689

1 INTRODUCTION

The efficiency of the ship's power plant depends on

the rational use of energy for the movement of the

vessel and the provision of internal needs. The main

source of energy at present is natural fuel, the price of

which is significant. The composition and structure of

the ship's power plant are optimized by the

developers for the most efficient and complete use of

the expended fuel resources [1, 2].

The use of various types of fuel, the disintegration

of the power of the ship's power plant, effective

control of energy flows has led to the emergence of

multi-generator power plants with various types of

drive engines (diesel generators, shaft generators,

turbo generators) and alternators (synchronous,

asynchronous, with permanent magnets), [3].

Obviously, it is possible to combine all types of

energy used on a ship by converting it into electricity,

which has led to the widespread use of ships with

electric propulsion [4].

Different characteristics of drive motors and many

power plants operating in parallel in the traditional

use of synchronous generators on ships creates

problems with the stability of multi-generator power

plants, [5]. The rigid geometric connection of the

magnetic flux of a synchronous generator with its

excitation winding increases the oscillatory properties

of the system of power plants operating in parallel.

An alternative solution to the problem of

increasing the stability of multi-generator ship power

plants can be the wider use of asynchronous

generators (AG) with a squirrel-cage rotor.

The advantages of asynchronous generators are

widely known [6, 7]. The AG has smaller dimensions

and weight, the design of the squirrel-cage generator

rotor is simpler, there are no rotating windings,

sliding contacts and semiconductor elements, there is

no current insulation on the rotor, which increases the

limiting heating temperature and ensures high

limiting rotor speeds. The high efficiency of the

generator due to the low value of the active resistance

of the rotor ensures its higher economical

characteristics. The AG has a sinusoidal shape of the

curve of the generated voltage, as well as the

symmetry of the three-phase voltage with uneven

load distribution.

Discrete Laws of Capacitor Control of Asynchronous

Generator Voltage

L. Vyshnevskyi, M. Mukha, D. Vyshnevskyi & A. Drankova

National University "Odessa Maritime

Academy", Odessa, Ukraine

ABSTRACT: In this article, the authors analysed the voltage stabilization system of an electrical installation with

an asynchronous generator and capacitor excitation. The properties and tuning parameters of the discrete-pulse

switching laws of three-phase sections of excitation capacitors of an asynchronous generator are considered.

The analysis is carried out and recommendations for the use of the described digital controllers are given.

http://www.transnav.eu

the

International Journal

on Marine Navigation

and Safety of Sea Transportation

Volume 18

Number 3

September 2024

DOI: 10.12716/1001.18.03.2

690

The small time constant of the generator leakage

circuits, the rapid decay of inrush currents and short

circuit currents ensures the safety of short circuits for

the generator. Regulation of the AG excitation

through the stator circuit makes it possible to create

high-speed and invariant voltage stabilization

systems. Simplicity and safety of switching to parallel

operation, absence of rotor oscillations with

significant load changes ensure the stability of parallel

operation in multi-generator power plants.

Such significant advantages explain the interest in

the developments of asynchronous generator sets.

Scientific and technical problems that hinder the

widespread use of capacitor-excited asynchronous

short-circuit generators in ship electrical installations

can be grouped into the following areas.

1. Excitation of the AG with additional reactive

power.

2. Choice of optimal design parameters of an

asynchronous machine operating in a generator

mode.

3. Creation of a controlled source of reactive power

with good technical and economic indicators.

4. Efficient control of a ship AG electrical installation

modes.

2 ANALYSIS OF DISCRETE CONTROL LAWS FOR

THREE-PHASE SECTIONS OF AG EXCITATION

CAPACITORS

This article discusses technical solutions for the third

problem: an analysis of several discrete control laws

for three-phase sections of AG excitation capacitors is

carried out. The authors of the article consider the

further development of the previously described

controller [8, 9], which implements the integral

discrete-pulse law of voltage stabilization of the AG

with capacitor excitation.

The diagram of a capacitor control device with N

three-phase sections of capacitors C

0,C1-CN in a ship

electrical installation is shown in Fig. 1. Stabilization

of the AG voltage when the AG load or the drive

engine (DE) speed changes is performed by

connecting capacitor sections in appropriate

combinations. The switching of capacitors is carried

out by thyristor switches depending on the deviation

of the generator voltage U

g from the set value U0:

∆U=U

0-Ug.

Figure 1. AG electrical installation with a discrete capacitor

voltage stabilization system: DE – drive engine, AG -

asynchronous generator

The AG excitation current is generated by the

connected capacitor sections. The initial excitation of

the generator is provided from a permanently

connected block of capacitors when the generator is

rotated by the drive motor. The capacitance value of

the permanently connected block of capacitors

provides the specified voltage of the AG at idle at the

rated rotation speed of the drive motor.

In this circuit, the sampling value of the control

action corresponds to the minimum capacitance of the

capacitors ∆C=C

1, which is the level quantization

interval. The level quantization interval ∆C is

determined by the accuracy of generator voltage

regulation at a constant current frequency.

The number of discrete values of the connected

capacitance n depends on the choice of individual

capacitor sections capacitances.

The minimum number of discrete value levels that

differ in the quantization interval ∆C will be minimal

if the section capacitances are the same. The

maximum number of discrete levels is achieved if the

ratio of the capacitor sections capacitances is

determined by the weights of the digits of the binary

number system:

123

: : ... 1: 2 : 4...2 ; 2= ≤≤

CCCC N n

The capacitor device shown in Fig. 1 is discrete not

only in terms of level, but also in time, i.e. is

impulsive. It belongs to the class of digital automatic

control systems with a limited number of bits N. The

quantization of the control signal in time is due to the

physical properties of capacitors and the technical

characteristics of semiconductor switches.

Uncoordinated capacitors inclusion leads to their

overcharging by pulsed currents, which can lead to

breakdown of switching elements and significant

electromagnetic interference. Therefore, the capacitors

inclusion in AC circuits is carried out when the

voltage on the key is zero, Fig. 1. The capacitor

disconnection from the network occurs when the

current through the capacitor stops. Because the

capacitance current leads the voltage by a quarter of a

period, so the capacitor is disconnected from the

network at its maximum charge.

The agreed switching times of the capacitors in

each phase do not coincide in time. Therefore, the

switching control period of a three-phase capacitor

section takes at least half of one period of the AC

network. Otherwise, bump less switching will become

impossible.

The average value of the generator three-phase

voltage U

g is measured by a voltage sensor during

each period of the generated current, [10]. At the end

of the measurement period, the voltage U

g is

compared with the set voltage U

0. To eliminate

voltage modulation caused by switching sections of

capacitors, a dead zone is introduced into the system

If this difference |∆U|=|U

g-U0 |>Uzis outside the

set dead zone U

z, then the deviation from it is

converted into an N-bit binary number A

n depending

on the voltage U

ud, which determines the discreteness

691

of the voltage deviation (controller sensitivity) by

level:

( )

{ }

/.= ∆−

n z ud

A Round U U U

(1)

If the generator voltage is in the dead zone

0-Uz≤Ug≤U0+Uz, then the number An is zero.

The discrete number A

n (1) is proportional to the

voltage deviation in the n-th period. A control binary

number C

n is used to control the number of capacitor

sections connected. Each digit c

i of the binary control

number C

n controls the corresponding switches that

switch one of the three-phase capacitor banks. If, c

i=1

then the switches connect the i-th block of capacitors

to the stator windings of the generator, and if c

i=0,

then the switches disconnect the i-th block of

capacitors.

Depending on the use of numbers A

n and Cn it is

possible to realize several discrete laws of AG voltage

control. Comparing their effectiveness is the purpose

of this article.

In works [8, 9], an integral discrete-pulse control

law is described, in which the binary number C

n+1,

which controls the switching of capacitor sections in

each current period n + 1, is determined by the sum of

binary numbers in the previous control period n:

= +

n nn

C CA

(2)

Transient processes in an electrical installation

with a system of discrete capacitor voltage

stabilization AG were studied on computer and

physical models created by the authors [9].

Let us consider the form of the main transient

processes and the influence of the tuning parameters

of a discrete-pulse voltage regulator on the dynamic

characteristics of an installation with an AG (the main

parameters of AG are presented in relative units:

)

On Fig. 2 shows the dynamic processes of current,

voltage, capacitance, and frequency in relative units in

an asynchronous electrical installation when 50% of

the active-inductive load is turned on with cosϕ=0.8 .

Here, in the four-digit integral voltage regulator (2),

the dead zone U

z=±0.05 and the discreteness level

ud=0.01 of the regulator are set. The voltage enters

the dead zone, and the stabilization process ends in a

quarter of a second.

The tuning parameters of the integrated discrete-

pulse controller are the number of discharges and the

number of switched sections of capacitors N, the dead

zone U

z and the controller sensitivity Uud. The

capacitance C

1 switched by the first digit of the control

number C

n must be matched to the dead band Uz.

On Fig. 3 shows the processes of voltage regulation

when the generator load is turned on with different

levels of the regulator discreteness (sensitivity) U

ud.

The controller sensitivity U

ud significantly affects

the response rate to voltage deviation and reduces the

transient process time to 0.1 ... 0.2 seconds. When

ud<0.005 the system loses its stability.

The voltage stabilization accuracy in the system is

set by the value of the dead zone U

z. On Fig. 2 and

Fig. 3 U

z=±0.05, so the stabilization accuracy is ±5%.

Doubling the static stabilization accuracy requires

halving the capacitance of the least significant bit and

increasing the number of bits and capacitor sections

per unit. This significantly increases the cost of the

regulator.

Figure 2. Switching on 50% resistive-inductive load with

cos(ϕ)=0.8: I

a - generator phase current; Ud - voltage sensor

output; ω - drive engine frequency rotation; Cn - the number

that controls the capacity; A

n - capacity addition; Ug -

generator phase voltage

Figure 3. Switching on 50% active-inductive load with

cos(ϕ)=0.8 at different levels of controller sensitivity

ud=0.007...0.5

If the change in the generator voltage with an

increase in the excitation capacitance by the value of

the least significant bit C

1 is greater than the dead

zone, then self-oscillations occur in the system relative

to the dead zone ±U

z, Fig. 4.

Figure 4. Decreasing and increasing the dead zone of the

controller U

692

On Fig. 5 shows the dynamic processes of voltage

establishment for various switched loads in a system

with an integral control law C

n+1=Cn+An.

The dynamic deviation when switching 25% of the

load is 6...7%, when switching 50% the deviation is

approximately 10...12%, and at 75...100% - about

14...16%. The time for the AG voltage to enter the

dead zone does not exceed 0.3 seconds.

Figure 5. Switching of various active-inductive loads with

cos(ϕ)=0.8 in a system with an integral control law

n+1=Cn+An

The moment of switching the load is random, and

the control is synchronized with the network. The

difference of these moments ∆t lies within the

network period: ∆t=0…T

0. The voltage transient

depends on ∆t, Fig. 6.

Figure 6. The moment offset of applying the perturbation

The difference in the regulation processes

manifests itself in the unsaturated section of the AG

iron magnetization curve, i.e. when the load is turned

on and the voltage is reduced. The oscillations

number can be from one to three.

The desire to speed up the stabilization process

when switching large loads and voltage dips leads to

forcing control by a non-linear increase in the addition

of capacitors, i.e. numbers A

n in the control law (2).

The acceleration of the stabilization process with

forcing is shown in Fig. 7. Here in the four-digit

controller A

n=8 with ∆U<-4Uud and An=12 with ∆U<-

ud.

The rate of the perturbation change is considered

by introducing a derivative into the control law. In the

discrete case, the rate of deviation ΔU change can be

considered by introducing the deviation difference on

adjacent control cycles into the law, for example:

+−

=+−

n n nn

C C AA

(3)

The law (3) implements the integration and

differentiation of the voltage deviation. The switching

processes of the active-inductive load of the AG with

control (3) are shown in Fig. 8.

Figure 7. Forcing control

Figure 8. Digital integral-differential control law

Cn+1=Cn+2An-An-1

Comparison of switching processes of 75% active-

inductive load using control laws (2) and (3) is shown

in Fig. 9.

Figure 9. Comparison of discrete control laws (2) and (3)

An increase in the differential component in the

law C

n+1=Cn+3An-2An-1 leads to unacceptable oscillation

of the transient process, Fig. 9.

The application of the combined control principle

based on the deviation of the controlled parameter

(voltage) and perturbation (load) requires an

additional measurement of the load current during

one voltage period T

0, Fig. 1. The effect of active Ig and

inductive I

l load current on the generator voltage Ug is

different. A resistive load causes the armature to react,

while an inductive load demagnetizes the generator.

Therefore, separate measurement of active and

reactive load is necessary.

The combined voltage control law of the generator

has the form:

693

( )

{ }

( )

{ }

Round Round

−−

=++ − + −

n n n g gn l ln

gn ln

C C A K I I KI I

(4)

where K

g, Kl-adjustment coefficients for active and

reactive current.

Comparison of the switching processes on the load

of the AG with control by deviation and with the

combined control law is shown in Fig. 10.

Figure 10. Integral and combined control law. Switching on

75% active-inductive load with cos(ϕ)=0.8

The complication of the control law by considering

the switched load slightly improved the dynamic

properties of the system.

3 CONCLUSIONS

1. The discrete-pulse law (2) implements the integral

control of the generator voltage deviation. The

correct choice of its tuning parameters provides

the requirements of regulatory documents for the

dynamic properties of the ship's electric power

plant (voltage dip less than 15%, transient duration

less than 0.5 sec.), [11].

2. When choosing the tuning parameters of the

voltage regulator, it is necessary to analyse the

processes with possible changes in the load (Fig. 5,

Fig. 8) and shifts in the moments of its switching

(Fig. 6). Digital integration of the sensor signal of a

three-phase circuit and the formation of a

synchronizing pulse allows you to optimize the

ratio of the analogue and digital parts of the sensor

according to the ease of implementation criterion.

3. Improving the dynamic properties of the AG

voltage stabilization system with the control law

(2) is achieved by introducing excitation forcing at

large deviations (Fig. 7) and control by the

deviation change rate (3), (Fig. 8 and Fig. 9). The

complication of the implementation of these

changes to the law (2) is insignificant and does not

require additional equipment.

4. The implementation of the combined principle of

control by deviation and disturbance (4) does not

lead to significant improvements in the dynamic

properties of the stabilization system (Fig. 10), but

requires additional measurements and controller

settings.

REFERENCES

[1] Energy efficiency handbook. Toward zero emission ship

operations. Helsinki: ABB Marine & Ports,

9AKK107680A8255.(2020).

https://new.abb.com/marine/energy-efficiency-

handbook

[2] White paper - ABB Marine and Cranes. Using Variable

Frequency Drives (VFD) to save energy and reduce

emissions in newbuilds and existing ships. Energy

efficient solutions. – ABB (2012).

[3] Miikka Jaurolaa, Anders Hedinb, Seppo Tikkanenaand

and Kalevi Huhtala “Optimising design and power

management in energy-efficient marine vesselpower

systems: a literature review”. JOURNAL OF MARINE

ENGINEERING & TECHNOLOGY, vol. 18, No. 2, p.p.

92–101. (2019).

https://doi.org/10.1080/20464177.2018.1505584.

[4] Veretennik A.M. Self-oscillatory processes in the

frequency stabilization system of a marine diesel

generator // Ship power plants: scientific and technical.

Sat. - Odessa: ONMA. - No. 12. – p.p. 30-36.(2005).

[5] Vishnevsky, M. Mukha, A. Drankova "Improving the

energy efficiency of ship electrical installations by using

asynchronous generators". Proceedings of the Institute

of Electrodynamics of the National Academy of Sciences

of Ukraine. - Kyiv, Vol. 56. pp. 40-46. doi:

10.15407/publishing 2020.56.040 (2020).

[6] Kitsis S.I. Asynchronous self-excited generators. - M.:

Energoatomizdat, - 328 p. (2003).

[7] Toroptsev N.D. Asynchronous generators for

autonomous electric power plants. - M.: NTF

"Energoprogress", - 87 p. (2004).

[8] Patent No. 113039 C2, UA, H02P9/46, "Method of

controlling the voltage of an asynchronous generator" /

L.V. Vishnevskyi, D.L. Vishnevskyi, Publ. in Bull. No.

22, 25.11. (2016).

[9] Vishnevsky L.V., Mukha M.Y, Dao Ming Quan. Voltage

control of autonomous asynchronous generator sets / //

Monograph - Odessa: NU "OMA". - 196 p. (2016).

[10] Vishnevsky, L., Mukha, M., Vishnevsky, O.,

Vishnevsky, D.: Voltage Sensor of the Autonomous

Generating Set. In: 2020 IEEE 15th International

Conference on Advanced Trends in Radioelectronics,

Telecommunications and Computer Engineering

(TCSET). pp. 712–715 (2020).

https://doi.org/10.1109/TCSET49122.2020.235525.

[11] ISO 8528-1:2005. Alternating current generator sets

driven by reciprocating internal combustion engines.