Preliminary assessment of innovative seawater reverse osmosis (SWRO) desalination powered by a hybrid solar photovoltaic (PV) - Tidal range energy system

Desalination 477 (2020) 114247

Contents lists available at ScienceDirect

Desalination journal homepage: www.elsevier.com/locate/desal

Preliminary assessment of innovative seawater reverse osmosis (SWRO) desalination powered by a hybrid solar photovoltaic (PV) - Tidal range energy system

T

⁎

Agustín M. Delgado-Torresa, Lourdes García-Rodríguezb, , María Jiménez del Moralb a Departamento de Ingeniería Industrial, Escuela Superior de Ingeniería y Tecnología (ESIT), Universidad de La Laguna (ULL), Avda. Astrofísico Francisco Sánchez s/n, 38206 La Laguna, Tenerife, Spain b Escuela Técnica Superior de Ingeniería, Universidad de Sevilla, ETSI, Camino de Los Descubrimientos, s/n, 41092 Sevilla, Spain

A R T I C LE I N FO

A B S T R A C T

Keywords: Tidal energy Tidal-powered desalination Reverse osmosis Hybrid solar PV/tidal

This paper proposes an innovative desalination technology for sustainable off-grid systems taking advantage of complementary features of tidal range and solar PhotoVoltaic (PV) energies. According to the literature survey, this proposal has not been considered before. Since fresh water production can be easily stored, the key issue in designing SeaWater Reverse Osmosis (SWRO) desalination plants is to minimise the capital cost required per m3 of fresh water produced throughout the plant lifetime. In addition, water cost of renewable energy - driven desalination strongly depends on decisions concerning battery capacity and nominal power installed, thus hybrid systems play and important role in this regard. The energy analysis performed quantifies the temporal complementarity of tidal and solar resources in an exemplary case study of a semiarid plant location at Broome, Australia. An estimation of the yearly energy production profile of the tidal range power plant is calculated with a zero-dimensional numerical model whereas the System Advisor Model (SAM) is used for the solar PV plant. The main result obtained is the great temporary complementarity of tidal and solar photovoltaic resources for SWRO application. For a given size of the PV generator the inclusion of the tidal range power plant implies an increase of the operating time of the desalination plant at nominal capacity between 1.8 and 2.8 times compared to the only solar PV driven case. This result depends on the SWRO nominal capacity. The recommended design for the case study consists in off-grid desalination plants, with minimum battery capacity if any, powered by a hybrid energy system with a ratio of installed desalination capacities of 55·103 m3/d per each hydraulic turbine of 10 MW. This system can be operated at full load a 42% of the year maximizing the yearly fresh water production.

1. Introduction This paper proposes an innovative Renewable Energy (RE) desalination technology based on a hybrid tidal range/solar PhotoVoltaic (PV) energy system for medium to large capacity Sea Water Reverse Osmosis (SWRO). The main objective is to prove the interesting prospects that said innovative combination offers to promote the development of sustainable desalination at a large scale due to both, 1) the predictable behavior and infallibility of tidal range energy throughout the year and 2) the solar resource availability at the same time that

peak water demands. These simultaneous features result in ensuring a minimum fresh water production nearly every day and relatively high percentage of time throughout the year during which the plant would operate at full load in comparison to other RE based desalination solutions. The assessment is illustrated through an off-grid system in Broome, Australia. However, as with any hybrid RE – based system, this hybrid option would have interesting prospects in regions with good solar and tidal range resources. For coastal regions with high tidal range, the higher the available solar resource the higher the suitability of the proposed configuration.

Abbreviations: AC, alternative current; BP, Booster Pump; DC, direct current; HW, high water; HPP, High Pressure Pump; IEA, International Energy Agency; LA, Level of Autonomy; LCE, Levelized Cost of Energy; LST, Local Standard Time; LW, Low Water; NPV, Net Present Value; NREL, National Renewable Energy Laboratory; OTEC, Ocean Thermal Energy Conversion; PEX, Pressure EXchanger; RO, reverse osmosis; PV, PhotoVoltaic; RE, Renewable Energy; SAM, System Advisor Model; SEC, Specific Energy Consumption; SWRO, SeaWater Reverse Osmosis; TDS, total dissolved solids; US, United States; 0D, zero – dimensional; 1D, one - dimensional; 2D, two - dimensional; 3D, three - dimensional ⁎ Corresponding author. E-mail addresses: [email protected] (A.M. Delgado-Torres), [email protected] (L. García-Rodríguez), [email protected] (M.J. del Moral). https://doi.org/10.1016/j.desal.2019.114247 Received 24 July 2019; Received in revised form 24 November 2019; Accepted 25 November 2019 0011-9164/ © 2019 Published by Elsevier B.V.

Desalination 477 (2020) 114247

A.M. Delgado-Torres, et al.

Nomenclature Abasin Ain Asluice cISWRO CISWRO D Ee,useless Ee,surplus Ee,useful fg g Gp Gb,N Gd,h GPV Hsea Hbasin H Hmin Hst lt

nt Qin Qout Qp Pe,PV Pe,tidal Pe,total

total basin area, m2 total filling flow area, m2 total sluice gate's area, m2 investment cost of the desalination plant per unit on nominal capacity, €/(m3/d) total investment cost of the desalination plant, € turbine diameter, m yearly useless energy, J yearly surplus energy, J yearly useful energy, J electricity grid frequency, Hz gravitational acceleration, m/s2 number of generator poles direct normal solar irradiance, W/m2 horizontal diffuse solar irradiance, W/m2 incident solar irradiance on the plane of the photovoltaic array, W/m2 tide level, m basin level, m available head, m minimum generation head, m generation starting head, m lifetime of the desalination plant, years

Pe,nom P⁎e Pe,useless Pe,surplus Tdry V10 Vp τ τtidal τPV τtidal+PV ηgen Δt

number of tidal turbines filling volumetric flow rate, m3/s volumetric discharge flow rate, m3/s output capacity of the desalination plant, m3/h electrical power output of the solar photovoltaic plant, W electrical power output of the tidal range power plant, W electrical power output of the hybrid tidal range/solar photovoltaic energy system, W nominal electrical power output of the tidal turbine, W given power level to compute the power availability factor, W useless electrical power output, W surplus electrical power output, W dry bulb temperature, °C wind speed at 10 m, m/s volume of fresh water produced, m3 power availability factor power availability factor computed only with tidal plant power output power availability factor computed only with solar photovoltaic plant power output power availability factor of the hybrid system generator efficiency time interval

factor is of particular relevance when dealing with energies that may have a high degree of intermittency as wind, solar or wave energy. Within this framework, this paper presents a preliminary assessment of the innovative combination of solar PV and tidal range energy to drive a SWRO desalination system. RO process has been the dominant technology in the increasing of the global desalination capacity since the early 1990s. At present, 84% of the desalination plants in operation are RO plants, which produce 69% of the desalinated water in the world [1]. For those readers less familiar with this technology the review paper presented by [4] is recommended where the basic principles and information about the characteristics of feed waters, the phenomenon of fouling and cleaning of membranes and the pre-treatment processes can be consulted. Information on the design of the RO system, costs, the post-treatment of the permeate and the management of the rejected brine is also provided. In the brief analysis about the use of renewable energies in [4] the solar PV and wind energies are the two main options to be combined with SWRO plants. In the case of Brackish Water RO (BWRO), solar PV is the most commonly used option, citing an interesting experience in Australia on the testing of six solar PV - BWRO systems [4]. In a more recent comprehensive review work [5] all the aspects of RO technology are characterized even in a deeper and more detailed way than in [4]. RO powered by RE is also addressed in [5] indicating that almost half of the desalination plants powered by solar energy in the world are RO plants of which three quarters use solar PV as a source of energy. In the field of renewables, some information on the use of ocean energy is presented in [5] describing some existing experiences on the use of wave energy to drive RO plants. Both [4,5] coincide in the most important technological challenges of the RO: the fouling of membranes, the degradation of polyamide membranes by contact with chlorine, the potential use of technology for the removal of certain organic pollutants and the discharge of brine. Finally, the possibility of reducing the environmental impact of the technology and the cost of desalinated water through the use of renewable energies is also pointed in [5]. In this regard, neither [4] nor [5] mention the use of the tidal range or the hybridization tidal range/ solar PV to drive a RO system. With regard to the off - grid RO desalination systems, most of the work (theoretical or experimental) is about low capacity systems

The significant increase in water demand in large regions of the world in recent years has made the desalination of seawater and brackish water a key element in the solution to the problem. To illustrate this, it is enough to look at some of the last global data published by [1], such as the following: 1) the global desalination capacity with plants in operation jumped from approximately 30 million m3/day in 2005 to almost 80 in 2015 and 2) there were approximately 8000 desalination plants in operation in 2005, which became almost 14,000 in 2015. Nowadays, there are 15,906 plants in operation with a global capacity of 95.37 million m3/day [1]. However, desalination processes are energy-intensive and the brine (concentrate) disposal can have an adverse hydro environmental impact depending on the discharge method employed. With regard to the energy demand it is currently satisfied with electrical or thermal energy generated by the combustion of fossil fuels. This explains some of the main energy related - disadvantages of water desalination as it is currently carried out: high operating costs and negative environmental impact associated with the emission of greenhouse gases. In other words, such a desalination global system scheme can be considered unsustainable. All the above justifies the interest in RE-driven desalination. This problem is discussed in [2] in relation to Australia. In that work the carbon footprint of desalination in that country is measured over the period 2005–2015 by rating the 20 largest existing desalination plants, being the reverse osmosis (RO) the dominant technology. The study assumes that the electricity consumed in the construction and operation of these plants is generated by conventional power plants. The authors conclude that the electricity sector is the main source of carbon emissions associated with desalination activity. For example, 96% of the total emissions in the operation phase of desalination plants corresponded to this sector in 2015. From the results of this study, the increase of renewable electricity share is considered by [2] as the key factor in solving the nonsustainability of the current desalination sector in Australia. However, there were only 131 desalination plants powered by REs in the world by the end of 2017 representing < 1% of global desalination capacity [3]. Approximately 80% of these plants are powered by solar PV, solar thermal and wind energy. Some of the reasons for the above are the temporal variability of the renewable resources, their limited availability for a given location and high capital costs. The first 2

Desalination 477 (2020) 114247

A.M. Delgado-Torres, et al.

combination of any of the forms of ocean energy for water desalination is not even mentioned in the early review papers published by [12,13]. After these, [14] is the first bibliographical review on desalination and REs in which the case of ocean energy - powered desalination is addressed. In said work, two studies on the application of ocean thermal energy in distillation plants and various experiences related to wavepowered RO are described. Finally, the proposal for a hybrid wind-wave energy system connected to a distillation plant is also mentioned, being the main advantage of this combination the increase of the regularity of the energy source. However, there is no design or system based on the use of the tidal range energy for desalination by RO in [14] either. The most recent information available so far on the use of ocean energy for desalination is found in the excellent work of [15] which is complemented in relation to wave energy by the also very recent paper of [16]. These two specific studies together with the information contained in the general review work of [3] serve to establish the current state of the ocean energy - desalination combination. The review presented by [15] is the first specific review paper on water desalination using ocean energy. This work deals with all forms of ocean energy and their application in different desalination technologies. In the case of the Ocean Thermal Energy Conversion (OTEC), the electricity generated could be used by a RO system. However, the description of the existing experiences and works in [15] indicates that the main option studied is the use of open or hybrid cycles for the production of electricity and water being the latter generated by distillation. Among current, tidal and wave energies the latter has been the main studied option to drive a RO system. It would be possible to produce both electrical energy and pressure energy simultaneously with the associated wave energy technology, both of which are necessary in the RO process. The desalination systems based on wave energy are described in [15] including the first commercial wave powered - RO system: the CETO project developed in Australia. Of special interest for this paper, there is no reference to RO systems driven by tidal range energy in [15]. In relation to the production of freshwater with wave energy the recent work of [16] is recommended for the readers interested in this option. Finally, a paper not cited in [15] but detected in the literature review performed in this paper is the experimental work of [17] where the results of a laboratory scale tidal current turbine (not a tidal range turbine) are presented. The same design is then used together with a RO unit installed on a ship sailing. The tidal current turbine generates pressurized seawater for the reverse osmosis membranes by means of a directly connected plunger pump. The use of ocean energy for desalination is also reviewed in the subsequent review paper of [3] on recent progress in RE - driven desalination. In the case of tidal energy, only one study in which RO would be used as a desalination technology is referenced. In said work, the economic evaluation of a system with direct coupling between the tidal turbine and the RO unit is presented. A double - basin tidal range power plant is assumed in the assessment to increase the power output generation time. However, no energy evaluation is carried out and the tidal plant is not characterized, simply assuming a load factor of 85% ad hoc to incorporate its contribution in the economic model of the whole system [18]. Therefore, from all the information given before, the verification of the novelty of the proposed hybrid configuration tidal range - solar PV to drive RO desalination systems is the first relevant result of this work since no previous study has been detected in the literature. In this context, the preliminary assessment of an off-grid high-capacity RE - driven SWRO desalination system is performed in this paper to investigate the complementarity of the solar PV and tidal range energy resources and its benefit for seawater desalination. The results of the study have been obtained for a specific location but the developed analysis may be applied to another different location if the information on the solar and tidal resources is available. In the first part of the paper the main characteristics of a RO system are set to determine its Specific

because this is the typical range of RE - driven RO desalination systems. One of the most recent studies is the techno-economic study of freshwater production by PV-RO in the Aegean islands presented by [6]. In these islands, the demand for freshwater is currently met by transporting it by ship. Five different scenarios for the PV-RO off-grid system with freshwater storage tank are considered in the study, two of them without batteries for electrical storage. The operating modes of the SWRO plant at nominal or variable load are also considered. In addition, a different energy management and control system is proposed for each scenario. Simulations are performed to optimize the unit cost of water obtaining the optimal values of the water storage tank volume, PV generator's size and battery system specifications for each scenario [6]. On the other side, a recent example about off-grid large capacity RE - desalination systems is the study presented by [7]. In this work the design of a solar PV system to drive a 90,920 m3/day desalination plant in Sharjah, United Arab Emirates is proposed. The on - grid and off – grid PV options are proposed with diesel generator and batteries in the latter, being the direct connection of the desalination plant to the grid the baseline case. The optimal configuration is determined minimizing the net present cost through the modeling and simulation of each element of the system. The results show that the on-grid PV option yield the lowest cost of energy but the off-grid system exhibit a renewable fraction considerably higher and thus the lowest CO2 emissions [7]. Within this framework, ocean energy can be an interesting option as an energy source for desalination processes for several reasons. First of all, it is an abundant source of energy. According to recent studies by the International Energy Agency (IEA) 20,000 TWh of electricity could be globally generated with this energy source [8]. Secondly, in some of its modalities it is a highly predictable resource for a given location even in the long term. That is the case of tidal range energy, which is the option discussed in this article. Finally, as an aspect of special interest to the case of seawater desalination, both the energy production facility and the desalination plant would be located on the coast. In general, there are four forms of usable energy in the oceans: tidal energy, wave energy, oceanic thermal energy and energy from the oceanic saline gradients. Updated information on these can be found in a recently published review paper [8] where information about fundamentals, resource assessment and technologies are globally reviewed, among other aspects. Up to now, the technological development of the ocean energy has focused mainly on wave and tidal energy. In the case of the latter there are two main options: tidal stream energy and tidal range. A specific review about the operating principles of both, their main issues and future developments is presented in [9]. It is also discussed the current status of both technologies giving detailed information of the tidal range power plants in operation. In the case of tidal current technology information on the most promising designs of devices (up to a total of 14) is provided. Finally, the tidal energy policy is also treated because it has been the main driver for the tidal energy development [9]. In general, tidal range option presents a greater technological maturity and interesting characteristics for this assessment study as its high degree of regularity and predictability [10]. These characteristics are derived from the natural marine tidal phenomena. This option is assessed in [11] where information on existing tidal range power plant projects is provided. More specifically, the two most important plants in operation with this technology are described: the 240 MW plant in La Rance (France) and the Lake Sihwa plant (South Korea) with 254 MW of installed power. These two plants represent 95% of the total tidal range power installed worldwide. A description of the existing tidal range turbines (bulb and Straflo turbines) is also presented in [11]. Their characteristics of efficiency, cost, maintenance and environmental problems are analyzed and the same is also presented about future designs of tidal range turbines. In addition to the already given information by [4,5], no design proposals of RO systems powered by tidal range or hybrid RE solar PV – tidal energy systems have been considered before according to the rest of the bibliographic review performed for this investigation. In fact, the 3

Desalination 477 (2020) 114247

A.M. Delgado-Torres, et al.

to analyze other options before such as the incorporation of the pumping and the maximum holding time in the tidal range power plant [20] and the variable load operation of the SWRO plant. The latter is particularly interesting due to the possibility of breaking down the total power consumption of the SWRO plant according to its different subprocesses and the own modularity of the RO technology [21,22].

Energy Consumption (SEC) for the feed water characteristics of the location. In this way, the nominal power consumption of the plant can be determined by its capacity thanks to the modularity of the RO technology. Next, the model used for the quasi-dynamic simulation of the tidal range power plant is presented. Since this is a preliminary evaluation, a zero-dimensional model has been applied using the hill chart method for the modeling of the tidal turbine. The output of this modeling approach is the power output profile of the plant throughout the year. In addition to the above, a model for the output calculation of the solar PV plant is also needed. In this work, the PVWatts model executed in the System Advisory Model (SAM) software is used. With the previous results, the total power output profile of the hybrid solar PV – tidal range power plant is obtained. Afterwards, the operation time of the SWRO plant at nominal load and the unusable energy for desalination are determined as a function of the plant's capacity. To express the operating time of the SWRO plant the power availability factor (τ) is defined. In the field of hybrid systems, other parameters have been used to express their performance as presented by [19]. In. general, these parameters can be of technical, economic or techno-economic nature as, for example, the Level of Autonomy (LA), the Net Present Value (NPV) or the Levelized Cost of Energy (LCE). However, given the specific application of desalination discussed in this paper, it has been considered more appropriate to use the defined power availability factor as it directly provides an idea of the freshwater produced by the SWRO plant. In addition, with the implemented modeling approach, it is possible to determine the τ values for the operation of the solar PV plant and the tidal range plant alone. This is an interesting aspect to evaluate the suitability of the proposed hybridization because it allows to perform the comparison with respect to the only PV case. With all the above it is possible to obtain the energy used in desalination, the amount of desalinated water and the not used energy (surplus and useless). Finally, a preliminary economic assessment of the unitary cost of investment of the desalination system is performed. The location chosen for the analysis in the northwest of Australia has good conditions from the point of view of tidal and solar resources. As will be seen, for fixed sizes of the solar PV generator and the tidal range turbine there is a size of the SWRO plant that maximizes the amount of water produced annually. Finally, as this paper is the preliminary assessment of a hybridization not proposed before, the options available for the management of the unusable energy by the desalination plant have not been considered. This aspect will be addressed in future works where the different possibilities for achieving higher power availability factor of the hybrid system will be analyzed. These include, of course, the use of batteries for electrical storage. However, the objective of the authors is

2. System description and calculation procedure 2.1. Basic layout of the hybrid solar PV/tidal range SWRO desalination system Fig. 1 shows the simplified general basic scheme of the proposed hybrid tidal range/solar PV desalination system. The solar PV plant generates electricity from solar energy with the main elements shown: the PV generator and the DC to AC inverter. On the other hand, a tidal barrage serves to separate the sea from the basin in the tidal range power plant. Tidal turbines are installed inside the tidal barrage to obtain power from the height difference between the basin and sea levels. The case represented in Fig. 1 corresponds to an ebb - mode generation situation in which the electrical energy is produced by the turbines when the basin level is above the sea level. This operation mode is the considered in this work so the basin is filled when the sea level is higher than the basin level through the idle turbines and the sluice gates installed in the tidal barrage for this purpose. The electrical power generated by the two plants will be consumed by the SWRO plant whenever possible. 2.2. SWRO plant and its specific energy consumption (SEC) The main characteristics and the energy consumption of the RO process is determined in this section. In case of on-grid desalination systems the best selection of its nominal capacity depends on the electric tariff since the surplus energy could be sold. In off-grid SWRO plants as the considered in this paper, the conventional design means an ON/OFF operation strategy at nominal load (desalination plant at full load). Therefore, the SWRO desalination plant is defined by the specific energy consumption so the nominal power consumption allows to calculate the hourly fresh water production once its nominal capacity is selected. For the preliminary design of the reverse osmosis plant in Broome, seawater with a total salinity of 37 g/L, density of 1025.5 kg/m3 and temperature of 24 °C has been considered. This is the most unfavorable case from the energy consumption point of view. The ROSA design software [23] was used to calculate the SEC of the core of the process –

Fig. 1. Basic scheme of the hybrid tidal range/solar photovoltaic seawater desalination system. 4

Desalination 477 (2020) 114247

A.M. Delgado-Torres, et al.

Along with all the above information it is also necessary an operating mode of the plant. In general, there are three methods of operation: ebb only, flood only and two-way generation (bidirectional). Flood generation is the less efficient option among these three methods and the bidirectional mode generates more energy. However, the latter is also the most complicated and costly because the turbines to be used must be optimized to operate with a high efficiency in both directions of the flow [9]. Since the aim of this paper is a preliminary assessment, the ebb only generation mode has been considered. More detailed information on the operation modes of the plants can be found in [9,11]. The algorithm used in the 0D modeling approach is based on the proposed by [29] but for simplicity the maximum holding time has not been included in the numerical modeling. In the ebb mode the power generation takes place only if the basin level (Hbasin) is higher than the tide's level (Hsea), i.e. it occurs only within the emptying period of the basin. Fig. 3 shows the qualitatively scheme of operation over a complete two tidal cycles. It is assumed that the tide level is known as a function of time and that Afilling and Abasin values have also been set considering the latter depth - independent. With this information the available head (H) is defined as:

see Fig. 2 -, which requires, among other input information, the detailed seawater composition. This composition has been calculated with the same relative ratio of individual components to that of seawater in the Canary Islands [21]. Tables 1 and 2 show, respectively, main input parameters and results obtained for the best design found. For the design several membrane models were considered corresponding the best selection to the SW30XLE-440i membrane module (see Table 2) since the use of higher permeability modules results in lower permeate quality. When the energy recovery system (efficiency 95%) is included in the design (Fig. 2), the specific energy consumption drops to 1.96 kWh/m3. Besides that, [24] provides representative values of auxiliary energy consumption although they strongly depend on plant location, so a total SEC of 3.5 kWh/m3 is selected. This fits the available value of Adelaide desalination plant with a specific energy consumption of 3.48 kWh/m3 [25]. 2.3. Modeling of the tidal range power plant For simplicity and due to the novelty of the hybrid tidal range/solar PV - SWRO configuration the design criterion adopted for the preliminary assessment consists in selecting a single hydraulic tidal range turbine and the associated area of the basin to analyze the performance of the energy system. This approach is supported by the fact that the power output of a tidal range power plant is the sum of the power produced by each single turbine. Concerning the hybrid power plant, once studied the percentage of the year in operation with power output greater than a given value, a reasonable nominal desalination capacity will be selected. The modeling of a tidal range power plant can be performed with numerical models of different complexity from zero-dimensional (0D) approaches to more computing-demanding three-dimensional (3D) schemes [10]. Two-dimensional and 3D models are appropriate for a more precise evaluation of the tidal resource available in a given location and, especially, to study the impact of the plant on its nearby area [20,26,27]. These are obviously more complex models from a mathematical point of view and more computationally demanding. On the other hand, 1D models are considered suitable for simulating relatively small projects and for evaluating the impact of the location of barrages in estuaries and plant operating strategies. Finally, 0D modeling does not provide results about the hydrodynamic impact that the existence of the plant would have on its environment [10]. Such a method requires minimal information about the specific location for its application and the output results can be considered acceptable for the purpose of a preliminary energy assessment [26–29]. Therefore, a 0D numerical model for the tidal range power plant based on the work of [29] has been implemented for this study. The steps of the development of this model are explained below. As it is necessary with any RE-based system, the first step for its energy production assessment is the characterization of the available resource. In the case of tidal energy, this characterization corresponds to the recording of information on the evolution of the tidal level over the period of time to be studied (Hsea). This information is indispensable for the simulation of the operation of a tidal range plant whatever the modeling approach used. On the other hand, it is also necessary to know the wetted surface area of the basin (Abasin) and the flow area throughout the hydraulic structure in the filling periods of the basin (Afilling). In this work it is assumed a constant Abasin for the simplicity derived from the preliminary assessment concept. This parameter can vary with time in regions with high tidal range energy resource which must be considered in more detailed studies. It is also considered that the filling of the basis is carried out through the idle turbines and sluice gates of total area Asluice Finally, the conversion of potential energy to mechanical or electrical energy takes place in the installed hydraulic turbines through which the water flows. The operation of these turbines requires a specific model that provides the power produced for given values of discharge flow (Qout) and available head (H).

H = (Hbasin − Hsea)

(1)

Generating period requires H positive and within an adequate range to achieve efficient operation of the selected hydraulic turbine (see Fig. 3). Starting head, Hst, and minimum head, Hmin, define said range. Negative values of H correspond to filling period. Otherwise the level of the basin remains unchanged until generating period starts or from it ends. Therefore, once H is evaluated at a given instant, ti:

H (ti ) = Hbasin (ti ) − Hsea (ti )

(2)

next value depends on the possible mode, either “filling” or “generating”, or remains unchanged. ■ Generating mode starts when H(ti) = Hst and remains until H (ti) > Hmin. This work considers starting and minimum heads of 2 m and 1 m, respectively. Generation mode means:

Hbasin (ti + 1) = Hbasin (ti ) −

nt ·Qout (ti )·∆t Abasin

(3)

where Qout is the turbine discharge flow rate whose value is determined with the turbine performance model. nt is the number of hydraulic tidal turbines installed in the plant. In this work nt = 1. ■ Filling mode (H(ti) < 0) leads to:

Hbasin (ti + 1) = Hbasin (ti) +

Qin (ti )·∆t Abasin

(4)

where Qin is the total filling flow rate computed as Qin = Cd·Ain · 2·g ·|H| and Cd is the discharge coefficient of the hydraulic structure. In this work Cd = 1.

Fig. 2. Conceptual diagram of the core of a SeaWater Reverse Osmosis (SWRO) desalination plant. HHP: High Pressure Pump; BP: Booster Pump; PEX: Energy recovery device based on Pressure EXchanger. 5

Desalination 477 (2020) 114247

A.M. Delgado-Torres, et al.

2.4. Tidal turbine selection and modeling

Table 1 Input parameters of the SWRO energy consumption calculation.

Once the 0D numerical model has been established for the simulation of the plant, a performance model of the installed turbine is also necessary in addition to the values of the wetted basin and total sluicing areas. The modeling of the turbine's operation is performed in this work with the widely employed in the most recent works on tidal range power plants: the hill chart method [20,29–33]. With this approach the power output of the turbine and discharge flow can be computed as a function of the available head if the maximum turbine capacity and diameter are known in addition to the generator's number of poles. Fig. 5 shows the result of applying this model to the turbine used in this work whose characteristics are given in Table 3. Along with this information, data of the 20 MW turbine considered in [29] for the design proposal of the Swansea Bay Lagoon project are also given. The power and flow discharge curves for both cases are presented to check the correct application of the hill chart method.

Parameter Feedwater TDS Feedwater temperature Feedwater density Membrane elements per pressure vessel Average flux of the system Recovery

37,000 mg/L 24 °C 1025.5 kg/m3 7 12.75 L/(m2·h) 42%

Table 2 Results of ROSA software with the two best membrane element models considered without brine energy recovery. Parameter

Feed pressure – gauge pressure - [bar] Average flux, L/(m2·h) System recovery, % Brine pressure [bar] Product TDS [mg/l] SEC [kWh/m3]

Membrane element model SW30HRLE – 440i

SW30XLE – 440i

53.34 12.75 41.99 51.76 188.7 4.41

51.26 12.75 42.0 49.69 225.1 4.24

2.5. Tidal resource at Broome (Australia) The time series of the tide level (Hsea) on the Broome coast has been freely obtained from the website of the Australian Government Bureau of Meteorology [34]. Within the framework of the Australian Sea Level Monitoring Project, the sea level in a number of locations along Australia's coastline is monitored including the location subject of this study. In the case of Broome available hourly values of the sea level covers from the year 1991 to 2019. This work is based on 2018 data. Therefore, the tidal resource information used is of empirical nature instead of an alternative approach based on the sum of the tidal constituents.

■ If ti-1 corresponds to filling mode and 0 ≤ H(ti) < Hst, or if ti-1 corresponds to generating mode and 0 ≤ H(ti) < Hmin the plant remains in holding mode and the effect on Hbasin is described by:

Hbasin (ti + 1) = Hbasin (ti )

(5)

The complete and detailed flow diagram of the implemented 0D model is given in Fig. 4.

Fig. 3. Conceptual diagrams of operation modes and corresponding production in a tidal range power plant in ebb – mode. HW: Holding water at High Water. LW: Holding water at Low Water. 6

Desalination 477 (2020) 114247

A.M. Delgado-Torres, et al.

, ,

,

(

,

INPUT , , , , Abasin, Asluice, nt, = + ∙ ∙ ( ⁄2) ,.., ,∆ = ( − ) ,…, , ); , ;

=

−

,

¿

,

,

< 0?

¿

NO

≥ 0?

NO

¿

YES

YES

¿

IF = "illing" OR = "ℎ " OR = "holding HW" OR =" "

¿

<

NO

?

IF

=

,

,

=

,

∙

∙ 2∙ ,

YES

"illing" OR

YES

IF

"holding HW"

= "holding LW" OR "holding HW" OR “generating”

= "holding LW" OR "generating"

∙ ,

+

∙∆

¿

<

= "holding HW"

= "holding LW"

=

,

YES

?

=

IF

= "illing" , =0 , =0

≥

¿

YES

< ? OR < < ?

,

,

,

=

= "generating"

, =0 =0

Tidal turbine model (hill chart method)

,

?

,

>0;

,

,

=

>0

,

=0

, ,

−

∙

,

∙∆

NO

OUTPUT ; Fig. 4. Detailed flowchart of the 0D modeling approach implemented.

available to model the evolution of the plant in each tidal cycle. This constraint makes difficult the analysis with the quasi – dynamic 0D simulation, especially when power generation should start (H = Hst) and should stop (H = Hmin) and wait to start filling (see Fig. 3). This problem does not exist when the tide level is determined through the

Extracts of the raw tidal data obtained from the Bureau of Meteorology are depicted in Figs. 6 and 7. The basic physics of the oscillation shown in Fig. 6 can be consulted, for example in [9].The interval shown in Fig. 7 corresponds to 22 December. Since the raw data are supplied in an hourly base, only 11–12 points would be 7

Desalination 477 (2020) 114247

A.M. Delgado-Torres, et al.

Fig. 5. Model of selected turbine of 10 MW in comparison to that selected for the Swansea Bay Lagoon project (20 MW).

Fig. 7. Comparison of raw and interpolated data of sea level (Hsea) at Broome (hour of the year in the x-axis).

Table 3 Tidal rage turbine specifications.

Turbine capacity [MW] Generator poles, Gp Turbine diameter, D [m] Electricity grid frequency, fg [Hz] Turbine speed [rpm]

Turbine considered in [29]

Turbine considered in this work

20 97 7.35 50

10 (generator capacity) 80 6.0 50

61.9

75

allow to complete 29/01/18 and 31/01/18. Finally, the average of both of them was assigned to the intermediate day 30/01/18. Analogous procedure was applied to other missing interval of time, since 18:00 h of 03/07 up to 2:00 h of 05/07, from 13:00 h of 11/05 to 14:00 h of 12/ 05 and to the missing days 28–29 of August. Fig. 8 shows the sea level data throughout the year 2018 at Broome. Another way to complete the time series when lack of data is found could be through the generation of said missing data with the sum of the tidal constituents. 2.6. Modeling of the solar PV plant and solar resource at Broome (Australia) For a preliminary analysis of the tidal range/solar PV combination a detailed model of the photovoltaic plant is not necessary. For this reason the PVWatts model available in the System Advisor Model (SAM) developed by National Renewable Energy Laboratory (NREL) has been applied [35]. The necessary input parameters to run the model and their values set to obtain the output power curves of the photovoltaic plant are shown in Table 4. In this work, the nominal power of the PV plant corresponds to 1.5 kWp per kW of nominal power of the tidal plant generator in order to avoid oversizing of the solar PV field but achieving the nominal production of the tidal plant in sunny days. Concerning the solar resource, hourly data of global irradiance on a North-oriented tilted surface according to latitude (18°) were calculated from the meteorological data of the station at Broome airport obtained from [36]. The calculated annual energy output of the solar PV plant at Broome is 26.5 GWh. 3. Results and discussion

Fig. 6. Exemplary behaviour of sea level (Hsea) where neap and spring tides are observed at Broome, Australia (hour of the year in the x-axis).

3.1. Modeling procedure and some exemplary cases

sum of its constituents which is sinusoidal in nature because the periodic nature of the tidal phenomenon. In order to maintain the use of measured tidal data, a spline interpolation has been applied to the raw data to generate tidal level data with a step of < 1 h (0.1 h in this work). In this way, the simulation will be more precise and the evolution of the plant will be analyzed in a more reliable way for a given operation strategy (see Fig. 7). Besides, data downloaded required minor corrections as follows. The lack of data from 12:00 h of 29/01/18 to 4:00 h of 31/01/18 makes necessary to copy analogous data corresponding to 12:00 h until 23:00 h of 28/01 and data from 0:00 h to 4:00 h of 01/02/18. They

With the information about the tidal and photovoltaic plants models and the tidal and solar resources described in previous sections, the procedure to generate the results is shown in the Fig. 9. In particular, the time series for normal direct solar irradiance, horizontal diffuse irradiance, dry bulb temperature and wind speed at a height of 10 m are required. This information, together with the specifications of the solar PV plant in Table 4 allows to generate its power output values. Similarly, starting from the time series of sea level and the specifications of the tidal plant elements and the tidal range turbine to be used, the time series of power generated by the tidal range power plant with the described 0D model are computed. Finally, the total power output of the hybrid system will be the direct sum, for each instant of time, of the 8

Desalination 477 (2020) 114247

A.M. Delgado-Torres, et al.

Fig. 8. Sea level (Hsea) at Broome throughout the year 2018. Table 4 Annual values calculated form the weather data of Broome, PV system parameters required by the PVWatts model and computed annual energy output. Annual global horizontal irradiation Annual direct normal irradiation Annual diffuse horizontal irradiation System nameplate size Module type DC to AC ratio Inverter efficiency Array type and tilt Azimuth Total system losses Annual energy output

INPUT Tidal range system: ( ), Abasin, Asluice Tidal turbine specifications ( ), ( ) Solar resource: , ( ), , ( ), PV system specifications

2292.2 kWh/m2 2544 kWh/m2 554.8 kWh/m2 15 MW Standard 1.2 96% Fixed open rack, 18° 0° (North oriented) 14.08% 26,476.3 MWh

( ), , ( ) ( ), ( ) PV system specifications

( ), Abasin, Asluice, tidal turbine specifications

,

powers of the two plants separately. The preliminary economic analysis is not performed within this procedure because it was considered that the energy analysis alone is first mandatory due to the novelty of the hybrid configuration. Next figure depicts the results about the power generation of the hybrid tidal/PV system as a function of time (in Local Standard Time, LST) for several days of December: behavior of the sea and basin levels (Hsea, Hbasin); the energy resources namely, available head (H) and solar resource at the tilted surface of the PV panels (GPV); the respective power output of the two different energy plants (Pe,tidal, Pe,PV) and the total power generated by the hybrid system (Pe,total). According to Section 2.2, the generation mode takes place from the starting height (Hst) and remains until the available head H decreases up to the minimum height (Hmin). Fig. 10 describes how the power generation of the tidal range power plant decreases as the available head decreases, attributable to the combined behavior of sea and basin levels. The most favorable days are shown on the left of the diagrams and unfavorable days on the right. Consequently, the chart on the top describes the evolution of a decreasing tidal range related with spring to neap tides transition. The third diagram quantifies the suitability of the proposed innovative desalination systems based on hybrid tidal range/solar PV systems since tidal power output can be available when solar resource is low or even nil. For example, in the first day shown (December 9) the solar PV plant could generate a power output equal to or > 8 MW for 5.4 h while the tidal plant could do so for 7.9 h. However, the hybrid plant would be able to do so for 16.2 h without interruption. For the last day of the series (December 16) the photovoltaic plant would be able to produce a power equal to or > 8 MW for 5 h and the tidal plant could never reach that power output level. However, the hybrid plant could supply that power for 6 h continuously. From the last chart, it is concluded that a total power output above 10 MW is not significant. Finally, it is remarkable that fresh water production is possible every day of the time range selected.

MODEL of the PV plant: PVWatts model with SAM

,

0D MODEL of the tidal range power plant in ebb – mode

( )

,

,

( ),

,

Tidal turbine model

( )

( )

Fig. 9. Description of the modeling procedure of hybrid tidal range/solar PV system.

In addition, results power generation in selected days of March, June and September are shown in Figs. 11–13 as a summary of representative results. All of them exhibit aforementioned behavior. Figs. 10–13 prove that the tidal range operation complements the solar PV plant production plant notably expanding the water production period throughout the year in off-grid desalination systems. This behavior is observed in all seasons but in said figures is clearly observed in the first days of the series. In that days the solar PV production is centered between the two generation periods of the tidal range plant. Besides, there are intervals in which both, tidal range and PV plants generate power at the same time. 9

Desalination 477 (2020) 114247

A.M. Delgado-Torres, et al.

Fig. 10. Full description of hybrid tidal range/solar PV system: Level of the sea and of the basin; tidal and solar resources represented by available head and solar irradiance on a North-oriented titled surface (18°); Power production of both 10 MW - tidal range and 15 MW peak - PV plants, and total power production in December at Broome.

Fig. 11. Power production of both 10 MW - tidal range and 15 MW peak PV plants and total power production in March at Broome.

10

Desalination 477 (2020) 114247

A.M. Delgado-Torres, et al.

Fig. 12. Power production of both 10 MW - tidal range and 15 MW peak PV plants and total power production in June at Broome.

Fig. 13. Power production of both 10 MW - tidal range and 15 MW peak PV plants and total power production in September at Broome.

3.2. Yearly results based on the power availability factor, τ

τPV (Pe∗) =

To quantify the net effect of said complementarity between the two power plants throughout the year the power availability factor τ(Pe∗) is defined. This factor represents the percentage of time during which the power output is equal to or greater than a given value Pe∗. Thus, it is expressed as follows for each of the three possibilities (only tidal, only solar PV and solar PV + tidal):

τtidal (Pe∗) =

∆ttidal (Pe, tidal ≥ Pe∗ )[h] 8760 h

∆tPV (Pe, PV ≥ Pe∗ )[h] 8760 h

τPV + tidal (Pe∗) =

∆tPV + tidal (Pe, total ≥ Pe∗ )[h] 8760 h

(7)

(8)

In equations above Δttidal, ΔtPV and ΔtPV+tidal are the total periods of time throughout the year, expressed in hours, during which the power output is equal to or greater than Pe∗ with each option and Pe, total = (Pe, tidal + Pe, PV). The results of the annual power availability factor τ(Pe∗) for a system with a 10 MW tidal range plant and a 15 MW peak solar photovoltaic plant are shown in Fig. 14. This figure is complemented with

(6)

11

Desalination 477 (2020) 114247

A.M. Delgado-Torres, et al.

and 11 MW of nominal power. Therefore, for the same SEC value the largest plant has a capacity 2.2 times greater than the smallest. In addition, neap and spring tide periods are shown for both winter and summer in order to show the influence of the tide on the production profile of the SWRO plant. As it is to be expected, in any case the 5 MW plant would operate for a longer period of time. This is the predicted result according to the Fig. 11. However, even if the 11 MW plant operates for less time, the total amount of water produced must be computed because, as has been said, its capacity is 2.2 times greater. In addition, for a fixed size of the desalination plant the integration of the tidal range energy may present an additional advantage. Given the predictability of this energy resource it would be possible to determine the increase of freshwater production in spring tide periods with respect to neap tide periods in advance. This variability, being manageable, could facilitate the storage of freshwater to ensure an approximately constant production throughout the entire period of neap-spring-neap tides. According to the previously showed results the suitable selection of the SWRO design and operation should be based on the annual profile of power production. As a guideline, the PV solar field needed in standalone desalination is normally at least twice the power consumption of the desalination plant. In wind-driving desalination, about a value of 2 is recommended for the ratio of wind power installed to the nominal consumption of the desalination plant in favorable location as the Canary Islands [21]. The results in Fig. 14 and Table 5 also allows to analyze related information as the following exemplary cases. Per each MW of tidal generator installed combined to 1.5 MWp of PV system a desalination plant with a specific energy consumption (SEC) of 3.5 kWh/m3 can be operated:

Fig. 14. Power availability factors for a hybrid system 10 MW tidal range and 15 MWp PV at Broome, Australia. Asluice = 100 m2, Abasin = 1.5 km2, Hmin = 1 m and Hst = 2 m. Values for the only solar 15 MWp PV and only tidal 10 MW plants separately are given.

some numerical given in Table 5. The expected decrease of τ with Pe∗ is obtained in the three options depicted. As can be seen, power availability factors of only tidal (τtidal(Pe∗)) and only solar PV (τPV(Pe∗)) plants for Pe∗ between 5 MW and 10 MW are similar. The difference between them is < 2.4 percentage points which is equivalent to < 210 h per year. However, within the same Pe∗ range the hybrid configuration solar PV + tidal range has annual power availability factors τPV+tidal(Pe∗) between 2.0 and 2.6 times the factor of the tidal plant τtidal(Pe∗) and 1.8 and 2.8 times the factor of solar photovoltaic plant τPV(Pe∗). This strong increase indicates an excellent temporal complementarity of both power plants throughout the year being this a very attractive feature for the SWRO application and one of the main results of this work. For values Pe∗ values higher than the maximum power output of the tidal plant the result is also the awaited (τtidal = 0%) because the latter can never produce a power > 10 MW (see Fig. 5). On the other hand, the PV plant is able to provide a power greater than or equal to Pe∗ up to a value somewhat higher than the nominal power of the tidal plant (~11.5 MW). However, also in this Pe∗ range the combination of both solar PV and tidal range power plants can yield the specified power Pe∗ during a certain time interval throughout the year (~11%) because Pe, ∗ ∗ ∗ PV + Pe, tidal > Pe although Pe, tidal < Pe . For Pe = 11 MW (> maximum tidal power output) the hybrid plant is able to amplify 8.4 times the power availability factor of the PV plant. Lastly, even when each plant separately is unable to supply a power equal to Pe∗ at any time of the year the combination of the two is capable of doing so, see the case of Pe∗ = 12 MW for example. To illustrate the annual results explained above in a shorter period of time Figs. 15 to 18 are presented below. They include the operation of the RO plant doing Pe∗ = Pe, SWRO for two different plant sizes: 5 MW

- 30% of the year at full load considering 0.99 MW of nominal power consumption (6789 m3/d). With similar power ratio, [21] calculated 51% for a wind-powered desalination system in an excellent location. - 51% if it consumes 0.6 MW of electricity (4114 m3/d). This corresponds to a ratio of 2.5 between PV generator (1.5 MW) and desalination plant, in which the single PV plant only would achieve a 27% throughout the year. - 55% for 0.5 MW (3428.6 m3/d). With similar power ratio 65% is feasible for a wind-powered desalination [21]. The power availability factors for the hybrid system and the only solar PV case are shown in Fig. 19 in a reduced Pe∗ range (5 to 10 MW). The same graph shows the resulting nominal capacity of the SWRO plant with a consumed nominal power equal to Pe∗. This capacity is shown for three different SEC values of the plant to enlarge the applicability of the results. The clear benefit of the analyzed hybridization is evident in this graph once again. For example, a plant with nominal capacity of 65,000 m3/d and an SEC of 3.5 kWh/m3 could operate at its nominal capacity for 1106 h per year (τPV(Pe∗ = 9.5 MW) = 12.6%) if it were only powered by the 15 MWp solar battery – less PV plant. However, when the 10 MW tidal range plant is incorporated the same SWRO plant is capable of producing at its nominal point a total of 2911 h (τPV+tidal(Pe∗ = 9.5 MW) = 33.2%). If the nominal capacity of

Table 5 Numerical values of the annual power availability factor for a hybrid system 10 MW tidal range and 15 MWp PV plant at Broome, Australia. Asluice = 100 m2, Abasin = 1.5 km2, Hmin = 1 m and Hst = 2 m. Ratios between values of the hybrid system and only tidal or PV cases are also given. Pe⁎ [MW]

5

6

7

8

9

10

11

12

13

14

15

τtidal [%] τPV [%] τPV+tidal [%] (τPV+tidal/τtidal) (τPV+tidal/τPV)

27.7 29.9 54.5 2.0 1.8

24.2 26.7 50.5 2.1 1.9

20.9 23.3 46.4 2.2 2.0

17.6 19.6 41.7 2.4 2.1

14.4 15.2 36.3 2.5 2.4

10.2 9.5 26.9 2.6 2.8

0.0 1.7 14.2 – 8.3

0.0 0.0 9.7 – –

0.0 0.0 6.9 – –

0.0 0.0 4.3 – –

0.0 0.0 2.3 – –

12

Desalination 477 (2020) 114247

A.M. Delgado-Torres, et al.

Fig. 15. Power output of the tidal range, solar PV and tidal range/solar PV plants and operation profile of the desalination system according to the power output of the hybrid plant and the nominal load of the SWRO plant (5 MW and 11 MW). Case: neap tide period in summer at Broome.

Fig. 16. Power output of the tidal range, solar PV and tidal range/solar PV plants and operation profile of the desalination system according to the power output of the hybrid plant and the nominal load of the SWRO plant (5 MW and 11 MW). Case: spring tide period in summer at Broome.

the SWRO plant was 45,000 m3/d, these values would be 2196 and 4256 h per year respectively.

production of each plant computing the numerical integration with the following results: hybrid plant Ee, total = 51.6 GWh, tidal range plant 25.13 GWh and solar PV plant 26.48 GWh. But in addition, the useless energy for a given nominal consumption of the RO plant can be also determined having said time series. If the total power output of the hybrid plant does not reach the nominal level of the RO plant, this energy cannot be consumed under an ON/OFF plant control scheme. Therefore, the yearly useless energy is given by:

3.3. Optimal size of the SWRO system for maximum freshwater production As stated in Sections 2.2 to 2.5 the models used for the determination of the energy production of tidal and solar photovoltaic plants provide the power output profile of both plants as well as the time series for the hybrid plant. This allows to determine the total energy 13

Desalination 477 (2020) 114247

A.M. Delgado-Torres, et al.

Fig. 17. Power output of the tidal range, solar PV and tidal range/solar PV plants and operation profile of the desalination plant according to the power output of the hybrid plant and the nominal load of the SWRO plant (5 MW and 11 MW). Case: neap tide period in winter at Broome.

Fig. 18. Power output of the tidal range, solar PV and tidal range/solar PV plants and operation profile of the desalination plant according to the power output of the hybrid plant and the nominal load of the SWRO plant (5 MW and 11 MW). Case: spring tide period in winter at Broome. 8760

Ee, useless =

0

8760

8760

∫ Pe,useless dt ≅ ∑ Pe,useless,i·Δt i=1

Ee, surplus = (9)

8760

∫ Pe,surplus dt ≅ ∑ Pe,surplus,i·Δt 0

i=1

(10)

where Pe, surplus, i = (Pe, total, i − Pe, SWRO) if Pe, total, i > Pe, SWRO, otherwise Pe, surplus, i = 0 MW. When surplus energy exists or Pe, total, i = Pe, SWRO the RO plant produces at its nominal load so the energy consumed for fresh water production is:

where Pe, useless, i = Pe, total, i if Pe, total, i < Pe, SWRO, otherwise Pe, useless, i = 0 MW. In the same way, if the total power produced is greater than the nominal consumption of the RO plant there is a surplus energy that can also be computed:

14

Desalination 477 (2020) 114247

A.M. Delgado-Torres, et al.

Fig. 19. Power availability factors for a hybrid system 10 MW tidal range and 15 MWp PV at Broome, Australia. Asluice = 100 m2, Abasin = 1.5 km2, Hmin = 1 m and Hst = 2 m. Values for the only solar 15 MWp PV are given. Nominal SWRO plant capacity as a function of nominal power consumption is also shown for three different values of the SEC.

Fig. 21. Investment cost of the SWRO plant per unit of fresh water produced in its entire lifetime.

•

•

• Fig. 20. Yearly energy output of the hybrid solar PV/tidal range energy system as a function of the nominal consumption of the SWRO plant.

Ee, useful = (Ee, total − Ee, useless − Ee, surplus )

(11)

Therefore, the values of Ee,surplus, Ee,useless and Ee,useful depend on the nominal power of the RO plant (Pe,SWRO), which is to be expected from the examples shown in Figs. 15 to 18. Those figures are helpful for the following analysis based on Fig. 20 where the results as a function of the SWRO plant's nominal consumption are shown. As it can be seen:

3.4. Effect on the unitary investment cost of the SWRO plant As it has been demonstrated, there is a high degree of complementarity between the solar PV and the tidal range plants during a high percentage of the year which considerably increases the operation time and hence the fresh water production at nominal capacity. This result implies a decreasing of the unitary investment cost of the SWRO plant since it is inversely proportional to the power availability factor (τ). Indeed, if cISWRO is the investment cost of the SWRO plant per unit of nominal capacity (Qp) then the total investment cost of the plant is CISWRO = cISWRO · Qp. On the other hand, if the time period during which the plant will produce at such capacity is known the total amount of fresh water produced during the plant's lifetime (lt) is Vp = τ · 8760 · lt · Qp. In this expression it is assumed that the yearly performances of tidal and solar PV plants are the same throughout their

• The dependence of the surplus energy is the expected. The higher •

by the RO approaches the nominal power of the tidal range power plant (10 MW) the useless energy increases abruptly because this is the maximum power output of the tidal plant. The sum of surplus and useless energies represents the energy output that cannot be consumed by the RO plant. This unusable energy presents a minimum value as a consequence of the opposite trends of the surplus and useless curves. It also presents a sharply increase at the nominal power of the tidal range plant (10 MW) because in this interval the useless energy becomes the dominant component. Since the sizes of the tidal and solar PV plants are fixed the total energy output is constant (51.6 GWh). For this reason, the difference between this and the unusable energy has a maximum located at the same nominal power value of the RO plant where the minimum unusable energy is found. This point is at 8 MW for the system evaluated in this work. For a given SEC value of the RO plant, since an ON/OFF operating strategy at nominal load has been considered, the curve of total yearly freshwater is directly proportional to the useful energy curve. Therefore, there is a maximum of fresh water production for a RO plant size of 8 MW which implies plant capacities between 50,000 and 60,000 m3/d for SEC values between 3.2 and 3.8 kWh/m3 (see Fig. 19). For nominal loads above the nominal output of the tidal plant the useful energy and hence the fresh water production drops abruptly in accordance with the increase in unusable energy.

the nominal power demanded by the RO plant the lower the time during which the total power output of the hybrid plant is greater than this demand and hence Ee,surplus decreases. This result can be also observed in the examples in Figs. 15 to 18. On the contrary, when the power demanded by the SWRO plant increases, the hybrid plant is not capable of reaching the power level demanded by the SWRO plant for longer periods of time. This explains the increase of Ee,useless with Pe,SWRO. This situation occurs more frequently in periods of neap tides. When the power demanded 15

Desalination 477 (2020) 114247

A.M. Delgado-Torres, et al.

Table 6 Effect of desalination capacity selection considering the hybrid tidal/PV generator. Numerical results derived from the 10 MW nominal tidal – 15 MWp nominal solar PV configuration of the hybrid system at Broome, Australia. Asluice = 100 m2, Abasin = 1.5 km2, Hmin = 1 m and Hst = 2.0 m. SEC = 3.5 kWh/m3.

SWRO nominal power consumption, MW Ratio of PV peak power to nominal tidal power Ratio of nominal tidal power to nominal SWRO power consumption Ratio of PV peak power to nominal SWRO power consumption Nominal capacity per nominal tidal power [(m3/d)/MW] SWRO plant + solar PV + tidal Annual operation of SWRO plant at nominal point [%] with solar PV Annual water production per nominal tidal power [(m3/y)/MW] Capital costs of desalination plant per m3 of fresh water production Same SWRO plant + solar PV only Annual operation of SWRO plant at nominal point [%] with solar PV Annual water production with solar PV only plant [m3/y] Capital costs of desalination plant per m3 of fresh water production -

+ tidal range (hybrid) 1000 €/(m3/d); lifetime: 15 y only 1000 €/(m /d); lifetime: 15 y 3

Case 2

Case 3

Case 4

Case 5

Case 6

5 1.5 2.00 3.00 3429

6 1.5 1.67 2.50 4114

7 1.5 1.43 2.14 4800

8 1.5 1.25 1.88 5486

9 1.5 1.11 1.67 6171

10 1.5 1.00 1.5 6857

54.5 681,771 0.34

50.5 758,914 0.36

46.4 812,100 0.39

41.7 835,543 0.44

36.3 817,509 0.50

26.8 671,886 0.68

29.9 374,114 0.61

26.7 401,057 0.68

23.3 408,460 0.78

19.6 392,869 0.93

15.2 343,311 1.20

9.5 238,314 1.92

4. Conclusions

entire lifetime. According to the above, the unitary investment cost of the RO plant is:

⎛ CISWRO ⎞ = cISWRO ⎜ ⎟ τ ·8760·lt ⎝ Vp ⎠

Case 1

The innovative stand-alone desalination plant analyzed driven by a battery-less solar PV/tidal range system exhibits outstanding prospects for developing medium to large capacity RE-powered desalination as follows. The daily power output profiles show that tidal range production expands the energy production of the hybrid system over the only PV based production, thus maintaining fresh water production at high demand. In fact, annual water production of a stand-alone desalination plant at Broome will be equivalent for a 10 MW tidal generator than for a 15 MWp PV one. Therefore, the water production of the analyzed hybrid system exhibits an excellent profile thorough the year in comparison to other options of RE-desalination technologies. For a fixed capacity and power consumed by SWRO plant, the results on an annual basis show that the incorporation of the tidal range power plant results in an increase in the annual operating time of the SWRO plant at its nominal point between 1.8 and 2.6 times the production time it would have if it were only driven by the PV solar plant. In addition, this amplification depends on the SWRO plant capacity and its specific energy consumption (SEC). This conclusion has been obtained for the case of 10 MW tidal range plant and 15 MWp for the PV solar plant at Broome (Australia), i.e. a power ratio of 1.5. For this case the fresh water production with the desalination plant at nominal point and with a SEC of 3.5 kWh/m3 is maximized with the hybrid system if the SWRO plant's capacity is of 5.5·103 m3/d per MW of nominal power of the tidal range power plant. In this point the hybrid RE-powered desalination system would operate 41.7% of the year whereas the same desalination driven only by the battery less PV plant would operate only 19.6% of the year. Moreover, designs described in the literature as “gradual capacity plants” would increase said percentage along with an optimized control of the hybrid tidal range/solar PV plant as a whole. The effect of the PV plant's size and the gradual capacity operation mode of the RO plant on the performance of the hybrid desalination system will be addressed in future works where the introduction of batteries as electrical storage will be discussed too.

(12)

The dependence of (CISWRO/Vp) on the nominal power consumption of the SWRO plant is shown in Fig. 21 for an investment cost of 1000 €/(m3/d) and a lifetime of 15 years. The results for the proposed hybrid configuration and for the only solar PV case are depicted. The effect of hybridization on the investment cost of the desalination plant per unit of desalinated water when compared with the case in which the power demand was supplied only with the PV plant is clearly appreciated. The decrease of this value for the hybrid case derives from the temporal complementarity of the solar and tidal resources and the lower τ values in the only PV case imply a higher unit investment cost. For Pe,SWRO between 5 MW and 9 MW the unitary cost in the hybrid case is between 44% and 58% below the corresponding to the only solar PV case. In addition, the unitary cost in the hybrid case shows a minor dependence on the size of the RO plant. When Pe,SWRO goes from 5 MW to 9 MW the unitary cost practically doubles (0.61 to 1.20 €/m3) in the only solar PV case whereas an increase of 47% is found in the hybrid case (0.34 to 0.50 €/m3). Fig. 21 also shows the expected sharp rise in unitary cost when Pe,SWRO approaches the nominal value of the tidal plant (10 MW). Table 6 shows some numerical cases showed in Fig. 21. In case of off-grid desalination plants the intervals in which both, tidal range and PV plants generate power at the same time can implies surplus energy that is wasted. Note that the reserve SWRO train normally installed in a desalination plant could operate in order to consume this surplus energy. Otherwise, some auxiliary consumptions as pumping due to permeate distribution or seawater intake could work discontinuously. Moreover, seawater pumping from the sea to the basin could also be an option. In fact, pumping water at flood and/or ebb tide is an option already considered in only tidal control sequences studies. The option of using batteries to store the excess or deficit of energy produced should be analyzed after having evaluated the above options due to their cost and associated energy losses. Only minimum battery capacity should be included to avoid negative effect of solar transient on desalination plant operation. In addition, decisions on whether the desalination plant should be operated in absence of enough tidal power will be based on adequate prediction of cloudy days. Besides that, it should be noted that an optimized control strategy of the tidal plant could improve those results. As an example, the approach presented in [20,37] could be adopted. Optimizing control strategy on both energy and desalination plants along with a thorough cost analysis are required prior to selecting the most suitable desalination plant capacity.

CRediT authorship contribution statement Agustín M. Delgado-Torres:Conceptualization, Methodology, Software, Writing - review & editing, Visualization.Lourdes GarcíaRodríguez:Conceptualization, Writing - original draft, Writing - review & editing, Supervision, Funding acquisition.María Jiménez del Moral:Conceptualization, Software, Methodology, Writing original draft.

16

Desalination 477 (2020) 114247

A.M. Delgado-Torres, et al.

Declaration of competing interest

[15] Z. Li, A. Siddiqi, L.D. Anadon, V. Narayanamurti, Towards sustainability in waterenergy nexus: ocean energy for seawater desalination, Renew. Sust. Energ. Rev. 82 (2018) 3833–3847, https://doi.org/10.1016/j.rser.2017.10.087. [16] J. Leijon, C. Boström, Freshwater production from the motion of ocean waves – a review, Desalination 435 (2018) 161–171, https://doi.org/10.1016/j.desal.2017. 10.049. [17] Y. Zhao, G., Su, X., Cao, Y., Su, J., Liu, Experiments on the hydrodynamic performance of horizontal axis tidal current turbine and desalination of sea water, Int. J. Energy Res. 40 (2016) 600–609. [18] C. Ling, Y. Wang, C. Min, Y. Zhang, Economic evaluation of reverse osmosis desalination system coupled with tidal energy, Front. Energy. 12 (2018) 297–304, https://doi.org/10.1007/s11708-017-0478-2. [19] R. Luna-Rubio, M. Trejo-Perea, D. Vargas-Vázquez, G.J. Ríos-Moreno, Optimal sizing of renewable hybrids energy systems: a review of methodologies, Sol. Energy 86 (2012) 1077–1088, https://doi.org/10.1016/j.solener.2011.10.016. [20] A. Angeloudis, S.C. Kramer, A. Avdis, M.D. Piggott, Optimising tidal range power plant operation, Appl. Energy 212 (2018) 680–690, https://doi.org/10.1016/j. apenergy.2017.12.052. [21] B. Peñate, F. Castellano, A. Bello, L. García-Rodríguez, Assessment of a stand-alone gradual capacity reverse osmosis desalination plant to adapt to wind power availability: a case study, Energy 36 (2011) 4372–4384, https://doi.org/10.1016/j. energy.2011.04.005. [22] B. Peñate, L. García-Rodríguez, Current trends and future prospects in the design of seawater reverse osmosis desalination technology, Desalination 284 (2012) 1–8, https://doi.org/10.1016/j.desal.2011.09.010. [23] DOW, (n.d.). http://www.dowwaterandprocess.com/support_training/design_ tools/rosa.htm (accessed June 1, 2019). [24] M. Wilf, L. Awerbuch, C. Bartels, M. Mickley, G. Pearce, N. Voutchkovv, The Guidebook to Membrane Desalination Technology: Reverse Osmosis, Nanofiltration and Hybrid Systems, Process, Design, Applications and Economics, L’Aquila, It (2007). [25] A.R. Evendenandrew, Sea water reverse osmosis - energy efficiency and recovery, Water Pract. Technol. 10 (2015) 187–195, https://doi.org/10.2166/wpt.2015.023. [26] A. Angeloudis, R. Ahmadian, R.A. Falconer, B. Bockelmann-Evans, Numerical model simulations for optimisation of tidal lagoon schemes, Appl. Energy 165 (2016) 522–536, https://doi.org/10.1016/j.apenergy.2015.12.079. [27] J. Xue, R. Ahmadian, R.A. Falconer, Optimising the operation of tidal range schemes, Energies 12 (2019), https://doi.org/10.3390/en12152870. [28] N. Yates, I. Walkington, R. Burrows, J. Wolf, The energy gains realisable through pumping for tidal range energy schemes, Renew. Energy 58 (2013) 79–84, https:// doi.org/10.1016/j.renene.2013.01.039. [29] A. Angeloudis, R.A. Falconer, Sensitivity of tidal lagoon and barrage hydrodynamic impacts and energy outputs to operational characteristics, Renew. Energy 114 (2017) 337–351, https://doi.org/10.1016/j.renene.2016.08.033. [30] G.A. Aggidis, D.S. Benzon, Operational optimisation of a tidal barrage across the Mersey estuary using 0-D modelling, Ocean Eng. 66 (2013) 69–81, https://doi.org/ 10.1016/j.oceaneng.2013.03.019. [31] R. Burrows, I.A. Walkington, N.C. Yates, T.S. Hedges, J. Wolf, J. Holt, The tidal range energy potential of the West Coast of the United Kingdom, Appl. Ocean Res. 31 (2009) 229–238, https://doi.org/10.1016/j.apor.2009.10.002. [32] G.A. Aggidis, O. Feather, Tidal range turbines and generation on the Solway Firth, Renew. Energy 43 (2012) 9–17, https://doi.org/10.1016/j.renene.2011.11.045. [33] S. Petley, G. Aggidis, Swansea Bay tidal lagoon annual energy estimation, Ocean Eng. 111 (2016) 348–357, https://doi.org/10.1016/j.oceaneng.2015.11.022. [34] Australian Baseline Sea Level Monitoring Project. Bureau of Meteorology, (n.d.). http://www.bom.gov.au/oceanography/projects/abslmp/data/index.shtml (accessed October 7, 2019). [35] NREL, (n.d.). https://sam.nrel.gov/ (accessed June 1, 2019). [36] Energy Plus, (n.d.). https://energyplus.net/weather (accessed July 23, 2019). [37] F. Harcourt, A. Angeloudis, M.D. Piggott, Utilising the flexible generation potential of tidal range power plants to optimise economic value, Appl. Energy 237 (2019) 873–884, https://doi.org/10.1016/j.apenergy.2018.12.091.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements L. García-Rodríguez wishes to thank the European Regional Development Fund, Interreg Atlantic Area, for its financial assistance within the framework of the EERES4WATER Project (Second Call, Priority 2, EAPA_1058/2018). References [1] E. Jones, M. Qadir, M.T.H. van Vliet, V. Smakhtin, S. mu Kang, The state of desalination and brine production: a global outlook, Sci. Total Environ. 657 (2019) 1343–1356, https://doi.org/10.1016/j.scitotenv.2018.12.076. [2] M. Heihsel, M. Lenzen, A. Malik, A. Geschke, The carbon footprint of desalination: an input-output analysis of seawater reverse osmosis desalination in Australia for 2005–2015, Desalination. 454 (2019) 71–81, https://doi.org/10.1016/j.desal. 2018.12.008. [3] M.A. Abdelkareem, M. El Haj Assad, E.T. Sayed, B. Soudan, Recent progress in the use of renewable energy sources to power water desalination plants, Desalination 435 (2018) 97–113, https://doi.org/10.1016/j.desal.2017.11.018. [4] L.F. Greenlee, D.F. Lawler, B.D. Freeman, B. Marrot, P. Moulin, Reverse osmosis desalination: water sources, technology, and today’s challenges, Water Res. 43 (2009) 2317–2348, https://doi.org/10.1016/j.watres.2009.03.010. [5] M. Qasim, M. Badrelzaman, N.N. Darwish, N.A. Darwish, N. Hilal, Reverse osmosis desalination: a state-of-the-art review, Desalination 459 (2019) 59–104, https:// doi.org/10.1016/j.desal.2019.02.008. [6] C.S. Karavas, K.G. Arvanitis, G. Papadakis, Optimal technical and economic configuration of photovoltaic powered reverse osmosis desalination systems operating in autonomous mode, Desalination 466 (2019) 97–106, https://doi.org/10.1016/j. desal.2019.05.007. [7] C. Ghenai, A. Merabet, T. Salameh, E.C. Pigem, Grid-tied and stand-alone hybrid solar power system for desalination plant, Desalination 435 (2018) 172–180, https://doi.org/10.1016/j.desal.2017.10.044. [8] M. Melikoglu, Current status and future of ocean energy sources: a global review, Ocean Eng. 148 (2018) 563–573, https://doi.org/10.1016/j.oceaneng.2017.11. 045. [9] F. O. Rourke, F. Boyle, A. Reynolds, Tidal energy update 2009, Appl. Energy 87 (2010) 398–409, https://doi.org/10.1016/j.apenergy.2009.08.014. [10] S.P. Neill, A. Angeloudis, P.E. Robins, I. Walkington, S.L. Ward, I. Masters, M.J. Lewis, M. Piano, A. Avdis, M.D. Piggott, G. Aggidis, P. Evans, T.A.A. Adcock, A. Židonis, R. Ahmadian, R. Falconer, Tidal range energy resource and optimization – past perspectives and future challenges, Renew. Energy 127 (2018) 763–778, https://doi.org/10.1016/j.renene.2018.05.007. [11] S. Waters, G. Aggidis, Tidal range technologies and state of the art in review, Renew. Sust. Energ. Rev. 59 (2016) 514–529, https://doi.org/10.1016/j.rser.2015. 12.347. [12] V. Belessiotis, E. Delyannis, The history of renewable energies for water desalination, Desalination 128 (2000) 147–159, https://doi.org/10.1016/S0011-9164(00) 00030-8. [13] V. Belessiotis, E. Delyannis, Water shortage and renewable energies (RE) desalination - possible technological applications, Desalination 139 (2001) 133–138, https://doi.org/10.1016/S0011-9164(01)00302-2. [14] L. García-Rodríguez, Renewable energy applications in desalination: state of the art, Sol. Energy 75 (2003) 381–393, https://doi.org/10.1016/j.solener.2003.08.005.

17