Browsing by Issue Date, starting with "2019-01-30"
Now showing 1 - 1 of 1
Results Per Page
Sort Options
- Automatic engine and propeller selection for mission and performance optimizationPublication . Ribau, Nídia Salvador; Gamboa, Pedro VieiraIn this dissertation, an already existing optimization software is employed to minimize the total energy consumed at a certain given mission. Initially, the optimization algorithm only returns the optimized design variables for the propeller specifications, and no mission performance constraints can be defined. Hence, on this thesis, two main objectives are stipulated. One is to enable the optimization of certain engine/motor design parameters to match the propeller. Thus it is required to create two data bases, one with the IC engine specifications and another with the electric motor specifications, in order to develop empirical models as functions of certain design variables. These engine design variables are then inputted into the optimization algorithm, to be optimized alongside the propeller parameters. The second objective established is to add certain mission performance constraints, to enable the user to constrain certain parameters inside the algorithm iterations. Both data bases were successfully created, and all empirical models obtained, although with a certain error associated with the coefficients of the functions. The design variables selected to be introduced in the algorithm, which are the inputs of the empirical models, were rated power and rated engine speed for the IC engine, and maximum allowed current and the motor speed constant for the electric motor. The mission constrains are also calculated and inputted inside the algorithm, optimizing according to the feasible space defined by the user. The updated software now returns, for a given set of mission constraints, the engine solution which matches the optimized propeller parameters, and selects a real engine from the database created. The results obtained confirmed the practicality of the engine empirical models, given good matches, although not perfect, to the optimum solution reached. The design variables and the objective function are converging correctly to a stabilized solution, according to the feasible space the user may choose to define.