SAWADKOSIN, P. Optimalizace tvaru strojních součástí s vlivem variabililty vstupních údajů [online]. Brno: Vysoké učení technické v Brně. Fakulta strojního inženýrství. 2019.

Posudek vedoucího

Novotný, Pavel

The main aim of the thesis is to evaluate the existing state of the solution of the given problem, to propose a suitable optimization algorithm and apply this algorithm to a real technical problem like thrust bearing of turbocharger. The work responds to current issues connected with maximizing the mechanical efficiency of turbochargers using computational methods. The background search lacks a more detailed description of the generally available methods and procedures to address the problem. From the very beginning, the author assumes that genetic algorithms will be used. However, there are undoubtedly other methods of optimization, that can be also used. The used approach to solving the influence of variability of input parameters is chosen logically, however, in the case of practical tasks, this approach is often not feasible. Large FEM or CFD models with demanding nonlinear numerical solutions can be taken as examples of such real world problems. The author's own contribution is lower, mostly the use of commercial software, especially the Matlab program. The ability to understand and interpret the results from a technical point of view is very weak, but this is probably due to the author's study focus. The practical benefits of the work are significant, enabling the use of computationally fast procedures in the early stages of designing the given machine component. Subsequently, more detailed models can be used, but on a limited area defined by input parameters. The formal processing of the work is without significant deficiencies, the image processing and equations do not show any major problems and in general it can be regarded as above average quality. After the overall evaluation, it can be stated that the student fulfilled the requirements for the thesis.

Posudek oponenta

Jonák, Martin

The master’s thesis focused on optimization of machine parts is formally complete and logically structured. I also appreciate the fact that the work contains a preface. From the factual point of view, the author's task was to develop algorithms suitable for optimization of selected properties of the turbocharger thrust bearing. At the beginning of the work, the review and the theory necessary for the optimization of the thrust bearing shape are carried out. The chosen optimization method is a genetic algorithm. Implementation is done in Matlab using native functions and Global Optimization Toolbox. Finally, the optimized geometry of the thrust bearing surface has been obtained for a series of 13 operating states in which the turbocharger can operate. Besides occasional typos (e.g. missing dx in equation 2.1, or missing parameter h0 on page 42 below), I reproach the author for a slight diversion from the topic, where the task is to develop optimization algorithms rather than using the already created one. Furthermore, more detailed description of the way, in which the weighting coefficients given in Tab.7.5 have been obtained, is missing. Overall, the work is at a good level and I recommend it to the defense.

eVSKP id 117400