Dispatchability, Energy Security, and Reduced Capital Cost in Tidal-Wind and Tidal-Solar Energy Farms

: The global tidal energy resource for electricity generation is small, and converting tidal kinetic energy to electricity is expensive compared to solar-photovoltaic or land-based wind turbine generators. However, as the renewable energy content in electricity supplies grows, the need to stabilise these supplies increases. This paper describes tidal energy’s potential to reduce intermittency and variability in electricity supplied from solar and wind power farms while lowering the capital expenditure needed to improve dispatchability. The paper provides a model and hypothetical case studies to demonstrate how sharing energy storage between tidal stream power generators and wind or solar power generators can mitigate the level, frequency, and duration of power loss from wind or solar PV farms. The improvements in dispatchability use tidal energy’s innate regularity and take account of tidal asymmetry and extended duration low-velocity neap tides. The case studies are based on a national assessment of Australian tidal energy resources carried out from 2018 to 2021.


Introduction
The need for electricity grid stability drives the cost of dispatchable power as electricity supplies increasingly rely on renewable energy (RE). For example, in Australia, the Retailer Reliability Obligation now requires National Electricity Market retailers to cover a share of peak demand and the Australian Energy Market Commission is changing energy market regulations to ensure electricity grid security and facilitate RE generation [1,2].
At present, fossil-fuel-powered generators that run continuously, coupled with rapid control of pumped hydro and natural gas power generation, can manage intermittent or variable renewable energy [3], or rapid changes in demand [4]. However, as RE displaces fossil fuel, it will be necessary to increase demand management and energy storage to help stabilise electricity grids. Batteries, pumped hydro, hydrogen, biomass, and thermal energy storage are likely candidates to help provide dispatchability, particularly in remote area power supplies. Tidal stream power is rarely considered in this context, but its regularity should make it a competitive alternative [5]. However, its capital cost is higher than many other commercially available RE technologies. Moreover, the tidal resource may be insufficient to meet the capacity needs of a remote power station that is not connected to a

Status and Projections for Tidal Power
The potential of tides to generate power worldwide is about 3 TW, of which 1 TW is dissipated in shallow waters suitable for the installation of TECs [7]. However, the most productive and accessible sites limit the extractable power to about 120 GW, distributed across the Americas, Europe, Russia, West Africa, Southeast Asia, East China, and Western Australia [8]. The International Energy Agency has estimated that in 2019 the world used 27,764 TWh of electricity. Extractable tidal energy represents about 5% of this global electricity consumption [9].
Tidal stream energy converters ranging from tens of kilowatts to megawatts are installed worldwide. China, France, Japan, the Netherlands, Korea, and the United Kingdom have megawatt-scale TECs in operation [10]. The last fifteen years have seen increased research and development into tidal stream power systems. The European Marine Energy Centre and the Fundy Ocean Research Center for Energy have provided facilities for collaborative testing in many of these projects. Notable examples for commercialising and upscaling tidal stream power include the Simec-Atlantis Meygen and Orbital Marine Power-Scotrenewables-SR2000 projects [11].
Between 2018 and 2021, the AUSTEn project [6] measured and analysed Australia's tidal stream energy resources. It estimated Australia's economically extractable tidal stream resource as 1.5 TWh/yr. While small, this is not insignificant if one considers tidal power's regularity and modest energy storage requirements. This project determined that a critical role for tidal energy in remote coastal locations could be to provide cost-efficient, secure electricity for defence, emergency services, and industries such as mining or chemical processing. Table 1 provides CAPEX values for RE generators in Australia. These valuations were based on projections for 2025 and included: concept and definition costs, design and development costs, manufacturing costs and installation costs [6,12,13]. Table 1 indicates that TECs and offshore wind electricity generators have similar costs, and that solar PV and land-based wind turbines require significantly lower capital and operational expenditure. The Offshore Renewable Energy (ORE) Catapult initiative [11] estimates that CAPEX and the cost of capital account for about 77% of the levelised cost of electricity (LCOE) Energies 2021, 14, 8504 3 of 28 for early-stage tidal stream power projects. Ongoing research aims to reduce the CAPEX and LCOE by improving the efficiency of TECs, for example, by optimising subsea hub connection layouts, incorporating energy storage, developing direct drive power takeoff generators, and improving rotor blade pitch control [11]. With further development and economies of scale, ORE Catapult anticipates an LCOE in the United Kingdom of £150/MWh for 100 MW cumulative deployment of tidal stream generation, reducing to £80/MWh for 1 GW cumulative deployment [14]. Such projections evolve continuously and are reassessed annually in Australia [12].

Tidal, Wind and Solar PV Generator CAPEX
The CAPEX required for dispatchable power is a significant issue for remote area power systems. The Australian national electricity grid spans 850,000 km, and consequently, power averaging from uncorrelated RE resources increases grid electricity dispatchability. Off-grid solar PV and wind farms do not have this capability, so the intermittency of the local RE resource can range from hours to days of reduced power [5,15]. Table 1 does not include the cost of energy storage needed to ensure dispatchability in a remote area power station supplying 100 per cent RE. This cost could double the CAPEX required. Combining tidal stream energy with solar PV or wind farm energy may be a more economical method for providing dispatchable electricity.

Tidal Arrays and Farms
Countries are starting to move from assessing single tidal turbine test sites to assessing arrays that may be integrated into a single structure or mounted on the seabed and interconnected. High power single structure examples include the Orbital Marine Power O2 with two 1 MW turbines supported by a floating vessel or the Tocardo semi-submersible FloMo, which integrates five T2 turbines for a 1.5 MW array. Simec-Atlantis has developed a 6 MW, seabed-fixed, grid-connected array of four turbines for their Meygen project in the Pentland Firth, and Nova Innovation has developed a seabed fixed, grid-connected, array of four 100 kW turbines for their Bluemull project.
The MeyGen array produced 37.5 GWh of renewable electricity in 2020. It [16] interconnects a Simec Atlantis Energy AR1500 TEC with three Andritz Hydro Hammerfest HS1500 TECs. Each TEC can operate to depths of 100 m and uses an 18 m diameter rotor, producing a rated 1.5 MW for a 3 m/s tidal flow rate. The Simec Atlantis Energy plans to upgrade the MeyGen project to 86 MW rated power. It will be one of three proposals that are eligible to participate in the UK £20 m/year contracts for difference (CfD) subsidy dedicated to tidal stream energy.
Th CfD scheme ensures wholesale price stability by offering a 15 year fixed rate for electricity produced by projects that have secured a leased plot, grid connection and consents. The eligible tidal stream projects for the next CfD round (AR4) include MeyGen, the Perpetuus Tidal Energy Centre and Morlais. These three tidal stream farms could potentially contribute 124 MW towards a projected 2031 cumulative installed capacity of 160 MW that has the potential to reduce the levelised cost of energy (LCoE) from about 240 £/MWh to less than 150 £/MWh [17].
Coles et al. [18] have modelled a tidal stream array in the Alderney Race consisting of 78 sub-arrays. Each has eight TEC's with a lateral spacing of five rotor diameters and a longitudinal spacing of 18 rotor diameters. Three observations from the Alderney Race study are particularly relevant to the AUSTEn study reported here. These include: (a) The impact of array spacing and tidal velocity on the diversion of flow (blockage) from the array. The study found that such blockages reduced flow speeds by up to 2.5 m s −1 , reducing the annual energy production by 61%. (b) The need to increase the capacity factor of one subarray from 11% to 18%; and (c) the caution that while increased rotor size might improve TEC efficiency, it would also increase gearbox and bearing load and therefore capital cost. These issues have in common the role of tidal flow rate, which is substantially lower for TEC farms in Clarence Strait and Banks Strait (Australia) than in the Alderney Race. The low tidal flow rates in the Australian sites would make it necessary to reduce the rated power to increase the capacity factor and lower the capital cost of the TECs [6,19]. However, these lower tidal flow rates could make tidal power farms less susceptible to blockage and able to use TECs with larger rotor diameters [20,21] that could facilitate the capacity factor increase and CAPEX reductions.
Tidal energy farms are at an early stage of commercial development. Significant reductions in capital cost may be expected as their size and power output grow and cumulative deployment increases. Goss et al. [22] review the indices and processes of cost reduction, grouping them as: learning rate, in which the industry develops more efficient production and operational procedures; economies of scale, as devices become larger; and economies of volume where the cost of distributing and maintaining devices reduces as their number increases. Learning rates and economies of scale for tidal energy converters are reasonably well understood, and we have used them in our 2025 projections for hypothetical tidal energy farm costs [6,12]. Economies of volume are relatively new as evidence accumulates from installed TEC arrays being developed and planned in the UK and Channel Islands [22]. Incorporating an economy of scale factor into the CAPEX assessment during wind farm design may create opportunities to choose between a few large scale tidal turbines and many smaller turbines for a similar capital cost per unit power coupled with access to increased opportunities for generating tidal power.

Intermittency and Variability in Tidal Farms
Novo et al. [23] have assessed a range of standalone TEC farm options at 17 sites in Japan to determine the optimal combination of TEC and spatial location needed to produce dispatchable power. They have determined that a "firm capacity of 2.61% of total installed capacity can be reached", with "capacity factors greater than 0.34". This would be equivalent to a 7.8% dispatchability evaluated as the percentage of rated average power whose supply can be guaranteed.
Lewis et al. [24] have also studied the use of phase diversity to provide dispatchability. They comment that three tidal power stations with phase differences of 120 degrees could provide constant power throughout a tidal cycle, given a suitable coastline. However, the availability of such sites is quite uncertain. The study included an analysis of fine timescale effects such as turbulent loading on TEC power curves. It suggested that energy generation estimates would have a less than 1% error and that standardised power curves could be used with low-resolution, typically half-hourly, hydrodynamic data for resource assessment.
Jurasz et al. [25] described the use of spatial and temporal complementarity in hybrid renewable power configurations to manage the impact of their intermittency and variability on dispatchability. The paper reviewed research on the design of solar, wind, hydroelectric, and diesel hybrids to improve RE availability and summarised the associated widespread use of statistical correlation techniques and failure statistics as well as two performance indices based on: (a) partial time complementarity and (b) load tracking that measured the ability of a simulated renewable power supply to manage load. The paper further suggests that new work is needed to include such resources as wind and tidal energy and to determine the differences between measured and model data. The AUSTEn study reported here partially addresses these needs by describing and applying a novel algorithm that measures complementarity as the minimum energy storage required to ensure a predetermined level of dispatchability from a hybrid RE supply (Equations (1)-(3)). The technique is used in a set of hypothetical case studies that demonstrate the impact of including tidal energy in wind or solar farms on RE farm dispatchability and CAPEX.

Materials and Methods
This study evaluated the relationship between CAPEX, energy storage and generated power dispatchability for hypothetical tidal farms at locations identified by the AUSTEn project [6,12,20,21]. Three farms included a wind turbine or solar PV component to provide generator capacity and a tidal stream component to provide dispatchable power. The paper will refer to these combinations as tidal-solar PV farms, or tidal-wind farms, or in general

Assessing Australian Sites Suitable for Developing Tidal Energy
Sites were selected (Figure 1), and their potential for extractable power was assessed using power curves derived from commercially available TECs. The assessments were repeated after the power curves were modified to match the tidal velocities available in Australia (Table 2).
farms. The case study evaluations comprised: • an assessment of Australian sites suitable for developing tidal stream energ taking account of environmental limits on extraction; • a TEC model matched to the available range of tidal stream velocities; • an assessment of the variability, intermittency, and meteorological constraints site; • an evaluation of the energy storage required for dispatchable power at each s • a model for determining the dependence of CAPEX for RE generation on tida and solar variability and intermittency.

Assessing Australian Sites Suitable for Developing Tidal Energy
Sites were selected (Figure 1), and their potential for extractable power was a using power curves derived from commercially available TECs. The assessmen repeated after the power curves were modified to match the tidal velocities avai Australia ( Table 2).
Three tidal model velocity datasets, all assessed against observations, were u the studies. The first was a low-resolution dataset taken from the AUSTEn Nation Energy Model to assess national tidal energy resources [19]. The second and thir high-resolution AUSTEn datasets from local-area models of two regions, Banks S Tasmania [20] and Clarence Strait in the Northern Territory [21].  Three tidal model velocity datasets, all assessed against observations, were used for the studies. The first was a low-resolution dataset taken from the AUSTEn National Tidal Energy Model to assess national tidal energy resources [19]. The second and third were high-resolution AUSTEn datasets from local-area models of two regions, Banks Strait in Tasmania [20] and Clarence Strait in the Northern Territory [21].
The assessment for tidal stream energy extraction was moderated to protect the marine environment and the tidal resource from the impacts of over-extraction. These impacts are complex and site-dependent, and a detailed assessment was beyond the scope of the AUSTEn project. A significant impact factor (SIF) of 20% was used for the hypothetical case studies described here as an approximate tidal stream energy extraction limit to prevent environmental and resource degradation [26]. The SIF did not include the TEC Betz limit, which further reduced the model extraction.

The Tidal Energy Converter Model
The parameters required for TEC selection included: power curves, rated power, rotor diameter, generator dimensions and operational draught. Specifications were obtained from a confidential survey of tidal energy companies, and these were used to develop generic power curves for rotor diameters ranging from 4 m to 26 m to match operating depths. The model assumed that all TECs would have yaw control to maintain their orientation towards the tidal current flow.
Commercially available TECs with rotors large enough to take advantage of Australia's relatively low-speed tidal currents have power ratings that are too high for them to operate near total capacity. They are built to withstand and exploit current speeds that do not occur in Australian waters, which makes them unnecessarily expensive for this continent. To address this issue the generic power curves (Figure 2a,b) were modified by limiting their rated operating speed to 2 m/s instead of the more typical value of 3.6 m/s for commercially available units. These power curves are for hypothetical reduced power rating TECS designed for the weaker available tidal stream energy resource [6]. The down-rated TECs were applied to regions including Banks Strait, The Tiwi Islands, Ardyaloon, Derby and Wadeye. The generic power curves (Figure 2a,b) reflect the TEC design adjustments needed to operate cost-effectively in Australian waters. The largest possible TEC was selected for any given depth because the CAPEX model indicated that their cost per unit power improved with rotor size up to about 26 m in diameter. A 5 m clearance was allowed below the sea surface for depths greater than 10 m to allow vessels of less than 2 m draught to pass and to protect the rotor from wave action. The TEC specifications were used to determine the clearance for the rotor above  The largest possible TEC was selected for any given depth because the CAPEX model indicated that their cost per unit power improved with rotor size up to about 26 m in diameter. A 5 m clearance was allowed below the sea surface for depths greater than 10 m to allow vessels of less than 2 m draught to pass and to protect the rotor from wave action. The TEC specifications were used to determine the clearance for the rotor above the seabed.
TECs to produce the power curves in Figure 2b are not in commercial production. Such turbines would need to have a power rating of about a third of the rated power shown in the generic power curves derived from commercially available TECs (Figure 2a). Savings for the modified TECs were based on limiting their rated operating speed to 2 m/s instead of the more typical value of 3.6 m/s for commercially available units. The gear ratios, rotor blade size, and power rating of distribution and power conversion equipment were then adjusted to match the modified TECs lower rated power and thus reduce the unit CAPEX.
The tidal flow rate (U) data comprised 1020 hourly data points representing 42.5 days, including two spring-neap tide cycles. The hourly energy delivered (E) was estimated from tidal velocity data using a linear interpolation between each velocity step in the power curves. The resulting time series was cropped to represent a lunar month and replicated to approximate a year of data.

Managing Variability and Intermittency Using Shared Battery Storage
In this paper, dispatchable power and energy are quantified as a power threshold (Pth) that is guaranteed, and a corresponding daily dispatched energy threshold (Eth). The daily dispatched energy is (Ed). Dispatchability (D) can be calculated as the ratio of the dispatchable power threshold (Pth) to the rated average power. Variability refers to predictable changes in generated power [3], and intermittency refers to unpredictable changes in generated power [3].
Variability and intermittency were managed using a combination of batteries and tidal stream power. The battery energy capacity required to manage the variation in power during diurnal and semidiurnal tidal and solar cycles was evaluated using spectral and temporal analyses. An energy storage sharing algorithm was used to determine the additional capacity required for managing intermittent drops in solar or wind power.
In Case Study 1, the RE intermittency was assessed using wind energy data from Swan Island in Banks Strait on the north coast of Tasmania [15]. In Case Study 2, RE intermittency was assessed using solar exposure data for Murrumujuk south of Clarence Strait and northeast of Darwin [27]. If a daily solar exposure data point was missing from a weather station record, it was replaced by the average of data points for that day from neighbouring weather stations within a 20 km radius. Tidal energy data were from the AUSTEn surveys of Banks Strait [20] and Clarence Strait [21]. Figure 3 shows examples of the spectra for electrical power generated from TEC farms in Banks Strait and Clarence Strait, while Figure 4 shows examples of intermittent and variable wind and solar power in Banks Strait [15,27].

The Tidal Energy Farm Diurnal and Semidiurnal Low Power Periods
A Fast Fourier Transform was applied to time series data for the average electrical power generated each hour by the TEC farms. The moduli of the Fourier coefficients provided a spectrum for the electrical power from which the low power TEC periods corresponding to diurnal and semidiurnal tidal cycles could be determined. Figure 3 shows the Fourier amplitudes vs frequency representing the TEC power spectra.
The spectra for Case Study 1 (Banks Strait) and Case Study 2 (Clarence Strait) show peaks at 1.95 and 3.87 cycles per day (cpd) corresponding to 12.3 h and 6.2 h periods for the rectified diurnal and semidiurnal tide cycles. The Clarence Strait spectrum also shows a 0.9 cpd line and its harmonic at 2.9 cpd produced by power rectification and the cubic transformation of velocity to power. The 3.1 h half-cycle period for the 3.87 cpd line in the Banks Strait and Clarence Strait spectra characterised the low power phase of the tide cycles and was used with Eth to calculate the energy storage capacity required to manage it. The spectra for Case Study 1 (Banks Strait) and Case Study 2 (Clarence Strait) show peaks at 1.95 and 3.87 cycles per day (cpd) corresponding to 12.3 h and 6.2 h periods for the rectified diurnal and semidiurnal tide cycles. The Clarence Strait spectrum also shows a 0.9 cpd line and its harmonic at 2.9 cpd produced by power rectification and the cubic transformation of velocity to power. The 3.1 h half-cycle period for the 3.87 cpd line in the Banks Strait and Clarence Strait spectra characterised the low power phase of the tide cycles and was used with Eth to calculate the energy storage capacity required to manage it.

Calculating the Energy Storage Required to Manage Intermittent Low Power Periods
A novel algorithm (Equations (1)- (3)) has been developed that applies phase shifting, linear superposition and a dispatchable power threshold to hybrid tidal-wind or tidalsolar power time-course data. This procedure develops critical energy deficit time windows [25] for each of which the net energy deficit is zero. The maximum energy deficit in this set is the required battery capacity for the system to provide dispatchable power.
The battery energy capacity (Ebattery) was calculated as the sum of Ev and Eint. The Ev component managed renewable energy resource variability, and the Eint component managed its intermittency. Steps 1 to 4 calculate these components using: 1. Equation (1) to evaluate Ev for the solar PV and tidal arrays; 2. Equation (2) to determine the year and day when Ed was at its minimum for the standalone wind or solar PV farm and the tidal hybrid farm; 3. Equation (3) to evaluate the Eint component for Ed to equal or exceed Eth; 4. The battery energy capacity was calculated either during step 2 or as Ev + Eint.
Equation (1) evaluated Ev by applying the dispatchability factor (D) to the rated average powers (Psolar) and (Ptidal) and the low power diurnal or semidiurnal periods (τsolar) and (τtidal).
Equation (2) evaluated the minimum value of Ebattery required to ensure that Ed equalled or exceeded Eth by dividing the generated renewable energy (Ere) into a daily

Calculating the Energy Storage Required to Manage Intermittent Low Power Periods
A novel algorithm (Equations (1)-(3)) has been developed that applies phase shifting, linear superposition and a dispatchable power threshold to hybrid tidal-wind or tidal-solar power time-course data. This procedure develops critical energy deficit time windows [25] for each of which the net energy deficit is zero. The maximum energy deficit in this set is the required battery capacity for the system to provide dispatchable power.
The battery energy capacity (Ebattery) was calculated as the sum of Ev and Eint. The Ev component managed renewable energy resource variability, and the Eint component managed its intermittency. Steps 1 to 4 calculate these components using: 1.
Equation (1) to evaluate Ev for the solar PV and tidal arrays; 2.
Equation (2) to determine the year and day when Ed was at its minimum for the standalone wind or solar PV farm and the tidal hybrid farm; 3.
Equation (3) to evaluate the Eint component for Ed to equal or exceed Eth; 4.
The battery energy capacity was calculated either during step 2 or as Ev + Eint.
Equation (1) evaluated Ev by applying the dispatchability factor (D) to the rated average powers (P solar ) and (P tidal ) and the low power diurnal or semidiurnal periods (τ solar ) and (τ tidal ).
Energies 2021, 14, 8504 9 of 28 Equation (2) evaluated the minimum value of Ebattery required to ensure that Ed equalled or exceeded Eth by dividing the generated renewable energy (Ere) into a daily stored energy component (Estore) and a daily dispatched energy component Ed. The evaluation used increments of one day over a period of one year.
Ere i = Esolar i or Ewind i + Etidal i generated renewable energy, avoiding battery overcharge, Estore i ≥ Ev, managing tidal and solar variability.
Equation (3) evaluated Eint from the Ere time series by checking the intermittent periods containing events in which Ere fell below Eth one or more times in each period. An energy deficit (Edeficit) accumulated during each period while Ere varied above and below Eth. Each intermittency period commenced when Edeficit fell below zero and finished when Edeficit returned to zero. The maximum absolute value of Edeficit in the Edeficit set was the battery energy capacity component Eint. where: For the tidal hybrid farm, Equations (2) and (3) included coincident tidal and solar or wind minima. This was ensured using Equation (2) to calculate a set of 30 time series for Ere in each of which the tidal energy was phase-shifted relative to the solar or wind energy in increments of one day. The required battery energy capacity was the maximum Ebattery value in the set. The year-long time series containing Edmin was used to evaluate Eint.
The battery energy capacity, Ebattery, can be evaluated rapidly from Equation (3) when assessing dispatchability for a range of Ebattery values. Using Equation (2) to determine Ebattery is slower; however, it can also be used to show the cumulative frequency for low energy days in the Ed time series.  Figure 5 shows that the minimum value for Ed is 21 MWh with no energy storage or tidal power. Figure 6 shows that the minimum Ed increases from 21 MWh to the Eth target value of 216 MWh when tidal power replaces 25% of the solar power generation and a 624 MWh energy storage battery is added. Figure 6 also shows the cyclic impacts of neap tides; Figure 7 is an average of the 30 tidal-solar power scenarios. It hides the neap tide cycles but accounts for the simultaneous occurrence of solar and tidal power minima. Figure 8 shows that the minimum value of Ed falls to 21 MWh if the tidal power component is replaced with solar power.

The Impact of Energy Storage and Tidal Power on Intermittent Low Power Periods
MWh when tidal power replaces 25% of the solar power generation and a 624 MWh energy storage battery is added. Figure 6 also shows the cyclic impacts of neap tides; Figure  7 is an average of the 30 tidal-solar power scenarios. It hides the neap tide cycles but accounts for the simultaneous occurrence of solar and tidal power minima. Figure 8 shows that the minimum value of Ed falls to 21 MWh if the tidal power component is replaced with solar power.        The principal factors that reduced intermittency in the electricity supplies associated with Figures 6 and 7 were the regularity of tidal power and sharing energy storage to manage phase coincidences between the solar and tidal power minima. Algorithms that The principal factors that reduced intermittency in the electricity supplies associated with Figures 6 and 7 were the regularity of tidal power and sharing energy storage to manage phase coincidences between the solar and tidal power minima. Algorithms that predicted the coincidence of tidal or solar minima did not further mitigate intermittent power drops below Pth.

Energy Storage and Cost
Vanadium batteries were projected to be more cost-effective than lithium, where more than six hours of energy storage were required [28,29]. Consequently, lithium batteries were only used where the tidal energy resource had short periods of low power during neap tides. Small scale pumped hydro was approximately the same cost per kilowatt-hour as lithium batteries in coastal Australia, where cliffs seldom rise above 200 m [29].
The battery CAPEX included two components: costs related to power rating and costs related to energy storage. Table 3 provides the CAPEX) values for the most cost-effective energy storage options projected for 2025 [12,28,29]. Equation (4) combined the battery cost components. $E battery was the capital-cost per kilowatt-hour, E battery was the battery energy storage capacity, $P battery was the capital-cost per kilowatt, and P battery was the battery power rating.
The cost estimate for vanadium redox batteries included construction, commissioning, and vanadium electrolyte costs in the energy storage component and power conversion, stack, and balance of plant costs in the power component [29,30]. The lithium battery cost estimate included construction, commissioning, and all stack costs in the energy storage component and power conversion and balance of plant costs in the power component. A factor of two is included in the lithium battery cost to account for the longer lifecycle of vanadium compared to lithium batteries [28].
In the case studies, the energy component of cost was five to ten times greater than the power component because of the extended periods of low power that needed to be managed. The vanadium constituent of vanadium redox batteries was the greatest contributor to this cost, and the price of vanadium is volatile [31]. This price uncertainty impacts the projected case study estimates of CAPEX reduction as they were based on a relatively low present-day price for the vanadium electrolyte. Increases in the price may increase the projected CAPEX savings.

Evaluating the CAPEX Reduction for Tidal Hybrid Farms
Equation (5) evaluated the difference in CAPEX for dispatchable power generated with or without tidal power. The CAPEX difference included power generation costs ($P) in units of ($/MW) and energy storage costs ($E) in units of ($/MWh). The power component depended on the dispatchable power threshold Pth that governed the battery power costs and the rated average tidal power (Ptidal) that governed the tidal and wind or solar PV generator costs. The energy component (E) included the battery energy storage costs. The CF variables are capacity factors. where: The "SolarWind" subscript identifies solar PV or wind turbine arrays. The "Tidal" subscript identifies TEC arrays. The "TidalHybrid" subscript identifies tidal-solar or tidal-wind hybrid arrays. The low power periods (τ) are evaluated from their respective battery energy.
Capacities and the dispatchable power threshold (Pth), using Equation (6): Figures 9-11 are derived from Equation (5). Each graph is evaluated for 1.0 MW average power, and the CAPEX saving scales linearly with average power. There is a change from lithium to vanadium batteries at a tidal low power duration equal to or greater than six hours. Figure 9 shows a critical boundary with CAPEX saving depending on increases in the dispatchable power threshold Pth and wind low power duration τwind.
Energies 2021, 14, x FOR PEER REVIEW where: The "SolarWind" subscript identifies solar PV or wind turbine arrays. The "Tidal" subscript identifies TEC arrays. The "TidalHybrid" subscript identifies tidal-solar or tidal-wind hybrid array The low power periods (τ) are evaluated from their respective battery energy capacities and the dispatchable power threshold (Pth), using Equation (6): = .
Figures 9-11 are derived from Equation (5). Each graph is evaluated for 1.0 M erage power, and the CAPEX saving scales linearly with average power. There is a from lithium to vanadium batteries at a tidal low power duration equal to or grea six hours. Figure 9 shows a critical boundary with CAPEX saving depending on in in the dispatchable power threshold Pth and wind low power duration τwind.  Figure 10 describes the CAPEX savings dependence on tidal hybrid farm low periods; for example, during neap tides or where there is tidal asymmetry. There periods; for example, during neap tides or where there is tidal asymmetry. There is ical boundary for Pth at 0.15 MW (15% dispatchability). The figure suggests that C saving should be possible for a broad range of tidal hybrid farm low power durat the standalone wind farm has extended periods of low power.  Figure 11 indicates the range of low wind farm power durations that can be m using tidal energy. It shows that CAPEX savings increase as hybrid farm low pow ration reduces and standalone wind farm low power duration increases.

Results
The case studies demonstrate that incorporating tidal energy in wind farms PV farm designs may improve energy security and reduce CAPEX. Preliminary v of these case studies were described in an ARENA-AUSTEn project report [6] w mates limited to tidal hybrid farms using a three-hour energy storage battery to m the tidal ebb and flood power. The case studies described here extend the range of storage, dispatchability and CAPEX reduction, introduce a battery sharing techniq account for neap tides and asymmetric tide cycles. They also include performan for case studies 1 and 2 using solar exposure and wind velocity data and the m described in Section 2.
Case Study 1 demonstrates a reduction in CAPEX from replacing the wind and high-capacity energy storage with a tidal hybrid farm and lower-capacity ener age. Case Study 2 demonstrates improved dispatchability and reduced CAPEX by  Figure 10 describes the CAPEX savings dependence on tidal hybrid farm low power periods; for example, during neap tides or where there is tidal asymmetry. There is a critical boundary for Pth at 0.15 MW (15% dispatchability). The figure suggests that CAPEX saving should be possible for a broad range of tidal hybrid farm low power durations if the standalone wind farm has extended periods of low power. Figure 11 indicates the range of low wind farm power durations that can be managed using tidal energy. It shows that CAPEX savings increase as hybrid farm low power duration reduces and standalone wind farm low power duration increases.

Results
The case studies demonstrate that incorporating tidal energy in wind farms or solar PV farm designs may improve energy security and reduce CAPEX. Preliminary versions of these case studies were described in an ARENA-AUSTEn project report [6] with estimates limited to tidal hybrid farms using a three-hour energy storage battery to manage the tidal ebb and flood power. The case studies described here extend the range of energy storage, dispatchability and CAPEX reduction, introduce a battery sharing technique, and account for neap tides and asymmetric tide cycles. They also include performance tests for case studies 1 and 2 using solar exposure and wind velocity data and the methods described in Section 2.
Case Study 1 demonstrates a reduction in CAPEX from replacing the wind energy and high-capacity energy storage with a tidal hybrid farm and lower-capacity energy storage. Case Study 2 demonstrates improved dispatchability and reduced CAPEX by sharing energy storage between tidal and solar PV energy generators, despite the impact of extended period neap tides. Case studies 3 and 4 describe potential applications where tidal energy by itself may provide dispatchable electricity.
The average wind farm power was set equal to the tidal farm power in Case Study 1, to demonstrate the separate and independent contributions of tidal and wind farms to dispatchability (Table 4). This contrasts with Case Study 2, where the solar and tidal energy resources share the battery capacity because the extended period neap tides mean that tidal power has no competitive advantage over solar power. Nevertheless, the temporal complementarity of the solar and tidal farm powers ensures that when they share a storage battery, they will provide more cost-effective dispatchability than either could provide separately. In both case studies, alternative configurations and optimisations are possible depending on a client's specification. Potential optimisations can be assessed using charts such as Figures 9-11 or by applying Equations (1)- (5).
Case studies one and two used the maximum economically available tidal power, and cost estimates were 2025 projected costs. Economies of scale and the need to compete cost-effectively with wind and solar power farms mandated the use of hypothetical 26 m turbines with a power cut out at tidal flow rates of 2 m/s [6]. Using a greater number of smaller TECs [22] might have achieved economies of volume (EoV); however, there was insufficient Australian data to estimate them. If CAPEX EoV's were achievable a factor of 0.9 would indicate an installed CAPEX of A$4065/kW for TECs with 20 m rotors, comparable with A$4076/kW for TECs with 26 m rotors. In case study 2, this would have increased the available tidal farm power to 12 MW.

Overview
The Tasmanian government plans to generate 200% of its electricity from renewable energy, including wind, solar PV and hydropower, by 2040 [32]. The Musselroe wind farm located next to Banks Strait is a significant resource contributing to this plan; it has 56 wind turbines with a total rated power of 168 MW, an average power of 57 MW, and from 2016 to 2020 produced on average 575 GWh/year [33][34][35]. However, it can experience intermittent drops in wind energy for durations of up to six days [4,15]. In contrast, tidal stream energy in Banks Strait may require as little as three hours of energy storage to provide dispatchability.
This case study describes the performance and cost impacts of introducing tidal energy into wind farms. The first example is an offshore wind farm; the second example is a land-based wind farm. The tidal-wind farms are hypothetical, drawing on wind energy data from Swan Island in Banks Strait [15], wind turbine data from the Musselroe wind farm [4,[33][34][35], and tidal stream data from the AUSTEn project ( Figure 12) [20]. The offshore and land-based farms have the same dispatchability and average power ratings but different capacity factors. The capacity factors are estimated values typical in this region [33] and reflect the average energy delivered but not the maximum period of low wind power. This region's maximum period of negligible wind power was six days per year for both the land-based and offshore wind farms [4,15].

The Neap Tide Influence on Battery Storage
The neap tide average power generation in Banks Strait had a minimum of approximately 50% of the rated average power. Consequently, the battery capacity to support up to 50% TEC array dispatchability was only needed for semidiurnal variability and was not required to support neap drops in tidal power. With wind power, the reverse is the case where variability does not need to be supported, but intermittency does. This suggests that the storage batteries required for wind and tidal power dispatchability may not need to be shared. Setting the hybrid farm dispatchability at 40% instead of 50% allowed the contributions of wind and tidal farms to dispatchability to be readily observed (Table  4).

System Specification, Performance, and Costs
The performance comparison in Figure 13 applies to both the offshore and landbased wind farm examples. It describes four configurations for the wind farm: with and without tidal energy support and with two energy storage levels. The required dispatchable energy threshold (Eth) is 154 MWh that is 40% dispatchability. Figure 13 shows that the wind farm's daily dispatched energy (Ed) requires 842 MWh energy storage to reach the Eth threshold value. Replacing 50% of the wind energy with tidal energy reduces the required battery energy capacity to 294 MWh, lowers the frequency of Ed values that are less than 400 MWh, and ensures that Ed exceeds Eth.

The Neap Tide Influence on Battery Storage
The neap tide average power generation in Banks Strait had a minimum of approximately 50% of the rated average power. Consequently, the battery capacity to support up to 50% TEC array dispatchability was only needed for semidiurnal variability and was not required to support neap drops in tidal power. With wind power, the reverse is the case where variability does not need to be supported, but intermittency does. This suggests that the storage batteries required for wind and tidal power dispatchability may not need to be shared. Setting the hybrid farm dispatchability at 40% instead of 50% allowed the contributions of wind and tidal farms to dispatchability to be readily observed (Table 4).

System Specification, Performance, and Costs
The performance comparison in Figure 13 applies to both the offshore and land-based wind farm examples. It describes four configurations for the wind farm: with and without tidal energy support and with two energy storage levels. The required dispatchable energy threshold (Eth) is 154 MWh that is 40% dispatchability.  Tables 4 and 5 compare the specifications and costs for an offshore wind farm using wind turbines with and without tidal energy. Both farm configurations produce an average power of 16 MW, of which 40% is dispatchable. The CAPEX reduction for the tidalwind farm is 20% (A$105m), principally due to a reduction in the battery energy capacity needed for dispatchability. The second example estimates the cost of a land-based wind farm with and without a tidal component. The wind farm uses wind turbines with capacity factors typical of landbased farms in the region. The average power and dispatchability for the land-based farm are identical to the offshore farm in the first example. However, the reduced capacity factor changes the rated peak power needed to achieve the required average power rating and alters the CAPEX saving associated with the tidal-wind farm. These changes are shown in Tables 6 and 7.  Figure 13 shows that the wind farm's daily dispatched energy (Ed) requires 842 MWh energy storage to reach the Eth threshold value. Replacing 50% of the wind energy with tidal energy reduces the required battery energy capacity to 294 MWh, lowers the frequency of Ed values that are less than 400 MWh, and ensures that Ed exceeds Eth. Tables 4 and 5 compare the specifications and costs for an offshore wind farm using wind turbines with and without tidal energy. Both farm configurations produce an average power of 16 MW, of which 40% is dispatchable. The CAPEX reduction for the tidal-wind farm is 20% (A$105m), principally due to a reduction in the battery energy capacity needed for dispatchability.

Tidal-Onshore Wind Specification and CAPEX
The second example estimates the cost of a land-based wind farm with and without a tidal component. The wind farm uses wind turbines with capacity factors typical of land-based farms in the region. The average power and dispatchability for the land-based farm are identical to the offshore farm in the first example. However, the reduced capacity factor changes the rated peak power needed to achieve the required average power rating and alters the CAPEX saving associated with the tidal-wind farm. These changes are shown in Tables 6 and 7. The CAPEX saving also depended on how the tidal component of the tidal-wind farm was connected to the end-user. If the connection was via a 33 km subsea power line to the national grid, the saving was 5% (A$20m). If it was connected to a local grid via a 10 km subsea power line, the saving increased to 11% (A$43m).

Overview
Case Study 1 illustrated how replacing high-capacity wind energy storage with tidal energy and lower-capacity energy storage can reduce the CAPEX required for dispatchable power. However, this may not be possible in regions where extended duration neap tides with low tidal velocity last more than a day. Under these circumstances, a battery sharing strategy may be helpful. Case study 2 illustrates how introducing tidal energy and battery sharing into a solar PV farm could mitigate solar power loss during the wet season while managing neap tides with negligible kinetic energy for one or two consecutive days per month. The application is for a chemical plant where a constant electricity supply can be critical, for example a plant for converting hydrogen and carbon dioxide to fuel.
For this case study the maximum available TEC power was used combined with enough solar power to provide a similar production capacity compared to projects that did not use tidal power. For example, the George Olah plant in Iceland, which produces 4000 tons of renewable methanol per year [36]; or the proposed Heroya project in Norway, designed to produce 8000 tonnes of e-fuel per year [36]. The dispatchability was set at 50 per cent, as a minimum requirement for cost-competitive power to efuel conversion [36].

Dispatchable Renewable Power for a Clarence Strait Power to Liquid Electro-Fuel Plant
In Darwin, a liquefaction plant at Wickham Point processes natural gas from the Ichthys gas field in the Timor sea. The plant could produce 278 Mt of carbon dioxide during its 40-year life, and about 30% of this would be concentrated carbon dioxide stripped from the natural gas by its liquefaction [37,38]. The proximity of this plant to tidal stream and solar energy resources suggests an opportunity to recycle some of this waste carbon dioxide as electro-fuel (e-fuel). The car manufacturer Audi pioneered carbon dioxide recycling in a 6 MW power to efuel conversion plant in Werlte, Germany. The Audi plant used electricity from four North Sea wind turbines, each rated at 3.6 MW peak power, and could produce up to 1000 t/year of synthetic methane from 2750 t/year of carbon dioxide [39]. The liquid fuel industry is also testing prototypes for producing e-fuels such as synthetic methanol, diesel or kerosene from hydrogen and waste carbon dioxide. The largest of the proposed commercial-scale projects aimed to produce 8000 t/year of liquid e-fuel in Heroya, Norway [36]. The intermittency of wind farm power may compel the plant operators to consider electricity price arbitrage strategies. For example, the Audi Werlte plant may use electricity purchased from the spot market on an hourly trade basis [40], reducing the cost of electricity and improving grid stability during periods of surplus renewable energy production. Including tidal energy in such a plant could further improve such strategies or lessen the need for them.
A power to liquid e-fuel plant near Glyde Point north of Darwin could receive concentrated waste carbon dioxide by pipeline from the natural gas liquefaction plant at Wickham Point. The plant could run on power from a solar PV farm at Glyde Point and TECs in Howard Channel 4 km offshore. Figure 14 shows the distribution of tidal energy along the Howard Channel. t/year of synthetic methane from 2750 t/year of carbon dioxide [39]. The liquid fuel industry is also testing prototypes for producing e-fuels such as synthetic methanol, diesel or kerosene from hydrogen and waste carbon dioxide. The largest of the proposed commercial-scale projects aimed to produce 8000 t/year of liquid e-fuel in Heroya, Norway [36]. The intermittency of wind farm power may compel the plant operators to consider electricity price arbitrage strategies. For example, the Audi Werlte plant may use electricity purchased from the spot market on an hourly trade basis [40], reducing the cost of electricity and improving grid stability during periods of surplus renewable energy production. Including tidal energy in such a plant could further improve such strategies or lessen the need for them. A power to liquid e-fuel plant near Glyde Point north of Darwin could receive concentrated waste carbon dioxide by pipeline from the natural gas liquefaction plant at Wickham Point. The plant could run on power from a solar PV farm at Glyde Point and TECs in Howard Channel 4 km offshore. Figure 14 shows the distribution of tidal energy along the Howard Channel.    Figure 15 shows the power supply time series for a tidal farm in Howard Channel comprising TECs with power curves that have been modified to reduce their rated power and the associated CAPEX. The solar PV panels are equipped with a single-axis tracking unit to provide about eight hours a day of solar power generation. The tidal and solar PV generators share a 624 MWh battery: to reduce solar power intermittency, provide electricity for continuous plant operation during the low power phase of tidal ebb and flow, and manage low tidal power during neap tides. The phase difference between the tidal and solar power time series was varied from 0 to 30 days in one-day increments to include potentially coincident tidal and solar power minima.

System Specification, Performance, and Costs
Energies 2021, 14, x FOR PEER REVIEW 20 and manage low tidal power during neap tides. The phase difference between the t and solar power time series was varied from 0 to 30 days in one-day increments to incl potentially coincident tidal and solar power minima. During the wet season in Clarence Strait, daily solar PV electricity generation randomly drop from an average of 432 MWh to as little as 21 MWh [27]. The assessm was carried out for 2018, the worst-case solar intermittency scenario between 2015 2020.
For this case study, Figure 16 shows that the daily energy supplies (Ed) from so PV farms with either zero or 624 MWh storage capacity do not reach the Eth thresh value of 216 MWh. Increasing the solar PV farm's energy storage to 1015 MWh ensu that Ed exceeds the 216 MWh threshold. Alternatively, replacing 25% of the solar PV ergy with tidal energy also ensures that the tidal-solar PV farm exceeds that thresh reducing the required energy storage to 624 MWh and lowering the frequency of Ed ues that are below 300 MWh. During the wet season in Clarence Strait, daily solar PV electricity generation can randomly drop from an average of 432 MWh to as little as 21 MWh [27]. The assessment was carried out for 2018, the worst-case solar intermittency scenario between 2015 and 2020.
For this case study, Figure 16 shows that the daily energy supplies (Ed) from solar PV farms with either zero or 624 MWh storage capacity do not reach the Eth threshold value of 216 MWh. Increasing the solar PV farm's energy storage to 1015 MWh ensures that Ed exceeds the 216 MWh threshold. Alternatively, replacing 25% of the solar PV energy with tidal energy also ensures that the tidal-solar PV farm exceeds that threshold, reducing the required energy storage to 624 MWh and lowering the frequency of Ed values that are below 300 MWh.  During the wet season in Clarence Strait, daily solar PV electricity generation can randomly drop from an average of 432 MWh to as little as 21 MWh [27]. The assessment was carried out for 2018, the worst-case solar intermittency scenario between 2015 and 2020.
For this case study, Figure 16 shows that the daily energy supplies (Ed) from solar PV farms with either zero or 624 MWh storage capacity do not reach the Eth threshold value of 216 MWh. Increasing the solar PV farm's energy storage to 1015 MWh ensures that Ed exceeds the 216 MWh threshold. Alternatively, replacing 25% of the solar PV energy with tidal energy also ensures that the tidal-solar PV farm exceeds that threshold, reducing the required energy storage to 624 MWh and lowering the frequency of Ed values that are below 300 MWh.    Tables 8 and 9 compare the specifications and cost of the standalone solar PV farm with a 1015 MWh battery against a tidal hybrid farm with a smaller 624 MWh battery and 25% of the solar energy replaced by tidal energy. Both farm configurations produce an average power of 18 MW, of which 50% is dispatchable. The tidal hybrid farm also reduced low power events delivering 94% of Ed at levels greater than 300 MWh, and it reduced the system CAPEX by 18% (A$83m). The CAPEX reduction was principally due to a reduction in the battery energy capacity needed for dispatchability. Given that the tidal-solar PV farm supplies energy to a chemical plant, it is also worth comparing annual values for the alternative products, including electricity, hydrogen, liquid e-fuel, and carbon dioxide ( Table 10). The carbon price has been set at the level recommended for transformational change by the IMF and World Bank. Hydrogen has an additional value not given in the table as its production can be rapidly controlled and used to take advantage of power peaks. The price values given are projected for 2025.  (Figures 17 and 18). Both islands have significant reserves of high-grade iron ore, and both rely on sea walls to facilitate mining access to these resources. Over several decades the Cockatoo mine has been closed for long periods and, on two occasions, has flooded. Leaving the mine flooded or dewatering the flooded area too quickly and without adequate control risks breaching the sea wall [46]. A local community has raised concerns that such a breach would result in silt covering nearby coral reefs [47]. Access to a continuous electricity supply could allow pumps to counter water seepage through the sea wall, saving the months needed to pump out the flooded mine and maintain it. Tidal powered pumps may be a cost-effective and more reliable alternative to diesel-powered pumping that could otherwise cost several hundred thousand dollars a month.
to counter water seepage through the sea wall, saving the months needed to pump ou the flooded mine and maintain it. Tidal powered pumps may be a cost-effective and mor reliable alternative to diesel-powered pumping that could otherwise cost several hundre thousand dollars a month.
The ARENA National Tidal Model did not have a sufficiently high spatial resolutio to detect strong tidal streams in the Cockatoo and Koolan Islands, despite the powerfu tidal currents in the neighbouring King Sound and Yampi Sound. However, environmen tal assessments of islands in the Buccaneer Archipelago describe tidal ranges of 4 m to 1 m and tidal currents up to 5 m/s [48]. A local study (Atlantis Resource Corporation, per sonal communication) has recorded flow rates of up to 3 m/s in The Gutter, southeast o Koolan Island. High-resolution surveys to the north and south of Koolan Island (Figur 17) and Cockatoo Island (Figure 18) would be needed to determine if tidal streams ar strong enough to power the Koolan mine's offices, workshops and accommodation, an for mine-dewatering on Cockatoo Island.

Renewable Power Plant Options, Specification, and Cost
Options 1 and 2 in Table 11 are for deep channels north and south of Koolan Island Option 3 is for a tidal turbine reef of eight TECs, each with 4 m rotor diameters [49] tha ing nearby coral reefs [47]. Access to a continuous electricity supply could allow pump to counter water seepage through the sea wall, saving the months needed to pump ou the flooded mine and maintain it. Tidal powered pumps may be a cost-effective and mor reliable alternative to diesel-powered pumping that could otherwise cost several hundre thousand dollars a month.
The ARENA National Tidal Model did not have a sufficiently high spatial resolutio to detect strong tidal streams in the Cockatoo and Koolan Islands, despite the powerfu tidal currents in the neighbouring King Sound and Yampi Sound. However, environmen tal assessments of islands in the Buccaneer Archipelago describe tidal ranges of 4 m to 1 m and tidal currents up to 5 m/s [48]. A local study (Atlantis Resource Corporation, per sonal communication) has recorded flow rates of up to 3 m/s in The Gutter, southeast o Koolan Island. High-resolution surveys to the north and south of Koolan Island (Figur 17) and Cockatoo Island (Figure 18) would be needed to determine if tidal streams ar strong enough to power the Koolan mine's offices, workshops and accommodation, an for mine-dewatering on Cockatoo Island.   Table 11 are for deep channels north and south of Koolan Island Option 3 is for a tidal turbine reef of eight TECs, each with 4 m rotor diameters [49] tha The ARENA National Tidal Model did not have a sufficiently high spatial resolution to detect strong tidal streams in the Cockatoo and Koolan Islands, despite the powerful tidal currents in the neighbouring King Sound and Yampi Sound. However, environmental assessments of islands in the Buccaneer Archipelago describe tidal ranges of 4 m to 11 m and tidal currents up to 5 m/s [48]. A local study (Atlantis Resource Corporation, personal communication) has recorded flow rates of up to 3 m/s in The Gutter, southeast of Koolan Island. High-resolution surveys to the north and south of Koolan Island ( Figure 17) and Cockatoo Island (Figure 18) would be needed to determine if tidal streams are strong enough to power the Koolan mine's offices, workshops and accommodation, and for mine-dewatering on Cockatoo Island.

Renewable Power Plant Options, Specification, and Cost
Options 1 and 2 in Table 11 are for deep channels north and south of Koolan Island. Option 3 is for a tidal turbine reef of eight TECs, each with 4 m rotor diameters [49] that may be suitable for locations near Round Island adjacent to the southeast coast of Koolan Island. Option 4 is for a small tidal farm in deeper water either north or south of Cockatoo Island. In this option, the rate of water seepage requiring pump out may fall during neap tides, so the energy storage required to manage neap tides has been reduced, compared to options 1 to 3. The four suggested options are contingent on further high-resolution surveys finding suitable tidal streams with maximum flow velocities of at least 2.5 m/s.  Table 12 shows the performance and cost data for a range of potential locations and site depths. The first three options were to supply office and accommodation power on Koolan Island. The fourth option is to supply electricity to dewater the mine on Cockatoo Island. Options 1 and 2 are for deep channels north and south of Koolan Island. Option 3 is for a tidal turbine reef of eight TECs, each with 4 m rotor diameters [49] that may be suitable for locations near Round Island adjacent to the southeast coast of Koolan Island. Option 4 is for a small tidal farm in deeper water either north or south of Cockatoo Island. In this option, the rate of water seepage requiring pump out may fall during neap tides, so the energy storage required to manage neap tides has been reduced, compared to options 1 to 3. The four suggested options are contingent on further high-resolution surveys finding suitable tidal streams with maximum flow velocities of at least 2.5 m/s.

Overview
Port Melville is a remote defence and shipping service facility located at the northern end of Apsley Strait between Melville and Bathurst Islands. It is a security regulated port operated by Northern Territory Port and Marine and can house 150 personnel. It provides a range of services and storage facilities for civilian and defence shipping. This case study describes the contribution tidal energy could make to meet a demand for reliable emergency power at the port. The tidal energy resource being considered is located 2 km to 3 km south of Port Melville at a narrow section of Apsley Strait.
The Port Melville tidal assessments suggest that tidal streams in this area may be suitable for electricity generation. However, the tidal flow is significantly stronger for an ebb tide than for a flood tide. The peak flooding tide from the north, at Melville Port wharf, is generally 0.75 m/s and 1.25 m/s during spring tides. The ebbing tide can rise over two hours to flow rates of 1.5 m/s to 2 m/s and up to 2.5 m/s during spring tides. The ebbing tide flow rates persist for three hours, after which they rapidly slow [50,51].
The ARENA National Tidal Model identified a strong tidal stream south of Harris Island at (11 • 25.8 S, 130 • 24.6 E,) ( Figure 19), but this may be located too close to Port Melville and interfere with its operations. The hypothetical installation is at the south end of a 45 m deep 500 m wide channel from Port Melville wharf to 2.5 km south of Port Melville. The channel narrows and becomes shallower at its southern end, which may increase the tidal stream flow rate. A hypothetical tidal farm was evaluated for this location (11 • 26.60 S, 130 • 24.78 E).

Port Melville Tidal Power Generator-Specification and Costs
Three TEC configurations are given. The 10 m and 26 m rotor turbine arrays are for deeper sites about 1.6 km south of Harris Island, and a tidal turbine reef [49] is proposed for shallower locations. Vanadium flow cell batteries are suggested to accommodate an estimated 60 h, half power, neap tide phase. Table 13 shows the specification, and Table 14 gives indicative capital costs for the three tidal turbine options. Each of these is contingent on confirming that the tidal stream resource has a semidiurnal peak velocity of at least 2.5 m/s.

Discussion
Despite Australia's moderate tidal stream flow velocities, tidal power's regularity makes it a renewable energy resource worth considering for critical applications. These might include chemical processing, emergency power, security, or supplying regional microgrids where intermittent power shutdowns may not be tolerable.
Case Study 1 described the use of tidal-wind farms to supply dispatchable power. Replacing 50% of the wind farm power with tidal power made it possible to provide 40% of the average power generation as dispatchable power. The replacement also reduced low power events above the dispatchability threshold and provided a CAPEX reduction of 20% (A$105m) for an offshore tidal-wind farm and 5% (A$20m) to 11% (A$43m) for a land-based hybrid farm depending on its location. The cost savings were principally due to the continuity and regularity of tidal power in Banks Strait and the reduced charge and discharge periods required for tidal-wind farm energy storage compared with 132 h of energy storage needed for a standalone wind farm.
Case Study 2 demonstrated that even with extended period neap tides and the much cheaper cost of solar power, using tidal power to replace 25% of a solar PV farm's capacity may provide cost-efficient dispatchability. The tidal and solar PV farm exceeded a 50% dispatchability threshold. It reduced low power events delivering 94% of Ed at levels greater than 300 MWh. It also reduced the system CAPEX by 18% (A$83m).
Case studies 3 and 4 suggested opportunities to provide small scale emergency power from standalone tidal energy converters. The principal advantages would be the ability to supply electricity for long periods from unattended remote or transportable TECs, or the potential for a renewable electricity supply on islands where a solar PV farm might cover too great an area.
The case studies are hypothetical and require a more detailed assessment of the tidal resource, capital, and operation and maintenance costs before implementation. Case studies such as those presented here are not conclusive, but they suggest directions for further work. Future studies could extend databases on the intermittency characteristics of solar and wind resources for high-security power in remote regions. They could also assess prototypes of the large rotor, reduced rated power, TECs outlined in this study for moderate tidal stream flow environments. If such TECs proved viable, they would significantly expand the global potential for using tidal stream power.

Conclusions
We have shown that the cost of supplying dispatchable power from variable sources, such as wind and solar energy farms, may be significantly reduced if tidal energy is also available and is exploited. Hybrid tidal-wind or tidal-solar energy farms may provide cost savings, power availability and levels of dispatchability that are higher than can be achieved by using any of these resources by themselves. Table 15 compares tidal-solar hybrid and tidal-wind hybrid examples, based on data measured from moderate power tidal sites in Australia. The improved dispatchability shown in Table 15 is because wind, solar and tidal energy have high levels of temporal complementarity; that is, when one resource is not available, it may in whole or part be replaced by another. The cost savings are primarily due to the reduced battery capacity needed for dispatchability. A novel algorithm has been described that allows combinations of tidal and wind or tidal and solar energy resources to be optimised for specified levels of dispatchability, power, battery capacity or capital cost.