The article analyses some properties of Kijima incomplete recovery models using a Weibull distribution for time to first failure. The maximum likelihood method is used for assessment of distribution parameters and recovery coefficient. Confidence limits have been identified using a Fisher information matrix. The authors consider cases of processing data from several identical elements and prove the inverse relationship between the deviation value and the number of elements. The paper examines two ways of assessing the leading function of the deteriorating component flow. A comparison is made between the new approach that represents the leading function of the flow as the ultimate sum and the approach that uses the statistical testing method. The paper suggests the method of calculation of the average direct time and reverse residual time based on the statistical testing method. Several demonstration examples are given.

The article deals with the philosophical aspect of reliability exemplified by unique mission-critical systems for which it is impossible to apply probabilistic and statistical methods of reliability theory.

The article describes a method helping to define the scope of recovery for failure-free operation, frequency of preventive replacements and repairs of technical devices under the given failure rate level.

The paper describes the methodology for calculating the cost of warranty obligations of products manufactured by a supplier that allows estimating a predicted repair cost over the period of warranty obligations and including it into the structure of a product’s price on the basis of the duration of the warranty obligations period. The cost calculation is performed for various types of warranty obligations with consideration of different stages of operation of a product. Generally, a product consists of constituent parts manufactured by a producer, and parts purchased from other manufacturers. Warranty periods for self-produced constituents and for purchased parts can either coincide or differ.

The mathematical model of electrical insulation failure is proposed. The model is based on the superposition of uniform and exponential probability distributions. Relation for time dependence of the insulation failure rate is obtained. Application of this relation provides correct analytical description of insulation reliability on a long time interval.

The paper describes the method of estimating the reliability of technical security systems, which allows to forecast their failure-free operation with high probability based on the limited operational information.

The problems related to information solutions deployment in the education system are connected with the development of complex software that solves both information and management tasks. The development of such software involves significant material costs. Beside the software development costs, the expenses related to elimination of the consequences of software failures must be taken into consideration as well. All those costs should be evaluated at the stage of the development of performance specifications for software. This article deals with the evaluation and minimization of the above costs.

The paper considers computational and experimental methods for defining the probability of functional failures for the electronic component base under electromagnetic pulse interference. The analysis of the properties of the pulse interference random process and the physical mechanisms of interference influence on electronic node elements is carried out. The ways of failure probability calculation based on the concept of the probabilistic interference nature are elaborated. It is shown that the analysis of interference influence enables to prove the choice of pulse equivalence conditions, which is vital for electromagnetic compatibility.

The paper deals with the issue of selecting parameters that secure the construction of a risk matrix. Main aspects affecting the error of presentation of risk matrix results are identified. The construction of a risk matrix with minimum error for given conditions is shown by means of a particular example. Directions for further research are outlined.