International Congress and Workshop on Industrial AI and eMaintenance 2023

Vision-based neural networks as artificial intelligence models have been critical in many manufacturing industries, including automotive, food, and aerospace. Machine vision and deep learning have provided practical, promising, and accessible innovations to address many problems during manufacturing and diagnostics. Recent studies offer beneficial results in implementing industrial artificial intelligence systems that require determining, comparing, and evaluating optimal technological solutions. The aerospace industry has to deal with the problem of diagnosing critical components that continually need the expertise of a trained human being. On the other hand, automatic diagnostics are becoming a critical technology to deal with this problem. However, these systems require a particular configuration of computer vision algorithms which, in the case of turbine trench-filling components, have yet to be added to the literature. Considering the above, we report in this paper a new methodology that uses a pre-trained deep neural network framework available online to automatically diagnose geometric nonconformities in aeronautical components that protect the turbines from vibrations and reduce noise from the engines to the aircraft deck. The method is based on the following stages: a computer vision system with a monochrome camera to acquire images, training of deep neural networks with transfer learning, and a stage to analyze nonconformities automatically. In addition, a benchmark of several pre-trained deep neural networks is presented to address the problem. According to the experimental results, the YOLO deep neural network topology significantly contributes to automatic diagnosis with high precision and exact rate. Finally, we show that integrating deep neural networks with control graphs is a promising strategy to optimize online diagnostics in the production lines of the aeronautical industry.

Asset management of railways is a data-driven process. Empowering asset management through utilisation of Artificial Intelligence (AI) and digital technologies for data-driven fact/based decision-making is highly dependent on the availability and accessibility of data. Additionally, data-driven approach puts demands on the quality of data and the relevance of the datasets to the contexts of analytics. This is to ensure the accuracy of the analytics and the precision of the predictions. One of the emerging approaches that can be utilised to augment data used for analytics and model learning process is Generative Artificial Intelligence (GAI). GAI can be useful in the various contexts of asset management of railways. This paper aims to provide some use cases in which GAI can be utilised for e.g. data augmentation that will lead to an improved accuracy and precision of decision-support. The identified use cases will provide a list of potential areas that can be used to develop a roadmap for implementation of GAI within the asset management of railways.

The constant needs to recruit and retain professional fighter pilots has been advocated by the rising trend of pilot age related cognitive impairments, pilot shortages and the continuous struggle to retain pilots is likely to get worse in the coming years. This crisis requires immediate actions and must be addressed by recruiting more pilots and by offering incentives to retain available pilots. The present investigation aims to address this problem with the introduction of a prevention strategy to ensure continued service of aging pilots. The continuous interaction between pilot, the flight information system, and the environment, have been recognized to affect pilot’s performance and error potentially hindering the safety of the aircraft. The research direction undertaken is intended to provide pilots with cognitive stimulation to raise flight safety levels and maintain flying skills proficiency. For this purpose, a novel neuron-based mirror platform has been proposed enabling understanding of another pilot’s actions and purposes behind pre-activated targeted responses from motorial cognitive domains. The study extends the application of mechanisms involved of mirror activation by combining a cognitive architecture for generating neuroergonomics mapping (NMP) programs. This interaction will enable pilots to develop learning experience based on observation and synchronized movements. NMP programs would also be tailored to ground-based training systems to enhance tactical air combat manoeuvring for active-duty fighter pilots.

This paper describes results from a research and development (R&D) project at Trafikverket (Swedish transport administration). The purpose of the project is to establish a risk-based framework for continuous safety improvement in railway infrastructure asset management. The approach is based on a combination of methodologies and tools described in dependability standards, such as Fault Tree Analysis (FTA), Failure Modes, Effects and Criticality Analysis (FMECA) and Event Tree Analysis (ETA). The empirical material is related to railway track with focus on safety critical rail defects potentially leading to rail breaks and derailment, but also their management within Trafikverket. Besides identifying risks, barriers and improvements, the approach complies with regulations and mandatory standards. Examples of these are Common Safety Method for Risk Evaluation and Assessment (CSM-RA, EU 402/2013) and EN 50126—RAMS (Reliability, Availability, Maintainability and Safety) for railway applications. In addition, the approach complies with regulatory requirements related to enterprise risk management and internal control, i.e., effectiveness, productivity, compliance and reporting. The approach also supports asset management in accordance with the ISO 55000-series.

Utilizing computational frameworks that involve simulation, modeling, and machine learning has gained popularity in lubricant industries to speed up research development. The frameworks serve as digital twins to aid the design of new lubricants by allowing the study of molecular assembling processes, analyzing various candidate chemicals, and understanding their physical properties under different application conditions to complement laboratory experiments. This work aims to evaluate the performance of the Proximal Policy Optimization (PPO) deep reinforcement learning (RL) agent in describing long-chain folding hydrocarbons, compounds commonly used as a main ingredient in lubricant industries. The hexadecane structure is a suitable benchmark molecule for assessing the RL agent. The policy learned by the RL agent encodes the intramolecular characteristics required to dictate the activity of individual molecules. Once trained on ab initio molecular dynamics trajectories, the RL molecular agents act in a virtual environment. Observing the dynamics of topological shapes and their properties then demonstrates the agents’ ability.

This paper illustrates a generalisable experimental approach to assess the causal effects of track maintenance actions. In our case, we assess the causal effects of tamping on railway track geometry. The tamping was conducted during the regular autumn tamping campaign in 2022 on track section 118 of the Swedish Iron ore line “Malmbanan”. The experimental setup provided frequent measurements of track geometry closely before and after the tamping action to estimate the potential causal effects of tamping on track geometry. Damill AB provided the track geometry data through an onboard measurement system mounted on a passenger train wagon travelling the track section frequently. The system provides position, speed, and track geometry variables, such as longitudinal level and lateral alignment, through GPS, accelerometers, and gyros. We based the analysis on hypothesis testing (Welch’s two-sample t-test) to test the statistical significance of tamping effects on track geometry variables. The analysis included a segmentation of the tamped track section to form experimental units allowing the application of hypothesis testing. The frequent measurements also allow for assessing how effects change over time. The analysis established a statistically significant reduction in the standard deviation of the longitudinal level and the standard deviation of the lateral alignment of the track after tamping. The experimental approach is generalizable for assessing the causal effects of other railway track maintenance actions.

In recent years, self-driving car technology has advanced rapidly due to significant investments in research and development by major automakers and technology companies. However, there are still challenges to be addressed, particularly when it comes to operating in harsh weather conditions such as the Arctic environment. The operation of sensor technologies used in self-driving cars can be significantly affected by such conditions, making it challenging to deploy them in these regions. Therefore, there is a necessity for further research and development of specialized solutions and technologies that can be used specifically for self-driving cars in Arctic environments. This paper addresses the following research questions: (RQ1) What are the technologies that enables the autonomous driving of self-driving cars, and (RQ2) how do they work? (RQ3) What are the key challenges that must be addressed to successfully implement self-driving cars in the Arctic region? (RQ4) What are the impacts of widespread adoption of self-driving cars, and how might they shape the future of transportation? The development of specialized solutions and technologies specifically designed for use in Arctic Environment is crucial to overcome these challenges. Further research and development are necessary to ensure that self-driving cars can be deployed safely and effectively in all weather conditions. As technology continues to evolve, it is likely that we will see even more advancements in self-driving car technology in the coming years.

Africa has 82,000 km of railway lines of which around 68,000 km are operational (est. 2007). South Africa accounts for a quarter of Africa’s operational railway lines which moves 63% of Africa’s railway freight by mass (est. 2007). In South Africa, freight railway lines are managed by Transnet, where freight is the main driver for Transnet’s growth. Transnet’s freight is currently centered around two lines, namely, the 861 km Sishen-Saldanha iron-ore line and the 621 km coal system linking Mpumalanga mines with the Richards Bay port which account for 60% of Transnet’s freight (Transnet SOC Ltd © LTPF (2016) Chapter 3). South Africa’s railway network is in rapid decline with freight transport declining by 25% over the last five years. For the financial year 2021–2022, the transported freight was merely 172.7 mt, a decline of 10 mt from 2007 and a decline of 54 mt from the peak in 2015. The welfare of South Africa’s railway infrastructure is critical for the world as South Africa stores 75% of the world’s known manganese used in stainless steel production. Maintaining an aged and declining railway network within a resource-constrained environment requires careful planning and prioritization of maintenance tasks. To support the management of Transnet’s railway infrastructure, we are developing a railway infrastructure performance digital twin. The performance digital twin is being designed for real-time assessment of the state of the railway infrastructure as well as the translation thereof in identifying and prioritizing maintenance tasks. Given available resources, utilization, and transport risks the digital twin will be able to inform maintenance schedules and maintenance decision-making. Only limited digital twin instances for railway infrastructure have been realized mainly within a first-world context.

Mining is becoming increasingly vulnerable to the effects of climate change (CC). The consequences of changing weather patterns, such as extreme weather events that can damage equipment, infrastructure, mining facilities, and operation interruption, are the source of the vulnerability. The new demand initiated by governments and international agreements put extra pressure on mining industries to update their policies to reduce greenhouse gas (GHG) emissions and adapt to CC, such as carbon pricing systems, renewable energy, and sustainable development. Most mining and exploration industries focus on reducing mining’s impact and climate mitigation on CC rather than adapting to extreme weather events. Therefore, it is important to study and investigate the impacts of CC on the mining sector. This paper aims to study the challenges and strategies for adapting and mitigating CC impacts on mining using a systematic literature review (SLR). These results showed that most of the proposed models and strategies in the mining field are in the conceptual phase, and fewer are practical models.

This paper describes results from a research and development (R&D) project at Trafikverket (Swedish transport administration). The purpose of the study was to systemize dependability improvements of Trafikverkt’s Control Command and Signalling (CCS) assets. A case study was conducted on level crossings that represent a critical part of the CCS system. The results of the study show that the systemic approach contributes to asset management as it contributes to short-term dependability and productivity improvements as well as to medium term specifications for system modifications and long-term specifications for next generation of level crossings. The approach is based on a combination of methodologies and tools described in dependability standards, e.g., Failure Modes, Effects & Criticality Analysis (FMECA). However, the approach also considers aspects from Design of Experiments (DoE) to support field tests aligned with other tasks in the railway infrastructure. Besides contributing to improvements, the approach comply with regulations and mandatory standards. Examples of these are Common Safety Method for Risk Evaluation and Assessment (CSM-RA, EU 402/2013) and EN 50126–RAMS (Reliability, Availability, Maintainability & Safety) for railway applications. In addition, the approach comply with regulatory requirements related to enterprise risk management and internal control, i.e., effectiveness, productivity, compliance and documentation. The approach also supports asset management in accordance with the ISO 55000-series.

Artificial Intelligence (AI) and digitalization are changing the landscape of performing projects in the construction industry. In construction projects, the responsibility for achieving the required scope within a specified timeframe and estimated cost lies with the project managers. To meet these stringent requirements, site managers must make on-site project decisions, relying heavily on their past experiences and intuition. Moreover, the complexity of the decision-making process is amplified by the need to manually review various information sources such as Building Information Modelling (BIM), Bill of Quantities (BOQ), construction drawings etc. The heterogeneity, complexity, availability, accessibility, and volume of these contents that need to be processed by the decision-maker pose risks to the decision-making processes that may adversely impact resource consumption, planned time, and cost. The emerging technologies related to AI and digitalization are expected to improve the effectiveness and efficiency of the decision-making process in construction projects. However, the development and implementation of such technologies are highly dependent on the identification and definition of contexts to which AI and digital technologies add additional value. Hence, the purpose of this research paper is to study and explore the decision-making process of site managers. This paper also provides the identified gaps and potential of utilizing AI and digital technologies to assist in decision-making. Further, this paper proposes a set of conceptual models that combine hardware and AI algorithms to support the site decision-making process. The findings provide insights into the complexities site managers face and offer innovative approaches to mitigate risks and improve decision making efficiency.

One vital contribution of Reliability-Centred Maintenance (RCM) is the definition of potential failure, which led to the concept of Condition-Based Maintenance (CBM) being accepted as one of the best ways of preventing functional failure. To enable CBM, the condition of an item must be monitored by Condition Monitoring (CM) of some critical functions. The CM results in collected data that represent the system’s condition in some way. Diagnostics and prognostics is then concerned with the interpretation of collected condition data and the conclusion drawn about the item’s current and future condition. On the basis of the diagnostic and prognostic information, decisions about appropriate CBM can be made. The purpose of the risk-based dependability assessment described in this paper, is to support a decision whether or not a railway infrastructure item should be covered by a new digitalised inspection solution for CM to enable improved CBM. Hence, the dependability assessment indicates ‘which’ functions or items that should be covered by a digitalised solution for inspection (CM) and ‘why’ they should be covered. ‘How’ the coverage should be solved, i.e., by which technical solutions, is not included in this paper. The proposed dependability assessment is based on stakeholder requirements, through a combination of a Failure Modes, Effects & Criticality Analysis (FMECA) and an Event Tree Analysis (ETA) with three major decision points where Maintenance Significant Items (MSIs) for further analysis are identified. This paper focuses on the ETA-part and the third MSI-selection of the proposed approach.

In this paper we describe recent Swedish trials to combine different data sources to support fault finding on railway vehicles and in railway operations in Sweden. The background to the trials, trials aims, the conduct of the trials and the data sources used are described in brief. The results of the trials are described using several case studies. Practical limitations and constraints are discussed together with the ways in which these might be overcome. Finally, the paper describes the way in which it is hoped to apply such methods to support future railway fleet management and operations.

Search missions for objects are relevant in both industrial and civilian context, such as searching for a missing child in a forest or to locating equipment in a building or large factory. To send out a drone swarm to quickly locate a misplaced item in a factory, a missing machine on a building site or a missing child in a forest is very similar. Image-based Machine Learning algorithms are now so powerful that they can be trained to identify objects with high accuracy in real time. The next challenge is to perform the search as efficiently as possible, using as little time and energy as possible. If we have information about the area to search, we can use heuristic and probabilistic methods to perform an efficient search. In this paper, we present a case study where we developed a method and approach to evaluate different search algorithms enabling the selection of the most suitable, i.e., most efficient search algorithm for the task at hand. A couple of probabilistic and heuristic search methods were implemented for testing purposes, and they are the following: Bayesian Search together with a Hill Climbing search algorithm and Bayesian Search together with an A-star search algorithm. A swarm adapted lawn mower search strategy is also implemented. In our case study, we see that the performance of the search heavily depends on the area to search in and domain knowledge, e.g., knowledge about how a child is expected to move through a forest area when lost. In our tests, we see that there are significant gains to be made by selecting a search algorithm suitable for the search context at hand.

One vital contribution of Reliability-Centred Maintenance (RCM) is the definition of potential failure, which led to the concept of Condition-Based Maintenance (CBM) being accepted as one of the best ways of preventing functional failure. To enable CBM, the condition of an item must be monitored by Condition Monitoring (CM) of some critical functions. The CM results in collected data that represent the system’s condition. Diagnostics and prognostics are then concerned with the interpretation of collected condition data and the conclusion drawn about the item’s current and future condition. On the basis of the diagnostic and prognostic information, decisions about appropriate CBM can be made. The purpose of the risk-based dependability assessment described in this paper, is to support a decision whether or not a railway infrastructure item should be covered by an additional digitalised inspection solution for CM to enable improved CBM. Hence, the dependability assessment indicates ‘which’ functions or items that should be covered by a digitalised solution for inspection (CM) and ‘why’ they should be covered. ‘How’ the coverage should be solved, i.e., by which technical solutions, is not included in this paper. The proposed dependability assessment is based on stakeholder requirements, through a combination of a Failure Modes, Effects & Criticality Analysis (FMECA) and an Event Tree Analysis (ETA) with three major decision points where Maintenance Significant Items (MSIs) for further analysis are identified. This paper focuses on the FMECA-part and the two related MSI-selections of the proposed approach. The approach has been used on an aggregated level to formulate goals of an innovation procurement. Later on, the methodology is expected to support verification and validation of proposed CBM-solutions on a more detailed level.

Railway infrastructures deteriorate under the influence of various physical, mechanical, and environmental including climate change. Climate change impacts during past years have led to various critical damage to railway infrastructure assets. Switch and crossing (S&C) are the sensitive components of the railway network, which is affected by climate change and severe events such as abnormal temperatures, snow and ice, and flooding. S&Cs’ Failures can lead to severe consequences, which often negatively influence the reliability and safety of the railway network. Reliable railway infrastructure asset clustering is essential for tactical and strategic decision-making to operate and maintain railway networks dealing with future climate scenarios. This study utilized machine learning (ML) models to cluster S&C failures by leveraging the historical maintenance data (number of failures, failure mode, time to repair, etc.), asset registry data (type, location, the criticality of the asset, etc.), inspection data, and weather data. Four different clusters have been identified considering the climatic pattern. The proposed model has been validated using S&Cs from the Swedish railway network. The clustering approach leads to uncertainty reduction in the model building, which has the potential to support robust and reliable decision-making in railway operation and maintenance management. Furthermore, this categorization helps infrastructure managers to implement climate adaptation actions leading to more resilient transport infrastructure.

Anti-friction bearings (AFB) are crucial structural components conveying rotating motions in a variety of mechanical systems. To avoid unscheduled breakdowns and fatal failures, remaining useful life (RUL) prediction is of great practical significance in industrial practice for prognostics health management, e.g., optimizing maintenance plan for component replacements. Recently, the artificial intelligence (AI) advancements have provided effective data-driven models for bearing prognostics using machine learning. In this paper, using the variational auto-encoder (VAE) networks as the regression backbone, the bearing RUL is estimated using envelope spectra via measured vibrational data. First, the envelope spectra are utilized for bearing fault detection and the network input features. After the fault is detected, the VAE is used for learning the probabilistic mapping from the spectral input to the estimated RUL value, given its good probabilistic and generative properties over the classical auto-encoder (AE) in content generation and variational inference. The application of the method to the run-to-failure measured vibration data from the experimental rig available online have shown its efficacy in bearing RUL estimation.

The fourth industrial revolution, colloquially referred to as “industry 4.0”, has garnered substantial global attention in recent years. There, Artificial intelligence (AI) driven industrial intelligence has been increasingly deployed in predictive maintenance (PdM), emerging as a vital enabler of smart manufacturing and industry 4.0. Since in recent years the number of articles focusing on Artificial Intelligence (AI) in PdM is high a review on the available literature reviews in this domain would be useful for the future researchers who would like to advance the research in this area and also for the persons who would like to apply PdM in their application domains. Therefore, this study identifies the AI revolution in PdM and focuses on the next stages available in the literature reviews in this area by quality assessment of secondary study. A well-known structured review approach (Systematic Literature Review, or SLR) was employed to perform this tertiary study. In addition, the Scale for the Assessment of Narrative Review Articles (SANRA) approach for evaluating the quality of review papers has been employed to support a few of the research questions. Here, This tertiary study scrutinizes four crucial aspects of secondary articles: (1) their specific research domains, (2) the annual trends in the quantity, variety, and quality (3) a footsteps of top researchers, and (4) the research constraints that review articles face during the time frame of 2015 to 2022. The results show that the majority of the application areas are applied to the manufacturing industry. It also leads to the identification of the revolution of AI in PdM as well. Our final findings indicate that Dr. Cheng et al.’s (2022) review has emerged as the predominant source of information in this field. As newcomers or industrial practitioners, we can benefit greatly from following his insights. The final outcome is that there is a lack of progress in SLR formulation and in adding explainable or interpretive AI methodologies in secondary studies.

There is an increased need for protection against unauthorized entry of drones as there has been an increased number of reports of UAV’s entering restricted areas. In this paper we explore an approach of using a swarm/team of drones that are able to cooperate, to autonomously engage and disable one or more unauthorized drones entering a restricted area. In our approach, we have investigated technologies for distributed decision-making and task allocation in real-time, in a dynamic simulated environment and developed descriptive models for how such technologies may be exploited in a mission designed for a drone swarm. This includes the definition of discrete tasks, how they interact and how they are composed to form such a mission, as well as the realization and execution of these tasks using machine learning models combined with behaviour trees. To evaluate our approach, we use a simulated environment for mission execution where relevant KPI’s related to the design of the mission have been used to measure how efficient our approach is in deterring or incapacitating unauthorized drones. The evaluation has been performed using Monte-Carlo simulations on a batch of randomized scenarios and measures of effectiveness has been used to measure each scenario instance and later compiled into a final assessment for the main scenario as well as each ingoing task. The results show a mission success in 93% of the simulated scenarios. Of these 93%, 58% of the scenarios resulted in the threat being neutralized and in 35% of the scenarios the threat was driven away from the critical area. We believe that the application of such measurements aids to validate the applicability of this capability in a real-world scenario and in order to assert the relevance of these parameters, future validations in real-world operational scenarios are warranted.

This paper focuses on the development of a platform used to identify optimal maintenance plans for railway tracks. To achieve this, a fully automated solution that can monitor the status of the tracks has been deployed by Hellenic Train. Sensors monitoring a variety of parameters (such as acceleration, vibration, position, cameras etc.) have been attached to the rolling stock frame, continuously monitoring the status of the tracks. Based on the collected measurements, Machine Learning schemes able to detect track defects and estimate the deterioration rate of track quality over time have been developed. The output of these models has been used as input to a set of optimization problems that have been formulated in order to estimate the candidate time periods during which maintenance activities can be scheduled under various constraints and cost functions.

Roads are one of the major physical infrastructures of Hepburn Shire Council (HSC) as of all other local councils. Every year HSC allocates and spends huge amount of budget on roads for maintenance and renewal. The road performance condition level has been the major priority for roads renewal selection. However, other criteria are under-considered, and there are gaps in significant analysis of the relation between roads age, condition, risk, and cost. In this study, decision-making framework or tool has developed using multi criteria technique (MCT) and analytic Hierarchy Process (AHP) for single objective optimisation i.e., to provide an agreed level of service optimising Maintenance and Renewal cost or improve the condition subjected to annual budget. This study adopted decision criteria as per community and council needs, by developing a model for criteria selection. Additionally, this study analysed the adopted HSC maintenance strategies, condition monitoring systems, performance conditions of the roads, and operational and renewal budget of HSC.

The concept of Metaverse is emerging in the industry. Metaverse is expected to be important in industrial asset management and sustainable operation and maintenance. Some of the potentials of implementing Metaverse in the industry can be related to virtual co-creation and design, remote and virtual inspection and maintenance, skills development and training, simulation, safety, and security. Additionally, Metaverse integrated with Artificial Intelligence (AI) and digital technologies will augment human perception, facilitating the Human-System-Interaction (HSI). The traditional HSI carries limitations regarding usability, immersiveness, and connectivity when it comes to the interaction between the virtual, augmented, and real world. An improved HSI in such cyberspace applications may lead to a better understanding of the system and eventually reduced faults. However, implementing Metaverse in industrial contexts is challenging and has not yet been explored thoroughly and systematically. Hence, this paper aims to systematically identify and investigate the various issues and challenges in the implementation of Metaverse in industrial contexts from an HSI perspective. The paper will further provide a taxonomy of these issues and challenges. The research methodology has been based on literature surveys, active and passive observations, and experiments done in the eMaintenanceLAB at Luleå University of Technology. The findings from this paper can be used to increase the effectiveness and efficiency of implementing Metaverse in various industrial contexts.

Biomimicry is an approach for solving industrial challenges by studying similar cases in nature and emulating bio-organisms’ responses. Thus, it helps to solve modern day technological problems using the solutions that bio-organisms have successfully used over the course of millions of years. In an ongoing research project, investigations are being carried out to explore the use of biomimicry approach for developing a framework for a human-centric condition management system. This framework is inspired by the knowledge of human cognition. It is expected that the system will be able to utilize various data and integrate it with analytical models and knowledge-based systems to help an equipment diagnose and recommend optimised operation and maintenance programs. This paper describes the proposed framework for this human-centric condition management system.

In this paper, we investigated the application of an autoencoder neural network for predictive analytics of control valves, which are crucial components in industrial processes with significant consequences in case of failure. The autoencoder was created using Python and Keras deep learning framework, comprising encoding and decoding sections. By comparing the difference between the input sensor data and its reconstructed output, referred to as the reconstruction error, we were able to identify anomalies. The result which is based on the data of an actual asset was compared with the random forest regressor to ensure the effectiveness of the approach. We have also proposed a practical approach to generate alerts when the magnitude exceeds a predefined threshold, thereby enabling proactive maintenance and avoiding unplanned shutdowns. We emphasized the diagnostic capability of the autoencoder in identifying anomalous sensors, which is not present in traditional regression approaches. Furthermore, we argued that this capability could be more valuable for a complex equipment with many input sensors. The proposed approach can be further improved to provide prognostic capability by forecasting the trend of the reconstruction error.

Life cycle cost (LCC) analysis is a key tool for effective infrastructure management. It is an essential decision support methodology for selection, design, development, construction, and maintenance of railway infrastructure system. Effective implementation of LCC analysis will assure cost-effective operation of railways from both investment and life-cycle perspectives. A major setback in the successful implementation of LCC by infrastructure managers is the availability of relevant, reliable, and structured data. Another challenge is the prediction of future behaviour of railway system with a change in the design or operation parameters. Different cost estimation methods and prediction models have been developed to deal with both challenges. However, there is a need to integrate prediction models as an integral part of LCC methodology, to account for possible changes in the model variables. This article presents an LCC based approach for requirement specification. It integrates degradation models with an LCC model to study the impact of change in design speed on key decision criteria such as track possession time, service life of track system, and LCC. The methodology is applied to an ongoing railway investment project in Sweden to investigate and quantify the impact of design speed change from 250 to 320 km/h. This is carried out to support specification of technical requirement for the design of track system. The results from the studied degradation models show that the correction factor for a change in speed varies between 0,79 and 0,96. Using this correction factor to compensate for changes in design speed, the service life of ballasted track system is estimated to decrease by an average of 15% from 30 years to approximately 25 years. Further, the LCC of the route under consideration will increase by an expected value of 30%.

Degrading linear assets such as railway track deteriorate and lose their functionality over time and usage. A reliable and effective predictive maintenance strategy is necessary to rehabilitate the functionality and reliability of these assets. Data analytics are required to be performed to extract the information used for the decision-making process and for implementing an optimized maintenance strategy. Accordingly, data pre-processing and data quality improvement are essential to remove errors in measurements and develop efficient data analysis methods. Inaccurate measurement positioning is a common error in track geometry measurements which causes track geometry single defects suffer from an uncontrolled shift called positional error. To reduce the positional errors of track geometry measurements, this paper presents two alignment methods i.e., modified correlation optimized warping (MCOW) and recursive segment-wise peak alignment (RSPA). MCOW is a profile-based method that align all the measurements with the same priority, while RSPA is a featured-based method that only focuses on the alignment of peaks with high amplitudes in the geometry measurements. To evaluate and compare the performance of these methods in aligning track geometry measurements, a case study was conducted. The results show that RSPA can precisely align the single defects, while the MCOW is more efficient when considering the same importance for aligning every single data-point.

For wind turbine operation and maintenance, wind turbine blade surface defect detection is a very important and challenging problem, as wind turbine blade surface defects seriously affect the efficiency and safety of the wind turbine. The performance of traditional methods depends heavily on the correlation between the handcrafted features and the features of the defects themselves, but surface defects are diverse which could make the traditional methods fail in reality. In this work, we present an automated framework to identify surface defects of blades using the advanced YOLOv5 algorithm, which can learn and extract blade surface defect features adaptively and accurately identify even very minor faults. The results of different algorithms are collected and compared based on a self-built dataset, which show that YOLOv5 has the best performance. In addition, YOLOv5 has significant advantages in terms of model size and training speed, and these advantages make the YOLOv5 model well suited for wind turbine blade surface defect recognition.

Unmanned Aerial Vehicle (UAV) swarms, also known as drone swarms, have been a subject of extensive research due to their potential to enhance monitoring, surveillance, and search missions. Coordinating several drones flying simultaneously presents a challenge in increasing their level of automation and intelligence to improve strategic organization. To address this challenge, we propose a solution that uses hill climbing, potential fields, and search strategies in conjunction with a probability map to coordinate a UAV swarm. The UAVs are autonomous and equipped with distributed intelligence to facilitate a cooperative search application. Our results show the effectiveness of the swarm, indicating that this approach is a promising approach to addressing this problem.

Unsupervised data-driven methods are attractive options for fault detection in rotating machinery since they do not require any failure data during training. However, in these data-driven approaches, engineering domain knowledge remains unexploited. Although engineering features are often used as inputs to machine learning models, thereby including domain knowledge, few methods exist for directly integrating domain knowledge about the expected machine fault behaviour into unsupervised fault detection methods. This paper presents a generic method for including domain knowledge into unsupervised, auto-encoder-based fault detection methods by regularising the Jacobian of the latent feature representation. This regularisation results in informative latent features that are sensitive to changes that are expected from a machine in a faulty condition. The proposed method is evaluated on a bearing fault detection task, both when using a low dimensional vector of engineering features and when using high dimensional frequency domain data. The analysis is conducted on two bearing fault data sets with different operating conditions, fault modes and signal-to-noise ratios. The proposed regularised auto-encoder yields improved ROC-AUC performance as compared to the unregularized baseline when evaluated on a latent-feature based fault indicator. The proposed method shows potential as a generic method for integrating engineering domain knowledge into fault detection problems.

Real-world data is typically dispersed among numerous businesses or governmental agencies, making it difficult to integrate them into data privacy laws like the General Data Protection Regulation of the European Union (GDPR). Two significant obstacles to the use of machine learning models in applications are the existence of such data islands and privacy issues. In this paper, we address these issues and propose ‘HFedRF: Horizontal Federated Random Forest’, a privacy-preserving federated model which is approximately lossless. Our proposed algorithm merges d random forests computed on d different devices and returns a global random forest which is used for prediction on local devices. In our methodology, we compare IIDs (Independent and Identically Distributed) and non-IIDs variant of our algorithm HFedRF with traditional machine learning (ML) methods i.e., decision tree and random forest. Our results show that we achieve benchmark comparable results with our algorithm for IID as well as non-IID settings of federated learning.

Rail surface defect detection is a relevant problem in the field of data-driven railway maintenance. Artificial intelligence and neural networks (NN) for axle box acceleration (ABA) or camera data show great potential for defect detection and classification. However, a sufficient amount of labeled training data is required, all the more if the defect severity is to be estimated. A unique dataset of time-synchronized ABA and camera data is employed that contains labeled defect instances. For the image analysis, RetinaNet as a single-stage object detector (with the backbone of ResNet-50 and a feature pyramid network) is used to achieve high classification performance for the two most common rail surface defects (squat and corrugation). Additionally, a machine learning-based method on ABA data to estimate defect severity levels (low, medium, heavy) is proposed. False positives are detected in the original labels by both classifiers during evaluation. The inspection of the false positives in image data reveals that defects have been overlooked in the initial labeling. The insights of this work help to reduce the dependency on labeled data by using only a few labeled samples and by exploiting complementary data sources instead of increasing the number of labeled instances.

Industry 4.0 presents nine technologies including Industrial Internet of Things (IIoT), Big Data and Analytics, Augmented Reality (AR), etc. Some of the technologies play an important role in the development of smart maintenance technologies. Previous research presents several technologies for smart maintenance. However, one problem is that the manufacturing industry still finds it challenging to implement smart maintenance technologies in a value-adding way. Open questionnaires and interviews have been used to collect information about the current needs of the manufacturing industry. Both the empirical findings of this paper, as well as previous research, show that knowledge is the most common challenge when implementing new technologies. Therefore, in this paper, we develop and present a testbed for how to approach smart maintenance technologies and to share technical knowledge to the manufacturing industry.

Digitalisation within industries is associated with many positive opportunities and simultaneously poses considerable cybersecurity-related threats. Cybersecurity is a critical concern for many industries, such as railway, aviation, mining, healthcare, and finance, where critical information is at risk of being compromised as well as the operations security. Today, researchers are looking into various solutions to tackle cybersecurity risks and still retain the desired function of the system. However, it is believed that these challenges can be approached by the integration of game theory and Cyber Kill Chain (CKC) which describe the different stages of a cyberattack to understand the chaotic situation of cybersecurity. Thus, the objective of the paper is to propose a game-based approach using game theory to address cybersecurity risks within industries by modelling the strategic interaction between attacker and defender in the Cyber Kill Chain (CKC). This approach aims to enhance understanding of the complex challenges and facilitate the development of effective cybersecurity solutions. This approach will help in evaluating the effectiveness of different security strategies. The proposed strategic approach uses a non-cooperative game which is based on mixed strategies. The authors have defined a scenario for simultaneous-move games by estimating values for various elements of the game. By analysing the behaviour of both attacker and defender, the proposed game-based approach can help industries to develop more effective and efficient security strategies. Further, the proposed approach will provide a better understanding of the complex challenges of cybersecurity in industrial contexts. It can also be used to develop appropriate strategies to mitigate cybersecurity risks.

The digitalization of manufacturing industry, known as e.g., Industry 4.0 or smart production, has opened new opportunities for real-time optimization of production systems. Also, this technological leap has provided new possibilities for the maintenance of production equipment to become data driven and in many cases predictive. This fourth industrial revolution is changing the role of humans at the shop floor. Visions of the dark factory arises, meaning fully automated factories where humans are redundant, both for physical processing and for decision making. The research on Smart maintenance shows great advances in predictive diagnostics and prognostic techniques. However, in manufacturing industry, studies have shown that up to 50–60% of equipment breakdowns are due to human errors. Some of these errors are partly addressed through the development of improved information aid, such as e.g., instructions through Augmented Reality and training in Virtual Reality. Still, the root cause of human errors in manufacturing industry haven’t been properly categorized in terms of e.g., neglect, lack of competence, unclear processes, or poor leadership. In this paper the potential of data driven maintenance is discussed from a human centric perspective. Considering the large part of failures being due to human factors and the possibilities of improvement through implementation of smart technologies, this paper argues for exploring the root causes of human errors in discrete item manufacturing systems and address the proper human centric technologies as a means of reducing these failures.

The maintenance and development of rail infrastructure in Africa is key to the future success of the African content. Development of a digital twin in the context of South African railway infrastructure will assist with the future need for effective maintenance planning and resource optimisation. As part of the process towards realisation of a railway infrastructure digital twin for South African railway infrastructure, the current study quantified the effect of various signal alignment strategies and errors on the ability of Singular-Spectrum Analysis to extract features that can be used when considering the evolution of track geometry over time. When minimal stretching is present in the data, the pairwise cross-correlation for the flattened matrices representing the two sets of elementary matrices for two subsequent measurement campaigns provide a latent space that looks promising for identifying changes in the track geometry as recorded by a track geometry car.

Wheel damages on railway vehicles caused by rolling contact fatigue or blocked wheels can cause severe problems for railway operators and infrastructure owners. Wheel impact load wayside detectors (WILD) are one of the means to assess the condition of a wheel in operation, but varying operating routes, irregular traffic patterns, and especially cross-border operations make this quite challenging. While the condition updates occur randomly, the detectors themselves are managed by different owners and principles. Thus, using the same type of data from not only different wayside locations but also different providers and authorities with varying fidelity and operational practices introduces uncertainties in data quality and consistency. This paper presents an approach for predicting wheel damage severity on a wagon fleet with irregular cross-border operations, achieving similar confidence levels as for regular traffic patterns on a national scale. The different sensor characteristics are explored between countries and within each country. The approach is implemented as a cloud-based solution which integrates wayside detector data from multiple locations provided by two different infrastructure owners in two countries. The solution estimates remaining useful life based on data from both countries and aggregates this to a single indication for the decision maker. The algorithm’s performance is showcased for vehicles with cross-border operations. The results indicate that the proposed approach confirms that irregularly provided measurement data with data quality and consistency issues are manageable and adequate decision-making performance.

The present paper explores the current need of a predictive model for Maintenance and Operations in Railway Systems that tackles the challenges of vertical separation. Railway Vehicles and Track (V-T) systems are responsible for large investment and maintenance costs, which should be optimised using a reliability-based Maintenance and Operation (M&O) decision model. The European railways face vertical separation, adding further complexity to M&O: while Train Operating Companies (TOCs) are maintaining their trains, track maintenance decisions are made by the Infrastructure Manager (IM). However, in this vertically separated system, no clear decision model seems to be in place to optimise the overall life cycle impacts of M&O decisions across the different railway agents. Therefore, a Collaborative Decision Model (CDM) is missing to align predictive M&O decisions. TOCs and IM are monitoring the evolution of their own assets and using sensor systems and signal processing techniques to identify and predict specific failures and support their M&O strategies in separate decision models. These M&O strategies, which very often have conflicting objectives, which may lead to sub-optimal overall life-cycle impacts. In fact, V-T systems have relevant joint behaviour in degradation, which significantly affect wear and damage of wheelsets and rails, as well as the life-cycle costs, reliability, availability and safety of the overall railway system. Thus, misalignments in M&O decisions can be reduced by using cooperative strategies between TOCs and IM. The current project PMO-RAIL will contribute towards an innovative reformulation of railway M&O problems, aiming to achieve a proof-of-concept that such a CDM framework to support PM&O scheduling decisions can provide better overall life-cycle impacts.

The paper discusses the independent cart technology, which utilizes linear motors to move carts along a predetermined track autonomously. This technology offers control of individual speed profiles for each section along the track, frictionless propulsion mechanism, and the ability to start and stop loads quickly. Nevertheless, the initial cost of these systems is substantial, and regular condition monitoring is required to ensure optimal performance and long-term economic benefits. The paper provides an overview of various condition monitoring and signal processing techniques for analysis, including data-driven modeling with machine learning algorithms. The article presents an experimental setup based on the independent cart system and outlines a strategy for data acquisition that emphasizes specific conditions during each run of the system. The collected data is critical in monitoring the independent cart system’s condition and developing expertise in identifying different types of faults and their precise locations, utilizing hybrid modeling approaches.

Belt conveyor systems are a primary means of bulk material transport in industrial applications due to their high bulk capacity and limited need for human involvement. Abnormal objects on the belt conveyor can be hazardous to the operation of the belt conveyor systems and/or downstream equipment. The dependability of production on a well-operating system in combination with the high degree of automation and limited inspection accessibility, establishes the need for a continuous and fully automated monitoring solution. In this paper, a monitoring solution comprising a camera, object detection and classification model, and decision support is presented and discussed. The detection and classification model is comprised of two steps: a classical brightness and contour detection algorithm using colour channel weighing, and a subsequent processing by a Convolutional Neural Network (CNN). The CNN performs a classification of the detections as True Positives (TP) or False Positives (FP). Further, the object size is estimated providing a measure for the risk imposed by the object. The solution makes use of an off-the-shelf industrial network camera that communicates with an edge computing device close to the installation site. The edge device is further connected to a SaaS solution for predictive maintenance and decision support where results (classified detections) are visualized in a dashboard. There, operators can assess classified detections as TP or FP, which provides a ground truth for subsequent retraining of the solution. Moreover, these actionable insights enable a warning and stopping mechanism that can be implemented when the operators trust the solution. The solution is implemented and tested at LKAB Narvik and operational since 2021. Initially, the solution was trained using artificially introduced objects and manually labelled video frames, followed by a validation phase to assess the performance of the solution. The solution exceeds targeted performance while having a low false positive rate.

Earlier studies have optimised the vibration-based parameters to identify the rotor defects only for the rotating machines. The artificial neural network (ANN) model was used earlier to classify the faults. The earlier optimised parameters are further examined for both rotor and bearing defects. These parameters are slightly modified in this research to accommodate bearing defects. The paper presents the study using an experimental vibration data from a laboratory-scaled rig.

Industry 4.0 promises huge effects on industrial performance, once critical equipment is equipped with sensors and interconnected, and big data sets and digital twins are established that allows for advanced data analytics using machine learning, cognitive computing, and information visualization techniques. Maintenance is an area of industrial activities that would greatly benefit from the implementation of Industry 4.0. But how far has the digital transformation progress come? In 2018, an interview study was performed with 14 representatives within the maintenance ecosystem during the Nordic maintenance fare held in Gothenburg. A similar study was performed at the fare held in 2022, in which 22 actors representing system providers, computerized maintenance management suppliers, researchers, and educators participated. The aim of the studies was to get a broad view on maintenance in the digital era, covering topics like enabling technologies, challenges as well as opportunities. This paper reports on the similarities and differences in results from the two interview studies and draws conclusions on the progress and directions of the digitalization in maintenance. The findings suggest that the progress is rather slow. Data management and decision-making capabilities forms the basis for digitalization of maintenance. The focus on sensor technology has somewhat been reduced, while the prediction was that it would have increased. Instead, the ability to communicate and share information is stressed. Advanced analytical capabilities are foreseen to have a breakthrough in five years’ time, as well as technologies for data gathering and communication. The challenges are mainly the same, i.e., related to competence, leadership, and strategy. This suggests that, to enable the digital transformation, we should focus on the formulation of appropriate business cases and initiation of pilot studies, supporting the implementation process and involving all people in the change, and securing the competence and skills by training, education, and recruitment of young people to maintenance positions.

Traditionally, solutions for Industrial Artificial Intelligence (IAI) in railways focus on productivity improvements and single-loop learning. This is mainly achieved by the implementation of IAI in the technical rail system and its operation, traffic management, maintenance, and modification. These productivity improvements are limited to doing things the right way or better according to existing regulations. However, to support the implementation of these solutions and keep pace with the fast technological development (e.g., by reducing the pacing problem), IAI should also be used to manage effectiveness improvements and double-loop learning. Hence, IAI should be used in the management of regulations (e.g., based on technical specifications for interoperability, TSI) according to process-related regulations for dependability and safety (e.g., EN 50126/28/29 and Common Safety Methods, CSM). Thereby, IAI can change the traditional evolutionary management of railway regulations, where it tends to expand gradually based on experienced risks, incidents, and accidents. In addition, IAI can also support management in how to decide upon what the right things to do are by triple-loop learning. This might be achieved by using relevant theories in managing risks related to internal control, i.e., effectiveness, productivity, compliance, and reporting. This paper presents an integrated enterprise risk management framework and approach for the future railway, including the use of four different levels of IAI for continuous improvement and organizational learning. The applied approach is deductively based on a literature review in databases for regulations, standards, and scientific publications. The work is inductively supported by empirical examples, mainly from the Reality lab digital railway at Trafikverket (the Swedish transport administration). The result is an integrated enterprise risk management framework that should be applied to support the management of requirements related to risk in the railway when working with continuous improvement supported by IAI.

In the process of developing digital twin for maintenance, there is a lack of reference to digital twin, architecture, and models in standards. In particular, the application also differs depending on needs of respective organisations. Before implementing the design and implementation of digital twin, it is necessary to define the user specifications and requirements. Hence, the first objective is to provide the digital twin terminology for maintenance based on the digital twin five-dimension model; physical, virtual, data, connection, and services. Due to the distinctive possibilities associated with the Digital Twin, their design and implementation are also wide-ranging. It can be classified based on dimensions. Hence, the second objective is to classify based on including several factors, such as, life cycle stages, completeness etc. In addition, the capability of digital representation to the physical asset is also of interest. Hence, the third objective is to assess the maturity level of DT. This paper provides a guideline in defining the DT with standardization, classification, and maturity level in the practical industrial applications.

The advanced technologies available in the development of Smart Maintenance within Industry 4.0 have the potential to significantly improve the efficiency of industrial maintenance. However, it is important to be careful when deciding which technologies to implement for a given application and when evaluating the quality of the data generated. Otherwise, what should be cost-effective solutions may end up being cost-driving. The use of domain knowledge in selecting, developing, implementing, setting up, and utilizing these technologies is increasingly important for achieving success. In this paper, we will elaborate on this topic by presenting and analyzing insights from industrial cases, drawing on the authors’ extensive experience in the field.

Machine learning algorithms are creating new approaches to address issues faced by the industry. These methods are data hungry for the right quality of data. Light Detection and Ranging (LiDAR) systems are becoming popular for collecting spatial data from vast areas. Processing of such large datasets and extraction of assets previously depended on tuning algorithms and development of features. Deep learning techniques free the user from the burden of generating and hand-designing features to extract assets of interest but require labelled data sets for training. Data augmentation is a popular technique for generating labelled data sets from available data and information. Data augmentation becomes the key to use machine learning by providing labelled point cloud data. Linear assets such as power lines, pipelines, railway tracks and roads extend over large areas and are designed as per set of specifications. Data augmentation for linear assets poses requirements set on specifications and simple data augmentation techniques do not fulfil the requirements. The objective of this paper is to explore available data augmentation techniques for image and point cloud and access their applicability to generate augmented point cloud data for linear assets specifically railways.

The use of greenhouse gas emissions as a criterion for decision-making within the rail industry is increasing. The demand for the consideration of this criterion affects the type of decision models acceptable by railway infrastructure managers in the planning, construction, and maintenance of railway assets. The total amount of greenhouse gas emitted from a track solution in tunnels during its service life depends on the track form (i.e., ballasted track or the ballastless track), the type of construction, maintenance machines used, the current traffic profile, and the length of the tunnel. However, the development in the design of ballastless track systems during recent decades to make them environmentally friendly is a motivation for infrastructure managers to rethink and consider the use of the system. This paper addresses the effect of applying the optimized ballastless track system Rheda 2000® in a railway tunnel (the Hallsberg-Stenkumla tunnel) as part of a new line project in Sweden. The greenhouse gas emissions, represented by life cycle CO2 equivalent emission is calculated using the climate impact software developed by the Swedish Transport Administration. The result is compared with the estimated emission from the conventional ballasted tracks. The study shows that CO2 equivalent emissions by a ballastless track during its life cycle is 10% lower than that of the ballasted track. The primary total emission driver for both track form solutions is the emissions generated at the manufacturing of rails. The second important emission factor for the ballasted track solution is the emission from the renewal of the track form during its life cycle. The second important emission factor for the ballastless track solution is concrete manufacturing. The model applied in the study is an integral part of an integrated decision support system for effectively selecting track solutions from a lifecycle perspective.

Condition-Based Maintenance (CBM) is a maintenance strategy that reduces equipment downtime, production loss, and maintenance cost based on changes in equipment condition (e.g., changes in vibration, changes in power usage, changes in operating performance, changes in temperatures, changes in noise levels, changes in chemical composition, increase in debris content and changes in the volume of material). In this study, we present the newly developed Condition Monitoring Model (CMM) based on an ensemble machine-learning model that utilizes the random forest, support vector machine, and artificial neural network classifiers, to classify data points from the normal state of a rotating machine. The efficacy of the model in adequately detecting and diagnosing faults in the rotating machine for maintenance planning is discussed in this paper. The developed model can efficiently avoid unnecessary maintenance and make timely actions by analyzing the received vibration signals from the rotating machine. An illustrative example is demonstrated to present the application of the model.

With the advancements in artificial intelligence and the emergence of shallow and deep learning algorithms, there is a growing demand for precise and efficient methods of predicting asset degradation across various industries. This has led to a resurgence of interest in artificial neural networks (ANNs) as a solution. In this study, the aim is to evaluate the potential of using ANNs, as well as specific types of ANNs that are equipped to handle sequential data, such as Recurrent Neural Networks (RNNs), Long Short-Term Memory (LSTMs), and Gated Recurrent Units (GRUs), for predicting individual track geometry degradation in railway systems. The performances of these ANNs were evaluated by comparing their ability to predict degradation patterns for 110 segments obtained from a 30-km track section in Northern Sweden. Hyperparameters, which include the number of hidden layers, the number of neurons per layer, the learning rate, the activation function, batch size, and the optimizer, play a crucial role in defining the architecture and behaviour of a neural network. Hyperparameter tuning can significantly impact the accuracy and generalization ability of the ANNs. Therefore, the impact of hyperparameter tuning on the performance of each algorithm was also explored. The results indicated that GRU outperformed simple RNN, LSTM, and feedforward ANN in terms of prediction accuracy in predicting track geometry degradation. The results provide insights into using different ANN algorithms to predict asset degradation, emphasizing the importance of proper hyperparameter tuning in achieving optimal performance.

Large-scale alumina refineries use strategic planning to forecast production plans for short, medium, and long term operational decisions. However, actual production deviates from the forecast due to reasons within Supplier, Input, Process, Output and Contractor (SIPOC) related variations including unplanned downtimes, issues with supply chain disruptions, availability of staff and demand fluctuations due to numerous factors including environmental changes, if any. The unreliable production process results in lost revenue and adversely affects the corporate image. This paper presents a statistical approach applying the Weibull model to identify the causes of production deviation and find improvement opportunities for reducing costs and risks while enhancing performance. An illustrative example from a chemical alumina refinery plant in Australia is presented. The various steps used in the analysis are discussed in this paper using illustrative example where production data is analysed and compared for diverse options of interventions for a robust and effective method for managers to better understand the gaps for monitoring and assuring plant performance.

Effective asset management (AM) is essential to achieve operational excellence in the context of large complex cyber-physical systems. It involves coordinated activities across all life-cycle stages, including information and service exchange within and between adjacent air operations domains, for effective air operations and collaboration. For enterprises and military air operations AM presuppose information and service exchange between the adjacent air operations domains. Important aspects of aviation AM are Cyber-physical aircraft systems (CPS), which include Integrated Vehicle Health Management (IVHM), CPS models, and Digital Twins (DTs), as central concepts used to predict and optimize asset performance. This paper provides an overview of the aviation AM phenomenology, environment, and challenges, and how they may impact envisioned realisations of effective AM solutions for support of enterprise air vehicle operations. A comprehensive approach is presented, comprising interdependent dimensions of the problem domain, and how they may be integrated into a framework and a platform concept addressing aviation AM needs.

Ontology development plays a vital role as it provides a structured way to represent and organize knowledge. It has the potential to connect and integrate data from different sources, enabling a new class of AI-based services and systems such as decision support systems and recommender systems. However, in large manufacturing industries, the development of such ontology can be challenging. This paper presents a use case of an application ontology development based on machine breakdown work orders coming from a Computerized Maintenance Management System (CMMS). Here, the ontology is developed using a Knowledge Meta Process: Methodology for Ontology-based Knowledge Management. This ontology development methodology involves steps such as feasibility study, requirement specification, identifying relevant concepts and relationships, selecting appropriate ontology languages and tools, and evaluating the resulting ontology. Additionally, this ontology is developed using an iterative process and in close collaboration with domain experts, which can help to ensure that the resulting ontology is accurate, complete, and useful for the intended application. The developed ontology can be shared and reused across different AI systems within the organization, facilitating interoperability and collaboration between them. Overall, having a well-defined ontology is critical for enabling AI systems to effectively process and understand information.

This paper summarises discoveries in the project “Climate Resilient Railway Infrastructure: A System Innovation Approach, CliMaint2Innovate ( www.ltu.se/climaint2innovate ). The project aimed to develop a system-level innovation roadmap and implementation plan for the CliMaint ( www.ltu.se/climaint ) system innovation. The innovation is a decision support system (DSS) that will target adaptation to climate change for railway infrastructure. To achieve the project aim, barriers and enablers related to five system innovation pillars have been explored, identified, and analysed. A literature review and two interview studies have been performed. In total, 45 experts from a population with knowledge in infrastructure and transportation networks have been selected for the study. The interviews have been analysed and the related barriers and enablers for implementation associated to each system innovation pillar have been identified. It was concluded that when identifying barriers and enablers to introducing new business models, two categories stand out: (1) barriers related to the external environment, e.g., policy and industry and low digitalisation readiness level, and (2) barriers related to the internal environment, e.g., tendering process, business models, and client-supplier relationships. The barriers and enablers regarding policy and regulation concerning the implementation of the CliMaint system innovation for climate change adaptation are related to policy, regulations, process, digitalisation, business models, and finance. The knowledge concerning barriers and enablers will further be used to suggest appropriate measures to accelerate the implementation, upscaling, and dissemination of the CliMaint system innovation.

Condition monitoring together with predictive maintenance of bearings and other equipment used by the industry avoids severe economic losses resulting from unexpected failures, greatly improves the system reliability and allows a more efficient usage of human experts’ time. This paper describes a predictive maintenance system, based on a data science approach. The system was developed and tested on a single real paper machine, and then verified with multiple external validations. Results show a proper behaviour of the approach on predicting different machine states with high accuracy.

This paper presents a dependability management framework and a system model for railways based on related regulations and standards. The applied approach is based on a literature review in databases for regulations, standards, and scientific publications. The result is a dependability management framework of regulations and standards that should be applied to fulfil requirements related to railway infrastructure availability throughout its life cycle. The dependability framework is supported by a generic rail system model, which summarises requirements, functions, characteristics, capabilities and performance. The framework and model are mainly exemplified with results from the Reality lab digital railway at Trafikverket (the Swedish transport administration). Besides compliance with existing regulations, the application of standards supports fulfilment of requirements related to internal control for authorities, i.e., effectiveness, productivity, and documentation. The proposed application of mentioned dependability standards should also support continuous asset management improvement according to the ISO 55000-series. The framework can also be used within research and development (R&D) projects to support results that consider mandatory requirements and facilitate implementation.

Maintenance management in modern industry has become an increasingly important and complex activity, due to the enhanced use of automation within a manufacturing infrastructure. The significance of effective maintenance activities is compounded by the introduction of new technologies to further enhance competitiveness and profitability in the global marketplace. These new technologies may be classified by the term Industry 4.0 and have primarily been applied to production facilities. However, the agility of this technology allows applications which improve maintenance performance—yet the rate of implementation in this area is much reduced. This has been due to several reasons which include, amongst others, a return on investment; training; a lack of strategic planning and mistrust within the traditional workforce. This paper will review the different technologies associated with Industry 4 (I4.0) and discuss the benefits to manufacturing, in particular maintenance management, and the possible barriers and enablers to adopting certain aspects of I4.0 to create a smart factory to improve productivity through better maintenance. In addition, the paper will review the important change of perspective which has emerged with Industry 5.0 (I5.0), which acknowledges the importance of human centric and sustainable technology. Finally, this paper attempts to understand if this change in perspective will transform and increase the use of I4.0 technologies, by providing a platform which promotes organisational and societal improvement.

The 2-Steps Smart Rotor Fault Diagnosis Model (SRFDM) is proposed. This consists of a supervised classic pattern recognition artificial neural network (ANN), which uses parameters extracted from the measured vibration signals from the machine. The Step-1 identifies the machine is healthy or faulty, and then the classification of faults in the Step-2 is performed. Earlier studies have used both time and frequency domain parameters as the input vectors to the ANN model. Currently these parameters are normalised with the speed synchronous vibration amplitude from the frequency domain analysis to remove the influence of the machine unbalance due to change in the machine speeds. Hence, the proposed model is likely to be applied to a typical machine irrespective of the machine operating speeds.

Climate change poses various challenges to the operation and maintenance of rail infrastructure assets. The impact of climate change and the implementation of associated measures depend on geographical locations, demography, etc. Therefore, a climate change impact analysis needs to be performed at both local and regional infrastructures. In this paper, the most common climatic impact on railway infrastructure assets of urban areas, including rail, switches and crossing, have been considered. The data were collected through a nationwide questionnaire survey on the Swedish railway infrastructure, railway maintenance entrepreneurs, and municipalities to assess the risks posed by the climate change. A framework of risk and vulnerability assessment is developed for railway infrastructure. The study indicates that track buckling and vegetation fires that occur due to the impact of hot temperatures and rail defects and breakage resulting from exposure to the cold temperatures pose a medium level of risk. On the other hand, supportability losses caused by cold temperatures are categorized with high level of risk. The impact analysis helps infrastructure managers identify and prioritize climate risks and develop appropriate climate adaptation measures and actions to cope with future climate change effects.

Oil and gas pipelines are extensively used in the energy industry. The optimal performance of these pipelines is essential to maintaining energy supply cost-effectively. In downstream operations, flow assurance is of paramount importance. Control valves are vital components of the oil and gas pipelines as they control and regulate the flow parameters. The valves have complex flow areas resulting in the generation of high shear forces which causes strong emulsification in the mixture. This emulsified mixture is difficult to separate and results in extra resources such as additional separation time and chemical additives resulting in a significant increase in the cost of the separation process. In the present work, the effect of the presence of a valve on enhanced mixing in oil and gas pipelines has been quantified. Novel indicators namely, Mixing Coefficient Mc, Modified Mixing Coefficient (MMc) and Velocity-involved Modified Mixing Coefficient (VMMc) based on the in-situ properties have been used for quantifying the mixing behaviour. The computational Fluid Dynamics based globe valve model has been simulated using different velocities and oil volume fractions. Various cross-sectional planes inside the valve and in the straight pipe are created and, Mc, MMc and VMMc are computed at those planes. The mixing behaviour of the valve has been quantified and a considerable increase in the mixing has been observed as compared to the straight pipe. Suggestions have been provided as to how to minimise the mixing effects through the design modifications.