The first section deals with scientific and applied issues related to the development
and operation of ship automated control system (ACS), current areas in the field of
modernization and adaptation of ACS ship steam boilers. Based on the analysis of existing
problems associated with the development and operation of ACS ship steam generating
units (SGU), the purpose and objectives of the dissertation research. The purpose and tasks
of research are put, the scientific hypothesis about urgency of use of intellectual methods
and algorithms in processes of development, modernization and adaptation of ACS of
water-tube SGU of wide constructive purpose is put forward.
In the second section the parametric identification of SGU subsystems is carried out.
The coefficients of the equations of the dynamics of the SGU parameters during its
operation at three thermal regimes are calculated and their degree of adequacy is
determined. New differential equations of local SGU systems are obtained, which allow to
estimate the influence of ACS control actions on the processes of chemical and
electrochemical corrosion in SGU elements. By means of full-scale experiments on
operating auxiliary SGU Mitsubishi MВ, analogs of SGU – production steam boilers GM
and DKVR and ship emulators, mathematical models were determined by control and
perturbation channels for main, auxiliary, utilization and combined SGU, and the accuracy
of the obtained models was evaluated.
In the third section the modeling of nonlinear SGU systems is performed. Based on
the simulation results performed in the System Identification Toolbox, a wide class of
nonlinear Wiener-Hammerstein models that implement a neural network, nonlinear
ARMAX models and other models that allow to describe the experimental processes of
controlled SGU parameters with a high degree of adequacy. The expediency of
introducing a nonlinear model of regular wave oscillations in the form of the van der Pol
differential equation (oscillator with nonlinear attenuation) as an additional component of
the external perturbation channel for effective analysis of the stability of ship ACS SGU by Rauss – Hurwitz method, which reduces the occurrence of emergencies. The combined
expert criterion of an estimation of efficiency of process of adjustment of ACS of
parameters of SGU is offered and tested and research of methods of adaptive management
of units of SGU is carried out.
The fourth section identifies methods for developing fuzzy output systems, which
allow to obtain optimal quality indicators of transients of control systems of SGU
parameters and to provide support of the set thermal regime of SPU. An expert system of
adaptation of setting parameters of a standard regulator has been developed, which
operates on the basis of identification of indicators of operational processes of SGU. The
efficiency of the proposed methods and expert systems that implement the theory of fuzzy
logic is shown. It is noted that the use of expert systems ensures the adaptation of the SGU
to the operating conditions and the maintenance of the specified thermal regime of the
SGU without significant deviations. Also, maintaining a given vapor pressure with the
proposed system helps to reduce emergencies by 25 % and slow down the growth of
chemical corrosion in the elements of the SGU, ie in general increase the reliability of the
SPU. The developed methods form the basis of a new concept of creating highly efficient
ACS.
In the fifth section, the development of neural network systems for identification and
control of SGU. A universal adaptive neural network control system for the combustion of
liquid fuel in SGU furnaces is proposed, which allows minimizing the content of harmful
emissions into the atmosphere. Methods of development of neural network systems of
monitoring and optimization of ecological indicators of SGU work are tested and their
efficiency in comparison with traditional ACS is shown. The method of development,
adjustment and operation of the neural network ACS of flue gas recirculation to the SPU
furnace, which reduces the NOx content to 54 %, is presented and tested. Developed neural
network ACS optimization of fuel consumption by controlling air intakes in the furnace
and flue SGU, correction of excess air, based on the steam load of the units SGU and
turbulizer can increase the gross efficiency of SGU up to 8 % depending on heat load and
heating surfaces.
Key words: ship steam generating installation, control system, mathematical model,
PI-controller, thermal regime, fuzzy control system, neural network regulator.
