Page 22 - Sayı-08

20 M.O. ŞERİFOĞLU, T. HIZARCI and H. AKYILDIZ

characterization of denied areas, passive detection of acoustic sources, sustained monitoring and
surveillance of marine regions, and multistatic acoustic detection (Cmre.nato.int, 2014).
3. Uncertainty
Uncertainty analysis investigates the effect of lack of knowledge or potential errors of the model.
Uncertainty is incomplete knowledge and information about a system as well as inaccuracy of the
behavior of systems. Uncertainty can be categorized into three groups as physical uncertainty,
statistical uncertainty, and modelling uncertainty (Thof & Murotsu, 1986). Physical uncertainty
caused by physical quantities, such as loads, material properties and environmental changes. This
uncertainty also be called aleatory and random uncertainty. This type of uncertainty cannot be
reduced or eliminated by means of collection of additional information because of that there is
always be unpredictability in the variables. However, uncertainty can be quantified by
examination of the data. The statistical uncertainty arises due to a lack of information. Distribution
parameters can be considered as random variable according to given data set. This uncertainty
also be called epistemic and systematical uncertainty. Caused by limited information or lack of
knowledge on a quantity. The model uncertainty caused by assumptions and unidentified
boundary conditions and their interaction with the model. To design and develop a model, a lot
of assumptions and hypotheses have to be defined. Even if these assumptions are chosen correctly,
model need to match with the real world conditions. (Liu, 1996).
An uncertainty analysis uses the occurrence levels to determine the possible outputs and
possibilities of the outputs. The probabilities of observing particular range of values of a random
variable are described or defined by a probability distribution.
Uncertainty analyses involve identifying characteristics of various probability distributions of
model input and output variables, and subsequently functions of those random output variables
that are performance indicators or measures.
Uncertainty analyses can be used for:

 Determination of probability and outputs range and tresh holds.
 Determination of standard deviation of the system and the effects of inputs to the outputs.
 Determination of the total relaibility of the system and estimating the possible outcomes.

3.1 Sensitivity
Sensitivity is another important parameter for the system reliability with the uncertainty.
Sensitivity analysis is a method to determine which variables, parameters or other inputs have the
most influence on the model output. This involves a study of the effect each of the different
parameters has on results of reliability analysis of the overall system. If the overall effects of
changing a variable are found to be small, then the variable can be treated deterministically.
However, where changes in a variable are found to affect the overall reliability significantly, then
it is important to model the variable by using the best available distribution.

GiDB|DERGi Sayı 8, 2017