wind turbines in the high voltage networks of czech … · (tk), fitted with the transformer...

44 Acta Electrotechnica et Informatica No. 3, Vol. 4, 2004

WIND TURBINES IN THE HIGH VOLTAGE NETWORKS OF CZECH REPUBLIC

Jan MÜHLBACHER, Pavel NOVÁK West Bohemia University in Pilsen, Faculty of Electrical Engineering, Department of Electrical Power Engineering and

Environmental Engineering, Univerzitni 8, 306 14 Pilsen, Czech Republic, Phone: (+420) 377 634 300, Fax: (+420) 377 634 310, E-mail: [email protected], [email protected]

SUMMARY It has been generated a lot of gigawatthours from new wind farms all over the world in the last years. Otherwise, there

appeared problems with connection and operation associated with this “boom” of wind power. Questions like voltage changes, flicker, harmonic currents, behaviour during network disturbances and regulation of reactive power are presently discussed.

The purpose of this article is to show possibilities of wind power in Czech Republic, with aspects of integration to the high voltage networks. The rules for operation of transmission and distribution lines are determined by Energy Regulatory Office. According to these rules, which include the connecting conditions and evaluation of wind turbine grid impacts, is made an evaluation of small wind farm grid connection.

Keywords: wind turbine, legislature, connecting conditions, grid impacts, wind farm, voltage change

1. INTRODUCTION

A continental position of Czech Republic (CR) and complicated topographical conditions implicate the decrease and variability of wind speed. Hence, the proper locations can be nearly always found in higher above sea levels, usually above 600 metres. After the reduction of the legislative, ecologically and wind potential insufficient locations the viable technical potential of CR is according to some judgments 700 – 1000 MW, which represents the production of 1.5 – 2.5 TWh yearly.

Although CR has a quite satisfactory viable potential of wind power, the proper locations are situated mostly at mountains or foothills regions. But these locations represent from the power system (grid) point of view the undersized ends of distribution grids. Therefore, it is important to ensure high-quality conditions when connecting the wind turbines (WT) to the grid, so the standard power supply quality for other customers would not be disturbed. 2. LEGISLATIVE CONDITIONS

According to the power law, the Czech distribution companies released the rules of the power systems operation. It means that on the whole territory of CR pay the same rules, which vary only in the details of appropriate company. These rules apply to all kinds of renewable energy sources. The manufacturers of renewable energy have as well the right of priority to be connected to the electric grid. The decisive requirements concerning the grid connection of WT are included in appendix no. 4 – “The rules for parallel operation of sources with LV or HV grids”. Except these rules, it must be of course adhered the valid norms, the business standards and the regulations for work safety, when operating renewable sources.

Next two chapters mention the most important conditions, that have to be fulfilled when connecting the WT to the grids, and whereby evaluate their grid impacts [1, 2]. 2.1 CONNECTING CONDITIONS 2.1.1 Switching device When connecting WT to grid, it has to be used the switching device, with the minimally ability of switching-off the load. Such a device can be e.g. a circuit-breaker, a switch-disconnector-fuse or a section-disconnecting switch. 2.1.2 Voltage changes at steady and switching

operations

These voltage changes must be in comparison with the voltage at their disconnection in the most adverse case:

− in LV grids 3% nU U∆ ≤ (1) – in HV grids 2% nU U∆ ≤ (2)

where: ∆U - voltage change Un - nominal network voltage

The values may be tolerated higher after the

settlement with distribution company, according to the starting-up duration. Respecting the minimisation of grid impacts, it is however necessary to prevent the simultaneous switching of more generators. The technical solution is realized by a time step of each switching operation, which is dependent upon the evoked voltage changes. For WT pays a special "factor of switching dependent

44Acta Electrotechnica et Informatica No. 3, Vol. 4, 2004

45Acta Electrotechnica et Informatica No. 3, Vol. 4, 200445

WIND TURBINES IN THE HIGH VOLTAGE NETWORKS OF CZECH REPUBLIC

Jan Mühlbacher, Pavel Novák

West Bohemia University in Pilsen, Faculty of Electrical Engineering, Department of Electrical Power Engineering and Environmental Engineering, Univerzitni 8, 306 14 Pilsen, Czech Republic, Phone: (+420) 377 634 300,

Fax: (+420) 377 634 310, E-mail: [email protected], [email protected]

SUMMARY

It has been generated a lot of gigawatthours from new wind farms all over the world in the last years. Otherwise, there appeared problems with connection and operation associated with this “boom” of wind power. Questions like voltage changes, flicker, harmonic currents, behaviour during network disturbances and regulation of reactive power are presently discussed.

The purpose of this article is to show possibilities of wind power in Czech Republic, with aspects of integration to the high voltage networks. The rules for operation of transmission and distribution lines are determined by Energy Regulatory Office. According to these rules, which include the connecting conditions and evaluation of wind turbine grid impacts, is made an evaluation of small wind farm grid connection.

Keywords: wind turbine, legislature, connecting conditions, grid impacts, wind farm, voltage change

1. INTRODUCTION

A continental position of Czech Republic (CR) and complicated topographical conditions implicate the decrease and variability of wind speed. Hence, the proper locations can be nearly always found in higher above sea levels, usually above 600 metres. After the reduction of the legislative, ecologically and wind potential insufficient locations the viable technical potential of CR is according to some judgments 700 – 1000 MW, which represents the production of 1.5 – 2.5 TWh yearly.

Although CR has a quite satisfactory viable potential of wind power, the proper locations are situated mostly at mountains or foothills regions. But these locations represent from the power system (grid) point of view the undersized ends of distribution grids. Therefore, it is important to ensure high-quality conditions when connecting the wind turbines (WT) to the grid, so the standard power supply quality for other customers would not be disturbed.

2. LEGISLATIVE CONDITIONS

According to the power law, the Czech distribution companies released the rules of the power systems operation. It means that on the whole territory of CR pay the same rules, which vary only in the details of appropriate company. These rules apply to all kinds of renewable energy sources. The manufacturers of renewable energy have as well the right of priority to be connected to the electric grid. The decisive requirements concerning the grid connection of WT are included in appendix no. 4 – “The rules for parallel operation of sources with LV or HV grids”. Except these rules, it must be of course adhered the valid norms, the business standards and the regulations for work safety, when operating renewable sources.

Next two chapters mention the most important conditions, that have to be fulfilled when connecting the WT to the grids, and whereby evaluate their grid impacts [1, 2].

2.1 Connecting Conditions

2.1.1 Switching device

When connecting WT to grid, it has to be used the switching device, with the minimally ability of switching-off the load. Such a device can be e.g. a circuit-breaker, a switch-disconnector-fuse or a section-disconnecting switch.

2.1.2 Voltage changes at steady and switching operations

These voltage changes must be in comparison with the voltage at their disconnection in the most adverse case:

· in LV grids

(1)

· in HV grids

2%

n

UU

D£

(2)

where:

(U - voltage change

Un - nominal network voltage

The values may be tolerated higher after the settlement with distribution company, according to the starting-up duration. Respecting the minimisation of grid impacts, it is however necessary to prevent the simultaneous switching of more generators. The technical solution is realized by a time step of each switching operation, which is dependent upon the evoked voltage changes. For WT pays a special "factor of switching dependent upon grid", whereby evaluates their switching operations and which also respects a short time transient phenomena.

2.1.3 Connecting the synchronous generators

At synchronous generators, it is necessary such a synchronizing unit, whereby the conditions for synchronizing may be fulfilled:

· voltage difference

n

U

%

10

U

£

D

(3)

· frequency difference

Hz

5

,

0

f

±

Acta Electrotechnica et Informatica No. 3, Vol. 4, 2004 45

upon grid", whereby evaluates their switching operations and which also respects a short time transient phenomena. 2.1.3 Connecting the synchronous generators

At synchronous generators, it is necessary such a synchronizing unit, whereby the conditions for synchronizing may be fulfilled: – voltage difference nU%10U ≤∆ (3) – frequency difference Hz5,0f ±

46 Wind Turbines in the High Voltage Networks of Czech Republic

are estimated and depend on the flicker factor c. It is necessary to fulfil the limiting value, in connection point: Plt ≤ 0.46 (Alt ≤ 0.1) (6) where: Plt - long time severity index Alt - flicker factor disturbance

2.2.2 Flicker

The flicker factor c of WT indicates an index of subjective perceptions at the luminous flux changes. Producer determines it. It depends above all on the operation regularity of the unit and at his classification used to be the WT critical. Their flicker factors are up to 40, according to experiences. Index Plt can be calculated according to:

( )⎥⎦

⎤⎢⎣

⎡ϕ+ψ⋅⋅= ikV

kV

Alt cosS

ScP (7)

where: SA - apparent power of the unit SkV - short-circuit power in connection point ψkV - phase angle of short-circuit impedance ϕi - phase angle of unit current

It pays for WT, that:

– at units with inverters is a tendency to the lower c values than at units with direct grid connected asynchronous or synchronous generators

– if the plant consists of n same generators, it happens a particular “compensation” of each generator flicker factor

2.2.3 Harmonic currents

The harmonics occur above all at units with inverters or frequency converters. If the voltages of harmonics are higher than limits, it comes into question:

– an installation of harmonic filters – a connection to the point with lower grid

impedance, so with higher short-circuit power

2.2.4 Impacts on CT signal

The CT system is usually operated with frequency between 180 and 1050 Hz. Its level must not in each grid point decrease about more than 10 % to 20 % below the desired value. From this reason it is necessary to consider these aspects:

– the synchronous or asynchronous generators, connected to the grid over the transformer, generate the lower signal decrease, the higher is the grid short-circuit power

– the evoked interfering voltage, whose frequency matches or is very closely to the locally used CT frequency, must not exceed 0,1 % Un

3. EXAMPLE OF GRID CONNECTION

EVALUATION [3]

The wind farm consists of 5 WT and is situated in Krusne Hory region, in one of the windiest and for wind power supply most promising region in Czech Republic. Upon the evaluation of wind speeds was chosen the most suitable type of WT. Vestas V52 has a nominal power of 850 kW and is applied with the most widely used technology – with variable speed asynchronous generator. Close to the location can be found a distribution substation (DS) and the HV distribution grid 22 kV. To this voltage level it is possible to connect a farm at most of five WT. Larger farm would be necessary to connect to 110 kV grid, which is 6 kilometres far, and though would considerably increase the cost of investments. 3.1 Technical concept of grid connection

The way of connection the farm to grid was considered according to figure 1, to the DS "Celnice" with the 22 kV cable 3x AXEKVCEY 1x70 mm2. The length of the cables was considered jointly 300 metres according to the distance of each WT. The distance from DS was considered 500 metres, according to hygienic rules.

Fig. 1 Scheme of wind farm grid connection At each WT is placed a transformation kiosk

(TK), fitted with the transformer 22/0.69 kV. In each kiosk is placed a switch-board of HV 22 kV, with the SF6 gas isolation to protect the transformer and to connect the HV 22 kV. In WT2 to WT5 are the switch-boards equipped with fuse sets. Switch-board in WT1 is equipped with the circuit-breaker, with current and voltage measuring transformers and with the directive protection. This protection is able to react at short-circuits, at over-currents, at unipolar ground faults in foreign unit, perhaps even at over-voltages from WT. At frequency decrease it prevents the “island running”.

For the cable connection to grid it is used an existing backlog in DS. The way of the cable connection in DS allows to distribution company to remote-control the switch-disconnector of the cable. Measuring of the power supply is carried out at HV side, DS is therefore necessary to equip with current and voltage measuring transformers. It is possible to read the values over modem line.


3.2 Grid impacts of wind farm

The calculation of grid impacts was simulated in a computer program. In the model were used the parameters declared by Vestas. The scheme of connection in figure 2 is described in table II

Fig.2 Scheme of 22 kV circuit from DS “Ostrov”

Fig. 2 Scheme of 22 kV circuit from DS “Ostrov”

DS Un [kV]

Sk[MVA]

Ik[kA] Z0/Z1 R/X

110/22 kV 23.1 276 6.9 1 0

Demand U [kV]

I [A]

S [kVA]

P [kW] cos ϕ

D1 23 83.84 3340 3173 0.95

D2 23 46.69 1860 1767 0.95

D3 23 77.57 3090 2935.5 0.95

WT1–WT5 Sn[kVA] cos ϕn k cos ϕk c

V52 894 0.95 3 0.3 3

Tab. 2 Parameters of each element where: Ik short-circuit current Sk short-circuit power cos ϕn nominal power factor cos ϕk short-circuit power factor K ratio between starting and nominal current C flicker factor

With the calculation, the voltage changes were

determined in each node up to the circuit outlet in DS “Ostrov”. The system was then classified in three operational situations; the results are in figure 3.

a) at steady operation b) at start up of WT1 and others disconnected c) at start up of WT1 and others at steady

operation

Fig. 3 Voltage changes in each node

As is seen in figure 3, it is met the condition for voltage changes at steady operation (according to chapter 2.1). In both starting-up variants, the unacceptable voltage changes however occur. If the switching would not be more often than once per day, it would be possible to tolerate these changes, after the settlement with distribution company. Otherwise it is necessary to adjust the generator start-up, e.g. by using a soft-starter (thyristor starter), a frequency converter, an Y-delta change-over switch or to start-up the WT by wind. In the worst case, it is then necessary to connect the farm to the point with higher short-circuit power. Harmonic currents declared by Vestas meet the limits given by distribution company, moreover the short-circuit power in connection place is sufficient. The CT signal will not be influenced as a result of using the block transformers. 4. CONCLUSION

Today, wind power is a fully established branch on the electricity market in EU. The energy gain is not the only criteria to be considered when installing new wind turbines. Any device connected to the electric grid must fulfil the Power Quality standards and this is a particularly interesting and important issue to be considered in the case of wind power installations.

The subject of this article was above all an outline of solving these problems in Czech Republic. They were mentioned the present and future rules for grid connections of wind turbines and the questions of their grid impacts. On a concrete was classified the possibility of connection the wind farm with a view to the evoked voltage changes, above all. REFERENCES [1] Distribution companies: The rules for parallel

operation of sources with LV or HV distribution grids 2004. (in czech)

[2] Procházka, K.: Current requirements and specifications for evaluation the connection of dispersed sources in distribution grids, EGC-EnerGoConsult Ltd., 2003. (in czech)

[3] Novák, P.: Project of wind farm in Bozi Dar region, diploma thesis, Pilsen 2003. (in czech)

[4] Zander, W.: Connection and operation of wind turbines in power systems, Büro für Energiewirtschaft und technische Planung GmbH, 2003. (in czech)

BIOGRAPHY Jan Mühlbacher, for biography see page 26 of this journal. Pavel Novák graduated at Faculty of Electrical Engineering, West Bohemia University in Pilsen in 2003. Now he is an internal post graduate (PhD.) student. Thesis title is “Integration of renewable energy sources to the electricity supply system”.


CHALCOGENIDE MATERIALS FOR ELECTROCHEMICAL SOLAR CELLS

Péter TURMEZEI Institute of Microelectronics and Technology, Budapest Polytechnic,

Tavaszmező street 17, H-1084 Budapest, Hungary

SUMMARY The problem of electrical energy storage can possibly be solved with the help of electrochemical solar cells, which are

suitable to generate either electrical energy or hydrogen gas under special conditions. The greatest problem of the electrochemical solar cell technology is to find novel materials which have appropriate properties for electrochemical energy conversion. In this work Cd4GeSe6, a novel material for electrochemical solar cells, will be presented. Keywords: Electrochemical solar cell, chalcogenide material 1. INTRODUCTION

Solar cell technology is a well-developed area of electronics; however, there are still some unsolved issues under research. Theoretically it is impossible to reach 100% efficiency, as the semiconductor materials or combination of them are suited only for specific spectral ranges. The energy of the rest of the spectrum cannot be utilized, because the light quanta (photons) in that range do not have enough energy to "activate" the charge carriers. A certain amount of this surplus photon energy is transformed into heat rather than into electrical energy. The maximum efficiency of the energy conversion can be determined by the help of

η = p pg

p p0

( )

( )

Wg

W N W dW

W N W dW

∞

∞

∫

∫ p (1)

formula. Where N(W) is the photon density function and Wg is the energy necessary to activate a charge carrier, in case of semiconductor materials Wg is equal to the band gap. If Wg is low many electron-hole pairs are generated, but their energy will be low, as well. However, if Wg is high, though the energy of the electrons will be high, the number of them will be little. Let’s approximate the spectrum of the sun with a black body, which is at 5800K temperature. In this case the efficiency in the function of Wg, based on equation (1), has the following shape (Fig. 1.).

It can be seen that the efficiency has its maximum, approximately 44%, around Wg =1.1eV. This is not the efficiency of the solar cell, but the efficiency of the photon-electron conversion. In case of semiconductor solar cells the voltage at the terminals is limited by the forward voltage. The maximum efficiency achieved in case of monocrystalline silicon solar cells is 24% in laboratory, but it is not more than 14-17% in batch production. The efficiency of GaAs solar cells is better, but their manufacturing costs are higher, as well.

Fig. 1 The maximum efficiency of the energy conversion

The electrochemical solar cells are relatively

simple solar cells. The photoelectric transformation of solar energy into electric energy or into storable chemical energy source (hydrogen) has three steps. The first step is excitation of electrons by photo absorption. This can be done either by means of semiconductors or by photosensitive electrochemical reaction. The second step is the separation of charge carrier pairs by electric field. Finally the energy in the electrons and holes is utilized by reduction or oxidation of the molecules in the electrolyte.

The electric field necessary for the charge separation does not only exist at solid state – semiconductor (having different potential barrier) interface, but also at electrolyte-semiconductor interface, too. The choice of electrolyte (to have appropriate redoxpotential) and semiconductor (to have appropriate forbidden gap) is important both for optimum utilization of solar energy and for development of high enough electric field at the interface. If the semiconductor is n type then due to the direction of the developed electric field the generated electrons are drifted towards inside the semiconductor, while the holes propagate towards the interface. In p-type semiconductor the direction of the developed electric field is opposite and so is the movement of the charge carriers. At the interface charge transport takes place into the redox system. Then, in case of n-type semiconductor the oxidized, in case of p-type material the reduced molecules diffuse away from the surface. The charge transport


takes place until the equilibrium state has been established. At this time potential difference can be measured between the semiconductor and the electrolyte. This potential difference can deliver power if it is connected into an electric circuit. The energy conversion works permanently if the redox system is regenerative, i.e. if the oxidized (reduced) system on the semiconductor is reduced (oxidized) back at the metal electrode. 2. THE ELECTROCHEMICAL SOLAR CELL

One of the greatest problems in solar cell applications is the storage of electrical energy. This problem can possibly be solved with the help of electrochemical solar cells, which are suitable to generate either electrical energy or hydrogen under special conditions [1]. The technology of electrochemical solar cells has some technical and scientific problems. One such problem is photocorrosion, which occurs at the electrolyte–semiconductor interface. Photocorrosion damages the semiconductor electrode during the operation of the solar cell. Direct energy conversion relies on the semiconductor material, which can absorb a fraction of the solar spectrum depending on its Wg bandgap energy. Unfortunately, many materials with adequate bandgaps are susceptible to photocorrosion, due to to destructive hole-based reactions. Also, the semiconductors less susceptible to photocorrosion, such as metal oxides like TiO2 and SnO2, exhibit a too large bandgap to permit significant collection of visible light.

An alternative to overcome the limited spectral sensitivity of the wide bandgap semiconductors, which are restricted to UV light, is surface modification with visible-light absorbing dye molecules. The sensitization of semiconductors using dyes is a century old, when it was used in the development of photography. Their application in solar energy conversion is more recent, and progressed considerably after the seventies, with the advanced in the development of dye sensitizers, especially bipyridyl Ru complexes with anchoring groups to attach them to the semiconductor surface.[2]

Ruthenium based dyes, and specially those casting pyridil ligands, have proven very effective in photovoltaic conversion. Their molecular properties, and the possibilities of related applications, are directly related to: (a) the energetic spectrum of the molecules that determines the optical absorption and emission characteristics; (b) the molecular electron density (and its polarizability) which is associated to the reactivity of the molecule.

Photoelectrochemical cells based on dye-sensitized semiconductor electrodes also include solutions containing a suitable redox couple and a counter-electrode. The illumination of the dye to an electronically excited state which is quenched by electrontransfer to the conduction band of the semiconductor, leaving the dye in an oxidized state.

The oxidized dye is reduced by the electron donor present in the electrolyte. The electrons in the conduction band are collected, flow trough the external circuit to arrive at the counter-electrode, where they cause the reverse reaction of the redox mediator. Thus, the photoelectrochemical cell is also regenerative and the process leads to direct conversion of sunlight into electricity .If only the above reactions took place, the solar cell would be stable, delivering photocurrent indefinitely. The maximum photovoltage, at open circuit potential (VOC), is the difference between the Fermi level of the semiconductor under illumination and the redox potential of the mediating redox couple. The photocurrent yield depends on the spectral and redox properties of the dye, its excited state lifetimes, the efficiency of charge injection, the ionic conductivity of the electrode to collect and channel the electrons through the external circuit.

The redox couple in the electrolyte also is crucial importance for stable operation of a dye-sensitized solar cell, because it must carry the charge between the photoelectrode and the counter-electrode for regeneration of the dye. After electron injection, the electron donor in the electrolyte must reduce the oxidized dye to the ground state as rapidly as possible. Thus, the choice of this charge mediator should take into account its redox potential, which must be suitable for regenerating the dye. Also, the redox couple must be fully reversible and should not exhibit significant absorption of visible light. Another important requirement is related to the solvent, which should permit the rapid diffusion of charge carriers, while not causing the desorption of the dye from the oxide surface.

The band diagram of the dyesensitized photoelectrochemical solar cell is shown in Fig. 2.

Fig. 2 The band diagram of the dyesensitized photoelectrochemical solar cell

A possible direction of this research is the search

for novel materials with appropriate properties for electrochemical applications. One of the important groups of such semiconductor compounds is the chalcogenides such as Cd4GeSe6.

50 Chalcogenide Materials for Electrochemical Solar Cells

In this work Cd4GeSe6, a novel material for electrochemical solar cells, will be presented. The properties of this material will be investigated, which has been scarcely done before, and that is why these properties are not known in detail. Cd4GeSe6 belongs to the agryrodite family, of which lattice parameters were determined [3]. The band gap and type of band transition was determined by absorption and the I-V characteristics was determined by photoelectro-chemical method [4]. Furthermore it was found that this material shows very good resistivity against photocorrosion [5]. The knowledge of the electrical parameters of the Cd4GeSe6–electrolyte junction is very important for solar cell applications. It was also determined in this work. The properties of the Cd4GeSe6 crystal–electrolyte junction are investigated with impedance analysis. The evaluation of the measured data was carried out with the help of a computer program developed by us in Pascal language. We used an equivalent circuit with physical meanings, this circuit was appropriate for the calculations [6]. 3. THE CHALCOGENIDE MATERIAL

A possible direction of semiconductor research is the search for novel materials with appropriate properties for different applications. One of the important groups of such semiconductor compounds is the group of chalcogenides, a well-known example of binary compounds. They are good photoconductors and have high absorption coefficient. Material properties can be improved and modified by forming ternary, quaternary etc. compounds of the above. Ternary chalcogenid materials, such as Cd4GeSe6, were synthesized in which new covalent chemical bonds appeared. Due to these covalent bonds these materials show higher resistance against corrosion. This novel property in itself makes novel applications, such as photoelectrochemical energy conversion electrode, possible.

In this work the properties of Cd4GeSe6 are investigated which are until now scarcely studied and therefore not known in details. The existing data in the literature differ over a wide range even for fundamental material parameters such as lattice parameters, band gap or type of band transition. This material belongs to the agrirodite family, which belongs to the monoclinic crystal class. The structure of chemically analogous compounds was investigated earlier [7]. The optical parameters of Cd4GeSe6 were scarcely studied possibly because of the difficulty of making larger pieces of single crystal.

The synthesis of Cd4GeSe6 crystal can be carried out from chalcogenide and dichalcogenide sources. The crystallizing period is several weeks long. The Cd4GeSe6 is a stable crystal and keeps its stability even at high temperatures under normal atmospheric conditions.

Fig. 3 The amplitude and phase diagrams of the

Cd4GeSe6 and 0.05 M H2SO4

4. PHOTOELECTROCHEMICAL INVESTIGATION

The band gap was determined by absorption [4]

and photoelectrochemical [8] methods and was found to be 1.7 and 1.75 eV respectively. A further reference [9] gives 1.5 eV band gap and indirect band transition. Ref. [10] gives 1.9 eV for band gap from photoluminescence measurement at 10 K. These strongly different photoelectrochemical and photoluminescence results are reviewed in Ref. [4].

The impedance measurements were performed in an electrochemical cell under potentiostatic control. The electrolytes were 0.05 M H2SO4 and solution. The impedance analysis was carried out with the perturbation of some mV. The modeling of an electrolyte – semiconductor junction is a difficult problem because the values of the circuit elements exhibited frequency dependence. In this work we determined the proper values of equivalent circuit components with their physical meaning for the transfer function of the junction. The parameters of the equivalent circuit are very important to know for device applications. A simple equivalent circuit with physical meaning was appropriate for the calculation [6]. It contains three parallel branches, one branch is a resistance R1, the second branch is a swinging circuit (R2C2) and the third branch is a capacitor (C3).

The evaluation of the measured data was carried out with the help of a computer program developed by us in Turbo Pascal language. The transfer function of the equivalent circuit has three solutions (one zero and two poles). In the first step these three


roots were fitted in the same time with the help of the least square method. The minimum of the error is determined with the help of gradient method, until the error becomes less than 1 Hz. The value of the constant in the transfer function was determined from the amplitude diagram with similar method [11]. The R1 resistance represents the charge transfer that is the electrochemical reaction at the interface. The value ranges from 6.2 to 7.1 kΩ/mm2. R2 and C2 represent the surface levels and deep centers where the values range between 0.6 and 1.6 kΩ/mm2 and between 5.3 and 7.6 nF/mm2, respectively. The element C3 means the space charge capacitance and its value is between 0.62 and 0.78 nF/mm2 without bias voltage. The space charge capacitance shows a little larger value in KOH then in H2SO4 solution. Measured and fitted amplitude and phase diagrams of the junction are shown in Fig. 3.

Fig. 4 The simple equivalent circuit

The junction of electrolyte–Cd4GeSe6 crystal

was investigated with impedance analysis. We set-up an equivalent circuit of this junction. The electrical parameters of the junction were determined which are very important to know for device applications. The space charge capacitance was found to be about 0.9 nF/mm2. The charge transfer resistance was about 5 kΩ/mm2. The values of the elements of the RC circuit which represent the surface levels are 1.7 kΩ/mm2 and 1.5 nF/mm2, respectively. This model describes the electrical behavior of the junction in the whole investigated frequency range. Furthermore the surface morphology of Cd4GeSe6 crystal was investigated after electrochemical treatment [5].

REFERENCES [1] Á. Nemcsics: Solar Cells and their Developing

Perspective; Academic Publisher, Budapest 2001

[2] Grätzel, M.; Nature 2001, 4/4, 338. [3] K.-F. Hesse, M. Czank and Á. Nemcsics; Z.

Kristallogr. 216, (2000) 14 [4] I. Kovách, Á. Nemcsics and Z. Lábadi;

Inorganic Materials, 39 (2003) 108 [5] Á. Nemcsics; to be published [6] Á. Nemcsics; Phys. Stat. Sol. (a), 173, (1999)

405 [7] P. Quenez and O. Gorchov; J. Cryst. Growth 26

(1974) 55 [8] W. F. Kuhs, R. Nitsche, K. Scheunemann;

Mater. Res. Bull. 14 (1979) 241 [9] B. von Krebs, J. Mandt; Z. Anorg. Allg. Chem.

388 (1972) 193 [10] P. Quenez, A. Maurer; J. Phys. 36 (1975) 83 [11] P. Turmezei, Á. Nemcsics; Phys. Stat. Sol. (c) 0

(2003) 967 BIOGRAPHY Péter Turmezei was born in Budapest, Hungary in 1949. He received the electrical engineering degree from the Technical University of Budapest, Hungary in 1973, the dr. techn. degree in 1986 and the PhD degree in 2003. He joined the Research Institute for Particle and Nuclear Physics of Central Research Institute for Physics in 1973, where his job was designing and testing nuclear measuring equipments. Since 1975 he worked at the Research Institute for Telecommunication on the field of analogue and digital circuit design, development of equipments and systems for telecommunication. Since 1987 he has been Associate Professor at the Institute of Microelectronics and Technology of the Kandó Kálmán Polytechnics of Budapest. Currently he is the head of the Institute. His latest research interests include semiconductor technology and solar cells.


THE PLASTIC SOLAR ABSORBERS AND POSSIBILITIES OF THEIR UTILIZATION

Ján TKÁČ Department of Electric Power Engineering, Faculty of Electrical Engineering and Informatics,

Technical University of Košice, Letná 9, 042 00 Košice, tel. 055/622 3478, E-mail: [email protected]

SUMMARY Utilization of solar energy becomes the more topical. The most spread application is heating the water. This contribution

deals with construction of plastic solar collector-absorber for low temperature utilization, especially for heating the water in swimming pools. This collector has a set construction. It enables the creation of required absorbing area on the surface of the ground without bearing constructions, and also the free movement of persons on the absorber surface. These solar absorbers also enable the other non-conventional possibilities of utilization. Carried-out measurements shown that solar absorbers are the most effective possibility of utilization of solar energy at present.

Keywords: solar energy, plastic absorber, energy profit 1. INDRODUCTION

Utilization of solar energy becomes the more actual for its advantages and perspectives. The present situation is characterized by spreading the possibilities of practical use of this renewable source from the most simple applications up to the most complex ones. Equipment that changes the solar energy on the other kinds of energy, reaches the outputs from some watts up to some megawatts. The main orientation is for the utility water heating and production of electrical energy. Solar collectors used for heating of the utility water are characterized with perfect construction and high effectiveness and also with the high production demands that are especially resulting from the high requirements on the quality of absorber [1]. The high quality of black absorption layer is reached by creation of coat with high absorptivity in the area of visible radiation and low emissivity in the area of infrared heating radiation. At low heat applications it is possible to simplify the construction in a great measure at preserving the required parameters by leaving out the transparent cover, supporting frame, heat insulator, or by decreasing the demands on absorption layer properties. The simplest solution is the utilization of single absorber that construction must be convenient for this application or must be adapted [3]. This utilization is especially suitable for heating of liquids to 30 °C i.e. in swimming pools where the required water temperature is 27 °C. For this application the textile-plastic absorber from firm Ekosolaris Kroměříž (Czech Republic) and new product Solar Plast (Slovak Republic) are suitable. 2. PLASTIC SOLAR ABSORBERS

In the last time more types of solar absorbers with utilization of plastics at their construction have been appeared on the European market. The most frequently used material is polyethylene. Advantages of plastics are simplicity of their

processing, possibility of high production and low prizes. The high chemical resistance of polyethylene enables the aggressive heat-transfer media to be also heated. Disadvantages of the utilization of plastics are resulting from their low resistance to ultra - violet part of solar radiation, their lower mechanical resistance at higher temperatures and non - selective surface. The first disadvantage may be influenced by utilization of UV stabilizer in the form of additions into basic material, the second one by construction and the third one by suitable way of service.

It is obvious that plastic absorbers do not reach the parameters of solar collector absorbers, but the non-conventional possibilities of their applications and also their prizes effect their broader utilization. The greatest applications of plastic absorbers are for water heating in swimming pools. 3. SOLAR ABSORBER SOLAR PLAST

Absorber Solar Plast, that samples have been produced in the Slovak Republic, belongs to the group of plastic absorbers. It is all-plastic solar absorber produced on the basis of polyethylene of high density, with high mechanical, heat and chemical resistance. Construction is in the form of set consisting from basic components in the shape of square with dimensions: 295 x 295 x 30mm /Fig. 1/.

Front wall surface of absorber is shaped with aim to decrease the reflectivity of surface at low incidence angles of solar radiation. Components are vertically assembled into columns /Fig. 2/ and by their parallel arrangement the bigger areas are created /Fig. 3/. Number of components in column and number of columns may be changed. Above-mentioned enables us to adapt the shape of absorber surface to possibilities of place or space of utilization. Absorber is intended for low heat temperature application and therefore it is constructed without transparent cover, heat insulator, supporting frame and selective absorption layer.


output

connecting and fixing points

input with seal

Fig. 1 Construction of absorber

Fig. 2 Connection of absorber parts into columns

OUTPUT

Max

. 10

pie

ces

INPUT

Fig. 3 Connection of absorber parts into plane

Mechanical properties of absorber are ensured by basic material properties and constructional solving of component. This component is mechanically strengthened in 13 points by press joints of front and rear walls. Achieved properties of absorber enable his location directly on the ground surface and free motion of persons on surface of absorption area filled with heat bearing medium. This predetermines its utilization in surroundings of swimming pools for creation of the access pavements and pass zones around swimming pool. 4. MEASUREMENTS ON ABSORBERS

SOLAR PLAST

The measurements of absorptivity and resistance tests against frost were carried out on samples of absorber. The energy profit was measured on absorber with area 1m2.

Technical data of absorber were completed with measured parameters. The value of absorption coefficient that was obtained at absorptivity measurement was A = 0,93 and that shows the excellent ability to absorb the solar radiation. Measurements of heat emissivity at temperature of working medium 80 °C have also shown the high degree of heat radiation from the absorber surface E = 0,96. The high emissivity of absorber surface effects that absorber reaches low stagnation temperature. This causes that system filled with water is not overheated even at the highest solar radiation intensity.

Test of absorber resistance to freezing were carried -out on the basic constructional component filled with water. This component was exposed to the influence of low temperature up to –30 °C. That has shown the resistance to possible frost during its service in spring or autumn.

54 The Plastic Solar Absorbers and Possibilities of Their Utilization

Technical parameters: Outer dimensions 295 x 295 x 30mm Inside volume 1,8 l Material PE HD with additions Maximum working temperature 90 °C Resistance to freezing –10 °C (–30 °C) Working pressure 0,16 MPa Testing pressure 0,5 MPa Absorption layer non-selective Absorptivity A 0,93 Emissivity E (80 °C) 0,96 Quality of absorber Q 0,97 Temperature of stagnation 39 °C Acceptance angle 125 ° Energy profit 945 Wh/m2

Measured results of energy profit and directional

characteristics measured during clear solar day are shown in Fig. 4.

Wh/m2 1000 800 600 400 200

10 20 30 40 50 60 70 80 90 °

Fig. 4 Dependence of acquired energy amount QS on incidence angle of solar rays

Fig. 5 Solar system with absorbers type Solar Plast

warm water

cold water

- warm water

direction of flow

temperature gauge

backflow valve

ball valve

discharge valve

air valve

hand three-way valve

controlled three-way valve

absorbers Solar Plast

regulator

pump

filter swimming pool


5. CONCLUSION

Solar absorbers Solar Plast are non-conventionally constructed multipurpose applicable absorbers working at water heating in swimming pools with high-energy profit 945 Wh/m2. This energy profit is enabled by their construction and by method of service at low temperatures of working medium and comparable high or also higher temperatures of surroundings. Service of swimming pools has the positive influence on benefit of energy. In this case absorbers receive except solar radiation also heat from surroundings. This fact was verified also by measurement at deflection of collector about 90° from solar rays direction. Tested absorber has received also thermal energy by permeation of heat and reflected solar radiation from surroundings. In the case when the swimming pool is not covered during the night it loses practically the whole received energy [2]. In the morning the temperature of both water and the ambient air water is low therefore the absorber will work with high effectiveness. During day the temperature of water and ambient air gradually rises. From above mentioned it results that absorber activity may be assumed in the area of maximum energy profit. Solar system with absorbers Solar Plast is shown in Fig. 5.

Utilization of Solar Plast collector in the function of pavement, fencing or pass zone from surroundings to swimming pool and also the

possibility of full recycling after finishing its life-time considerably increases its utility value. Carried-out measurements shown that solar absorbers are the most effective possibility of utilization of solar energy at present. REFERENCES [1] Marko, Š. et al.: Energetické zdroje a premeny,

Alfa Bratislava, 1988. [2] Kittler, R. – Mikler, J.: Základy využívania

slnečného žiarenia, VEDA Bratislava, 1986. [3] Dickinson, W. C. – Cheremisinoff, P.N.: Solar

energy technology handbook, Marcel Dekker Inc., New York, 1993.

BIOGRAPHY Ján Tkáč was born in Vranov, Slovakia. He received his Ing. (MSc.) degree in 1976 at the department of Electric Power Engineering of the Faculty of Electrical Engineering and Informatics at Technical University in Košice. He defended his CSc. (PhD.) in the field of high voltage technique in the year 1986. Since 1978 he is working as assistant professor on the Department of Electric Power engineering. His scientific research is focusing from the year 1985 on solar energy and renewable energy sources.


SPECIFICATION ISSUES OF COMMUNICATION AND CODE MOBILITY1

Martin TOMÁŠEK Department of Computers and Informatics, Faculty of Electrical Engineering and Informatics,

Technical University of Košice, Letná 9, 042 00 Košice, Slovak Republic, phone: +421 55 6023178, E-mail: [email protected]

SUMMARY Presented process calculus for agent communication and mobility can be used to express distributed systems based on

agent technology and mobile code applications in general. Agents are abstraction of the functional part of the system architecture and they are modeled as process terms. Agent actions model interactions within the distributed system: local/remote communication and mobility. Places are abstraction of the computational environment where the agents are evaluated and where interactions take place. Distributed system is modeled as a parallel composition of places where each place is evolving asynchronously. Formal operational semantics defines rules to describe behavior within the distributed system and provides a guideline for implementations. Via a series of examples we show that mobile code applications can be naturally modeled. Keywords: distributed system, mobile code, process calculus, agent

1 This paper was supported by the grant Nr. 1/0176/03 of the Slovak Grant Agency.

1. INTRODUCTION

Mobile agent [1] is an autonomous program that decides which places of the distributed application visits and what operations uses there. Distributed systems based on mobile agents are more flexible than static ones: they support mobile users and can reduce network bandwidth [2]. It means the user just sends an agent then disconnects from network and finally receives the agent with result upon new connection.

Formal description and specification of such systems is very important for modeling and successful implementation of the application. If we think of most important system characteristics, we identify communication and mobility as a key point. There are a lot of techniques to describe mobile processes and communication in existence. Very powerful tools for describing parallelism, communication and mobility are process algebras [3] and other formal techniques [4].

In this paper we present process algebra to describe mobile agents and their communication strategies. We provide basic abstraction of the distributed system and its parts and we define syntactic and semantics rules for modeling mobile applications. At the end we provide a formal description of three mobile code paradigms to illustrate the flexibility and expressiveness of the presented abstraction. Some very typical applications that implement code mobility are showed too. 2. ARCHITECTURE ABSTRACTION

We can identify three main entities from the abstraction of distributed system architecture: agents, interactions and places.

Agents are abstraction of the functional part of

the system. They are evaluated in distributed computational environment and they are performing basic actions in their evolution.

Interactions are events presented between two agents or more agents in the computational environment. Basic agent actions are communication and mobility.

Places are abstraction of distributed computational environment. Whole distributed system is a set of places. Each place consists of agents and they are evaluated there. Interactions between agents can rise within one place or between two or more places. 3. ABSTRACT SYNTAX

We define terms of process algebra for modeling agents that can interact by performing three basic actions (read, write and move). The agents are modeled as process terms. The constructions for building agent terms are taken from Milner’s CCS [5] and π-calculus [6, 7] and correspond to basic notions of process algebras [3].

Distributed system is defined as a parallel composition of independent places within a network. Each place is represented by its name and an agent term defining agents located inside the place. We define operator || for parallel composition of places and its notion is very similar to | operator for parallel composition of agents.

Abstract syntax of the calculus is following:

::α = (actions) | x (perform name)

( )| p xr (read name) ( )| p yw (write name)


( )| p Qm (move agent) ::P = (agents)

| 0 (inactivity) .| Pα (action composition)

1 2|| P P (parallel composition)

1 2| P P+ (choice) | A x〈 〉 (agent invocation) ::S = (system)

[ ]| pP (place) 1 2||| S S (system composition)

Symbols , , ,x y p … are called names and N is

the set of all names. Names are an abstraction of manipulated data within agent interactions. Abbreviation x is a sequence of names and { }x is a set of names in x .

Symbol α denominates the actions provided by the agents. Action x performs an operation represented by name x . Action ( )p xr reads a name that was sent by another agent to place and stores it in name

px . Action ( )p xw outputs name x in

place named as . Action moves agent term to the place and the term is computed there.

p ( )p PmP p P

Agents are defined as process terms very similar way as in other standard calculi and they are denominated as symbols. The inactivity defines an agent with no activity. Term

, ,P Q … 0.Pα is an

action composition and its notion is that when an action α is performed the term continues as . Parallel composition defines two independent agents and that can be computed in parallel. Agent term is nondeterministic choice where an agent can be computed either as or .

P1 2|P P

1P 2P

1P P+ 21P 2P

We assume that each agent abstraction A is

defined by equation ( )def

AA x P= where all free names of are contained in AP x . Process abstraction is then a term without free names while ( )A x binds names of x . Agent invocation A y〈 〉 is then the use of term where all occasions of names from AP x are substituted by y

The distributed system is composed of places Place [ ]pP is defined by its name and agent term

which is computed inside the place. System is parallel composition of independent

places in and . Given a system , we assume the existence of function

pP

1 ||S S21S 2S S

sites which returns the set of places of . The composition is defined only if , thus we can

consider a system just as a set of disjunctive places.

S 1 ||S S2= ∅

P

1 2( ) ( )sites S sites S∩

4. OPERATIONAL SEMANTICS

Presented semantics describes possible evolution of agents, places and whole distributed system without providing the actual allocation of processes and names. We will define operational semantics of the system in a notion of evaluating of the actions. 4.1 Agent Semantics

The rules of agent semantics describe the evolution of an agent. We present labeled transition P α ′⎯⎯→ where agent is derived from agent

by performing action P′

P α . Structural rules of the agent semantics are following:

( )( ). p xp x P P⎯⎯⎯→rr (A1)

( )( ). p xp x P P⎯⎯⎯→ww (A2)

( )( ). p Qp Q P P⎯⎯⎯→mm (A3)

P PP Q P

α

α

′⎯⎯→′+ ⎯⎯→

(A4)

P PQ P P

α

α

′⎯⎯→′+ ⎯⎯→

(A5)

| |P P

P Q P Q

α

α

′⎯⎯→′⎯⎯→

(A6)

| |P P

Q P Q P

α

α

′⎯⎯→′⎯⎯→

(A7)

{ / } ( )defP y x P A x P

A y P

α

α

′⎯⎯→=

′〈 〉 ⎯⎯→ (A8)

Rules (A1), (A2) and (A3) describe how the

actions are evaluated by agents. Rules (A4) and (A5) describe behavior of nondeterministic composition of agents, while rules (A6) and (A7) describe semantics of parallel composition of agents. Last rule (A8) describes invocation of agent named A .

We will use the standard notion to indicate the simultaneous of any free occurrence of

{ / }P y x

{ }x x∈ with corresponding in . { }y y∈ P 4.2 Distributed System Semantics

Semantics of the distributed system is defined by reduction relation ( ) rules which present basic computational paradigm for agent interactions within the system and evolution of the system. In addition the structural congruence ( ) is defined for the system semantics. Reduction rules are following:

→

≡

58 Specification Issues of Communication and Code Mobility

( )

[ ] [ | ]

p Q

p p

P PP P Q

′⎯⎯⎯→′→

m

(S1)

2

1 2 1

( )1 1

1 2 1 2[ ] || [ ] [ ] || [ | ]

p Q

2p p p

P PP P P P Q

′⎯⎯⎯⎯→′→

m

p

(S2)

( ) ( )1 1 2

1 2 1 2[ | ] [ | ]

p px y

p p

P P PP P P P

′ ′⎯⎯⎯→ ⎯⎯⎯→′ ′→

r w2P (S3)

1 1

1 2 1 2

( ) ( )1 1 2

1 2 1 2[ ] || [ ] [ { / }] || [ ]

p px y2

p p p

P P P PP P P y x P

′ ′⎯⎯⎯→ ⎯⎯⎯→′→

r w

p′ (S4)

1 1

1 2 1 2

[ ] [ ][ | ] [ | ]

p p

p p

P PP P P P

′→′→

(S5)

1 1 1 2

1 2 1 2

( ) ( )|| ||

S S sites S sites SS S S S

′→ ∩′→

= ∅ (S6)

1 1 2 2S S S S S SS S

′≡ → ≡′→

(S7)

Reduction rules clearly distinct between local

and remote interactions performed by located agents and provide a formal model to guide the implementation.

Rule (S1) describes movement and evaluation of an agent. Agent evaluates the agent at the same place. Agent Q is running in parallel with agents located at place . Rule (S2) is very similar to the rule (S1) while agent moves the agent Q to another place where it is evaluated in parallel with existing agents ( ) there.

P Q

pP

2p

2PRule (S3) describes synchronous communication

between two agents located at the same place. The communication is synchronized when both peers want to interact (read or write) within the same place. It means two communication actions ( )p xr and will interact when and then the name will be substituted for all occurrences of name

( )p y′w p p′=yx in term followed by ( )p xr prefix. Rule

(S4) is very similar to rule (S3) while communicating agents are located on different places.

Rule (S5) describes asynchronous evolution of subcomponents of the place. It means each site of the system is working autonomously.

How the reduction behaves in presence of operator of parallel composition of places is defined by rule (S6).

The reduction behaves with respect to structural congruence as we can see in rule (S7). Structural congruence is defined following way:

1 2 2|| ||S S S S≡ 1 (C1)

1 2 3 1 2 3( || ) || || ( || )S S S S S S≡ (C2)

The rule (C1) shows the operator || is commutative and rule (C2) shows the the operator || is associative.

5. SPECIFICATION OF MOBILE CODE

APPLICATIONS

According to the classification proposed in [8], we can single out three paradigms, apart from the traditional client-server paradigm, which are largely used to build mobile code applications:

• remote evaluation, • code on demand and • mobile agent.

However we think of distributed systems based

on mobile agents, our model of communication and mobility can describe all three programming paradigms. Now we will show expression of the three mobile code paradigms and some practical examples of mobile code applications. 5.1 Expressing Mobile Code Paradigms

Remote evaluation is performed when a client sends a piece of code to the server and server evaluates the code and client can get the results back from the server.

We define term C that sends a request for remote evaluation to the ’s place

lientServer s . Request

consists of a code and a name of the C ’s place . Then the reads the result into the name and continues as C .

P lientc Clienty

Term reads the request from his local place

Servers . The received code is stored in name x and

the name of C ’s place is stored in name . Then the code in

lient px is evaluated and the result is

sent back to the C ’s place. The is performing an independent work in .

rlient Server

SWe define the following terms where the whole

system defined by term is a parallel composition of ’s place and C ’s place:

SystemServer lient

( , ). ( ).

( , ). . ( ) |

[ ] || [

def

s cdef

s p

c s

Client P c y C

Server x p x r S

System Client Server

=

=

=

w r

r w

]

Code on demand describes the situation where a

client wants to perform a code that is presented by the server. Client asks for a code and server sends it to the client where it can be evaluated.

We define term Clie that sends a request to the ’s place

ntServer s . The request consists of a name of the ’s place c Then the reads the code from local place into the name

Client Clientx . Finally the

code is evaluated and Clie continues as . nt C


Term reads the request from his local place

Servers . The received name of Clie ’s place is

stored in . Then the sends a code to the ’s place. The is performing an independent work in .

ntp Server P

Client ServerS

We define the following terms where the whole system defined by term is a parallel composition of ’s place and ’s place:

SystemServer Client

( ). ( ). .

( ). ( ) |

[ ] || [

def

s cdef

s p

c s

Client c x x C

Server p P S


=

=

=

w r

r w

]

Mobile agent is a paradigm where an

autonomous code (agent) is sent from the client to the server. By autonomous we mean that the client and server do not need to synchronize the agent invocation and the agent is running independently and concurrently within the server’s place.

We define an abstraction ( )Agent x of a mobile agent and term Clie is moving the agent to the

’s place nt

Server s . Term is performing its independent

work and it is able to receive the agent which is then running in parallel with other S ’s actions in its local place

ServerS

ervers .

We define the following terms where the whole system defined by term is a parallel composition of ’s place and ’s place:

SystemServer Client

( )

( ).

[ ] || [

def

def

sdef

c s

Agent x P

Client Agent z C

Server SSystem Client Server

=

= 〈 〉

==

m

]

5.2 Example: Remote Procedure Call

This example show that we are able to model very traditional mobile code application that is performing remote procedure call.

A client sends a request to a server and waits for response. The request consists of procedure name and its real parameters that should be performed by a server and the address of the client’s place where to send a result.

Term sends the request with name of the procedure , its real parameters and the name of the ’s place to the ’s place

ClientProc z

Client c Server s . term reads from its local place Server s the request into the x (name of the procedure), (parameters of the procedure) and (name of the Clie ’s place). Then in parallel it runs the recursively and continues as procedure stored in

yp nt

Serverx

with parameters. When procedure y x is finished

the result is sent back to the ’s place which name is stored in .

r Clientp

The whole distributed system is defined in term where ’s place and ’s place

are computed in parallel: System Client Server

( )

( , , ). ( ).

( , , ).( | . ( ))

[ ] || [ ]

def

def

s cdef

s p

c s

Proc x P

Client Proc z c y C

Server x y p Server x y r


=

=

= 〈

=

w r

r w〉

5.3 Example: Dynamic Data Searching

This example shows the model of simple mobile agent system. We define a mobile agent, which travels from place to place and searches for an information.

A user defined by term User needs additional information on a data represented by name . User launches mobile agent Seeker that dynamically travels among nodes looking for result information

in distributed database and stores it in . If the information is found it is sent back to the User otherwise the Seeker continues in next place. User is waiting for the result and in parallel it continues with other independent work U .

z

r y

Agent , where ( , , )Seeker x h l x is searched information, is home place of the User and is local place of the agent, is reading the data from the local place . It reads either searching result or the name of the next place where to search. In the first case it sends the result stored in back to the

. In he second case it moves a new instance of the agent to the new place and ends

h l

l

yUser

pThe whole system is defined in term

where each independent place is sending either result information or the next place for searching:

System

1

1 1 1

1( , , ).( ( ) | )

( , , ) ( , ). ( )( , ). ( , , )

[ ] || [ ( , ) ( , )] ||

|| [ ( , ) ( , )]n n n

def

p u

def

l h

l p

u p p i p

p p i p

User Seeker z u p y U

Seeker x h l x y yx p Seeker x h p

System User z r z p

z r x p

= 〈 〉

= ++ 〈 〉

= +

+

m r

r wr m

w w

w w

6. CONCLUSIONS AND FUTURE WORK

Modeling rules presented in the paper seem to be very suitable tool for formal description of distributed systems based on agent technology and technology of mobile code. The formal semantics is useful for discussing the design of modeled application and provides guidelines for its implementations in programming languages.

60 Specification Issues of Communication and Code Mobility

Primitive actions defined in the model present communication and mobility as key interactions for mobile agents. Abstraction of places, their parallel composition and performing interactions within places are very natural for distributed system architectures. These approaches in our model differ from very general π-calculus and ambient calculus [9].

Security properties of distributed system are also very important area and research on presented apparatus continues in this field. For example, presence of typing information [10, 11] within the names can provide privacy and security properties. In addition implementation of spi-calculus [12] primitives can add usage of secure communication protocols to the model.

We also work on multi-agent system platform [13] where mobile agents can work together to solve the common tasks. We use these models to define and to make verification of communication schemes [14, 15] for mobile agents coordination and cooperation within the multi-agent environment. REFERENCES [1] Jennings, N. R. – Wooldridge, M. J.:

Applications of Intelligent Agents. In N. R. Jennings, M. J. Wooldridge (eds.): Agent Technology: Foundations, Applications, and Markets, Springer, Berlin, Heidelberg, New York, 1998, pp. 3 – 28

[2] Harrison, C. G. – Chess, D. M. – Kershenbaum, A.: Mobile Agents: Are They a Good Idea? Technical report, IBM Research Division, T. J. Watson Research Center, March 1995

[3] Baeten, J. C. M. – Weijland, W. P.: Process Algebra. Cambridge University Press, Cambridge, New York, Port Chester, Melbourne, Sydney, 1990

[4] Šimoňák, S. – Hudák, Š.: Petri Net Semantics for ACP Terms. Vol. 4, No. 1, Košice, 2004, pp. 55 – 59

[5] Milner, R.: Communication and Concurrency. Prentice Hall, 1989

[6] Milner, R. – Parrow, J. – Walker, D.: A Calculus of Mobile Processes, Part I and II. Information and Computation, 100, 1992, pp. 1 – 77

[7] Milner, R.: Communicating and Mobile Systems: the π-Calculus. Cambridge University Press, Cambridge, New York, Melbourne, 1999

[8] Fuggetta, A. – Picco, G. P. – Vigna, G.: Understanding Code Mobility. Software Engineering, 24(5), May 1998, pp. 342 – 361

[9] Cardeli, L. – Gordon, A. D.: Mobile ambients. Theoretical Computer Science 240, 2000, pp. 177 – 213

[10] Pierce, B. – Sangiorgi, D.: Typing and Subtyping for Mobile Processes. In Proceedings of LICS ’93, IEEE Press, 1993

[11] Kobayashi, N. – Pierce, B. – Turner, D.: Linearity and the π-Calculus. In Proceedings of POPL ’96, 1996

[12] Adabi, M. – Gordon, A. D.: A Calculus for Cryptographic Protocols: the Spi-Calculus. In Proceedings of the Fourth ACM Conference on Computer and Communications Security, ACM Press, April 1997, pp. 36 – 47

[13] Paralič, M.: Mobile Agents Based on Concurrent Constraint Programming. In: Lecture Notes in Computer Science, Vol. 1897, Zurich, 2000, pp. 62 – 75

[14] Tomášek, M.: Concepts for Mobile Agents Interaction. In M. Jelšina, J. Kollar (eds.): Proceedings of Scientific Conference with International Participation Computer Engeneering and Informatics ’99 (CEI ’99), Košice – Herľany, Slovakia, October 14 – 15, 1999, pp. 259 – 264

[15] Tomášek, M.: Distributed System Based on Technology of Mobile Agents. Acta Electrotechnica et Informatica, Vol. 1, No. 1, Košice, 2001, pp. 55 – 60

BIOGRAPHY

Martin Tomášek was born 1975 in Košice, Slovakia. He received the master degree in computer science in 1998 at the Faculty of Electrical Engineering and Informatics of the Technical University of Košice, Slovakia. He is presently a assistant professor at the Department of Computers and Informatics of the Faculty of Electrical Engineering and Informatics of the Technical University of Košice, Slovakia and his study field is software and information systems. The subject of his PhD dissertation focuses on formal expressing dynamics of mobile programs. He authored several publications on mobile agents, distributed systems and he participates in research and educational projects of the university. His research interests include mobile code, software agents, distributed systems and formal description of communication and code mobility.


LOGICKÉ ŘÍZENÍ

(LOGICAL CONTROL)

Josef BOKR Department of Information and Computer Science, Faculty of Applied Sciences, University of West Bohemia,

Univerzitní 8, 306 14 Plzeň, Czech Republic, E-mail: [email protected]

SUMMARY The paper deals with logic control technological devices. An impossibility of Glushkov′s and Bellman′s concepcions of

logic control is shown and thus the new idea is proposed. Keywords: system of automatical logic control, finite automaton, logical transmitter 1. ÚVOD

Předmětem oboru, obvykle zvaného „Logické systémy“, jsou jednak logické převodníky (viz dodatek) a jednak systémy automatického logického řízení (SALŘ). A stává se i, že SALŘ jsou pouhou záminkou (úvodem) ke studiu logických převodníků; např. [1,2].

Pojetí, nyní již klasické, logického řízení podle Gluškova (1965), který navrhl SALŘ jako zpětnovazební kompozici dynamických logických převodníků: deterministického řídicího (ŘP) a operačního, přičemž operační převodník odpovídá nedeterministickému řízenému technologickému převodníku (TP), se pokládá za korektní. Vedle zmíněné koncepce existuje pojetí SALŘ podle Bellmana (1956) − nezávisle zavedené v [3] − kde ŘP je statický a TP je dynamický převodník. Bellmanova koncepce se pokládá za fundamentální ideu řízení vůbec.

Snadno lze ukázat, že Bellmanův ŘP je mi-nimální, co do počtu stavů, formou Gluškovova ŘP Mealyho typu, což bezprostředně vyplývá z konceptu logického řízení.

Obě pojetí logického řízení, což je snad pře-kvapivé, nejsou s to se vyrovnat s následující situací: dospěje-li řízený TP po absolvování předepsané stavové trajektorie do, pro řízení, stabilního stavu, není možné, aby TP zmíněný stav opustil a pokračoval v pohybu po požadované trajektorii, neboť ŘP produkuje řízení podle stávajícího stavu TP, a nelze očekávat, že deterministický ŘP dokáže podle stabilního stavu produkovat různá řízení tak, že jedno udržuje TP ve stabilním stavu a druhé převádí TP do stavu jiného (následovníku) stavové trajektorie.

Příčina uvedené neutěšené situace je nasnadě; vždyť, co „nutí“ projektanty SALŘ pokládat řízené technologické zařízení za dynamický převodník? Zřejmě rutina; návrháři SALŘ totiž jednak projektují ŘP stejně jako strukturní modely dynamických logických převodníků (zpětnovazební logické obvody) kánonickou dekompozicí [4], a jednak se nevědomky při identifikaci technologické aparatury ztotožňují s jejím ŘP a pokládají pak TP za

dynamický. Avšak řídit dynamický převodník nemá smysl (!), je přece dynamický. 2. TRADIČNÍ LOGICKÉ ŘÍZENÍ

Uvažujme SALŘ podle obr. 1 a nechť je, bez újmy na obecnosti, modelem TP nedeterministický dynamický semiautomat

TP = 〈 U, S, δT 〉 a modelem deterministického Gluškovova dynamického ŘPG či Bellmanova statického ŘPB buď konečný automat

ŘPG = 〈 S, Q, U, δG , λG 〉 nebo, položíme-li Q = {qp},

ŘPB = 〈 S, U, λB 〉

kde U, S a Q je příslušně abeceda řízení, stavová TP a ŘPG , δT je přechodová relace

δT : S × U × S : 〈 s, u, s′ 〉 ,

δG a λG je odpovídající přechodová a výstupní funkce Gluškovova ŘPG

δG : Q × S → Q : 〈q, s〉 q′

λG : Q × S → U : 〈q, s〉 u a λB je výstupní funkce Bellmanova ŘPB

λB : S → U : s u

ŘP u TP s

Obr. 1 Systém automatického logického řízení (SALŘ)

Fig. 1 System of the automatical logic control (SALC)

62 Logické řízení

Konečnosemiautomatovým modelem SALŘ s ŘPG nebo s ŘPB je dyáda příslušně

SALŘG = 〈 S × Q , ∆G 〉 či

SALŘB = 〈 S , ∆B 〉

kde ∆G či ∆B je odpovídající přechodová relace

∆G : (S × Q) × (S × Q) : 〈s, q, proj3 δT (s, λG (q, s), s′), δG (q, s)〉

nebo

∆B : S × S : 〈s, proj3 δT (s, λB (s), s′)〉

viz obr. 2, kde × symbolizují stavy ŘPG .

Nyní je již zřejmé, že vystačíme s Bellmanovým ŘPB.

Vyšetřeme fragment stavové trajektorie TP v SALŘG nebo v SALŘB (obr. 3). Protože na TP v SALŘ je sik = proj3 δ T (si,k-1, ui, sik), sik = proj3 δT (si,k , ui, sik) a sj = proj3δT (si,k , uj , sj ), pro ui ≠ uj , pak také na ŘPG je λG (qik , sik) ∈ {ui , uj} a právě tak na ŘPB je λB (sik) ∈ {ui , uj}, což ovšem znamená zařadit do SALŘ nedeterministický ŘP. δT : s1′ u q′ q s q′ u s2′ q′ u sk′ δG : s/u

q q′ . δB : s/s

. qp

Obr. 2 Součinnost TP a ŘPG či ŘPB

Fig. 2 Cooperation of a technological transmitter (TP) and the Glushkov (ŘPG) or the Bellman control

(ŘPB) transmitters

ui qik ui ui uj uj si,k-1 si,k sj

qi,k-2 qi,k-1 qi k qj

Obr. 3 Stavová trajektorie na TP Fig. 3 A state trajectory at the TP

3. LOGICKÉ ŘÍZENÍ

Ukážeme, že nedeterminismus ŘP jak podle Gluškova, tak podle Bellmana způsobuje „převodníkové“ chápání TP; byl-li by totiž TP dynamický převodník, pak nemá smysl TP řídit, neboť se zjevně řídí sám (viz dodatek).

Ilustrujme proto na příkladu samopalu AK libovolného vzoru („kalašnikovi“), přitažlivém zejména pro vojáky bývalé československé i české či slovenské armády, že TP není dynamický, ale jen potenciálně dynamický převodník. Nechť je náboj v nábojové komoře, tj. uzamčený závěr se nachází v přední úvrati pouzdra závěru. Po spuštění spouště uvolní spoušťadlo úderník, úderník aktivuje roznětku a dojde k výstřelu. Dříve než střela opustí hlaveň samopalu, odvede se malá část spalin z hlavně plynovým kanálem do válce vratného pístu. Vratný píst uvolní závorník závěru a přesune závěr unášející vytahovačem nábojnici do zadní úvratě v pouzdru závěru. Po uvolnění vytahovače vyhodí vyhazovač nábojnici a podavač přemístí další náboj do komory závěru. Stlačená vratná pružina převede závěr do přední úvratě atd. potud, pokud je spuštěná spoušť. Je-li spoušť uvolněná, zablokuje se úderník a k opětovnému výstřelu dojde až po spuštění spouště (obr. 4). Použijeme-li cvičné střelivo, není tlak spalin v hlavni dostatečný k vrácení závěru do zadní úvratě (nepoužijeme-li ovšem úsťový nástavec) a pro opakování střelby by pak bylo nezbytné přesunovat závěr do zadní úvrati ručně; střelba dávkami je tak vyloučena.

Činný samopal (dynamický logický převodník) je tak agregací statického logického převodníku (vratného pístu s vratnou pružinou) a potenciálně dynamické řízené pušky [4]. Pokud si čtenář nedokáže představit vratný píst spolu s vratnou pružinou jako statický převodník, ať oba chápe jako „pohybovody“. A ještě, střelec spuštěním spouště − nesilový podnět − iniciuje posloupně stavové přechody v samopalu: ze stavu výstřel do stavu nabití a naopak. Stav výstřel vykonává silou zpětného pístu přechod do stavu nabití, a stav nabití vykonává silou vratné pružiny přechod do stavu výstřel. Metaforicky: vratný píst s vratnou pružinou „oživuje“ „mrtvou“ pušku.

Nabízí se tedy dynamický logický převodník interpretovaný jako SALŘ ve tvaru agregace statického logického převodníku Ř pojatého jako řídicí a potenciálně dynamického řízeného objektu T (obr. 3), kde zadané řízení u vybírá subjektem požadovanou stavovou trajektorii v T a stav s, resp. buzení e, je vykonavatelem jednotlivých přechodů stavové trajektorie.

Konečnoautomatovým modelem statického deterministického Ř je uspořádaná trojice

Ř = 〈 U × S , E , λ 〉

kde U, S, E je příslušně abeceda řízení, stavů, buzení a λ je výstupní funkce

λ : U × S→ E : 〈u, s〉 e


Obr. 4 Schéma samopalu AK – 47 Fig.4 Schema of the subgun AK - 47

a pseudokonečnoautomatovým modelem nedetermi-nistického T (konečný automat je totiž modelem dynamických převodníků) je uspořádaná triáda

T = 〈 E , S , δT 〉

kde δT je přechodová relace

δT : S × E × S : 〈 s, e, s′〉.

Konečný automat

〈U , S , δ 〉

kde δ je přechodová relace

δ : S × U × S : 〈s, u, s′〉

je modelem nedeterministického SALŘ z obr. 5, přičemž

( ) ( )3 3 , , , , Tproj s u s proj s e sδ δ′ ′= . z u Ř e T s Obr. 5 SALŘ jako agregát řídicího převodníku Ř a

technologického „nepřevodníku“ T Fig. 5 SALC as an aggregate of a control trans-mitter Ř and a technological „not transmitter“ T

Zmíněné pojetí logického řízení se snadno vyrovná s nedeterminismem ŘP z části 2. Skutečně, protože řízení ui a uj (ui ≠ uj) zadává subjekt, potom je nasnadě modifikovat fragment stavové trajektorie z obr. 2 − obr. 5 − a psát:

( ) (3 , 1 3 i,k- , , s ,ik i k i ik Ts proj s u s projδ δ−= = 1)

) (, 1 3, , , , ,i i k ik ik ik i iku s s s proj s u sδ− = = ( ) ( )3 , 3 i,k , , s , ,i k i ik T i ik ikproj s u s proj u s sδ δ= = a

( ) (1 3 1 3 , , s ,j j j js proj s u s projδ δ+ += = jT )1,j j ju s s + .

ui sik ui si ,k-1 uj sj si , k-1 sik sj

Obr. 6 Modifikovaná stavová trajektorie

na I z obr. 3 Fig. 6 A modified state trajectory at I from Fig. 3

Lze ovšem namítnout, že statický řídicí pře-

vodník Ř neuspěje, budou-li na T instalovaná impulsová čidla snímající stavy T; postačí však impulsová čidla vybavit podpůrnými paměťovými moduly [3]. 4. PŘÍKLAD

Až dosud jsme předpokládali, že nedeter-minismus T je dílem neměřitelného a tedy implicitního nahodilého rušení T. Jsou-li však poruchy z působící na T měřitelné a tedy explicitní, tj. disponujeme-li poruchovou abecedou Z deterministického T, je pseudokonečnoautomatovým modelem T uspořádaná triáda

Tdet = 〈 E × Z , S , δTdet 〉

kde δT det je přechodová funkce, bez újmy na obecnosti,

δT det : S × E × Z → S : 〈s, e, z〉 s′.

plynový kanálek vratný píst závěr pouzdro závěru

hlaveň

náboj

spušťadlo

zásobník

64 Logické řízení

Konečný automat

〈 U × Z , S , δ 〉

kde δ je přechodová funkce

δ : S × U × Z → S : 〈s, u, z〉 s′ ,

je modelem deterministického SALŘ, přičemž

δ (s, u, z) = δT (s, e, z). i+1 W I P S2 F S1 i D V M K i-1 . ν1 ν2

Obr. 7 Odstředivé lití Fig. 7 Centrifugal casting

a) s1′ s2′ s3′ ez s1s2s3

stop

start

1 2 S S start

1 2S S start

1 2 S S start

1 2S S

IKD IVD IVD WVD WVD WVD WMD WMD IMD IMD IK D

IK D IK D IK D b)

u s1s2s3

zastavení IK D

spuštění IKD, IVD, WVD,

WVD, WMD, IMD

spuštěníIK D

e stop start stop

Tab. 1 a) Přechodová tabulka řízeného odstředivého lití, b) výstupní tabulka řídicí

automatiky Tab. 1 a) Transition table of the controled

centrifugal casting, b) response table of the control automatics

Navrhněte proto statickou řídicí automatiku pro

zařízení na odstředivé odlévání litinových trubek (obr. 7). Dosáhne-li nosič (D) tavicí pánve (P) polohy i, přemístí se vozík s formou (F) z polohy V rychlostí v2 do polohy K. Tavicí pánev se naklání z pozice I do pozice W dotud, dokud nezačne litina vytékat z pánve (S2). Když tok litiny dosáhne pravý okraj formy (S1), vozík se pohybuje pomalu zleva doprava rychlostí v1 (v1


připouštějí ve stavové trajektorii v TP výskyt kromě koncového i jiných stabilních stavů, což vede k nedeterministickému ŘP

Východisko ze vzniklé situace je nasnadě. Řízením u se provádí pouze výběr (iniciace) stavové trajektorie na nikoliv dynamickém TP, ale potenciálně dynamickém T a každý výchozí stav přechodu s stavové trajektorie je nezbytné chápat jako jeho vykonavatele. Snad proto, že uvedená koncepce logického řízení je očividná, může čtenáři připadat autorův podíl na fyzikálně zdůvodněném pojetí logického řízení zanedbatelný. LITERATURA [1] Lazarev,V. G. - Pijl, E. I.: Sintez

upravljajuščich avtomatov. Energoatomizdat, Moskva 1989

[2] Šalyto, A. A.: Logičeskoe upravlenie (metody apparatnoj i programmnoj realizacii algoritmov). Nauka, Sankt Peterburg 2000

[3] Bokr, J. et al: Logické řízení technologických procesů. SNTL, Praha 1986

[4] Bokr, J. - Jáneš, V.: Logické systémy. Vyd. ČVUT, Praha 1999

[5] Bokr, J.: Kánonická dekompozice. Acta Electrotechnica et Informatica, No. 1, vol. 4 2004, pp. 60 – 65

[6] Bokr, J.: Upravlenie logičeskim objektom i kanoničeskaja dekompozicija. Avtomatika i vyčislitelnaja technika, N. 1, 2000, pp. 12 – 23

[7] Bokr, J. et al: Logické řízení technologických objektů. Sešity Inorga, N 148 – 149, 1988

[8] Kalman, R. E. – Falb, P. L. – Arbib, M. A.: Topics in Mathematical System Theory. Mc Graw-Hill Book Co., New York, Sydney 1969

6. DODATEK

Logickým převodníkem rozumíme logické zařízení převádějící zadané vstupní slovo na požadované slovo výstupní. Nechť je stacionární deterministický dynamický převodník dán konečnoautomatovým modelem

〈 X , S , Y , δ , λ , sp 〉

kde X, S, Y je příslušně vstupní, stavová, výstupní abeceda, δ : S × X → S : 〈s, x〉 s′ je přechodová a λ : S × [X] → Y : 〈s, [x]〉 y je výstupní funkce s tím, že [ ] { }X X= Λ a [x] = xΛ (Λ je prázdné slovo) a sp je počáteční stav. Chápeme-li δ (s, x) = s′ jako model změny stavu z s na s′, je jedinou příčinou změny podnět x; avšak změna stavů znamená realizaci přechodu z s do s′, tj. jak je obvyklé, je nezbytné přechod spustit podnětem x a také vykonat stavem s. Chováním („převodovou“ funkcí) B dynamického převodníku rozumíme funkci (m ≥ 1) B : {sp} × Xm → Ym : 〈sp , xi 1 xi 2 ... xi m〉 yj1 yj 2 ... yj m = ( ) ( )( )1 1 , , , p i p i i2s x s xλ λ δ⎡ ⎤⎡ ⎤⎣ ⎦⎣ ⎦ x ...

( )( )( )( )1 2 , 1 ... , , ,... , ,p i i i m ims x x x xλ δ δ δ −⎛ ⎞⎜ ⎟⎝ ⎠ a může být jak kombinační, tak i sekvenční. Pouze, je-li převodník Mealyho typu (λ : S × × X →Y : 〈s, x〉 y) a chová-li se kombinačně, lze položit S = {sp}, tj. minimalizovat Mealyho kombinační převodník co do počtu stavů, a formálně (nikoliv fakticky) ignorovat přechodovou funkci δ, neboť δ :{sp}×X → {sp}: 〈sp ,x〉 sp modeluje virtuální stavový přechod, a modifikovat funkci výstupní λ :{sp}× X →Y : 〈sp , x〉 y tak, že λ : X →Y : x y. Mealyho minimální kombinační převodník je tedy statickým převodníkem a jeho konečnoautomatový model má tvar 〈X , Y , λ , sp〉 či 〈X , Y , λ〉 , přičemž pro jeho chování B platí

{ } p 1 2 : : s , .. m mp iB s X Y x x x⎡ ⎤ ⎡ ⎤× → ⎣ ⎦⎣ ⎦ .i im ( ) ( ) ( )1 2 1 2 ... ... j j jm i i imy y y x x xλ λ λ= .

V dynamickém logickém převodníku jsou zajímavé zejména stavové trajektorie:

− ( )( )( )( )... , , , ... , xps x xδ δ δ a ( )( )( )( )... , , , ... , xps x xδ δ δ = = ( )( )( )( ) ... , , , ... , x , xps x xδ δ δ δ⎛ ⎞⎜ ⎟⎝ ⎠ , tj. posloupně „spontánní“ stavové přechody ze stavu sp do stabilního pro podnět x stavu ,

− ( )( )( )( ) ... , , , ... , x , xps x xδ δ δ δ⎛ ⎞⎜ ⎟⎝ ⎠ = = sp a ( ) ( )( ), , , , , , ...,p p ps s x s x xδ δ δ ( )( )( )( )... , , , ... , xps x xδ δ δ jsou navzájem různé, tj. „spontánní“ stavové cykly. Je „podivné“ hovořit o samovolných posloupných stavových přechodech či cyklech v deterministickém převodníku; postačí si totiž uvědomit, že výchozí stav

( ) ( )( ), , , , , , ...,p p ps s x s x xδ δ δ ( )( )( )( )... , , , ... , xps x xδ δ δ

přechodu je vykonavatelem příslušného přechodu, a podnět x přechod pouze iniciuje.

Nyní je již také zřejmé, že nelze ztotožnit jednak statický a kombinační, a jednak dynamický a sekvenční logický převodník. BIOGRAPHY Josef Bokr was born on 1940. In 1965 he graduated with honor at Moscow Power Institute with specialisation in mathematical computing devices and apparatus. He received Ph.D (CSc) degree with a thesis Logic Control in 1990 and was done an associate professor. His scientific research is focused on logic system and automata theory.


Pokyny pre autorov príspevkov do časopisu Acta Electrotechnica et Informatica

Na publikovanie do časopisu Acta Ele

wind turbines in the high voltage networks of czech … · (tk), fitted with the transformer...

Documents