Nikolai Kuznetsov
Full Doctor of Engineering Sciences
|
Date of birth |
30 May 1955 |
|
Place of birth |
Kiev, Ukraine |
|
|
Married, two children |
|
|
|
|
1971 -- 1976 |
Student of Kiev University |
|
1976 -- 1979 |
Post-graduate course in Kiev
University |
|
1979 |
Defence of the first dissertation
(candidate of Mathematics) |
|
1979 -- 1984 |
Junior researcher V.M.Glushkov
Institute of Cybernetics, Kiev |
|
1984 -- 1989 |
Senior researcher V.M.Glushkov
Institute of Cybernetics, Kiev |
|
1989 until now |
Leeding
researcher V.M.Glushkov Institute of Cybernetics, Kiev |
|
1987 |
Defence of the second dissertation
(Full Doctor of Engineering Sciences) |
-
Research grant of Alexander von
Humboldt Foundation Research centre KFA-Jülich, Germany (May
1990 – March 1992)
-
Research centre GRS-Garching, Germany
(November
1994 – February 1995)
-
Research work in GRS-Garching, Germany (November
1992 – March 1993)
-
Research grant of Royal Society, U.K.
STORM Research Centre, UNL (September, October
1995 - June, July 2001)
-
Research work in
ETH Zürich,
Switzerland (1996,
1997)
Field of scientific
interests:
-
Mathematical methods of reliability
theory and queuing theory,
-
Monte Carlo method, fast simulation
methods, methods of fault tree analysis
Nikolai Kuznetsov is author of 5 books and
more than 80 scientific articles
Main results
It is
well known that Monte Carlo simulation method is the most widely used
techniques to evaluate system characteristics, among them reliability.
Frequently it is the only practical method for evaluating the
reliability of systems being too large and complex to be analysed by
analytical (explicit or asymptotic) methods. Two types of Monte Carlo
simulation have been developed: direct and indirect simulation. The
principles of direct Monte Carlo methods (which are sometimes called
standard, crude or naïve) are well known. At the same time, modern
systems possess not only structural complexity, but also high
reliability. This is the case when standard Monte Carlo simulation is
impractical because of the excessive amount of computing time used.
That’s why so much attention has been given in the last few years to
special simulation techniques, known as
variance reduction
techniques, fast simulation methods or analytical-statistical methods.
N.Kuznetsov
has developed several methods permitting to create fast simulation
algorithms for the evaluation of highly reliable systems with complex
interconnections between components and without any assumptions about
exponentiality of the distribution functions. Such algorithms make it
possible to decrease the variance of estimate (and hence the computer
time needed) in two orders of magnitude. Algorithms have already been
developed for fast evaluation of reliability indexes of systems with
different dependencies between components, with two and more phases of
system operation, with different types of components (revealed failures,
periodically detected failures, failures per demand of components), and
with variable loading on the system, etc. Fast simulation has been
successfully applied to real engineering systems containing about 200
components.
The fault
tree is amongst the most useful models for the description of an
accident development and is one of the main models used in system
reliability analysis. The recurrent nature of the fault tree
construction enables to carry out reliability analysis of large and
complex systems by means of analytical or statistical methods based on
computer algorithms. One more important feature of fault trees is the
possibility of description and analysis of systems with noncoherent
structures and with common cause failures. The fault tree can be used as
the aid in determining of the possible and the most probable causes of
an accident (“weakest links” of the system). Moreover it is a very
helpful diagnostic tool for the qualitative system evaluation what is
necessary for grounded decisions permitting to avoid the additional
expenditure on the elaboration of the system design. In the most complex
case of systems with common cause failures (the great number of
replicated gates) the usual fault tree techniques based on bottom-up
approach with standard modularization techniques, and with truncation of
the low-probability cut sets, can lead to inefficient use of computer
time, and also a loose upper bound for the probability of system failure
due to truncated cut sets.
For the
analysis of large and complex fault trees N.Kuznetsov has proposed a new
method based on the multi-level representation of fault trees with a
great number of replicated gates. This method makes it possible to find
and simplify cut sets at the level of gates as the process descends from
higher to lower levels in the fault tree representation. Further
analytical reliability evaluation is offered on the basis of module cut
sets. This newly implemented approach reduces both the computer time and
memory required for the cut set evaluation, and produces a tighter upper
bound for the truncation error. The method of multi-level representation
has been used to investigate fault trees containing up to 4000 gates.
Main resent publications
-
Kovalenko I.N., Kuznetsov
N.Yu., Shurenkov V.M. (1996) Models of Random Processes: a Handbook
for Mathematicians and Engineers, New York: CRC Press, 446 p.
-
Kovalenko I.N., Kuznetsov
N.Yu., Pegg P.A. (1997) The Mathematical Theory of Reliability of
Time Dependent Systems, with Practical Applications, Chichester:
Wiley, 303 p.
-
Kuznetsov N.Yu. (1994).
Fault trees – problems and the modern state of investigations,
Cybernetics and Systems Analysis, 30, No 4, pp. 419-439.
-
Hennings W., Kuznetsov
N.Yu. FAMOCUTN and CUTQN – computer codes for fast analytical
evaluation of large fault trees with replicated and negated gates,
IEEE Transactions on Reliability, 1995, 44, No 3, pp. 368-376.
-
Kuznetsov N.Yu. (1998).
Estimating the failure probability of a Markovian system during
regeneration period by the importance sampling method, Cybernetics
and Systems Analysis, 34, No 2, pp. 216-222.
-
Kovalenko
I.N., Kuznetsov N.Yu. (1999). Analysis of the deviation of the
nonstationary unavailability of a repairablele system from its
stationary value, Cybernetics and Systems Analysis, 35, No 2, pp.
240-252.
-
Kuznetsov N.Yu. (1999).
Fast simulation of the failure probability of a system on the busy
period for nonexponential distributions defining the process of
failure and repair of components, Cybernetics and Systems Analysis,
35, No 3, pp. 413-423.
-
Kuznetsov N.Yu. (1999).
Fast Simulation of failure probability of a system consisting of
components with considerably differing reliability, Cybernetics and
Systems Analysis, 35, No 6, pp. 884-891.
-
Kuznetsov N.Yu. (2000).
Finding the probability of uninterrupted operation of a main
pipeline system by an analytical-statistical method (a serial
model), Cybernetics and Systems Analysis, 36, No 4, pp. 587-596
-
Kuznetsov
N.Yu. (2002). Fast simulation of steady-state availability of non-Markovian
systems, Cybernetics and Systems Analysis, 38 No 1 pp. 89-98.
E-mail:
Official
address:
40, Prospect Glushkova,
V.M.Glushkov Institute of
Cybernetics,
03680, Kiev 187, Ukraine
Telephone: 380 - 44 - 5266381
|
