TRANSACTIONS ON ELECTRICAL ENGINEERING

CONTENTS Weber, A. R., Weissbacher, J., Steiner, G., Horn, M.: An Accurate Auto-Tuning Procedure for Encoderless AC Motor Drives in Industrial Environments . . . . . . . . . . . . . . . . . . . . . .

1– 7

Mlynařík, L., Doleček, R.: The Effect of Fault States of the Twelve-pulse Rectifier During the Recuperation . . . . . . . .

8 – 12

Bachorec, T., Král, P.: Simulation of Magnetodynamic Forces Acting on the Eccentric Rotor of the Generator . . . . . . . . .

13 – 17

Benes, I., Valasek, J.: Application of Resilient Power Pilot Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

18 – 22

Miksiewicz, R.: Field-Circuit Modelling of Self-Excited Induction Generators . . . . . . . . . . . . . . . . . . . . . . . .

23 – 27

Vol. 3 (2014)

No.

1

ERGO NOMEN

pp.

1 - 27

TRANSACTIONS ON ELECTRICAL ENGINEERING Publisher:

ERGO NOMEN, o.p.s., K13114 FEE CTU in Prague, Technicka 1902/2, 166 27 Praha 6, Czech Republic E-mail: [email protected]

Editorial Office:

PIVONKA Pavel BAUER Jan HAVLICEK Radek MERICKA Jiri NOVA Ivana VONDRICH Jiri ZDENEK Jiri

Periodicity: Language: Scope: On-line version:

Quarterly English International scientific journal of electrical engineering www.transoneleng.org ISSN 1805-3386

Each paper in the journal is evaluated by two reviewers under the supervision of the International Editorial Board. International Editorial Board Editor in Chief: Prof. LETTL Jiri, Czech Technical University in Prague, Czech Republic Members: Prof. BAUER Palo, Delft University of Technology, Netherlands Prof. BRANDSTETTER Pavel, VSB-Technical University of Ostrava, Czech Republic Prof. DOLEZEL Ivo, The Academy of Sciences of the Czech Republic, Czech Republic Prof. DUDRIK Jaroslav, Technical University of Kosice, Slovakia Prof. NAGY Istvan, Budapest University of Technology, Hungary Prof. NOVAK Jaroslav, University of Pardubice, Czech Republic Prof. ORLOWSKA-KOWALSKA Teresa, Wroclaw University of Technology, Poland Prof. PEROUTKA Zdenek, University of West Bohemia, Czech Republic Prof. PONICK Bernd, Leibniz University of Hannover, Germany Prof. RICHTER Ales, Technical University of Liberec, Czech Republic Prof. RYVKIN Sergey, Russian Academy of Sciences, Russia Prof. SKALICKY Jiri, Brno University of Technology, Czech Republic Prof. VITTEK Jan, University of Zilina, Slovakia Prof. WEISS Helmut, University of Leoben, Austria

Responsibility for the contents of all the published papers and technical notes is upon the authors. Template in MS WORD and basic typographic rules to be followed see www.transoneleng.org.

Copyright:

©2014 ERGO NOMEN, o.p.s. All right reserved.

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

1

An Accurate Auto-Tuning Procedure for Encoderless AC Motor Drives in Industrial Environments Andreas R. Weber 1), Joachim Weissbacher 2), Gerald Steiner 1), Martin Horn 3) 1)

Institute of Electrical Measurement and Measurement Signal Processing, Graz University of Technology, Austria, e-mail: [email protected], [email protected] 2)

3)

Bernecker + Rainer Industrie Elektronik Ges.m.b.H, Eggelsberg, Austria, e-mail: [email protected]

Institute of Smart System-Technologies, Control & Mechatronic Systems Group, Klagenfurt University, Austria, e-mail: [email protected]

Abstract — Modern ac motor drives are based on field oriented vector control with feedback and feedforward control units. The feedback control unit for position and speed consists of two cascaded standard PI-controllers. These controllers require information of the position and the derivative of the position, respectively. Typically a shaft encoder provides this information but more often a position observer is used instead. The feedforward control unit uses the set position trajectory in combination with mechanical parameters and essentially imposes the overall dynamics. This work presents a new method for self-commissioning of speed controller, position controller and feedforward control unit for drives without a shaft encoder. Performance and feasibility of the proposed method are demonstrated by experimental results. Keywords — self-commissioning, feedforward control, encoderless, sensorless, flux observer, field oriented vector control, ac motor drives, frequency response measurement, maximum peak criteria

I. INTRODUCTION Modern drive technology is often based on field oriented control structures. A typical controller structure of a field oriented vector controlled drive is shown in Fig. 1. The controller structure is composed of three cascaded controller loops. The innermost is the current controller loop which is composed of a standard PI-controller with electromagnetic force (emf) feedforward compensation. The controller parameters are adjusted automatically by the knowledge of the electrical motor parameters like resistances and inductances. Set value for the current controller (q-direction) is calculated by the superimposed speed controller whose actual value is given by the time derivative of the position signal. In order to suppress noise due to the quantization of the position signal a low pass filter is used in speed feedback loop. The corresponding set value of the speed controller is given by the position controller. Both, position and speed

controller are standard PI-controllers which are only used as proportional elements. The information of the known set position trajectory can be used to increase the performance of the command response. Based on identified mechanical parameters additive reference values for speed and quadrature current are calculated in the feedforward controller unit. The described field oriented control structure assumes exact information about the flux position. Typically a shaft encoder is used to provide the servo drive with this necessary position information. For lower investment costs, lower maintenance costs and increased reliability, the shaft encoder, the necessary cables and evaluation unit are saved and replaced by an observer. The usage of such an observer has an impact on the plant and therefore on the dynamics of the whole system. Since the commissioning of the current controller and the observer is assumed to be already completed, the parameters of the speed controller, position controller and feedforward controller unit have to be determined. Goal of this work is to present a method for self-commissioning of these controllers and the feedforward controller unit which is characterized by simplicity for the user in terms of tuning parameters. In contrast to [1] where the mechanical parameters of two mass systems are identified (parametric model) in order to design a controller, in this work a nonparametric model is the basis for the design of the controllers. The parameters of the feedforward controller unit are identified by way of a point to point movement and an offline calculation. The work is organized as follows: In Section II the basics of the used observer is explained and compared. In Section III the identification and determination of feedback and feedforward controller parameters is shown. In Section IV experimental results are presented and Section V concludes the article.

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

2

Figure 1: Typical field oriented controller structure with feedforward controller unit

II. ENCODERLESS CONTROL A. Postion observer based on a flux estimator for a permanent magnet synchronous motor (PMSM) The used position observer is a common flux observer in the stator fixed coordinate system ( ). Different methods of flux observation have been published, and some examples are reported in [2], [3], [4], [5]. Generally, the observer is estimated by the fundamental voltage equations.

(1) (2)

(5) The problem with the open integration of the observer structure is avoided with a low pass filter and a feedback of the estimated flux error vector multiplied with a gain factor K. The error vector is the difference between the observed flux vector and a calculated reference flux vector. So unwanted influences based on current measurement errors, inadequateness in the voltage generation (e.g. nonlinear inverter voltage drop) and uncertain motor parameters are strongly reduced. The basic structure of the observer is shown in Fig. 2 as signal flow chart. The performance of the described position observer is shown in [6] and [7].

Transformed into the flux vector (3)

(4) where is the stator resistance, the stator inductance, the measured current vector and the voltage vector. With the stator fixed components and the position of the flux vector and commutation angle can be calculated via the arcus tangens function.

Figure 2: Signal flow chart of the flux vector estimator

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

3

AC Motors with position or speed observers, which are based on the fundamental equations, are not applicable at standstill or very low speed. Due to this fact operation at certain speed is necessary for identification of the overall system consisting of mechanics and observer [1], [8]. Since drive controllers are not parameterized yet, a provisory controller parameterization is necessary.

lowpass as a free design parameter. It should be mentioned that should be chosen in this case much greater than the sum of the small time constants (current controller loop) and dead times of the servo drive. Experiments on different mechanical setups showed that is a reasonable start value for sufficient suppression of any existing resonances and high frequency noise.

A. Provisory controller parameterization

B. Identification procedure

In [9] a self-tuning speed control concept for drives with encoder is presented where a provisory PI speed controller with low bandwidth is used for performing experiments for estimation of the frequency response function (FRF). In contrast to [9] a proportional speed and position controller parameterized for a nominal value that equals to the motor inertia (known in the majority of cases out of datasheets) is used instead of the published PI speed controller based on the total inertia. The advantage of the missing integral part in the speed controller is that in case of a strong deviation of the actual inertia from its nominal value no unstable speed loop behavior can occur. Additionally to the given cascaded controller structure of the used servo drive there is a low pass (speed filter) in the speed feedback loop in order to smooth the quantization effects caused through the speed calculation

With the controller parameters of the previous chapter and a smooth position trajectory with a sufficiently small acceleration the drive can be operated at desired constant speed. The closed loop system is excited with a pseudo random binary signal (PRBS) as a disturbance signal in the torque generating component of the stator current [1], [8]. A PRBS of order can be generated by

III. AUTOTUNING

(6) with a filter time constant and as sample time. The overall identification structure can be seen in Fig. 3.

(9) with appropriate coefficients , , where is also the number of used shift registers. Useful coefficients for different orders can be found in discrete time [11]. The PRBS is periodic with steps. With a cycle time the necessary measurement time for one period is given by

(10) As depicted in Fig. 3, the current (Input) and the differentiated position signal (Output) of the observer are used for calculation of the frequency response function. Therefore a discrete fourier transformation (DFT) of both input and output sequence sampled at with sample time is performed. Element-wise division of the complex coefficients leads to the desired frequency response function

(11)

Figure 3: Identification structure

By use of the "Magnitude Optimum" [10] both speed and position controller gains can be calculated as

(7) (8) with as the motor torque constant, the known motor inertia and the filter time constant of the

at frequencies . If exactly one period ( samples) is used after transient behavior has decayed no windowing technique and advanced signal processing is necessary to attenuate the leakage-effect. By use of an efficient fast fourier transformation (FFT) algorithm for calculation of DFT the estimation of the FRF can also be implemented on an embedded system with limited memory and calculation power. In Fig. 4 a FRF for a typical one mass system is depicted. As a reference the calculated FRF of the system with the installed encoder is shown. It can clearly be seen that at low frequencies both characteristics are very similar but at higher frequencies ( ) a differentiating behavior of the observer is obvious. The problem with increasing noise at higher frequencies can be avoided with the low pass filter (6) in feedback of the speed loop.

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

4

Figure 4: Comparison of measured FRF (one mass system) with observer and encoder and a one mass model approximation for low frequencies

C. Speed controller tuning Equation (11) can be interpreted as a nonparametric model of the plant for which the speed controller is designed in the following. Usually a parametric model is derived from the measurement data and used for further calculations. This parameter identification is a challenging task in view of an automatic procedure. Therefore a nonparametric model for numerically calculating the controller parameters is chosen. The open loop FRF of the speed controller is a series connection of the speed filter (6) evaluated at , the plant (11) and the controller gain .

(12) The Maximum Peak Criteria [12]

(13) is one possible way to formulate the design specifications in frequency domain in terms of gain margin (GM) and phase margin (PM) for the closed loop FRF Figure 5: Proposed tuning algorithm for position and speed controller

(14) For a given upper bound a minimum gain and phase margin of the open loop system is guaranteed. In this work a value of resulting in and is chosen.

The goal of the proposed tuning algorithm is to find a maximum value of the gain factor subject to the speed filter time constant and condition (13). In Fig. 5 (upper part) the main steps of the algorithm are presented graphically.

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

5

D. Position controller tuning The identification of the position controller parameter is based on the calculated FRF of the closed speed loop and an integrator from actual speed to actual position. Similar to Sec. III-C a gain factor of the position controller can be calculated for the open position loop FRF

(15) as depticted in Fig. 5 (lower part). E. Feedforward controller tuning Since the positioning profile is usually known in advance, this fact can be used to improve the tracking behavior of the drive by means of speed and torque feedforward (feedforward controller unit). Whereas the speed reference is calculated by simply differentiating the position the torque reference is based on a simple one mass system with static and dynamic friction and can be calculated as:

(19) In contrast to [13] where the identification of the feedforward parameters is carried out online, here the optimization is done offline in the rest time task. This has the advantage of less computation effort in the cyclic task. IV. APPLICATION RESULTS The used motor is a three phase permanent magnet synchronous motor (PMSM) from B&R connected with an extra flywheel ( ). The motor is fed by a B&R ACOPOS 8V1090.00-2 servo drive. Motor and servo drive parameters are described in the Appendix. The identified FRF of the speed controller plant is shown in Fig. 6

(17) (16)

It should be mentioned that the calculated feedforward torque has to be transformed into an equivalent current by

(17) with the motor torque constant (see Fig. 1). To identify the parameters and movements with acceleration, constant speed and deceleration phases (speed trapezoidal) are necessary. Since the position and speed controller are already parameterized only the characteristics of the trajectory (distance, speed, acceleration) have to be determined. Usually the maximum distance is given by the hardware limits of the mechanics and the maximum speed by the nominal speed of the motor. The acceleration can be ) which is calculated via an estimated inertia ( based on a one mass approximation of the measured FRF (see Fig. 4) for small :

(18) The set speed, set acceleration and resulting current signal for this excitation profile are used for identifying by solving the the parameters optimization problem:

Figure 6: Comparison of measured FRF of the two mass system with observer and encoder

and represents a typical two mass system with a resonance frequency at , an anti-resonance frequency at and an approximated mass moment of inertia of . The parameters identified with the presented method for the speed an position controller are given in Tab. I.

Table I: Controller Parameters

In Fig. 7 the FRF of the plant with observer, open and closed speed loop is shown. Fig. 8 represents the FRF of the position controller plant (closed speed loop in series with an integrator) as well as the open and closed position loop FRF. In the upper part of Fig. 9, the speed profile for the feedforward parameter identification is shown. In the

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

6

identified parameter set is listed in Tab. II. By means of a typical positioning profile the performance improvement with the identified parameters is demonstrated. Table II: Feedforward Parameters

Figure 7: Calculated FRF for speed controller tuning

The profile includes an acceleration phase from standstill, a constant speed phase and a deceleration phase to standstill. Note that the used observer can not be used in standstill, so an open loop current vector control is used for the region from standstill to . In this area the position controller error is forced to zero. Fig. 10 shows the reference position, reference speed and position error of a point to point movement with and without feedforward control. The acceleration and deceleration are set to a value that the necessary torque is half of the nominal torque of the motor. The occurred position error is described in units per revolution (1000 Units per revolution). The improvement with feedforward control is more than sufficient for a motor and servo combination without shaft encoder and can be useful for a lot of industrial applications.

Figure 8: Calculated FRF for position controller tuning

Figure 9: Speed profile (top) and measured vs. identified torque (bottom)

lower part the necessary measured torque for the given profile and the identified torque are depicted. The

Figure 10: Position profile (top), speed profile (middle) and position error for a point to point movement with and without feedforward control (bottom)

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

7

V. CONCLUSION This paper presents an acccurate method for parameterizing the position controller, speed controller and feedforward control unit in a field oriented controller structure without the position information of a shaft encoder. The need for an accurate position information and the possibility to observe this information via a flux observer are discussed. The structure of this position observer is described. A method to identify the FRF of the plant in combination with the position observer is shown. With the described maximum peak criteria the controller gains are determined. Furthermore a procedure for identifying the feedforward controller parameters is shown. Finally the performance and the feasibility of the proposed solution are demonstrated by means of a typical movement profile in industrial drive technology.

VI. APPENDIX MOTOR: PMSM, B&R 8LSA36.E0030D200-0, NUMBER OF POLE-PAIRS = 2, RATED VOLTAGE = 400V, RATED CURRENT = 1.9A, RATED SPEED = 3000MIN , RATED TORQUE = 2.7NM INVERTER: B&R ACOPOS 8V1090.00-2, RATED VOLTAGE = 3X400-480V, RATED CURRENT = 8.8A, WWW.BR-AUTOMATION.COM

REFERENCES [1] H. Zoubek and M. Pacas, “A method for speed-sensorless identification of two-mass-systems,” in Energy Conversion Congress and Exposition (ECCE), 2010 IEEE, Sept. 2010, pp. 4461 –4468. [2] S. Beineke, J. Schirmer, J. Lutz, H. Wertz, A. Baehr, and J. Kiel, “Implementation and applications of sensorless control for synchronous machines in industrial inverters,” in Sensorless Control for Electrical Drives (SLED), 2010 First Symposium on, July 2010, pp. 64 –71.

[3] L. Yongdong and Z. Hao, “Sensorless control of permanent magnet synchronous motor — a survey,” in Proc. IEEE Vehicle Power and Propulsion Conf. VPPC ’08, 2008, pp. 1–8. [4] X. Dan and S. Zhengqiang, “Sensorless control of surface permanent magnet synchronous motor using a structured adaptive flux observer,” in Power Electronics and Motion Control Conference, 2004. IPEMC 2004. The 4th International, vol. 2, Aug. 2004, pp. 1023 –1027 Vol.2. [5] Y. Imaeda, S. Doki, M. Hasegawa, K. Matsui, M. Tomita, and T. Ohnuma, “Pmsm position sensorless control with extended flux observer,” in IECON 2011 - 37th Annual Conference on IEEE Industrial Electronics Society, Nov. 2011, pp. 4721 –4726. [6] J. Weissbacher, A. Weber, G. Steiner, and M. Horn, “A simple method for self-commissioning of industrial ac motor drives without shaft encoder,” in MECHATRONIKA, 2012 15th International Symposium, 2012, pp. 1–6. [7] A. Weber and G. Steiner, “An accurate identification and compensation method for nonlinear inverter characteristics for ac motor drives,” in I2MTC 2012 Graz, 13-16 May 2012, pp. 821– 826. [8] H. Zoubek and M. Pacas, “An identification method for multimass-systems in speed sensorless operation,” in Industrial Electronics (ISIE), 2011 IEEE International Symposium on, June 2011, pp. 1895 –1900. [9] H. Wertz and F. Schutte, “Self-tuning speed control for servo drives with imperfect mechanical load,” in Industry Applications Conference, 2000. Conference Record of the 2000 IEEE, vol. 3, 2000, pp. 1497 –1504 vol.3. [10] D. Schroeder, Elektrische Antriebe Regelung von Antriebssystemen, ser. Elektrische Antriebe. Springer, 2001. [11] R. Isermann, Identifikation Dynamischer Systeme 1: Grundlegende Methoden. Springer Verlag, 1992. [12] S. Skogestad and I. Postlethwaite, Multivariable feedback control: analysis and design. Wiley, 1996. [13] F. Mink, A. Baehr, and S. Beineke, “Self-commissioning feedforward control for industrial servo drive,” in Advanced Electromechanical Motion Systems Electric Drives Joint Symposium, 2009. ELECTROMOTION 2009. 8th International Symposium on, 2009, pp. 1–6.

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

8

The Effect of Fault States of the Twelve-pulse Rectifier During the Recuperation Ladislav Mlynařík1) , Radovan Doleček2) 1, 2)

University of Pardubice, The Jan Perner Transport Faculty, Department of electrical and electronic engineering and signalling in transport (KEEZ), Czech Republic, e-mail: [email protected], [email protected]

Abstract — The paper deals with analysis of states from the viewpoint of the rectifier DC side . In the case of diode breakdown the interphase short-circuit occurs and the transformer is disconnected from the primary side by a high-speed switch. After disconnecting the DC current of a recuperative traction vehicle influences the circuit. The fault states of the twelve-pulse rectifier in a substation of 3 kV DC traction system used at railways in the Czech Republic are analyzed. The situation at the rectifier diode breakdown and at the same time recuperative vehicles in power supply section is examined. The vehicle with recuperation can increase strain of elements in power supply system under certain conditions during fault states of the rectifier. The simulation computer model was created for analysis of current waveforms in trolley line at this situation. All simulations were created by the program PSpice. The main goal is to analyze the endanger of power supply system elements and to present the recommendation for increasing the reliability of this mentioned power supply system.

system and it cannot be returned back through the traction substation to the main network [1-2]. A different situation can arise in the case of two diode breakdowns with opposite polarity in one of the bridges (six-pulse rectifier). In this situation of the catenary short-circuit and thus also the short-circuit of the vehicle with recuperation occurs. This short-circuit represents the consumption for this vehicle. Figure 1 shows an example in such conditions. There took place the diode breakdown (D5 and D2) creating the short-circuit source, which is represented in the circuit DC link by capacitor CV of the vehicle with recuperation (the flowing current is shown by red color). The current is reduced by the substation inductance LS, catenary resistance RC and catenary inductance LC.

Keywords — component; twelve-pulse rectifier; diode breakdown; recuperation; traction substation; transformer

I. INTRODUCTION The operation of electrified railway lines depends on the stability and reliability of the electric energy distribution at the traction power supply system. Every fault in this traction system leads to a reduction of electrical transport with following significant decreasing of railway line capacity. These faults can be represented by a damage of the catenary due to weather conditions or defective co-operation between the pantograph and contact wire and also faults in the traction substation. The switch off overcurrent or undervoltage protections of the 3 kV DC traction substation is easily soluble by the remote control and the electrified railway lines can be operated in short time again. A bigger problem can occur at the faults of the power rectifier units located in the traction substation where the number of these units can be limited due to the previous faults. This situation can bring the effect to decrease the railway line capacity or in the extreme case stopping of the railway transport. . II. VEHICLE WITH RECUPERATION The condition at the 3 kV DC traction system is that electric energy got by the recuperation must be consumed at the same time by other electric traction vehicle or vehicles in the feeding connected section of the traction

Fig.1: The recuperative current flow direction through the rectifier

The switch before the primary winding of the rectifier transformer in Fig. 1 disconnects rectifier unit from the AC side in the case of diode breakdown. The DC high-speed switches are placed in output of the substation and they separate each outlet of the catenary from the DC bus 3 kV. A rectifier unit can have at its output a few high-speed switches for several sections. The isolator switch enables to disconnect the rectifier from the DC bus placed between the DC bus and rectifier. III. EQUIVALENT CIRCUIT The equivalent circuit of the traction system has been derived for the necessary simulations. The simulations of the effect of the vehicle with recuperation on the power rectifier with the diode breakdowns and the traction substation transformer were researched for the reason of the current situation at the Czech Railways at the 3 kV DC traction system.

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

A. Equivalent circuit for recuperation The equivalent circuit for the recuperation allows the simulation of diode breakdowns in the rectifier and is depicted in Fig. 2. The middle part of the equivalent circuit represents vehicle with recuperation by capacitor Cl in the DC link . If the voltage of this capacitor overpasses the actual values of catenary voltage then the vehicle current flows to the catenary (recuperation). In the circuit this current with the value Prek/Uc [A] is supplied by the voltage dependent current source G. The Zener diode Dz has set the Zener voltage value of 3 600 V and it represents the ability of the vehicle to control the voltage at the pantograph (just at this value). The left part of the equivalent circuit represents the model of the traction substation at the fault. The first stage of the simulation of the left traction substation has the same function as the right traction substation in the circuit. The simulation of a short-circuit of the vehicle with recuperation at the diode breakdown is done by the circuit switch S1+S2 [3-6]. The meaning of other elements and symbols in Fig. 2 is: Prek is recuperation power coming to the catenary, Dh, Vh and Rh represent one of the substations which works against the substation represented by Ls2+D2h+Vh2 (twoway feeding). The resistor Rh (its value approximates infinity) is only for simulation, the inductances Ls a Ls2 are air chokes with 4 mH added at the output of each substation from the reason of increase of the speed of short-circuit current rise. The Lv2 + Rv2 simulate the consumption of the recuperative power and represent the vehicle with fixed speed. The S3 switches on the recuperation by current source G. The S4 and Ri are only auxiliaries elements for simulation. The Ll and Cl represent input filter of a modern recuperative vehicle.

9

transformer at no-load mode were found out by experimental measurements. The short-circuit current does not pass through the secondary winding of the transformer in Fig. 1. IV. ANALYSIS OF SITUATION AT FAULT RECTIFIER AT RECUPERATION

The short-circuit of the rectifier happens due to the diode breakdown as it was mentioned at the situation in Fig. 1. At the situation when the vehicle with recuperation is located in the feeding section of the catenary, this source will be short-circuited and its current increases the load of parts of the traction system. A. Estimation of current waveforms flowing back to traction substation The current waveforms which can flow back to the traction substation, were analyzed by simulation using the equivalent circuit from Fig. 2. The example of the simulations is the situation with two passing vehicles in half of the section between traction substations (20 km), Fig. 3. One vehicle is breaking (i.e. recuperation with power 6 MW) and the second vehicle is starting. At the first stage of the simulation the second vehicle is fed only by traction substations (each of the traction substation supplies the half current for the vehicle). The first vehicle starts to recuperate in time of 100 ms and the voltage of the catenary reaches the voltage value of 3600 V [7-8]. The recuperative current of the first vehicle covers all other consumptions and therefore the current of the traction substations is zero.

Fig.3: The situation with two passing vehicles

Fig.2: The equivalent circuit for simulation of faults at recuperation

B. Model transformer The three-phase transformer with two secondary windings and the nominal power of 5 000 VA was used for measurement of characteristics of the DC current flowing through the secondary windings of the transformer. The primary winding has the Y-connection, transformer ratio is 1:1 and the transformer phase angel corresponds to the real transformer with two secondary windings of the type RESIBLOC Yyn0d1. The same leakage inductances of two secondary windings of the transformer and also the same phase-to-phase voltages of output of both secondary windings are necessary conditions for transformer feeding the twelve-pulse power rectifier at the 3 kV DC traction system. The same phase-to-phase voltages and a little bit different inductances of both secondary windings of the

The diode breakdown (D2 and D5) in the rectifier of the first traction substation is simulated at the time of 200 ms. The short-circuit causes a voltage drop on the vehicle pantograph. Because this voltage is lower than the voltage of the second traction substation, this traction substation begins together with the vehicle with recuperation to produce a current for the second vehicle and also for the short-circuit caused by the diode breakdown (D2 and D5). This stage will last until the turn off of the high-speed circuit breaker in the traction substation in which the fault occurred. The high-speed circuit breaker in the neighboring traction substation also switches off by using of the control coupling of these high-speed circuit breakers. If the high-speed circuit breaker is set over the value of the steady state short-circuit current, the stage of switching off will last until switch off of the disconnector which is located between the damaged rectifier and DC bus. If the recuperative current flows through this disconnector during operating control it can be damaged.

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

10

• The maximal theoretical current flowing back to the traction substation 4.70 kA (green color) • The pantograph voltage of vehicle with recuperation 2.12 kV

Fig.5: An example of the short-circuit current passing through the transformer secondary winding at the diode breakdown and recuperation of the vehicles in the feeding section Fig.4: The theoretical current waveform flowing back to the traction substation

V. ANALYSIS OF THE TRANSFORMER INDUCTANCE In the case of recuperative current flowing through the transformer secondary windings it is obvious that the inductance between two phases will slow down the current rise to the traction substation with the faulty rectifier. The theoretical analysis of the relations between the inductances of the traction transformer each winding was done to verify the recuperative current waveforms flowing through the transformer secondary winding. The analysis of failure modes without detailed knowledge s of these transformer parameters is not possible to complete. Therefore the passing of this recuperative current was analyzed experimentally in the laboratory. The speed of the short-circuit current rise and the value of inductance between two phases at the real traction transformer during the fault of the rectifier and energy from the vehicle with recuperation at the rectifier fault is deduced from the results. The inductance of 5.27 H at the Y-connection winding and 5.79 H at the ∆connection winding were found out by measurements at no-load mode. The measurements of the current flow time response through the pairs of the transformer secondary windings to the voltage jump was done for verification of the values of this inductance obtained from laboratory measurements. During the measurements it was monitored the time (a period) when the DC current reached 63 % of its steady value. This time was 0.5 s. In the case of the first system order, it would be possible to evaluate the value of the inductance in the circuit on the basis of the L / R ratio. Similar measurements were done at the secondary winding at the ∆-connection and the obtained value was 0.7 s. From the time response measurements of the current flows through the pair of the transformer secondary windings to the voltage jump the value of the inductance of 0.60 H at the Y-connection and 0.91 H at the ∆connection was determined.

VI. THE ANALYSIS OF TRANSFORMER CURRENT IN THE CASE OF DC POWER SUPPLY Theoretically it is possible to describe this situation by the equivalent circuit in Fig. 6. This circuit describes the real transformer fed by DC voltage. The secondary circuit is created by the elements L2 and R2 representing the magnetic losses of the transformer. The difference of the inductance values found out by the AC measuring at noload mode and by the current response to the voltage jump comes out at the loaded transformer when the mutual magnetic coupling of the primary and the secondary circuit becomes evident.

Fig.6: The equivalent circuit for the theoretical derivation of the current time response to voltage jump U1

The derived equation for the current waveform was solved numerically by substitution of the values corresponding to parameters of the laboratory transformer. For the circuit in Fig. 6, the magnetic coupling simulation of the transformer primary and secondary windings (respectively the magnetic coupling between the primary winding and fictive secondary winding simulated the influence of eddy currents in the magnetic circuit) is valid

u1 = R1 i1 + L1 0 = R2 i2 + L2

di1 di +M 2 , dt dt

di2 di +M 1, dt dt

(1)

(2)

where

L2 = Lh + L2σ ,

(3)

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

After differentiation of equations 1 and 2 we get

0 = R1

di1 d 2i d 2i + L1 21 + M 22 , dt dt dt

(4)

0 = R2

di2 d 2i d 2i + L2 22 + M 21 . dt dt dt

(5)

The linear differential equation of the second order with constant coefficients has the form

 L R  di  L L M 2  d 2i1  u1 = R1 i1 +  L1 + 2 1  1 +  1 2 − . R2  dt  R2 R2  dt 2  (6) We get for the laboratory transformer numerical values of the characteristic equation solution 2. λ1 = 0.095 and λ2= -14.34. Supposing real different solutions of the characteristic equation we get a general solution of the differential equation in form

i1O = K1 e λ1 t + K 2 e λ2 t .

(7)

After determining of the coefficients K1 and K2 we obtain the equation for the current waveform

i1 (t ) = −0,149 e −0,095 t − 0,151e −14,34 t + 0,3.

(8)

The coefficients in the equation 8 were obtained by substituting of the actual transformer parameters to the equation 7. The current response time waveform to voltage jump is obtained by representation of the solved equation 8, Fig. 7, where i is the instantaneous value of the current.

Fig.7: The theoretical waveform of the current

The dependence of this time response on the value of the resistor R2 (in Fig. 6) is evident from the derived equations. This resistance represents at the real transformer no-load mode the effect of losses, especially eddy currents in the magnetic circuit of the transformer. The above mentioned calculations were further verified by measuring of the time response of the current to the voltage jump on the laboratory transformer and by the simulation of the equivalent circuit using the connection in Fig. 6. The main inductance between two phases is not indicative exactly for the speed of the DC short-circuits current rise in the transformer secondary winding as it was proved by previous calculations. They do not contain

11

the losses in the magnetic circuit of the transformer. The speed of the rise of the DC short-circuit current will be higher in the case of the recuperation. The waveform of the short-circuit current rise at the time of the recuperation will rather correspond to the curve in Fig. 7, because the real transformer is loaded by the losses in the magnetic circuit (cos ϕ at no-load mode is 0.73). At the first stage it will be a very quick process practically limited only by the transformer leakage inductance and the recuperating current will increase the current strain of the transformer in the traction substation. VII. CONCLUSION In the first part of this paper, the values of short-circuit currents flowing through 3 kV DC traction system were numerically simulated. Independently the situation, when the short-circuit DC current flowing through the transformer secondary winding was analyzed. The inductances of this transformer slow down the rise of the DC short-circuit current. However, in the first stage of this DC short-circuit current rise, the transformer leakage inductances which allow a rapid rise of this current, become evident. The whole interval of the short-circuit current rise at the real transformer is further reduced by the effect of oversaturation of the magnetic circuit at connection to the DC source. It is possible to conclude by the circuit theory that the worst situation is in the case of vehicle with recuperation very close to the traction substation when the short-circuit current is not limited by the traction line resistors. The simulations show, the voltage drop on the pantograph below 2 kV occurs by this close short-circuit. The vehicle has a good opportunity to recognize this nonstandard situation and stop the recuperation. Potentially the most hazard situations can be considered, when a vehicle with recuperation is close enough to pass through the catenary the high current, but far enough that the voltage on the pantograph does not fall below 2 kV. If the DC source in the contact line passes through a higher current to the traction substation than the set of the highspeed circuit breaker, this circuit breaker will be turned off. The current values by the effect of the inductances in the circuit reach their steady maximum. This situation is dangerous especially for the machine isolator (circuit breaker) located between the 3 kV DC bus and damaged rectifier. The circuit breaker starts to disconnect immediately after the switch off of the transformer primary switch (responding to diode breakdown in the rectifier). Therefore, it is always appropriate to add the protection to the rectifier evaluating the direction of the DC current flow. This protection is already installed in some built traction substations. Usage of this protection makes sure disconnection of the damaged rectifier unit from the catenary. The only vehicle, that would recuperative even if very low value of voltage at the pantograph, could complicate this situation (i.e. the situation of recuperation close to the traction substation when the current flowing back to the traction substation could be up to tens of milliseconds with very high values). The recuperation at the low catenary voltage is always suitable to interrupt from

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

viewpoint of the traction substation protection from dangerous current flow back to this traction substation. ACKNOWLEDGMENT The research was supported by the TACR grant “Competence Center of Railway Vehicles” No. TE01020038. REFERENCES [1] K. Hlava: Elektromagnetická kompatibilita (EMC) drážních zařízení. University of Pardubice, Pardubice, 2004, ISBN 807194-637-0. [2] F. R. Holmstrom, D. Turner, E. Fernald, “Rail transit EMIEMC“, in: Electromagnetic Compatibility Magazine, Vol.1, Issue1, pp.79-82, 2012, ISSN 2162-2264. [3] M. V. Petkova, “Integration for EMC and network rail's management proces“, in: Railway Electrification Infrastructure and Systems, pp. 232-239, London, 2011, ISBN 978-1-84919512-6. [4] D. Gonzalez, F. Manzanedo, “Optimal design of a D.C. railway power supply system“, in: Electric Power Conference, 2008, pp.1–6, 2009, Vancouver, BC, ISBN 978-1-4244-2895-3.

12

[5] R. Langella, A. Sollazzo, A. Testa, “Modeling waveform distortion produced by DC locomotive conversion system“. Part 1: Locomotive model, in: Harmonics and Quality of Power, 2004, pp. 477-482, ISBN 0-7803-8746-5. [6] “Calculating of peak value and steady-state value of short-circuit current at external short-circuit on bridge rectifier with silicon diodes“, Report, ABB, Poland, 2006. [7] J. Pavelka, Z. Čeřovský, J. Lettl: Výkonová elektronika. ČVUT, Prague, 2007, ISBN 978-80-01-03626-6. [8] J. Ibl: Průmyslová elektronika II.část: Rtuťové usměrňovače. SNTL, Prague, 1955. [9] Calculating of peak value and steady-state value of short-circuit current at external short-circuit on bridge rectifier with silicon diodes. ABB Poland, 2006. [10] C. C. Herskind, H. L. Kellogg: Rectifier fault currents. Transactions AIEE, pp. 145-150, 1945. [11] C. C. Herskind, H. L. Kellogg, A. Schmidt: Rectifier fault currents-II. Transactions AIEE, pp. 243-252, 1949. [12] K. Hlava: Parametry odběru elektrické energie dvanáctipulzním trakčním usměrňovačem v závislosti na jeho zatížení. The Scientific and Technological Anthology of Czech railways n. 14, Prague, General office of Czech railways, 2002. [Online] . [Accessed: 03 Jan. 2010] ISSN 1214-9047.

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

13

Simulation of Magnetodynamic Forces Acting on the Eccentric Rotor of the Generator Tibor Bachorec 1), Petr Král 2) 1)

2)

SVS FEM s.r.o., Brno, Czech Republic, e-mail: [email protected] M.L.S. Holice s. r.o., Olomouc, Czech Republic, e-mail: [email protected]

Abstract — The paper presents results of coupled electromechanical simulations for the generator transition to nominal and special operating conditions. Furthermore, results are given for calculations of the magnetodynamic forces and their spectral components acting on the rotor of the generator depending on its considered eccentricity. The simulations were performed on the electromechanical finite element model of a four-pole generator 130 kVA, 400V, 50Hz, 1500 rev./min in the ANSYS/Maxwell program.

ment of the ANSYS/Maxwell program. The simulations considered the rotor motion, nonlinear properties of ferromagnetic material and coupling between the load circuit and the finite element representation of the stator winding.

Keywords — Simulation, generator, rotor, eccentricity, magnetodynamic, short-circuit, no-load

I. INTRODUCTION At present increasing demands are put on the performance and efficiency of electric machines. At the same time, however, smaller dimensions, weight and lower production costs are expected. These expectations can be successfully faced only by exploitation of new materials and implementation of modern technologies considerably shortening the processes from development to production. Numerical simulations, mainly based on the finite element method undoubtedly belong to them. Nowadays they can comprehensively take into account all the specifics of the electric machines simulation nonlinearities, losses, temperature dependence of materials and magnetomechanical interaction. The last one has recently become interesting from the viewpoint of prediction of vibrations and all related aspects.

Fig.2. Diagram of the stator winding The simulations referred to a four-pole generator with salient poles 130 kVA, 400 V, 50 Hz (Fig. 1). The stator outer diameter was 400 mm, the rotor diameter 269.1 mm and its length was 290 mm. The minimum size of the air gap was 0.9 mm. The stator winding was double-layered in two parallel branches with loop coils (Fig. 2). The simulations took into account the induction of electromotive forces in the stator windings due to the rotor motion, as well as the currents flowing through the windings due to the generator load. During the simulations simultaneous calculations using the finite element and circuit model were carried out (Fig. 3). The circuit was created using the Circuit Editor, which is a part of the Maxwell program. Thanks to this coupling it was possible to simulate various load conditions of the generator.

Fig.1. Cross-section of the generator II. MODEL OF THE GENERATOR Electrodynamic simulations were performed on a twodimensional coupled electromechanical finite element model of the generator which was created in the environ

Fig.3. Schema of the stator winding and the load coupling

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

14

Calculations were carried out on a quarter symmetric model (Fig. 4), with antiperiodic boundary conditions on symmetry planes and zero vector potential Az = 0 applied on the stator surface. The assumed speed of the rotor was 1500 rev./min.

Fig.6. Transition to NOC - phase currents

Fig.4. Symmetric calculations

part

of

the

model

used

for

2.00 1.75 1.50

B (tesla)

1.25

Fig.7. Line and phase voltage and phase current - NOC

1.00 0.75 0.50 0.25 0.00

0

500

1000

1500 2000 2500 H (A_per_meter)

3000

3500

4000

Fig.5. BH characteristics of the considered material

The rotor winding was excited by a current source, which took into account the relevant simulated conditions of the generator. Induction of eddy currents was considered only in the damper rods whose ends were coupled with an equivalent impedance obtained from ANSYS/RMxprt pre-calculation. After successful functionality tests of the model, the simulations were carried out for the nominal, no-load and short-circuit operating conditions of the generator certified by the manufacturer.

Fig.8. Detail of phase current – NOC

A. Simulation of nominal operating conditions During the simulation of the nominal operating conditions (NOC) switches SW1 to SW3 (Fig. 3) were permanently turned on and switches SW4 and SW5 turned off. The generator was connected to a nominal load represented by delta connection of RL components.

Fig.9. Spectrum of phase current – NOC

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

15

Results are shown in Fig. 6 to 9. After reaching the steady state, rms values of line voltages, phase voltages and phase currents were determined and these values were compared with the manufacturer's database. B. Simulation of short circuit and no load conditions Both of these conditions were considered at constant rotor speed, i.e. 1500 rev./min. The short circuit was simulated from NOC by turning on the switches SW4 and SW5 at time t = 100 ms (Fig. 10-11). Simultaneously with this the excitation of the rotor coils was reduced to the value prescribed for this conditions of the generator. The simulation of phase interruption was carried out similarly. After reaching the NOC the switch SW3 was turned off (at time t = 100 ms). Fig. 12 shows the voltage spike on the phase line C due to this load .

Fig.10. Short circuit from the NOC - phase currents

Fig.13. Phase and line voltage – no load

The generator and load were disconnected during the simulation of no the load condition (Fig. 13). The rotor coils were excited by values of currents prescribed for this condition of the generator. Simulated time courses are shown in Fig. 10 to 13. The rms values of the calculated voltages and currents were compared with the manufacturer's database. In both cases good correspondence between the simulated and database values were found whereby the functionality of the FEM model was verified for the next class of simulations – simulations of magnetodynamic forces acting on the eccentric rotor.

III. SIMULATION OF MAGNETODYNAMIC FORCES ACTING ON THE ECCENTRIC ROTOR OF THE GENERATOR A balanced rotor is an important factor in terms of operation, reliability and lifetime of electric machines. Imbalance causes increased bearing load, rotor load and vibrations. The magnetodynamic forces showing magnetic pull between the rotor and stator grow with increasing eccentricity of the rotor. The magnitude of the magnetodynamic forces is increasing in the direction of the diminishing air gap. As a result of the variable air gap (imbalanced rotor) the magnetodynamic forces become uneven around the circumference of the electric machine.

Fig.11. Phase currents – short circuit

Fig.12. Interruption of the phase C during the NOC phase voltages

Fig.14. Time courses of the magnetodynamic forces in dependence on the eccentricity

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

TABLE I. Mean value of the magnetic pull in dependence on the eccentricity Eccentricity [mm]

Magnetodynamic force [N]

0.0

6.5

0.1

686.5

0.2

1385.5

0.3

2100.1

0.4

2841.1

0.5

3610.6

0.6

4458.5

The magnetodynamic forces affect the shaft bending stiffness which is reduced by their action and therefore the resonant frequency of the shaft is decreased. The accuracy of these forces calculation has therefore a significant impact on the accuracy of the rotor resonant frequencies determination.

FFT

2_Load_Excentricity

1000.00 Curve Info Excnt='0.1mm'

16

Calculations of the magnetodynamic forces were carried out by the FEM using the model validated by the previous simulations. But unlike them it was not possible to utilize symmetry and the calculations had to be performed on a model of the full cross section of the generator, as shown in Fig.1. Calculation parameter was the rotor eccentricity which was set from 0.1 to 0.6 mm i.e. in the range of 11 – 66 % of the nominal air gap size. The speed of the rotor was constant and the nominal load of the generator was considered. After transients, time courses of the magnetodynamic forces acting on the rotor (Fig. 14) were recorded, their mean values were evaluated (tab. I) and the Fourier analysis of their spectrum was performed (Fig. 1516). This showed relatively significant presence of higher harmonic components up to the frequency of 2 kHz. The magnitude of these components is minor in comparison with the fundamental harmonic frequency and cannot probably reduce significantly the shaft stiffness, however, these frequencies are potential source of an acoustic vibration.

IV. POWER BALANCE The power balance was performed for the NOC. Eddy current losses in damper rods, winding losses and ferromagnetic losses in steel parts were evaluated. The FEM programs usually offer several algorithms to calculate losses in ferromagnetics - based on the hysteresis loop, on the basis of a simplified algorithm of equivalent hysteresis loop and on the basis of the Steinmetz formula. In our case the last one was utilized: 2

100.00

2

1.5

F [newton]

Pv = Khf(Bm) + Kc(fBm) + Ke(fBm)

(1)

with the following values of loss coefficients:

10.00

Kh=178.6, Kc=1.4, Ke=1.8. 1.00 0.00

1.00

2.00 Freq [kHz]

3.00

4.00

Fig.15. Force spectrum for eccentricity e = 0.1 mm

FFT

2_Load_Excentricity

1000.00 Curve Info Excnt='0.2mm'

• 2D model does not reflect losses on the ends of the winding, • The selected algorithm for losses calculation did not take into account their re-dampening effect on the electromagnetic field, • Detailed information relating to the winding, particularly to the cross-sections of wires was not available.

F [newton]

100.00

10.00

1.00 0.00

The overview of the calculated values is in Tab. II. Distribution of the instant loss density in the ferromagnetics is in Fig. 17. In comparison with active power of the generator 105.2 kW these losses made approximately 3 %. Based on the catalogue data, the efficiency of the generator is in the range of 92-95 %. Our calculated efficiency is higher probably for the following reasons:

1.00

2.00 Freq [kHz]

3.00

Fig.16. Force spectrum for eccentricity e = 0.2 mm

4.00

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

Fig.17. Distribution of the instant loss density in the ferromagnetic parts @time=230 ms

17

For the validated generator model simulations of the electrodynamic forces were performed in dependence on the eccentricity of the rotor. Their time courses were registered and the Fourier analysis of their spectrum was performed. This showed relatively significant presence of higher harmonics up to the frequency of 2 kHz. At the power balance, in comparison with the reference data, somewhat higher efficiency of the generator was calculated. The difference could be partly due to the neglect of some 3D effects in the model (e.g. the effect of winding ends) and partly by the lack of information on the windings. The simulations demonstrated real possibility of the electrodynamic analysis taking into account the magnetomechanical interactions. They also showed relative ease of simulation of any condition of the electric machine represented by a system with distributed parameters (FEM model) and controlled by changes of electrical, mechanical or geometrical parameters.

TABLE II. Power balance of the generator [W] Ferromagnetic losses

920.5

Eddy current losses

627.5

Winding losses

1653.9

Sum of losses

3201.9

V. CONCLUSION In the environment of the ANSYS/Maxwell program the electromechanical model of the generator was created and verified by simulations of load cases certified by its manufacturer. In these simulations good agreement between the reference and calculated values of the phase voltages and currents was found, the differences ranged at the level of a few per cents.

REFERENCES [1] ANSYS Maxwell Technical Notes, ANSYS, Inc., 2012. [2] ANSYS Mechanical APDL Theory Reference, ANSYS, Inc., 2012. [3] Partner Alternators LSA 44.2 – 4 Pole, Leroy Somer Datasheet, 2011. [4] J. Pyrhonen, T. Jokinen, V. Hrabovcova, “Design of Rotating Electrical Machines”, J. Wiley & Sons, 2008. [5] Joshua Lorenz, J. T. Fowler, “Synchronous Generator Subtransient Reactance Prediction Using Transient Circuit Coupled Electromagnetic Analyses & Odd Periodic Symmetry”, Int. ANSYS Users’ Conference, 2006. [6] P. J. Duijsen, U. Killat, J. Otto, ”Parameter extraction in FEM models for dynamic system simulations”, 19th CAD-FEM Users’ Meeting, 2001. [7] Silvester, P., Ferrari, R., ”Finite Elements for Electrical Engineers”, Cambridge University Press, 1996.

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

18

Application of Resilient Power Pilot Project Ivan Benes 1) and Jarmil Valasek 2 1)

2)

AF-CITYPLAN, Prague, Czech Republic, Praha, Czech Republic, e-mail: [email protected] Population Protection Institute, Ministry of Interior - General Directorate of Fire and Rescue Services of the Czech Republic, Lazne Bohdanec, Czech Republic, e-mail: [email protected]

Abstract — The presented pilot project increases the resilience of power distribution against a national transmission grid blackout through adaptive distribution grid islanding. This function should be a basic function of smart grids. In 2011 the project team implemented the pilot project on a local micro-grid. The project demonstrates the possibility of crisis demand side management that enables necessary electricity for residents and critical infrastructure to be provided through adaptive distribution grid islanding. Keywords — blackout, critical infrastructure resiliency, smart grids, island operation, black start, population protection

I. INTRODUCTION The Indian blackout of August 2012 has shown the basic problems that arise when a country’s rapid development leads to demand far outstripping supply. That means power officials must manage the grid by shutting down power to small sections of the country on a rotating basis. But doing so requires quick action from government officials who are often loath to shut off power to important constituencies. Thus, selective crisis demand side management could be a welcome option. Japan suffered from a massive earthquake and tsunami disaster in March 2011, which caused wide-spread damage to nuclear power plants. Areas around of Tokyo had rotating blackouts suddenly in March and then power shortages in summer. During rotating blackouts, energy supply and other elements of the infrastructure were in chaos. With smart distribution technologies, we could avoid such chaos and change rolling blackouts into priority supply of critical infrastructure while ensuring minimal service to all customers. Energy flows and exchanges define the life of all living organisms on the Earth. The same mechanism can be seen in such super-organisms as human society and civilization. No human activity can take place without energy and its transformation. From the point of view of basic biophysics and thermodynamics the wealth of a specific society is primarily determined by size and effectiveness of its energy transformations. Without energy it is possible to secure neither some basic physiological human needs nor the need of safe being. Access to energy in its various forms is a basic condition for the “life” of any society. That is why the governments of various countries pay so much attention to energy security. The strategic planning of resources goes together with ensuring their availability in the time and at the place they are required. The “smartest” kind of energy – electric power – can be very quickly distributed and switched to

any place where a power line is installed and transformed easily to any other kind of energy. But it has one unpleasant feature: it cannot be stored so consumption and production must be balanced all the time. All the networks and switches are capable of redirecting the energy in seconds, but when such a super-system fails, it means that millions of people and all key businesses may be out of energy for a long time – plunging into a black-out. Blackouts are a "Sword of Damocles" of our civilization as illustrated figure 1. Due to our dependence on electricity, society could be threatened within a few seconds. Coincidence of weaknesses

Climate Change & Disasters

Intentional attack

BLACKOUT

Physiological needs: breathing, warmth, water, food, shelter, sleep, ….

Maslow's hierarchy of needs

Safety needs: protection, security, order, law, limits, stability

Belongingness needs

Esteem needs

Selfrealization

Due to interdependency we couldn’t ensure basic human needs without electricity

Fig.1 Blackout is sword of Damocles over our life

Basic physiological and safety needs would not be possible to satisfy without electricity. If the imbalance of production and consumption in the electricity supply is not immediately removed, the power system breaks down in a few seconds and results in a blackout. This situation may then last for days or weeks. The current public distribution networks are passive and are not able to provide electricity supply from local sources without a connection to the transmission grid. Thus the blackout will hit all regions. Local power distributors could manage this unacceptable risk if they would be able to stay in operation islanding through utilization of local generation, even if system services from the transmission grid are lost (i.e. during the system failure). However, the requirement for the island operation goes against the common practice. For example reference [1] stated that islands pose a significant risk to equipment and

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

need to be quickly detected and eliminated. However, this pure view in terms of distribution companies is inconsistent with a broader view of human security. Reference [2] considers islanding as the effective means to increase the resiliency of the National Critical Infrastructure. Better understanding of critical interdependencies among core infrastructures is namely one of the most important requirements to mitigate the impact of extreme events and improve survivability. Reference [3] contains an analysis of 2003 Italy blackout. This blackout became a reality after about 2 minutes after the disconnection of Italy from the UCTE grid except Sardinia and a limited number of load islands. All other generators got tripped within both the distribution grid and the high voltage grid. The restoration process was based on creating as many islands as possible in the southern areas. This highlights the benefit of island operation not only for consumer, but also to restore operation of the transmission system. A lot of theoretical works were done on the vulnerabilities of the power system. This complex problem is subject of research and development as well as doctoral thesis. References [4] and [5] are one of the best of them.

19

transmission substations with vulnerable HV transformers, it is almost impossible to increase surveillance on power lines that lead over country with free access to its pylons. A collapse of the transmission system means a loss of system services that balance the frequency and voltage level, and thus distribution grids cannot remain in operation. Standard condition

Alert condition

Self actualisa tion

Sebereal izace

Esteem needs

Esteem needs

Love and belonging

Love and belonging

Security needs

Security needs

Physiological needs

Physiological needs

State of emergency

FastSebere recovery alizace regardless of Uznání economic damages Love and belonging Security needs

T < 24 hours

Physiological needs

Blackouts: USA 2003, Italy 2003, …

Standard condition

Alert condition

Selfactualisaiion Esteem needs

Sebereali zace

Love and belonging

Love and belonging

Security needs

Security needs

State of emergency

Recovery

Esteem needs

T > 5 days

II. METHODS AND DISCUSSION Reference [1] was the first from past projects sponsored by Czech Ministry of Interior which have been devoted to an analysis of protecting the population from possible failures of critical infrastructure systems. These preliminary works were focused on relevant disasters, impact analysis and risk analysis. Blackouts hit all customers immediately in seconds because electricity is an un-storable commodity. The current practice in treating blackouts is acceptable only if the power outage remains for less than 24 hours. After that time serious problems arise for most customers and the life of community is widely disturbed. Figure 2 explains why a minimal service level for the power supply is important not only for critical infrastructure but, due to economic losses also for national economy and inhabitants. Emergency supply with using smart grid technology can reduce this risk. If we consider that 30% of power production is connected to distribution networks (e.g. municipal district heating plants and independent power producers), it seems to be an attractive option to ensure emergency power supplies through the use of distributed power generation. The lost system services coming from transmission could be replaced by a special function of smart grids that will provide crisis demand side management. It will allow not only the supply of the critical infrastructure but also supply a limited but indispensable amount of power for inhabitants and business. References [2], [3] presents comprehensive risk analysis of the entire Czech energy system, which was done while solving the problem of emergency power supply in a crisis situation. At figure 3 the risk scoring has shown. The most vulnerable energy infrastructure is the transmission grid. Due to the N-1 practice (power system is resilient to failure of only one key element), multiple malfunctions (or multiple attacks) can cause serious system failure: a blackout. From the impact point of view the most critical part of the power system is the transmission grid. While it is possible to increase surveillance and the resilience of power plants and

Recovery

Physiological needs

Physiological needs

Collective security formation Community desintegration (plunder gangs)

Experience: New Orleans (Katrina 2005), Haiti a Chile (earthquake 2010),…

Standard condition

Selfactualisaiion Esteem needs

Alert condition

Sebereali zace Esteem needs

State of emergency

Sebere Critical alizace infrastructure Uznání resiliency

Recovery

FastSeberea recovery lizace regardless of Uznání economic damages

Love and belonging

Love and belonging

Sounáležitost

Love and belonging

Security needs

Security needs

Security needs

Security needs

Physiological needs

Physiological needs

Physiological needs

Physiological needs

T > 5 days

Fig.2 Principle of critical infrastructure resiliency

Energy Critical Inf rastructure Risk Scoring

0%

20%

40%

60%

80%

100%

transmission line, N>2 transmission station, N>2 big powerplant, N>2 nuclear plant with leakage ground oil storage radioactive storage ground gas storage compressor plant transmission gas pipeline transmission oil pipeline hydrostation + dam oil products pipelines underground oil storage underground gas storage transmission line distribution station, N>2 big powerplant distribution line, N>2 transmission station reduction gas station heating plant high presure gas pipeline small powerplant distribution cable, N>2 district heating grid distribution station low presure gas pipeline heat exchanger

Fig.3 Results of criticality analysis - risk scoring of energy systems

Reference [4] presents the consequent project “Resilience of the distribution system against national grid

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

blackout to improve human safety” which was partly sponsored by Ministry of Trade and Industry in the framework of the “Sustainable prosperity" programme. The goal of the five-year project with a budget of 2 million EUR was to improve security of citizens, protect the environment and decrease the damages caused by long-term power failures. The results will help to diminish the unacceptable risk associated with possible crises in the power supply. The project team consists of five experienced organizations (AF-CITYPLAN, ViP, EGU CB, T-Soft, AF Consult Czech Republic). During the first phase the project team had executed a search and analysis of possibilities to achieve the appointed goal. In the second phase the project team developed the overall conception. In the third phase the technical requirements for new equipment were defined. The technical development of new equipment was the subject of the fourth phase and technical documentation was created in the fifth phase. Based on acquired data the economic analysis was performed during the sixth phase. In 2011 project team implemented the pilot project on a local micro-grid. This pilot project has demonstrated the functionality of all necessary equipment and the concept as a whole. The project put emphasis on the interoperability between regional crisis management (responsibility for population protection) and utility management (responsibility for power supply and equipment protection). This research project has been discussed many times with officials from the Czech parliament and Czech government. The vision of island operation as an “island of life” for fulfilling basic human needs has been understood by General directorate of the Fire Rescue Service of Czech Republic. This organization is responsible to solve all non-military crisis situations, like fires, floods, blackouts, car accidents, extreme weather, etc. Reference [5] presents the complementary research project which was solved at the same time. In the power island it would be very desirable that power source will be capable of starting from dark, e.g. without the external electric grid. The principal function of a system black start is to restore the power system. This project included not only the technical and economic feasibility of black-start equipment, but also analysed and modelled transitional phenomena (ferroresonance). Figure 4 shows that during island formation and expansion of the network, it may occur to transience due to dangerous ferroresonance accompanied by overvoltage. The study and modelling of these transient phenomena can prevent their occurrence and reduce the possibility of equipment damage. III. RESULTS The pilot project was implemented in the wastewater treatment plant of the City of Ceske Budejovice (about 100 thousands inhabitants). The reason for this choice was the fact that the project was interesting for management and also won the support of the mayor. Water supply and wastewater treatment are sensitive and vulnerable to power failure. Based on the severe flooding in the year 2002 the City of Ceske Budejovice was willing to reduce its vulnerability. This can be achieved using smart design

20

and specific synergies. In particular, the water supply has to be independent from the external power grid. Before, the city was supplied by only one water pipeline from the distant dam. During the flood in 2002, there was a danger that this pipe would be destroyed. During this flood, the wastewater treatment plant was destroyed. Based on this experience it was decided to drill a new water well in the adjacent area of the wastewater disposal plant as an emergency source of potable water for the city.

Transformer

Transformer Line 1

Line 2

Generator

Fig.4 Transients phenomena during switching island operation

During reconstruction, the wastewater treatment plant was equipped with biogas cogeneration units. They can operate independently of the external grid in island operation mode, so that during a blackout the wastewater treatment plant can remain in operation. Nevertheless in mode island operation mode it is necessary to reduce the load of wastewater treatment plant. Then this local power and heat cogeneration can supply more than just the wastewater treatment plant. During a power outage and after the reduction of unnecessary consumption it can supply electricity to power pumps for emergency sources of potable water. The necessary equipment was installed and the actual test was presented at a workshop in September 2011. The City of Ceske Budejovice has now ensured operation of the wastewater treatment plant as well as an emergency source of fresh water independently of the electrical network. This project shows how a disaster experience creates a willingness to implement potential smart solutions to achieve resilience for the future. Figure 5 shows the rack with the necessary technical equipment and automation. Figure 6 shows a display of normal operation. Cogeneration units generate electricity into the public grid and switching station is supplied from the grid by means of 2 transformers. Balancing automation constantly calculates the difference between local generation output and consumption. During a grid blackout the automation immediately creates an island. Figure 7 shows like local switching station is disconnected from the external grid and connected to the local cogeneration. Record on the

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

figure 8 registered that island is created within 200 milliseconds. The load management automation ensures balancing by switching off the less important consumption.

Islanding automation

21

blackouts would be replaced by rolling gray-outs which leads to enhanced population protection. The pilot project has demonstrated this functionality. During a blackout the local distribution grid operates as an island. Through forced demand side management the homes can have in operation lights, refrigerators, TV and some others small necessary appliances, but dispensable appliances like kettles, washing machines, irons cannot be used. The load is controlled by means of remote load limitation using smart meters. When the smart meters will be installed, the distributor can remotely switch the households to the emergency, without the material loss, keeping people with the basic living conditions. Based on economic calculations and cost-benefit analysis we can state, that such resiliency of the power supply is affordable. Figure 10 present the results.

CPU & balance automation Blackout occurs

Blackout disappears

Fig. 5 Arrangement of pilot project automation

Developed islanding automation

Developed islanding automation

End of island operation

ISLANDING

RETURN TO THE GRID

Start of island operation

0,2 s

Difference between local generation and consumption

Fig. 8 Snapshot of island operation

Key issue is interoperability between power dispatching and municipal crisis plan

Priorities

This line has the lowest priority

SCADA

~

Crisis plan

Transmission grid

Fig. 6 Control panel display – normal operation

Available output

Prioritization Dispatching

~

Distribution grid

~

Load management according to priorities

Local plant

District heat

Functions: Islanding on the distribution level and crisis demand side management Local distribution grid operates like insulated island system

Fig. 7 Control panel display – island operation

If the pilot project is realized in the public grid it is possible to use smart meters to control demand side management. The idea is presented on the figure 9. The smart meter will be able to decrease consumption instead switching it off entirely. The existing practice of rolling

Site generation

Demand side management Inessential consumption necessary

Fig. 9 Concept of power distribution resilience

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

15 (2011)

Including Smart Meter Just functionality

22

results are presented. Researchers would like to thank both Ministries for financial support of these projects: Project No RN20042005001The Protection of Population and its Ties to Critical Infrastructure in the Sphere of the Energy Systems, supported by MI. Project No VD20072008A05 - Systemic Solution of Emergency Power Supply in Crisis Situation, supported by MI. Project No2A-1TP1/065 Resilience of the Distribution System Against National Grid Blackout to Improve Human Safety (RESPO - RESilient POwer), supported by MIT. Project No 2A-2TP1/003 Enhancement of Black Start Ability to Improve the Czech Power System Reliability and Resilience, supported by MIT.

Fig. 10 Cost of power resiliency

Exemplary cooperation between experts across the General Directorate of Fire and Rescue Services and Departments and the Ministry of Industry and Trade led to an agreement to utilize these results to update the State Energy Policy. The new Czech Energy Policy incorporates demand to increase the resiliency of power distribution with island operation into the practice. This requirement is strongly supported by Ministry of Interior, which is responsible for population protection. IV. CONCLUSION The results of various analyses have shown that (not only in the Czech Republic) black-outs might become a reality and that they might have several causes. Although current high-voltage network is reliable, it may break due to for example excessive demand for wind-power transfer from one part of Europe to another or due to a hostile attack. The concept of dynamic islanding comes from the assessment of threats and risks. It decreases the vulnerability of citizens by decreasing the impact of a long-term break of the transmission network. The solution in the RESPO project is based on the fact that the island type of operation existed already 100 years ago, when there was no long-distance transmission system (and of course the quality of the power supply was poor). The utilization of local sources together with an upgrade of equipment and distribution networks may help to progressively decrease the impact of possible national blackout. The pilot project has demonstrated the feasibility of power supply islanding and crisis demand side management according to the prioritization of critical infrastructure. The research results were demonstrated by the successful pilot smart grid implementation. Recommendations based on this research were included into the proposal for an updated Czech energy policy. All bigger cities should have a resilient distribution system with the ability to switch into island operation during national transmission system outage. ACKNOWLEDGEMENT Thanks to the foresight of the Ministry of Industry and Trade (MIT) and the Ministry of the Interior (MI), synergistic effects of research on energy and security challenges were achieved through four projects, which

REFERENCES [1] R. A. Walling, N. W. Miller, “Distributed generation islandingimplications on power system dynamic performance”, in Power Engineering Society Summer Meeting, IEEE 2002, volume 1, pp. 92 – 96 [2] J. A. Hollman, J. R. Martí, J. Jatskevich and K.D. Srivastava, “Dynamic islanding of critical infrastructures: a suitable strategy to survive and mitigate extreme events”, in International Journal of Emergency Management, Vol. 4, No. 1, 2007, pp. 45 – 58 [3] UCTE (2003) “Interim report of the investigation committee on the 28 September blackout in Italy”, The Union for the Co-ordination of Transmission of Electricity (UCTE), Brussels, 2003, p. 115 [4] A. J. Holmgren, “Quantitative Vulnerability Analysis of Electric Power Networks”, Doctoral thesis in safety analysis, Royal Institute of Technology, Sweden, Division of Safety Analysis, Dept of Transport and Economics, RIT. SE-100 44 Stockholm, Sweden, 2006 [5] DeTao Mao, “Survival from Disaster: Interdependencies Management in Critical Infrastructure Networks”, The Faculty of Graduate Studies (Electrical and Computer Engineering), The University of British Columbia (Vancouver), 2009 [6] J. Valasek, I. Benes, J. Rosa, “Project No RN20042005001 The Protection of Population and its Ties to Critical Infrastructure in the Sphere of the Energy Systems”, research report, Population Protection Institute, 2005, unpublished. [7] J. Valasek, “Risks Understanding”, The Science for Population Protection, Vol. 2008, No. 0, Ministry of Interior - GD FRS, Protection Population Institute Lazne Bohdanec, ISSN 1803-568X. [8] I. Benes, J. Rosa, “Project No VD20072008A05 - Systemic Solution of Emergency Power Supply in Crisis Situation”, research report, AF-CITYPLAN, 2008, unpublished. [9] I. Benes et al., “Project No2A-1TP1/065 Resilience of the Distribution System Against National Grid Blackout to Improve Human Safety (RESPO - RESilient POwer)”, research report, AFCITYPLAN, 2011, unpublished. [10] I. Benes at al., “Project No 2A-2TP1/003 Enhancement of Black Start Ability to Improve the Czech Power System Reliability and Resilience”, research report, AF-CITYPLAN, 2009, unpublished.

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

23

Field-Circuit Modelling of Self-Excited Induction Generators Miksiewicz Roman Silesian University of Technology, Gliwice, Poland, [email protected]

Abstract — In the paper computational models of an induction self-excited generator are presented. Circuit models taking into consideration nonlinearity of the magnetic circuit enable calculations of the generator static characteristic at an autonomic operation. The field-circuit model using Maxwell 2D software allows determination of time curves of any electrical variable in different conditions of the generator operation. There are presented basic determined static characteristics using both models and time curves of currents, voltages, torque during self excitation and under symmetrical load and one-phase short circuit, in cooperation with a rectifier system and during connection of the excited generator to network. Keywords — induction generator; squirrel cage motor; cuircuit-field modelling.

magnetizing reactance Xm; it depends on magnetizing current Im. This equivalent circuit diagram does not take in consideration iron losses. For the considered induction machine the parameters of the equivalent circuit diagram were determined using the RMxprt Maxwell software. The data of the induction motor are the following: rated power 7,5 kW; rated voltage 380 VAC; rated rotational speed 965 rpm; winding connection delta. On the basis of the circuit calculations characteristics Xm(Im) and values of other parameters of the equivalent circuit were determined; they are :

Rs = 1,215 ; X σ s = 2,72 ; Rr' = 2,29 ; X σ' r = 4,42

Io

I. INTRODUCTION Both squirrel-cage and slip-ring machines are used as wind generators. Squirrel-cage induction machines are used especially in smaller wind power plants mainly due their greater reliability nevertheless their worse control characteristics. They require to use condensers (Fig. 1) and existence of residual magnetism for their excitation. At autonomic operation the generated voltage frequency depends on rotational speed and load what is the essential disadvantage of the induction squirrel-cage generators. In order to assure constant frequency additional powerelectronics equipment with conversion are used, for example AC-DC-AC. Many publications [1-7; 9] discuss problems related to features of the induction generators at autonomic operation and cooperation with a network. n G

3

Ro C

Fig. 1. Circuit diagram of an induction generator at autonomic operation.

II. CIRCUIT MODEL The equivalent circuit diagram of the circuit model is presented in Fig. 2. The circuit diagram takes into consideration non-linearity of the magnetic circuit as

fr X o

Rs

Is

fr X σs

'

Rr

fr f r - nr

'

fr Xσ r

Uo XC fr

fr X m

Ro

Fig. 2. Equivalent circuit diagram of an induction generator for one phase.

In the equivalent circuit diagram fr means relative frequency - frequency of the stator voltage in comparison with the rated frequency of the motor. The prepared algorithm enables by use of Mathcad software for determination of static characteristics taking into consideration non-linearity of magnetization characteristics determination of characteristics at a constant rotational speed or constant frequency of voltage at the load [5]. The relative frequency may be determined by comparison of the zero real part of the equivalent impedance seen from terminals of the magnetizing reactance to Re(Z z ) = 0 . The stationary working point is calculated by comparison of the imaginary part of the equivalent impedance to the magnetizing reactance:

Im(Z z ) = f r X m .

'

Ir

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

A. Self-excitation of the generator In Fig. 3 there are presented results of calculations of line voltage and phase current time curves at constant rotational speed n = 1040 rpm, during self-excitation of the loaded generator. In order to make possible the generator self-excitation during digital calculations nonzero initial conditions (current) in one of phase windings were set. 400

0.05

a)

0.07

0.09

0.11

0.13

0.15

0.11

0.13

0.15

Torque (Nm)

− 30 − 60 − 90 − 120 − 150

Time (s) t [s]

30

b) 15

Current (A) i(t) [A]

III. FIELD-CIRCUIT CALCULATIONS In the field-circuit calculations the Maxwell software with Transient solver was used. Using this software it is possible to determine various transient states of the generator at autonomic operation at symmetric and nonsymmetric loads and also at cooperation with rectifiers. It is also possible to make calculations for various nonsymmetric states. In 2-dimensional calculations there were taken into consideration the resistance of the stator and leakage inductances of the stator and rotor, determined using RMxprt, the same as in the circuit calculations. The software does not take iron losses into consideration.

24

0.05

0.07

0.09

− 15

a) 200

Voltage (V)

− 30 0

0.05

0.1

0.15

0.2

Time (s)

400

c)

− 200

Voltage (V)

200 − 400

Time (s) 30

0.05

0.07

0.09

0.11

0.13

b) − 200

Current (A)

20 10

− 400 0

0.05

0.1

0.15

Time (s)

0.2

t [s] Fig. 4. Torque, phase current and line voltage time curves at rotational speed change (n1 = 1040 rpm) (n2= 1140 rpm).

− 10 − 20

.

− 30

Time (s) t [s]

Fig. 3. Line voltage and phase current time curves during self-excitation of the generator.

B. Transient state at rotational speed change. In Fig. 4 curves for transient state (a generator with resistance load and separate operation) are presented at discrete change of the rotational speed from n = 1040 rpm to n = 1140 rpm. There were presented the torque, phase current and line voltage time curves. In the stationary state preceding voltage frequency change was of 49.74 Hz and after the speed change it increased to f = 54.42 Hz.

C. Transient state at single-phase short circuit. For the presented field-circuit model it may be performed calculations for an unsymmetrical load. In Fig. 5 there are presented results of calculations (electromagnetic torque, phase currents, terminal voltages) for the following conditions of the generator operation: constant rotational speed n = 1040 rpm; symmetrical load by receiver resistance Ro=30 Ω and next single-phase short circuit of the receiver. As a result of the unsymmetrical short circuit large pulsating component arises in the torque time curve.

0.15

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

25

600

0.06

0.08

0.1

0.12

0.14

0.16

a)

− 40

Voltage (V)

Torque (Nm)

a)

− 80 − 120

400

200

− 160 0.06

− 200

0.07

Current (A)

Current (A)

15

0.1

0.12

0.08

0.09

t [s]

b)

0.08

0.09

40

b)

30

0.06

0.08

Time (s)

Time (s)

0.14

20

0.06

0.07

0.16 − 20

− 15 − 40

Time (s) − 30

c)

Time (s)

Current (A)

t [s]

600

c) 300 Voltage (V)

20

10

0.06

0.07

0.08

0.09

− 10

0.06

0.08

0.1

0.12

0.14

0.16 − 20

Time (s)

− 300 0.06

0.07

0.08

− 600 Time (s) Fig. 5. Torque, phase current and line voltage time curves before and after the single phase short circuit.

Torque (Nm)

d) − 50

− 100

− 150

Time (s)

Fig. 7. Steady state of: a) line voltage, b) phase currents, c) line currents, d) electromagnetic torque time curves.

Fig. 6. Circuit diagram of a generator with rectifiers of the Schematic Capture module.

D. Cooperation of the generator with rectifiers The circuit diagram of the Schematic Capture module of the Maxwell software is presented in Fig. 6. Calculations were performed for the resistance load Ro=30 Ω, capacitors C=90 µF, at the constant rotational speed n = 1040 rpm. Exemplary voltage at the receiver, phase current of the winding and line current time curves

0.09

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

E. Switching on of excited generator to a network The Maxwell software enables calculations of various transient states. It was considered the case of the nonexcited generator connection to a network and setting a chosen torque. The circuit diagram of the circuit system is presented in Fig. 8.

200

a) 0.05

Torque (Nm)

before rectifiers and electromagnetic torque are presented in Fig. 7. Knowledge of these curves allows accurate determining of their RMS values and permissible load.

26

0.1

0.15

0.2

0.25

− 200 − 400 − 600 − 800 − 1 × 10

3

Time (s)

Fig. 8. The circuit diagram of the generator in the Schematic Capture module at connection to a network.

In Fig. 9 there are presented the torque, rotational speed and current time curves of the excited generator (initial rotational speed of 1040 rpm) during its switching on to a network and loading by torque T = 100 Nm. As can be seen from presented time curves no problems occur with synchronization while large switching currents can be expected.

Speed (rpm)

b)

1.1 × 10

3

1.05 × 10

3

1 × 10

3

950

900 0.05

0.1

0.15

0.2

0.25

Timet [s] (s) 300

c) 200

0.05

0.1

0.15

0.2

0.25

0.2

0.25

− 100 − 200 − 300

Time (s) 400

d) Current (A)

IV. STATIC CHARACTERISTICS OF THE GENERATOR Static characteristics of the generator may be determined on the basis of above calculated parameters of the equivalent circuit diagram and on the basis of the field-circuit calculations for stationary operational conditions. In Fig. 10 there are presented external characteristics U = f(I) determined using both above mentioned methods for two values of capacitors at the constant rotational speed n = 1040 rpm and resistance load. On the basis of the circuit calculations it is easy to determine static characteristics at a constant frequency what is more troublesome for the field-circuit calculations and it requires repeated time-consuming recalculations. Comparing the determined static characteristics using both methods it may be noticed slight differences between them. It influences accuracy of the equivalent circuit diagram parameter determination using the circuit method and stability of these parameters, however from practical point of view it should be used for initial determination of capacitors and generator windings circuit calculations. The circuit calculations give greater variability of voltage at generator terminals.

Current (A)

100

200

0.05

0.1

0.15

− 200

− 400

Time (s)

Fig. 9. Time curves : a) electromagnetic torque, b) speed, c) phase currents, d) line currents at switching on the excited generator to a network.

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

V. SUMMARY The circuit and circuit-field calculations show good conformity of calculation results. Therefore it is possible to use the circuit model in the initial stage of designing. The prepared field-circuit model allows for calculations of various transient states types and analysis of results both from point of view of the generator and for various systems.

500

a)

U [V]

Voltage (V)

400

300

200

100

REFERENCES

C=90 uF C=70 uF 0

5

10

Current (A)

15

500

b) 400

Voltage (V)

300

200

100

C=90 uF C=70 uF 0

2 × 10

3

3

4 × 10

6 × 10

3

8 × 10

3

1 × 10

4

Power (W) P [W] 500

c)

U [V] Voltage (V)

400

300

200

100

C=90 uF C=70 uF 0

5

10

15

Current (A) I [A]

500

d) 400

Voltage (V)

27

300

200

100

C=90 uF C=70 uF 0

2 ×10

3

4 × 10

3

6 ×10

3

8 ×10

3

1 ×10

PPower [W] (W) Fig. 10. Static characteristics calculated using the field-circuit model a) U = f(I), b) U = f(P), and using the circuit model c) U = f(I), d) U = f(P).

4

[1] R.,C. Bansal, "Three-Phase Self-Excited Induction Generators: An Overview". IEEE Transactions on Energy Conversion, Vol. 20, No. 2, June 2005, pp.292-299 [2] J. Cidrás, and A. E. Feijó, and C.C. González, "Synchronization of Asynchronous Wind Turbines". IEEE Transactions on Power Systems, Vol. 17, No. 4, November 2002, pp.1162-1169 [3] K. S. Sandhu, and S. P. Jain, "Steady State Operation of SelfExcited Induction Generator with Varying Wind Speeds". International Journal of Circuits, Systems and Signal Processing, Issue 1, Volume 2, 2008 [4] M. G. Simões, and S. Chakraborty, and R. Wood, "Induction Generators for Small Wind Energy Systems". Third Quarter 2006 IEEE Power Electronics Society Newsletter 19, Illinois St., U.S.A. [5] P. Dybowski, and W. Orlewski, "Badania generatora indukcyjnego wzbudzanego kondensatorami". Zeszyty Problemowe – Maszyny Elektryczne Nr 77/2007 23, AGH Kraków, str.23-26 [6] J. Hickiewicz, and J. Moch, "Praca generatora indukcyjnego przy niesymetrycznym obciążeniu". Prace Naukowe Instytutu Maszyn, Napędów i Pomiarów Elektrycznych Politechniki Wrocławskiej. Studia i Materiały, Nr 28, Wrocław 2008, str.412-419 [7] B. Jakubowski, and K. Pieńkowski, "Analiza warunków wzbudzenia autonomicznego generatora indukcyjnego". Prace Naukowe Instytutu Maszyn, Napędów i Pomiarów Elektrycznych Politechniki Wrocławskiej. Studia i Materiały, Nr 30, 2010 [8] R. Miksiewicz, "Maszyny elektryczne. Zagadnienia obliczeniowe z wykorzystaniem programu Mathcad". Wydawnictwo Politechniki Śląskiej, Gliwice 2000 r. [9] R. Miksiewicz, "Modelowanie samowzbudnych prądnic indukcyjnych". Zeszyty Problemowe Maszyny Elektryczne BOBRME-Komel Nr. 98 1/2013, Katowice 2013, ss.121-126 [10] R. Miksiewicz, and M. Pasko: Modelling of Self-Excited Induction Generators. XX International Symposium on Electric Machinery ISEM 2013, Prague September 2013, pp.15-19

Transactions on Electrical Engineering, Vol. 3 (2014), No. 1

Weber, A. R., Weissbacher, J., Steiner, G., Horn, M.: An Accurate Auto-Tuning Procedure for Encoderless AC Motor Drives in Industrial Environments Modern ac motor drives are based on field oriented vector control with feedback and feedforward control units. The feedback control unit for position and speed consists of two cascaded standard PI-controllers. These controllers require information of the position and the derivative of the position, respectively. Typically a shaft encoder provides this information but more often a position observer is used instead. The feedforward control unit uses the set position trajectory in combination with mechanical parameters and essentially imposes the overall dynamics. This work presents a new method for selfcommissioning of speed controller, position controller and feedforward control unit for drives without a shaft encoder. Performance and feasibility of the proposed method are demonstrated by experimental results.

Mlynařík, L., Doleček, R.: The Effect of Fault States of the Twelve-pulse Rectifier During the Recuperation The paper deals with analysis of states from the viewpoint of the rectifier DC side. In the case of diode breakdown the interphase short-circuit occurs and the transformer is disconnected from the primary side by a high-speed switch. After disconnecting the DC current of a recuperative traction vehicle influences the circuit. The fault states of the twelve-pulse rectifier in a substation of 3 kV DC traction system used at railways in the Czech Republic are analyzed. The situation at the rectifier diode breakdown and at the same time recuperative vehicles in power supply section is examined. The vehicle with recuperation can increase strain of elements in power supply system under certain conditions during fault states of the rectifier. The simulation computer model was created for analysis of current waveforms in trolley line at this situation. All simulations were created by the program PSpice. The main goal is to analyze the endanger of power supply system elements and to present the recommendation for increasing the reliability of this mentioned power supply system.

Bachorec, T., Král, P.: Simulation of Magnetodynamic Forces Acting on the Eccentric Rotor of the Generator The paper presents results of coupled electromechanical simulations for the generator transition to nominal and special operating conditions. Furthermore, results are given for calculations of the magnetodynamic forces and their spectral components acting on the rotor of the generator depending on its considered eccentricity. The simulations were performed on the electromechanical finite element model of a four-pole generator 130 kVA, 400V, 50Hz, 1500 rev./min in the ANSYS/Maxwell program.

Benes, I., Valasek, J.: Application of Resilient Power Pilot Project The presented pilot project increases the resilience of power distribution against a national transmission grid blackout through adaptive distribution grid islanding. This function should be a basic function of smart grids. In 2011 the project team implemented the pilot project on a local micro-grid. The project demonstrates the possibility of crisis demand side management that enables necessary electricity for residents and critical infrastructure to be provided through adaptive distribution grid islanding.

Miksiewicz, R.: Field-Circuit Modelling of Self-Excited Induction Generators In the paper computational models of an induction self-excited generator are presented. Circuit models taking into consideration nonlinearity of the magnetic circuit enable calculations of the generator static characteristic at an autonomic operation. The field-circuit model using Maxwell 2D software allows determination of time curves of any electrical variable in different conditions of the generator operation. There are presented basic determined static characteristics using both models and time curves of currents, voltages, torque during self excitation and under symmetrical load and one-phase short circuit, in cooperation with a rectifier system and during connection of the excited generator to network.

_____________________________________________________________________________________________ TRANSACTIONS ON ELECTRICAL ENGINEERING VOL. 3, NO. 1 HAS BEEN PUBLISHED ON 31ST OF MARCH 2014