Hasil untuk "Industrial directories"

Deng Chenggang, Li Mingfan

To address the issue of low recognition accuracy of yarn tube types in practical industrial scenarios, this study proposes a SimAM-ResNet18-based image recognition method for cone yarns. Different from the traditional yarn recognition method based on Resnet, the framework introduced in this study combines parameter free attention and swish activation to improve the recognition accuracy and robustness under industrial conditions. First, a high-resolution image acquisition system was designed and implemented. The acquired images were preprocessed using bilateral filtering, Gamma correction, HSI color space extraction, and rapid template matching of edge points to enhance image features. Then, the Swish activation function and SimAM attention mechanism were integrated into the ResNet18 network, effectively improving the model's focus on key regions and its feature representation capabilities. On a dataset composed of 1800 real-world images collected from a textile production line, the proposed model achieved a recognition accuracy of 98.3%, a precision of 0.969, a recall of 0.972, and an F1-score of 0.970, significantly outperforming mainstream models such as MobileNetV2, EfficientNet-B0, and SENet18. Without retraining, the model maintained an accuracy of 92.8% under challenging conditions such as angle variation and illumination changes, demonstrating strong generalization capability and practical industrial value.

Francesco Ragusa, Rosario Leonardi, Michele Mazzamuto et al.

Understanding human behavior from complementary egocentric (ego) and exocentric (exo) points of view enables the development of systems that can support workers in industrial environments and enhance their safety. However, progress in this area is hindered by the lack of datasets capturing both views in realistic industrial scenarios. To address this gap, we propose ENIGMA-360, a new ego-exo dataset acquired in a real industrial scenario. The dataset is composed of 180 egocentric and 180 exocentric procedural videos temporally synchronized offering complementary information of the same scene. The 360 videos have been labeled with temporal and spatial annotations, enabling the study of different aspects of human behavior in industrial domain. We provide baseline experiments for 3 foundational tasks for human behavior understanding: 1) Temporal Action Segmentation, 2) Keystep Recognition and 3) Egocentric Human-Object Interaction Detection, showing the limits of state-of-the-art approaches on this challenging scenario. These results highlight the need for new models capable of robust ego-exo understanding in real-world environments. We publicly release the dataset and its annotations at https://fpv-iplab.github.io/ENIGMA-360/.

Annemarie Jutte, Uraz Odyurt

Industrial Cyber-Physical Systems (CPS) are sensitive infrastructure from both safety and economics perspectives, making their reliability critically important. Machine Learning (ML), specifically deep learning, is increasingly integrated in industrial CPS, but the inherent complexity of ML models results in non-transparent operation. Rigorous evaluation is needed to prevent models from exhibiting unexpected behaviour on future, unseen data. Explainable AI (XAI) can be used to uncover model reasoning, allowing a more extensive analysis of behaviour. We apply XAI to improve predictive performance of ML models intended for an industrial CPS use-case. We analyse the effects of components from time-series data decomposition on model predictions using SHAP values. Through this method, we observe evidence on the lack of sufficient contextual information during model training. By increasing the window size of data instances, informed by the XAI findings for this use-case, we are able to improve model performance.

Jintai Li, Songqiang Chen, Shuo Jin et al.

Code summaries are essential for helping developers understand code functionality and reducing maintenance and collaboration costs. Although recent advances in large language models (LLMs) have significantly improved automatic code summarization, the practical usefulness of generated summaries in industrial settings remains insufficiently explored. In collaboration with documentation experts from the industrial HarmonyOS project, we conducted a questionnaire study showing that over 57.4% of code summaries produced by state-of-the-art approaches were rejected due to violations of developers' expectations for industrial documentation. Beyond semantic similarity to reference summaries, developers emphasize additional requirements, including the use of appropriate domain terminology, explicit function categorization, and the avoidance of redundant implementation details. To address these expectations, we propose ExpSum, an expectation-aware code summarization approach that integrates function metadata abstraction, informative metadata filtering, context-aware domain knowledge retrieval, and constraint-driven prompting to guide LLMs in generating structured, expectation-aligned summaries. We evaluate ExpSum on the HarmonyOS project and widely used code summarization benchmarks. Experimental results show that ExpSum consistently outperforms all baselines, achieving improvements of up to 26.71% in BLEU-4 and 20.10% in ROUGE-L on HarmonyOS. Furthermore, LLM-based evaluations indicate that ExpSum-generated summaries better align with developer expectations across other projects, demonstrating its effectiveness for industrial code documentation.

Liz Li, Wei Zhu

While retrieval augmented generation (RAG) has been swiftly adopted in industrial applications based on large language models (LLMs), there is no consensus on what are the best practices for building a RAG system in terms of what are the components, how to organize these components and how to implement each component for the industrial applications, especially in the medical domain. In this work, we first carefully analyze each component of the RAG system and propose practical alternatives for each component. Then, we conduct systematic evaluations on three types of tasks, revealing the best practices for improving the RAG system and how LLM-based RAG systems make trade-offs between performance and efficiency.

Maxim Skoretskyi , Vadym Kaptur , Andriy Berezkin et al.

Мета статті. Розроблення алгоритму визначення пріоритетів засобам радіоелектронної боротьби для подавлення приймального тракту конкретного радіоелектронного об’єкту на основі визначення коефіцієнтів подібності між характеристиками засобів і характеристиками цілей. Методи дослідження. Під час написання статті застосовано метод формалізованого аналізу подібностей, що дає змогу оцінити потенційну ефективність впливу засобів радіоелектронної боротьби на приймальні тракти виявлених радіоелектронних засобів. Запропонований підхід дав змогу врахувати алгоритм попередньої фільтрації та формування матриці близькості, що ґрунтується на нормативно заданих коефіцієнтах сумісності. Отримані результати дослідження. Розроблено алгоритм визначення пріоритетів засобам радіоелектронної боротьби для подавлення конкретного радіоелектронного об’єкту, що враховує подібність тактико-технічних характеристик засобів до характеристик цілей, а також інші важливі оперативні обставини. Запропонований алгоритм дає змогу ранжувати засоби радіоелектронної боротьби для кожної цілі на основі інтегральної оцінки подібності. Алгоритм ураховує частотну сумісність, геометричні характеристики сектору дії, очікувану потужність перешкоди, а також зайнятість і придатність засобів до групової роботи. Елементи наукової новизни. Запропонований алгоритм описує новий спосіб визначення пріоритетів засобам радіоелектронної боротьби, який дає змогу адаптувати вибір засобів до динамічних змін у бойовій обстановці. Вперше застосовано метод формалізованого аналізу подібностей в частині перетворення технічних параметрів у метрику ефективності взаємодії. Теоретична й практична значущість викладеного у статті. Отримані результати мають прикладне значення для побудови адаптивних систем управління засобами радіоелектронної боротьби у зоні бойових дій. Модель може бути інтегрована в існуючі системи автоматизованого управління засобами радіоелектронної боротьби, підвищуючи точність і ефективність їх роботи за мінімізації зайвих витрат ресурсів.

Enmao Wang, Qiming Huang, Klishin V.I. et al.

To optimize the selection of surfactants for coal seam water injection, this study examines the effects of ionic surfactants (DTAB and SDS) on water seepage in low-rank long-flame coal. A three-dimensional Eulerian two-phase flow model is established via ANSYS Fluent software to simulate the surfactant migration and is then validated through triaxial seepage experiments. The following crucial results are obtained: For long-flame coal from Houwenjialiang Coal Mine in Ordos City, Inner Mongolia Autonomous Region, China, the seepage coefficient increases with higher water injection pressure when using two surfactant solutions. Subjected to the water injection pressure of 2 MPa, the seepage coefficients of SDS and DTAB surfactants in order are obtained as 0.06 and 0.08, respectively, and the enhancement effect of cationic surfactant DTAB on coal seam water injection is superior to that of anionic surfactant SDS. In the presence of external pressure, the liquid seeps upwards along the pressure direction and with a steady flow, and seeps around along the pores and cracks of the coal body. The obtained results also reveal that the adsorption effect of SDS surfactant on coal is strong, resulting in a weaker water injection effect with increased water injection time into the coal seam.

Dian Afif Arifah, Ratih Andhika Akbar Rahma, Triana Harmini et al.

Introduction: As the center-cultured region in Indonesia, Ponorogo Regency is dominated by traditional manufacturing industries which support regional economic growth. Most production in this sector is labor-intensive and depends on manual handling processes, which may increase the risk of work-related musculoskeletal disorders (WMSDs). This study aims to develop a model to evaluate and predict ergonomic hazards using a neural network algorithm, focusing on the relationship between manual handling postures and musculoskeletal pain in 12 body regions. Method: A cross-sectional study involved data of 250 workers measured using used Nordic Musculoskeletal questionnaire and manual handling exposure checklist based on SNI 9011:2021. A neural network model was developed based on GLM’s output to explore the complex interrelationships between manual handling postures (X variables) and musculoskeletal pain across 12 body regions (Y variables). Result: The outputs identified carrying object over 9 meters (X10), one-handed lifting (X3), and trunk twisting (X2), with X10 confirmed as the most predictor for multiple outcomes, affecting six regions. Neural network models demonstrated adequate learning capacity with stable architecture, proved by average CEE values ranging from 0.21 to 0.54. The models showed improved predictive accuracy across epochs. Conclusion: The finding shows that NN modelling may be expanded to include broader industries in Indonesia's traditional manufacturing sector as an integrated data-based information system application. However, further validation using external datasets is recommended to enhance generalizability.

Chaozheng Wang, Zezhou Yang, Shuzheng Gao et al.

Code completion, a crucial practice in industrial settings, helps developers improve programming efficiency by automatically suggesting code snippets during development. With the emergence of Large Code Models (LCMs), this field has witnessed significant advancements. Due to the natural differences between open-source and industrial codebases, such as coding patterns and unique internal dependencies, it is a common practice for developers to conduct domain adaptation when adopting LCMs in industry. There exist multiple adaptation approaches, among which retrieval-augmented generation (RAG) and fine-tuning are the two most popular paradigms. However, no prior research has explored the trade-off of the two approaches in industrial scenarios. To mitigate the gap, we comprehensively compare the two paradigms including Retrieval-Augmented Generation (RAG) and Fine-tuning (FT), for industrial code completion in this paper. In collaboration with Tencent's WXG department, we collect over 160,000 internal C++ files as our codebase. We then compare the two types of adaptation approaches from three dimensions that are concerned by industrial practitioners, including effectiveness, efficiency, and parameter sensitivity, using six LCMs. Our findings reveal that RAG, when implemented with appropriate embedding models that map code snippets into dense vector representations, can achieve higher accuracy than fine-tuning alone. Specifically, BM25 presents superior retrieval effectiveness and efficiency among studied RAG methods. Moreover, RAG and fine-tuning are orthogonal and their combination leads to further improvement. We also observe that RAG demonstrates better scalability than FT, showing more sustained performance gains with larger scales of codebase.

Huajie Liang, Di Wang, Yuchao Lu et al.

With the advancement of Industry 4.0, intelligent manufacturing extensively employs sensors for real-time multidimensional data collection, playing a crucial role in equipment monitoring, process optimisation, and efficiency enhancement. Industrial data exhibit characteristics such as multi-source heterogeneity, nonlinearity, strong coupling, and temporal interactions, while also being affected by noise interference. These complexities make it challenging for traditional anomaly detection methods to extract key features, impacting detection accuracy and stability. Traditional machine learning approaches often struggle with such complex data due to limitations in processing capacity and generalisation ability, making them inadequate for practical applications. While deep learning feature extraction modules have demonstrated remarkable performance in image and text processing, they remain ineffective when applied to multi-source heterogeneous industrial data lacking explicit correlations. Moreover, existing multi-source heterogeneous data processing techniques still rely on dimensionality reduction and feature selection, which can lead to information loss and difficulty in capturing high-order interactions. To address these challenges, this study applies the EAPCR and Time-EAPCR models proposed in previous research and introduces a new model, Time-EAPCR-T, where Transformer replaces the LSTM module in the time-series processing component of Time-EAPCR. This modification effectively addresses multi-source data heterogeneity, facilitates efficient multi-source feature fusion, and enhances the temporal feature extraction capabilities of multi-source industrial data.Experimental results demonstrate that the proposed method outperforms existing approaches across four industrial datasets, highlighting its broad application potential.

Changhong Fu, Hua Lin, Haobo Zuo et al.

Text spotting for industrial panels is a key task for intelligent monitoring. However, achieving efficient and accurate text spotting for complex industrial panels remains challenging due to issues such as cross-scale localization and ambiguous boundaries in dense text regions. Moreover, most existing methods primarily focus on representing a single text shape, neglecting a comprehensive exploration of multi-scale feature information across different texts. To address these issues, this work proposes a novel multi-scale dense text spotter for edge AI-based vision system (EdgeSpotter) to achieve accurate and robust industrial panel monitoring. Specifically, a novel Transformer with efficient mixer is developed to learn the interdependencies among multi-level features, integrating multi-layer spatial and semantic cues. In addition, a new feature sampling with catmull-rom splines is designed, which explicitly encodes the shape, position, and semantic information of text, thereby alleviating missed detections and reducing recognition errors caused by multi-scale or dense text regions. Furthermore, a new benchmark dataset for industrial panel monitoring (IPM) is constructed. Extensive qualitative and quantitative evaluations on this challenging benchmark dataset validate the superior performance of the proposed method in different challenging panel monitoring tasks. Finally, practical tests based on the self-designed edge AI-based vision system demonstrate the practicality of the method. The code and demo will be available at https://github.com/vision4robotics/EdgeSpotter.

Pingyi Fan, Anbai Jiang, Shuwei Zhang et al.

With the rapid deployment of SCADA systems, how to effectively analyze industrial signals and detect abnormal states is an urgent need for the industry. Due to the significant heterogeneity of these signals, which we summarize as the M5 problem, previous works only focus on small sub-problems and employ specialized models, failing to utilize the synergies between modalities and the powerful scaling law. However, we argue that the M5 signals can be modeled in a unified manner due to the intrinsic similarity. As a result, we propose FISHER, a Foundation model for multi-modal Industrial Signal compreHEnsive Representation. To support arbitrary sampling rates, FISHER considers the increment of sampling rate as the concatenation of sub-band information. Specifically, FISHER takes the STFT sub-band as the modeling unit and adopts a teacher student SSL framework for pre-training. We also develop the RMIS benchmark, which evaluates the representations of M5 industrial signals on multiple health management tasks. Compared with top SSL models, FISHER showcases versatile and outstanding capabilities with a general performance gain up to 4.2%, along with much more efficient scaling curves. We also investigate the scaling law on downstream tasks and derive potential avenues for future work. Both FISHER and RMIS are now open-sourced.

Samir K. Khrais, Tamer Elia Yared, Noor Majid Saifan et al.

Surface mining is one of the hazardous industries that have several risky operations, including transportation, treatment, and mineral extraction. To avoid the risk of disaster, it is important to evaluate safety procedures and determine expected hazards. The aim of this study is to develop a thorough safety evaluation model for the surface mining industry based on the analytic hierarchy process (AHP), one important multi-criteria decision-making approach. A total of 11 criteria and 36 sub-criteria that are both independent and homogeneous were involved in the decision problem. Further, a deep sensitivity analysis was conducted to assess the stability of the ranking preference. The findings indicate that four out of the eleven criteria are particularly significant. To test the model’s applicability and effectiveness, a case study was conducted involving three surface mining companies located in the north of Jordan. The results demonstrate that the model is reliable, applicable, and effective in addressing real-world problems.

Hiroshi Nakashima, Isamu Kabe, Satoko Iwasawa et al.

Objectives: For occupational health staff, the health status of the worksite is an important matter, and a single index for presenting this health status is desired. We applied a stroke prediction model to employees of a Japanese non-iron metal company working at 10 worksites to present health status of the worksite. Methods: We applied a stroke prediction model of the Japan Public Health Center-based Prospective Study to 2,807 male employees without history of cardiovascular disease. We additionally applied models from the Japan Arteriosclerosis Longitudinal Study and from the Suita Study for validation. As the expected value for each employee at a worksite, we calculated the mean of employees’ predicted 10-year stroke risk for each worksite. To adjust difference in age distribution, the stroke risk of each worksite was age-adjusted using the direct method. The expected values were presented as the representative value of a worksite with the 95% confidence interval calculated using the bootstrap method. Logistic regression analysis was conducted to explore the reason why a worksite exhibits a high risk. We examined if partial regression coefficients of the worst worksite were affected by modifiable risk factors. Results: Three models predicted similar stroke risks for 10 worksites. Difference in the predicted stroke risk was observed among the worksites even after age-adjustment. Diabetes mellitus was found to affect partial regression coefficient of the worst worksite in any of three prediction models. Conclusion: The stroke prediction model was observed to be a comprehensive tool for presenting a worksite’s health status.

Yurii Husak , Viktor Vlasiuk , Ivan Starynskyi

Метою статті є розроблення способу оцінювання оперативності управління та прогнозування її зміни під час застосування угруповань військ (сил). У дослідженні використано елементи теорії масового обслуговування, теорії ймовірностей, методи аналізу, експертних оцінок, статистичної обробки даних та порівняння. Запропонований спосіб оцінювання дав змогу побудувати граф станів, в яких може перебувати пункт управління і визначити інтенсивності переходів між ними та ймовірності перебування у кожному стані. Основною властивістю функціонування органів управління, яка характеризує їх здатність здійснювати управління військами (силами) в терміни, що забезпечують успішне виконання поставлених завдань, в умовах негативного впливу зовнішніх і внутрішніх факторів є оперативність управління. Враховуючи, що інтенсивність впливу більшості з цих факторів змінюється у часі, оперативність управління теж не буде постійною. Враховуючи складність об’єкту дослідження, який функціонує в багатофакторному просторі, його аналіз найкраще здійснити шляхом математичного моделювання. Для цього в статті розроблено математичну модель, що ґрунтується на використанні системи масового обслуговування з різнорідними каналами. Різнорідність каналів пов’язана з різним можливим впливом на пункт управління – вогневим, диверсійних розвідувальних груп противника, кібернетичних засобів, надпотужного високочастотного випромінювання, обмеженої технічної надійності засобів автоматизованих систем управління тощо. Зважаючи на те, що структура пункту управління являє собою сукупність відносно однорідних елементів, використання згаданого методу передбачає побудову графа станів. Для цього в статті визначені можливі стани умовного пункту управління, орієнтовні значення інтенсивності переходів між ними, ймовірностей виникнення зовнішніх і внутрішніх факторів. Відповідно до теорії масового обслуговування кожен стан на визначений момент часу (етап ведення операції) буде характеризуватися ймовірністю перебування в ньому, а переходи між станами описуватися системою диференціальних рівнянь. Такий підхід дає змогу враховувати ймовірнісний вплив зовнішніх і внутрішніх факторів, присутніх у процесі функціонування пункту управління. Удосконалено математичну модель для опису процесу зміни стану пункту управління під впливом різних факторів. Оперативність управління розраховується через тривалість циклу управління та ймовірність перебування пункту управління в працездатному стані. Теоретичним значенням викладеного у статті є розробка способу оцінювання оперативного управління угрупованнями військ. Практичне значення зводиться до можливості застосування цієї методики органами військового управління Збройних Сил України та науково-дослідними установами для виявлення проблем під час управління застосуванням угруповань військ.

Wei Gao, Jingqiang Wang, Xinv Zhu et al.

We would like industrial robots to handle unstructured environments with cameras and perception pipelines. In contrast to traditional industrial robots that replay offline-crafted trajectories, online behavior planning is required for these perception-guided industrial applications. Aside from perception and planning algorithms, deploying perception-guided manipulators also requires substantial effort in integration. One approach is writing scripts in a traditional language (such as Python) to construct the planning problem and perform integration with other algorithmic modules & external devices. While scripting in Python is feasible for a handful of robots and applications, deploying perception-guided manipulation at scale (e.g., more than 10000 robot workstations in over 2000 customer sites) becomes intractable. To resolve this challenge, we propose a Domain-Specific Language (DSL) for perception-guided manipulation applications. To scale up the deployment,our DSL provides: 1) an easily accessible interface to construct & solve a sub-class of Task and Motion Planning (TAMP) problems that are important in practical applications; and 2) a mechanism to implement flexible control flow to perform integration and address customized requirements of distinct industrial application. Combined with an intuitive graphical programming frontend, our DSL is mainly used by machine operators without coding experience in traditional programming languages. Within hours of training, operators are capable of orchestrating interesting sophisticated manipulation behaviors with our DSL. Extensive practical deployments demonstrate the efficacy of our method.

Christian Nikolajew, Holger Eichelberger

For Industry 4.0 applications, communication protocols and data formats even for legacy devices are fundamental. In this paper, we focus on the Modbus/TCP protocol, which is, e.g., used in energy metering. Allowing Industry 4.0 applications to include data from such protocols without need for programming would increase flexibility and, in turn, improve development efficiency. As one particular approach, we discuss the automated generation of Modbus/TCP connectors for our Open Source oktoflow platform and compare the performance of handcrafted as well as generated connectors in different settings, including industrial energy metering devices.

Rakesh Gangadharaiah, Johnell O. Brooks, Patrick J. Rosopa et al.

Due to the advancements in real-time information communication technologies and sharing economies, rideshare services have gained significant momentum by offering dynamic and/or on-demand services. Rideshare service companies evolved from personal rideshare, where riders traveled solo or with known individuals, into pooled rideshare (PR), where riders can travel with one to multiple unknown riders. Similar to other shared economy services, pooled rideshare is beneficial as it efficiently utilizes resources, resulting in reduced energy usage, as well as reduced costs for the riders. However, previous research has demonstrated that riders have concerns about using pooled rideshare, especially regarding personal safety. A U.S. national survey with 5385 participants was used to understand human factor-related barriers and user preferences to develop a novel Pooled Rideshare Acceptance Model (PRAM). This model used a covariance-based structural equation model (CB-SEM) to identify the relationships between willingness to consider PR factors (<i>time/cost, privacy, safety</i>, <i>service experience</i>, and <i>traffic/environment</i>) and optimizing one’s experience of PR factors (<i>vehicle technology/accessibility, convenience</i>, <i>comfort/ease of use</i>, and <i>passenger safety</i>), resulting in the higher-order factor <i>trust service</i>. We examined the factors’ relative contribution to one’s <i>willingness/attitude towards PR</i> and <i>user acceptance of PR</i>. <i>Privacy</i>, <i>safety</i>, <i>trust service</i>, and <i>convenience</i> were statistically significant factors in the model, as were the <i>comfort/ease of use</i> factor and the <i>service experience</i>, <i>traffic/environment</i>, and <i>passenger safety</i> factors. The only two non-significant factors in the model were <i>time/cost</i> and <i>vehicle technology/accessibility</i>; it is only when a rider feels safe that individuals then consider the additional non-significant variables of time, cost, technology, and accessibility. <i>Privacy</i>, <i>safety,</i> and <i>service experience</i> were factors that discouraged the use of PR, whereas the <i>convenience</i> factor greatly encouraged the acceptance of PR. Despite the <i>time/cost</i> factor’s lack of significance, individual items related to time and cost were crucial when viewed within the context of convenience. This highlights that while user perceptions of privacy and safety are paramount to their attitude towards PR, once safety concerns are addressed, and services are deemed convenient, time and cost elements significantly enhance their trust in pooled rideshare services. This study provides a comprehensive understanding of user acceptance of PR services and offers actionable insights for policymakers and rideshare companies to improve their services and increase user adoption.

Rui Zhao, Zhongze Yang, Dong Liang et al.

This paper explores the application of automated machine learning (AutoML) techniques to the construction industry, a sector vital to the global economy. Traditional ML model construction methods were complex, time-consuming, reliant on data science expertise, and expensive. AutoML shows the potential to automate many tasks in ML construction and to create outperformed ML models. This paper aims to verify the feasibility of applying AutoML to industrial datasets for the smart construction domain, with a specific case study demonstrating its effectiveness. Two data challenges that were unique to industrial construction datasets are focused on, in addition to the normal steps of dataset preparation, model training, and evaluation. A real-world application case of construction project type prediction is provided to illustrate the accessibility of AutoML. By leveraging AutoML, construction professionals without data science expertise can now utilize software to process industrial data into ML models that assist in project management. The findings in this paper may bridge the gap between data-intensive smart construction practices and the emerging field of AutoML, encouraging its adoption for improved decision-making, project outcomes, and efficiency

Priskila Hananingrum, Ais Assana Athqia, Y. Denny A. Wahyudiono

Introduction: Mental workload is one of the most important aspects that affects the health and safety of workers. The Maintenance and Repair Division and Warship Division are divisions in PT. PAL which has a high job demand of the workers in it. Methods: This study was an observational analytic study with a cross-sectional design. The independent variables used in this study were age, gender, job placement, and social relationship, while the dependent variable was mental workload. The sample used was the total population of all managers in both divisions, totaling 12 respondents. The data was collected using a general questionnaire and the NASA-TLX method was used to measure mental workload. The data analysis technique used was the correlation test. Results: In the Maintenance and Repair Division, most managers were in the age range of 46 – 55 years old (50%) and 4 managers had an overloaded mental workload (66.7%). In the Warship Division, most of the managers were 46 – 55 years old (66.6%) and 4 managers (66.7%) had a moderate workload. Age has a relationship with mental workload in the Maintenance and Repair Division (0,612) and Warships Division (-0,316). Gender shows no relation with mental workload in the Warship Division (0,196). Job placement (-0.632) and social relationship (0.316) have a relation with mental workload in the Warship Division. Conclusion: Age has a relationship with mental workload in both divisions while there is no relationship between gender and mental workload among the managers in the Warship Division. Job placement has a strong negative relationship while social relationship has a strong positive with the mental workload in the Warship Division.