Analytical approach to dimension transient currents in asymmetric VSC monopole

(1)

IN

DEGREE PROJECT ELECTRICAL ENGINEERING, SECOND CYCLE, 30 CREDITS

,

STOCKHOLM SWEDEN 2019

Analytical approach to dimension

transient currents in asymmetric

VSC monopole

SADRA VARSHOCHI

(2)

Abstract

As an important part of robust design of an HVDC station, all HVDC equipment should stay safe and sound during steady state mode as well as transient condition. In order to have a proper dimensioning for the equipment, maximum transient stress and current should be derived and calculated by performing transient current studies.

The transient current studies should be done twice during the project process, tender and delivery phase. Based on the nature of these two phases, simplified formulas derived from Thevenin circuit is used during tender work while comprehensive HVDC models and simulation in PSCAD is used in delivery phase of a project.

The scope of the Exam work is to derive formulas for the dimensioning transient currents for asymmetric VSC monopoles and create a calculation tool. In order to have a better platform for future developments such as optimized dimensioning of equipment vs transient current, the result is a calculation tool in Matlab with the collected formulas derived from Thevenin voltage and impedances of the equivalent circuit for decisive faults cases.

Keywords

(3)

Abstract

Som en viktig del av en robust design av en station bör alla HVDC-utrustningar vara säker under normal och transient drift. Maximal överström nivå borde beräknas för varje utrustning för att få korrekt dimensionering av varje utrustning. Denna brukar utföras i transient överström studier.

De transient överström studier brukar göras två gånger under varje projektprocessen, anbuds- och leveransfasen. Baserat på karaktären av dessa två faser används förenklade formler från Thevenin-kretsen under anbudsarbete, medan omfattande HVDC-modeller och simulering i PSCAD används i leveransfasen av ett projekt.

Exjobb arbetet här visar hur förenklade formler för dimensionering av transient strömmar för asymmetriska VSC-monopoler skapas.

För att få en bättre plattform för framtida utvecklingar såsom optimerad dimensionering av utrustning vs transient ström har ett verktyg som baserat på Matlab skaffats.

Nyckelord

(4)

1 Introduction

As a part of dimensioning of equipment in HVDC station, equipment should survive and stay sound during transient condition as well as steady state operation. The purpose of the transient current study is to determine the maximum appearing current of each high voltage equipment in the HVDC station which helps to select a proper size for equipment and also calculate proper protection settings for fault clearing.

As a first step of transient current studies, decisive fault cases are identified for each project based of configuration of HVDC station at the first place in order to dimension equipment on the following area:

· AC bus

· AC converter bus including converter transformer

· Converter area including converter reactors and the IGBTs · DC pole bus

In the next of the studies and depending on the project phase, simplified calculation or simulation should be done in order to identify decisive current and stress of equipment in each aforementioned areas. As a trial and error process, the size of equipment is changed to minimize the transient current and also the cost of the project. The simplified of this this process is shown in Figure 1.

(6)

The transient current studies should be done twice during the project process, tender and delivery phase. The studies can help to have a reasonable estimation of equipment costs during tender phase while the result of the studies in delivery phase defines the actual dimension of equipment and also the protection settings. Based on the needs at these two stages, different approaches are used to find maximum stress and appearing current of station apparatuses. Simplified formulas derived from Thevenin circuit in Microsoft Excel are used during tender work while comprehensive HVDC models in PSCAD are used in delivery phase.

The scope of the Exam work is to derive formulas for the dimensioning transient currents for asymmetric VSC monopoles and create a calculation tool in Matlab with the collected formulas in order to have a fast and accurate preliminary calculation tool for transient currents. The work started by drawing current path for decisive fault cases. Later on the formulas was derived with the help of Thevenin voltage and impedances of the equivalent circuit for each fault case. First chapter describes the basics of HVDC, converter technologies and difference between line commuted converters versus voltage source converter. Later on different converter topologies are presented.

Simulation software, HVDC model and HVDC main circuit parameters are presented in the second chapter. This chapter included different fault locations and the necessity of considering of each fault cases.

Different fault cases are studied in different chapters where normal fault behavior, fault current path, primary protection are discussed.

(7)

2 HVDC introduction

2.1 HVDC Transmission

The first commercial HVDC use was in 1954 between in Swedish mainland and Gotland Island. This system used Mercury arc values to provide 20 MW. Since then the need of HVDC links has increased steadily and new technologies has been introduced to power system such as thyristor valves in 1972 and voltage source converters in 1997, [1] and [2].

As described in [1] HVDC applications can be categorized as following:

1. Underwater cables for distances longer than 30 km. The reason is high reactive compensation required for ac cables for this distance

2. Connecting two asynchronous system. Different nominal frequencies or system stability problem makes use of ac connections difficult

3. Bulk power transmission for links longer than 600 km

Rapid control of transmitted power makes HVDC systems perfect solutions with significant impact for power system stability issues.

The main component of HVDC stations is HVDC converters which can work as rectifier to converter ac power to dc or as inverter to converter dc power to ac. Generally there are two type of converter technologies, current source converter (CSC) and voltage source converters (VSC). Current source converters are also known as line commutated converters (LCC) which utilized mercury arc valve or thyristor valves. These types of converters were modified to capacitor commutated converters (CCC) in order to be used in weak systems and back to back applications, [3]. Voltage source converters are based on Insulated gate bipolar transistors, IGBT which consist of a controllable switch placed in parallel with a diode. These types of converters have higher losses but they provide better flexibility and controllability. According to [4] the following items can be mentioned as differences between CSC and VSC:

· VSC has independent control on active and reactive power while CSC can control only active power

· Dynamic response of VSC is faster than CSC

· VSC is not dependent on network short circuit capability · Black start capability

· VSC requires no reactive power support from network

2.2 Converter technologies

(8)

In order to have higher dc voltage, better performance and removing some of harmonics, two six-pulse bridges can be connected in series with two ac voltage sources 30˚ phase displaced. This phased displaced feeding can be achieved by using a transformer with wye connected on secondary side for one six-pulse bridge while the other bridge is fed from a transformer with delta connected on secondary side,Figure 2.

Figure 2 Thyristor valve arrangement for 12 pulse converter

Commutation of thyristor valves in line commutated converters require strong network with short circuit capacity of at least twice of converter rating, [3]. Since thyristors can only operate with current lagging the voltage, line commutated converters need reactive power compensation. This reactive compensation can be done by capacitive shunt banks or series capacitors integrated in the converter (CCC). As a result, the reactive compensation is done is steps in a discrete form and the lack or surplus of generated reactive power should be provided by ac network, Figure 3.

Figure 3 LCC reactive power exchange

(9)

Simplest possible voltage source converter is a two-level converter which each arm consists of couple of series connected IGBTs in order to achieve required DC voltage. All switches in each arm are switched simultaneously and the switching method is pulse width modulation as mentioned before. In this method of switching, a sinusoidal is used as a reference voltage and a triangular wave is used as a carrier. The position of carrier to respect to reference voltage determines positive dc voltage or negative dc voltage should be applied to ac side, Figure 4. As it can be seen from figure below that the output of two level converters has lots of harmonics which requires extensive filtering in order to deliver desired fundamental harmonic to ac net.

Figure 4 Two level converter with two level PWM waveform

In order to lower the need of filtering and also lower the switching losses, converter needs to have an output as closer to sinusoidal voltage than two level converter output, [4]. One of the ways to lower the switching losses and have a closer output to sinusoidal is to lower the blocking voltage and increase the number of level. A three level converter as such shown in Figure 5 is one approach to lower converter losses and have a smoother output voltage. Decreased blocking voltage reduces the switching losses and provides an output closer to sinusoidal waveform.

(10)

The latest generation of converters with lower switching losses and smoother outputs is modular multilevel converters or MMC. These types of converters are consisting of n number of identical two level converters connected in series to achieve desired dc voltage. Each submodule or cell is consist of a capacitor, two diodes and two IGBTs which can insert the capacitor voltage to output or bypass it meaning the cell does not contribute to output voltage.

Figure 6 Output voltage of MMC

(11)

2.3 Configurations

Basically there are two type of configurations for HVDC transmissions, monopolar and bipolar. In monopole configuration there is only one converter stations at each end. This configuration is the simplest and least expensive configuration for transmitting power, [3]. On the other hand, bipole configurations are more expensive but provide more flexibility and redundancy and can be used when higher transmission capacity is required.

The simplest monopole configuration is called asymmetric monopole where each terminal is consisting of only converter. The transmission can be done by use of only one converter, one high voltage line conductor or cable and ground can be used as a return path for the current. Although using ground as a return path can lower the cost of HVDC, the flow of constant dc current through ground can causes damage to environment such as corrosion of buried pipes and earth system of substation, [5]. Such cases which ground cannot be used as current return path, a low voltage electrode line can be used. In cases such as fresh water cable existence or high earth reactivity, [3], ground return cannot be used. The alternative solution for these cases is to use a metallic neutral or low voltage cable to return the dc current.

Figure 8 Asymmetric monopole a) simple ground return b) electrode line return c) metallic return

The most popular configuration for voltage source converters is symmetric monopole configuration. In this case configuration each terminal has one converter and the terminals are connected to each other with two line conductors or cables. In this configuration half of the dc voltage appears on each conductor, one positive and the other one negative.

(12)

Bipole configuration is most popular configuration for CSC technology. In this configuration each terminal has two converters which transmit power via two conductors. As mentioned before, this configuration can be used for cases where higher power capacity is required or more redundancy is needed. Bipole configuration can work as a single asymmetric monopole configuration when one of poles trips.

Figure 10 Bipole configuration

2.4 Thesis description

Depending on the fault case, fault current can be decisive for some equipment. The purpose of this thesis work is to derive formulas of decisive currents of an asymmetric monopole station for different fault cases.

AC.B. F1 F2 F3 F4 F5 Zone 1 Zone 2 Zone 3 Zone 4 Idnv Idpv Ivn Iv Ipcc

Figure 11 Studied fault cases in the thesis

Faults can occur on every location of HVDC station. However, by identifying the most critical and decisive fault cases which give highest current on each location, limited number of fault cases can be studied in order to find maximum fault current of each location. A HVDC station consists of the following zones:

(13)

· Zone 2, AC converter bus including converter transformer

· Zone 3, Converter area including converter reactors and the IGBTs · Zone 4, DC pole bus

In order to achieve highest current of aforementioned area, five fault cases can be studied, Figure 11.

· F1, three phase to ground fault, decisive fault case for zone 1 · F2, three phase to ground fault, decisive fault case for zone 2

· F3, single phase to ground fault, decisive fault case for zone 2 and 3 · F4, pole to ground fault, decisive fault case for zone 4

· F5, pole to neutral bus fault, decisive fault case for zone 4 The detail of studied currents for each fault case can be seen in Table 1.

Table 1 List of fault cases that will be considered for the transient current study

Fault

Case Fault Type

Dimensioning Current

Fault 1 Three-phase to ground fault on the a.c. network bus. _I_pcc Fault 2 Three-phase to ground fault on the converter bus between

the transformer and the converter reactor. Iv Fault 3 Single phase to ground fault on the converter bus between

the transformer and the converter reactor. Iv_t

Ivn, Idn_v Fault 4 Pole to ground fault on the valve side of the smoothing

reactor. Ivp, Idp_v

Fault 5 Pole to neutral fault on the line side of the smoothing

reactor. Ivp, Idp_v

The case study for the thesis work was an HVDC station with circuit parameters shown in Table 2. The transient currents study was performed for worst case steady state conditions. This includes maximum short circuit capacity in the networks, maximum steady state ac voltage at the point of common connection, -5% tolerances for inductances of the main circuit equipment, lowest frequency at steady state and maximum transmitted power on rectifier station.

Table 2 Main circuit data of HVDC station

Main circuit data, common data Bottom Brook

Bridge type CTL VSC

converter Nominal d.c. voltage (UdN), pole-neutral kV 200 Maximum fundamental frequency short circuit current kA 31.5

(14)

The simulation software PSCAD/EMTDC is used. A PSCAD/EMTDC model is including both converter stations.

The fault cases are based on maximum voltages, rated power, minimum frequency and -5 % tolerances for inductances of the main circuit equipment. The ac breaker is left open after the fault appears. The help resistor breaker is assumed to stay open.

Initial condition for the fault cases is that the transmission system is in steady state. The faults are initiated at the time instants which result in the worst stresses for the respective equipment. The worst instant for fault application is found with the help of a ‘multiple run’ component in PSCAD: The component automatically runs a particular fault case several times, with the fault applied at different times during a period. Meanwhile, the peak values of the fault currents of each run are written to an output text file. 20 different runs, with the fault application times evenly distributed over one period, will be made for each fault case. The run that results in the highest fault current peaks is then selected for each fault case.

All fault cases are also simulated under two reactive power (Q) conditions: full absorption of reactive power from the network and full export of reactive power to the network. Among these scenarios, the case resulting in the worst transient current peak is then chosen.

The operational scenarios are summarized below: Bottom Brook Rectifier

P = 250.01 MW QPcc = 125 Mvar Bottom Brook Rectifier

P = 250.03 MW QPcc = -125 Mvar

Nominal a.c. bus voltage (PCC) kV 230 Maximum a.c. bus voltage (PCC) kV 253

AC network X/R 16.6

Nominal frequency Hz 60

Minimum frequency Hz 59.5

Rated power transmission capability, PdN MW 250 Converter reactor arm inductance mH 18

DC smoothing reactor mH 10

Number of cells / arm (N) 14

Cell capacitor size mF 1.0

Transformer rated power MVA 300

Transformer ratio kV/kV 230/121

Transformer reactance % 14

Tap-changer location line side

Min Tap-changer position on Max AC voltage T.B.D Max Tap-changer position on Min AC voltage T.B.D

External filter MVAr 30

(15)

3 AC bus three phase fault

3.1 General description

AC bus fault refers to fault on primary side of transformer. Normally the HVDC scope is up to AC breaker and most of protections are tripping this breaker in order to clear the fault. As a result, a current measurement should be place right after the AC breaker in order to convert the whole area between AC breaker and converter transformer. This area can be seen in Figure 12.

Figure 12 AC bus protection zone in Single Line Diagram of an example HVDC station

The fault can feed from both AC network and deblocked converter. Since the AC network has higher contribution to the fault current than deblocked converter, only AC network current should be considered for this fault.

In order to have maximum possible fault current, a three phase fault between AC breaker and converter transformer is considered.

3.2 Primary protection

Generally there are two protections which are tripping the station if a fault occurs on AC bus, AC bus differential protection and pre insertion overload protection. AC bus differential protection is protecting the equipment on AC bus from the faults occur on differential zone while pre-insertion resistor overload protection is protecting pre insertion resistor from overheating. Pre insertion resistor is used to limit the inrush current during energization of converter. The breaker or disconnector across the resistor bypasses the resistor after some time when inrush current goes lower than acceptable current level of valves. The pre insertion overload protection calculates the dissipated energy to the resistor and trip the AC breaker reaches resistor’s limit. Since harmonics can overload the resistor, this protection uses measuring cores of CTs.

(16)

Figure 13 Measuring points of AC bus differential zone of a HVDC station

Dispute differential protection which is sensitive to the faults within the protection zone, pre insertion protection can react to any faults after the pre insertion such as DC pole to ground fault during energization. Since differential protection is faster and more precise on the fault location than pre insertion overload protection, it can be consider as primary protection toward faults on AC bus.

Protective actions after detecting a fault consist of blocking the valves, opening the breaker start a breaker failure protection in order to send the trip order to next breaker in case of breaker failure.

3.3 Derived formula

As mentioned earlier, a three phase fault applied to AC bus. The fault current consists of network contribution and converter contribution. Since the converter contribution is much less than network contribution, only network contribution can be considered.

Figure 14 Network and converter contribution of AC bus three phase fault a) AC bus voltage (kV) b) Network contribution (kA) c) converter contribution (kA)

(17)

and initial condition prior to the fault. In other words, a k factor is needed to compensate for initial offset of the current when the fault occurs and √2 is to convert rms to peak value.

Figure 15 Fault current path of a three phase fault on AC bus. AC network represented with its Thevenin equivalent circuit

Normally ⁄ ratio of a network should be provided by HVDC customer. This ratio is based on nominal value of the system. These values are nominal voltage, network short circuit capacity and nominal frequency.

Considering a constant ⁄ ratio, higher network voltage can increase symmetric fault current and evidently it increases maximum transient fault meaning in order to have maximum fault current, highest operational network voltage provided by HVDC customer. Table 3 shows an example of voltage data provided by HVDC customer. In order to calculate maximum transient current of AC bus faults, 253kV and 380kV should be used for station 1 and station 2 respectively.

Table 3 An example of network voltage data provided by HVDC customer

AC voltage (kV) Bottom Brook

Nominal voltage 230

Voltage range, steady state 207-253 Voltage range, temporary for t ≤ 30min N.A.

(18)

Table 4 An example of network frequency data provided by HVDC customer

AC frequency (Hz) Bottom Brook

Nominal frequency 60

Frequency range, steady state 59.5-60.5

(19)

4 Converter bus three phase fault

Converter bus fault is referred to the fault on the secondary side of transformer and AC side of converter reactors. Since the converter is grounded on the neutral DC bus and transformer has a delta winding configuration on secondary side, half of positive pole to neutral voltage appears as a dc offset on phase to ground voltage of converter bus. As a result of this dc offset, converter voltage cannot be measured by normal capacitive voltage transformer. Instead of CVT, voltage of converter bus can be measured with the help of direct voltage dividers. This dc offset can be seen in Figure 16where converter voltage during and before a three phase fault is shown.

Figure 16 Converter bus voltage before and during of a three phase fault a) converter voltage (kV) b) Pole to neutral voltage (kV) c) Deblock status of

converter

Generally the built of voltage on DC side starts when converter is energized. During energization, rectified ac voltage by diodes will charge the pole capacitors on the dc side. When converter is deblcoked, converter keeps the dc voltage constant. During three phase fault on converter bus, converter stays deblocked for a while after fault occurrence. While converter is deblocked and three phases is present, dc voltage drops significantly. Even though converter dynamic control tries to restore the converter bus voltage and maintains dc voltage, deblocked switches can create a path for dc pole capacitor to discharge into the fault.

When converter is blocked, the only remaining path between fault and dc pole capacitors will be the diodes. However, higher voltage of dc side compare to converter voltage bus creates a reversed biased for the diodes. As a result, the dc can be seen isolated from the fault.

(20)

Figure 17 converter bus three phase fault a) converter voltage (kV) b) converter current contribution (kA) c) ac network current transferred to secondary side of

transformer (kA)

In practice converter bus is protected by number of protections. Some of these protections are primary protections for converter bus area such as converter bus differential protection and some are backup protections such as converter overcurrent protection.

Figure 18 shows differential zone of converter bus.

Figure 18 Converter bus protection zone in Single Line Diagram of an example HVDC station

As mentioned in previous section, converter bus is protected by couple of different protections. Depending on HVDC configuration and operational point, one or more than one protection can detect a three phase fault on converter bus.

(21)

the differential zone, transformer CTs should be the same as the current exits the zone, DCOCTs in positive and negative arms of respective phase, Figure 19.

Figure 19 Measuring points of converter bus differential protection of a HVDC station

Normally the fault current fed from the converter to a three fault can trigger the converter overcurrent protection. As a primary action from the converter overcurrent protection, protection temporary blocks the valves so they don’t get exposed to a higher current than their rated valve and then it opens the main breaker to clear the fault.

Pre insertion resistor overload protection is another protection which can trip the station in a certain condition when three phase fault occurs on converter bus. If the three phase fault is present before energizing the station, Pre insertion breaker can also detect the fault. Even though the fault current is low because of having an extra resistor in the circuit, pre insertion resistor is exposed to a current which is more than normal inrush current. As soon as calculated energy of pre insertion resistor reaches its maximum allowed energy, pre insertion resistor overload protection trips the station.

Depending on fault clearing time of fast protections some other protections can work as back up protection. DC pole differential protection and dc pole unbalance voltage protection can be named as backup protections for converter over current protection and converter bus differential protections in case of a converter bus three phase fault.

DC abnormal voltage protection measures the voltage of pole to neutral voltage. If this voltage is lower than defined settings for the protection, it permanently blocks the converter and opens the ac circuit breaker. When a three phase fault occurs on converter voltage, deblocked valves can provide a path for dc pole capacitor to discharge into the fault. The capacitor discharge reduces the voltage across the pole capacitor. This voltage is equal to pole voltage. As a result of this discharge, pole voltage suffers from abnormal variations which lead to have a voltage below nominal value. As mentioned before, abnormal voltage protection is triggered as soon as a low voltage criterion is met.

(22)

The method to calculate and derive a formula for three phase converter bus fault is the same as the method described in section 3.3. As explained before, fault current can be calculated with two steps, symmetrical current and then peak of asymmetrical current. The only difference between this section and AC bus three phase fault calculation is ⁄ . This ratio in ac bus fault calculation is the ⁄ of ac network and it should be provided by HVDC customer or grid owner. For converter bus three phase fault, ⁄ will be a combination of network ratio and all other HVDC major equipment located between the fault and AC network. The only equipment which can have a major effect on the fault current is the transformer. In other words, the ⁄ of the network should be combined with ⁄ of converter transformer. Converter transformer ⁄ ratio should be defined by HVDC before ordering the transformer. This value is provided later on the transformer is manufactured by transformer supplier. This value can is defined based on nominal values but in some cases the frequency of the tests could be slightly below nominal frequency.

(23)

Figure 20 Fault current path of a three phase fault on converter bus. AC network represented with its Thevenin equivalent circuit

The current path shown in

Figure 20 consists of the path from AC network and DC side. The path from DC side is originated from DC voltage shift. As mentioned before, when three phase fault occurs on converter bus and converter is still deblocked, the voltage of neutral changes to a negative value. As a result, the arrester on neutral bus starts conducting. However, when converter is blocked by protections, the negative voltage on neutral bus disappears and leads to decay of current on dc side, Figure 17. Since the protections of the converter react to the fault fast and the dc contribution to the fault does not superpose on network contribution, the converter contribution can be ignored.

(24)

5 Converter bus single phase fault

The single phase to ground fault on converter bus is referred to faults between converter reactor and transformer secondary side on one of the phases. Shortly after detecting the fault, converter over current protection blocks the valve. However, diode rectification creates a negative voltage on neutral bus. Having negative voltage on neutral forces the arrester to conduct on neutral bus. The fault current on the secondary side of transformer needs to close its path through the neutral bus or electrode. During the time that diodes on negative of healthy phases are conducting, the voltage on the neutral bus is proportional to converter bus voltage which has lost its dc offset. During the fault, even though the diodes of healthy phases might be reversed biased, they conduct the fault current. This make the voltage on neutral bus to become positive until the other healthy phase start conduct the fault current and take the voltage down to a negative value. This positive voltage can make the diode of the faulty phase to be forward bias and force them to conduct which creates an alternative path for fault current.

Figure 21 Converter bus single phase fault, a) converter voltage (kV) b) neutral voltage (kV) c) negative valve arm currents (kA)

Since the voltage on positive pole does not change much during the fault and after the converter is blocked, all diodes on positive arms are reversed biased. As a result, no current flow from converter bus to positive pole.

(25)

Figure 22 Converter bus single phase fault, a) negative arm currents (kA) b) transformer secondary current (kA) c) network contribution current (kA)

The delayed zero crossing of the current on ac bus should be taken into account of calculating the instance of ordering the main breaker to open otherwise the breaker won’t be able to break the current until the current reaches the zero. If the time between opening of main contacts of breaker and the time the current reaches zero is more than breaker specification, it can causes catastrophic damage to the beaker chambers and main contacts.

As mentioned in previous section, converter bus is protected by couple of different protections. Depending on HVDC configuration and operational point, one or more protections can detect single phase fault on converter bus. Since the fault occurs within converter differential zone, differential protection can detect and clear the fault by opening the main breaker. The measuring point of this protection includes transformer bushing CTs and DCOCTs on converter arms. As a basic functionality of a differential protection, sum of currents enter the differential zone, transformer CTs should be the same as the current exits the zone, DCOCTs in positive and negative arms of respective phase, Figure 23.

(26)

Converter overcurrent protection is another protection which can detect the fault and it can be seen as the fastest protection which block the converter and trip the main breaker. This protection uses the arm current measurements on both positive and negative arms and as mentioned in section 4.2, it protects the converter in two different approaches, temporary blocking and permanent blocking.

If the station is energized against a single phase to ground fault, pre insertion resistor overload protection can sees the fault and it can order the breaker to open. However, this protection as mentioned earlier is slower than converter bus differential protection and it cannot be seen as primary protection for this type of faults.

There are other protections which might be in included in the protection system such as neutral bus differential protection, pole differential protection and electrode line open circuit and these protections can react to converter bus single phase faults.

Pole differential protection follows the same principle of all differential protections. It measures the difference between current goes out from positive pole and enters the neutral bus.

Neutral bus differential protection is another differential protection which might react to converter bus single phase faults if any of the CTs or their range is missing or chosen improperly. This protection calculates the currents enter the neutral bus, negative arms, and the currents exit the bus, electrode, pole capacitors and neutral arrester.

(27)

Figure 24 Converter bus single phase fault a) converter voltage (kV) b) positive bus voltage (kV) c) positive arm currents (kA)

Neutral bus voltage is rectified voltage of two healthy phases. Considering converter reactors between converter bus and neutral bus, diodes on negative arms continue conducting even though the converter voltage becomes more than neutral voltage. In this case, neutral voltage follows converter bus voltage to a positive value. During the time neutral bus voltage is positive, diodes on faulty phase start conducting, Figure 21.

Since diodes only on negative arms are conducting during the fault, the fault current on two healthy phases are always negative, from converter toward transformer, while the current on faulty phase is always positive, from transformer toward fault. This causes the dc components on primary side currents which creates the delayed zero crossing of the current. The fault can be seen in Figure 25.

(28)

Based on the fault current behavior, this fault case cannot be studied as a single phase to ground fault. There reason is that in fault analysis based on symmetrical components, fault current of healthy phases should be zero while the whole fault current should find its way through faulty phase. However, in this case, fundamental frequency of fault current in all three phase has more or less the same amplitude but two healthy phases have negative dc offset because the diodes can conduct in one way and the faulty phase has positive offset because it is the sum two healthy phases in reverse direction. The assumption for the method used here is to assume diodes are conducting all the times without instruction of current in healthy phases.

Another difference between the method used here compare to standard one phase to ground fault analysis is that the fault currents does not have initial dc offset, Figure 26.

The method used here is to consider this fault case as a three phase fault but without dc component. After calculating the fault current, two of currents are given negative dc offsets to create fault current of healthy phases with diode conduction. The sum of healthy phases current can simulate the faulty current with opposite direction.

(29)

Figure 27 faults currents a) simulated b) hand calculation

(30)

6 Pole to neutral fault

Pole to neutral fault refers to fault between pole bus and neutral bus on the DC side. In reality, this fault can happen when station is energized while portable groundings in both positive and negative buses are left on the DC side. Another possibility is that both wall bushings in pole bus and neutral bus fail at the same time. During the fault, the voltage of pole bus reaches to zero. Normally this type of faults can be detected by different protections based on converter operational mode. If portable grounds are the cause of the fault, Pre insertion resistor overload protection which described in section 3.2 can detect the fault. If the fault occurs when converter is already energized, protections on dc side such as pole bus differential protections and neutral bus differential protection can detect the faults. If the fault occurs when converter is deblocked, converter protection can detect the fault alongside dc side protections. Converter thermal overload protection and converter over current protection are the protections which react to pole to neutral bus faults.

As a result of pole to neutral fault when dc pole voltage reaches to zero, the dc offset on converter bus disappears. The remaining voltage on converter bus is the voltage of ac side minus the voltage drop on transformer impedance, Figure 28. On the other hand, the fault current is dependent on the converter voltage and during of conduction of diodes in positive and negative arms.

Figure 28 Pole bus to neutral bus fault a) converter voltage (kV) b) positive bus voltage (kV) c) negative bus voltage (kV)

(31)

Figure 29 Pole bus to neutral bus fault a) converter voltage (kV) b) positive arm currents (kA) c) negative arm currents (kA)

Generally there are two types of protections on dc side, current base and voltage base. In asymmetrical configuration, faults on dc side leads to high current thus it is more efficient to use current base protections while in symmetrical configuration, dc faults leads to abnormal voltage changes.

Although pole to neutral bus leads to both abnormal voltage changes and huge fault currents in both symmetrical and asymmetrical configuration, protections of dc side are tuned to react on abnormal currents in asymmetrical converters. According to this philosophy, dc side of asymmetrical converters is protected by differential protections. There are two separate differential protections on pole bus and neutral bus which are protecting their respective buses.

These protections follow the general principle of all differential protections, calculating the differential current of their protection zone and react to differential currents higher than specified protection settings. The protection of these two protections can be seen in Figure 30.

Figure 30 Pole bus and neutral bus side differential protection zones in a HVDC station

(32)

detect pole to neutral faults and they are more used in order to prevent maloperation of dc differential protections during normal operation or other type of faults such as converter single phase faults.

Figure 31 Measuring points of pole bus and neutral bus differential protection in a HVDC station

As mentioned in previous section, converter operational points affect the fault detection. Pre insertion resistor overload protection can detect the fault if the station is energized against existing pole to neutral bus fault. While pole differential protections and converter protections detect the fault when station is energized or deblocked.

Generally the converter overcurrent protection orders the converter to temporary blocks its valves when the current through valves are higher than protection settings. In transient cases, the current through the valve reaches the protection level and as a result, the protection orders temporary blocking which forces the current goes down. When the current goes below protection settings, the temporary block order is released and converter can continue its normal operation. However, pole to neutral fault derive huge current through deblocked valves. As soon as the current reaches protection settings, protection orders temporary block order to valve. Despite of transient cases which high current is reduced by blocking the converter valves, blocking the valves cannot reduce the fault current effectively since the fault current closes its path through diodes. As a protective action in converter overcurrent protection, protection orders a permanent block order when the temporary block cannot lowers the current to safe operation current. This protective action is the fastest protective action which blocks and trips the main breaker when pole to neutral bus fault occurs.

(33)

contribution while the other one is cable discharge or the ground current though neutral bus. These currents can be seen in Figure 32.

Figure 32 Currents on pole and neutral differential protections a) sum of positive arm currents (kA) b) pole bus current fed from cable (kA) c) sum of

negative arm currents (kA) d) neutral current (kA) 6.3 Derived formula

The behavior of HVDC system during a pole to neutral fault can be described as a second order model. This method is shown for DC traction supplies in [8] and for HVDC stations in [4]. According to [8] the rate of rise and prospective peak fault current of pole to neutral bus fault is relatively high because the fault current is only limited by converter and source impedance. As it can be seen from Figure 33, the fault current closes its path on dc side, through smoothing reactor. The second order model method to calculate maximum peak current presented in [8] is based on calculating damping ratio and undamped natural frequency and later on changing the damping ratio based on the size of smoothing reactor on DC side.

(34)

Figure 33 fault current path of pole to neutral fault

As mentioned before, steady state and peak current should be calculated first in order to calculate the damping ratio and undamped natural frequency. After detecting the fault, converter is ordered to block so the converter acts like half bridge diode rectifier.

The next step is to calculate the maximum overcurrent and later try to identify the behavior of the system based on damping ratio and undamped natural frequency.

(35)

7 Pole to ground fault

Pole to ground fault refers to ground faults between converter and smoothing reactor. As a result, the rise rate of fault current is higher than pole to neutral faults where smoothing reactor affects the damping ratio. During the fault, the fault current closes its path through electrode or ground connection on neutral bus. Compare to pole to neutral bus fault, more resistance exists on the fault path which makes the steady state fault current smaller bus the overshoot and max current can be higher since there is less reactance in the fault circuit. It should be noted that if the neutral connection is grounded, there is no significant difference between fault current of pole to neutral fault and pole to ground fault.

Another difference between pole to neutral bus fault and pole to ground fault current is that since there is less reactance in the fault path, more oscillations can be seen on the current and the voltage, Figure 34 and Figure 35.

It can be seen from Figure 34 that the voltage of neutral bus is limited by arrester during a short while after the fault. This is because when the pole to ground fault occurs, the voltage on the other pole will be shifted. However, the arrester and also the ground connection on neutral bus prevent the neutral voltage to drift a lot from its rating.

The fault behavior is more or less identical to pole to neutral bus fault, fault is feed from ac network, dc component of converter voltage disappears during the fault which causes both diodes on positive and negative arms conduct. As mentioned in previous chapter, different protection can detect this type of fault, Pre insertion resistor overload protection, converter over current protection and pole bus differential protection.

(36)

Figure 35 Pole bus to ground fault a) converter voltage (kV) b) pole bus current (kA) c) neutral bus current (kA)

Since the nature of pole to ground fault is the same as pole to neutral bus, more or less the same the protections which react to pole to neutral bus fault can react to pole to ground faults.

Converter protections such as converter overcurrent protection are the first protections which detect the fault. First action from converter overcurrent protection is to block the valve temporary in order to protect the valve from high current and also interrupt the fault current. As mentioned before pole to ground fault removes the dc component on the converter voltage. As a result, the diodes on both positive and negative arms are forward bias and start conducting. In other words, temporary blocking of valves cannot interrupt the fault current. Another protection which can react to the fault is pole bus differential protection. This protection works based on differential current. In case of fault on pole bus, the current fed from ac network is different in size and direction compare to cable discharge current which feeds the fault from dc side. Figure 38 illustrates the valve currents, dc current from ac side and cable side.

(37)

Figure 37 pole bus to ground, measuring points of protections

Another protection which can react to the fault is pole bus differential protection. This protection works based on differential current. In case of fault on pole bus, the current fed from ac network is different in size and direction compare to cable discharge current which feeds the fault from dc side. Figure 38 illustrates the valve currents, dc current from ac side and cable side.

Figure 38 pole bus to ground fault currents a) positive and negative arms (kA) b) cable discharge (kA) c) dc current from ac network (kA)

(38)

Figure 39 pole to ground fault, current path

(39)

8 Conclusions

The faults defined in this thesis are the ones which can result to maximum transient current on different zones of HVDC. In this thesis work, the protections are not simulated in PSCAD model and their clearing time and their protective behavior is not considered when calculating the maximum fault current. The summery of protection which can be triggered for each fault cases are shown in Table 5.

Table 5 protection summery

Fault cases Primary Protections Secondary protection

AC bus 3 phase fault AC bus diff. prot. PIR overload prot.1 Under voltage prot. Converter bus 3 phase fault Convert bus diff. prot.

Converter overcurrent prot. PIR overload prot. 1 Converter bus 1 phase fault Convert bus diff. prot.

Converter overcurrent prot. Restricted earth fault prot.

PIR overload prot.1 DC diff. prot.

DC pole bus diff prot. neutral bus diff. prot. DC positive pole to ground Converter overcurrent prot. DC diff. prot._{DC pole bus diff. prot.} DC positive pole to neutral Converter overcurrent prot. DC pole bus diff. prot._{neutral bus diff. prot.}

The result of the work can be seen in Table 6 where result of hand calculation is compared with PSCAD simulation for two different HVDC stations. PSCAD column shows the simulation results on two projects. The derived formulas are defined in Matlab as simple Matlab scripts and the results gained by Matlab scripts are shown in Matlab columns.

The models to simulate transient currents has been developed by ABB with purpose of doing all required detailed studies which are needed for delivery projects such as transient current studies, transient over voltage studies. Meaning the models are not developed only of the soul of this thesis worked hence the accuracy of simulated results are as accurate as they can be.

Maritime is the project which is used as case study. NSL project is another HVDC project which is used as verification of thesis work output. It is noteworthy to mentioned that Maritime project has an electrode line connection while NSL is solidly grounded on neutral bus.

(40)

Table 6 Thesis result

Fault cases Currents Maritime (kA) NSL (kA)

PSCAD Matlab PSCAD Matlab

AC bus 3 phase fault Ipcc 68.89 71.42 111.72 112.54 Converter bus 3 phase fault Iv 24.26 25.76 18.37 18.19 Converter bus 1 phase fault

Iv 17.17 17.29 20.68 22.08 Ivn 12.51 11.53 16.13 14.72 Idnv 17.05 17.29 20.68 22.08 DC positive pole to ground Ivp_Idpv 10.33_10.58 _14.3911.75 _15.9313.15 13.26_16.24 DC positive pole to neutral Ivp_Idpv 12.54_14.09 12.60_15.50 13.45_15.89 13.31_16.16

It can be seen from the table that the derived formulas for AC bus three phase fault and Converter bus three phase faults are very accurate and they can be used to determine maximum current on AC and converter busbars.

Considering the number of transients and dynamics which is involved in Converter bus single phase fault, conventional fault current analysis could not be used. Using the new method which is method in this thesis to calculate fault current for this fault can be the reason for having some error between simulated and hand calculation results.

The method which is used for calculating DC positive pole to ground and neutral buses are accurate and trustworthy. By looking at above table, one can see the results gained from simulations are close to hand calculated values. The only difference is for DC positive pole to ground in Maritime. The explanation here is that Maritime project has an electrode connection and the resistance of electrode can reduce the current in software simulation. However, this is not very important because decisive current on positive pole bus is achieved by DC positive pole to neutral fault.

As it was mentioned before, the method to calculate maximum current for converter single phase to ground fault is not conventional and it does not follow any standard, it is developed in this thesis. This can be an improvement for this thesis work to justify the method used here or as a better improvement, to propose a better method to calculate max current on converter bus and arms currents connected to neutral bus.

Depending on the requirements and project phase, different approaches should be used in order to calculate maximum transient currents. Software simulation gives a better and more accurate result of fault currents. However, setting up the simulation environment takes time. On the other hand, hand calculation gives faster results with higher error.

(41)

9 References

[1] Kundur, Prabha. “Power System Stability and Control” Mc Graw Hill.

[2] Asplund, G.,Svensson, K., Jiang, H., Lindberg, J., Pålsson, R., “DC transmission

based on voltage source converters”,CIGRÉ session, Paris, 1998, paper reference 14-302

[3] M. P. Bahrman and B. K. Johnson, "The ABCs of HVDC transmission technologies," in IEEE Power and Energy Magazine, vol. 5, no. 2, pp. 32-44, March-April 2007. doi: 10.1109/MPAE.2007.329194

[4] Bucher, Matthias Karl. “Transient Fault Currents in HVDC VSC Networks During

Pole-to-Ground Faults”. ETH-Zürich (2014)

[5] V. Jankov and M. Stobart, "HVDC system performance with a neutral conductor,

"2010 International Conference on High Voltage Engineering and Application, New

Orleans, LA, 2010, pp. 188-191.

doi: 10.1109/ICHVE.2010.5640834

[6] “Liquid-Immersed Distribution, Power, and Regulating Transformers” IEEE

C57.12.00-2000

[7] “Short circuit currents in threeohgase ac systems - Part 0: Calculation of currents” IEC 60909-0 second edition 2016-01

[8] C. L. Pires, S. I. Nabeta and J. R. Cardoso, "Second-order model for remote and close-up short-circuit faults currents on DC traction supply," in IET Power

Electronics, vol. 1, no. 3, pp. 348-355, September 2008.

(42)

Analytical approach to dimension transient currents in asymmetric VSC monopole

Analytical approach to dimension

transient currents in asymmetric

VSC monopole

SADRA VARSHOCHI

Abstract

Abstract

Table of Contents

1 Introduction

2 HVDC introduction

3 AC bus three phase fault

4 Converter bus three phase fault

5 Converter bus single phase fault

6 Pole to neutral fault

7 Pole to ground fault

8 Conclusions

9 References