Shchapov P. Method of improved control and diagnostics confidence of stochastic parameters for objects of different physical nature

Українська версія

Thesis for the degree of Doctor of Science (DSc)

State registration number

0509U000598

Applicant for

Specialization

  • 05.11.13 - Прилади і методи контролю та визначення складу речовин

08-10-2009

Specialized Academic Board

Д 64.050.09

National Technical University "Kharkiv Polytechnic Institute"

Essay

Object of study is the process of formation of the uncertainty in obtaining information on the controlled parameters, which reduces the reliability of monitoring and diagnostics. Purpose is support of the theoretical foundations of development and implementation of methods for improving the reliability of multiparameter monitoring and functional diagnostics. Methods are probability theory, mathematical statistics, stochastic processes, information theory of measurement. The results are increased accuracy of control of bulk materials, diagnostics of technological equipment, quality control of liquid insulation. Novelty - for static objects, such as biochemically complicated loose materials, a method was developed for taking statistically uniform nonstandard samples of control parameter; a selection method was developed of informative, as much as possible, subset of regulated values with discrete heterogeneity used in polynomial transformation models at indeterminate, due to short learning samples, model coefficients; a theory and method were developed for confidence maximum optimal multialternative recognition of control parameter levels using a subset of stochastic regulated values. Such an algorithmic transformation was proved to be able to supplement polynomial transformations, thus improving the confidence of any multiparametric control; some methods were developed to improve the confidence of control and functional diagnostics of dynamic objects on the basis of stochastic signals with complicated types of nonstationarity; a method was developed to form a maximum informative system of regulated values with mixed heterogeneity as represented by multidimensional nonstationary timing series. Results are implemented in the state budgetary research: some practical problems were studied in improvement of confidence for spot control of composition parameters of loose agricultural products; medical functional diagnostics and dynamic control of transitional processing conditions; operational control of high-voltage electric equipment filled with oil; in the educational process of NTU "KPI". Scope is enterprises of machine building, power generation and medical organizations.

Files

Similar theses