Modern Applications of Power Electronics to Marine Propulsion Systems C G Hodge BSc MSc CEng FIMarEST Rolls-Royce plc, PO Box 3, Filton, Bristol, BS34 7QE, United Kingdom Phone: +44 (0) 117 979 5508, Fax: +44 (0) 117 979 6424 E-mail: c.hodge@blueyonder.co.uk Abstract – This paper reviews the current state of marine electric propulsion systems and the various system developments currently being pursued to achieve greater power density and explains why these developments are economically and operationally desirable. The paper also reviews the specific power electronic converter applications that are supporting the system developments with particular reference to converter design. The paper concludes by considering the likely future direction of development for power electronic converters. Indexing Terms – Electric Warship, More Electric Applications, Integrated Full Electric Propulsion. I. INTRODUCTION Marine propulsion systems are undergoing rapid development with a significant thrust towards the use of electrical propulsion and the integration of auxiliary and propulsion power systems. Whether the vessel in question is a large cruise liner driven by external podded motors, a small highly powered warship with inboard motors and shafts or an off-shore oil support vessel the need for very flexible, robust and compact power systems is ever increasing. This need is being satisfied through the recent advances, verging on revolution, in power electronic device construction and their application for power conversion and control. The recent developments in switching device implementation have been fundamental to the developments underway in marine electrical power systems. AS well as the Insulated Gate Bi-Polar Transistor (IGBT) Devices such as the Integrated Gate Commutating Thyristor (IGCT) and MOS Turn Off Thyristor (MTO) are beginning to be considered for use in marine electrical power systems. However, to allow space for discussion of the developments also underway in converter topology and control, this paper will not discuss these devices further. II. INTEGRATED FULL ELECTRIC PROPULSION A traditional mechanical marine propulsion system uses prime movers to drive propellers via shafts and in the most common arrangements each shaft is driven by its own dedicated prime mover. Unless complicated combining gearboxes are used it is not possible to cross-connect the shafts. The variation of propulsion power with speed is – to a first approximation – a cubic relationship. As a result 80% power provides 93% speed but 25% speed demands only around 1.5% power. Since many vessels rarely use full speed this results in the prime movers running for lengthy periods at fractional loads that reduce efficiency and increase operating costs. Electric propulsion – which is perhaps more accurately described as electrical transmission (between the mechanical prime movers and propellers) – offers a relatively straight forward method of cross connecting shafts in order to alleviate the light load running of the prime movers when the vessel operates at reduced speed. Figure 1 illustrates a conventional X Electrical System normally run split DG X DG X GT GB GT Shafts cannot be cross-connected GT GB GT X X DG X DG Figure 1: Mechanical Propulsion System mechanical propulsion system for a vessel with 2 shafts and an equivalent electrical propulsion system. The advantage of the latter is that a single prime mover can provide power to either shaft and – with the adoption of modern power electronics – astern operation and ship manoeuvring is possible without recourse to controllable pitch propellers or complicated reversing gearboxes. A further advantage of electrical propulsion is that the electrical generators need not be limited to supplying propulsion power alone – they can also be arranged to provide the electrical power required for the ship’s domestic and other uses. This auxiliary electrical power system is often termed the ship’s service power system and by combining this with the electrical propulsion system a common set of prime movers can provide power to either the propulsion or ship’s service duties. This arrangement is termed an integrated power system. When the electrical propulsion system is rated at full power (with no direct drive mechanical boost arrangement) the overall system is designated as Integrated Full Electric Propulsion (IFEP), figure 2 illustrates such an IFEP GEN CONV 20 MW GEN 2 MW CONV CONV CONV PMPM mass as desirable characteristics. There are several programmes across the world aimed at developing compact direct drive propulsion motors, notably in the United States of America, France and the United Kingdom however these programmes – though fascinating – are outside the scope of this paper. The need for compact and efficient equipment applies equally to the power electronic converters that are used to drive the propulsion motors in IFEP installations and the developments underway in this area will now be reviewed. IV. GENERAL ASPECTS OF MARINE POWER ELECTRONIC CONVERTERS Devices Used: Historically the converters used in marine electrical propulsion systems have been based on thyristors and these applications have obtained their commutating voltages either from the supply-side voltage source (as in cycloconverters) or from the motor back-emf (as in current source converters). Although this – as a type of natural commutation – brings high conversion efficiency such thyristor-based converters are in general too large for use in space sensitive applications and in addition each have significant (though differing) limitations on their frequency output, control and waveform quality. As a result current source converters and cycloconverters are now becoming displaced by voltage source Pulse Width Modulated (PWM) Converters that use devices capable of switching current off as well as on – such as IGBTs. Pulse Width Modulation: Figure 3 illustrates the construction of a PWM pulse train that would provide an effective sinusoidal voltage to a load. As with all PWM strategies the width of each pulse and succeeding “gap” are controlled to present the correct average voltage to the load. In the simplest strategy – as illustrated in Figure 3 – a saw tooth switching 300 200 BATTERY CONV CONV CONV PMPM GEN 8 MW GEN 800 V DC 6.6 kV AC 20 MW Figure 2: IFEP Propulsion System system. Integration of the two power systems can reduce the total number of prime movers that need to be installed – bringing reduced vessel acquisition costs – and where the ship’s service load is a significant proportion of the propulsion requirement, additional fuel savings can be achieved by using the ship’s service power as a base load for the prime movers – thereby ensuring that the engines are loaded sufficiently to retain fuel efficiency. Given their high hotel load, this latter aspect is the main reason why cruise liners have all but universally adopted IFEP in recent years. Even simple reduction gearboxes are generally avoided with IFEP installations as direct drive motors are feasible and thereby the additional purchase cost, maintenance burden and inefficiencies associated with a gearbox are avoided. III. THE IFEP EQUIPMENT REQUIREMENT Ships are always limited with respect to internal space and reduced equipment volumes have direct benefits in providing capacity for other purposes, not least additional cargo, passengers or in the case of warships weaponry. One benefit of electrical propulsion is that it provides the opportunity to move the electric motors towards the stern of the vessel leaving the higher value central hull volume for use directly associated with the vessel’s purpose. However, for hydrodynamic reasons, the stern of a vessel reduces in cross section and thus, as a result of its reduced displacement, the stern provides less uplift than other sections of the hull form. Consequently, equipment placed towards the stern needs to be relatively light if unacceptable stresses are to be avoided within the hull. Therefore marine electrical propulsion motors have both low volume and 100 Vol ts 0 0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018 -100 -200 -300 Time (s) Desired Voltage Saw Tooth PWM Figure 3: PWM Waveforms function is used to determine when the switches operate. Series Connection of Switches: The initial problem with PWM converters for marine propulsion use was that the rating of the available power switches were insufficient to provide the necessary power, which can often be 20 MW per shaft or higher. As described later, Alstom was the first company to overcome this problem and developed highly rated converters using series connected devices. One consequence of the use of series connected IGBTs is that the IGBT temperature must be tightly controlled in order to preserve the switching characteristics of each IGBT. Voltage Clamped Switching: As IGBT ratings have increased other topologies that avoid the need for series connection have become feasible; such as Neutral Point Clamped (NPC) Converters. Figure 4 shows a +V DC the losses generated in the capacitor network as it increases in size. As a result a 5 level diode clamped converter probably represents the limit of this topology. V. CONSTRAINTS OF MARINE POWER ELECTRONIC CONVERTERS Harmonic Distortion: Any converter’s prime function is to transform power between differing systems, AC to DC is perhaps the simplest converter (a DC rectifier) but even that shares a characteristic with all other converters – it switches current and voltage. The switching action allows the most favourably suited input phases to be connected to the appropriate output lines such that the desired form of output power is provided. Again in the case of the rectifier this simply means connecting the most positive input phase to the positive DC rail and the most negative of the input phases to the negative DC rail – this can be achieved passively through the use of diodes but nevertheless a switching process remains which results in discontinuous currents being drawn from the supply. As the current drawn from the supply is discontinuous it is – by definition – non-sinusoidal and it therefore creates non-sinusoidal voltage drops in the supply system. These non-sinusoidal voltage drops mean that the supply voltage (as seen by the converter itself and other connected equipment) is no longer sinusoidal – the converter has distorted the supply system. This feature applies to all converters: rectifiers, cycloconverters, current and voltage source converters. The most common form of distortion is that which results from DC rectification since this is a feature of the rectifiers found as the front ends of current and voltage source converters, however cycloconverters also exhibit supply side distortion and of a more complicated form. Since the supply system voltage and load current waveforms (although no longer sinusoidal) remain periodic, Fourier analysis can be used to decompose them into spectra of frequency components – each an integer multiple of the fundamental frequency. The non-fundamental frequency components of the supply voltage and load current are termed harmonics and the overall distortion is termed harmonic distortion. It is of note that current harmonics are the cause and voltage harmonics the affect. Thus in order to prevent voltage harmonics interfering with the operation of parallel-connected equipment, the current harmonics must be reduced to an acceptable level by filtration. Figure 5 illustrates the frequency components of the current waveform that result in the supply system that feeds a DC rectifier. Because voltage harmonic distortion arises through voltage drops created in supply system impedances, high impedance systems – such as those found in marine use – are more 0V DC One phase output connection -V DC Figure 4: NPC Topology simplified schematic of a NPC topology. As can be seen an artificial neutral voltage – obtained by a capacitor ladder bridge – clamps the inter-device voltage via diodes such that the maximum voltage seen across the switches is half the circuit maximum. If the converter’s devices are able to withstand half the circuit voltage then any need to switch them in synchronism is avoided. At present converters rated at 30 MVA are achievable with the NPC topology. As well as providing a robust converter design NPC converters are also inherently more efficient as the devices switch through lower voltage differentials and thus the switching loss is reduced. A final advantage – again accruing from the reduced voltage swing – is that the output waveform more closely approximates to a sinusoid; it contains a lower harmonic content. It is of course not necessary to limit the clamping voltage to the mid point of the DC link voltage alone, more extensive capacitive networks can generate any desired number of intermediate voltages which can then be used to asynchronously operate any number of switches in a converter limb safely, the converter is then termed a diode clamped converter. However the advantages offered by increasing the number of levels offers reducing returns with respect to the quality of the waveform and the efficiency increases gained by the reduction of the switching loss are rapidly overtaken by susceptible to voltage distortion and one of the perennial challenges for the marine power system integrator is to predict the levels of harmonic distortion at the design phase so that suitable counter measures such as harmonic filters can be incorporated in order that the supply system retains an adequate level of power quality. The size of harmonic filtration required in any given marine power system is clearly a system level issue but it is rarely insignificant. For a typical IFEP system with the ships service system interconnected to the propulsion system by simple transformers the harmonic filters will occupy around 60% of the volume of the propulsion converters themselves. This adds weight and cost and reduces overall system efficiency. Hence the harmonic performance of any given converter design is a crucial issue. dV/dt: The output of a converter is often used to drive a propulsion motor and here – particularly for compact marine propulsion – the performance of the motor insulation under the PWM waveform generated by the converter is crucial. The motor’s winding insulation needs to withstand the sequence of steep sided pulses of voltage emulating the required sinusoidal voltage if failure is not to result. The most crucial aspect of the voltage waveform with regard to insulation stress is the rate of change of voltage and PWM converters are capable of extremely high rates of change of voltage. This aspect of insulation degradation and potential failure is often, on first thought, found to be surprising. Nevertheless insulation failure is one of the commonest reason for failure in the several compact propulsion motor developments that are taking place in both Europe and the United States of America. It currently appears that a peak voltage of around 5kV is a maximum –at least empirically. This limit also sets the DC link voltage in a PWM converter and therefore the United States medium voltage standard of 4.16 kV is attractive to marine system designers as it provides around 5 kV when rectified. Because the rate of change of voltage as seen in a converter’s output is a key contributor to the stress and resultant degradation of the motor insulation many converters use output filters to reduce their rates of change of voltage. Occasionally these output filters are designed to remove the majority of the output harmonics and they may then be termed sinusoidal filters. However, more commonly, the output filter is designed only to limit the rate of change of voltage to a specified level and the filter is then termed a dV/dt filter. dV/dt filters are naturally much smaller than sinusoidal filters although the exact size – and thus their cost and the amount by which they reduce overall system efficiency – is one of the key system design considerations. Motor Regeneration: Because marine electrical propulsion systems drive vehicles that gain kinetic energy as they move, the maneuvering of ships is complicated by the need to dissipate the stored kinetic energy when reducing speed. The interaction of the propeller with the slipstream of the water passing it further complicates the issue. In general when wishing to reduce speed, simply reversing the direction of rotation of the propeller, or even just stopping it, is never the best strategy. The best thrust in the opposite direction to the ship’s motion is achieved by turning the propeller at just less than synchronous speed – that is at just less than that rotational speed which would match propeller advance and ship speed when considering the propeller as a simple screw. When this is done, as well as the thrust reversing so does the torque on the shaft and therefore so does the power flow from the motor that starts to regenerate power back into its converter. The problem of controlling and dissipating the regenerated power during maneuvering is a significant issue for the system integrator in the early phases of the design and it usually required as dedicated dynamic braking resistor and controller so that the significant regenerated power does not destabilize the upstream power system. Switching Frequency and Switching Loss: The higher the switching frequency the better the quality of the voltage waveform supplied to the load and thus the current it draws. This has a direct impact on noise and vibration and also on the filtration requirement of the converter output. Ideally the marine propulsion system designer would like to operate at the switch’s maximum frequency. However each switching cycle represents an energy loss as the voltage across the switching device builds up at the same time as the current decays and thus there is energy dissipated 2.5 2 1.5 1 0.5 0 0 5 10 15 20 25 30 35 -0.5 Time Volts Current Power Energy Figure 6: Switching Loss within the switch. In order to limit the switching loss current-day marine propulsion power electronic converters are limited to around 2kHz – far below the theoretical capability of modern IGBTs. Figure 6 illustrates the switching loss inherent in a forcecommutated device. Reduction of switching loss is one of the ambitions of current converter development. VI. EXAMPLES OF MODERN MARINE POWER ELECTRONIC CONVERTERS Alstom Series Connected Converter: Alstom developed the first high power marine power electronic PWM converter based on series connected IGBTs to provide the necessary voltage withstand for a 25MVA Disk 3 Outer Supply Bridges Filter Filter Filter Filter Inner Disk 4 Outer Inner Power Paragon (a part of L3 Communications) which uses 5 intermediate voltage levels (again generated by a capacitor bridge). DC Link SIGMA 1 AC Supply SIGMA 2 Channel 1 Channel 2 Figure 9: Power Paragon Diode Clamped Converter Filter Outer Disk 1 Filter Inner Filter Outer Disk 2 Filter Inner Figure 7: Alstom Converter requirement. The main technical challenge was synchronizing the switching of the IGBTs so that the situation where insufficient devices were attempting to withstand the system voltage was avoided and also providing robust voltage sharing circuits so that each IGBT was acceptably stressed. The general modular topology of the Alstom Converter is shown in Figure 7. The Alstom converter development has been a significant success and has been used by both the United States of America and the United Kingdom within their current Naval programmes. Voltage Clamped Converters: The NPC topology has been developed by ASI Robicon of the United States of America and also by PMES of the United Kingdom. Figure 8 illustrates a NPC converter. An NPC converter is, of course, the simplest case of a multilevel converter using three voltage levels and Figure 9 illustrates a Diode Clamped converter developed by DCPOS Isolated Voltage Converter: Power Paragon and ASI Robicon (a division of the USA based High Voltage Engineering Corporation) have each conceived a type of multi-level converter that does not rely on a capacitor network to generate intermediate clamping voltages. Instead a transformer with a single primary winding and multiple secondary windings is used to create electrically isolated voltages which can then be independently switched by IGBTs or other power switches which are nominally series connected on the output side. Although the converter tends to be relatively large the topology does offer the possibility of generating extremely accurate sinusoidal output voltages through the use of interleaved switching across the multiple modular switches. The converter topology exhibits an excellent harmonic distortion performance and this can offset its larger volume through the reduced need for harmonic filtration. VII. FUTURE MARINE POWER ELECTRONIC CONVERTERS Resonant Converters: Resonant Converters are a class of converters that offer great benefit for the future. The concept is that an auxiliary circuit around the main power switch creates a resonant flow of current in a circuit that forces either the switch current or voltage to be near to zero during the commutation interval. In this way the switching loss is largely removed because the volt-amp product is also near to zero. When the current is held near to zero the circuit is said to be current resonant, conversely when the voltage is held near to zero the circuit is said to be voltage resonant. Current resonant circuits resemble the external commutation circuits used with thyristors – and indeed they provide the same function, the load current is diverted out of the switch. Voltage resonant circuits resemble passive IGBT Gate Drive PEC IGBT Gate Drive PEC IGBT Gate Drive PEC IGBT Gate Drive PEC OUTPUT IGBT Gate Drive PEC IGBT Gate Drive PEC IGBT Gate Drive PEC DCNEG IGBT Gate Drive PEC Figure 8: PMES NPC Converter snubber circuits – the main difference now being that the voltage resonant circuits operate in advance of the switching transient whereas the snubber circuit follows behind. In both cases the timing of the firing of the auxiliary resonant circuit is crucial in relation to the timing of the commutation of the main switch. Once the switching loss is largely removed the designer has a choice between increasing converter efficiency, increasing converter switching-frequency, increasing converter current output (so that the conduction loss rises to replace the reduced switching loss) or perhaps a combination of the three. There is however a problem with resonant conversion for marine propulsion uses. As previously explained marine propulsion systems operate for considerable periods at part load and in this condition the converter’ switching loss is significantly reduced anyway. However the additional passive circuit losses introduced by the resonant circuitry remain at the same level as if the converter was at full power. Thus a marine propulsion converter using resonance can actually increase overall losses rather than reduce them. One answer is to design the resonant circuits so that they can be switched out of operation once their losses outweigh the savings – however this can only be achieved for current resonant converters and these are the more bulky and expensive version of resonant converter. SPCo Development: An innovative form of multi-level converter is being developed by SPCo of the USA where MOS Turn Off Thyristors (MTO) switch a base voltage between various intermediate levels and this is then used as a PWM voltage source by a final resonantly switched IGBT based output stage. This converter ought to offer improved efficiency since both the switching and conducting losses are significantly reduced. The switching losses are reduced because the MTOs controlling the base voltage switch infrequently and the IGBTs performing the final PWM modulation switch resonantly. The conducting losses are reduced because the MTOs – which replace IGBTs in comparable topologies – are thyristor based and therefore exhibit a very low forward conducting volt drop. Figure 10 illustrates the SPCo converter. Matrix Converter: Another form of converter under development is the Matrix Converter and this is receiving significant attention in both the USA and UK with much work being conducted by the UK’s Nottingham University as described at [1]. The Matrix Converter gains it name from the appearance of the converter’s electrical connections. When drawn as shown in Figure 11 the nine bi-directional switches 3 Ph Input Bi-Directional Switches 3 Ph Output Figure 11: Matrix Converter appear as a matrix of connections. The bridge produces any desired output waveform by connecting each of the input phases (VA, VB, and VC) to each output phase (Va, Vb, and Vc) in turn within a switching interval of fixed duration. Within the fixed switching interval the periods of output-connection for the three input phases is varied in a manner similar to PWM such that the desired output is produced in terms of the average of the input voltages. In order that any desired output voltage can always be produced without regard to the instantaneous phase of the three input voltages the output voltage must be limited to a level below the input voltage envelope. As can be seen from Figure 12 300 200 100 Soft-Switching Section Volts 0 0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018 -100 -200 -300 Time (s) IGBT 3300-V 1200-A Figure 12: Matrix Converter Voltage Envelope (1) 7-kVDC 5-Level Super GTO 5000-V 3000-A Super GTO 2500-V 1500-A Figure 10: SPCo Converter the input voltage envelope restricts the output voltage to 50% of the input. The minimum of the maximum (positive) is half of the peak value, occurring when the two positive voltages are equal. Similarly the maximum of the minimum (negative) is half of the peak value, occurring when the two negative voltages are equal. This is a serious limitation for the Matrix Converter when it is intended for use in a standard power system with standard motors. However it may be overcome by using one of two differing techniques. The first is to distort the desired output phase voltages by use of a combination of third harmonics of both the input voltage frequency and the desired output frequency. The required additional harmonics are: Input Voltage Third Harmonic: -1/6 Output Voltage Third Harmonic: +√3/6 When this is done the phase voltages are as shown in Figure 13 and the peak desired output voltage may be raised to 0.866 of the input. Thus the maximum output 300 200 100 0 0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018 -100 -200 -300 Time (s) Figure 13: Matrix Converter Voltage Envelope (2) voltage can be increased from 50% to 86.6% an increase of 36.6 percentage points or 73.2%. Two important points are worth noting with regard to this technique. First, although the phase voltages are distorted this does not appear in the output line voltages because of the natural third harmonic cancellation that occurs in a balanced three-phase system when moving from phase to line voltages. Second, this additional voltage increase has not been achieved at any other performance aspect of the converter, its overall efficiency remains unaltered as does the quality of the output voltage and the level of distortion imposed on the input system. The second technique uses a fictitious DC link and divides the modulation function into two stages, the first deals with input voltages, and the second with the generation of the output voltages. It is termed a fictitious DC link because in order to maximise the output voltage the first modulation stage chooses to work between the most positive and most negative input voltages, as would a standard uncontrolled DC rectifier. The overall maximum voltage output that can be achieved with this technique is 6√3/π2 or 105.3%. Although the modulation function is developed by considering two distinct phases of time-continuous operation this cannot be the case when the modulation function is applied to the converter – there is only one switch between the input and the output and no intermediate stage. In practice the two separate switching strategies are combined into one unified requirement before the modulation function is applied. Techniques to achieve this are covered at [2] and [3]. There are however significant disadvantages. Although this technique overcomes the voltage limitation of the Matrix Converter completely, this gain is bought at the expense of the overall harmonic performance of the converter: the harmonics reflected into the input voltage or generated in the output current (or indeed both) must increase. It is perhaps not surprising - since the modulation function is in part derived through consideration of DC rectification - that the largest increase in harmonics are those associated with this process itself (5th, 7th, 11th etc). An additional source of harmonic distortion arises when the voltage is forced above 86.6%, when this is done it is no longer possible for the average output voltage in each switching interval to equal the target voltage and additional low frequency harmonics result. It is interesting to note that a Matrix Converter operating at minimum harmonic distortion under a fictitious DC link scheme will be subject to the same voltage limit as a converter operating under the third harmonic strategy. There are practical difficulties in developing the overall modulation function through two stages in real time so that implementation of a converter producing continuously varying frequency and voltage (as required in IFEP applications) would be difficult. Clearly a Matrix Converter would benefit from a symmetrical bi-directional switch. SAIC Multi-Port Inversion: SAIC – and in particular Dr Rudy Limpaecher – have been developing an extremely innovative converter termed the Multi-Port Converter. The circuit schematic diagram for an MPC converter is shown in Figure 14 together with the current waveforms for one charge and discharge cycle at Figure 15. The technique has been more fully explained at [4], [5] and [6] but, in summary, it utilises Input Redistribution Output Volts Filter Figure 14: SAIC Multi-Port Converter Filter the fact that in any balanced multi-phase system the Ici,I1i 1200 1000 800 600 400 200 0 0 100 200 Vc I3o 1000 Capacitor Voltage (V) I3i 500 0 Charge Currents (A) natural commutation – and extremely good harmonic performance both with respect to the quality of the waveform offered to the motor and the level of harmonic pollution reflected back into the supply system. VIII. CONCLUSION The development of marine electrical power systems has been extremely rapid over the last two decades and this has been due to developments in the topology and construction of both power electronic switches and power converters themselves. These under-pinning device and equipment developments have formed the enabling technology for the implementation of IFEP systems in both warships and commercial vessels. As such the power electronic community has exerted a revolutionary effect on marine engineering resulting in a common development thread and ambition across the world. ACKNOWLEDGMENT The author acknowledges the support of Rolls-Royce plc and the many other companies and corporations without whose help this paper could not have been written. REFERENCES [1] Dr Jon Clare, Dr Pat Wheeler, “Matrix Converters Study”, UK MoD IFEP Report, September 1999 unpublished. [2] Ziogas PD, Khan S I and Rashid MH, “Some Improved Forced Commutated Cycloconverter Structures”, IEEE Transactions on Industrial Electronics Vol 1A-21 No 5, September - October 1985, pp1242-1253. [3] Ziogas PD, Khan S I and Rashid MH, “Analysis and Design of Forced Commutated Cycloconverter Structures with improved Transfer Characteristics”, IEEE Transactions on Industrial Electronics Vol 1E-33 No 3, August 1986, pp271280. [4] Cdr C G Hodge and Cdr D J Mattick, ’The Electric Warship III’ Trans IMarE, Vol 110, Part 2, The Institute of Marine Engineers (1998). [5] Cdr J M Newell, Cdr D J Mattick Royal Navy and C G Hodge, ’The Electric Warship IV’ Trans IMarE, Vol 111, Part 2, The Institute of Marine Engineers (1999). [6] C G Hodge and D J Mattick OBE, ’The Electric Warship V’ Trans IMarE, Vol 112, Part 2, The Institute of Marine Engineers (2000). I2i Ico,I2o 300 400 500 600 -500 -1000 µsec Ico,I1o Charge Cycle Disc harge Cycle Figure 15: Multi-Port Converter Waveforms main phase currents at any time sum to zero, simplistically: for any given time interval there will have been as much positive current as negative flowing in the circuit. The charges delivered and received balance precisely. Hence these positive and negative currents can all be supplied from a single capacitor whose positive plate and negative plate charges self evidently always sum to zero. When considering a 3 phase to 3 phase MPI converter within the time intervals of its operation; there will always be a phase which delivers, or receives the highest charge, while the remaining two phases are of opposite sign and combine to balance the overall charge transfer. Therefore, when operating, one input or output phase is connected to one side of the capacitor throughout the charge or discharge cycle while the other two are Central Capacitor Charging Inductor InputThyristors Freewheeling Thyristor Discharging Inductor Input Filter Capacitor Input Inductor Output Thyristors 1.45 h X 1.27 d X1.98 w = 3.64 cu meter Figure 16: Conceptual Multi-Port Converter connected in turn to the opposite plate. The only control requirement during this phase is to determine at which point the two minor phases should be commutated. Figure 16 shows the general arrangement of a conceptual MPI. The general operating principles have also received independent endorsement from third party assessments. The main advantages of this converter topology are very low switching losses – all switching is achieved by thyristors using un-forced
Please download to view
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
...

Ispsd 2002 Paper

by argonaftis

on

Report

Category:

Documents

Download: 0

Comment: 0

21

views

Comments

Description

Download Ispsd 2002 Paper

Transcript

Modern Applications of Power Electronics to Marine Propulsion Systems C G Hodge BSc MSc CEng FIMarEST Rolls-Royce plc, PO Box 3, Filton, Bristol, BS34 7QE, United Kingdom Phone: +44 (0) 117 979 5508, Fax: +44 (0) 117 979 6424 E-mail: c.hodge@blueyonder.co.uk Abstract – This paper reviews the current state of marine electric propulsion systems and the various system developments currently being pursued to achieve greater power density and explains why these developments are economically and operationally desirable. The paper also reviews the specific power electronic converter applications that are supporting the system developments with particular reference to converter design. The paper concludes by considering the likely future direction of development for power electronic converters. Indexing Terms – Electric Warship, More Electric Applications, Integrated Full Electric Propulsion. I. INTRODUCTION Marine propulsion systems are undergoing rapid development with a significant thrust towards the use of electrical propulsion and the integration of auxiliary and propulsion power systems. Whether the vessel in question is a large cruise liner driven by external podded motors, a small highly powered warship with inboard motors and shafts or an off-shore oil support vessel the need for very flexible, robust and compact power systems is ever increasing. This need is being satisfied through the recent advances, verging on revolution, in power electronic device construction and their application for power conversion and control. The recent developments in switching device implementation have been fundamental to the developments underway in marine electrical power systems. AS well as the Insulated Gate Bi-Polar Transistor (IGBT) Devices such as the Integrated Gate Commutating Thyristor (IGCT) and MOS Turn Off Thyristor (MTO) are beginning to be considered for use in marine electrical power systems. However, to allow space for discussion of the developments also underway in converter topology and control, this paper will not discuss these devices further. II. INTEGRATED FULL ELECTRIC PROPULSION A traditional mechanical marine propulsion system uses prime movers to drive propellers via shafts and in the most common arrangements each shaft is driven by its own dedicated prime mover. Unless complicated combining gearboxes are used it is not possible to cross-connect the shafts. The variation of propulsion power with speed is – to a first approximation – a cubic relationship. As a result 80% power provides 93% speed but 25% speed demands only around 1.5% power. Since many vessels rarely use full speed this results in the prime movers running for lengthy periods at fractional loads that reduce efficiency and increase operating costs. Electric propulsion – which is perhaps more accurately described as electrical transmission (between the mechanical prime movers and propellers) – offers a relatively straight forward method of cross connecting shafts in order to alleviate the light load running of the prime movers when the vessel operates at reduced speed. Figure 1 illustrates a conventional X Electrical System normally run split DG X DG X GT GB GT Shafts cannot be cross-connected GT GB GT X X DG X DG Figure 1: Mechanical Propulsion System mechanical propulsion system for a vessel with 2 shafts and an equivalent electrical propulsion system. The advantage of the latter is that a single prime mover can provide power to either shaft and – with the adoption of modern power electronics – astern operation and ship manoeuvring is possible without recourse to controllable pitch propellers or complicated reversing gearboxes. A further advantage of electrical propulsion is that the electrical generators need not be limited to supplying propulsion power alone – they can also be arranged to provide the electrical power required for the ship’s domestic and other uses. This auxiliary electrical power system is often termed the ship’s service power system and by combining this with the electrical propulsion system a common set of prime movers can provide power to either the propulsion or ship’s service duties. This arrangement is termed an integrated power system. When the electrical propulsion system is rated at full power (with no direct drive mechanical boost arrangement) the overall system is designated as Integrated Full Electric Propulsion (IFEP), figure 2 illustrates such an IFEP GEN CONV 20 MW GEN 2 MW CONV CONV CONV PMPM mass as desirable characteristics. There are several programmes across the world aimed at developing compact direct drive propulsion motors, notably in the United States of America, France and the United Kingdom however these programmes – though fascinating – are outside the scope of this paper. The need for compact and efficient equipment applies equally to the power electronic converters that are used to drive the propulsion motors in IFEP installations and the developments underway in this area will now be reviewed. IV. GENERAL ASPECTS OF MARINE POWER ELECTRONIC CONVERTERS Devices Used: Historically the converters used in marine electrical propulsion systems have been based on thyristors and these applications have obtained their commutating voltages either from the supply-side voltage source (as in cycloconverters) or from the motor back-emf (as in current source converters). Although this – as a type of natural commutation – brings high conversion efficiency such thyristor-based converters are in general too large for use in space sensitive applications and in addition each have significant (though differing) limitations on their frequency output, control and waveform quality. As a result current source converters and cycloconverters are now becoming displaced by voltage source Pulse Width Modulated (PWM) Converters that use devices capable of switching current off as well as on – such as IGBTs. Pulse Width Modulation: Figure 3 illustrates the construction of a PWM pulse train that would provide an effective sinusoidal voltage to a load. As with all PWM strategies the width of each pulse and succeeding “gap” are controlled to present the correct average voltage to the load. In the simplest strategy – as illustrated in Figure 3 – a saw tooth switching 300 200 BATTERY CONV CONV CONV PMPM GEN 8 MW GEN 800 V DC 6.6 kV AC 20 MW Figure 2: IFEP Propulsion System system. Integration of the two power systems can reduce the total number of prime movers that need to be installed – bringing reduced vessel acquisition costs – and where the ship’s service load is a significant proportion of the propulsion requirement, additional fuel savings can be achieved by using the ship’s service power as a base load for the prime movers – thereby ensuring that the engines are loaded sufficiently to retain fuel efficiency. Given their high hotel load, this latter aspect is the main reason why cruise liners have all but universally adopted IFEP in recent years. Even simple reduction gearboxes are generally avoided with IFEP installations as direct drive motors are feasible and thereby the additional purchase cost, maintenance burden and inefficiencies associated with a gearbox are avoided. III. THE IFEP EQUIPMENT REQUIREMENT Ships are always limited with respect to internal space and reduced equipment volumes have direct benefits in providing capacity for other purposes, not least additional cargo, passengers or in the case of warships weaponry. One benefit of electrical propulsion is that it provides the opportunity to move the electric motors towards the stern of the vessel leaving the higher value central hull volume for use directly associated with the vessel’s purpose. However, for hydrodynamic reasons, the stern of a vessel reduces in cross section and thus, as a result of its reduced displacement, the stern provides less uplift than other sections of the hull form. Consequently, equipment placed towards the stern needs to be relatively light if unacceptable stresses are to be avoided within the hull. Therefore marine electrical propulsion motors have both low volume and 100 Vol ts 0 0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018 -100 -200 -300 Time (s) Desired Voltage Saw Tooth PWM Figure 3: PWM Waveforms function is used to determine when the switches operate. Series Connection of Switches: The initial problem with PWM converters for marine propulsion use was that the rating of the available power switches were insufficient to provide the necessary power, which can often be 20 MW per shaft or higher. As described later, Alstom was the first company to overcome this problem and developed highly rated converters using series connected devices. One consequence of the use of series connected IGBTs is that the IGBT temperature must be tightly controlled in order to preserve the switching characteristics of each IGBT. Voltage Clamped Switching: As IGBT ratings have increased other topologies that avoid the need for series connection have become feasible; such as Neutral Point Clamped (NPC) Converters. Figure 4 shows a +V DC the losses generated in the capacitor network as it increases in size. As a result a 5 level diode clamped converter probably represents the limit of this topology. V. CONSTRAINTS OF MARINE POWER ELECTRONIC CONVERTERS Harmonic Distortion: Any converter’s prime function is to transform power between differing systems, AC to DC is perhaps the simplest converter (a DC rectifier) but even that shares a characteristic with all other converters – it switches current and voltage. The switching action allows the most favourably suited input phases to be connected to the appropriate output lines such that the desired form of output power is provided. Again in the case of the rectifier this simply means connecting the most positive input phase to the positive DC rail and the most negative of the input phases to the negative DC rail – this can be achieved passively through the use of diodes but nevertheless a switching process remains which results in discontinuous currents being drawn from the supply. As the current drawn from the supply is discontinuous it is – by definition – non-sinusoidal and it therefore creates non-sinusoidal voltage drops in the supply system. These non-sinusoidal voltage drops mean that the supply voltage (as seen by the converter itself and other connected equipment) is no longer sinusoidal – the converter has distorted the supply system. This feature applies to all converters: rectifiers, cycloconverters, current and voltage source converters. The most common form of distortion is that which results from DC rectification since this is a feature of the rectifiers found as the front ends of current and voltage source converters, however cycloconverters also exhibit supply side distortion and of a more complicated form. Since the supply system voltage and load current waveforms (although no longer sinusoidal) remain periodic, Fourier analysis can be used to decompose them into spectra of frequency components – each an integer multiple of the fundamental frequency. The non-fundamental frequency components of the supply voltage and load current are termed harmonics and the overall distortion is termed harmonic distortion. It is of note that current harmonics are the cause and voltage harmonics the affect. Thus in order to prevent voltage harmonics interfering with the operation of parallel-connected equipment, the current harmonics must be reduced to an acceptable level by filtration. Figure 5 illustrates the frequency components of the current waveform that result in the supply system that feeds a DC rectifier. Because voltage harmonic distortion arises through voltage drops created in supply system impedances, high impedance systems – such as those found in marine use – are more 0V DC One phase output connection -V DC Figure 4: NPC Topology simplified schematic of a NPC topology. As can be seen an artificial neutral voltage – obtained by a capacitor ladder bridge – clamps the inter-device voltage via diodes such that the maximum voltage seen across the switches is half the circuit maximum. If the converter’s devices are able to withstand half the circuit voltage then any need to switch them in synchronism is avoided. At present converters rated at 30 MVA are achievable with the NPC topology. As well as providing a robust converter design NPC converters are also inherently more efficient as the devices switch through lower voltage differentials and thus the switching loss is reduced. A final advantage – again accruing from the reduced voltage swing – is that the output waveform more closely approximates to a sinusoid; it contains a lower harmonic content. It is of course not necessary to limit the clamping voltage to the mid point of the DC link voltage alone, more extensive capacitive networks can generate any desired number of intermediate voltages which can then be used to asynchronously operate any number of switches in a converter limb safely, the converter is then termed a diode clamped converter. However the advantages offered by increasing the number of levels offers reducing returns with respect to the quality of the waveform and the efficiency increases gained by the reduction of the switching loss are rapidly overtaken by susceptible to voltage distortion and one of the perennial challenges for the marine power system integrator is to predict the levels of harmonic distortion at the design phase so that suitable counter measures such as harmonic filters can be incorporated in order that the supply system retains an adequate level of power quality. The size of harmonic filtration required in any given marine power system is clearly a system level issue but it is rarely insignificant. For a typical IFEP system with the ships service system interconnected to the propulsion system by simple transformers the harmonic filters will occupy around 60% of the volume of the propulsion converters themselves. This adds weight and cost and reduces overall system efficiency. Hence the harmonic performance of any given converter design is a crucial issue. dV/dt: The output of a converter is often used to drive a propulsion motor and here – particularly for compact marine propulsion – the performance of the motor insulation under the PWM waveform generated by the converter is crucial. The motor’s winding insulation needs to withstand the sequence of steep sided pulses of voltage emulating the required sinusoidal voltage if failure is not to result. The most crucial aspect of the voltage waveform with regard to insulation stress is the rate of change of voltage and PWM converters are capable of extremely high rates of change of voltage. This aspect of insulation degradation and potential failure is often, on first thought, found to be surprising. Nevertheless insulation failure is one of the commonest reason for failure in the several compact propulsion motor developments that are taking place in both Europe and the United States of America. It currently appears that a peak voltage of around 5kV is a maximum –at least empirically. This limit also sets the DC link voltage in a PWM converter and therefore the United States medium voltage standard of 4.16 kV is attractive to marine system designers as it provides around 5 kV when rectified. Because the rate of change of voltage as seen in a converter’s output is a key contributor to the stress and resultant degradation of the motor insulation many converters use output filters to reduce their rates of change of voltage. Occasionally these output filters are designed to remove the majority of the output harmonics and they may then be termed sinusoidal filters. However, more commonly, the output filter is designed only to limit the rate of change of voltage to a specified level and the filter is then termed a dV/dt filter. dV/dt filters are naturally much smaller than sinusoidal filters although the exact size – and thus their cost and the amount by which they reduce overall system efficiency – is one of the key system design considerations. Motor Regeneration: Because marine electrical propulsion systems drive vehicles that gain kinetic energy as they move, the maneuvering of ships is complicated by the need to dissipate the stored kinetic energy when reducing speed. The interaction of the propeller with the slipstream of the water passing it further complicates the issue. In general when wishing to reduce speed, simply reversing the direction of rotation of the propeller, or even just stopping it, is never the best strategy. The best thrust in the opposite direction to the ship’s motion is achieved by turning the propeller at just less than synchronous speed – that is at just less than that rotational speed which would match propeller advance and ship speed when considering the propeller as a simple screw. When this is done, as well as the thrust reversing so does the torque on the shaft and therefore so does the power flow from the motor that starts to regenerate power back into its converter. The problem of controlling and dissipating the regenerated power during maneuvering is a significant issue for the system integrator in the early phases of the design and it usually required as dedicated dynamic braking resistor and controller so that the significant regenerated power does not destabilize the upstream power system. Switching Frequency and Switching Loss: The higher the switching frequency the better the quality of the voltage waveform supplied to the load and thus the current it draws. This has a direct impact on noise and vibration and also on the filtration requirement of the converter output. Ideally the marine propulsion system designer would like to operate at the switch’s maximum frequency. However each switching cycle represents an energy loss as the voltage across the switching device builds up at the same time as the current decays and thus there is energy dissipated 2.5 2 1.5 1 0.5 0 0 5 10 15 20 25 30 35 -0.5 Time Volts Current Power Energy Figure 6: Switching Loss within the switch. In order to limit the switching loss current-day marine propulsion power electronic converters are limited to around 2kHz – far below the theoretical capability of modern IGBTs. Figure 6 illustrates the switching loss inherent in a forcecommutated device. Reduction of switching loss is one of the ambitions of current converter development. VI. EXAMPLES OF MODERN MARINE POWER ELECTRONIC CONVERTERS Alstom Series Connected Converter: Alstom developed the first high power marine power electronic PWM converter based on series connected IGBTs to provide the necessary voltage withstand for a 25MVA Disk 3 Outer Supply Bridges Filter Filter Filter Filter Inner Disk 4 Outer Inner Power Paragon (a part of L3 Communications) which uses 5 intermediate voltage levels (again generated by a capacitor bridge). DC Link SIGMA 1 AC Supply SIGMA 2 Channel 1 Channel 2 Figure 9: Power Paragon Diode Clamped Converter Filter Outer Disk 1 Filter Inner Filter Outer Disk 2 Filter Inner Figure 7: Alstom Converter requirement. The main technical challenge was synchronizing the switching of the IGBTs so that the situation where insufficient devices were attempting to withstand the system voltage was avoided and also providing robust voltage sharing circuits so that each IGBT was acceptably stressed. The general modular topology of the Alstom Converter is shown in Figure 7. The Alstom converter development has been a significant success and has been used by both the United States of America and the United Kingdom within their current Naval programmes. Voltage Clamped Converters: The NPC topology has been developed by ASI Robicon of the United States of America and also by PMES of the United Kingdom. Figure 8 illustrates a NPC converter. An NPC converter is, of course, the simplest case of a multilevel converter using three voltage levels and Figure 9 illustrates a Diode Clamped converter developed by DCPOS Isolated Voltage Converter: Power Paragon and ASI Robicon (a division of the USA based High Voltage Engineering Corporation) have each conceived a type of multi-level converter that does not rely on a capacitor network to generate intermediate clamping voltages. Instead a transformer with a single primary winding and multiple secondary windings is used to create electrically isolated voltages which can then be independently switched by IGBTs or other power switches which are nominally series connected on the output side. Although the converter tends to be relatively large the topology does offer the possibility of generating extremely accurate sinusoidal output voltages through the use of interleaved switching across the multiple modular switches. The converter topology exhibits an excellent harmonic distortion performance and this can offset its larger volume through the reduced need for harmonic filtration. VII. FUTURE MARINE POWER ELECTRONIC CONVERTERS Resonant Converters: Resonant Converters are a class of converters that offer great benefit for the future. The concept is that an auxiliary circuit around the main power switch creates a resonant flow of current in a circuit that forces either the switch current or voltage to be near to zero during the commutation interval. In this way the switching loss is largely removed because the volt-amp product is also near to zero. When the current is held near to zero the circuit is said to be current resonant, conversely when the voltage is held near to zero the circuit is said to be voltage resonant. Current resonant circuits resemble the external commutation circuits used with thyristors – and indeed they provide the same function, the load current is diverted out of the switch. Voltage resonant circuits resemble passive IGBT Gate Drive PEC IGBT Gate Drive PEC IGBT Gate Drive PEC IGBT Gate Drive PEC OUTPUT IGBT Gate Drive PEC IGBT Gate Drive PEC IGBT Gate Drive PEC DCNEG IGBT Gate Drive PEC Figure 8: PMES NPC Converter snubber circuits – the main difference now being that the voltage resonant circuits operate in advance of the switching transient whereas the snubber circuit follows behind. In both cases the timing of the firing of the auxiliary resonant circuit is crucial in relation to the timing of the commutation of the main switch. Once the switching loss is largely removed the designer has a choice between increasing converter efficiency, increasing converter switching-frequency, increasing converter current output (so that the conduction loss rises to replace the reduced switching loss) or perhaps a combination of the three. There is however a problem with resonant conversion for marine propulsion uses. As previously explained marine propulsion systems operate for considerable periods at part load and in this condition the converter’ switching loss is significantly reduced anyway. However the additional passive circuit losses introduced by the resonant circuitry remain at the same level as if the converter was at full power. Thus a marine propulsion converter using resonance can actually increase overall losses rather than reduce them. One answer is to design the resonant circuits so that they can be switched out of operation once their losses outweigh the savings – however this can only be achieved for current resonant converters and these are the more bulky and expensive version of resonant converter. SPCo Development: An innovative form of multi-level converter is being developed by SPCo of the USA where MOS Turn Off Thyristors (MTO) switch a base voltage between various intermediate levels and this is then used as a PWM voltage source by a final resonantly switched IGBT based output stage. This converter ought to offer improved efficiency since both the switching and conducting losses are significantly reduced. The switching losses are reduced because the MTOs controlling the base voltage switch infrequently and the IGBTs performing the final PWM modulation switch resonantly. The conducting losses are reduced because the MTOs – which replace IGBTs in comparable topologies – are thyristor based and therefore exhibit a very low forward conducting volt drop. Figure 10 illustrates the SPCo converter. Matrix Converter: Another form of converter under development is the Matrix Converter and this is receiving significant attention in both the USA and UK with much work being conducted by the UK’s Nottingham University as described at [1]. The Matrix Converter gains it name from the appearance of the converter’s electrical connections. When drawn as shown in Figure 11 the nine bi-directional switches 3 Ph Input Bi-Directional Switches 3 Ph Output Figure 11: Matrix Converter appear as a matrix of connections. The bridge produces any desired output waveform by connecting each of the input phases (VA, VB, and VC) to each output phase (Va, Vb, and Vc) in turn within a switching interval of fixed duration. Within the fixed switching interval the periods of output-connection for the three input phases is varied in a manner similar to PWM such that the desired output is produced in terms of the average of the input voltages. In order that any desired output voltage can always be produced without regard to the instantaneous phase of the three input voltages the output voltage must be limited to a level below the input voltage envelope. As can be seen from Figure 12 300 200 100 Soft-Switching Section Volts 0 0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018 -100 -200 -300 Time (s) IGBT 3300-V 1200-A Figure 12: Matrix Converter Voltage Envelope (1) 7-kVDC 5-Level Super GTO 5000-V 3000-A Super GTO 2500-V 1500-A Figure 10: SPCo Converter the input voltage envelope restricts the output voltage to 50% of the input. The minimum of the maximum (positive) is half of the peak value, occurring when the two positive voltages are equal. Similarly the maximum of the minimum (negative) is half of the peak value, occurring when the two negative voltages are equal. This is a serious limitation for the Matrix Converter when it is intended for use in a standard power system with standard motors. However it may be overcome by using one of two differing techniques. The first is to distort the desired output phase voltages by use of a combination of third harmonics of both the input voltage frequency and the desired output frequency. The required additional harmonics are: Input Voltage Third Harmonic: -1/6 Output Voltage Third Harmonic: +√3/6 When this is done the phase voltages are as shown in Figure 13 and the peak desired output voltage may be raised to 0.866 of the input. Thus the maximum output 300 200 100 0 0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0.018 -100 -200 -300 Time (s) Figure 13: Matrix Converter Voltage Envelope (2) voltage can be increased from 50% to 86.6% an increase of 36.6 percentage points or 73.2%. Two important points are worth noting with regard to this technique. First, although the phase voltages are distorted this does not appear in the output line voltages because of the natural third harmonic cancellation that occurs in a balanced three-phase system when moving from phase to line voltages. Second, this additional voltage increase has not been achieved at any other performance aspect of the converter, its overall efficiency remains unaltered as does the quality of the output voltage and the level of distortion imposed on the input system. The second technique uses a fictitious DC link and divides the modulation function into two stages, the first deals with input voltages, and the second with the generation of the output voltages. It is termed a fictitious DC link because in order to maximise the output voltage the first modulation stage chooses to work between the most positive and most negative input voltages, as would a standard uncontrolled DC rectifier. The overall maximum voltage output that can be achieved with this technique is 6√3/π2 or 105.3%. Although the modulation function is developed by considering two distinct phases of time-continuous operation this cannot be the case when the modulation function is applied to the converter – there is only one switch between the input and the output and no intermediate stage. In practice the two separate switching strategies are combined into one unified requirement before the modulation function is applied. Techniques to achieve this are covered at [2] and [3]. There are however significant disadvantages. Although this technique overcomes the voltage limitation of the Matrix Converter completely, this gain is bought at the expense of the overall harmonic performance of the converter: the harmonics reflected into the input voltage or generated in the output current (or indeed both) must increase. It is perhaps not surprising - since the modulation function is in part derived through consideration of DC rectification - that the largest increase in harmonics are those associated with this process itself (5th, 7th, 11th etc). An additional source of harmonic distortion arises when the voltage is forced above 86.6%, when this is done it is no longer possible for the average output voltage in each switching interval to equal the target voltage and additional low frequency harmonics result. It is interesting to note that a Matrix Converter operating at minimum harmonic distortion under a fictitious DC link scheme will be subject to the same voltage limit as a converter operating under the third harmonic strategy. There are practical difficulties in developing the overall modulation function through two stages in real time so that implementation of a converter producing continuously varying frequency and voltage (as required in IFEP applications) would be difficult. Clearly a Matrix Converter would benefit from a symmetrical bi-directional switch. SAIC Multi-Port Inversion: SAIC – and in particular Dr Rudy Limpaecher – have been developing an extremely innovative converter termed the Multi-Port Converter. The circuit schematic diagram for an MPC converter is shown in Figure 14 together with the current waveforms for one charge and discharge cycle at Figure 15. The technique has been more fully explained at [4], [5] and [6] but, in summary, it utilises Input Redistribution Output Volts Filter Figure 14: SAIC Multi-Port Converter Filter the fact that in any balanced multi-phase system the Ici,I1i 1200 1000 800 600 400 200 0 0 100 200 Vc I3o 1000 Capacitor Voltage (V) I3i 500 0 Charge Currents (A) natural commutation – and extremely good harmonic performance both with respect to the quality of the waveform offered to the motor and the level of harmonic pollution reflected back into the supply system. VIII. CONCLUSION The development of marine electrical power systems has been extremely rapid over the last two decades and this has been due to developments in the topology and construction of both power electronic switches and power converters themselves. These under-pinning device and equipment developments have formed the enabling technology for the implementation of IFEP systems in both warships and commercial vessels. As such the power electronic community has exerted a revolutionary effect on marine engineering resulting in a common development thread and ambition across the world. ACKNOWLEDGMENT The author acknowledges the support of Rolls-Royce plc and the many other companies and corporations without whose help this paper could not have been written. REFERENCES [1] Dr Jon Clare, Dr Pat Wheeler, “Matrix Converters Study”, UK MoD IFEP Report, September 1999 unpublished. [2] Ziogas PD, Khan S I and Rashid MH, “Some Improved Forced Commutated Cycloconverter Structures”, IEEE Transactions on Industrial Electronics Vol 1A-21 No 5, September - October 1985, pp1242-1253. [3] Ziogas PD, Khan S I and Rashid MH, “Analysis and Design of Forced Commutated Cycloconverter Structures with improved Transfer Characteristics”, IEEE Transactions on Industrial Electronics Vol 1E-33 No 3, August 1986, pp271280. [4] Cdr C G Hodge and Cdr D J Mattick, ’The Electric Warship III’ Trans IMarE, Vol 110, Part 2, The Institute of Marine Engineers (1998). [5] Cdr J M Newell, Cdr D J Mattick Royal Navy and C G Hodge, ’The Electric Warship IV’ Trans IMarE, Vol 111, Part 2, The Institute of Marine Engineers (1999). [6] C G Hodge and D J Mattick OBE, ’The Electric Warship V’ Trans IMarE, Vol 112, Part 2, The Institute of Marine Engineers (2000). I2i Ico,I2o 300 400 500 600 -500 -1000 µsec Ico,I1o Charge Cycle Disc harge Cycle Figure 15: Multi-Port Converter Waveforms main phase currents at any time sum to zero, simplistically: for any given time interval there will have been as much positive current as negative flowing in the circuit. The charges delivered and received balance precisely. Hence these positive and negative currents can all be supplied from a single capacitor whose positive plate and negative plate charges self evidently always sum to zero. When considering a 3 phase to 3 phase MPI converter within the time intervals of its operation; there will always be a phase which delivers, or receives the highest charge, while the remaining two phases are of opposite sign and combine to balance the overall charge transfer. Therefore, when operating, one input or output phase is connected to one side of the capacitor throughout the charge or discharge cycle while the other two are Central Capacitor Charging Inductor InputThyristors Freewheeling Thyristor Discharging Inductor Input Filter Capacitor Input Inductor Output Thyristors 1.45 h X 1.27 d X1.98 w = 3.64 cu meter Figure 16: Conceptual Multi-Port Converter connected in turn to the opposite plate. The only control requirement during this phase is to determine at which point the two minor phases should be commutated. Figure 16 shows the general arrangement of a conceptual MPI. The general operating principles have also received independent endorsement from third party assessments. The main advantages of this converter topology are very low switching losses – all switching is achieved by thyristors using un-forced
Fly UP