DEFKALION-THALIS project (MIS:380164) November 2014 - Summary 1. Introduction Electric power plant onboard ships has always been a rather complicated power system, comprising DC and AC subsystems of several operating voltage and frequency levels, especially in sophisticated structures with electric propulsion. The aforementioned complicacy is worsened even further in the All Electric Ship (AES) systems, referring to full electric propulsion and extended electrification of all shipboard installations. On the other hand, similarly to continental grids, several steady- and transient-state phenomena, especially concerning power quality problems, emerge, and their consequences have to be thoroughly studied, analyzed and investigated. The main general goals of DEFKALION project can be summarized below: - Establish a Centre of Excellence in PQ issues of ship electric energy systems. - In-depth interdisciplinary investigation of complex Power Quality (PQ) problems that influence ship operation, in the light of intense electrification of all ship systems. - Propose solutions for PQ problems, in the light of energy saving and environmental issues. - Design and implementation of a monitoring system for analysis and identification of PQ problems in naval power systems within the framework of improved energy efficiency and by the integration of this system with the existing on board power management tools (Power Management Systems PMS or Electric Power Management and Control Systems EPMACS). - Improve the quality of the education provided by Universities, considering the needs of the modern job market for Naval Architecture and Marine Engineering, especially Marine Electrical Engineering, as well as the recent developments and requirements in energy efficiency. The specific targets of DEFKALION project consisting the corresponding work-packages are: - Study of the optimum shaft generator configuration/exploitation, considering fuel cost, fuel consumption, emission rates, optimum performance, production and operation cost. - Study of mitigation practices for PQ problems due to thruster starting and operation. - Study of mitigation practices for PQ problems due to pod operation, especially during maneuvering. - Study of PQ problems due to ship grounding practices. - Study of PQ problems due to lightning strikes directly or nearby ships. - By combining the above items, study and analysis of PQ phenomena during the different types of ship operation (maneuvering, mooring, disembarkation etc) as well as design of an integrated monitoring system for PQ. 2. Work Description - Implementation The project DEFKALION builds on a previous project, already completed successfully, on the "Investigation of Power Quality phenomena in shipboard installations focusing on voltage deviation problems (funded within the frame of the PYTHAGORAS-II funding scheme). DEFKALION aims at performing a thorough investigation and analysis of Pοwer Quality (PQ) phenomena during the different types of ship operation, as well as at designing an integrated
monitoring system for PQ. The ultimate goal is a greener, safer, more reliable and more economic electrified ship. This target is well aligned, in a synergetic manner, with the MARINELIVE project funded by EU (contract No 264057, FP7-CAPACITIES-REGPOT-2010-1, Project Coordinator: John Prousalidis), within the frame of which a Centre of Excellence has been formed in the scientific field of All Electric Ship. Below is a brief summary of the work that has been achieved thus far based on the Work- Packages (WPs). WP2 Investigation of PQ problems due to shaft generator operation Shaft generator (SG) systems are gaining increased interest in ship propulsion, due to a number of advantages over conventional diesel generator (DG) systems. Generally, SGs are mounted on the propeller shaft between main propulsion engine and propeller and thus convert part of the power produced by the main engine to electrical. SGs are thus used in conjunction with DGs and other non-conventional sources for generating power on ships. Therefore, their design should be carried out taking into consideration performance and efficiency issues. In addition, the evaluation of their electromagnetic characteristics plays an important role in the analysis of SG systems. In this work, a low-speed direct-driven PMSG (400rpm) is designed and compared with a conventional salient-pole synchronous generator (SPSG) (1200rpm) of the same nominal power but greater speed, for implementation on an actual Ro-Ro ship. These machines are geometrically optimized and compared in terms of their main operating characteristics. Furthermore for the proposed candidate SG systems (direct driven PMSG with back to back converter & gearbox driven-high speed SPSG) are evaluated in terms of weight and volume. In addition, an innovative control system has been proposed in order to ensure constant frequency and voltage under variable engine speed and to enable operation of the shaft generator as motor in case of failure of the main engine or running in parallel with DGs. Finally, the benefits of implementing the proposed control scheme compared to the existing configuration have been examined, in terms of fuel savings. It was calculated that for a typical journey a 3% reduction in fuel consumption can be achieved with the proposed operational scenario. The analysis has been implemented in a real RoRo-Trailer-Passenger ship, constructed in Germany 12 years ago. Generally, there are three main types of SG topologies [1]: Power Take Off/Gear Constant Ratio (PTO/GCR), which consists of flexible coupling, stepup gear, torsionally rigid coupling, and generator as shown in Figure WP2_1. In this type, the electric frequency of the generator is proportional to the speed of the propulsion engine, which means that constant frequency production is possible only when the ship navigates at sea. In order to produce constant frequency, a controllable pitch propeller is installed or an induction generator (IG) is used. In the second condition, the rotor can rotate on variable speed higher than the synchronous speed, producing constant electric frequency on the stator. However, the IG must initially operate as a motor in order to be magnetized and then switch to generator mode, supplying with active power the electric network of the ship. Furthermore, IGs absorb reactive power, necessitating the installation of a synchronous condenser.
Alternatively, this topology can operate on standalone condition, with a small margin on the electric frequency (50-60Hz), supplying only some special loads (such as bow thrusters), which are tolerant on frequency fluctuations. Power Take Off/Renk Constant Frequency (PTO/RCF), which includes an RCF speedcontrolled gear and generator. The schematic of this topology is the same as PTO/CGR, shown in Figure WP2_1. In this topology, the gearbox maintains the generator s rotor speed constant. As a result, the electric frequency of the generator remains stable, enabling the parallel operation with diesel generators. Wound rotor synchronous generators are the most common electrical machine type for this SG topology. One of their main benefits is the ability to produce variable reactive power, unlike induction generators, maintaining the grid voltage constant by adjusting properly their field current. Furthermore, the protection of the machine is possible in short circuit operation by controlling their excitation. Power Take Off/Constant Frequency Electrical (PTO/CFE), which includes a step-up gear, generator and electrical control equipment, as shown in Figure WP2_2. Alternatively, a slowrunning generator, directly mounted to the front end of the main engine, with electrical control equipment, could be established, as shown in Figure WP2_3. In the first case, there is a gearbox with constant ratio and the electrical control system is responsible to convert the variable electric frequency to constant frequency. The power converter generally consists of a rectifier and an inverter which take up the role of keeping the electric frequency and voltage of the grid bus constant. In the second case, the gearbox can be eliminated by choosing a slow-running alternator with a high number of poles. In this topology, PMSGs can be considered. Figure WP2_1. Schematic of PTO/GCR & PTO/RCF with conventional or controllable pitch propeller Figure WP2_2. Schematic of PTO/CFE with gearbox and conventional or controllable pitch propeller
Figure WP2_3. Schematic of PTO/CFE with a slow running generator and conventional or controllable pitch propeller Nowadays, PMSGs are gaining increased interest in SG systems. Their main benefits include the elimination of the field winding, brushes and slip-rings, higher efficiency, since copper losses in the rotor are eliminated, and larger torque density. Radial field Surface Mounted Permanent Magnet (SMPM) generators are the most common topology though more complex topologies, such as axial field PM generators and transverse flux PM generators, are also considered. PMSGs with fractional-slot concentrated winding (FSCW) are gaining great interest in direct-drive applications, such as wind turbines and electric vehicles. Additionally, recent trends in power electronics might enhance the establishment of direct-driven PMSGs in SG applications, connected to the electric grid via an AC/DC/AC converter. However, there are several issues to be addressed regarding PMSG operation, the most notable of which being the lack of control of the excitation, especially under short circuit operation. Also, the rotor geometrical configuration (surface/internal magnets) should be chosen according to the nominal speed and the respective risk of demagnetization. Initially, a preliminary design process for both synchronous generator configurations has been implemented, applying fundamental machine equations. During the preliminary design process, the main specifications of the synchronous machine have been determined i.e. nominal speed and power. Specifically, the nominal speed is modified if the machine is directly mounted to the main engine or driven by a gearbox. Also, the generator s nominal power and phase voltage is determined according to ship s electric network data. Following this, the machine s main dimensions have been calculated, taking into consideration typical electric and magnetic loadings. For the final design, a 2D finite element (FE) model was employed, in order to modify the generator s initial geometrical configuration for a detailed calculation of machine electromagnetic properties. Afterwards, a parametric design process was developed in order to perform sensitivity analysis with variables being the machine s main geometrical parameters. The optimization algorithm executed is shown in Figure WP2_4. The algorithm s main variables for the PMSG and the SPSG are: X 1 =[tw, mw] X 2 =[If, tw, dsy]
where tw and mw are the tooth and magnet width respectively. In addition, If, tw and dsy are the field current, tooth width and stator yoke. Using the parametric design algorithm a PMSG was designed for PTO/CFE SG application. In a first step, the main specifications of the machine are determined: nominal speed 400 RPM and power 2.4 MW. Since the generator is designed for direct-drive application, the nominal speed was determined at the same value as the main propulsion engine of an actual Ro-Ro ship. Figure WP2_4. Flowchart of the proposed algorithm A surface-mounted PM configuration is chosen, because of its manufacturing simplicity and low cost. A 20-pole machine is considered, in order to maintain the frequency relatively high, which is quite important for the power converter s design and operation, reducing the dc lick capacitor value. A double-layer FSCW is proposed, owing to the lower iron losses in the rotor and more sinusoidal EMF waveform than the single-layer winding. In Figure WP2_5 the PMSG s magnetic flux density distribution for the final PMSG is illustrated for nominal operating conditions. Figure WP2_5. PMSG Magnetic flux distribution under full load operation
In a second step, a Salient-Pole synchronous machine of the PTO/RCF topology is designed. The generator s nominal power and voltage are the same as the PMSG s, with nominal speed 1200 RPM. This is a typical configuration for used in ship shaft generator systems. A 6-pole machine was designed to produce 60 Hz nominal electric frequency. A full-pitch distributed winding with 2 slots per pole per phase was adopted. The machine s magnetic flux density distribution for the final SPSG is illustrated in Figure WP2_6 for nominal operating conditions. The final PMSG and SPSG parameters are tabulated in Table WP2_1. Figure WP2_6. SPSG Magnetic flux distribution under full load operation Table WP2_1.PMSG&SPSG parameters PMSG SPSG P(MW) 2.4 2.4 n (rpm) 400 1200 Te (knm) 57.3 19.1 f (Hz) 66.7 60 Nominal Voltage (V) 600 450 Nominal Current Density (A/mm 2 ) 4 4 Poles 20 6 Slots 24 36 Airgap Diameter (cm) 150 90 Airgap length (mm) 5 5 Active length (cm) 48 63 Tooth percentage (%) 43 44 Magnet arc (%) 76 - Magnet length (cm) 2 - Nominal current (Α) 2566 3421 Field Current (A) - 340
Stator yoke (cm) - 2 Motor weight including housing (tn) 9.25 5.81 Volume (m 3 ) 1.86 1.27 The comparison of the main operating characteristics between the two machines has shown that the PMSG configuration is advantageous in terms of torque ripple, EMF harmonic content and efficiency. Furthermore, for the proposed SG systems, total weight and volume are calculated. For the direct-driven PMSG SG system, the motor weight is presented in Table WP2_1. Moreover, this configuration employs a 2.4MW back to back converter for the voltage and frequency regulation, as well as the control of the produced active and reactive power of the PMSG. The total weight and volume of such converter system is about 4.3tn and 5.8m 3, according to similar converters applied to wind generators [7]. Therefore, the total weight and volume of the low-speed direct driven PMSG SG system are m total =13.55tn and V total =5.8m 3, respectively. For the SPSG SG system, a 1:3 2.4MW gearbox for the shaft speed motor adaptation and an AVR system for the field winding are required. The typical weight and volume values are considered as 7.5tn and 6m 3, respectively [8]. Therefore, the total weight and volume of the high-speed gearbox driven SPSG SG system are m total =14.1tn and V total =7.27m 3, respectively. It should be noted that the elimination of the gearbox can reduce significantly the total SG system weight and volume, regardless the motor weight increase, due to the higher torque demand. The PMSG configuration provides a higher torque to mass ratio, due to the presence of permanent magnets and concentrated winding. As a result, a low-speed PMSG with fractional-slot non-overlapping winding, connected to the ship s electric grid via a power converter, can be considered as an attractive option for SG systems. The proposed control system for the proposed direct-driven PMSG, as shown in Figure WP2_7, for the shaft generators consists of the machine side converter control, which includes active power control during generator mode and speed control during motor mode, and the grid side converter control that includes the grid voltage control and DC link voltage control. The control system allows variable speed operation of the main engine without affecting the frequency of the grid, allowing the shaft generators to operate in parallel with the DGs. Moreover, the proposed SG system could be implemented efficiently into vessels with lower shaft speeds, for example in cargo ships, with the utilization of a capacitor with higher capacitance values, or with PMSG with higher number of poles, in order to keep the ripple in the dc link in low levels.
Figure WP2_7. Overview of Shaft Generator control system Control on the machine side converter is based on the technique of Field Oriented Control, which is a common method for torque control of PM machines. On the grid side converter control is performed on a d q reference frame, synchronously rotating at grid s frequency, by two parallel controllers [4]. The modulation technique (Space Vector Modulation) of the grid side converter gives the opportunity to apply this particular drive system to electric grid with different frequencies (50 or 60 Hz), as the angle imposed in the d-q transformations (Park and Clarke transformations), in order to produce the reference voltages for the inverter pulses is regulated by phase lock loop (PLL) in electric grid phase voltage [9]. As part of this work, data was collected from the electrical system of a modern cargo ship (RoRo- Trailer-Passenger). This is a twelve year old ship built in Germany. The ship has 4 diesel engines 12.000 kw, 3 diesel generators 2.100 kva, 2 shaft generators 2.400 kva and one emergency diesel generator 1.125 kva [3]. A model was built in MATLAB/SIMULINK using permanent magnet shaft generators and the proposed control system. Simulations were carried out for the following four operating conditions: Operation of the electrical system of the ship with only the diesel generators Parallel operation of shaft generators and diesel generators Isolated operation of the shaft generators Emergency operation where the shaft machine operates as a motor powered by diesel engines In the following figures the results for the parallel operation of the SGs and DGs are given. For the results of the other operating conditions please refer to the complete report. The simulation duration was 4 seconds and the thrusters were switched on after 1.5 seconds. The percentages of the generators loads are given below.
Power (KW) Sea service (0-1.5 sec) Maneuvering service (1.5-4 sec) Diesel generators 1680 18.5% 89%,89%, Shaft generators 1920 80%, 80% 54.7%,70.6% 800 600 400 Grid Voltage Vabc (V) 200 0-200 -400-600 -800 0 1 2 3 4 time (sec) Figure WP2_8. Grid voltage Vabc Figure WP2_9. FFT Analysis Current and Voltage Figure WP2_10. Active and Reactive power
Voltage drop of 10% has been observed while inserting the thrusters which correspond to 69% of the existing ship s load. The voltage returns to 98% of its nominal value at 0.8 sec. The specifications define the limit of voltage drop at 16% of its nominal value for 20 sec. Total harmonic distortion of 2.42 % for the voltage and 3.18 % for the current. The specifications define a maximum of 5% for the total harmonic distortion for both the voltage and current [5]. Furthermore, in order to validate the flexibility of the proposed control model, the same simulation with a 50Hz electric grid, that is the DGs nominal speed is 1000rpm from 1200rpm, is implemented. The respective grid voltage response is illustrated in figure WP2_11. Grid Voltage, Vab (V) 800 600 400 200 0-200 -400-600 Grid Voltage, Vab (V) 600 400 200 0-200 -400-800 0 0.5 1 1.5 2 2.5 3 3.5 4 Time (sec) -600 2 2.005 2.01 2.015 2.02 2.025 2.03 2.035 2.04 Time (sec) (a) (b) Figure WP2_11. Phase voltage response for 50Hz electric grid. Following the above mentioned simulations, the specific fuel oil consumption (sfoc) was calculated for the main engine and diesel generator for a typical journey of 7 hours (6 hours sea service, 1 hour maneuvering). The fuel consumption was calculated for the existing configuration of the examined vessel. Optimization of the existing operational scenario was carried out by utilizing the proposed control system, which enables the parallel operation of SGs and DGs and the fuel consumption was again calculated after the aforementioned optimization. For the existing operational scenario the following generators loading were used. Sea Operation 1 shaft generator, 1 diesel generator SG DG ΜΕ Load % 66.86 73.25 95.6 Sfoc g/kwh 192.14 173.08 Maneuvering 2 shaft generators, 3 diesel generators
SG1 SG2 DG1 DG2 DG3 ΜΕ1 ΜΕ2 Load % 4.9 8.3 50 50 51 95.4 95.7 Sfoc g/kwh 197.23 197.23 196.88 173 173.12 According the above data the total fuel consumption was calculated as 3.478.118 gr. For the proposed operational scenario the following generators loading were used. Sea Operation 2 shaft generators SG1 SG2 ΜΕ1 ME2 Load % 71.33 71.33 95.71 95.71 Sfoc g/kwh 173.12 173.12 Maneuvering 2 shaft generators, 1 diesel generator SG1 SG2 DG1 ΜΕ1 ΜΕ2 Load % 40 40 81.36 98.2 98.2 Sfoc g/kwh 191.89 174.19 174.19 According the above data the total fuel consumption for the proposed operational scenario was calculated as 3.375.092 gr. This corresponds to a 3% reduction in fuel consumption. Furthermore, the combination of the proposed SG system with a waste heat recovery system, utilizing a combined cycle engine, can boost further the energy efficiency of the whole system, reducing the fuel consumption. References [1] MAN B&W, Shaft generators for the MC and ME engine. [2] Ayman M. EL-Refaie, "Fractional- Slot Concentrated- Windings Synchronous Permanent Magnet Machines: Opportunities and Challenges", IEEE Transactions on Industrial Electronics, 2010. [3] Paloumpis G., Transient Analysis of a ship s Power Grid, Diploma Thesis, School of Naval Architecture and Marine Engineering, National Technical University of Athens, 2011. [4] S. Merzoug, and F. Naceri, Comparison of field-oriented control and direct torque control for permanent magnet synchronous motor (PMSM), World Academy of Science, Engineering and Technology, 2008. [5] IEEE Industry Applications Society/Power Engineering Society, 1992, IEEE Recommended Practices and Requirements for Harmonic Control in Electrical Power Systems. [6] G.S. Stavrakakis, G.N.Kariniotakis, A general simulation algorithm for the accurate assessment of isolated diesel-wind turbines system interaction, IEEE Transactions on Energy Conversion, 1995.
[7] ABB, ABB low voltage wind turbine converters, ACS800, 0.6 to 6MW, 2013. [8] Wärtsilä Single Input Gear, http://www.wartsila.com/products/marine-oil-gas/propulsorsgears/gears/wartsila-single-input-gear. [9] F.A. Ramirez, M.A. Arjona, and C. Hernandez, Emulation of a single phase dspic based gridconnected wind energy conversion system, XIX International Conference on Electrical Machines (ICEM), 6-8 Sept. 2010. Notes & Comments on WP2 by the Evaluation Committee The case study looked at a Ro-Ro Ferry with a Power Take Off (PTO) generator system. The work comprised mainly the comparison of two different configurations one with Permanent magnet synchronous generator and another one with salient pole synchronous generator. The work was well received by the panel and they were happy with the progress and considered it almost complete as per schedule. The panel commented on: a. the applicability to lower shaft speeds typically found on other types of ships. b. practical design considerations such as volume and mass should also be appreciated. c. The possibilities of the combination of PTI with waste heat recovery. d. The presence of 3rd harmonics between the generator and the conversion equipment is beneficial for 100% protection of the generator. There should not be attempts to minimize the third harmonic. This does not affect the power quality improvements in the distribution system of the ship. e. The integration of these systems to both 50 and 60 Hz power systems could be investigated. Answering Notes by the Research Team The Research took into consideration the comments and suggestions made by the Evaluation Committee. The following notes have been made in the interval between the interim and final evaluation:
a. Few other ship types (e.g. old container ships) also exist with a PTI installed. Ship designers and ship owners have restarted considering of PTI in few new builds. Thus, it is expected that experience gained in this project via the considered ro-ro ferry (the ship grid of which covers several case studies of other ships, too) will be exploited in the near future towards the optimization of the generators installed onboard. b. Indeed, the research team made the effort to focus on small sized configurations. To this end, the most appealing alternative that was investigated comprised a permanent magnet synchronous motor, i.e. one of the smallest and most efficient electric motors. c. Considering that in the framework of another project (acronym ECOMARINE a waste heat recovery unit based on TEG technology has been developed and constructed a mutual exploitation of results and conclusions is planned to take place. d. The remark is well noted and since the 3 rd harmonic is beneficial no special effort to reduce is to take place. e. This remark can be easily answered affirmatively considering that both frequencies are met depending on the ship type and its mission, while power converters can integrate the shaft generator systems to either operating frequency WP2 is considered concluded since the end of 2014, still the remarks made will be taken into account to the maximum possible extent especially in the final report WP3 - Investigation of PQ problems due to thruster operation Electrically driven auxiliary propulsion systems (also referred to as thrusters), most often with controllable itch propellers, have been introduced in the aft or stern part of several ship types, increasing their maneuverability and collision avoidance capabilities, as shown in Figure WP3_1. Thrusters are often three-phase induction motors of high power demand, of the order of 0.5 up to 2.5 ΜW, which increases considerably the electric power demand from the electric power generation set. Moreover, during start-up, the thruster motor -like any other motor- absorbs a high transient inrush current (varying, in general, between 4-7 times the rated current), during the acceleration of the rotor. The rotor s inertia is increased by the submerged propeller. Further, during the inrush current phenomenon, the thruster motor active and reactive power demands are high, too, while the corresponding transient power factor is fairly low, as the reactive power required is significantly higher than in steady-state. This high energy demand at a low power
factor cannot be easily covered by the vessel s generator sets, leading to their possible overloading or even tripping. In addition, as a result of the inrush current, large voltage drops take place in the entire power distribution network, introducing symmetrical voltage dips to all three-phases. Figure WP3_1: Aft and stern thrusters +++++++++++++++++++++++++++++++++++++++++++++++++++++ Investigation of torque demands during transient operation with Boundary Element Methods (BEM) Calculation of loads excited by propulsor s components during dynamic conditions: Azimuthing angle Non-linear path Blade rotation about spindle axis (controllable pitch) Propeller rotational speed
Quasi steady method for bow thrusters The transient phenomena from cold start lasts 1-2 min Relatively slow variations & simple geometry It can be approached with a quasi steady method, i.e. a sequence of steady state calculations from propeller series It provides a fast and reliable way to bypass computationally intense methods Case study Results of bow thrusters Bow thruster Parameters Tube: Length = 2xD,tip clearance = 5% Propeller: 1m, z=3, AE/A0=0.7, P/D= 1.0 Scenario: 40-80 RPM // blade angle 7 O -22.5 O Advance coefficient: 0.1
Simulation of propeller rotation in rectilinear advancement (J=0.768) Torque around propeller axis calculated by UBEM and by Quasi-steady
As it can be seen from Figure, the quasi-steady method, though faster, results in similar torque demands, which are, in general, smaller than those calculated via the UBEM method. ++++++++++++++++++++++++++++++++++++++++++++++++++++++ In this work a thruster propulsion system consisting of two pairs of squirrel cage Induction motor thrusters of 148kW nominal power each was examined. An optimization procedure comprising an extended sensitivity analysis concerning the shape of both stator and rotor teeth of an induction motor was proposed. The procedure was applied to NEMA designs B, C and D to determine the optimal design for an induction motor used in ship thruster propulsion. The final design was acquired after a series of analyses using the finite element method. The results were validated using a simple per-phase motor equivalent circuit. The design procedure proved to be suitable and adequately efficient. The final design fully aligns with the defined restrictions and specifications of the application. In addition, in this work-package, a model of a thruster system was developed corresponding to a newly built ship, using Simulink / Matlab simulation software. The behavior of the grid and generators, for direct on line start-up of the thruster when connected to the bus, was studied. An electronic power converter has been proposed in order to start smoothly the large induction motor. The electronic soft starter may be bypassed after the startup in order to save energy. Induction motors classification has been proposed according to NEMA in classes as given below: Applying sensitivity analysis technique of geometrical parameters of the motors, as shown in the Flow Diagram below, resulted in optimal geometries for each class.
Following the preliminary design process the design of the electrical motor has been finalized, so that it matches the specifications, using Finite Element Analysis software called FEMM for classes B, C, and D. In Figure WP3_2 the results of the Finite Element Analysis are shown indicatively for class D before and after the sensitivity analysis. Figure WP3_2: Flux density distribution at maximum torque before and after the sensitivity analysis
The initial design resulted in an uneven distribution of the flux density. The stator iron had a flux density beyond the desired value of 1.8T. After the sensitivity analysis the flux is more evenly distributed. The results from the sensitivity analysis for the torque of the motor are given below: In Figure WP3_3 the torque vs slip graph is shown before and after the sensitivity analysis where it is evident that the maximum torque has increased by19% from 826 Nm and is 1020 Nm. The nominal slip has the same value of 0.006, which meets the speed requirements.
Figure WP3_3: Torque vs slip graph Torque slip results has been compared for each class, as shown in figures WP3_4 and WP3_5 where it can be seen that ΝΕΜΑ design C motor has higher torque capabilities and NEMA design D motor has higher slip at maximum torque but only NEMA design D motor is able to produce starting torque 1.3*Tn at starting current 5.8*In. Therefore NEMA design D has been selected that best meets the specifications. Figure WP3_4: Torque slip graph of NEMA class design motors
Figure WP3_5: Starting torque starting current graph of NEMA class design motors Having determined the NEMA design class of the motor and the geometrical characteristics that meet design specifications, while maintaining the load of the motor to desired levels, the phase equivalent circuit of the motor s steady state operation was determined. The estimation of the parameters was done using the method of least squares and as a result the characteristic torque - slip curve of the motor gave a clear indication of the predicted total behavior, as shown in figure WP3_6.
Figure WP3_6: Torque slip curve comparison The deviation between the two curves is due to the use of a simple equivalent circuit which does not take into consideration the skewing of the rotor slots and the end-rings. As mentioned earlier in a ship s electrical powergrid, power demand is covered by diesel generator sets and shaft generators. High instantaneous reactive power demand from thrusters can stress the ship s generators to its limits. The generator AVR has to be designed for fast and accurate response. For the smooth start-up of the thrusters two types of power electronic devices have been studied, a three-phase back to back Silicon Controlled Rectifier (SCR) (figure WP3_7) and a three-phase AC/DC/AC converter (figure WP3_8). Figure WP3_7: A three-phase back to back Silicon Controlled Rectifier (SCR)
Figure WP3_8: A three-phase AC/DC/AC converter In order to control these converters two types of control techniques have been studied. For the three-phase back to back Silicon Controlled Rectifier (SCR) voltage and current control has been modelled and for the three-phase AC/DC/AC converter scalar and vector control were modelled. A real power grid network from a ferry, was simulated using the software package MATLAB/SIMULINK (figure WP3_9). Two 2.1MVA gensets and two 2.4MVA shaft generators (AC bus at 440V/60Hz) were modelled. From onboard measurements it was known which loads where operating during the manoeuvring and included to the simulations. Thruster s 1MW induction machine equivalent circuit was estimated from the manufacturer s datasheet and a precise model has been created. In order to study the effects of thruster start-up on the electric power quality, three different operating conditions were simulated, thruster Direct on Line (DoL) operation, thruster soft starting through the three-phase back to back Silicon Controlled Rectifier (SCR) and thruster soft starting through the three-phase AC/DC/AC converter. Figure WP3_10 shows the current and voltage waveform during thruster start-up in DoL operation. As a voltage drop appears, the generators AVR start to increase magnetization in order to compensate for the increased load currents of the generators. As the motor reaches the pitching moment, the stator currents quickly reduce and since the generators are now temporarily over-magnetized, a voltage overshoot occurs. As a consequence, high inrush currents during this transient period may cause a significant voltage disturbance in the power system.
Figure WP3_9: Ship electric power system model Figure WP3_10: Current and voltage waveform during motor start up in DOL operation For the thruster soft starting through the three-phase back to back Silicon Controlled Rectifier (SCR) operating condition two control schemes (voltage and current control) were modelled. Figure WP3_11 shows the results for the two schemes. For the Timed Voltage Ramp scheme open-loop control is used. Also there is no need for current measurement and it has a shorter startup time. For the Current Limiting scheme the thruster startup current is limited, the voltage drop is minimized and it has improved power quality. To implement the logic of scalar and vector control, an inverter AC / DC / AC was modeled. The first stage of the converter consists of a diode bridge rectifier, while the inverter is constructed with IGBT semiconductor elements. Furthermore it should be noted that the motor accelerates having a linear load on its shaft that reaches a torque of around 2500Nm after 8 seconds. Figure WP3_12 shows the results for the thruster soft start operation using a three-phase AC/DC/AC converter between scalar and vector control. For the scalar control large motor current variations can be observed causing in turn major variations of the bus voltage. The current variations are comparable in magnitude to the DoL operating condition. Also, the rotor speed is not smooth.
Figure WP3_11: Speed, current and voltage waveforms comparison between voltage (left) and current (right) control during thruster soft start operation
Figure WP3_12: Speed, current and voltage waveforms comparison between scalar (left) and vector (right) control during thruster soft start operation Vector control succeeds in precisely controlling the speed of the motor with reduced initial current. The voltage drop and the voltage transient is reduced. Unlike the scalar control, the motor speed is smooth with minimal torque ripple. From the study it can be seen that it is possible to reduce current and voltage transients during thruster start-up through power electronics soft starters. In particular, the three-phase back to back rectifier with current control, is considered as a satisfactory solution with low cost and volume in order to address the problems associated with thruster start-up. In addition, the use of AC/DC/AC converters, using vector control can cope with even more stringent restrictions regarding limiting the starting current but with the disadvantage of significantly increased cost and complexity. Notes & Comments on WP3 by the Evaluation Committee The first stage was the data collection of such systems. The second stage was a hydro-dynamic study of the transient torque demand of a CPP thruster. The third stage comprised the design of the thruster electric motor and its associated power interface, the modeling of this system, and the analysis of the power quality of it during the transient start-up phase. The panel is happy with the good progress. The panel commented on: a. In future projects it would be good if there is an advanced control mechanism for pitch angle change rate for the CPP and motor load/ overload behaviour b. The case studied is an ideal one (i.e. starting-up begins at zero blade pitch angle). Extra overloads are experienced during start up, for example when the bow thruster is a solid pitch bladed propeller, or a piece of wood or rope or other obstacles are present inside the tunnel thus impeding motion. c. In future work, the way to integrate this into a dynamic positioning system could be developed.
d. The impact of the thruster into the entire system in the light of energy storage devices needs to be considered. Answering Notes by the Research Team WP3 is considered concluded WP4 Investigation of PQ problems due to pod operation A pod is a propeller propulsion device installed externally to the ship s hull. The pod system comprises a short propulsion shaft, on which the propulsion electric motor is mounted. In most cases, propulsion is achieved by means of fixed pitch propellers powered by an AC synchronous motor (conventional, permanent magnet, or high temperature superconductive) or asynchronous electric motor, fitted inside the pod. Amongst the highest failure indices are those of main electric motor parts: stator windings and core rotor windings slip ring electric connectors These high failure rates are attributed mainly to the high temperature development in windings, leading to excessive overheating stresses, indicating that the motors have not been properly selected according to the stressful conditions they are subject to operate in. This work package deals with the design and optimization of the geometry of a surface mounted permanent magnet synchronous motor for application as a podded main propulsion motor. Specifically, it is an Azipod system that is an outboard marine propulsion unit, in which the engine is out of the hull, is rotatable around itself and the shaft is directly connected to the propeller. ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Investigation of torque demands during transient operation with Boundary Element Methods (BEM) Calculation of loads excited by propulsor s components during dynamic conditions: Azimuthing angle Non-linear path
Blade rotation about spindle axis (controllable pitch) Propeller rotational speed Geometrical modelling First the Real Model is considered
Geometric Breakdown (patches) Motor housing Strut Propeller Blades Boss Cap Fillet Fin Estimating torque demands of electric driven pods and thrusters 26 41 National Technical University of Athens Then the Surface model is developed via Offset e.g. propeller blades or Mathematical Equations e.g. housing Finally the Surface grid is generated via interpolation schemes.
The increased number of elements results to increased Accuracy of the model Converteam Schottel SSP
Motion Generation: this is achieved via superposition of the motion of the following components: 1. Propeller rotation 2. Blade rotation around spindle axis 3. Propulsor rotation around vertical axis 4. Translation Case studies Podded propulsor Parameters Configuration: Azipod XO 2100 Propeller: 6m, z=4, AE/A0=0.7, P/D= 1.0
Strut airfoil: NACA 0012 Scenarios: Rectilinear, Yaw +/-20 o Advance speed: 7.2 m/s @ 93 RPM Grid: Sparse - 1648 elements Azimuthing angle scenarios: Rectilinear motion (0 o yaw) Yaw +20 o (const. rate turn) Yaw -20 o (const. rate turn) X-axis velocity pattern: Step 1-30: Const. acceleration Step 31-240: Const. x-axis velocity
Results of podded propulsors Torque around propeller axis Oscillation due to strut propeller interaction Strut yaw affects loads
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ As part of the work, the design specifications have been selected, given the constraints of the application, and the design and optimization of the geometry of the machine has been carried out, with the main purpose to increase performance, while maintaining high efficiency and power quality. The motors used in this application are characterized by very high power (of the order of 4-5 MW), low rotational speed (approximately 200 rpm) and very high efficiency (due to the large size the theoretical efficiency can reach 98%). The design specifications chosen for the machine are as follows: The power was chosen to be 4,5MW, eg sufficient for small and medium passenger vessels with a maximum cruising speed of 21 knots. Motor speed (and thus the propeller) 200 RPM. Voltage 690V, 50Hz. This voltage level is common in electric marine propulsion systems even for propulsion motors of the order of MW, in order to have low insulation requirements.
Moreover, due to the outer positioning of the motor directly to the propeller, the motor was chosen to have a maximum outer diameter of 1,5m, so as not to obstruct water flow. Based on the selected design requirements set out above, the preliminary design of the machine was carried out, which concerns the selection of the air-gap, the configuration of the rotor and stator and the calculation of the magnetic and electric loading. An important consideration was the selection of a single layer winding or a double layer winding. Single-layer windings have coils wrapped every two teeth, while in the double layer windings coils are wound on each tooth, as shown in Figure WP4_1. Figure WP4_1: Single layer (a) and double layer (b) winding For the design of this motor double layer winding was selected, due to its lower torque ripple and low losses. Indeed, some early simulations using the software package FEMM showed that for the same magnetic circuit single layer winding gives much higher values of magnetic flux, which leads to saturation of the iron and high losses. Also, for this application the motor does not require constant power over a wide range of speeds, so this advantage of the single-layer winding was not considered important. In figure WP4_2 below we see two instances of the same motor with a double layer winding (left) and single-layer (right). The saturation of the magnetic circuit of the second motor is apparent. Figure WP4_2: Comparison of the magnetic field of the motor with double layer (left) and single layer (right) winding Other disadvantages of the single-layer winding are increased iron losses and high harmonic distortion of the back EMF.
Figure WP4_3 shows the cross section of the motor after the preliminary design showing details such as materials selection and winding distribution in slots. Figure WP4_3: Cross-section of preliminary design of motor During optimization of the motor design, sensitivity analysis was performed of the operating characteristics of the machine changing various parameters of the geometry. This analysis was done for three values of current, nominal, 1/4 of the nominal and twice the nominal, so as to study the behavior of the machine in under-load or over-load conditions. The selection of the parameter values of the geometry is carried out for rated current condition, while the simulations with 1/4 of the nominal current and double the nominal current are used to check the design for normal torque values, volume losses etc. Two parameters that were optimized were the magnet width expressed as percentage of the pole pitch and the tooth width. Through Finite Element analysis several motor characteristics were computed, such as average and maximum torque, torque ripple, back-emf, iron losses, etc. It was found that the larger the width of the surface mounted magnet, the greater the output torque of the motor. The torque increase is not linear and shows a "knee" for magnet widths over 80-85% of the pole pitch. Therefore one needs to consider how the increase of the motor performance justifies the higher cost. Figure WP4_4 shows the result of the sensitivity analysis of the torque versus tooth width and magnet width expressed as percentage of the pole pitch.
2.4 x 10 5 Μέση ροπή (Nm) 2.3 2.2 2.1 X: 33 Y: 0.85 Z: 2.246e+05 1 2 0.9 0.8 magnet ang (Πλάτος μαγνήτη/πολικό βήμα) 0.7 Figure WP4_4: Graph of torque versus tooth width and magnet width expressed as percentage of the pole pitch In the second stage of optimization again the same machine characteristics were considered for three different values of current, but this time by varying the width of the tooth tip as a function of the tooth thickness of the tooth (k1) and the thickness of the tooth tip as a function of the length of the tooth. Figure WP4_5 shows the result of the sensitivity analysis of the average torque versus the width and the thickness of the tooth tip. 30 32 34 36 38 Πλάτος δοντιού (mm) 40 Μέση ροπή (Nm) 2.6 2.4 2.2 2 x 10 5 X: 2 Y: 0.05 Z: 2.513e+05 1.8 0.25 0.2 2 0.15 1.8 1.6 0.1 1.4 1.2 0.05 1 k2 (Πάχος πέδιλου/μήκος δοντιού) k1 (Πλάτος πέδιλου/πάχος δοντιού) Figure WP4_5: Graph of torque versus the width and the thickness of the tooth tip It has been observed that increasing the width of the tooth tip of the stator teeth provides increased torque. As the width of the tooth tip approaches the width of the surface mounted magnet the leakage flux is reduced and there is a better utilization of the permanent magnets.
Following the optimization procedure the basic dimensions of the final design have been selected and shown in the table below. Motor active length Stator outer diameter Stator inner diameter Inner rotor radius Outer rotor radius Thickness of magnets Width of magnets Stator tooth width Stator tooth length Tooth tip width Airgap length Stator and rotor iron Magnet type 2586 mm 1365,6 mm 1325,9 mm 573,8 mm 610 mm 12 mm 91,6 mm 33 mm 35 mm 66 mm 6 mm Thyssen M 235-35 A NdFeB 40 MGOe Coil turns per phase 8 The motor displayed the expected behavior in overload conditions, and has shown to provide a satisfactory response to overload conditions, with a slight decrease in performance but with a large increase of voltage harmonic distortion. In a next step, the equivalent circuit parameters are calculated, in order to develop a convenient dynamic model for the designed pod. Initially, the armature resistance is calculated analytically. The equation calculating the armature resistance is: RR pphaaaaaa = ρρ cccc LL pphaaaaaa SS wwwwwwww (1) where p cu is the copper special thermal resistance in 80 o C, L phase is the total wire length of the winding and S wire is the wire diameter. For the calculation of the wire diameter the estimated fill factor is taken into consideration, while for the calculation of the wire length, the motor active length and the end winding length are utilized.
The flux linkage induced by the Permanent Magnets (PMs) and the d-q axis synchronous stator inductances for various electric loadings are calculated by the developed parametric fixed-step FE model. In case of calculation of the PM induced flux, the armature currents are set to zero, while for the calculation of the synchronous inductances, the PM material is replaced by air. In a next step, the d-axis of the rotor is aligned with the magnetic axis of the phase A and the flux linkage of the phase A is calculated by the following equation: QQ 2 pp ΛΛ aa = 2 pp LL FFFF nn qq kk nn qq 1 AA PPPP SS zz dddd qq qq=1 SS qq (2) where p is the pair of poles, n q the conductors per slot, n pp the number of parallel coils in the winding, k q = 1 is the q th slot is included in phase a winding and the current is positive, k q = -1, if the q th slot is included in phase a winding and the current is negative and k q = 0, if the q th slot does not include conductors of the phase a winding, S q is the slot area and A z is the magnetic vector potential for a fundamental area ds. The d-q axis inductances for various electric loadings are calculated by the following equation: LL dd = ΛΛ aa = ΛΛ dd II aa = II dd (3) where I a is the phase A current, which is equivalent to the d-axis current in this case. It should be noted that the d and q axis inductances are assumed to be equal, due to the rotor structure of the surface-mounted Permanent Magnet Synchronous Motor (PMSM). The magnetic flux distribution for the PM flux and d-q inductances calculation is illustrated in Figure WP4_6. The synchronous inductance with current density for the designed PMSM is shown in Figure WP4_7, where it can be observed that the magnetic circuit is not saturated for adequate high electric loads (10A/mm 2 ), highlighting the benefits of the implemented sensitivity analysis. (a) (b) Figure WP4_6: Magnetic flux distribution for the magneto-static FE solutions for the computation of PM induced flux (a) and d-q axis inductances at nominal load (b)
d&q axis inductances (mh) 0.25 0.2 0.15 0.1 0.05 0 2 3 4 5 6 7 8 9 10 11 12 Current density, J (A/mm2) Figure WP4_7: D&q axis inductances with current density for the designed PMSM. Following the calculation procedure analyzed above, the equivalent circuit parameters for the designed PMSM-pod are tabulated in the table below. Poles 30 Nominal Phase Voltage Nominal Current Nominal produced Torque 400 V 4.595 ka 251 knm Rotor moment of inertia 4056 kg.m 2 Nominal speed Armature resistance d&q axis stator inductances at nominal current Induced flux linkage by PMs 200 rpm 1.1 mω 0.19 mh 0.84 Wb Notes & Comments on WP4 by the Evaluation Committee The first stage was the data collection of such systems. The second stage was a hydro-dynamic study of the transient torque demand of a azimuth propulsor. The third stage comprised the design of the azipod electric motor and its associated