AMAZON multi-meters discounts AMAZON oscilloscope discounts
6. GENERAL HVDC CHARACTERISTICS
6.1 Losses
The loss figure for a LCC HVDC scheme is approximately 0.8% (per end) under full transmitted power conditions, including the link, valves, filters and converter transformers. This is consistent with the 1.6% total converter losses reported by the International Council on large electrical systems (CIGRE) with no load losses about half this figure.
VSC schemes have reported power losses ranging from 1.8% to 3% of converter (per end) full load rating, including the IGBT, filter and interface transformer depending upon the topology employed.
6.2 Economics
The costs of any transmission system depend on a wide variety of factors such as the power rating, converter design, compensation requirements, losses and the type of transmission medium. Coupled with technical issues there is also the difficulty in providing comparable cost figures given the necessity to also take into account the exact nature of the geographic/environmental conditions along the route, land requirements and market fluctuations. CIGRE has how ever analyzed and published manufacturer cost data for LCC HVDC systems, and a typical cost structure is shown in the pie chart in FIG. 16. In general terms LCC HVDC is currently considered economically superior to VSC trans mission schemes in terms of capital cost and lower power losses. However each scheme must be analyzed on a case-by-case basis. The situation is for example complicated by the fact that if converter station land purchase costs are taken into account the required footprint is less for the VSC configuration.
6.3 Reliability, Availability, and Redundancy
HVDC systems require a greater and more complicated number of components than their equivalent AC transmission scheme. This increases the possibility of forced outages arising from equipment failure.
Recent results issued by the CIGRE working group B4.04 on LCC HVDC reliability over a 2005/2006 observation period indicate an average availability of 94% for 2005 and 93.4% for 2006. The data suggests the aver age number of converter station forced outages was 7.5 outages per year in 2006 with an average duration of 62.8 hours per outage. It should be noted that only converter station equipment outages are considered in this data which excludes the transmission line or cable link. This equates to a 2006 mean time before failure (MTBF, see Section 23) for an LCC HVDC system of approximately 50 days.
No general reliability data is currently published for VSC schemes.
However the number of installed VSC installations is much lower than for LCC schemes and comparisons would in any case need to be treated with caution. The minimum energy availability from operation of the Sound Cross and Murray Link (see Section 26.4.2.2) indicates availability in excess of 98%.
Equipment such as thyristor valves, cooling plant, auxiliary power sup plies and control systems usually incorporate a degree of redundancy. For example an additional series thyristor may be added into the phase arm stack of series connected devices so as to achieve this.
As well as the incorporation of equipment redundancy there is also the issue of spares holdings to be determined so as to enhance overall HVDC link availability and avoid long outages associated with failure of long lead time items (reordering delivery times of specialist components).
======
FIG. 16 Cost breakdown of HVDC schemes (courtesy of CIGRE).
Freight insurance 5.00% Erection 8.00% Engineering 10.00% Civil works 14.00% AC filter 10.00% Valves 20% Converter transformer 16.00%
Control 7.00% Other equipment 10.0%
======
7. CONTROL SCHEMES
7.1 Current Source HVDC
Current source HVDC transmission schemes under steady state conditions utilize one terminal to control the DC link voltage and the other terminal to regulate the DC current by controlling its voltage relative to that maintained by the other terminal. The converter DC output voltage may be controlled by:
_ Changing the ratio between the DC and AC voltage by varying the firing angle.
_ Changing the converter AC voltage via an on-load tap changer on the converter transformer.
7.2 Voltage Source HVDC
There are a number of steady state control functions for the VSC a combination of which may be applied to the rectifier or inverter sections of the converter:
_ DC voltage control.
_ Active power control.
_ Reactive power control.
_ AC bus voltage control.
_ Frequency control.
The control functions may be realized through the following:
_ Control of the VSC phase angle (for DC voltage control and active power control).
_ Control of the VSC voltage magnitude which in the case of pulse width modulation (PWM) schemes is realized through manipulation of a signal known as the modulation index (for reactive power control and AC bus voltage control).
_ Direct control of the main valve firing oscillator frequency. Also indirectly by regulating the active power delivered to or taken from the AC system (for frequency control).
8. AC AND DC INTERACTION
8.1 AC System
Strength In a power system comprising AC and DC systems the system performance and the possibility of adverse interaction between the AC and DC systems can be qualitatively assessed using the following factors:
_ Effective short circuit ratio (ESCR).
_ Effective inertia constant.
8.1.1 Effective Short Circuit Ratio
A simple means of measuring and comparing relative strengths of AC systems compared to the connected DC transmission system is the short circuit ratio (SCR) defined as:
...where S and Pd are the short circuit MVA fault level of the AC system and the DC converter MW rating, respectively.
Shunt capacitors including AC filters connected at the AC terminals of a DC link can significantly increase the effective AC system impedance, and hence decrease the short circuit ratio (SCR). To allow for this, the effective short circuit ratio (ESCR) is defined as:
…where Qc is the value of the three-phase fundamental MVAr of the shunt capacitors connected to the converter AC busbars expressed as per unit of the DC link power, Pd, at rated AC voltage. Note that the definition of ESCR is primarily applicable to the fixed and mechanically switched reactive power compensators. For dynamic power electronics based compensators, the controller dynamics and speed of response determine the system stiffness.
Generally speaking the risk of thyristor device commutation failure is increased with lower ESCR, and the resulting over-voltages will be higher.
Traditionally the AC system short circuit strength has been classified as:
_ High (if ESCR is greater than 3).
_ Low (if ESCR is between 2 and 3).
_ Very Low (if ESCR is less than 2).
This initial review provides an assessment of whether there are potential AC/DC system interaction problems and is primarily valid for current source LCC HVDC schemes. It has been shown that CCCs can be applied to relatively weak AC systems with an ESCR of around 1. VSC HVDC converters may be used with AC systems with very low (theoretically zero) ESCRs that is isolated systems with no normal local power generation such as the Valhall offshore oil platform in Norway.
8.1.2 Effective Inertia Constant
The ability of the AC system to maintain the required voltage and frequency depends on the rotational inertia, H, of the AC system as described in Section 1. For satisfactory performance the AC system should have a mini mum inertia relative to the size of the DC link. A measure of relative rotational inertia is the effective DC inertia constant, Hdc, defined as:
Hdc =Total rotational inertia of the AC system ðMW sÞ MW rating of the DC system
An effective inertia constant of at least 2-3 seconds is reported to be required for satisfactory operation.
9. HVDC SYSTEM PERFORMANCE ISSUES
This section describes the uniquely different performance issues associated with links involving current and VSC technologies. Transient performance issues are of a specialist nature and CCC HVDC solutions are little used in practice and so are not described here. The design and manufacture of HVDC converters is highly specialized, but there are sufficient internationally recognized manufacturers (ABB, AREVA, Siemens, etc.), with a strong track record, for competitive tenders for such equipment to be sought. It is however essential that the purchaser concentrates on specifying the system engineering and interfacing the equipment with existing networks rather than on over-specifying particular features within the converter equipment itself.
Worse still would be to cherry pick these features as a mismatch from different vendor's individual equipment literature.
9.1 Commutation Failure
A fundamental characteristic of the thyristor is that for successful turn-off; the voltage across the device has to remain negative for a minimum period of time after extinction of its current so as to remove stored charge and to allow the valve to become capable of blocking the forward voltage. Should the voltage across the device become positive prematurely the device may turn on again _ even without a firing pulse. This will result in failure of the commutation process. For example under these conditions, the converter bridge circuit will act as a short circuit on the DC side and transmitted active power becomes zero for that bridge. The negative valve (thyristor) voltage is provided during the time corresponding to the extinction angle _ referred to as the commutation margin. A larger extinction angle minimizes the risk of commutation failure. The risk of commutation failure is increased with low, or very low, SCR schemes. Single or successive commutation device failures may be due to:
_ abrupt changes in magnitude or phase angle of the commutation voltage;
_ a sudden increase in DC current; and/or
_ malfunctioning of firing control.
Current source LCC systems are designed such that single failures should not cause any adverse impact on the interconnected AC systems and the commutation may restart even before full removal of the disturbance (typically within two cycles). Under severe voltage depressions the commutation may only recommence once the disturbance is removed. The recovery from a commutation failure may be prolonged if the valve design is such that the valve firing circuits are blocked at some preset undervoltage levels and only deblock on return to normal voltage conditions. Commutation failure due to firing circuit malfunction normally results in a forced outage and even longer disruption of transmitted power flow. Commonly used commutation failure mitigation techniques are as follows:
_ Temporary increase of inverter extinction angle by 10_12_ before AC switching operations or immediately after fault inception.
_ Temporary increase of rectifier firing angle during disturbances on the rectifier network.
_ Voltage dependent current order limiter (VDCOL) which reduces the DC current order, and hence the reactive power consumption upon reduction of the AC system voltage.
_ The use of fast acting reactive controllers such as synchronous condensers and static VAR compensators (SVCs) to help alleviate the risk of commutation failure.
VSC schemes employ devices with inherent turn-off capability and there fore commutation failure can normally only occur in these cases under loss of gate firing control.
9.2 Temporary Over-voltages
In AC/DC systems with low ESCR, high over-voltages can be experienced during converter blocking, fault clearing or other disturbances. Temporary voltage peaks can be created at the converter stations by the superposition of the fundamental frequency and low order harmonic voltages. The fundamental frequency component of the over-voltage is due to a mismatch between the converter station reactive power supply and the instantaneous reactive demand of the AC and DC systems.
It is reported that the most onerous condition for over-voltages occurs following a close-up three phase fault -- assuming that the converters are blocking during the fault, and with the shunt reactive compensation left on. This condition produces full magnetizing inrush current on all transformers after removal of the fault which in turn results in substantial fundamental and harmonic over-voltages. This over-voltage condition may be the determining factor for valve, surge arrestor and insulation voltage ratings. Maximum reported design values for power frequency over-voltages associated with LCC HVDC systems range from 1.075 pu to 1.5 pu. The majority of designs installed to date are designed for temporary over-voltages in the range 1.25-1.4 pu. For systems with SCRs in the range 3_5, the maximum temporary over-voltage likely to be encountered is 1.35 pu if the total HVDC trans mitted power is blocked due to a fault on the HVDC transmission line.
Temporary current source converter over-voltages may be attenuated by different means such as:
_ additional low order harmonic filters;
_ modified HVDC control systems;
_ generator excitation control systems;
_ synchronous condensers, SVCs or static compensators (STATCOMs);
_ fast switching of capacitors and AC filters connected to inverter and/or rectifier sides of the converter; and
_ detuning and retuning AC harmonic filters to low order harmonics.
Modern LCC HVDC systems can be in principle designed such that temporary over-voltages do not impose a limit on the speed of fault recovery.
The reduced size of the necessary harmonic filters and shunt capacitors used in a VSC system helps reduce over-voltages experienced during AC or DC system disturbances. This minimizes the need for the numerous over voltage mitigation measures discussed earlier. In addition, the reduced level of non-characteristic harmonics produced with these systems reduces the occurrence of harmonic resonances. A significant scenario where over-voltages may occur is reported to be when the VSC is absorbing a significant amount of reactive power prior to ceasing operation.
9.3 Response to AC Faults
Balanced or unbalanced AC system faults in proximity to current source LCC systems will cause voltage dips and temporary reduction of DC power transfer. Depending upon severity, such voltage dips could cause commutation failure of some or all connected valves. At the LCC HVDC rectifier side, although in theory operation can occur at very low voltages without commutation failure, excessive reactive power requirements in these situations may result in converter blocking under special control measures. The basis for deciding the recovery control sequence to follow is normally derived from the ability of the AC system to survive the resulting disturbance and the need for power change within a specified time. These are highly dependent upon the maximum permissible duration (critical clearing time) of faults.
There are two main control strategies:
1. To recover the power as fast as possible (however the required reactive power consumption could result in the reduction of AC system voltage and the possibility of commutation failure as discussed in Section 9.1).
2. Temporarily block the inverter if the voltage falls below a certain level (after the fault clearing the converters are deblocked and the current is ramped back up under a predefined ramp rate).
Practical LCC HVDC installations are typically designed to recover 90% of their pre-fault power following an AC fault in the range 25_500 ms (most have a range of 100_200 ms).
The ability of a VSC to generate its own reactive power, and the practical elimination of commutation failure risk, can significantly improve AC fault performance over the HVDC link and allow for a faster recovery. However IGBTs used in HVDC schemes have a much lower over-current capability than LCC HVDC thyristors. Some VSC designs therefore use a control sequence to shut down the converter until the fault is cleared and thereby protect the IGBTs and freewheeling diodes from excessive over-currents.
9.4 Response to DC Faults
DC line faults are almost always single pole to ground faults which blocks the power transfer on the faulted pole. In practical HVDC schemes time is allowed for deionization of the air surrounding the arc prior to a restart. The typical time for an OHL fault clearing and return to rated power transfer is reported to be in the order of 200_300 ms. The recovery time for DC links connected to weak AC systems is longer. It is common practice to apply a staged recovery that has the effect of artificially increasing the effective short circuit ratio (ESCR) in this situation.
Sustained DC faults will lead to blocking of the valve group and usually shutting down the pole. Restoration is normally achieved under operator control. Some schemes allow isolation and bypassing of the faulted zone so that the remaining sound groups may be restored to full power. However the time taken to achieve this is often too long (i.e. several seconds) in comparison with transient stability or generator ride through capability at the AC ends.
With VSCs the anti-parallel diodes associated with the IGBT devices will allow current to flow indefinitely even when the IGBTs are blocked. The occurrence of DC line faults therefore requires opening of the AC circuit breakers at each end of the link in order to clear the fault (unless appropriate HVDC breakers which are not normally commercially available, are provided).
In the case of internal converter faults the VSC protection systems should isolate the faulty element and shutdown the HVDC transmission system.
Restoration therefore takes relatively longer in a VSC scheme (in the order of 10 seconds) as the DC capacitors have to be recharged to rated voltage.
In the context of HVDC schemes associated with cable systems any DC line faults would be assumed to be permanent failures requiring investigation prior to restoration of rated power flow. The total time to return to rated power transfer in such schemes will not therefore be limited to the current or VSC control characteristics.
9.5 HVDC for System Performance Support
Sections 24 and 25 refer to power quality disturbances and the purity of the voltage and current waveforms. In general an HVDC link has the ability to recover from disturbances and allow the AC system to reestablish its synchronizing power. Its ability to accommodate a controlled fast change in power can make it possible to arrest large fluctuations in frequency by matching the generation to the load in the area to which the DC link is connected. This feature can thus be used to maintain the system transient stability during contingency conditions. A fast change of power by DC control can also be used to damp electromechanical oscillations for nearby generators. In either case the AC systems connected to the HVDC link would need to accept the changes in DC link power without adverse effects.
In practice the contribution of an HVDC link to system stability is usually made in the form of a power modulation control or emergency power control. These controls (often referred to as remedial action schemes, RAS) typically involve a ramping up or down of DC power, discrete power changes or DC power reversal, in response to local or remote disturbances such as changes in generator rotator speed or busbar frequency. Functions such as generator inter-tripping may also form part of a RAS where power change functions may be insufficient. The RAS is activated by a signal which best indicates the occurrence of the disturbance and the necessity for DC power link changes. This parameter may often be associated with system frequency, although for LCC HVDC schemes, sensitivity to frequency has to be deliberately added into the control system.
For faults that may cause rotor instability, it can be beneficial to increase the post-fault rectifier side DC link power in response to generator phase angle. The thyristor valves used in LCC HVDC links are rated to withstand considerable overloads without adverse effects. In practice this feature can be used to increase rectifier DC link power in post-fault operation in response to generator phase angle changes, and thereby reduce risks of rotor angle instability. Typically practical schemes may apply modulation of 20-40% of the DC link rating for this purpose. However schemes with temporary overloading of as high as 67% have been designed.
The ability of a VSC HVDC scheme to accommodate fast changes in either active or reactive power (subject to design limitations and AC system constraints) can improve its capability to provide transient stability beyond that of a LCC HVDC link. For example with power reversal time as an example, the VSC is relatively faster, it can operate without telecommunication delay, it does not require voltage polarity reversal and it avoids reactive power consumption during operation. However as mentioned before currently available voltage source schemes have lower inherent over-current capability compared to current source schemes. Therefore special short term overloading as a means of improving system stability would require this special VSC feature to be clearly specified.
9.6 Fault Current Contribution
A current source LCC HVDC scheme provides practically no fault current contribution. This is primarily due to the use of a voltage dependent current magnitude limiter control function which reduces the current order during induced voltage dips. In addition thyristor based LCC schemes cannot supply fault energy into the network.
In contrast VSC schemes make some contribution to fault current -- albeit significantly lower than that achieved with an equivalent AC interconnector. It is reported that VSC fault contribution varies in inverse proportion to the SCR up to a maximum determined by the SCR and the rated current of the converter. The amount of contribution depends on the control strategy and operation modes used.
9.7 Sub-synchronous Resonance
Sub-synchronous torsional interaction (SSTI) occurs if HVDC control systems adversely interact with the rotating masses of nearby turbine generators.
This could excite a turbine generator torsional mode of operation and cause damage to the associated mechanical drive train. The phenomena usually occur in the frequency range between 4 Hz and 40 Hz. It has been shown that special DC power modulation control for damping low frequency electromechanical oscillations between sending end and receiving end systems, typically in the range 0_4 Hz, can introduce significant sub-synchronous torsional activity. Modern LCC HVDC scheme design takes this now well understood phenomena into account using supplemental sub-synchronous damping controls (SSDCs) so as to avoid the risk of SSTI.
SSTI is sensitive to many system parameters such as transmitted DC power the electrical distance between the generator and converter station, and the relative strength of the individual generator to that of the total AC system. The highest interaction occurs with the weakest AC systems. The SSDC must therefore be designed to accommodate a wide range of changes in network configuration.
10. EMERGING TRENDS AND TECHNOLOGIES
10.1 Ultra-High Voltage (UHV) DC
An exciting application of HVDC schemes is the connection of large (B6,000 MW), but remotely located (B2,000 km), hydro schemes to major load centers. Power losses in the DC link reduce with the inverse of the square of the voltage (P/V2). Improvements to the economics of the DC link (which currently use voltages in the range of 500-600 kV) are being employed by using 800 kV UHV technology. Active areas of research include the effect of pollution on DC insulation, converter transformer reliability, corona performance, air insulation design and various other aspects including the availability of suitable test facilities.
10.2 Offshore DC Grids
Very large offshore wind farms (up to 10 GW) are being planned worldwide.
These benefit from DC submarine cable links to the receiving grid so as to overcome the issues discussed in Section 2.1. VSC HVDC is the currently preferred technology (e.g. Troll A and Valhall offshore oil platforms and BorWin 1 offshore wind farm) for such reasons as footprint, black start capability, grid code compliance issues and suitability for multi-terminal connection. Issues to be overcome include the following:
_ The use of multi-terminal configurations for connection of wind farms to two or more AC systems and realization of the associated control schemes.
_ The need for high speed DC breakers for fast fault isolation, clearance and grid restoration.
_ The need for more efficient and compact converter configurations.
10.3 DC Circuit Breakers
Very reliable lower voltage DC breakers are an established and mature technology in the traction industry. Section 13 describes the use of the natural current zero in AC circuit breaker design. HVDC circuit breakers are difficult to build because of the absence of this effect and the requirement for supplementary active or passive devices that force the current to zero. The most common DC circuit breaker configurations are as follows:
_ Conventional AC breaker in parallel with a resonant circuit and metal oxide varistor.
_ A solid state breaker comprising a bidirectional switch based upon IGBT and diode technology.
_ A hybrid DC breaker comprising a conventional AC breaker in parallel with a solid state breaker.
HVDC breakers have not yet been developed for fault current interruption. However in practice DC breakers have been applied for applications where DC loops must be opened to bypass converters.
10.4 Novel Circuit Topologies
A new VSC HVDC converter topology is being developed by AREVA using the concept of power electronics building blocks (PEBBs). A chain link-type circuit similar to the MMC is deployed to generate a controllable waveform from a segmented DC capacitor as shown in FIG. 17. The converter rating may be increased by scaling up the number of chain links.
FIG. 17 PEBB VSC HVDC arrangement using a chain link circuit.
10.5 Emerging Semiconductor Switching Devices
Valve semiconductor switching devices currently follow either thyristor or transistor-like, silicon based structures. Thyristor devices are turned on by low power current pulses to the gate, whereas transistor-like devices are turned on by a voltage pulse. Current switched devices include thyristors, gate turn off thyristors (GTOs) and insulated gate commutated thyristors (IGCTs). Voltage switched devices include insulated gate bipolar transistors (IGBTs) and injection enhanced gate transistors (IEGTs which have lower losses than IGBTs and relatively higher ratings up to some 4.5 kV and 4 kA). The maximum power ratings of GTOs and IGCTs are lower than those for a thyristor _ typically in the range 6 kV and 6 kA. GTOs are not expected for use in HVDC converters because of their higher cost and losses.
IGBTs have a lower power range than GTOs/IGCTs.
Significantly increased device voltage ratings and switching speeds appear to be possible by using a combination of semiconductor materials (e.g. silicon carbide and gallium nitride) with engineered band-gaps.
However the difficulty of reliably manufacturing such devices must not be underestimated. A recent forecast of the future voltage ratings of silicon (Si) and silicon carbide (SiC) devices is shown in FIG. 18.
10.6 Active Filters
Power electronic devices may be used in combination with software algorithms to generate active filter responses and avoid the need for passive capacitor/ inductor/resistor filter circuits. This is common practice in modern electronic radio circuitry and has the potential, when applied to power converters, to greatly reduce losses and station footprint requirements in comparison with equivalent passive filtering schemes. Active AC and DC filters have been installed in practical LCC HVDC schemes. The active DC filters mitigate harmonic currents on the DC OHL link and the active AC filters are also available to improve reactive power exchange with the AC grid and to improve dynamic stability. In all known applications the active filters are currently placed in series with passive filters so as to form a hybrid. This minimizes the power rating of the active filter and maximizes the overall full filter performance.
FIG. 18 Anticipated voltage ratings of silicon (Si) and silicon carbide
(SiC) switches (CIGRE).