Search

Much research is conducted on the importance of success factors. This study contributes to the body of knowledge by using artificial intelligence (AI), specifically machine learning (ML), to analyse success factors through data from construction projects. Previously conducted studies have explored the use of AI to predict project success and identify important success factors in projects; however, to the extent of the authors’ knowledge, no studies have implemented the same method as this study. This study conducts quantitative analysis on a sample of 160 Norwegian construction projects, with data obtained from a detailed questionnaire delivered to relevant project team members. The method utilises ML through a Random Forest Classifier (RFC). The findings obtained from the analysis show that it is possible to use AI and ML on a limited dataset. Furthermore, the findings show that it is possible to identify the most important success factors for the projects in question with the developed model. The findings suggest that a group of selected processes is more important than others to achieve success. The identified success factors support the theoretically acknowledged importance of thorough and early planning and analysis, complexity throughout the project, leadership involvement, and processes supporting project success.

In this paper, model-based systems engineering (MBSE) and discrete event simulation (DES) are combined to assess the performance of an offshore production system at an early stage. Various systems engineering tools are applied to an industrial case concerning the retrieval of deep-sea minerals, and a simulation engine is developed to calculate the annual production output. A mean production of 1 Million tonnes of ore per year is estimated for an operation in the Norwegian Sea using Monte Carlo simulation. Depending on the limiting design wave height of the marine operations, the estimated production output ranges from 280,000 tonnes to 1.8 Million tonnes per year. The constrained parameter of the production system is particularly the wave height operational limit of the ship-to-ship transfer operation. We present the learning outcome from applying MBSE and DES to this case and discuss important aspects for improved performance.

In this paper, model-based systems engineering (MBSE) and discrete event simulation (DES) are combined to assess the performance of an offshore production system at an early stage. Various systems engineering tools are applied to an industrial case concerning the retrieval of deep-sea minerals, and a simulation engine is developed to calculate the annual production output. A mean production of 1 Million tonnes of ore per year is estimated for an operation in the Norwegian Sea using Monte Carlo simulation. Depending on the limiting design wave height of the marine operations, the estimated production output ranges from 280,000 tonnes to 1.8 Million tonnes per year. The constrained parameter of the production system is particularly the wave height operational limit of the ship-to-ship transfer operation. We present the learning outcome from applying MBSE and DES to this case and discuss important aspects for improved performance.

In this publication, we present how the AltaRica 3.0 modelling language can be used to efficiently design a model of an offshore windmill farm and evaluate its performance. The system we consider is composed of combinations of series-parallel components, combining different states for components and different modes for parts of the system and implements complex reconfiguration strategies.Knowing the syntax and semantics of languages such as AltaRica 3.0 is however not sufficient to efficiently design models. First, models should make it possible to efficiently calculate performance indicators. Second, individual models should be designed quickly (and without bugs!) and modelling knowledge should be capitalized from models to models. With both respects, architectural and behavioural modelling patterns are of great help. The AltaRica 3.0 model we propose in this article for the assessment of an offshore windmill farm achieves both goals. We show that the design of the model is very efficient thanks to the advanced structural constructs of the AltaRica 3.0 modelling language. Finally, we use assessment tools available for AltaRica 3.0, e.g. the stochastic simulator, to evaluate the model of the system.

International audience; SYNTHESIS permet d'évaluer, puis d'optimiser les performances de sûreté d'un système complexe. Elle passe par la décomposition en briques fonctionnelles des composants. Un algorithme dédié est utilisé pour la génération d'indicateurs qualitatifs et quantitatifs représentatifs de ces performances.

International audience; This communication presents the underlying principles of the new version of Open-PSA model exchange format, which relies on the S2ML+X paradigm. S2ML stands for System Structure Modeling Language. It unifies structuring constructs coming from object-oriented and prototype-oriented programming languages. The new version of Open-PSA format, based on S2ML for its structural part, improves very significantly the previous one. Both its theoretical foundations and examples of use are provided.

International audience; Le maintien en condition opérationnelle de systèmes industriels est un facteur de compétitivité important ; et une bonne stratégie de maintenance du système assure une disponibilité élevée des équipements, tout en minimisant les coûts liés aux interventions. La modélisation et la simulation de tels systèmes, incluant leurs stratégies associées de maintenances, permet d'en étudier les indicateurs de performance sous-jacents (disponibilité, coûts, etc.). Dans cette communication, nous présentons un schéma de modélisation AltaRica 3.0 de combinaison de différentes politiques de maintenance de différents composants d'un système.

International audience; Operational condition strategies of complex systems is one of the major actual challenges. An appropriated maintenance strategy ensures a high availability of the system, while minimizing costs due to interventions. Modeling and simulation of such systems, with their strategies of maintenances, enables the study of their performance indicators, such as availability, costs, etc. This publication presents an AltaRica 3.0 modeling pattern for the combination of several maintenance policies.; Le maintien en condition opérationnelle de systèmes industriels est un facteur de compétitivité important ; et une bonne stratégie de maintenance du système assure une disponibilité élevée des équipements, tout en minimisant les coûts liés aux interventions. La modélisation et la simulation de tels systèmes, incluant leurs stratégies associées de maintenances, permet d’en étudier les indicateurs de performance sous-jacents (disponibilité, coûts, etc.). Dans cette communication, nous présentons un schéma de modélisation AltaRica 3.0 de combinaison de différentes politiques de maintenance de différents composants d’un système.

International audience

International audience; AltaRica 3.0 is an event-based, object-oriented modeling language dedicated to (probabilistic) safety analyses of complex systems. Several assessment tools have been developed for AltaRica 3.0, including a stepwise simulator. This tool is of a great help for the design and the validation of AltaRica 3.0 models. In this article, we show how the AltaRica 3.0 stepwise simulator has been greatly enhanced by the introduction of an abstract notion of time. This important result paves the way to the design of efficient model-checking algorithms, e.g. generators of sequences of events leading to a failure state.; AltaRica 3.0 est un langage de modélisation événementiel et orienté objet, dédié aux analyses probabilistes du risque de systèmes complexes. Plusieurs outils d'évaluation ont été développés pour AltaRica 3.0, y compris un simulateur pas-à-pas. Cet outil est d'une grande aide pour la conception et la validation des modèles AltaRica 3.0.Dans cet article, nous montrons comment le simulateurpas-à-pas d'AltaRica 3.0 a été grandement renforcée par l'introduction d'une notion d'abstraction du temps. Ce résultat important ouvre la voie vers la conception d'algorithmes efficaces de vérification de modèles, par exemple, pour la conception de générateurs de séquences d'événements menant à un état donné de défaillance.

International audience; Summary This article presents an efficient way to design AltaRica 3.0 models with different features: maintenance policy, limited number of repairers, etc. We use some advanced primitives of the AltaRica 3.0 modeling language to simplify the design of models with such features. This making the design activity efficient and easier. We also show that these features can be defined as modeling patterns. These modeling patterns could be reused to design other systems.; Cette communication présente comment modéliser efficacement, en AltaRica 3.0, des politiques de maintenance de plusieurs composants d'un système en prenant en compte différentes caractéristiques, telles que un nombre limité de réparateurs, des réparations conditionnelles, etc. Nous montrons comment certaines primitives évoluées du langage, peuvent être utilisées pour rendre l'activité de modélisation plus efficace et simple. Puis nous montrons comment rendre générique ces 'types' de modélisation afin de définir des schémas de modélisation. Ces schémas pouvant être ensuite repris pour modéliser d'autres systèmes.

International audience; Summary This article presents an efficient way to design AltaRica 3.0 models with different features: maintenance policy, limited number of repairers, etc. We use some advanced primitives of the AltaRica 3.0 modeling language to simplify the design of models with such features. This making the design activity efficient and easier. We also show that these features can be defined as modeling patterns. These modeling patterns could be reused to design other systems.; Cette communication présente comment modéliser efficacement, en AltaRica 3.0, des politiques de maintenance de plusieurs composants d'un système en prenant en compte différentes caractéristiques, telles que un nombre limité de réparateurs, des réparations conditionnelles, etc. Nous montrons comment certaines primitives évoluées du langage, peuvent être utilisées pour rendre l'activité de modélisation plus efficace et simple. Puis nous montrons comment rendre générique ces 'types' de modélisation afin de définir des schémas de modélisation. Ces schémas pouvant être ensuite repris pour modéliser d'autres systèmes.

International audience; AltaRica 3.0 is an event-based, object-oriented modeling language dedicated to (probabilistic) safety analyses of complex systems. Several assessment tools have been developed for AltaRica 3.0, including a stepwise simulator. This tool is of a great help for the design and the validation of AltaRica 3.0 models. In this article, we show how the AltaRica 3.0 stepwise simulator has been greatly enhanced by the introduction of an abstract notion of time. This important result paves the way to the design of efficient model-checking algorithms, e.g. generators of sequences of events leading to a failure state.; AltaRica 3.0 est un langage de modélisation événementiel et orienté objet, dédié aux analyses probabilistes du risque de systèmes complexes. Plusieurs outils d'évaluation ont été développés pour AltaRica 3.0, y compris un simulateur pas-à-pas. Cet outil est d'une grande aide pour la conception et la validation des modèles AltaRica 3.0.Dans cet article, nous montrons comment le simulateurpas-à-pas d'AltaRica 3.0 a été grandement renforcée par l'introduction d'une notion d'abstraction du temps. Ce résultat important ouvre la voie vers la conception d'algorithmes efficaces de vérification de modèles, par exemple, pour la conception de générateurs de séquences d'événements menant à un état donné de défaillance.

International audience; Summary This article invites the reader to a guided tour of AltaRica Wizard, a new integrated modeling and simulation environment for AltaRica 3.0. AltaRica 3.0 is an object-oriented modeling language dedicated to reliability engineering of complex technical systems. AltaRica Wizard provides the expected functionalities of a development environment for such a language. It manages AltaRica 3.0 models into projects that gather the different source files of a model, as well as the configuration and result files of assessment tools. It eases authoring, modifying, debugging and assessing of models by providing a unique, unified graphical interface. At the time we write these lines AltaRica Wizard embeds three tool chains to assess AltaRica 3.0 models: a stepwise simulator, a compiler to fault trees and a stochastic simulator.; Cette communication propose au lecteur une visite guidée de AltaRica Wizard, un nouvel environnement intégré de modélisation et simulation pour AltaRica 3.0. AltaRica 3.0 est un langage de modélisation orienté objet dédié aux analyses probabilistes du risque de systèmes complexes. AltaRica Wizard fourni l'ensemble des fonctionnalités attendues d'un environnement de conception pour un tel langage. Il permet de gérer les modèles AltaRica 3.0 en mode projets, qui rassemblent les différents fichiers source d'un modèle, ainsi que les fichiers de configuration et de résultats issus des outils d'évaluation. Il facilite l'écriture, la modification, la correction et l'évaluation des modèles en fournissant une interface graphique unifiée. AltaRica Wizard est actuellement relié à trois chaînes outillées pour évaluer les modèles AltaRica 3.0 : un simulateur pas-à-pas, un générateur d'arbres de défaillances ainsi qu'un simulateur stochastique.

International audience; Summary This article invites the reader to a guided tour of AltaRica Wizard, a new integrated modeling and simulation environment for AltaRica 3.0. AltaRica 3.0 is an object-oriented modeling language dedicated to reliability engineering of complex technical systems. AltaRica Wizard provides the expected functionalities of a development environment for such a language. It manages AltaRica 3.0 models into projects that gather the different source files of a model, as well as the configuration and result files of assessment tools. It eases authoring, modifying, debugging and assessing of models by providing a unique, unified graphical interface. At the time we write these lines AltaRica Wizard embeds three tool chains to assess AltaRica 3.0 models: a stepwise simulator, a compiler to fault trees and a stochastic simulator.; Cette communication propose au lecteur une visite guidée de AltaRica Wizard, un nouvel environnement intégré de modélisation et simulation pour AltaRica 3.0. AltaRica 3.0 est un langage de modélisation orienté objet dédié aux analyses probabilistes du risque de systèmes complexes. AltaRica Wizard fourni l'ensemble des fonctionnalités attendues d'un environnement de conception pour un tel langage. Il permet de gérer les modèles AltaRica 3.0 en mode projets, qui rassemblent les différents fichiers source d'un modèle, ainsi que les fichiers de configuration et de résultats issus des outils d'évaluation. Il facilite l'écriture, la modification, la correction et l'évaluation des modèles en fournissant une interface graphique unifiée. AltaRica Wizard est actuellement relié à trois chaînes outillées pour évaluer les modèles AltaRica 3.0 : un simulateur pas-à-pas, un générateur d'arbres de défaillances ainsi qu'un simulateur stochastique.

National audience; We observed a need for collaborative means between professional expertise early in the process, especially between the system architecture design and RAMS1 assessments. In this paper, we introduce concepts and conceptual tools to improve the consistency between models and allow refinement of the models used. Our approach will be illustrated on a system on board a fighting helicopter.; Nous avons observé un besoin de moyens collaboratifs entre les expertises métiers au plus tôt dans les processus, notamment entre la conception d'architectures système et les évaluations de la sûreté de fonctionnement. Dans cet article, nous introduisons des concepts et des outils conceptuels qui visent à améliorer la cohérence entre les modèles et permettent un raffinement des modèles utilisés. Notre approche est illustrée ici sur un système embarqué dans un hélicoptère de combat.

International audience; An event-based stochastic simulator is mainly composed of three mechanisms: the mechanism to update the model, to schedule the transitions, and to compute the indicators. Simulated models use these three mechanisms at various degrees. When an user wants to select a stochastic simulation tool for inclusion in its process, criteria should be established. They can take into account, in addition to the quality of results, the performance of the tool for models of the size and type of those that will be studied.To facilitate the choice of the tool according to the intended use, an evaluation kit is proposed, composed of test-cases of parameterizable sizes and corresponding to various mechanisms of the stochastic simulator tool. These test-cases are from various areas of safety assessment or automata simulation. The results of their simulations are comparable with other tools (Markov chains, fault trees, analytical formulas ...). The user can evaluate or estimate, according to its intended use, the performance and the quality of the results of the event-based stochastic simulation tool, according to the size and characteristics of its own models.; Un simulateur stochastique évènementiel est principalement composé de trois mécanismes : le mécanisme de mise à jour du modèle, de l’échéancier, et de calcul des indicateurs. Les modèles simulés font appel à des degrés divers à ces trois mécanismes. Lorsqu’un utilisateur souhaite sélectionner un tel outil afin de l’inclure dans un processus de conception, il doit établir des critères pouvant prendre en compte, en plus de la qualité des résultats obtenus, la performance de l’outil pour des modèles de la taille et du type de ceux qu’il étudiera.Pour faciliter ce choix, il est proposé un kit d’évaluation, composé de cas-tests de taille paramétrable et sollicitant différentes composantes du simulateur stochastique évènementiel. Ces cas-tests sont issus de divers domaines de la sûreté de fonctionnement ou de la simulation d’automates. Les résultats de leurs simulations sont comparables à l’aide d’autres outils (chaînes de Markov, arbres de défaillances, formules analytiques, …). L’utilisateur peut ainsi évaluer ou estimer, suivant son utilisation prévue, la performance et la qualité des résultats de l’outil de simulation stochastique évènementielle correspondant à la taille et aux caractéristiques de ses modèles propres.

International audience; An event-based stochastic simulator is mainly composed of three mechanisms: the mechanism to update the model, to schedule the transitions, and to compute the indicators. Simulated models use these three mechanisms at various degrees. When an user wants to select a stochastic simulation tool for inclusion in its process, criteria should be established. They can take into account, in addition to the quality of results, the performance of the tool for models of the size and type of those that will be studied.To facilitate the choice of the tool according to the intended use, an evaluation kit is proposed, composed of test-cases of parameterizable sizes and corresponding to various mechanisms of the stochastic simulator tool. These test-cases are from various areas of safety assessment or automata simulation. The results of their simulations are comparable with other tools (Markov chains, fault trees, analytical formulas ...). The user can evaluate or estimate, according to its intended use, the performance and the quality of the results of the event-based stochastic simulation tool, according to the size and characteristics of its own models.; Un simulateur stochastique évènementiel est principalement composé de trois mécanismes : le mécanisme de mise à jour du modèle, de l’échéancier, et de calcul des indicateurs. Les modèles simulés font appel à des degrés divers à ces trois mécanismes. Lorsqu’un utilisateur souhaite sélectionner un tel outil afin de l’inclure dans un processus de conception, il doit établir des critères pouvant prendre en compte, en plus de la qualité des résultats obtenus, la performance de l’outil pour des modèles de la taille et du type de ceux qu’il étudiera.Pour faciliter ce choix, il est proposé un kit d’évaluation, composé de cas-tests de taille paramétrable et sollicitant différentes composantes du simulateur stochastique évènementiel. Ces cas-tests sont issus de divers domaines de la sûreté de fonctionnement ou de la simulation d’automates. Les résultats de leurs simulations sont comparables à l’aide d’autres outils (chaînes de Markov, arbres de défaillances, formules analytiques, …). L’utilisateur peut ainsi évaluer ou estimer, suivant son utilisation prévue, la performance et la qualité des résultats de l’outil de simulation stochastique évènementielle correspondant à la taille et aux caractéristiques de ses modèles propres.

International audience; The complexity of modern critical systems is growing rapidly while the industry is submitted to more and more pressure for reducing costs and time-to-market. Traditional development methods " in disciplinary silos " used to design and analyze such complex systems are reaching their limits. RAMS engineers face more and more difficulties to satisfy demands of reliability evaluation especially at early stages of system design. In this context we offer to take advantage of Model-Driven Engineering (MDE) approaches to reduce construction time of reliability models and improve their consistency with system models. Model-Driven Engineering is a promising approach used to develop and analyze complex systems from different domains. In this paper, we exploit MDE to support PSA analysis and illustrate the approach on a case study from avionics industry. This experimentation has enabled the suggestion of a seamless methodology to support iterative Probabilistic Safety Analysis, thus improving the cooperation with system designers.

This document defines the S2ML language, version 1.0. S2ML is developed in the framework of the OpenAltaRica project, leaded by IRT SystemX, Palaiseau, France. S2ML is a freely available, prototype-oriented language for both representing the structure complex systems and structuring models of these systems. It aims at providing a minimal but sufficient set of constructs for these purposes. S2ML 1.0 Specification 2

International audience; Considering their increasing complexity, industrial systems are, in general, specified in a natural language. Especially transportation systems where the design phase results an ambiguous and laborious system specification. The objective of this paper is to present ScOLA, a formal modeling language based on scenarios and built on railway system specifications. Its novelty is based on its restriction to the core concepts of the specification and its multiple representations (textual and graphical), and also on its formal semantics. The language offers means to understand what the system was supposed to do and to be as well as to support a dialog with experts so to be sure that we got everything correctly. The methodology is applied on the railway automation solution Trainguard MT CBTC of Siemens.

International audience

La Fédération Interuniversitaire de l’Enseignement à Distance (FIED) est heureuse et fière d’être associée à ce numéro dont le sujet est au cœur de ses préoccupations. Publier des articles sur la distance dans une revue de qualité et dans l’une des rares revues francophones du domaine est pour la fédération un motif de satisfaction : la FIED regroupe les universités (et bientôt d’autres établissements d’enseignement supérieur) françaises (ou francophones) offrant des formations à distance en ...

International audience; .

International audience; Nous présentons dans cet article une méthode d'évaluation de l'efficience basée sur la modélisation temporelle et stochastique et de la simulation de la propagation des fautes dans les systèmes industriels. Le TRS (Taux de Rendement Synthétique) est devenu au travers de la norme NF E60-182 l'un des indicateurs majeurs de l'efficience du pilotage des systèmes de production. Il intègre essentiellement 3 notions (qualité, productivité et disponibilité). Si son expression pour un composant est assez simple, sa modélisation pour l'évaluation du comportement fonctionnel et dysfonctionnel l'est beaucoup moins (prise en compte des facteurs d'échelle, des désynchronisations et du facteur humain). Afin de permettre une prise en compte des contributions individuelles de chaque composante du TRS, nous avons utilisé les automates de mode comme haut langage de description. Les attendus débouchent en amont sur un apport formel pour l'établissement d'une méthodologie d'analyse et de conception, et en aval sur une fourniture d'indicateurs décisionnels. Les résultats sont implantés autour d'un démonstrateur basé sur AltaRica DataFlow langage à la fois formel et graphique et véritable outil de modélisation / simulation.

International audience

Hovedfokuset i denne oppgaven er på utviklingen av en additivt produsert monteringsløsning for HYPerspectral Smallsat for Ocean observation (HYPSO)-2 nyttelasten som en del av HYPSO prosjektet ved Norges Teknisk-Naturvitenskapelige Universitet (NTNU). Hovedformålet med arbeidet som presenteres er å evaluere muligheten for å bruke additiv produksjon i CubeSats og å se hvilke ytelsesøkninger innføringen av denne produksjonsteknologien kan gi prosjektet. Under utviklingen av delen ble topologioptimalisering brukt for å utnytte designfriheten additiv produksjon tillater. Denne optimeringsprosessen startet med funksjonsintegrasjon hvor designrommet som skulle optimaliseres ble laget. Deretter ble topologioptimaliseringssimuleringene gjort for å generere konsepter med lavere og lavere masse. Til slutt ble konseptene simulert for å finne forbedringsområder for manuelle endringer. Den resulterende delen fra denne optimaliseringsprosessen oppfylte kravene til de simulerte verdiene samtidig som det nye designet reduserte vekten med 81,5% sammenlignet med HYPSO-1-designet. Dette viser at additiv produksjon kan muliggjøre store ytelsesøkninger, spesielt når det gjelder massereduksjon, så lenge de tilknyttede risikoene håndteres. Siden massereduksjon er av stor betydning for satellitter, er denne produksjonsteknologien lovende i utviklingen av CubeSats. The main focus of this thesis is on the development of an additively manufactured mounting solution for the HYPerspectral Smallsat for Ocean observation (HYPSO)-2 payload as a part of the HYPSO project at Norwegian University of Science and Technology (NTNU). The main purpose of the work presented is to evaluate the feasibility of using additive manufacturing in CubeSats and to see what performance increases the addition of this manufacturing technology can bring to the project. When developing the part topology optimization was used to leverage the design freedom gained from additive manufacturing. This optimization process started with function integration where the design space to be optimized was created. Then, the topology optimization simulations were done to generate lower and lower mass concepts. This was followed by simulations of the generated concepts to find areas of improvement for manual changes. The resulting part from this optimization process met the requirements for the simulated values while reducing the weight by 81.5% compared to the HYPSO-1 design. This shows that additive manufacturing can enable great performance increases, especially in terms of mass reduction, as long as the associated risks are managed. As mass reduction is of high importance for satellites this manufacturing technology therefore shows great promise in the development of CubeSats.

Det er forventet at prediktivt vedlikehold basert på maskinlæring vil redusere kostnader relatert til maskinhavari og vedlikehold med opptil 10%. Mange studier forsøker å definere metoder og algoritmer for å beregne gjenstående levetid på ulike maskindeler eller systemer. Dette er svært utfordrende og ofte baserer slike systemer seg på ideelle datasett. Disse datasettene har gjerne målinger fra systemet i god tilstand, havari og alt i mellom. I prosessindustrien som er avhengig av høy oppetid, forsøker man derimot å unngå uplanlagt vedlikehold så godt man kan. Derfor mangler det ofte målinger fra systemer i dårligere tilstand. Det finnes da store mengder målinger fra systemet i god eller tilnærmet god tilstand. Dette gjør det vanskelig å beregne gjenstående levetid på utstyret, men åpner for å trene maskinlæringsmodeller til å predikere normaltilstanden for et utstyr slik at man kan oppdage avvik. Dermed undersøker denne masteroppgaven hvordan man kan bruke svikthistorikk og sensormålinger som allerede samles inn for maskinbeskyttelse og vibrasjonsanalyse til å oppdage sviktmønstre tidligere enn man kan i dag. Dette forsøkes i et case-studie som undersøker dataene fra ti væskeringskompressorer som er i kontinuerlig drift hos Glencore Nikkelverk. Oppgaven presenterer og forklarer behovet for å undersøke sviktmodene og tilhørende symptomer før man kan lage maskinlæringsmodeller. Basert på undersøkelsene velges det to sviktmoder, som det ønskes å lage avviksmodeller for. En enklere modell for å oppdage behov for vedlikehold i en platekjøler er laget, og det vises til sensoranalyse for å finne ut hvilke målinger som kan gi informasjon om hva som fører til lekkasje i mekaniske tetninger. Til slutt presenteres en teoretisk mulighet for rotårsaksanalyse. Det diskuteres også hvordan erfaringene som er gjort i denne case-studien kan overføres til annet utstyr og gi et startpunkt for videre utvikling av prediktivt vedlikehold. Predictive maintenance (PdM) is expected to reduce the maintenance cost by up to 10\% and reduce the frequency of unplanned production disturbances. However, these figures are not verified yet and the industry faces many challenges in implementing PdM. Especially, the availability of labeled data sets, which are necessary for predicting the remaining useful lifetime of equipment, is low. Still, long-short-term-memory (LSTM) neural networks allow early fault detection and root cause analysis based on unlabeled condition monitoring data, which is readily available in the chemical process industry. PdM requires the integration of IoT networks, databases, software, historical data and knowledge of day-to-day operations of the equipment under study. In addition, the lack of labeled condition monitoring data increases the need for system knowledge to create and evaluate early fault detection models. Thus, the thesis discusses the need for, and presents, a failure mode and symptoms analysis to understand the equipment, possible degradation indicators and how to evaluate the model while also assessing the data quality of the sources of information used. A case-study that focuses on ten liquid ring compressors, sensor data analysis and model development for fault detection for two different failure modes are then presented. The thesis also theoretically discusses how these models can be used to find a failure's root cause with feature importance analysis. Lastly, the thesis shows how the experience gained in the case-study can be used when assessing which equipment is likely to be successfully modeled with long-short-term-memory neural networks to detect early fault development and future research necessary in more difficult modeling situations.

Systems are becoming more complex over time, requiring interdisciplinary groups to develop them. Moreover, systems are more interconnected, meaning that different systems should be designed to work together with others in a System on System environment. Therefore, more engineering groups are moving from paper-based to model-based system engineering approaches. Even though using models is not new in engineering, the MBSE concept became popular after 2007 when INCOSE introduced the MBSE initiative. From now on, the main limitation MBSE has had to overcome is the integration of different models that use different languages. Even though software suites nowadays offer a high level of integration, those solutions are out of the reach for several users in terms of cost. They do not offer the flexibility that uses different but well-known model languages. This limitation is more remarkable for RAMS modeling. Even though the concept of MBSA has been used since 1990, integrating those techniques with a pragmatic approach like MBSE is not easy. The development of approaches focused on translating pragmatic models to a formal structure becomes relevant to making MBSE languages compatible with the formal ones used to get relevant RAMS calculations. Therefore, this work proposes the activity approach as an additional step in the system architecture development. The activity approach is thought to use logic and structure so that it could be used as the link between the two different model types. This document uses the Cube Architecture framework as the first step of Model-Based System Engineer and introduces the activity concept in it. The case study is the HYPSO project, developed by the NTNU small Satellite lab. Satellites have to deal with unique conditions not seen on earth. Therefore, it is required prior to the development of the activity approach to analyze the hazard and the respective hazardous events, as well as their consequences. Activity approach logic is translated to a formal structure. Subsequently, AltaRica is used to model the activity approach in a formal language.

Dette prosjektet har som formål å studere hvordan modellering og simulering kan bidra til å bedre forstå industrielle vedlikeholdsproblemer. Hensikten er å hjelpe ledere til å forstå deres problemer bedre, finne kilden til problemene, og potensielle løsninger. Til tross for en betydelig utvikling av vedlikeholdsstyringskonsepter og optimaliseringsverktøy, er det er ofte vektlagt skille mellom teori og praksis. For å adressere dette har vi utviklet et modellering og simuleringsverktøy ved bruk av konsepter fra diskret hendelsessimulering (DES) og systemdynamikk (SD). Videre har vi gjennomført fem eksperimenter for å studere egenskaper ved vedlikeholdsprosessen. Det første eksperimentet danner basisgrunnlaget for de følgende eksperimentene. Eksperimentet simulerer en akkumulering av et preventivt vedlikehold (PM) etterslep som følge av en ubalanse mellom kapasiteten til å utføre vedlikehold og PM programmene. I eksperiment to og tre studerer vi to ulike måter å redusere etterslepet. Først øker vi PM intervallet og deretter øker vi kapasiteten til å utføre vedlikehold. Begge tiltakene reduserer PM etterslepet. På en annen side reduserer det siste også det faktiske PM intervallet, som igjen øker mengden PM aktiviteter som blir gjennomført. I eksperiment fire studerer vi hvordan systemet responderer på en uventet hendelse. Resultatet viser at en høyst optimalisert vedlikeholdsprosess kan være svært sårbar for uventede hendelser og forsinkelser. Til slutt, i eksperiment fem, undersøker vi hvordan unøyaktig prioritering av PM-aktiviteter påvirker vedlikeholdsprosessen. Resultatene viser at unøyaktig prioritering kan føre til en betydelig risiko. Eksperimentet viser hvordan PM ettersleppet tilsynelatende er stabilt, og indikerer et velfungerende PM program, men uten at det blir gjennomført PM-aktiviteter. Prinsippet for håndtering av data i kombinasjon med en korrektiv virkelighet, fører til gjentatt planlegging og forskyvning av PM aktiviteter frem til enheten svikter. Videre forsking burde fortsette utviklingen av et flerperspektiv-modelleringsrammeverk og -metodikk for å forstå industrielle problemer bedre. For videre utvikling av dette anbefaler vi å utforske kombinasjonen DES og SD. Metodikken burde legge vekt på utvikling av metaformodeller og utforsking av dynamikken til variabler. Vi forestiller oss at modeller utviklet i et slikt modelleringsrammeverk kan bidra som både et læringsverktøy, og som et kommunikasjonsverktøy mellom interessenter. For videre utforsking av utfordringer med vedlikeholdsstyring anbefaler vi å utvide grensene til systemet, og inkludere for eksempel produksjon, reservedeler, kritikalitet, og myke variabler, som kompetanse og press. Dessuten, studere dynamikken til variablene som i denne innledende forskingen er betraktet som statisk. For eksempel studere nøyaktigheten i prioritering av PM i sammenheng med PM etterslep, datakvalitet, og policy for planlegging. Et samarbeid med industrielle virksomheter vil være til fordel ved videre forsking og utvikling. Et slikt samarbeid vil forbedre utviklingen ved å inkludere faktorer som påvirker deres problemer, og bidra til å sørge for nytteverdien av et slikt verktøy og modelleringsrammeverk for industrien. This project aims to study how modeling and simulation can help to understand industrial maintenance problems better. The purpose is to help managers understand their problems better, the root cause, and potential solutions. Despite substantial development of maintenance managing models and optimization tools, there is a frequently emphasized gap between theory and practice. To address this, we developed a proof-of-concept modeling and simulation tool using a combined framework of discrete event simulation (DES) and system dynamics (SD). Further on, we performed five different experiments to study features of the maintenance process. The first experiment serves as a base case for the following experiments. The experiment simulates the accumulation of preventive maintenance (PM) backlog due to an imbalance between the capacity to carry out maintenance and the PM programs. In experiments two and three, we study two different ways of reducing the backlog. First, we increase the PM interval, and secondly, increase the capacity to perform maintenance. Both approaches reduced the PM backlog. The latter, however, also reduces the actual PM interval, which increases the number of PM activities performed. Further on, in experiment four, we study how the system responds to an unexpected event. The results show that a highly optimized maintenance process might be extremely fragile to unexpected events and delays. Lastly, in experiment five, we investigate how inaccurate prioritization of PM activities affect the maintenance process. The experimental results show that inaccurate prioritization might lead to substantial risk exposure. The experiment also shows how the PM backlog appears to be in equilibrium and indicates a well-functioning PM program without any PM activities performed. In combination with the data management policy, the corrective reality leads to rescheduling of PM activities until the item's failure. Future research should continue the development of a multi-perspective modeling framework and methodology to understand industrial problems better. For the continued development of a modeling framework, we recommend exploring the combination of DES and SD. The methodology should emphasize the creation of metaphor models and exploration of the dynamics of variables. We envision models created with such a modeling framework to serve as both a learning laboratory and as a communication tool among stakeholders. Furthermore, for addressing the challenges of managing maintenance, we recommend future research to expand the system boundary, including, for instance, production, spare part policy, item criticality, and soft variables such as competence and pressure. Moreover, study the dynamics of variables considered static in this preliminary—for instance, PM prioritizing accuracy along with the PM backlog, data quality, and scheduling policy. Additionally, further development and research would benefit from collaboration with industrial companies. A collaboration would improve the development to include the factors affecting their specific problems and to ensure the usefulness of such modeling framework and tools to the industry.

Den økende forekomsten av kompleksitet i tekniske systemer krever godt samspill mellom de involverte interessentene. Industrielt systemdesign (systems engineering) er en tverrfaglig tilnærming til utvikling av balanserte systemløsninger som oppfyller ulike interessenters behov. Det er en praksis som prøver å løse komplekse og teknologisk utfordrende problemer. Modellbasert industrielt systemdesign (model-based systems engineering) er en voksende tilnærming til industrielt systemdesign hvor modellen av et system er sentrum for alle aktiviteter. Fordelene med denne tilnærmingen er mange. Selv om de fleste systemer er komplekse og dynamiske, finnes det imidlertid få modeller som er komplekse eller dynamiske. For det meste er de enkle og statiske. Denne oppgaven er fokusert rundt ScOLa, et domenespesifikt modelleringsspråk som er opprettet med formålet om å støtte systemarkitekturstudier, og gjøre det mulig å beskrive og spille scenarier. Modellering i ScOLa har blitt utført på en eksisterende planovergang for å danne et inntrykk og vurdere fordelene og nytten av denne typen modellering. I løpet av dette prosjektet har kunnskap om ScOLa blitt tilegnet. Diskusjonen i slutten av oppgaven er fokusert på hva ScOLa tilbyr sammenlignet med andre typer modeller i systemarkitekturstudier. The increasing emergence of complexity in engineering systems requires good interaction between the involved stakeholders. Systems engineering is an interdisciplinary approach to develop balanced system solutions that meets diverse stakeholders needs. It is a practice that addresses complex and technologically challenging problems. Model based systems engineering is an emerging approach to systems engineering where the model of a system is the center of all system engineering activities. The benefits of this approach are many. However, even though most systems are complex and dynamic, there exists few models that are complex or dynamic. They are mostly simple and static. This thesis is focused around ScOLa, a domain specific modelling language that is created with the intention of supporting system architecture studies and make it possible to describe and play scenarios. ScOLa has been conducted to an existing level crossing system, to form an impression and evaluate the benefits and usefulness of this type of modelling. Through the project and the experiment, knowledge about ScOLa has been acquired. The discussion is focused around what ScOLa offers compared to other types of models in system architecture studies.

Model-based design is a common methodology in the development of embedded complex control systems. Control system engineers typically prefer to use MATLAB® Simulink® and suitable automatic code generators for the development and deployment of software. Embedded systems are subject to random hardware faults; bit-flips, for example, may affect random access memory (RAM) cells and central processing unit (CPU) registers and cause data errors that may propagate to critical system outputs and result in system failures. From a dependability perspective, the design space of control systems includes the selection of a suitable (reliable) implementation of a control algorithm. Such algorithm can be implemented with model-based software development frameworks, such as Simulink using different, but functionally equivalent implementations. However, these functional equivalents may exhibit completely different reliability properties. This thesis proposes an analytical method for the evaluation of the reliability properties of control systems that are designed with Simulink models. The method is based on a transformation of the assembly code, which is generated from the Simulink model, into a formal stochastic error propagation model as well as its quantification through underlying Markov chain models and state-of-the-art probabilistic model-checking techniques. The application of the method to the functionally equivalent implementations can determine which one is less vulnerable to data errors due to random hardware faults. Fault tolerance is significant to dependable system design. Control systems can be protected with fault tolerance mechanisms to increase the reliability. Redundancy is the key underlying concept for achieving fault tolerance that is usually implemented at the hardware or software level. In the case of model-based development, redundancy mechanisms are preferable for direct application at the model level (Simulink model level). This thesis introduces a systematic classification of fault-tolerant design patterns. Such patterns can be applied to the Simulink model to tolerate random hardware faults, and taken into account during the control system design. In addition, it is more transparent and convenient for control system engineers to directly protect vulnerable parts with fault tolerance mechanisms at the model level. The rigorous reliability assessment of the embedded control systems must be conducted at the assembly level based on the modeling of data errors that occurred in RAM and CPU. However, the scalability of the assembly-level assessment method is challenging and even problematic in view of the state space explosion (SSE) problem of the underlying Markov chain models. The computational complexity may increase exponentially as the assembly code size increases. Moreover, the transformation from the Simulink models to the assembly code is a complicated procedure. It is also more convenient for control engineers to already be able to estimate reliability properties and implement possible reliability improvements at the model level in the early design phase, when the model-based design is actually applied. Therefore, this thesis proposes a model-level reliability evaluation of Simulink models to address the aforementioned problems. The efficiency of the proposed modellevel evaluation is verified by a comparison of the reliability properties that are assessed at the assembly and model levels.:1. Introduction 2. Preliminaries 3. Reliability evaluation of control algorithm implementations at the assembly level 4. Fault-tolerant design patterns 5. MORE: MOdel-based REdundancy for Simulink models 6. Model-level assessment of Simulink models 7. Conclusion

The main objective of this thesis is do maintenance optimization for a system using AltaRica simulation tools. Maintenance optimization is based on the assessment of maintenance policy. Stochastic simulation is applied to assessment the performance of the maintenance policy. The simulation process consists of several steps. First, build the AltaRica model. Second, compile the AltaRica model using compiler from AltaRica project. Third, prepare indicator description file and generate the simulator. Fourth, prepare mission description file and execute the simulator. After this process, the simulation is conducted and the result will be stored in a csv file. Maintenance optimization is based on the simulation process. Among a large number of candidate solutions for maintenance policy, metaheuristic algorithm can lift the searching efficiency for the optimum solution. Some algorithms help to skip unnecessary simulation and avoid repeated work. The combination of AltaRica simulation tools and metaheuristic algorithm realizes the efficient automation of maintenance optimization process. This project makes the connection of these two parts and achieves the objective.

Performing reliability assessments always relies on utilizing data. Most often, this data is provided in the form of historic failure dates. To understand this data, models are used to derive reliability characteristics from it. These models can be parametric, trying to describe the system by means of mathematical equations. They can also be empirical, letting the raw data describe the system without assuming a certain outcome. Handling parametric models is convenient, as they are described by often just one value. Empirical probability distributions are built on all available data and hence requires them to be fully defined. Handling this amount of data is cumbersome. Part of this thesis is proposing different methods to represent the empirical reliability estimator. These representations try to combine convenient usage while keeping accuracy. Representing the empirical reliability graph by a reduced amount of linear segments is proposed and discussed. This is an efficient way to compress huge datasets to a low number of descriptive points to interpolate in. Furthermore, the feasibility to use polynomial regression on the empirical probability distribution is evaluated. The computational efficiency of all methods is compared. For all practical purposes, the time to retrieve a reliability estimate is negligible. Parametric and empirical approaches are applied to various datasets and the results discussed. The empirical methods outperform the exponential estimator in all cases. The given experiment hypothesis is validated on each of the four experiments: The empirical probability distributions do match sufficiently well the reference reliability and the computational efficiency is negligible for all practical purposes.

This master thesis investigates definitions and classifications of the common cause failure from various industries. From the investigation, four core aspects of the definition for the common cause failure are extracted, and improved classifications for root causes and coupling factors of the common cause failure are suggested. This thesis also explores a framework for the inclusion of the impact of the common cause failure in risk and reliability evaluations. From the investigation of the framework, the relationship among definition of the common cause failure, causes of the common cause failure, common cause failure parametric model, parameter estimation and system model is provided. Especially, this thesis focuses on the common cause failure parametric model that is used to quantify common cause failure effect on the risk assessment. The thesis investigates popular common cause failure parametric models (Beta Factor model, Alpha Factor model, Multiple Greek Letter model, and Multiple Beta Factor model.), and discusses their features. After the investigation of common cause failure parametric model, this thesis focuses on the Beta Factor Model and its parameter estimation. For the parameter estimation, two conventional approaches (data based and checklist based) and a relatively new approach (PDS method) are presented. The main contribution of this thesis is ⅰ) to identify challenges of the PDS estimator in the PDS method and ⅱ) to suggest improved approach for determining the value of PDS estimator. This thesis identifies underlying assumptions and constraints in PDS estimators, and proposes improved approach to determine CCF group size. Case study is conducted to apply this improved approach. This new approach is realized as a computer program using Microsoft Excel and its embedded Visual Basic, so that the industries can easily use the new approach.

This thesis links two fields : system safety analyses and formal methods.We aim at checking the consistensy of safety analyses based on formal models that represent a system at different levels of detail. To reach this objective, we introduce a refinement notion in the AltaRica modelling process : a detailed model refines an abstract model if the abstract model simulates the detailed model. The AltaRica model refinement verification is supported by the MecV model-checker. This allows to perform multi-system safety analyses using models with heterogeneous levels of detail : the main system is detailed whereas the interfaced systems remain abstract. This approach has been applied to the analysis of a rudder control system linked to an electrical power generation and distribution system., Cette thèse se situe au croisement de deux domaines : la sûreté de fonctionnement des systèmes critiques et les méthodes formelles. Nous cherchons à établir la cohérence des analyses de sûreté de fonctionnement réalisées à l'aide de modèles représentant un même système à des niveaux de détail différents. Pour cela, nous proposons une notion de raffinement dans le cadre de la conception de modèles AltaRica : un modèle détaillé raffine un modèle abstrait si le modèle abstrait simule le modèle détaillé. La vérification du raffinement de modèles AltaRica est supportée par l'outil de model-checking MecV. Ceci permet de réaliser des analyses multi-systèmes à l'aide de modèles à des niveaux de détail hétérogènes : le système au centre de l'étude est détaillé tandis que les systèmes en interface sont abstraits. Cette approche a été appliquée à l'étude d'un système de contrôle de gouverne de direction d'un avion connecté à un système de génération et distribution électrique.