OCHULO, I. Optimalizace tvaru mazací mezery hydrodynamického ložiska [online]. Brno: Vysoké učení technické v Brně. Fakulta strojního inženýrství. 2021.

Posudek vedoucího

Novotný, Pavel

The master thesis includes 50 pages excluding the introduction and list of references, figures and Matlab codes used. Ochelo Ikechi has revised his previous work. Unfortunately, there was again a problem with the timing of the activities, with substantial parts of the thesis was created and been written during the last week before submission. The thesis does not contain new knowledge, only a combination of many different materials, often unrelated. Most of the text in the thesis is taken from the literature, mainly from the book of Stachowiak et. al, 2014. Analytical methods, one-dimensional solutions and two-dimensional solutions are combined. The two-dimensional solution is not necessary and was not assumed. There is insufficient review of the possibilities of mathematical optimization of the problem. The student has stuck to a single solution using genetic algorithms and has not considered other options. The key analytical relations for calculating the bearing load capacity, friction force and mass flow (eqs. 4.20–4.23) are incorrect. Constants for converting values from dimensionless quantities to physical quantities are missing. The results of the calculations are suspect, showing signs of incomplete convergence. The optimized bearing (result) is not better than the initial state. Technically, the student probably did not understand the point of the whole problem, that is, to find the optimum bearing shape for a predefined load. The documented Matlab scripts contain errors of principle (basic mathematics), e.g. in the stopping of convergence or numerical integration of quantities distributed over a surface. The thesis is constructed in an unsystematic way, the objectives of the thesis are not met and the thesis as a whole does not meet the requirements for a thesis at the master's level.

Posudek oponenta

Vacula, Jiří

The diploma thesis revision of Mr. Ikechi Ochulo includes 50 pages of his work (from introduction to conclusion). The work deals with the optimization of the general shape of the thrust bearing geometry by numerical approach. The revision of the diploma thesis still includes many flaws and mistypes making the uninitiated reader completely uncapable of understanding the steps. Mainly, meeting the goals of the thesis was not achieved: 1) „Program for optimizing the bearing working surface of general shape.” The supplemented Matlab .m files include fundamental errors, e.g.: In “GenProfReyEqn.m” - Line 104: Convergence criterion is not based on the target function residual error. Convergence criterion is only based on number of iterations. Convergence of the individual geometric designs in not ensured. The optimization procedure is likely based on non-converged individuals. - Line 130: Load capacity is computed as multiplication of pressure [Pa] and the dimensionless term “delxbar” [-]. However, physically the load capacity is defined as force [N]. Units are not consistent. - „Cycle for“ between lines 135 and 141 is searching for maximum unit load capacity on bearing film instead of performing integration of the unit load capacities in the oil film. Load capacity is thus computed incorrectly. 2) „Application of the optimization procedure in the design of the working surface of the thrust bearing of a turbocharger.“ The optimal variant is summarized in Table 8. As stated above, the computation of the load capacity is wrong, the data in the table is not correct. In addition, friction coefficient F/W of the optimized profile is higher than of the original variant making the results unconvincing, moreover supporting the idea of non-converged results during the optimization process. 3) The goal „Review of suitable optimization methods for solving the problem” was only very partially achieved. The review shown on pages 17–19 is poor. Optimization methods are not profoundly summarized to obtain a review. The section 2 is basically focused on genetic algorithms mainly, not providing the reader any overview of the methods as such. One may find many mistakes in the work, e.g.: - in equation 4.19 the constant B is different that constant A. However, in the code “GenProfReyEqn.m” in line 95 the constant B is replaced by constant A. - equation 4.20 does not produce force in Newtons [N] but force per unit length [N/m]. - in equation 4.21 there is even a sum of terms with different units: [N/m] + [Pa]. - equation 4.22 uses notation typically used for mass flow rate but right hand side is volume flow rate. - there other flaws such as figure 14 and figure 15 are aligned to the right side of the paper, no axis titles in figure 20 as well as no description of the figure 20, „List of Figures” (page no. 64) and „List of Tables” (page no. 65) are not presented in “Contents”, etc. It seems that the problem to be solved was not understood in principle. The thesis is classified by mark F especially because of not achieving the goals and fundamental flaws in the presented equations and algorithms making the work at this state inapplicable for any bearing design.

eVSKP id 137282