The diploma thesis deals with the creation of loans provisions in banks. The need for adjustments is based on the existence of credit risk and this thesis presents techniques of credit risk measurement and applied credit risk indicators. Before calculating provisions, it is necessary to classify receivables, whose parameters are also discussed in the thesis. The practical part focuses on methods of calculating provisions. The method of coefficients, discounted cash flow method and statistical methods are compared. Markov model and gross roll rate model were selected from statistical methods. The findings were evaluated and recommendations were made based on the findings.

This work addressed aerodynamic modeling methods for prediction of post-stall flight dynamics of large transport aircraft. This was accomplished by applying historically successful modeling methods used on high-performance military aircraft to a transport configuration. The overall research approach involved integrating forced oscillation and rotary balance wind tunnel data into an aerodynamic model using several methods of blending these data. The complete aerodynamic model was integrated into a six degree-of-freedom simulation. Experimental data from free-spin wind tunnel testing was used to validate the aerodynamic modeling methods by comparing aerodynamic force and moment coefficients and also to validate the simulation performance by comparing spin mode characteristics and time histories. The aerodynamic model prediction of spin dynamics was generally very good using all of the blending methods studied. In addition, key spin mode characteristics were predicted with a high degree of accuracy. Overall, using the Hybrid Kalviste method of blending forced oscillation and rotary balance data produced the closest match to the free-spin data when comparing aerodynamic coefficients and spin mode characteristics. Several issues were encountered with the blending methods that were exacerbated by nonlinearities and asymmetries in the dynamic aerodynamic data. A new method of looking up dynamic aerodynamic data was proposed to address shortcomings in the blending methods and recommendations were provided on addressing issues with the dynamic aerodynamic data.

Light vehicle rollover accidents on soft surfaces can be modeled assuming constant drag with linear motion equations and other engineering principles. The concept of using segment average results to evaluate roll mechanics parameters throughout a roll sequence, and specifically, segment duration to evaluate vehicle trajectory between ground impacts is developed. The trajectory model is presented, explained and compared to values obtained by analyzing digital video of rollover crash tests. Detailed film analysis procedures are developed to obtain data from rollover crash tests that are not otherwise documented. Elevation of the center of gravity of vehicles is obtained where instrumentation does not explicitly yield this data. Instantaneous center of gravity elevation data throughout a roll sequence provides the opportunity to calculate descend distances as a vehicle travels from one ground contact to another. This data is used to quantify severity of ground impacts as a vehicle interact with the ground throughout a roll sequence. Segment average analysis is a reasonable method for determining general roll mechanics parameters. Because of the chaotic nature of rollover accidents, the range of effective drag factors for a given roll surface may be quite large. Choosing an average of typical drag factors is a reasonable approach for a first-order approximation although certain parameters may be predicted less accurately than if actual values were known. The trajectory results demonstrate the influence of drag factor descent height calculations. Typical constant drag factors tend to overestimate descent height early in a roll sequence and underestimate descent height later in the sequence. The trajectory model is a useful tool to aid in understanding rollover mechanics although a rolling vehicle may be in contact with the ground for a significant fraction of a roll segment. The model should not be used at locations in roll sequences where there are extremes in translational center of gravity decelerations. These extremes include the segments immediately following overturn where there are large angular accelerations and large differences between the tangential velocity of the vehicle perimeter and the translational velocity of the center of gravity, as well as segments that include vehicle impacts with irregular topography.