Yes / This paper presents the development and implementation of an Analytical Target Cascading (ATC) Multi-disciplinary Design Optimisation (MDO) framework for the steady state engine calibration optimisation problem. The case is made that the MDO / ATC offers a convenient framework for the engine calibration optimisation problem based on steady state engine test data collected at specified engine speed / load points, which is naturally structured on 2 hierarchical levels: the “Global” level, associated with performance over a drive cycle, and “Local” level, relating to engine operation at each speed / load point. The case study of a gasoline engine equipped with variable camshaft timing (VCT) was considered to study the application of the ATC framework to a calibration optimisation problem. The paper describes the analysis and mathematical formulation of the VCT calibration optimisation as an ATC framework, and its Matlab implementation with gradient based and evolutionary optimisation algorithms. The results and performance of the ATC are discussed comparatively with the conventional two-stage approach to steady state calibration optimisation. The main conclusion from this research is that ATC offers a powerful and efficient approach for engine calibration optimisation, delivering better solutions at both “Global” and “Local” levels. Further advantages of the ATC framework is that it is flexible and scalable to the complexity of the calibration problem, and enables calibrator preference to be incorporated a priori in the optimisation problem formulation, delivering important time saving for the overall calibration development process. / The research work presented in this paper was funded by UK Technology Strategy Board (TSB) through the CREO (Carbon Reduction through Engine Optimisation) project.

No / This paper presents the development of an Analytical Target Cascading (ATC) Multidisciplinary Design Optimization (MDO) framework for a steady-state engine calibration optimization problem. The implementation novelty of this research is the use of the ATC framework to formulate the complex multi-objective engine calibration problem, delivering a considerable enhancement compared to the conventional 2-stage calibration optimization approach [1]. A case study of a steady-state calibration optimization of a Gasoline Direct Injection (GDI) engine was used for the calibration problem analysis as ATC. The case study results provided useful insight on the efficiency of the ATC approach in delivering superior calibration solutions, in terms of “global” system level objectives (e.g. improved fuel economy and reduced particulate emissions), while meeting “local” subsystem level requirements (such as combustion stability and exhaust gas temperature constraints). The ATC structure facilitated the articulation of engineering preference for smooth calibration maps via the ATC linking variables, with the potential to deliver important time saving for the overall calibration development process.

Automotive engines are becoming increasingly technically complex and associated legal emissions standards more restrictive, making the task of identifying optimum actuator settings to use significantly more difficult. Given these challenges, this research aims to develop a process for engine calibration optimisation by exploiting advanced mathematical methods. Validation of this work is based upon a case study describing a steady-state Diesel engine calibration problem. The calibration optimisation problem seeks an optimal combination of actuator settings that minimises fuel consumption, while simultaneously meeting or exceeding the legal emissions constraints over a specified drive cycle. As another engineering target, the engine control maps are required as smooth as possible. The Multidisciplinary Design Optimisation (MDO) Frameworks have been studied to develop the optimisation process for the steady state Diesel engine calibration optimisation problem. Two MDO strategies are proposed for formulating and addressing this optimisation problem, which are All At Once (AAO), Collaborative Optimisation. An innovative MDO formulation has been developed based on the Collaborative Optimisation application for Diesel engine calibration. Form the MDO implementations, the fuel consumption have been significantly improved, while keep the emission at same level compare with the bench mark solution provided by sponsoring company. More importantly, this research has shown the ability of MDO methodologies that manage and organize the Diesel engine calibration optimisation problem more effectively.

Automotive engines are becoming increasingly technically complex and associated legal emissions standards more restrictive, making the task of identifying optimum actuator settings to use significantly more difficult. Given these challenges, this research aims to develop a process for engine calibration optimisation by exploiting advanced mathematical methods. Validation of this work is based upon a case study describing a steady-state Diesel engine calibration problem. The calibration optimisation problem seeks an optimal combination of actuator settings that minimises fuel consumption, while simultaneously meeting or exceeding the legal emissions constraints over a specified drive cycle. As another engineering target, the engine control maps are required as smooth as possible. The Multidisciplinary Design Optimisation (MDO) Frameworks have been studied to develop the optimisation process for the steady state Diesel engine calibration optimisation problem. Two MDO strategies are proposed for formulating and addressing this optimisation problem, which are All At Once (AAO), Collaborative Optimisation. An innovative MDO formulation has been developed based on the Collaborative Optimisation application for Diesel engine calibration. Form the MDO implementations, the fuel consumption have been significantly improved, while keep the emission at same level compare with the bench mark solution provided by sponsoring company. More importantly, this research has shown the ability of MDO methodologies that manage and organize the Diesel engine calibration optimisation problem more effectively. / Jaguar Land Rover

Le monde industriel, et en particulier l’industrie automobile, cherche à représenter au mieux le réel pour concevoir des outils et produits les plus adaptés aux enjeux et marchés actuels. Dans cette optique, le groupe Volvo a développé de puissants outils pour la simulation de la dynamique des véhicules industriels. Ces outils permettent notamment l’optimisation de composants véhicules ou de stratégies de contrôle. De nombreuses activités de recherche portent sur des technologies innovantes permettant de réduire la consommation des véhicules industriels et d’accroitre la sécurité de leurs usages dans différents environnements. En particulier, le développement des systèmes d’aide à la conduite automobile ITS et ADAS. Afin de pouvoir développer ces systèmes, un environnement de simulation permettant de prendre en compte les différents facteurs pouvant influencer la conduite d’un véhicule doit être mis en place. L’étude se concentre sur la simulation de l’environnement du véhicule et des interactions entre le véhicule et son environnement direct, i.e. le véhicule qui le précède. Les interactions entre le véhicule étudié et le véhicule qui le précède sont modélisées à l’aide de modèles mathématiques, nommés lois de poursuites. De nombreux modèles existent dans la littérature mais peu concernent le comportement des véhicules industriels. Une étude détaillée de ces modèles et des méthodes de calage est réalisée. L’environnement du véhicule peut être représenté par deux catégories de paramètres : statiques (intersections, nombre de voies…) et dynamiques (état du réseau). A partir d’une base de données de trajets usuels, ces paramètres sont calculés, puis utilisés pour générer de manière automatisée des scénarios de simulation réalistes. / The industrial world, and in particular the automotive industry, is seeking to best represent the real world in order to design tools and products that are best adapted to current challenges and markets, by reducing development times and prototyping costs. With this in mind, the Volvo Group has developed powerful tools to simulate the dynamics of industrial vehicles. These tools allow the optimization of vehicle components or control strategies. Many research activities focus on innovative technologies to reduce the consumption of industrial vehicles and increase the safety of their use in different environments. Particularly, the development of ITS and ADAS is booming. In order to be able to develop these systems, a simulation environment must be set up to take into account the various factors that can influence the driving of a vehicle. The work focuses on simulating the vehicle environment and the interactions between the vehicle and its direct environment, i.e. the vehicle in front of it. The interactions between the vehicle under study and the vehicle in front of it are modelled using mathematical models, called car-following models. Many models exist in the literature, but few of them deals specifically with heavy duty vehicles. A specific focus on these models and their calibration is realized. The vehicle environment can be represented by two categories of parameters: static (intersections, number of lanes) and dynamic parameters (state of the network). From a database of usuals roads, these parameters are computed, then, they are used to automatically generate realist traffic simulation scenarios.