This study investigates a novel methodology for the preliminary design of aeroengines. This involves the modelling of the disciplines that affect the engine's requirements and constraints, their implementation in software format and their coupling into a single unit. Subsequently, this unit is interfaced with an optimiser software. The resulting multidisciplinary optimisation (MDO) tool allows the automation of the traditional, human-based preliminary design process. The investigation of the above-mentioned novel methodology is carried out through the development of a "pilot" MDO tool and its subsequent utilisation in three case studies, characterised by different optimisation scenarios. The selection of each case study is motivated by current research questions, such as aviation's contribution to climate change or the attractiveness of specific novel propulsion concepts. The outcome of the pilot MDO study is considered successful and has been well received by several academic and industrial aero-engine organisations. The choice of the disciplines and of their modelling fidelity allowed a realistic representation of the main disciplinary interactions and tradeoffs that characterise the important phase of preliminary design. The computational effort involved in the solution of the optimisation studies was found to be acceptable, and no major reprogramming was required when different optimisation scenarios were considered. The case studies were investigated with an ease and comprehensiveness that would not have been achievable through a human-based parametric analysis. The positive experience with the pilot MDO tool suggests that an automated methodology for the preliminary design of aero-engines is feasible, applicable and valuable. Its adoption can provide substantial advantages over the traditional human-based approach, such as a reduction in human effort, costs and risk. From this perspective, the pilot study constitutes a first step towards the development of a full-scale MDO tooL usable by aero-engine manufacturers. In the near future, issues like climate change could drive significant modifications in airframe and engine design. A preliminary design MDO tool is therefore timely, and has the potential of making a significant contribution.

Cette thèse s’inscrit dans le cadre d’une problématique globale de recherche qui porte sur la modélisation et l’optimisation des chaînes logistiques. Il s’agit de développer, pour deux maillons successifs d’une chaîne logistique, des modèles pour résoudre le problème du lot économique commun en intégrant les aspects liés à la production, à la qualités des produits et à la maintenance.Le choix d’une politique de production continue, jusque là adoptée dans la littérature, est à l’origine d’accumulation de stock chez le fournisseur ou chez le client selon que les livraisons sont respectivement retardées ou immédiatement transférées.Dans notre travail nous proposons une nouvelle alternative de production basée sur le principe «ne produire que le besoin du client à livrer immédiatement». Globalement, l’idée se résume à produire la commande par lots séparés de taille optimale immédiatement transférés au client, cherchant à réduire le stock moyen dans le système. Dans un mauvais scénario, correspondant à une éventuelle amplification du nombre de setup, le gain sur le stockage couvrirait la perte au niveau du setup lorsque le coût unitaire de celui-ci est relativement réduit.Cherchant un meilleur compromis entre les coûts de stockage, de setup et de livraison, cette solution est généralisée par la suite au cas où la taille du lot produit est différente des tailles de la commande et des livraisons.L’étude est étendue au cas d’une production imparfaite caractérisée par un taux moyen fixe de non conformités. Les stratégies continue et lot-par-lot sont reformulées en intégrant des actions de maintenance dont le but est de réduire le coût de non-qualité et d’optimiser le coût total / In this work we deal with the single vendor single buyer integrated production inventory problem. In the beginning, two production strategies are compared. The first one, proposed by Ben-Daya and Hariga (2004), suggests that the buyer orders batches of size nQ every time his on hand inventory reaches the reorder point r after the reception of the ultimate lot of the last order. The vendor from his side produces continuously nQ and makes equal shipments of size Q. The second policy, which we develop, propose that to satisfy the same ordered quantity, the vendor produces separately smaller batches of size Q (lot-for-lot), n times. The total expected cost is adopted as the decision variable for the choice of the best strategy.Afterwards, the lot-for-lot strategy is generalised to the cases where the sizes of the produced lot and the shipment are different from the ordered quantity. Henceforth, the vendor manufactures separately smaller batches of size iQ and makes equal shipments of size jQ (1≤ j ≤ i ≤ n) limited or not to the transportation capacity. The optimal solution corresponds to the best combination (n, i, j, Q, r) yielding the minimal total average cost per time unit incurred by the vendor and the buyer.In the last contribution, we consider the case of an imperfect process where the production unit is assumed to randomly shift from an in-control to an out-of-control state characterized by a fixed production rate of non-conforming items. To resolve the problem, we propose and compare two different strategies integrating production, inventory and maintenance policies. The maintenance actions are performed to prevent or to reduce the losses related to imperfect items

No description available.

No description available.

No description available.

The effects of air and exhaust gas dilution on the CAI combustion of a range of fuels including three gasoline compositions, four primary reference fuels, and two alcohols are experimentally investigated using a single cylinder research engine. Two of the three gasolines tested are manufactured from standard gasoline during engine operation by a novel fuel system, designed to improve the performance of both controlled autoignition and spark ignition engines. A series of experimental tests are performed to establish the satisfactory combined air and exhaust gas dilution regions for each fuel. Detailed in-cylinder pressure and exhaust gas speciation measurements are taken, and the fuels are compared and contrasted for their performance in terms of power output, fuel consumption, and harmful exhaust emissions. Results show that alcohol fuels are superior to hydrocarbon fuels for controlled autoignition combustion because their autoignition characteristics are less affected by the presence of exhaust gas species. Furthermore, the timing of autoignition is shown to be of minor importance for achieving efficient and stable controlled autoignition combustion, contrary to widely held beliefs. In addition, the novel fuel system is developed and commissioned for use on a single cylinder research engine operating with a spark ignition system. The two gasoline fuels produced by the system are evaluated individually for their knocking combustion characteristics over a range of compression ratios and spark advances. Results from these tests indicate that the fuel system used in conjunction with a specially modified production engine may allow the normal compression ratio of that engine to be increased by up to 1.0, increasing its efficiency.

No description available.

No description available.

No description available.

No description available.