FUTURE ENERGY STORAGE SOLUTIONS IN MARINE INSTALLATIONS - FESSMI - FINAL REPORT

Transcription

1 FUTURE ENERGY STORAGE SOLUTIONS IN MARINE INSTALLATIONS - FESSMI - FINAL REPORT Kimmo Kauhaniemi, Jagdesh Kumar, Jari Lahtinen, Aushiq Memon, Hayder Al-Hakeem, Omid Palizban, Lauri Kumpulainen, N. Rajkumar University of Vaasa Olli Pyrhönen, Antti Pinomaa, Tuomo Lindh, Pasi Peltoniemi, Andrey Lana, Henri Montonen, Kyösti Tikkanen Lappeenranta University of Technology

2 2 CONTENTS SUMMARY 3 FOREWORDS 5 1. INTRODUCTION 6 2. WP 1 LIFE CYCLE ANALYSIS FOR HYBRID MARINE TECHNOLOGY Cruise ferry OSV TUG WP 2 HYBRID VESSEL HARBOUR SUPPORT SYSTEM MODELLING AND ANALYSIS Harbour Area Smart Grid Protection studies WP 3 HYBRID VESSEL ELECTRIC SYSTEM MODELLING AND ANALYSIS Diesel generator set Battery Energy Storage System Propulsion Drive Cruise ferry OSV WP 4 REMOTE VESSEL DATA MANAGEMENT WP 5 DEMONSTRATION SYSTEM DEVELOPMENT AND TESTING AC network emulator hardware and communication structure DC network emulator hardware and communication structure Real time simulators LIST OF REPORTS AND PUBLICATIONS 34

3 3 SUMMARY In this research project new energy efficient solutions for marine vessels taking full advantage of hybrid technology and battery energy storages were studied and developed. In addition, harbour support infrastructure, which allows vessel energy storage interaction with on-shore grid and renewable resources was investigated. Battery technology is developing fast due to increased utilization of electric vehicles. Until now battery energy storage has not been used in large extent in marine applications, but there is an increasing interest towards higher utilization of electric storages for reducing emissions, and fuel savings, and improved energy efficiency of modern all-electric and hybrid vessels. The utilization of batteries offers power balancing, reserve power and short response times for marine vessel power systems. In this project, both technical and economic feasibility of using battery energy storage technology in vessels were studied by simulations and laboratory experiments. Since load profiles and power systems have large variance in different vessel types, three different vessel types were selected for detailed case studies. Based on the results got from vessel hybridization studies, it was shown that by adding batteries to vessels even with current battery price levels, remarkable total savings, measured and estimated with following measures; vessel CAPEX, OPEX, and improved fuel efficiency, can be reached in each vessel type. Moreover, the effect of different vessel power distribution systems on total savings were analyzed in the studies. Dynamic behavior of the vessel power system was analyzed by building simulation models of the case vessels and simulating the models with actual power data retrieved from the case vessels. Based on the simulation results battery energy storage on board can improve fuel efficiency by allowing diesel generator set to be run on better fuel efficiency region. Dynamic performance of the vessel increases as the time constant of the battery is much shorter than the time constant of a diesel engine. For instance, off-shore support vessel (one of the analyzed vessels) can shutdown one of its diesel engines during dynamic positioning while using battery energy storage to level load fluctuations without compromising its performance. Two emulator systems for hybrid diesel electric vessel power train were created in the project. The first one is for testing the AC power distribution and another for the DC distribution.

4 4 The emulator is a tool for research of energy balancing between diesel generators and a battery energy storage. Different control laws were tested for both system energy balance and battery energy storage control as a part of the system. Also, the emulator is used to verify the simulation models of the hybrid power trains developed in WP3. The aim of the simulations and emulations is further to study the fuel economy, dynamic performance and the stability of developed control laws. In addition to on-board energy storage systems, future vessel types can take advantage of harbour based renewable energy production and energy storages. For that purpose an innovative design concept of harbour area smart grid (HASG) was developed in this project, which can supply shore to ship power during stay of vessels at a harbour as well as support hybrid and fully electrified vessels. In the project, feasible alternative topologies for the HASG was developed considering different operating scenarios. Simulation models were used for the dimensioning of the key elements and checking the system performance.

5 5 FOREWORDS This report summarizes the results of the FESSMI project, which is executed as a joint action between Lappeenranta University of Technology (LUT) and University of Vaasa (UVA). At the LUT the responsible leader of this project was Prof. Olli Pyrhönen and in the UVA Prof. Kimmo Kauhaniemi. The project was a part of the INKA-program of Tekes, and the funding came from the European Regional Development Fund (ERDF). In addition to the public funding, the project was also funded by the following industrial partners (representatives in steering group given in parentheses): Wärtsilä (Tuomas Linna) Danfoss (Hannu Sarén) VEO (Ari Pätsi) CLS Engineering (Timo Hanhimäki) We are very grateful for all the partners and authorities for the financial support to this project. Also, the active engagement of the industrial partners is acknowledged. Authors

6 6 1. INTRODUCTION This project investigated new short shipping concepts for environmental friendly and energy efficient vessel drive technology, where large-scale on-board battery storage combined with new control concepts and topologies play a key role. In addition to that, research was focused for harbour-based support system technology for hybrid and fully electrified vessel types, where connection to renewable energy production, harbor-based energy storages and system grid integration are investigated on the conceptual level. Finally, the project aimed for such results, which can be later demonstrated in a separate full vessel scale demonstration project on an EU-level project. The project is divided into five work packages. The work packages of the project are: WP 1 Life cycle analysis for hybrid marine technology (LUT, UVA) WP 2 Hybrid vessel harbour support system modelling and analysis (UVA) WP 3 Hybrid vessel electric system modelling and analysis (LUT) WP 4 Remote vessel data management (UVA) WP 5 Demonstration system development and testing (LUT) Main results from all the WPs are summarized in the following chapters.

7 7 2. WP 1 LIFE CYCLE ANALYSIS FOR HYBRID MARINE TECHNOLOGY Battery technology is developing fast due to increased utilization of electric vehicles. Marine vessels are new and interesting application area for battery energy storages. The utilization of batteries offers power balancing, reserve power and short response times for marine vessel power system. Expected large volumes in electric vehicle battery production is bringing technology price down, which gives new possibilities also to other application areas, like marine vessels. FESSMI project investigated possibilities to utilize battery technology in marine applications. Since load profiles and power systems have large variance in different vessel types, they are studied case by case. Three cases, from different marine segments, were studied; hybridization of cruise ferry, off-shore / platform support vessel (OSV/PSV), and TUG Cruise ferry The first study case focuses on battery storage options for cruise ferry type of vessel. Both technical and economical feasibility have been investigated. Battery technology is not the only option for electric energy storage. Electric energy can also be stored into fly wheels or supercapacitors etc. These technologies offer very high maximum power and low energy density compared to batteries. In cruise ferry case, the batteries offer the best power-energy density combination, thus other technologies are not investigated in this study. There are various battery technologies on the market. Important features to be considered in marine battery technology are available cycle life, energy density, maximum peak power as well as present and future estimated price. Most favorable battery chemistry solutions are based on lithium-ion-batteries. Strong driver for lithium-ion-battery market and technology development are electric vehicles, where lithium technology is selected without exception. There are several different versions of lithium-ion batteries. Most promising technologies for vehicles as well as for marine applications are lithium iron phosphate (LFP), lithium nickel manganese cobalt oxide (NMC) and lithium titanate (LTO). In all these battery chemistries one electrode is made of graphite (negative in LFP and NMC, positive in LTO) and other electrode material is according to battery name. The essential technical differences in these battery types are maximum discharge rate C (LFP; C=1, NMC; C=2, LTO; C=10), number of life cycle (LFP; , NMC; , LTO; ) and energy density (LFP;

8 Wh/kg, NMC; Wh/kg, LTO; Wh/kg). Battery investment cost together with cycle life and technical performance is an important measure for applicability in marine applications. The lithium battery technology price reduction is dependent on the used materials. While LTO cell price estimate for year 2020 is 600 USD/kWh (2014: 800 USD/kWh), the respective estimate for LFP is 440 USD/kWh (2014: 700 USD/kWh) and for NMC 300 USD/kWh (2014: 600 USD/kWh). In cruise ferry CAPEX and OPEX analysis NMC battery technology price and life time estimates were used due to highest energy density and most favorable cost condition. The battery feasibility in marine vessel depends heavily on the loading profile of the vessel. Load profiles differ in great deal in different vessel types. While off-shore support vessels (OSV) have high instantaneous power demand peaks e.g. due to dynamic positioning requirements, the cruise ferry has dynamically much lower requirements. These kind of differences do not allow to develop general rules for marine battery selection, but more detailed analysis is needed, where the specific requirements of the loading profile is taken into account. The basis in this study is the loading profile of cruise ferry under study. The vessel sails through archipelago between two harbours once per day, which gives a unique 24 h loading profile to the vessel loading. The harbor time for the vessel is just one hour at both terminal harbors, which requires high power capacity to the battery charging. Essential technical questions in the feasibility analysis are possible battery energy capacity, maximum power, system topology and charging method among others. When analyzing economical feasibility, battery system price, life-time estimate and energy capacity are most important parameters. Battery capacity relates to the possibility to reduce the total engine power, which could be replaced by the battery stored energy. The most relevant savings are reduced combustion-engine-related investment and higher fuel efficiency. The analyzing methodology in the study was based on comparison of four different scenarios, where battery size and concept has been varied. The different scenarios were: SCENARIO 1: Battery capacity 1.5 MWh, fixed mounting SCENARIO 2: Battery capacity 2 MWh, mobile batteries SCENARIO 3: Battery capacity 9 MWh, fixed mounting SCENARIO 4: Battery capacity 6 MWh, mobile batteries

9 9 The fixed batteries are installed into the ship machine room, while mobile batteries are located on a cargo deck, which enables fast change in harbor and low speed 24 h charging. The battery capacity has effects on the battery usage during the cruising. Small battery (Scenarios 1 and 2) is used for smoothing the combustion engine power profile, where fast power changes are handled with the battery capacity. In scenarios of large battery (Scenarios 3 and 4), the battery can be used to power leveling and diesel engine working point optimization. In all the analyzed scenarios, it is assumed that in the harbor the ship is in cold ironing mode, where all the power is supplied from the on-shore grid. Battery system in MWh-range requires large volume of space. Important issue is then battery installation options. This was analyzed for both fixed batteries and mobile versions. Mobile batteries would be a favorable solution for existing ship, since batteries could be located on a cargo deck and battery handling could be done using normal harbor cargo handling equipment. This, on the other hand, would cause some reduction in the cargo capacity; approximately 1% for container installed 2 MWh battery module increasing linearly as a function total battery energy capacity. Another benefit would be possibility to charge battery modules in the harbor within a longer time period. Longer charging time might increase the battery life time and harbor charging would allow usage of renewable energy, e.g. using local wind or solar power. Battery connection and disconnection would cause some extra costs, but those have not been analyzed in detail. Fixed batteries are an alternative solution for a newbuild ship. In Scenario 3 the large battery capacity would allow reduction of one diesel engine. This would free enough space for battery package in the engine room. When looking overall ship power system CAPEX, scenario 3 would also reduce the overall cost compared to present cruise ferry power, if the expected near future battery system prices are used. According to this analysis, the CAPEX for scenario 3 would be lower compared to present installation by the year In Scenarios 1 & 2, power system CAPEX would obviously increase, since in those scenarios the number of engines remains the same. The OPEX analysis is the most important factor, when economical feasibility is assessed. The biggest cost factor relates to the fuel consumption, other factors are additional electricity costs (due to cold ironing in the harbor), maintenance costs and emission (CO 2) costs. Im-

10 10 portant question in the fuel efficiency analysis is the optimal load sharing between the batteries and the combustion engines. Two different operational principles were used. In case of small battery (Scenarios 1 and 2), single engine operation is assumed. It means, that instead of running 2 engines at partial loads, one engine is running in the optimal point, and additional power is taken from a battery. In case of large battery energy capacity (Scenarios 3 and 4), zero emission operation is used. This means, that near the harbor engines are not running, but the power is taken from the battery alone. When battery discharge reaches 80%, normal engine operation is started. In practice, this might not be acceptable according to current safety regulation, but might change in future, when energy storages are in large use in vessels. The OPEX analysis gave yearly savings to all scenarios due to reduced 1) fuel consumption, 2) emissions and 3) maintenance. On the other hand some electricity costs were added due to hotel load and battery charging in the harbor. The payback times for scenarios were: 6 a for Scenario 1, 9 a for Scenario 2, 7.5 a for Scenario 3 and 20 a for Scenario 4. It is to be mentioned, that if reduced freight capacity is taken into account, the economics of Scenario 4 get worse. In the calculation battery system price 622 /MWh corresponding 2016 price level was used. The battery prices are getting lower, which will give shorter payback time in the future. The final comparative analysis was done between medium-voltage DC and AC vessel power systems including battery storage. The main benefit in DC compared to AC system is the possibility to run the engine at optimal speed as a function of a varying load. Specially at the partial loads the engine fuel consumption can be reduced using variable speed. The analysis show, that with a large battery (8 MWh, 14 MW for DC, 20 MW for AC) combustion engines can be driven in the optimal working point all the time giving fuel savings 8% for DC and 5.8% for AC compared to the original. In case of small battery, the purpose is to stabilize the engine loading profile. The analyzed smoothening operation requires battery capacity of 400 kwh and 5.8 MW. This requirement could be fulfilled with 600 kwh LTO (10C), 2 MW NMC (3C) or 6 MWH LFP (1C). The battery life time in different cases would be 4.5 a (LTO), 6 a (NMC) or 4.5 a (LFP). The fuel saving in AC battery system case found to be 0.3% while DC battery system produced fuel saving 5.8%. The above-mentioned results, related analysis methodology and used assumptions can be found more in detail in Life cycle analysis for hybrid marine technology Case: Cruise ferry public research report.

11 OSV The second study case focused on battery storage options for off-shore/platform support vessels (OSV/PSV). This vessel type has highly dynamic loading due to dynamic positioning requirements. This emphasis the benefit of batteries to response faster to power transient compared to diesel or gas engine. There are various battery technologies in the market. Important features to be considered in marine battery technology are available cycle life, energy density, maximum peak power as well as present and future estimated price. Most favorable battery chemistry solutions are based on lithium-ion-batteries. Strong driver for lithium-ion battery market and technology development are EVs where lithium technology is selected without exception. The technological variation in several lithium battery types relate to available maximum peak power vs. stored energy and maximum life cycle. High cycle life and peak power correlates straightforward also to higher battery cell prices. The battery feasibility in marine vessel depends heavily on the loading profile of the vessel, since load profiles differ in great deal in different vessel types. Off-shore support vessels have high instantaneous power demand peaks especially during dynamic positioning. For OSV, high peak power battery would be optimal from technical point of view, but the higher battery price due to peak power demand increases the battery cost. For that reason, analysis has been performed for three different lithium battery types; lithium iron phosphate (LFP), lithium nickel manganese cobalt oxide (NMC) and lithium titanate (LTO). LTO batteries provide highest peak power and cycle life, while NMC and LFP offer lower energy storage costs due to lower cell prices. The basis of the analysis in this study was the loading profile of OSV under study. Two weeks of operational vessel data has been used to simulate the detailed battery operation in different driving modes. Also annual data was used for analyzing overall profitability. Based on simulation analysis, a sufficient battery capacity was selected for different battery chemistries. Accordingly, investment costs (CAPEX) and operational costs (OPEX) for this particular OSV case were derived. A few alternative load sharing control strategies between engines and batteries were included into analysis for the sake of generality.

12 12 The main component in the OPEX are fuel costs. The fuel savings of ESS was investigated both in AC (alternating current, 50 Hz) and DC (direct current) grid cases. The essential advantage in DC grid is the possibility to have variable speed in diesel-generators. The operational analysis of annual load profile showed, that battery storage brings significant savings to the ship operation both in fuel consumption and engine running hours. In DC grid case the fuel consumption was reduced by 22% and engine running hours by 41%, while in AC case the corresponding numbers were 16% for fuel savings and 42% for engine running hours. The battery cycle life analysis resulted between a depending on the battery type and grid topology. The additional savings in OPEX are based on reduced maintenance costs due to lower engine running hours and also due to reduced CO 2 emissions. The fuel price range /MWh and emission price 0 20 /tn has been assumed. Today, there is no CO 2 emission cost, so the emission cost has been included mainly for possible marine vessel emission control in the future. The maintenance costs for the ICE was assumed to be around 20 /h. For both AC and DC grid option, the total OPEX saving vary between k per year. The CAPEX consists mainly on battery system and necessary electrical components and system technology. As an example, additional battery investment for a newbuild ship is in the range of k using present LTO battery prices. Combining OPEX and CAPEX calculations payback time for different battery options can be analysed. The shortest payback time, when emission cost is 0, for retrofit with present battery prices and high fuel costs (LNG: 60 /MWh) is 1.73 years using LTO battery technology or 1.86 years using NMC batteries. If fuel price is low (LNG: 20 /MWh), the corresponding payback times are 3.17 a for LTO and 3.40 a for NMC. The estimated decrease in battery prices (2030) would reduce the payback times further. For a new build ship at 2030 the most favorable alternative would be LTO battery with DC grid resulting payback time 0.66a for high fuel price. Even low fuel price would result short payback time 1.31 a. The required battery storage capacity and needed floor space depends on the chemistry. For retrofit LTO would be the most reasonable option due to lowest battery size, only 1.5 m x 2.6 m floor area with cabinet height 2.2 m for battery unit would be needed. The total battery mass would be 5.44 tn. Also converter could be installed onto OSV under study. Other battery chemistries require more floor area, which is a drawback in retrofit installation. If newbuild ship is considered, then all battery chemistry alternatives can be considered.

13 TUG The third case study was on the battery energy storage feasibility in TUG type of marine vessel. Actual TUG operational profiles were used for battery loading analysis. Vessel operation in tens of different TUG tasks were used in the analysis, the operational data containing over 50 hours and 150 nmi of vessel operation was given with 1 minute resolution. Hybrid system scenarios on different diesel engines, PTI/PTO, PDS and BESS were analyzed. The inputs for analysis were: load profiles, engine efficiency and fuel consumption, PDS components efficiency, engines and PDS investment cost, fuel and maintenance cost, hybrid vessel configurations. Operation of TUG PDS in different operation modes was investigated. These are full electric modes: sail out, stand-by and light assist, and hybrid modes: peak shaving, load leveling and boost (light/medium/heavy assist). Next, suitable battery types were analyzed: high energy type NMC - lithium nickel manganese cobalt oxide (Graphite/NMC) and high power LTO - lithium titanate (LTO/NMC). The main requirement for hybrid vessel PDS configuration was equal bollard pull capability. Six cases were formed, with engines from 0.8 MW to 2 MW, and PTI/PTO from 250 kw to 1.5 MW, and BESS from 100 kwh to 600 kwh. Vessel CAPEX was calculated from PDS configuration, and OPEX was computed by time-domain computation based on actual vessel profile. Hybrid LVDC PDS was found to be techno-economically feasible with two of six configurations with equal bollard pull capability examined. In first case, 8% CAPEX reduction for vessel powerplant and power distribution, and 54% OPEX reduction were estimated. Here, the results showed 40% less fuel consumption and 77% less running hours and therefore remarkable reduction of pollution. In second feasible case, 14% CAPEX reduction and 37% OPEX reduction are estimated, with 25% less fuel consumption and 59% less running hours. In both studied cases fuel saving were minor factor. Fuel savings during BESS lifetime were slightly higher that BESS re-stock costs and justify the investment. Maintenance savings are major system benefit. Depending on case, kwh capacity was found sufficient with requirement of 5C capability of energy storage cells. Due to vessel low annual amount of running hours, battery should be used heavily to use cycling potential in calendar life, battery life time estimate was a.

14 14 3. WP 2 HYBRID VESSEL HARBOUR SUPPORT SYSTEM MODELLING AND ANALYSIS Electrification by renewable energy based Distributed Generation (DG) and Battery Energy Storage (BES) on board, as well as onshore, is inevitably the best solution to increase the use of pollution-free energy, especially at harbour areas. The process of shutting down diesel engines of vessels, and getting power supply from shoreside for ships auxiliary services during a stay at harbours is historically known as cold ironing or onshore power supply or shore to ship power. In the future, the power supply is also needed for charging the BESS on board, thus increasing the amount of power needed to electrify harbour area. In this regard, the aim of this work package (WP) is to develop the design of harbour area grid in such a way that it can supply shore to ship power supply for the vessels during a stay at harbours as well as facilitate the charging of the batteries. In this WP, a comprehensive state of the art survey about the technology development in shipping, future marine solutions, and shore to ship power supply was made. As a result of this part a paper was produced, where the review section explains the existing practice, current standards, barriers and technical challenges in implementing shore ship power supply. The paper also highlights on adopted voltage levels and frequencies of the power supply of various types of ships on board and their associated voltage, power and cable requirements from shoreside Harbour Area Smart Grid As a main result of this WP, an innovative design concept of the Harbour Area Smart Grid (HASG) was developed that meets the future technical requirements. In practice, several alternate designs of harbour grids are possible, but two appropriate design models having features of slow and fast charging of batteries have been developed. The key features of the distinctive design models proposed for the HASG are to support the vessels for cold ironing, provide a facility for charging the batteries and employing the renewables and battery storage at harbours. Each proposed design model can be suitable for a particular case depending upon technical requirements of the ships, the space, and infrastructure available on ports. The proposed charging scenarios can open new business models for ship owners and port administrators.

15 15 The designed concept also utilizes local renewable based DG. In this way, the new HASG concept will not only provide a pollution-free environment at the harbour area but also bring many benefits to the owners of hybrid vessels and seaports. The reliability and efficiency of the power supply will also be increased for harbour area consumers, terminal operators, port administration, and power utilities by employing power generation by the DGs at the harbour area. An example of the single line diagram of the HASG is shown in the following Figure. The HASG supports the harbour area load, cold ironing load and facilitates the hybrid and electric vessels for charging the batteries in the harbour area. In this design, the Harbour Area Bus of 20 kv is the bus for common coupling, which is supplied by the main grid, the HASG, and the wind power supply near to the harbour area. The HASG consists of photovoltaic as a DG and the BESs at various locations for the sake of reliability and stability of the whole power system. The hybrid vessel bus and the frequency converter stations are located nearby the seaports in the harbour area. The frequency converter is used to convert 20 kv, 50 Hz to 20 kv 60 Hz for supplying power to the vessels operating at 60 Hz during cold ironing. Two shoreside transformers are connected with hybrid bus and the frequency converter station to step down the voltage from 20 kv to 6.6 kv at 50 Hz and 60 Hz respectively. A shoreside transformer is required according to the HVSC standard to isolate galvanically one ship from another ship or consumer. However, this may not be mandatory in case if a dedicated high voltage shore supply transformer is available onboard. Thus, two dedicated ship buses or double bus bar are required to provide power to the ships at 50 Hz and 60 Hz. The other promising feature of the HASG is to support the future hybrid vessels at the harbour with the charging of the batteries for the electric and hybrid vessels. The batteries can be charged in a container (movable) during off-peak time and replaced with discharged batteries of the hybrid vessels during the process of cold ironing at harbour area. There are several other possibilities of charging the batteries other than the container-based solutions, which are presented with detail in a separate report.

16 16 Main Grid Wind Turbine Battery Energy Storage 110 kv Bus 0.69 kv Bus 115/21 kv 20/0.69 kv Harbour Area Smart Grid Harbour Area Bus 20 kv, 50 Hz 20/0.69 kv 20/0.69 kv Harbour Area Load Frequency Converter 50/60 Hz 0.69/20 kv Battery Energy Storage Photovoltaic 0.69 kv Bus Battery Energy Storage Hybrid Vessel Bus, 20 kv, 50 Hz 20/6.6 kv 20/6.6 kv 20/0.69 kv 6.6 kv, 50 Hz 6.6 kv, 60 Hz 0.69 kv Bus for charging batteries Ship Bus, 6.6 kv, 50 or 60 Hz Cold Ironing Normal charging of batteries AC/DC Converter Battery Energy Storage AC/DC Converter Battery Energy Storage AC/DC Converter Battery Energy Storage Batteries in Container Figure 1. Harbour Area Smart Grid The designed HASG is modelled in PSCAD/EMTDC and performance of harbour grid including slow and fast charger models was validated by simulating different case studies in PSCAD/EMTDC. The response of the HASG was observed both under steady state and transient state of operation. The steady state analysis of the HASG showed that the voltage and frequency at all buses are maintained within limits. Especially onshore bus voltage is within the limit of 3.5% of the voltage drop of the nominal voltage as specified in the current standard document High Voltage Shore Connection (HVSC). For transient analysis, two case studies are considered, one case of disturbance from the grid side and another case of dynamically varying load from the ship side. The results showed that all bus voltages remain within limits of ±10% of their nominal values, while onshore bus has voltage within specified limits as mentioned in the HVSC standards. Furthermore, frequency of onshore power supply remained constant, and the total harmonic distortion (THD) of charger current and voltage were within the limit of 5%.

17 Protection studies Normally, the harbor area supply system will be operated in grid-connected mode for a stable operation. However, during the faults on the main grid, only the DGs and BESS will be available to supply power, in this situation the HASG is said to be operating in islanded mode. The overall structures essentially make HASG as AC Microgrids with DGs, energy storage and loads which can be operated in grid-connected or islanded mode. Therefore, the standards related to DGs and AC Microgrids are very much relevant to the proposed alternatives in addition to the standards applicable for shore-to-ship connection of marine installations. The main objective of the protection studies made in this WP was to select and analyze the protection schemes for the various faults scenarios for one of the proposed HASG from here onwards called as harbor area AC Microgrid. It is known from the previous research that the traditional protection schemes based on OC relays and fuses with single-settings will work satisfactorily in grid-connected mode due to sufficient short-circuit current from the main grid. However, in islanded mode, more sensitive protection strategies will be required due to limited available short-circuit current from DGs. Therefore, the main focus of the studies made was on the islanded-mode fault scenarios. Nevertheless, the selected cases for grid-connected mode were also considered. In the studies made the following questions have been addressed: What types of faults may occur in the developed harbor area AC Microgrid during different operational modes and what is the magnitude of short-circuit current in those situations? What type of protection strategies will be adequate for the considered fault types? How effective are the considered protection strategies during grid-connected and islanded mode of operation? The studied HASG (see Figure 2) is a seven-bus supply system consisting of two main-grid buses (110 kv and 20 kv), two harbor area buses (20 kv and 0.69 kv), one charger bus (0.69 kv), one port bus (0.69 kv) and one ship bus (6.6 kv). The three units of container-based vessel-bess each of 2 MW rating are connected with the 0.69 kv charger bus. These container-based vessel-batteries are charged overnight with 0.1C slow charge rate (10 hours for

18 18 full charge). The supply to the cold-ironing load of 2 MW, 6.6 kv with two different frequencies of 50Hz and 60 Hz is also provided via the 20 kv harbor bus. Figure 2. HASG used in protection studies. The detailed results of the simulation studies made are presented in a separate report. According to the simulation studies made for both grid-connected and islanded modes of operation in different DG-scenarios it is clear that the traditional OC protection relays and fuses with single setting provide complete protection to HASG in grid-connected mode only and their response is either slow or ineffective for islanded-mode of operation. The adaptive protection settings lower than the grid-connected mode settings are necessary for islanded mode of operation even if DGs in islanded-mode provide maximum fault contribution up to two-times the rated current.

19 19 The proposed adaptive OC protection can be implemented by using fast communication links between relays according to IEC communication standard. The scheme can be implemented equally well for both 50 Hz and 60 Hz vessel types. However for 60 Hz vessel type due the presence the 50 Hz/60 Hz frequency converter at 0.69 kv harbor bus the relays at cold ironing connection require the fixed lower settings for both grid-connected and islanded mode because of limited fault rating of frequency converter which is only twice the rated full load current. The full-range fuses provide the fastest possible fault protection for the chargers and the frequency converters for input terminal faults in grid-connected mode. The fuses and gridconnected OC relay settings are not effective for islanded mode even if converter-based DGs provide maximum fault current contribution of twice the rated current. The fault analysis results were produced only for 50 Hz vessel supply in details. For 60 Hz vessel type the modelling of the fault current limiter control of the frequency converter needs still some further development to be done in future projects. As a final conclusion it can be said that the adaptive protection based on IEC communication standard will be required for the proposed harbor area AC Microgrid to enable its operation in both grid-connected and islanded mode. The use of adaptive protection and communication links will help detect, locate and clear the fault in the minimum required time. Nevertheless some delays due to mathematical conversion (instantaneous to rms value) and communication are unavoidable; therefore some delay tolerances/safety margins will have to be included for the correct protection coordination. The current standards for low voltage ride through (LVRT) also called fault ride through (FRT) allow the fault contribution from DGs for only initial 150 ms or so after the fault depending on individual country standards. After the 150 ms fault ride through time is passed the DGs are no more bound to be connected to the network. But for the proposed harbor area AC Microgrid operation in islanded mode the DGs should provide fault current contribution for at least 2 s after the fault in order to ensure complete protection coordination between relays. To meet this requirement all of the DGs should be equipped with full-scale converters and the LVRT standards of DGs need to be revised for islanded mode operation.

20 20 4. WP 3 HYBRID VESSEL ELECTRIC SYSTEM MODELLING AND ANALYSIS The 3 rd work package (WP3) concentrates on the modeling of the electrical grid of a marine vessel and control of the power balance. Special interest is put on adding energy storage (battery) onboard. Simulation blocks were created from the components of the vessel power grid. Such blocks were generator set, propulsion, hotel load, cables, and battery energy storage system (BESS). These model blocks are described in following subchapters. Cruise ferry and offshore supply vessel (OSV) were modelled in Simulink using these blocks. Models were simulated using different control methods for generator sets and BESS. Droop control and isochronous control was used for gensets. Isochronous control and power control based on average load power estimation was used for BESS. Actual recorded propulsion power data from the case vessel was used as an input to the simulation. Figure 3 describes the top level of the simulation model of the cruise ferry. Figure 3. Top level of the simulation model of the cruise ferry under study. It consists of three diesel generator (DG) models (blue blocks), two propulsion loads (orange blocks), hotel load and battery energy storage (BESS, the green block) and AC grid cables, which are Pi-section lines.

21 Diesel generator set The diesel genset model consists of the diesel engine model and the generator model. Figure 4 describes the top level of diesel genset model. The generator model in Figure 4 is a standard Simscape library block parametrized according to the given initial data. The diesel engine model consists of an engine model and an actuator model. Step response of the diesel engine model (actuator and engine) is presented in Figure 5. A step is fed to the input (fuel request) and the output is (torque) is presented. Figure 4. Diesel genset model consists of a diesel engine model and a generator model Step response Input Output Time [s] Figure 5. Step response of the diesel engine model (actuator included).

22 Battery Energy Storage System The top level of the BESS model is presented in Figure 6. The model contains a battery model and a grid converter model. The battery model is a standard Simscape battery model, which can be parametrized to fit each simulation case from a mask. The grid converter is controlled by a power reference signal, which is coming from an external power reference block. The power reference is fed to the current controller of the grid converter as current reference. The grid converter keeps track of the grid angle by a phase-locked loop. The converter model is simplified to lighten the calculation and does not contain switches. They are simply replaced with current sources. Figure 6. Top level of the battery energy storage system model Propulsion Drive At first, it s worth mentioning that this model does not contain model of the actual drive. This simplified model just draws the specified power from the grid. The simplified propulsion drive model consists of a power transformer, two diode bridges, DC-link model and a current source. The propulsion power data can be loaded from a file. As the grid converter, this model also does not contain switches. The propulsion drive is modelled by a controlled current source which sinks power from the DC-link. The propulsion data is recorded from the actual vessel. The same data is used in all simulation cases, for comparison. Additionally, there exists a hotel load which is all the other loads in the ship (pumps, fans, lighting, etc).

23 23 Figure 7. Top level of propulsion model Cruise ferry The model of the cruise ferry is simulated using droop control, isochronous control and running the diesel engines only on optimal economy point. Droop control is used in the actual case vessel to control diesel generators. Hence, it is also simulated here to establish a baseline. Isochronous control is used in simulations with battery energy storage, because it allows more flexible load sharing among generators compared to droop control. Two isochronous control cases are simulated with two different power reference signals to the BESS converter. 1 st Case: Power reference to the BESS converter is difference between current load power and non-causal average load power, see equation (4.1), = (4.1) where is calculated by non-causal moving average filter (4.2) from the total power requirement. () = 1 [( 99) + + ( + 100)] (4.2)

24 24 Of course, it is not possible to use future values of power consumption in an actual vessel. However cruise ferry is operating on the same route every day. In this case it could be possible to use power demand estimator to calculate power demand ahead based on the vessel s position, speed limits, weather conditions, captain s decisions, etc. 2 nd Case: BESS converter power reference is calculated as a difference between the current load power and the load power from the last 400 s, see equation (4.3)., = (4.3) This means that BESS is used to level sudden load fluctuations, but if the load power stays on that level the BESS power reference approaches zero. Diesel gensets then produce the power required by the load. Third investigated case is running diesel gensets in their optimal fuel economy point. BESS makes it possible to run gensets only on their optimal point as the battery energy storage maintains grid frequency and power balance. This case is a modification of the isochronous case. Genset control is similar to the isochronous case, only constant is given to gensets. Simulation results indicate that BESS can be used to balance marine vessel PDS. Time constant of a battery is much faster than time constant of a diesel engine. This helps when trying to keep the grid frequency constant as the load changes rapidly. The difference between the two isochronous cases is the power reference to the BESS converter. In the 1 st case it is calculated from the difference between the current load power and the load power estimate, which takes future values into account. In the 2 nd case the BESS converter power reference is calculated from the difference between current load power and averaged load power from the last 400 s. The 1 st case causes the gensets to increase power by charging the BESS just before load power demand increases. When the load power demand actually rises, BESS changes from charging to discharging. The time constant of the BESS is fast compared to the time constant of the gensets. This keeps the network frequency more stable. In the 2 nd case, when a load spike occurs, the BESS reacts to this, but if the spike is high enough and the nominal power of the BESS is not enough to cover this power demand, frequency starts

25 25 to drop and due to the long time constant of gensets it takes to time to respond to this power demand. The 1 st isochronous case makes better use of the short time constant of the BESS. BESS can be used to increase fuel efficiency by allowing the gensets to run on high fuel efficiency region. In this case BESS is handling the grid frequency control by sinking or sourcing power. This control method is prone to BESS converter fault. Fault situation was simulated where BESS converter fails, and gensets go to normal isochronous operation. The dynamic behavior of this size cruise ferry is limited. The load in the case of this vessel is quite static. Also, the massive rotational inertia in the gensets helps as an energy storage OSV Model of the OSV is simulated running gensets in isochronous control mode. Input data to the model is dynamic positioning power data recorded from the generators of the actual case vessel. Simulations are done using one and two gensets without BESS and using different isochronous control parameters. In the second case BESS is added, and the model is simulated while running one genset and BESS. BESS is controlled by running it in isochronous control mode, as it would be a diesel genset. Other used method is average power estimate. The average consumed power over a period of time is estimated and the difference between current consumed power and this estimate is fed to the BESS as a power reference signal. Simulation results of these cases are presented below in the Figures 8 and 9. Simulation results suggest that OSV can shut down one genset while in dynamic positioning (DP) mode and use BESS to compensate load fluctuation. Load power in dynamic positioning mode is not high, but it contains fast changes. In our data, load was between 400 kw and 700 kw, with some spikes up to 1100 kw. Although only genset could produce the needed power, it cannot respond to the rapid load changes causing frequency fluctuation in the ACgrid. BESS can be used to level the load fluctuation while the diesel genset produces the required base load. This also improves fuel efficiency in the DP-mode because the genset can operate on a higher load (and higher efficiency) area.

26 26 1 st Case: OSV is using 1 and 2 generators, no BESS is added to the system (Figure 8). Figure 8. Generator speed (network frequency) on the left and generator torque on the right while running the OSV model without BESS using 1 and 2 generators. 2 nd Case: OSV is using one genset and BESS is used to level load fluctuation (Figure 9). Speed [p.u.] Power [p.u.] Speed [p.u.] Torque [p.u.] Figure 9. Generator speed (network frequency) on the left and generator and BESS power on the right while running the model using only one genset and BESS.

27 27 5. WP 4 REMOTE VESSEL DATA MANAGEMENT In WP4 the target was to study the IOT applications and communication protocols in vessels. As a separate task also the communication based link between the laboratories between LUT and Vaasa was supposed to be investigated. At the beginning the system platform provided by CLS Engineering was tested at the University of Vaasa. The core of the system is COSMOSX10 unit which is capable of collecting data and handling the communication to cloud where the data is then stored. The COS- MOSX10 is based on Beaglebone Black with Linux operating system. In the tested system the connection to cloud was established with 4G modem and measurement of analog and digital signals was successfully demonstrated. The platform supports wide variety of protocols but at the time of the tests there was not yet IEC support available so data from power grid automation system cannot be collected. As an interesting option the possibility to extend the processing power of the data collection platform by using a separate FPGA board for processing the measured data was developed at conceptual level with only some tests of the available interfaces. In this kind of system certain features can be extracted from data and then only limited amount of data is necessary to transfer over the communication link. The FPGA is claimed to have better energy efficiency when considering tasks requiring more computing power. Relating to this aspect a detailed study was also made where the power consumption of the FPGA circuit was analysed. The focus was on analysis methods and the main outcome wss that the software based power analyser is still rather inaccurate. Considering the IoT platform it was realized during the project that there is no specific needs for further development of the existing platform. On the other hand, processing the data for some specific purpose in the context of the project would provide some novelty. Therefore the potential applications were tried to find in the discussions with project partners. As a conclusion the optimization of the engine load turned out to be the most interesting application, but due to the lack of time and resources only brief background survey was made. The basic idea is that by controlling the battery output in a suitable way the peaks of the engine load can be reduced. Possibly some optimization method can be used which utilizes

28 28 e.g. weather data. There exist also various advanced methods for fuel consumption estimation that can also be utilized. Developing this kind of solution would be suitable topic for some next projects. Due to the delays in commissioning of the engine laboratory at the University of Vaasa the remote connection between the laboratories was not possible to accomplish during this project. Instead of that a brief analysis of potential ways to make the connection was made. At the top level the remote connection to the automation system is possible. On the other hand, direct transfer of some specific measurement quantity could be more practical depending on the operating scenario. Some further research will needed to plan the actual implementation.

29 29 6. WP 5 DEMONSTRATION SYSTEM DEVELOPMENT AND TESTING In the 5 rd work package (WP5), a hybrid vessel power train emulator was constructed. The purpose of the laboratory setup is to give answers to two different research question classes. The first class deals with the vessel hybrid power train development. The second class is more general. It deals with questions of hardware-in-loop (HIL) simulations where loads and prime movers are replaced by electric motor drives or by power electronics alone. In addition, the laboratory setup aims to form a concept to be used in conjunction with Vaasa Energy Labs later on. In the research of the power train development, the purpose of the laboratory setup is A) to verify simulation models, B) to test energy balancing and control of vessel hybrid power train when battery energy storage is used, and C) to test different normal operation modes and fault cases. In the research of the HIL simulation, the research concentrates in finding the benefits, limitations and dynamic performance of hardware-in-loop simulation. In normal simulations, made without HIL, there exists no limitations of presenting dynamics of mechanical loads. In mechanical level HIL setup, the dynamics of load emulator and the drive under study (DUT), in this case e.g. propulsion drive, form a dynamic system that inevitably differs from actual drive by nature. The dynamics of HIL setup is affected both by the DUT drive and by the load emulator drive. In order to achieve similar dynamics between HIL and actual drive at control bandwidth, the performance and stability issues have to be carefully considered. Further, it is required that different controller tuning of load emulator drive has to be used for different DUT controller parameters. The research is used to give guidelines for dimensioning and for controller tuning of such HIL setups. The laboratory level demonstration systems offers tools to verify the simulation results and to understand model uncertainties and technical challenges, which a real marine hybrid system will include. From scientific point of view, laboratory analysis of simulation results are essential to be able to produce high level scientific articles about the studies generated in the earlier work packages. The laboratory test will include following subtasks:

30 30 Laboratory analysis and testing of different operational modes Comparison of laboratory tests and simulation results, model improvements Evaluation of dynamic performance and accuracy of emulator 6.1. AC network emulator hardware and communication structure The emulator consists of diesel emulator (800 kw induction motor) connected to synchronous generator (800 kva), a LLC-type (Wärtsilä Low Loss Concept) AC-network, battery energy storage system, propulsion emulator, and multi-purpose emulator (4Q inverter) that is used as the second genset and propulsion emulator. Rotating propulsion emulator compromises of 110 kw induction motor and propulsion load emulator (315 kw induction generator). Main components are presented in schema of Fig. 10 and in Fig. 11. ACS 850 AM 200 kw *C1 *C2 GENSET EMULATOR PROPULSION EMULATOR AM 315 RPM ACS 880 AM 800 kw AM 110 RPM DCS 880 SG 800 kva PLC 12p 400 V n. 200 kva BATTERY ENERGY STORAGE 135 kwh, 180 kva 4Q (AFE/GRID) PLC Mains connection *C3 *C4 Mains connection Figure 10. A laboratory setup for testing a power train of a future hybrid vessel. The main components in addition to network itself are a genset emulator unit *C1, a propulsion emulator *C2, a multi-purpose emulator *C3 and a battery energy storage.

31 31 Figure 11. Devices, software and communication media of the AC-network hybrid power drive emulator. The main controller of the emulator setup is an industrial PC. The propulsion emulator and the battery energy storage have their own programmable controllers so that they can execute, if so set, their own emulator algorithms in a distributed manner but all the emulations can be calculated in main controller also. The distributed emulations are required only if very fast dynamics is emulated. At the moment, all emulations are executed in main controller, which limits the control cycle of load emulator to 1-2 milliseconds and of battery energy storage control to 100 milliseconds. The emulations can be implemented by programs made using C- or IEC languages. In addition, Simulink models can be used when compiled using PLC coder from Mathworks inc. or using Bechoff TwinCat 3 Matlab/Simulink interface. The latter option enables one to visualize simulations in real-time. The emulation results can be graphically displayed in real time and recorded with 1 millisecond time resolution.

32 DC network emulator hardware and communication structure The DC network emulator is smaller scale emulator than AC network emulator, but more complex. The DC network emulator consists of two diesel genset emulators. In the first emulator a 55-kW induction motor is used as diesel emulator. This motor is connected to synchronous generator (110 kva) feeding the DC network via diode rectifier. Another, multi-purpose emulator can be used as a diesel genset and propulsion emulator. It consists of two back-to-back connected 5.5-kW induction machines. One machine is connected to mains via 4Q-inverter setup so that it can be used as a load (generator) as well as a motor. Similar 4-kW setup is as a main propeller load emulator. The setup does not include actual battery energy storage but, instead, uses a grid inverter to emulate battery. Battery characteristics can be programmed as emulator code in the main controller or battery energy storage of AC-setup can be used in parallel with the emulator when battery energy storage is connected with mains instead of vessel network. In this case, emulator code calculates scaling between voltages and currents of actual BESS and emulated one. This way the battery emulator can emulate both controlled BESS or BESS that is directly connected to DC vessel network. It is worth mentioning that emulation is not restricted to battery storage but other storages, such as super capacitor storages can be emulated as well. The DC- and AC-network emulators share same main controller, DAQ and supervisory control (See Fig. 12). Figure 12. Hardware and software components of the DC-network hybrid power drive emulator.