Search

Designing a hybrid powertrain remains a complex task. It is an intricate system involving numerous variables that are spread over different levels: architecture, component technologies, sizing, and control. There is currently a lack of frameworks or tools that help in exploring the entire design space and in finding the global optimal solution throughout these levels. This article proposes a systematic methodology that tries to answer a part of this need. Starting from a set of chosen components, the methodology automatically generates all the possible graphs of architectures using constraint-programming techniques. A tailored representation is developed to picture the graphs. They are then transformed into other types of representation (tables describing the connections and the powertrain modes). Based on these representations, the architectures are automatically filtered and the most promising ones are selected. They are automatically assessed and optimized using a specifically developed general hybrid model that calculates the performance and fuel consumption of all the generated architectures. This model is inserted inside a bi-level optimization process: a Genetic Algorithm is used on the sizing and components level, while the Dynamic Programming is used on the control level. A case study is performed and the capability of the methodology is proven.

Volvo Construction Equipment is one of the leading construction machinery manufacturers in the world. To stay amongst the leaders, research and development projects for new technologies are crucial. The most important path of development today is the reduction of emissions produced by these heavy duty vehicles. To tackle this challenge, several technologies are already used in industry. One example are hybrid machines that combine a conventional diesel engine with batteries, resulting in reduced engine size and pollutants. Another option are full battery-electric vehicles, which can reduce the on-site emissions to zero. The electrochemical processes within batteries are however comparable slow and result in long recharge times. A new focus of development within the industry are hybrid systems combining fuel cells and batteries. Since hydrogen can be refueled almost as fast as convenient fuel, it solves the issue of long recharge times. Additionally, the reaction is emission free, since there is no combustion process and the only byproduct that is emitted from the fuel cell is chemically clean water. This thesis aims to propose an architecture and packaging solution to replace the diesel engine in a large size wheel loader with a fuel cell power system. This also includes all respective auxiliary systems, i.e. energy storage, cooling and electric systems. Achieving the same performance as a conventional large size wheel loader as well as keeping the spatial envelope the same are the main objectives of this work. To achieve these goals, an extensive study on the most common drive cycles is carried out to understand the power demand of the machine. After the selection of an energy storage system based on a MATLAB simulation script, a cooling system is modelled and scaled to fulfill the operating requirements of the different components. Eventually, all systems are modeled and installed into the wheel loader in CATIA V5. The study showed, that the new system architecture of the vehicle can be fitted into the existing engine bay with a slight extension of the rear frame and hood. Two power optimized batteries are combined with one fuel cell. Hydrogen tanks with a filling volume of 478 [L] can be installed in the machine, covering 50% of the customer population curve without degradation of performance. This includes one refill of the wheel loader during the day. The performance parameters match the conventional machine up to a high degree, concluding that the conversion of a large size wheel loader into a fuel cell powered wheel loader is feasible.

To meet the vehicle fleet-wide average CO2 targets, the stringent pollutant emissions standards, and the clients’ new demands, the automakers realized the inevitable need to offer more hybrid and electric powertrains. Designing a hybrid powertrain remains however a complex task. It is an intricate system involving numerous variables that are spread over different levels: architecture, component technologies, sizing, and control. The industry lacks frameworks or tools that help in exploring the entire design space and in finding the global optimal solution on all these levels. This thesis proposes a systematic methodology that tries to answer a part of this need. Starting from a set of chosen components, the methodology automatically generates all the possible graphs of architectures using constraint-programming techniques. A tailored representation is developed to picture these graphs. The gearbox elements (clutches, synchronizer units) are represented with a level of details appropriate to generate the new-trend dedicated hybrid gearboxes, without making the problem too complex. The graphs are then transformed into other types of representation: 0ABC Table (describing the mechanical connections between the components), Modes Table (describing the available modes in the architectures) and Modes Table + (describing for each available mode the global efficiency and ratio of the power flow between all the components). Based on these representations, the architectures are filtered and the most promising ones are selected. They are automatically assessed and optimized using a general hybrid model specifically developed to calculate the performance and fuel consumption of all the generated architectures. This model is inserted inside a bi-level optimization process: Genetic Algorithm GA is used on the sizing and components level, while Dynamic Programming DP is used on the control level. A case study is performed and the capability of the methodology is proven. It succeeded in automatically generating all the graphs of possible architectures, and filtering dismissed architectures that were then proven not efficient. It also selected the most promising architectures for optimization. The results show that the proposed methodology succeeded in finding an architecture better than the ones proposed without the methodology (consumption about 5% lower); Pour répondre aux objectifs de consommation des flottes de véhicules, au normes d’émissions de polluants et aux nouvelles demandes de l’usager, les constructeurs automobiles doivent développer des motorisations hybrides et électriques. Réaliser une chaine de traction hybride reste cependant une tâche difficile. Ces systèmes sont complexes et possèdent de nombreuses variables réparties sur différents niveaux : architecture, technologie des composants, dimensionnement et contrôle/commande. L’industrie manque encore d’environnements et d’outils pouvant aider à l’exploration de l’ensemble de l’espace de dimensionnement et à trouver la meilleure solution parmi tous ces niveaux. Cette thèse propose une méthodologie systématique pour répondre au moins partiellement à ce besoin. Partant d’un ensemble de composants, cette méthodologie permet de générer automatiquement tous les graphes d’architectures possibles en utilisant la technique de programmation par contraintes. Une représentation dédiée est développée pour visualiser ces graphes. Les éléments de boites de vitesse (embrayages, synchroniseurs) sont représentés avec un niveau de détails approprié pour générer de nouvelles transmission mécaniques sans trop complexifier le problème. Les graphes obtenus sont ensuite transformés en d’autres types de représentation : 0ABC Table (décrivant les connections mécaniques entre les composants), Modes Table (décrivant les modes de fonctionnement disponibles dans les architectures) et Modes Table + (décrivant pour chaque mode le rendement et le rapport de réduction global des chemins de transfert de l’énergie entre tous les composants). Sur la base de cette représentation, les nombreuses architectures générées sont filtrées et seules les plus prometteuses sont sélectionnées. Elles sont ensuite automatiquement évaluées et optimisées avec un modèle général spécifiquement développé pour calculer les performances et la consommation de toute les architectures générées. Ce modèle est inséré dans un processus d’optimisation à deux niveaux ; un algorithme génétique GA est utilisé pour le dimensionnement des composants et la programmation dynamique est utilisée au niveau contrôle (gestion de l’énergie) du système. Un cas d’étude est ensuite réalisé pour montrer le potentiel de cette méthodologie. Nous générons ainsi automatiquement toutes les architectures qui incluent un ensemble de composants défini à l’avance, et le filtrage automatique élimine les architectures présupposées non efficaces et sélectionnent les plus prometteuses pour l’optimisation. Les résultats montrent que la méthodologie proposée permet d’aboutir à une architecture meilleure (consommation diminuée de 5%) que celles imaginées de prime abord (en dehors de toute méthodologie)

To meet the vehicle fleet-wide average CO2 targets, the stringent pollutant emissions standards, and the clients’ new demands, the automakers realized the inevitable need to offer more hybrid and electric powertrains. Designing a hybrid powertrain remains however a complex task. It is an intricate system involving numerous variables that are spread over different levels: architecture, component technologies, sizing, and control. The industry lacks frameworks or tools that help in exploring the entire design space and in finding the global optimal solution on all these levels. This thesis proposes a systematic methodology that tries to answer a part of this need. Starting from a set of chosen components, the methodology automatically generates all the possible graphs of architectures using constraint-programming techniques. A tailored representation is developed to picture these graphs. The gearbox elements (clutches, synchronizer units) are represented with a level of details appropriate to generate the new-trend dedicated hybrid gearboxes, without making the problem too complex. The graphs are then transformed into other types of representation: 0ABC Table (describing the mechanical connections between the components), Modes Table (describing the available modes in the architectures) and Modes Table + (describing for each available mode the global efficiency and ratio of the power flow between all the components). Based on these representations, the architectures are filtered and the most promising ones are selected. They are automatically assessed and optimized using a general hybrid model specifically developed to calculate the performance and fuel consumption of all the generated architectures. This model is inserted inside a bi-level optimization process: Genetic Algorithm GA is used on the sizing and components level, while Dynamic Programming DP is used on the control level. A case study is performed and the capability of the methodology is proven. It succeeded in automatically generating all the graphs of possible architectures, and filtering dismissed architectures that were then proven not efficient. It also selected the most promising architectures for optimization. The results show that the proposed methodology succeeded in finding an architecture better than the ones proposed without the methodology (consumption about 5% lower), Pour répondre aux objectifs de consommation des flottes de véhicules, au normes d’émissions de polluants et aux nouvelles demandes de l’usager, les constructeurs automobiles doivent développer des motorisations hybrides et électriques. Réaliser une chaine de traction hybride reste cependant une tâche difficile. Ces systèmes sont complexes et possèdent de nombreuses variables réparties sur différents niveaux : architecture, technologie des composants, dimensionnement et contrôle/commande. L’industrie manque encore d’environnements et d’outils pouvant aider à l’exploration de l’ensemble de l’espace de dimensionnement et à trouver la meilleure solution parmi tous ces niveaux. Cette thèse propose une méthodologie systématique pour répondre au moins partiellement à ce besoin. Partant d’un ensemble de composants, cette méthodologie permet de générer automatiquement tous les graphes d’architectures possibles en utilisant la technique de programmation par contraintes. Une représentation dédiée est développée pour visualiser ces graphes. Les éléments de boites de vitesse (embrayages, synchroniseurs) sont représentés avec un niveau de détails approprié pour générer de nouvelles transmission mécaniques sans trop complexifier le problème. Les graphes obtenus sont ensuite transformés en d’autres types de représentation : 0ABC Table (décrivant les connections mécaniques entre les composants), Modes Table (décrivant les modes de fonctionnement disponibles dans les architectures) et Modes Table + (décrivant pour chaque mode le rendement et le rapport de réduction global des chemins de transfert de l’énergie entre tous les composants). Sur la base de cette représentation, les nombreuses architectures générées sont filtrées et seules les plus prometteuses sont sélectionnées. Elles sont ensuite automatiquement évaluées et optimisées avec un modèle général spécifiquement développé pour calculer les performances et la consommation de toute les architectures générées. Ce modèle est inséré dans un processus d’optimisation à deux niveaux ; un algorithme génétique GA est utilisé pour le dimensionnement des composants et la programmation dynamique est utilisée au niveau contrôle (gestion de l’énergie) du système. Un cas d’étude est ensuite réalisé pour montrer le potentiel de cette méthodologie. Nous générons ainsi automatiquement toutes les architectures qui incluent un ensemble de composants défini à l’avance, et le filtrage automatique élimine les architectures présupposées non efficaces et sélectionnent les plus prometteuses pour l’optimisation. Les résultats montrent que la méthodologie proposée permet d’aboutir à une architecture meilleure (consommation diminuée de 5%) que celles imaginées de prime abord (en dehors de toute méthodologie)

International audience; With the emergence of environmental legislations in many countries, the importance placed upon environmental protection has been raised to a new level, especially for industrial activities. Considering environmental issues as early as possible, starting with the design stage, is expected in order to better manage and diminish adverse environmental impact. Commensurate progress has been made in method/tool development for use in environmental impact estimation; however, very few of these methods allow integrating this estimation early in the design process – a critical point of deciding for potential product concepts and suppliers. In this paper, we propose a tool that integrates environmental impact estimation into architecture and supplier identification, in order to conjointly consider requirements satisfaction as well as uncertainty due to new module and new supplier integration. This tool is developed to support OEM (Original Equipment Manufacturer) decision-making in the context of an extended enterprise. A case study is presented to illustrate a plausible implementation.

There is a current need for Unmanned Aerial Vehicle (UAV) systems architecture generation and assessment models. Architecture assessment models that presently exist tend to be fractional and do not account for all dynamic attributes that should be considered in the architecture assessment. There are infinite possibilities of architecture assessment modeling methods that can be used to assess conceptual systems architecture. Developing a good architecture assessment model for evaluating the functional and systems architecture reduces system ambiguity while increasing the tangibility of the system. This paper will present an assessment model specifically for UAV systems and this model can also be adapted to virtually any type of complex adaptive system. On new complex UAV systems it is essential evaluate the appropriate Level of Autonomy (LOA) required to satisfy customer and mission requirements. The assessment model detailed in this paper combines known design heuristics with quantitative, qualitative, and visual representations that assess the probability that the generated architecture will meet performance and capability requirements. The architecture assessment model will assess and generate a combined score that will indicate if the architecture is acceptable or unacceptable. Future UAV systems may reach a LOA that would be considered as Artificial Intelligence (AI); and this pursuit will increase the need for advanced LOA architecture modeling and assessment.