found regarding the transmissibility through the three main axes for the same type of glove: the glove in gel seems to perform better in shear than in compression. The transmissibility in compression is around 20% higher than that provided by the manufacturers of the certified gloves. The usage of specific excitation curves during laboratory tests could help in providing a more accurate estimation of the transmissibility of the gloves when used with a specific tool.
Industrial safety. Industrial accident prevention, Medicine (General)
У роботі розглянуті наявні підходи впливу на роботу алгоритмів штучного інтелекту, зокрема машинного навчання, що застосовуються в системах комп’ютерного зору для виявлення, класифікації та ідентифікації об’єктів. На даний час найпопулярнішою та найперспективнішою технологією розпізнавання образів є штучні нейронні мережі. Комп’ютерний зір застосовується у військовій справі для виявлення візуальних об’єктів певних класів: людей, озброєння та військової техніки, військових об’єктів тощо. Вхідними даними для аналізу можуть бути: фотографії, відеокадри чи відео потік реального часу, що отримані з космічних, повітряних або наземних засобів розвідки. Для боротьби з системами автоматичного виявлення об’єктів можливо застосовувати підходи, що здатні впливати на моделі машинного навчання, які використовуються у цих системах. Атака на моделі машинного навчання – це спеціальні дії щодо впливу на її елементи з метою досягти бажаної поведінки системи або перешкодити її коректній роботі. За результатами аналізу досліджень різних авторів визначено, що майже кожен алгоритм машинного навчання має певні вразливості. Під час виконання завдань інженерної підтримки військ щодо маскування військових об’єктів, найбільш доступними способами впливу на системи комп’ютерного зору, для введення їх в оману, є зміна фізичних властивостей об’єкта, що маскується, шляхом нанесення на його поверхню спеціальних покриттів і матеріалів. У якості покриттів можливо використовувати згенеровані змагальні патч-зображення, шляхом накладання або наклеювання їх на об’єкт та які здатні вносити завади в роботу алгоритмів засобу розвідки, прицілювання або наведення. Це особливо важливо в перспективі створення автономних систем зброї, які здатні виявляти, ідентифікувати цілі та самостійно приймати рішення на їх ураження.
Monica Neagu, Fabia Grisi, Alfio Pulvirenti
et al.
Aerogels have recently started to be considered as “advanced materials”; therefore, as a general consideration, aerogels’ toxicity testing should focus on their functionality which resides in their nanoscale open internal porosity. To assess the hazards of organic aerogels, testing at three levels may characterize their biophysical, in vitro and in vivo toxicity, defining distinct categories of aerogels. At the first level of testing, their abiotic characteristics are investigated, and the best aerogel(s) is forwarded to be tested at level 2, wherein in vitro methodologies may mainly evaluate the aerogels’ cellular behavior. Within level 2 of testing, the main characteristics of toxicity are investigated and the selected aerogels are introduced to in vivo animal models at level 3. In the animal model testing, target organs are investigated along with systemic parameters of toxicity. Some study cases are presented for organic or anorganic aerogels. Within this tiered workflow, aerogels-based materials can be tested in terms of human health hazard.
Industrial safety. Industrial accident prevention, Medicine (General)
The duty on the part of manufacturers to incorporate features in equipment for the sole purpose of protecting workers and bystanders from injury grew out of the advent of worker’s compensation legislation in the second decade of the twentieth century. The new legal landscape suddenly made safe design an issue that impacted the bottom line of industrial employers through reduced insurance premiums. This was the impetus for the newly formed National Safety Council in 1913, which drew its members primarily from industry. The real sea change, however, can be traced to a paper published in the ASME Journal two years later by Carl Hansen, titled “Standardization of Safety Principles”[1]. In it, Hansen proposed the novel idea that it was the responsibility of design engineers to address the hazards their machines presented. His proposal was not made in a vacuum. The responses of leading figures in engineering, the insurance industry, and worker safety were published in the Journal as well, and the unanimous consensus was agreement with Hansen’s new ethic. In Hansen’s proposal can be found almost all of the basic concepts of the Safety Hierarchy as it was presented in the 1955 edition of the National Safety Council’s Accident Prevention Manual for Industrial Operations [2]. This states that a manufacturer has a duty to evaluate and address foreseeable hazards, including those related to human error, by design first and foremost. Only if no feasible design solutions can be found may the manufacturer rely on methods that control exposure of the hazard to users and bystanders, such as guarding. Only if no feasible design or exposure-control solutions can be found may the manufacturer rely on personal protective equipment. Since that time the hierarchy, only slightly modified, has penetrated every industry and has come to define the modern basis of safe design in the United States. In 2000, a lobbying group called the Association for Manufacturing Technology sponsored a technical paper which resulted in the promulgation of ANSI B11.0-2020 Safety of Machinery – General Requirements and Risk Assessment [3]. This standard represents a radical departure from established safety principles in that it inserts an undefined process by which manufacturers or purchasers can unilaterally decide the risk associated with any given piece of machinery is acceptable, and thereby opt out of the requirements of the safety hierarchy. The threshold of ‘acceptable risk’ is left uncodified and can be calculated based on a host of factors as vague as ‘culture’ and the ‘context of their own circumstances’. The effect of this approach on the stakeholders who would suffer the most dire consequences, the injured workers, is not considered. I argue that ANSI B11.0-2020 and its related standards are retrograde in their effect on safety in the workplace, and cannot be reconciled with the last 108 years of safe design principles as developed in the United States. This paper will provide a historical review of safe engineering design principles and analyze the provisions and implications of ANSI B11.0-2020. The basis for a revised ANSI B11.0 will be presented.
The risk of coal mine accidents rises significantly with mining depth, making it urgent for accident prevention to be supported by both scientific analysis and advanced technologies. Hence, a comprehensive grasp of the research progress and differences in hotspots of coal mine accidents in China serves as a guide to find the shortcomings of studies in the field, promote the effectiveness of coal mine disaster management, and enhance the prevention and control ability of coal mine accidents. This paper analyzes Chinese and foreign literature based on data mining algorithms (LSI + Apriori), and the findings indicate that: (1) 99% of the available achievements are published in Chinese or English-language journals, with the research history conforming to the stage of Chinese coal industry development, which is characterized by “statistical description, risk evaluation, mechanism research, and intelligent reasoning”. (2) Chinese authors are the primary contributors that lead and contribute to the continued development of coal mine accident research in China globally. Over 81% of the authors and over 60% of the new authors annually are from China. (3) The emphasis of the Chinese and English studies is different. Specifically, the Chinese studies focus on the analysis of accident patterns and causes at the macroscale, while the English studies concentrate on the occupational injuries of miners at the small-scale and the mechanism of typical coal mine disasters (gas and coal spontaneous combustion). (4) The research process in Chinese is generally later than that in English due to the joint influence of the target audience, industrial policy, and scientific research evaluation system. After 2018, the Chinese studies focus significantly on AI technology in deep mining regarding accident rules, regional variation analysis, risk monitoring and early warning, as well as knowledge intelligence services, while the hotspots of English studies remain unchanged. Furthermore, both Chinese and English studies around 2019 focus on “public opinion”, with Chinese ones focusing on serving the government to guide the correct direction of public opinion while English studies focus on critical research of news authenticity and China’s safety strategy.