Developments in shore based testing to meet latest marine power system integration challenges

Transcription

1 Developments in shore based testing to meet latest marine power system integration challenges G Bellamy* BSc(Hons) MIET CEng & B Salter* BEng MIET CEng * GE Energy Power Conversion UK Ltd, Rugby, UK This technical paper is prepared for the purposes of the 2016 International Naval Engineering Conference (INEC) and based on the information collected through various sources. Many variables may impact the technical considerations and data as well as the study results illustrated in the documents. GE s intention is only to give examples of potential applications and related impact of certain enhanced technology and products and these technical documents are not meant to provide any guarantee relating to the studies and related data. SYNOPSIS Integrating complex power and propulsion systems can be a thought provoking process, whether it is full electric or hybrid. Many of these challenges are now well understood and have seen the benefit of comprehensive land based testing. Recent trends towards transformerless drives, novel hybrid systems, and a broader use of power electronic functionality has of course introduced many benefits. But combining this with a more competitive landscape and solutions often now comprising multiple types of power electronics and sub-systems from different vendors, has introduced new issues to the marine power and propulsion community. In tandem with these challenges, customers are looking for greater functionality from their power electronic equipment, particularly as a power source as much as a load. The new functionality has its own de-risking requirements particularly in terms of fault behaviour and this, combined with an increased emphasis on the importance of environmental and long term testing has extended the range of risks addressed by shore based testing. This paper aims to capture some of these risks and show how the latest developments in full scale integrated shore based testing are being used to get it right first time. INTRODUCTION Naval electric propulsion tends to change and develop in significant steps, unlike the commercial shipping world which typically takes a more incremental approach (1). This fact, combined with the critical nature of the Power and Propulsion (P&P) system on a warship means that de-risking the P&P design is a key consideration in a warship design and build programme. Full Scale Shore Based Testing (FSSBT) is a key tool in the de-risking process; notable examples include the Type 23 Frigate (Figure 1 Type 23 Testing 1980's T45 Destroyer (RN) and the DDG1000 (USN) programmes. These test regimes helped to de-risk very significant challenges: Introduction of hybrid electrical propulsion to surface warship. Introduction of Integrated Full Electric Propulsion (IFEP) to RN warship. Introduction of very high power IFEP to USN warship. Author s Biographies Graham Bellamy is presently the Director for GE s Marine Power Test Facility and Naval Technology, with 18 years experience of Naval Power and Propulsion systems. Previous roles include Chief Engineer for Naval systems and lead roles for major Naval P&P projects in the UK and overseas including Albion Class, Wave Class, Chief Engineer for the Electromagnetic catapult (EMCAT) program, engineering manager USN DDG1000 destroyers P&P system and responsibility for teams of engineers delivering P&P for QEC carriers, T26, MARS tankers, Naval land based test facilities and other programs. Ben Salter is a Chartered Electrical Engineer with over 25 years experience in the application of power electronics, electrical machines and control systems. His current role is Technical Solutions Director Naval for GE Power Conversion where he involved in the development and deployment of electrical technology in the Naval field.

2 However, with the desire to move to highly flexible transformerless Active Front End (AFE) drive systems with Power Take-In (PTI) and Power Take-Off (PTO) and multiple vendor equipment and sub-systems, significant challenges remain; both of a technical and of a broader nature: Stability of multiple AFE drive systems from different vendors. Highly flexible PTI/PTO hybrid system functional de-risking. Management of common mode and high power, high frequency switching in complex Electro-Magnetic Compatibility (EMC) environments. Environmental Envelope Testing. Software Validation. Suitably Qualified & Experienced Personnel (SQEP) establishment and maintenance of skills. Some of these issues can manifest themselves to a degree at the equipment level but in general the impact can only be exposed at the sub-system, system and indeed the platform level. Overlaying all these issues is the imperative of affordability. It is essential that there must be a maximum of de-risking with a minimum of outlay and this means that the integration strategy must make an accurate assessment of the level of residual risk and reduce this to an acceptable level; which is not necessarily to a minimum. This raises two key questions: what is the nature of the key risks and how do we mitigate them? Typically design engineers tend to focus on the hard technical risks within the boundary of the design. In the case of P&P systems much focus has been given to electrical system level risks such as the management of load rejection, management of short circuit faults and the associated protection scheme, system stability and even system level faults such as mal-synching. It is essential that such proving is done. However, experience at sea suggests that these issues do not result in the majority of Operational Defects (OPDEFs). On one hand this can be seen as a success for the de-risking process. On the other hand, the issues that have impacted have often been at the electrical system / platform system boundary and these risks could equally have been mitigated at the test stage at relatively little cost. Humidity range testing, temperature range testing, out of range degradation testing and even support system fault testing, such as cooling water overpressure could all have been incorporated if the risks had been identified early enough. Having identified the key risks, the real question is how they can be mitigated at minimum cost. Many approaches are possible but this paper will consider 4 approaches: 1. Laboratory Scale Testing. 2. Extended Testing on Platform. 3. Mathematical Modelling. 4. Full Scale Shore Based Testing. Figure 1 Type 23 Testing 1980's

3 APPROACH 1 LABORATORY SCALE TESTING Laboratory scale testing is a useful tool for prototype development and control software development and testing at the product level. It can also be a useful tool for targeting specific non-designed behaviour in a limited area of a system. A good example of this is the use of laboratory testing for de-risking resonant modes on the Queen Elizabeth Class carrier shaftline for the RN. However, when it comes to modelling system behaviour and the interaction of multiple elements, lab based testing is of limited value beyond the concept proving stage. This is because it is very hard to be confident of the fidelity of significant aspects of a lab model when scaled up to full size. System level non-designed behaviour is typically not faithfully reproduced at lab scale unless it is the specific target of the lab model design. An example of this is common mode coupling, where stray capacitances are not directly scalable from a lab model to a real system. These factors mean that this is not a preferred option for de-risking system level issues for a specific platform. APPROACH 2 EXTENDED TESTING ON PLATFORM This form of testing provides the highest possible fidelity but it is rarely used as a stand alone de-risking approach. This is because, by definition, it can only occur very late in the build process at the set to work phase. This means that the cost of fixing any problem that is identified is very high and the impact on programme can be severe. This approach to de-risking is generally taken where the overall risks are judged to be low. This could be because the P&P system is judged to include only incremental design changes from one already in service or because the Extended Platform Testing follows previous de-risking using one or more of the other options. APPROACH 3 MATHEMATICAL MODELLING This is the most widely used de-risking approach for P&P systems. It is less costly than lab scale modelling and can be used early in the design process. Its ability to look at system level interactions means that it can be a very powerful de-risking tool. A key benefit is that a very wide range of de-risking scenarios can be performed inexpensively which can explore the boundaries of the performance envelope comprehensively. The key challenge facing mathematical modelling is the well-known computer problem of GIGO Garbage In, Garbage Out. The value of the modelling results in a given scenario depends heavily on the fidelity of the subsystem/equipment models; few of which are reliable at the performance boundary unless backed up by test data. Mathematical models are particularly questionable when dealing with non-designed behaviour. For example, the total capacitance to ground, including stray capacitance, within an item of equipment or along a specific cable run adjacent to other cables and earthed structures is simply not known and would be difficult to the point of impossibility to model from first principles. In such cases, test data on representative equipment is the only credible approach. APPROACH 4 FULL SCALE SHORE BASED TESTING This approach has several key advantages over the other de-risking methods. The maximum benefit of FSSBT can be realised by careful timing to ensure that it is early enough in the programme to enable impacted risks to be fixed at relatively low cost, but late enough in the programme to ensure that truly representative equipment is tested. A wide range de-risking can be achieved by this approach. This includes: Programme de-risking since ship standard equipment has a clear deadline well in advance of the platform critical path. Full scale system integration functional de-risking. Full scale system non-designed behaviour de-risking. SQEP de-risking since formal and informal training can take place on real equipment in a real system. System Software Functional Validation including Platform Management software functionality. Software Change Management Validation. Cultural de-risking encouraging the formation of an alliancing attitude from the different vendor technical teams as they work together to solve problems with technical goodwill. Proving Interfaces (Physical, Cabling, Control, Cooling etc.). Proving Installation Commissioning Operating and Maintenance Procedures.

4 Proving QPS, EMC, Earth currents, N&V, System Performance and defining Operating Envelopes. Undertake High risk tests in a controlled environment, e.g.) Short circuits, mal-synchronisation. Whilst nearly all programs face budget challenges today and the thought of additional costs associated with enhanced testing can often be on the like to have list, addressing some of these topics early can provide significant through life cost savings if planned into the overall build, commission, sea trial and platform through life strategy. However, if this level of de-risking is to be achieved as part of an affordable programme then the costs must be moderated and the activities focussed to achieve the optimal balance. Cost minimisation can be achieved in three key ways: 1. Re-use of infrastructure. 2. Maximise use of modelling. 3. Maximise use of emulation. REUSE OF INFRASTRUCTURE A great deal of the cost of FSSBT is tied up in the infrastructure: buildings, cranes, cooling systems, load systems, power supplies, SQEP staff etc. (see Figure 3 Prime Mover Emulator). All these elements are expensive to establish, so once an investment is made on one programme, it is enormously beneficial to re-use it on others. Attempting to re-start from scratch could, in itself, rule out FSSBT for a programme on affordability grounds. MAXIMISE USE OF MATHEMATICAL MODELLING In many cases the functional de-risking should be seen as primarily being achieved by the modelling. The purpose of the FSSBT is to validate the models and to prove the fidelity of their behaviour, especially at key boundaries on the performance envelope. The costs of the FSSBT are reduced by only testing these key boundary points and not repeating the full range of modelled scenarios. MAXIMISE USE OF EMULATION Emulation is a key way of reducing the cost of linking the full scale Equipment Under Test (EUT) to the wider system; electrical or mechanical. The extremely high control bandwidth of power electronic converters means that a very wide range of mechanical systems can be emulated by an electric machine and a drive. At the most fundamental level, virtually any P&P test site will include some form of four quadrant load that can both absorb power from and deliver power to the propulsion shaftline. The ability to emulate propulsion system dynamic loads is well understood and has been demonstrated with multiple references up to the largest scale for four quadrant loads (4QL) such as that shown in Fig 2. In this case the Converter fed 20MW AIM (Advanced Induction Motor) as used to propel T45, DDG1000 and QEC is connected via a flexible coupling and gear box to a 4QL machine and converter at the GE Marine Power Test Facility (MPTF). By applying the correct magnitude, direction and rate of change of torque in response to control system demands, the machine can effectively emulate an array of Robinson curves in all 4 quadrants as though the propulsion system is at sea and in various states. Including simulation of acceleration, deceleration, crash reversals, prop out of water or cyclic loading to name a few. In this case a representative test IFEP half ship system was tested, this allows not only the drive train aspects to be validated but the behaviour of the power and control system behind the drive train to be validated through various evolutions as though it were at sea.

5 Figure 2 20MW - 4 Quadrant Propulsion Load Emulator However, emulation can go far beyond this example. The full torque envelope, inertia, governor characteristics, overload behaviour etc. of an engine can be replicated without the need to provide the fuel system, cooling system, exhaust system and noise management infrastructure that a real engine requires (see Figure 3 Prime Mover Emulator). This means that the dynamic system interactions can be fully de-risked at a fraction of the cost of installing an engine on site. This is particularly valuable where multiple identical engines might be desired. One real engine could be installed for de-risking non-designed behaviours but the others can be emulated at lower cost, this effectively becomes the ultimate hardware in the loop test regime. A similar approach can be taken in the electrical system for elements such as auxiliary loads, where the infrastructure of the test site itself can be used to emulate platform loads. Fuel Systems or Steam Gen Plant Lubrication Systems Start Systems Air / Electric etc VSD Control Provides Dynamic Emulation & Synthetic Inertia Enclosure GT Uptake / downtakes Exhaust System Prime Movers Diesel Engines, Gas Turbine, Steam Turbine etc G G Civils Accoustic Management Systems Eng Mngt / Fire Suppression etc.. = Prime Mover Dynamic Model M G Maintenance & Running Cost Emissions Control & Environmental Impact Figure 3 Prime Mover Emulator Having considered these issues theoretically let us now look at couple of examples of modern de-risking that have taken place, or are about to take place at a full scale test site to address some of these topics. COST EFFECTIVE DE-RISKING OF THE MARITIME REACH AND SUSTAINMENT AT SEA (MARS) ADVANCED HYBRID PTI/PTO SYSTEM One example of a cost effective and targeted risk approach was employed for the new RFA MARS tankers advanced hybrid system. The shore testing effort here was focused on key sub-system integration risk and made use of an affordable shore test solution and available existing infrastructure. This approach was targeted at subsystem drive train and power system behaviour and integration of the PTI / PTO as it would be experienced in the MARS system. Significant de-risking was achieved by the use of mechanical emulation and dynamic loads with non-contract but equivalent system hardware at MW scale and utilising existing systems and infrastructure. The MARS solution uses a Combined Diesel electric or Diesel (CODLOD) and is a subject of other papers in the proceedings so not described in detail here. However in summary the system utilises a converter fed induction motor that can operate in both a PTI or PTO mode in conjunction with a Main diesel engine, gear box and controllable pitch propeller (see Figure 4).

6 The shore test not only demonstrated the hybrid propulsion mode behaviour but was focused on the performance envelope requirement of the converter system when used as a static converter and induction generator. In this mode the converter equipment is effectively the network and supporting the complete platforms power system and everything connected to it. Some of the challenges here include the behaviour of the converter and its software and control of KW and KVAr sharing when working in parallel with diesel generators, other PTO converters or standalone. The test arrangement was also configured to validate its performance in response to transient overloads such as DOL motor starts and transformer inrush. The system also proved its ability to actively manage network quality of power supply and voltage regulation in response to large reactive loads and associated filter voltage drops along with its ability to provide sufficient short circuit current to allow discrimination whilst maintaining control in the event of network faults. These cover but a few of the areas validated by the testing and was achieved utilising an arrangement at the GE MPTF as shown below. (see Figures 5,6). PTO Generation DE G/B IM PTI Propulsion Energy Re-circ mode when closed HRG Figure 4 MARS Ship System Figure 5 Equivalent Shore Test System The system utilised an equivalent induction motor coupled to a permanent magnet motor in a multifunctional arrangement which could either act as a 4QL to emulate propulsion load characteristics or prime mover emulation mode to drive the induction motor in PTO mode and represent the Diesel Engine (DE) behaviour. These were both fed by the same standard product as used for the platform in the form of 690V Multi MW transformerless AFE MV3000 converters With the bus tie closed the shore testing allowed a back to back mode for PTI testing and endurance runs and allowed class certification in a low fuel cost mode such that energy was re-circulated and only the losses supplied. Unlike other back to back test motor loading arrangements such as common DC bus schemes or DC loading schemes, this AC AFE arrangement allowed the full drive train motor and converter plus control functionality and power system behaviour to be fully validated in one arrangement. In PTO test mode the converter was coupled to the facility s diesel generator to demonstrate synchronisation, load sharing and transient behaviour as would be experienced on the platform. By coupling to the wider facility system, response steps of resistive and reactive loads via existing load banks, and DOL starts from a range of motor sizes and load such as pumps/fans to achieve current and network voltage regulation validation. The arrangement also allowed tests such as network short circuits, the stability aspects of multiple AFEs on the network. Tests to prove response to earth faults and enhanced protection co-ordination functions were undertaken in a controlled environment that would otherwise not have been performed easily on the vessel. Enhanced operating envelope testing was also undertaken such as ambient air and humidity trials and class certification for temperature rise performance at upper limits of ship primary cooling water temperature at very little cost but whilst bringing significant de-risking benefits to the platform integration.

7 Figure 6 MARS Shore Test Arrangement INTEGRATION CHALLENGES OF MULTI-VENDOR POWER ELECTRONICS & EMC IN A COMPLEX ENVIRONMENT As described previously the trend towards AFE transformerless power electronics on marine power systems and a more competitive procurement approach, utilising often several different Original Equipment Manufacturer s (OEM) converters on a common bus, has posed several new challenges to the industry. Often the use of VSDs now go beyond that first seen on a single line diagram and are also within an equipment black box, e.g.) a ship s chilled water plant unit with a single connection point may contain AFE drives to support high speed compressor motors and active bearings within the equipment. Whilst this approach can provide many benefits, the lack of galvanic isolation historically provided by transformers and different filtering and switching strategies now employed by manufacturers, requires specific consideration as to its integration with the power system and compatibility with other equipment connected at the common point of coupling and beyond. Some of these aspects are not well covered by known design practices and standards hence require a relatively deep understanding of the technology and the system to integrate it most effectively. One example of this is the behaviour in response to zero sequence current from earth faults seen by all equipment on the now common supply, another critical factor to a fighting ship is the electrical system s Quality of Power Supply (QPS). Whilst harmonics are generally well understood for 6 pulse type converters and are covered well by STANAG and Class Rules etc. up to the 50 th harmonic, much higher frequency harmonics are present with transformerless PWM converters both differentially and in the common mode world. The common mode performance is particularly dependent upon the system arrangement to which it is installed and the modes and concept of operation to which the platform will be used. Whilst the individual product performance can be well understood and modelled under a set of fixed parameters, in a system environment these become an element of unknown behaviour and can change dramatically both as function of the other equipment connected and the system arrangement. For example, factors affecting performance include: the interaction of filters and the overall system installation, the capacitance to ground of the equipment in the system including generator and motor parasitic capacitances to ground, Other components connected to ground such as voltage suppressors and filters, the converter s modulation strategy and capacitance both physical and parasitic, the cable lengths and installation techniques used,, the overall grounding methodology employed. The authors have experienced real world examples where shipbuilders have connected numerous off the shelf transformerless AFE drives to a power system where the net QPS and other performance parameters simply did not match what was modelled or expected. A typical example was attributed to resonances due to the interaction of different filtering of drives that each OEM targeted to its own converters PWM requirements but having no visibility of its application and other converters in the system.

8 If a clear and competent integrator role is identified then modelling can play a significant part here but more often than not the system is very complex and the models lack fidelity in this area. Either the required source data needed simply isn t available, e.g. restricted due to intellectual property constraints, or just isn t accurate enough to be able to manage all the variables (described earlier) that it may see in service. On larger complex systems one approach being used to mitigate this is by compatibility testing and bringing all of the equipment together on shore with a representative loading environment. This in itself poses several challenges. The system s Equipment Under Test (EUT) must not be adversely affected by the emulating Test Equipment (TE) system which itself is usually power electronic based and hence requires careful consideration in the high frequency world. With this addressed the overall performance can be easily measured if an installation close enough to the platforms design is built e.g. equivalent cable lengths and equipment /installation and various evolutions are undertaken to reflect the vessel s concept of operation. But when it doesn t behave as expected, identifying the problem can be fraught with complications if not planned from the outset. In one example of an on- going project with multiple transformerless AFE vendors, one way to achieve this has been by careful design of the shore test facility to a level of detail not normally undertaken. To allow common mode paths and the interaction of multiple AFEs to be clearly identified and measured requires the ability to fully isolate individual items of equipment from the common earth plane. An approach to this on a large shore facility test arrangement is shown in Figure 7. Other EUT Tx-less AFE Converter EUT Tx-less AFE Converter Pipe Work Isolation TE Tx-less AFE Converter Power Cabling From EUT EUT Earth Bar M (EUT) Low Capacitance Isolated Flex Coupling M (TE) Power Cabling From TE TE Earth Bar Mode Measurement EUT Earth Bar TE Earth Bar Power Cabling to TE Power Cabling to EUT Concrete Foundation Re-bar / Earth Grid GRP Unistrut- Insulated Cable Tray Switch to common or separate TE & EUT earth rings EUT = Equipment under test (Part of platform system being tested) TE = Test Equipment (Part of Loading / emulation test system) Figure 7 Typical Installation for Mode Assessment In this example the EUT is installed as a half ship set with cable lengths, type and techniques as used on the vessel, most importantly the EUT has the capability to be isolated from the TE. Achieving this with a common facility system supporting the test bed such as cooling, auxiliary systems and physical structure requires careful design of layout,isolation techniques and materials used such that sneak earth paths can be readily identified and isolated whilst managing risks with step and touch voltages associated with different earth potentials. The safety aspect is typically managed by the use of a common earth grid for manned operation and a split arrangement used in an unmanned configuration with permanently connected transducers to the facility s high speed data acquisition system. In the case of the motor drive train, flexible couplings, pipe work and mount isolation are carefully chosen to provide a low capacitance isolation into the MHz range. This arrangement also allows individual equipment to have a single dedicated HF earth connection to be achieved such that common mode contributors can be readily identified at a single point.

9 CONCLUSION Full Scale Shore Based Testing for Naval electrical propulsion has moved on from the pioneering stage seen in previous decades. The complexity of modern power electronic based systems, the functional flexibility required of them and above all the affordability pressures of the modern environment has proved a challenge to those responsible for delivering such systems at an acceptably low risk. Shore Based Testing has developed and adapted to maximise re-use of flexibly configured test equipment to ensure affordability and has taken significant steps forward in platform fidelity where the risk justifies the effort. These measures have helped to keep Shore Based Testing at the cutting edge of modern Naval platform derisking and planned Full Scale Shore Based Testing programmes around the world emphasise its value for the future. ACKNOWLEDGEMENT The views expressed are those of the authors and do not represent those of GE Power Conversion, United Kingdom Ministry of Defence or any other of the organisations referenced herein. REFERENCES 1. N Smith & B Longepe : De-risking the electric ship: from theory to reality, Proceedings of the International Naval Engineering Conference, Glasgow United Kingdom, April