ІНТЕРВАЛЬНІ СТАТИСТИКИ ВИЗНАЧЕННЯ ДІАГНОСТИЧНИХ ПАРАМЕТРІВ ЕНЕРГЕТИЧНИХ УСТАНОВОК ПРИ НЕГАУСІВСЬКОМУ РОЗПОДІЛІ ВИХІДНИХ ДАНИХ

<jats:p>An approach is proposed for determining interval estimates of diagnostic parameters of the technical condition of power plants, in particular those built on the basis of gas turbine engines, during bench tests and long-term operation under conditions of a non-Gaussian distribution of output data. The relevance of solving the specified scientific and applied problem lies in the need to determine the compliance of bench test results with the requirements of the technical conditions and the current analysis of the technical condition of the specified objects, to establish the possibility of their operation within the established resource. An analysis and comparison of methods for interval estimation of diagnostic parameter values for the technical condition of power plants, expressed as time series, were performed, accounting for departures of the output data from normality. Time series of values of diagnostic parameters of the technical condition of power plants are realizations of random processes, the stochastic component of which is caused by random disturbances and the nature of energy conversion processes. It has been established that deviations in the distribution of the initial data from normality can significantly shift the intervals for estimating the first moments of the distribution of the values of the measuring diagnostic parameters of power plants. Accordingly, such a shift in interval estimates requires correcting bench test results and assessing the technical condition of power plants during the long-term operation of the specified objects. An improved method of analytical solution of the set scientific and applied problem is proposed, which allows obtaining interval estimates of diagnostic parameters of power plants under conditions of non-Gaussian distribution of the initial data. </jats:p>