Integrated operation of renewable energy sources and water resources

Energy Conversion and Management 160 (2018) 439–454

Contents lists available at ScienceDirect

Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman

Integrated operation of renewable energy sources and water resources Yu-Ching Tsai a b

a,b

b

b

, Yea-Kuang Chan , Fu-Kuang Ko , Jing-Tang Yang

a,⁎

T

Department of Mechanical Engineering, National Taiwan University, No. 1, Sec. 4, Roosevelt Rd., Taipei 10617, Taiwan Institute of Nuclear Energy Research, No. 1000, Wenhua Rd., Jiaan Village, Longtan Township, Taoyuan County 32546, Taiwan

A R T I C L E I N F O

A B S T R A C T

Keywords: Renewable energy Hydropower Desalination Energy model

To respond to the global climate change, Taiwan has announced an ambitious target for the development of renewable energy, which is characterized by solar power 20 GW and wind power 4.2 GW by 2025, but the intermittency of renewable energy sources might have serious impacts on the existing power grid. Not only the energy system but also water resources will be impacted by the global climate change. In Taiwan, the strength of rainfall increases but the frequency of rain decreases; this factor combined with a disadvantageous topography to store rainfall worsens the water-shortage issues. As a solution of the aforementioned issues related to the renewable energy sources and water resources concurrently, an integrated system and its operating model for renewable energy sources and water resources are proposed according to which hydropower, pumped-storage hydropower, solar power, wind power, desalination plants and the conjunctive use of water between two reservoirs are considered. A mathematical model is established to describe how the system works under various input data. The results show that, with a retrofit of existing old units and the addition of 102-MW new units, the hydropower unit of the proposed system can eliminate a requirement of 853-MW gas-fired power plants during peak loading in the reference case; the cost, US$45 million per year, of power generation can be saved. With 1099-MW pumped-storage hydropower units added, the proposed system and its operating model further enhance the peak-loading support; relative to a battery-storage system in the reference case, the cost of energy storage can save US$166 million per year. As for the desalination plants in the proposed system, the cost of producing water still exceeds that of the planned reservoir in the reference case because of its greater cost of operation. On considering the total benefit from the water and energy sector, the extra expense, US$41 million per year, for desalination can, however, be readily compensated; the proposed system can save more, US$171 million per year, than the reference case.

1. Introduction In a context of a global climate change, renewable energy has been greatly promoted all over the world. In 2016, the total installed capacity of the existing power generating system in Taiwan was 49.9 GW; the total installed capacity of wind power and solar power was less than 2 GW. The plan announced by the Taiwan government to develop renewable energy by 2025 is ambitious; the capacity setting of solar power is 20 GW, wind power 4.2 GW, hydropower 2.15 GW, biomass 0.813 GW and geothermal 0.2 GW; their total power generation will share 20% of the power demand in Taiwan by 2025. With their rapid deployment, the unit costs of renewable energy sources (RES) are decreasing, but the issues to overcome the intermittent nature of RES are becoming serious. When the penetration of RES attains 5–10%, the impact of the intermittency on the power grid is no more ignorable [1]. The global climate change affects not only energy policy but also water issues. Taking Taiwan as an example, the topography of Taiwan is ⁎

precipitous; the rivers are short and rapid; the sedimentation of existing reservoirs is serious because of violent typhoons every year. With the increased strength and the decreased frequency of rainfall resulting from the global climate change, the reserve of water becomes increasingly difficult. This consequence worsens the problem of water shortage in recent years despite the annual rainfall being sufficient for the annual demand in Taiwan. Power generation is invariably accompanied by water consumption, which might occur during the fabrication of equipment or during power generation. Several researchers have contributed their efforts to the amount and cost of water consumption for power generation [2]. In this work, this issue is not addressed, but the problems of supply and demand of water and power are explored. Without considering hydropower, the issues of supply and demand for the energy and water sectors were formerly solved independently, but, as these two sectors can be combined to extract benefit greater than working them independently, there is increasing research on trying to capture the nexus

Corresponding author. E-mail address: [email protected] (J.-T. Yang).

https://doi.org/10.1016/j.enconman.2018.01.062 Received 16 July 2017; Received in revised form 17 January 2018; Accepted 25 January 2018 0196-8904/ © 2018 Elsevier Ltd. All rights reserved.

Energy Conversion and Management 160 (2018) 439–454

Y.-C. Tsai et al.

Nomenclature RES Sr,i Ed,hihi Ehydro,PSH

AEBS AEPSH APhydro

annual power generation of a battery system, kWh/year annual power generation of PSH, kWh/year annual power generation during the period of peak-power demand by additional hydro units, kWh/year AShydro required support of annual power generation from existing NGCC plants because of operation of hydropower units of ISRWR, kWh/year ASPSH required support for annual power generation from existing natural gas combined-cycle power plants because of the operation of PSH of ISRWR, kWh/year AWr annual production of water in the new reservoir, Mm3/ year CostIS additional annual cost required by ISRWR for water supply, power storage and power generating services, US$ CostR additional annual cost required by the Case Ref. for water supply, power storage and power generating services, US$ CBES installed capacity of battery energy-storage system, MW Cdes desalination unit capacity, Mm3/h (million cubic meter per hour) Chydro, i proposed hydro-turbine capacity of unit i, MW Chydro, PSH proposed pumped-storage hydroelectric capacity, MW Ebalance power demand after subtracting wind and solar power, MW Ed power demand, MW PSH pumped-storage hydroelectricity Qdes water supplied by the desalination plant, Mm3/h Qdome,i domestic water supply by reservoir i, Mm3/h Qeco outflow for ecological consideration, Mm3/h Qflush,i flushing flow rate from reservoir i, Mm3/h Qhydro,i water through hydro-turbine of unit i, Mm3/h Qhydro,max maximal flow for hydropower under a specified water level, Mm3/h Qhydro,PSH hydro flow rate of pumped-storage hydroelectricity, Mm3/ h Qpump pumping flow rate of pumped-storage hydroelectricity, Mm3/h Qagri,i the agricultural water supply by reservoir i, Mm3/h Qbalance balance of inflow and outflow in Shigang reservoir, Mm3/ h Qd water demand, Mm3/h Qpump,max maximum rate of pumping flow of pumped-storage hydroelectricity, Mm3/h Qtri,i inflow from tributaries upstream of reservoir i, Mm3/h Qunion positive for water support from Shigang to Liyutan, and negative for water support from Liyutan to Shigang, Mm3/ h RCC electricity price of existing natural gas combined-cycle

ELlolo ELr,i Epump ELhi ELhihi ISRWR LCBS LCdes LChydro LCPSH LCr Mm3 Sr,hihi Svirtual TPSH TBES t XPSH Xhydro ϕdes ϕhydro,i ϕhydro,PSH ϕpump

power plants, US$/kWh renewable energy resources water amount of reservoir i, Mm3 upper limit of power demand, MW hydro power generated by pumped-storage hydroelectricity, MW below this limit of water level, hydropower is prohibited, m water level of reservoir i, m pumping power of pumped-storage hydroelectricity, MW above this elevation, hydropower is fully operated, m above this elevation, flushing is required, m integrated system for renewable energy sources and water resources levelized cost of power production for a battery storage system, US$/kWh levelized cost of water production for desalination plant, US$/m3 levelized cost of power production for an additional hydropower unit, US$/kWh levelized cost of power production for PSH, US$/kWh levelized cost of water production for new reservoir, US $/m3 million cubic meter amount of water in Techi reservoir for water level at ELhihi, Mm3 virtual amount of water storage of Liyutan reservoir, Mm3 duration of pumping, h duration of BES discharge, h time, h output ratio of PSH, (0–1, dimensionless) output ratio of hydropower units, (0–1, dimensionless) specific energy consumption of desalination, kWh m−3 proposed hydro-turbine efficiency at design point for unit i, % proposed hydro-turbine efficiency at design point for PSH, % proposed pumping efficiency of PSH,%

Subscripts i

j k land off solar

1 through 8 denote units or reservoirs for Techi, ChinShan, Guguan, Tianlun, Maan, Shigang, Liyutan and TienHwa-Hu reservoir, respectively the upstream reservoir of reservoir i, j = i − 1 reservoir in which hydropower unit i discharges water land-based wind power offshore wind power solar power

power plants with PSH and a high penetration of wind power in Ireland. Chen et al. [6] proposed a mathematical model to maximize the utilization of RES and to minimize the use of diesel generators in an island with the help of PSH. Portero et al. [7] demonstrated a combination of seawater PSH with wind power, which might decrease the cost of power generation. Ma et al. [8] studied how to determine an optimal combination among wind power, solar power and PSH. In a remote area or island or Middle-east country, some researchers are trying to integrate renewable energy with desalination techniques to solve the scarcity of power and water. Georgiou et al. [9] used multicriteria analyses to evaluate the economic and environmental issues of a small-scale desalination plant with power sources in varied combinations. Mentis et al. [10] tried to supply the entire water demand of arid islands in the South Aegean Sea with desalination plants; they

between the RES and water resources. The most typical way to integrate the water resources and RES is through pumped-storage hydro (PSH), according to which the RES become stored within an upper reservoir of PHS during periods of excess supply and released for subsequent use in periods of high demand. There are more than 20 GW and 7.4 GW of PSH at the planning stage in USA and EU, respectively [3]. Many researchers are exploring how to use PSH to harness the excess wind power in an economic manner. For example, Anagnostopoulos et al. [4] used a simulation model to evaluate the critical factors for a PSH to recover the wind energy rejection in the power system of Greece; the results showed that the installed capacity of wind power, the available capacity of the reservoir and the operating strategies of the hydro-turbine are the key factors. Tuohy et al. [5] performed an economic analysis on replacing some gas-fired 440

Energy Conversion and Management 160 (2018) 439–454

Y.-C. Tsai et al.

which purpose a mathematical model is established to simulate its operation. Extending the previous study [20], several projects under the consideration of the Taiwan government, such as solar power with capacity 20 GW, a battery energy-storage system, gas-fired power plants, a new PSH embedded within the existing hydropower system and a mechanism for mutual support between reservoirs, are compared in this study to evaluate their economic feasibility. To accommodate the aforementioned system changes, the operational strategies used previously become greatly modified to fulfill an expectedly greater rate of penetration of RES. In the previous study, hourly data between 2011 and 2014 were defined into four cases. Although a similar approach of case study is adopted here, the data in 2015, which were characterized by a serious drought, are included in this research. To examine the practicality of ISRWR better, in this study appear fewer assumptions about the input data and system operation. The potential benefits are also evaluated through an economic comparison between ISRWR and the reference case. The following chapters are arranged to describe the method of constructing ISRWR (Section 2), the mathematics of the operating model through a case study of Taiwan (Section 3), and the results of the analysis (Section 4), before the conclusions (Section 5).

concluded that the least cost of water production was achievable with wind power as the power source. Ismail et al. [11] conducted a parametric study of a desalination system to determine the parameters that had the most effect on the performance; the economic feasibility of various power sources, such as RES and gaseous fossil fuel, were also compared. In the thermoeconomic assessment of a solar polygeneration plant by Leiva-Illanes et al. [12], concentrated solar power and a desalination plant were considered; this work demonstrated that such a polygeneration plant is more cost-effective than stand-alone plants. A more complicated linkage between the energy and water sectors can involve the integration of RES, a PSH, a desalination plant and other technologies. An application of this kind is intended to seek decreased cost and increased stability of water production, or to enhance the penetration of variable renewable energy. Spyrou et al. [13] and Segurado et al. [14] studied systems that consisted of renewable energies, PSH and desalination plants on a remote island; they demonstrated that the integration of these facilities can decrease the cost of power and water production. Novosel et al. [15] and Perković et al. [16] undertook studies of the integration of RES, brine pumped storage and a desalination plant for Jordan; their studies were characterized by detailed descriptions of mechanisms of power and water supply; the results showed that the introduction of a desalination plant is beneficial to the utilization of excess RES. Some researchers considered desalination plants as part of their system of energy management or plan for energy development. Al-Nory et al. [17] utilized a desalination plant as a medium of demand response to achieve an improved management of the intermittency of renewable energies. Saif et al. [18] addressed an expansion plan of desalination plants on considering the problems of a power-supply chain and the carbon tax in the Emirate Abu Dhabi, in which nuclear power, solar power, cogeneration plants and gas-turbine plants are considered in various scenarios. Most aforementioned research connected renewable energy and water supply through a desalination plant or a PSH. As the applications of these studies are mainly for those areas that have an extreme lack of water resources, an existing pattern of rainfall and reservoir operation, which can greatly affect the strategies for the system operation, are typically excluded from their consideration. Besides, most studies treat only the short- term intermittency of the renewable energy, which can endure several hours or days. Even for those studies in which the seasonal variation of RES was considered, the manner of the solution relied on hydrogen technology [19]. A previous study [20] established an operating model to simulate a system comprising an energy system and a water system; this model is characterized by the integration of traditional hydropower, offshore wind power, a desalination plant and traditional power plants, including consideration of the patterns of rainfall and water demand and the details of reservoir operation. The basic idea of this previous work was to enhance the capability of sharing peak power of existing hydropower during rainy seasons; a desalination plant then compensated for the small rate of water storage of a reservoir after that operation during dry seasons. As the new plan to develop renewable energy in Taiwan announced in 2016 features a great capacity of solar power, the profile of power supply could alter greatly. For example, if from the actual profile of power demand captured in 2011 is subtracted the simulated solar and wind power with capacities 20 and 4.2 GW, respectively, the peak power during summer shifts from 14 h to 20 h; the valley of power demand shifts from 6 h to 13 h, as Fig. 1 shows. This new profile of power demand would decrease the effectiveness of the previous strategies for the sharing of peak power. The new profile shows also that, even with a large capacity of RES, a decreased maximum output from traditional power plans in a day seems not obvious. Moreover, the mountain of the power output curve becomes sharper and the difference between mountain and valley becomes larger, which increases the stress on the dispatch of traditional fossil power plants. An integrated system for RES and water resources (ISRWR) is proposed to tackle the aforementioned issues related to RES and water, for

2. Methods The concept of ISRWR is not only applicable to Taiwan, actually, it can be widely applied to those countries which have hydropower systems and renewable power plants. Fig. 2 shows the method for the constructing and analysis of ISRWR. With the development target of RES, a new profile of power demand can be simulated by using the historical data and the simulated power curves of RES, which is demonstrated in Section 3.4. Within this new profile of power demand, the peak-loading power demand is the target that ISRWR tries to handle. To construct the allocation of related facilities for ISRWR, the existing hydropower system and reservoirs are the necessary parts, and a PSH system or other planned water facilities are optional. As the existing hydropower system and reservoirs may not be specially operated for the peak-loading period, a new operating model needs to be designed for this purpose, and the historical rainfall patterns shall be applied to check whether there will be a water shortage problem under the new operating model. If there is no water shortage problem, the benefit analysis for the new operating model can be performed. The analysis result of benefit shall be compared with the alternative system, which can provide the same services as the ISRWR constructed above, to determine whether the ISRWR is a superior solution. However, if the water shortage occurs under the new operating model, either the enhancement of the system, such as the addition of desalination plant and water storage facility, or the enhancement of the operating strategies shall be applied to solve this issue. At this stage, the modification of operating model is inevitable, and the new patterns of water supply are generated and then compared with the historical water demand again. This process will be iterated until the water shortage problem is solved, and the final result will then be examined through the benefit analysis. The guarantees of the peak-loading sharing capability and no water

Fig. 1. Profile of actual power demand after subtraction of simulated solar power and wind power.

441

Energy Conversion and Management 160 (2018) 439–454

Y.-C. Tsai et al.

Fig. 2. The methods for the establishment of ISRWR.

3.1. Power and water systems in Taiwan

shortage problem shall be always kept in mind through the whole process of evaluation. Besides, if the extreme patterns of historical rainfall are considered, the result of ISRWR can be more practical under the challenges of climate changes. Among the processes shown in Fig. 2, the most difficult one is the construction of the operating model. The system configuration of ISRWR shows only the physical linkages between facilities, however, the evaluation of the feasibility of ISRWR is impossible without an operating model. In this study, a case study in Taiwan is performed to give a detail picture of ISRWR, which helps to demonstrate the practicability of ISRWR.

In Taiwan, the installed capacity of power generating systems is 49.9 GW in total; at the end of 2015 this capacity comprised coal-fired power plants (35.3%), gas-fired power plants (31.7%), nuclear power plants (10.3%), oil-fired power plants (8.0%), PSH (5.2%), conventional hydroelectric power plants (4.2%) and other renewable energies (5.2%) [21]. In the same year, the installed capacities of solar power and wind power were 1210 MW and 682 MW, which are proposed to be increased to 20 GW and 4.2 GW, respectively, by the end of 2025. Among the existing conventional hydropower plants (capacity 2089 MW), more than half (1148 MW) are located in the Dajia river. These hydro-turbine units named Techi, Chinshan, Guguan, Tianlun and Maan units, respectively, are distributed along the Dajia river in a sequence and are located between the Techi and Shigang reservoirs (Fig. 3). These units are the main components of the target system of this study. The topography of rivers in Taiwan is characterized by short distance and steep river course; the rainfall is generally concentrated during the summer season. These factors combined make it difficult to preserve the water resources; problems of water shortage have occurred several times even though the annual rainfall in Taiwan is theoretically sufficient for the water demand. As the reservoirs along the Dajia river are considered in ISRWR, the water demand that it serves, which

3. A case study of Taiwan In the following subsections, the existing power and water systems in Taiwan are introduced, and the geographic location in which ISRWR are situated is also addressed. After accumulating a full picture of the problem formulation, the system configurations of ISRWR, the determination of reference case, the data used for analysis, the strategies of operation, the mathematical model, and the way to perform the benefit analysis are explained in sequence.

442

Energy Conversion and Management 160 (2018) 439–454

Y.-C. Tsai et al.

study with some modifications from the existing plan, such as that the total capacity is increased to 1099 MW and the duration of water pumping is decreased to 4 h. The desalination plant can provide 0.013 million cubic meters of water per hour (Mm3/h) according to a specific rate of energy consumption 3.3 kWh m−3. The capacities of solar and wind power are 20 GW and 4.2 GW, respectively; to the wind power is further contributed the land-based wind power (1.2 GW) and offshore wind power (3 GW). A reference case is established to evaluate the potential benefits of ISRWR; its specification is listed in Table 1, with a detailed description in Section 3.3. 3.3. Reference case The proposed system and its operating model attempt to provide an improved system for the supply of water and power. For the purpose of comparison, several alternatives that are also under consideration by the government are depicted in the reference case (denoted as Case Ref.), as shown in Table 1. The existing power and water systems are apparently unable to overcome the foreseen challenges. To solve the power and water issues addressed in this paper, the government has already evaluated some plans to manage them. Those plans that are highly likely to be implemented by the government are set to the reference case in this study, and they include a new reservoir, the addition of gas-fired power plants and the battery-storage system. The system configurations in ISRWR, such as a desalination plant, a PSH and an increased capacity of hydropower units, might not be a primary option or might be in a competitive status with those selected in the reference case. Since the water transport mechanism between Shigang and Liyutan reservoir is going to be constructed, thus it is considered in the reference case and ISRWR. The reference system includes all reservoirs in Fig. 4; the mechanism of water dispatch between Shigang and Liyutan reservoirs is applied also in the reference case. The retrofits of the Techi and Tianlun units are still applied to the reference case, but the other units are maintained at their current capacities. A new reservoir is required in the reference case to replace the desalination plant for an auxiliary water supply. The specifications of this new reservoir refer to the project “Tien Hwa Hu Reservoir” [23]; its capability of water supply is 94.9 Mm3 per year. Although this reservoir is under planning in Taiwan, the final decision of building is pending because of serious protest from local people. To confront the challenges arising from an intermittent RES, the Taiwan government plans to improve the existing power grid. Although not yet having an exact target, a battery-storage system has been considered as a possible solution. A battery-storage system is assigned in the reference case to replace the character of PSH for the energy storage of ISRWR. As ISRWR enhances the capability of peak-power sharing using hydroelectric power units, the reference case must build new fossil power plants to attain the same effect. As there is no plan to build new gas-turbine power plants to tackle peak-power demand in Taiwan, the natural-gas combined-cycle (NGCC) units are selected to accept this responsibility in the reference case. To make comparable the analysis results from ISRWR and the reference case, both systems are expanded to the same capability of power supply and water supply, which is similar to the approach used by Tsai et al. [20]. As most hydroelectric power is released during the peak-power demand of period T2, the ISRWR requires the existing NGCC to make up the power supply during the non-peak-power demand period and the power loss due to the smaller charging efficiency of PSH relative to the battery system. In contrast, the reference case requires the new NGCC units to provide the capability of peak-power supply the same as in the ISRWR. From the history of peak-power generating data, the power generation of all hydroelectric power units in Taiwan might have been as little as 389 MW in 2011; this value is selected as the minimum ensured power output of hydroelectric power units in the reference case. In the watersupply sector, the desalination plant of the ISRWR must supply

Fig. 3. Locations of rivers and reservoirs in ISRWR.

include the domestic and agricultural water demand of 2.8 million people in Taichung city, is also considered. 3.2. System configuration of ISRWR Fig. 4 shows the main components of the ISRWR, described as follows. Existing reservoirs named Techi, Chinshan, Guguan, Tianlun and Maan in a sequence are considered, each of which is equipped with hydroelectric power units. A proposed PSH that uses Techi and Guguan reservoirs as the upper and lower reservoirs, respectively, is integrated within ISRWR; this special feature differs from the content of the previous study. A downstream reservoir called Shigang serves to store the water released from Maan reservoir; the stored water is subsequently released for the domestic and agriculture demands of Taichung city. As the water shortage problem will occur when the existing hydroelectric power units are fully operated during the peak-loading period in summer, which was demonstrated in the previous study [20], a desalination plant is still required in this study to satisfy part of the aforementioned water demand. The other features specially embedded within ISRWR include a nearby reservoir called Liyutan (Fig. 3), which partially satisfies the domestic water demand of Taichung city, two plants for water treatment, which are capable of receiving water from Shigang reservoir and Liyutan reservoir, and solar power and wind power, which account for the entire capacity expected to be installed in Taiwan by 2025. All power generated by solar power, wind power, hydropower and traditional gas-fired power plants is fed into the power grid; the power demand of all infrastructures is satisfied by the power grid. In summary, the additions to the existing power and water systems include a PSH, a larger capacity of solar power, a larger capacity of wind power, a desalination plant and a mechanism of water transport between Shigang and Liyutan reservoirs. Table 1 lists the major parameters of the ISRWR. Besides the effective capacity of Shigang reservoir that is assumed to be restored to 1.7 Mm3 according to a desilting plan, the others are determined from data published in 2015 [22]. The existing hydroelectric units of Techi and Tianlun are assumed to be retrofitted because of their ageing; there are in total 102 MW new units added to each hydroelectric power station, which is required according to the operating strategies of ISRWR. To shift part of the solar power generated at noon to the second peakpower demand at night, as shown in Fig. 1, the government is evaluating a plan to build a new PSH. This new PSH is considered in this 443

Energy Conversion and Management 160 (2018) 439–454

Y.-C. Tsai et al.

Fig. 4. System configuration of ISRWR.

94.9 Mm3 water per year even if the actual water demand in the ISRWR is far less than this amount. In the power-storage sector, the battery system in the reference case is assumed to discharge the same power as the PSH in the ISRWR.

important than its absolute value. Although the power demand is expected to increase in 2025, the shape of this curve is not expected to alter greatly according to a review of the historical data. Even though the target of RES in 2025 is simulated, the historical curves for power demand are still appropriate for this study. For the water sector, as ISRWR must satisfy all water demand in Taichung city, the total water demand by 2031, according to the latest prediction by the government, must be reflected in the historical water-demand curves. Although the target of RES by 2031 is not yet announced, the simulation performed in this study suffices to demonstrate the feasibility of ISRWR under a high penetration of RES.

3.4. Input data In this study, the historical data related to power and water sectors are applied to verify the feasibility of an operating model. In the power sector, a related issue that ISRWR tries to solve is the capability of peakpower sharing; the shape of the power- demand curve is far more 444

Energy Conversion and Management 160 (2018) 439–454

Y.-C. Tsai et al.

As data for 2015 are unavailable, the data of power demand during 2011–2014 are used for Case A to Case D; Case E is assumed to use the same power data as Case D. To simulate the profiles of solar power, those profiles in the actual data are divided by their installed capacities and multiplied by 20 GW. For the land-based wind power, the same manner is adopted to simulate the output of capacity 1.2-GW. As only two offshore wind turbines were installed in Taiwan by the end of 2016, no datum for offshore wind power is available yet. The offshore wind data near the coast of Chiayi county by 2014 and the performance curve of SWT-3.6-120 (Siemens, 3.6 MW, rotor diameter 120 m) together are selected to simulate the 3-GW offshore wind power, just as was used in the previous study [20].

Table 1 Specifications of ISRWR and the reference case. Parameter Reservoir Cr,1 Cr,2 Cr,3 Cr,4 Cr,5 Cr,6 Cr,7 Cr,8 Hydroelectric Chydro,1 Chydro,2 Chydro,3 Chydro,4 Chydro,5 ϕhydro,1 ϕhydro,2 ϕhydro,3 ϕhydro,4 ϕhydro,5

Description

ISRWR

Reference case

Effective capacity of Techi reservoir, Mm3 Effective capacity of Chinshan reservoir, Mm3 Effective capacity of Guguan reservoir, Mm3 Effective capacity of Tianlun reservoir, Mm3 Effective capacity of Maan reservoir, Mm3 Effective capacity of Shigang reservoir, Mm3 Effective capacity of Liyutan reservoir, Mm3 Effective capacity of Tien-Hwa-Hu reservoir, Mm3

148.80 0.40

V V

5.12

V

0.27

V

0.17 1.70

V V

118.30

V

None

47.91

unit Techi hydroelectric power unit, MW Chinshan hydroelectric power unit, MW Guguan hydroelectric power unit, MW Tianlun hydroelectric power unit, MW Maan hydroelectric power unit, MW Proposed Techi hydro-turbine efficiency, % Chin-Shan hydro-turbine efficiency, % Guguan hydro-turbine efficiency, % Proposed Tianlum hydro-turbine efficiency, % Maan hydro-turbine efficiency, %

Pumped-storage hydropower Chydro,PSH Maan hydroelectric power unit, MW ϕhydro,psh Proposed PSH hydro-turbine efficiency, % ϕpump Proposed PSH pumping efficiency, % TPSH Duration of pumping, h Desalination plant Cdes The desalination plant’s capacity, Mm3/h ϕdes Specific energy consumption rate, kWh m−3 RES Csolar Cwind

Solar power capacity, MW Wind power capacity, MW

Miscellaneous Water transport mechanism between Shigang and Liyutan reservoirs Capacity of battery energy storage system, CBES MW Duration of BES discharging, h TBES CNGCC Additional capacity for a natural-gas combined-cycle plant, MW

3.5. Strategies of operation

234.0 406.4 261.2 248.5 153.0 90.0

V 368.0 217.6 V 133.5 V

91.0 91.0 87.0

V V V

88.0

V

1099.0 90.0 86.5 4.0

None – – –

0.013 3.3

None –

20,000 4200

V V

Yes

V

None

1099.0

– None

4.0 853.0

The operating model of ISRWR was constructed with MATLAB; Fig. 5 shows the strategies of operation. Several input data are required to determine the corresponding responses of the model; they include system parameters as stated in Table 1, solar power (Esolar), offshore wind power (Ewind,off), land-based wind power (Ewind,land), actual power demand (Ed), domestic water supply from Shigang reservoir (Qdome,6) and from Liyutan reservoir (Qdome,7), agricultural water supply from Shigang reservoir (Qagri,6) and Liyutan reservoir (Qagri,7) and inflow profiles of all reservoirs. The most benefit that can be extracted from the traditional hydropower of the ISRWR is to decrease the requirement of additional natural-gas combined-cycle power plants, which is used for the peak-load during the summer season. The mission of satisfying the peak-load during summer is hence important for the operating model. As Fig. 6 shows, the power-demand profile can be readily divided into three sections – before summer, during summer and after summer – according to its amplitude; they are assigned to be periods T1, T2 and T3 in this study, respectively. From a rough evaluation, T2 is located between 3789 h and 6915 h; two non- peak-load periods are then formed. To extract increased benefit from the system, the duration of T2 should be expanded optimally, but, when the period of T2 is set too long, a water shortage might occur because of the rainfall pattern. The periods are hence defined as T1 (time < 3600 h), T2 (3600 h ≤ time < 7460 h) and T3 (time ≥ 7460 h), respectively, after the iterative process. To incorporate the new infrastructures and the new limitations of operation, six main modes of operation are applied for periods T1 and T3 (Modes 1–6), and six main modes of operation likewise for period T2 (Modes 21–26). The main concept of operation during T1 and T3 is to preserve optimally the water in the reservoir; to share the burden of water supply, the desalination plant is fully operated when a water shortage occurs. The reserved water during periods T1 and T3 is then turned to enhance the peak-power support during period T2. As this study includes a PSH, a dispatch mechanism between reservoirs and a large amount of solar power, more complicated interactions among all infrastructure are inevitable. The following sections address the details of the mathematical models for periods T1 to T3, and evaluate the benefits of the ISRWR in comparison with the reference case.

V: The setting of the reference case is the same as that in ISRWR.

Most input data used in this study result from statistical data published by the government. The inflow data of each reservoir during 2011–2015 are selected and denoted as Cases A to E [24]; rather than an average value of all years, the flow gathered from the tributaries upstream of each reservoir for each year is considered. As shown in Table 2, the inflow patterns selected are characterized as extremely low rainfall (Case A), extremely high rainfall (Case B), normal rainfall (Case C), extremely low rainfall during the last half-year (Case D), and extremely low rainfall during the first half-year (Case E). A continuous pattern from Case A to Case E is suitable to verify the practicality of the proposed model of operation under extreme climate conditions. The profiles of water demand are collected from public data during 2011–2015 [25]. To satisfy the domestic water demand of Taichung city in 2031, the aforementioned data of domestic water demand are all elevated proportionally to attain an annual amount 573.78 Mm3. As agriculture is not expected to expand in Taiwan, the agricultural water demand is assumed to maintain the same as in the actual data.

Table 2 Descriptions of Case A to Case E.

445

Case

Rainfall pattern

Annual rainfall, Mm3

A B C D E

Extremely low rainfall Extremely high rainfall Normal rainfall Extremely low rainfall during the last-half year Extremely low rainfall during the first-half year

1216 3820 3271 2021 2065

Energy Conversion and Management 160 (2018) 439–454

Y.-C. Tsai et al.

Fig. 5. Strategies of operation.

the status of a typhoon, the preliminary balance of water in Shigang reservoir (Qbalance) and the storage rate of water of Shigang reservoir (Sr,6), are used to determine the applicable mode. When ELr,1 exceeds the high-high limit (1407 m), the model further tests the status of a typhoon. If the rate of water inflow into Techi reservoir (Qtri,1) is greater than 2.52 Mm3/h, which is an indicator of the existence of a typhoon, Mode 1 becomes applied; otherwise, Mode 2 is applied. As the desalination plant is not operated in Modes 1 and 2, the flow rate of desalination water is zero (Qdes = 0). When ELr,1 is between the high-high limit (ELhihi) and low limit (ELlolo), Eq. (1) determines the preliminary balance between inflow and outflow in Shigang reservoir,

Fig. 6. Power-demand profile of Taiwan in 2014.

3.6. Mathematical model during periods T1 & T3

Qbalance (t ) = Qtri,6 (t ) + Qhydro,5 (t −9) + Qflush,5 (t −15)−Qdome,6 (t )−Qagri,6 (t )

To capture satisfactorily the operation of major infrastructure, the operating strategies are described in detail for a desalination plant, a PSH unit and a hydroelectric power unit. The supply and demand for water and power sectors are also evaluated to determine the behavior of ISRWR.

(1) in which Qtri,6 is the inflow from tributaries upstream of Shigang reservoir; Qhydro,5(t-9) is the water released through Maan hydroelectric power unit 9 h previously; Qflush,5(t-15) is the flushing flow through Maan reservoir 15 h previously; Qdome,6 and Qagri,6 are the demands for domestic and agricultural water satisfied by the Shigang reservoir, respectively. When ELlolo ≤ ELr,1 < ELhihi and Qbalance ≧ 0, it means that the Techi

3.6.1. Desalination plant As shown in the upper part of Fig. 5, for time (t) within period T1 or T3, four parameters such as the water level of Techi reservoir (ELr,1), 446

Energy Conversion and Management 160 (2018) 439–454

Y.-C. Tsai et al.

inadequate for the expected amount of water released, Eq. (3c) must apply. The calculations of Eqs. (2b) and (3b) require an iterative process, similar to a calculation of the hydropower in the previous study [20]. As the water level of Guguan reservoir is no longer steady, the variation of water levels for upper and lower reservoirs of the PSH must be considered in the aforementioned calculations.

hydroelectric power unit is operable; there is adequate inflow water into the Shigang reservoir. The desalination plant will not be operated when the rate of water storage of Shigang reservoir (Sr,6) exceeds 40% of its effective capacity (Cr,6), which is denoted as Mode 3. When Sr,6 is less than 40% of Cr,6, which is denoted as Mode 4, the desalination plant might be fully operated (Qdes = Cdes) if Qbalance is inadequate to fill the residual volume of Shigang reservoir. When ELlolo ≤ ELr,1 < ELhihi and Qbalance < 0, which is denoted as Mode 5, the inflow to Shigang reservoir is inadequate; the desalination plant must be fully operated (Qdes = Cdes) to prevent a critical decrease of Sr,6. When ELr,1 < ELlolo, which is denoted as Mode 6, it means that the Techi reservoir lacks water; the desalination plant must be fully operated (Qdes = Cdes).

3.6.3. Hydroelectric power unit The key parameter for the hydroelectric power unit is the flow for generation (Qhydro), as shown in Eqs. (6) and (7).

a ⎧0 b Qhydro,i (t ) = Qhydro,max (ELr ,i (t −1),ELr ,k (t −1),Qhydro,design ) ⎨ Xhydro × Qhydro,max (ELr ,i (t −1),ELr ,k (t −1),Qhydro,design) c ⎩

3.6.2. The PSH unit The operation of PSH is described with pumping flow (Qpump) and hydroelectric flow (Qhydro,PSH), in Eq. (2) and Eq. (3), respectively.

⎧0 ⎪Qpump,max (Sr ,1 (t −1),Sr ,3 (t −1),XPSH ) Qpump (t ) =

(6) in which Eq. (6) is for Techi (i = 1) and Guguan (i = 3) hydroelectric units, and the water level is a variable parameter; Qhydro.max is a function used to calculate the maximum flow for hydropower, subject to the specific water levels of reservoir i (ELr,i), the specific water levels of its downstream reservoir (ELr,k) and the designed flow rate of the hydroturbine (Qhydro,design); Xhydro is the ratio of power output as listed in Table 4,

a b

⎨ (Ed,hihi−Ebalance (t ))/ Chydro,PSH × Qpump,max (Sr ,1 (t −1),Sr ,3 (t c ⎪ ⎩ −1),XPSH ) (2)

a ⎧0 b Qhydro,i (t ) = Qhydro,design ⎨ Qtri,i (t ) + Qhydro,j (t −Td1,j ) + Qflush,j (T −Td2,j )−Qflush,i c ⎩

In Eq. (2a), the PSH pumps no water; in Eq. (2b), the maximum pumping flow of PSH is determined by function Qpump,max, which is related to the rate of water storage of Techi reservoir (Sr,1), the rate of water storage of Guguan reservoir (Sr,3) and the ratio of operation (XPSH); in Eq. (2c), the operation of the PSH is further limited by the high-high-limit of power demand (Ed,hihi), the preliminary balance of power (Ebalance) and the capacity of the PSH units (Chydro,PSH).

a ⎧0 Qhydro,PSH (t ) = Qhydro,max (Sr ,1 (t −1),Sr ,3 (t −1),XPSH ) b ⎨ c ⎩ SPSH (t −1)

in which Eq. (7) is for Chinshan (i = 2), Tianlun (i = 4) and Maan (i = 5) units; the water level is a constant value as the inflow is designed to equal the outflow of the reservoir at any instant in the proposed model; Qtri.i is the inflow of reservoir i that derives from its upstream tributaries; Qhydro.j is the tail water released Td1.j hours previously from the upstream hydroelectric power units; Qflush.j is the flushing water released Td2.j hours previously from the upstream reservoir; Qflush.i is the flushing water released from reservoir i. Besides the water released through the hydroelectric turbine, another path of water release for the reservoirs is through flushing, as shown in Fig. 4. Eq. (8) shows how to calculate the required flushing flow,

(3)

According to Eq. (3a), the PSH releases no water; in Eq. (3b), the maximum generating flow of PSH is determined by function Qpump,hydro, which runs the calculation of Eq. (2b) in a reverse way; in Eq. (3c), the released water of the PSH equals the water stored in the virtual reservoir (SPSH), which is always greater than zero to ensure that the water released by PSH is always less than what it pumps. The calculations of the preliminary balance of power (Ebalance) and the water stored in a virtual reservoir (SPSH) mentioned in Eq. (2c) and Eq. (3c) are shown as Eq. (4) and Eq. (5), respectively.

Ebalance (t ) = Ed (t )−Esolar (t )−Ewind,off (t )−Ewind,land (t )

(4)

SPSH (t ) = SPSH (t −1) + Qpump (t )−Qhydro,PSH (t )

(5)

(7)

a ⎧ Qtri,i (t ) + Qpump (t )−Qhydro,1 (t )−Qhydro,PSH (t ) ⎪ ⎪ −[S r ,hihi−Sr ,1 (t −1)] Qflush,i (t ) = ⎨Q (t ) + Q tri,i hydro,j (t −Td1,j ) + Qflush,j (T −Td2,j )−Qhydro,i b ⎪ ⎪Qeco c ⎩

(8)

in which Sr,hihi is the high-high limit of the water storage of Techi reservoir; Qeco is the water flow required for ecological consideration; subscript j represents the upstream reservoir of reservoir i. For Mode 1, all hydroelectric units do not operate because a typhoon is present; Eqs. (6a) and (7a) are applied. At the same time, the Techi reservoir follows Eq. (8a) to maintain a rate of water storage

in which Ed is the actual power demand; Esolar is the simulated solar power; Ewind,off and Ewind,land are the simulated land-based wind power and offshore wind power, respectively. When a typhoon exists (Mode 1) or the water level of Techi reservoir is too low (Mode 6), the PSH does not operate as shown in Eqs. (2a) and (3a). For Modes 2–5, several criteria combine to determine the applicable equation of Qpump. When the residual volume in Techi reservoir is inadequate to accommodate Qpump, or the water stored in Guguan reservoir is less than Qpump, or SPSH attains its high limit, or Ebalance exceeds 30,000 MW, the PSH does not operate; Eq. (2a) is applied. When Ebalance is less than 30,000 MW and Ebalance plus pumping power exceeds 30,000 MW, Eq. (2c) is applied to constrain the power consumption of the PSH to prevent the creation of an additional peak-power point. For other conditions, the operation of PSH follows Eq. (2b); XPSH as listed in Table 3. As for the generating operation of the PSH, Eq. (3b) applies for most conditions, but, if the water stored in a virtual reservoir (SPSH) is

Table 3 Ratio of operation for PSH. Hours 10:00 11:00 12:00 13:00 14:00 19:00 20:00 21:00 22:00 23:00

447

a.m. ∼ 11:00 a.m. ∼ 12:00 a.m. ∼ 13:00 a.m. ∼ 14:00 a.m. ∼ 15:00 a.m. ∼ 20:00 a.m. ∼ 21:00 a.m. ∼ 22:00 a.m. ∼ 23:00 a.m. ∼ 24:00

a.m. a.m. a.m. a.m. a.m. a.m. a.m. a.m. a.m. a.m.

Operation

Output ratio

Pumping Pumping Pumping Pumping Pumping Generating Generating Generating Generating Generating

0.6 0.7 0.8 1 1 0.8 1 0.9 0.7 1

Energy Conversion and Management 160 (2018) 439–454

Y.-C. Tsai et al.

Table 4 Ratio of operation for a hydroelectric power unit. Period

Hours

T1 & T3 T2

All 18:00p.m. – 19:00p.m. – 20:00p.m. – 21:00p.m. – 22:00p.m. – The others

3:00p.m. 4:00p.m. 5:00p.m. 8:00p.m. 12:00p.m.

Qflush,6 (t ) = Sr ,6 (t −1) + Qtri,6 (t ) + Qhydro,5 (t −Td1,5) + Qflush,5 (t −Td2,5) −Qdome,6 (t )−Qagri,6 (t )−Qunion (t )−Cr ,6 Mon. – Fri.

Sat. – Sun.

0.12 0.3 1.0 1.0 0.9 0.3 0

0.12 0.2 0.6 1.0 0.7 0.5 0

Sr ,6 (t ) = Sr ,6 (t −1) + Qtri,6 (t ) + Qhydro,5 (t −Td1,5) + Qflush,5 (t −Td2,5) −Qdome,6 (t )−Qagri,6 (t )−Qflush,6 (t )−Qunion (t )

(11)

in which Cr,6 is the effective capacity of Shigang reservoir. A parameter denoted as virtual water storage (Svirtual) is created to take into account the accumulated amount of Qunion, as shown in Eq. (12),

Svirtual (t −1) + Qunion (t ) a Svirtual (t ) = ⎧ b ⎨ ⎩Cr ,7−SOr ,7 (t )

within a safe range. The other reservoirs follow Eq. (8b) to maintain the total outflow equal to the total inflow. For Mode 2, Techi and Guguan hydroelectric units are fully operational (Eq. (6b)) to prevent flooding over the reservoir; Eq. (8a) becomes applied to Techi reservoir in the case that the inflow is extremely large. For other hydroelectric units, if the inflow of the reservoir is greater than the designed flow rate of the hydro-turbine plus Qeco, Eqs. (7b) and (8b) become applied. Otherwise, the minimum flushing flow (Eq. (8c)) is maintained for each reservoir; the flow for hydropower is limited by Eq. (7c). For mode 3, as the water level of the Shigang reservoir is still high, the Techi unit is not operated to preserve the water; Eq. (6a) is applied. For Modes 4 and 5, the inflow and water level of the Shigang reservoir are low; the Techi unit follows Eq. (6c) to release water. For Mode 6, because of the extremely low water level, the Techi unit cannot be operated (Eq. 6a). The operating strategies of the other units for Modes 3 to 6 are similar to those addressed in Mode 2.

(12)

in which Cr,7 is the effective capacity of Liyutan reservoir; SOr,7 is the actual water storage value of Liyutan reservoir. The water storage of Liyutan reservoir (Sr,7) is calculated through a sum of SOr,7 and Svirtual. When this sum is less than Cr,7, Eq. (12a) is normally applied; otherwise, the calculation of Svirtual turns to use Eq. (12b) to prevent flooding of the Liyutan reservoir. 3.6.5. Power supply and demand The calculation of the hydroelectric power of PHS (Ehydro,PSH) and pumping power (Epump) are shown in Eqs. (13) and (14), respectively,

Ehydro,PSH (t ) = ϕhydro,PSH (t ) × Qhydro,PSH (t ) × 9.81 × 1000/3600 × [ELr ,1 (t )−ELr ,3 (t )]

(13)

Epump (t ) = Qhydro,PSH (t ) × 9.81 × 1000/3600 × [ELr ,1 (t )−ELr ,3 (t )] / ϕpump (t )

3.6.4. Water supply and demand As the profiles of water demand are already known, the determination of the sufficiency of water supply relies on the rate of water storage of Shigang reservoir, as shown in Eq. (9),

(14)

in which ϕhydro,PSH and ϕpump are the turbine efficiencies for generation and pumping, respectively. The calculation of the power (Ehydro,PSH) generated from a traditional hydroelectric power unit is similar to Eq. (13). The solar, wind and actual power demands constitute the known profile as described in Section 3.4 and are treated as input data. With the addition of the desalination plant, the proposed system increases the power consumption while the desalination plant is operating. The power consumption is determined by the produced desalination water (Qdes) and the specific rate of energy consumption (ϕdes), as shown in Eq. (15).

Sr ,6 (t ) = Sr ,6 (t −1) + Qtri,6 (t ) + Qhydro,5 (t −Td1,5) + Qflush,5 (t −Td2,5) −Qdome,6 (t )−Qagri,6 (t ) + Qdes (t )−Qeco (t )

(10)

(9)

in which Sr represents the rate of water storage of the reservoir; Qdome and Qagri represent the demands for domestic and agricultural water, respectively. The concept of Eq. (9) is to calculate the water balance between inflow and outflow without considering water transport to or from Liyutan reservoir. If Sr remains at a positive value, the water supply of Shigang reservoir suffices; otherwise, a water shortage occurs. To facilitate water transport between Shigang and Liyutan reservoirs, a tunnel to link the water-treatment plants downstream of the aforementioned two reservoirs is under planning by the government. In this study, the conjunctive use of water between two reservoirs is considered; the maximum rate of flow in this tunnel is 0.0625 Mm3/h. During periods T1 and T3, Liyutan reservoir provides an increased water supply, which helps to decrease further the burden of water supply of Shigang reservoir. Parameter Qunion serves to define the amount of water transport between the two reservoirs. A positive value of Qunion represents that the water flow is from Shigang reservoir to Liyutan reservoir, and a negative value represents the reverse direction of water transport. The value of Qunion is determined by the rate of water storage of Shigang reservoir (Sr,6), the demands for domestic (Qdome,7) and agricultural (Qagri,7) water supplied by Liyutan reservoir, the capacity of the water-treatment plants and the actual rate of water storage of Liyutan reservoir. After Qunion is evaluated, the rate of flushing flow of Shigang reservoir (Qflush,6) is calculated with Eq. (10); the water storage of Shigang reservoir (Sr,6) becomes recalculated with Eq. (11),

Edes (t ) = Qdes (t ) × ϕdes × 1000

(15)

3.7. Mathematical model during period T2 The description of a model during period T2 is similar to those models for periods T1 and T3, but, for those strategies similar to Section 3.2, the content is addressed briefly. 3.7.1. Desalination plant As shown in the lower part of Fig. 5, when the time (t) is within period T2, three parameters – water level of Techi reservoir (ELr,1), the status of a typhoon and whether the time is under the peak-load period – serve to determine the applicable mode. The criteria to determine Modes 21 and 22 are the same as for Modes 1 and 2, respectively; the desalination plant is not operated. When ELr,1 exceeds the high-high limit (1407 m), mode 23 becomes applied; the desalination plant is still not operated. When ELr,1 is between the high-high limit (ELhihi) and the low limit (ELlolo), Table 4 is viewed to determine the applicable mode. If the ratio of operation for the hydroelectric power unit is greater than zero, it is defined as a non-peak-power period, and Mode 24 is applied; otherwise, Mode 25 is applied. The desalination plant is basically not operated in Modes 24 and 25, but becomes fully operational in Mode 25 when the rate of water storage of Shigang reservoir is less than 70%; 448

Energy Conversion and Management 160 (2018) 439–454

Y.-C. Tsai et al.

reservoir is under the expected value at a particular instant. For Mode 26, because of the extremely low water level, the Techi unit cannot be operated (Eq. (6a)). The operational strategies of the other units for Modes 22–26 are similar to those addressed in Mode 2.

Mode 26 is applied when ELr,1 < ELlolo, which represents an extremely low water level of Techi reservoir; the desalination plant is fully operated. 3.7.2. PSH unit When a typhoon occurs (Mode 21) or the water level of Techi reservoir is too low (Mode 26), the PSH does not operate as shown in Eq. (2a) and (3a). For the other modes, the operating strategies are the same as Modes 2–5.

3.7.4. Water supply and demand During period T2, the major difference in the water supply is the direction of Qunion. As the water released from Shigang reservoir is typically larger than the downstream water demand, some excess water is transported to the water-treatment plant downstream of the Liyutan reservoir. This condition makes Qunion have a positive value and helps to preserve the water in Liyutan reservoir. When the water level of Shigang reservoir is too low, Qunion becomes negative to compensate part of the water shortage.

3.7.3. Hydroelectric power unit The basic idea for the operation of hydroelectric units during period T2 is to release as much water as possible when the system passes through the peak-power demand period, as shown in Table 4. For Mode 21, the hydroelectric units are not operated because of the occurrence of a typhoon (Eq. (6a) for Techi units and (7a) for other units); the flushing action is required (Eq. (8a) for Techi units and (8b) for other units). For Mode 22, because of the high water level in Techi reservoir, the Techi unit is fully operated (Eq. (6b)); a flushing action might be required (Eq. (8a)). For Mode 23, the Techi unit is still fully operated (Eq. (6b)), but the rate of flushing flow is controlled at its minimum limit (Eq. (8c)). For Mode 24, as the system is under the peak-power demand period, the Techi unit is operated based on Eq. (6c); the rate of flushing flow is controlled at its minimum limit (Eq. (8c)). For Mode 25, as the system is under the non-peak power- demand period, the Techi unit is basically not operated (Eq. (8a)), but a mechanism is designed to initiate the operation of the Techi unit when the water level in Shigang

3.7.5. Power supply and demand The formulae used to calculate the power during T2 are the same as those shown in Section 3.6.5.

3.8. Calculation of the overall benefit of the system Before comparing the overall benefit of the ISRWR with the reference case, a similar concept of LCOE (levelized cost of electricity) is adopted to calculate the cost of power production, water production and energy storage, as shown in Eq. (16),

Table 5 Parameters for economic analysis. Sector

Parameters

ISRWR

Case Ref.

Financial setting

Discount rate Inflation rate of O&M Inflation rate of LNG fuel price & electricity cost Type of additional water supply unit Capacity of additional water supply units, Mm3/h Annual water production by additional water supply units, Mm3 an−1 Capital cost, million US$ per Mm3/h Power consumption for 1 m3 water, kWh m−3 Fixed O&M cost, million US$ Mm−3 Variable O&M cost, million US$ Mm−3 Electricity cost, US$ kWh−1 Cost recovery period, an Type of additional energy storage units, MW Capacity of additional energy storage units, MW Duration of power discharge, h Charging power/discharging power Discharging power by additional energy storage units, GWh Charging power by additional energy storage units, GWh Capital cost of added energy storage units, million US$ MW−1 Fixed O&M cost of added units, million US$ MW−1 Variable O&M cost of added units million US$ MWh−1 Electricity cost for charging, US$ kWh−1 Annual power requirement from NGCC plants during charging period, GWh Cost recovery period, an Type of added unit, MW Capacity of added units, MW Total hydropower capacity, excluding PSH, MW Total system capacity, excluding PSH, MW Peak power production by added units during T2, GWh Annual power production by the hydropower units, GWh Annual power requirement from NGCC plants during non-peak period, GWh Total power production of the system, GWh Capital cost of added units, million US$ MW−1 Fixed O&M cost of added units, million US$ MW−1 Variable O&M cost of added units, million US$ MWh−1 Fuel cost for NGCC, US$ kWh−1 Generating cost for existing NGCC, US$ kWh−1 Cost recovery period, an

5.0% 2.0% 1.6% Desalination plants 0.0130 95 41,577 3.3 0.048 0.155 0.084 20 PSH 1099 4 1.35 1574 2119 1.487 0.020 0.00 0.085 262 50 Hydropower 102 1303 1303 54 2065 466 2531 1.594 0.024 0 – 0.068 50

5.0% 2.0% 1.6% Reservoir 0.0109 95 72,859 0 0.032 0.061 – 50 Battery System 1099 4 1.18 1574 1857 2.968 0.040 0.000008 0.085 0 15 NGCC 853 1202 2055 292 2239 0 2531 0.977 0.011 0.000004 0.061 – 30

Water sector

Power storage sector

Power generation sector

449

Energy Conversion and Management 160 (2018) 439–454

Y.-C. Tsai et al. n

LCOE =

∑

n

[(Iy + My + Fy )/(1 + r ) y]

y=1

∑

rationing for the domestic water demand, which occurred in Cases A (2011) and E (2015). As for the water transport between two reservoirs, the annual values of accumulated Qunion for all cases range between 4 and 215 Mm3; they are all positive values. This fact shows that Shigang reservoir supports more water to Liyutan reservoir than it receives from Liyutan reservoir, so that this mechanism for water exchange makes unnecessary the water-storage tank in the previous study [20].

[Py /(1 + r ) y]

y=1

(16)

in which n represents the life span of each infrastructure for economic analysis; y represents year y within a life span; Iy, My, Fy and Py represent the capital cost, operating and maintenance costs, fuel cost and annual power or water produced in year y, respectively; r represents the discount rate. The parameters used for Eq. (16) are listed in Table 5. Some such values are taken from the previous study [20]; others are updated from the cost analysis report of EIA [26,27], the project report of the new PSH unit [28], and the results of calculations in this study. To make comparable the results from ISRWR and the reference case, both systems must expand their capabilities of power supply, water supply and energy storage to the same extent. A calculation of the benefit focuses on the costs of those extra power, water and energystorage requirements; they are defined in Eqs. (17) and (18) for ISRWR and the reference case, respectively,

4.2. Analysis of the power supply The target of the power supply is to enhance the capability of peakpower support through the hydroelectric units and PSH. According to Table 6, the peak value of the actual power demand (Ed) is 34,821 MW, which occurs at 15 h, as shown in item 1. Item 2 demonstrates the maximum values of Ebalance, which equals Ed minus solar power and wind power; the peak point is 33,013 MW that occurs at 20 h. The difference of the peak values between items 1 and 2 indicates that solar and wind power together can decrease the actual demand for peak power by only 1807 MW, which is far below their installed capacities. Understanding the capability of peak-power sharing of the hydroelectric units alone requires a combination of the power generated by all hydroelectric units (Edajia) and the power consumed by the desalination plant (Edes) to Ebalance, as shown in item 3. The peak value is 31,752 MW and occurs at 20 h. The minimum difference between items 2 and 3 is 1242 MW, which demonstrates that the installed hydroelectric units (1303 MW) can effectively share the power output during

CostIS = LCdes × AWdes + LCPSH × AEPSH + LChydro × APhydro + RNGCC × (ASPSH + AShydro )

(17)

in which CostIS is the annual cost of ISRWR; LCdes is the levelized cost of desalination water; AWdes is the annual water produced by the desalination plant; LCPSH is the LCOE of the PSH unit; AEPSH is the annual power produced by the PSH unit; LChydro is the LCOE of additional hydroelectric units proposed in ISRWR, and APhydro is the annual power supplied by these hydroelectric units during a peak-power demand period; RNGCC is the cost of electricity from existing NGCC units; ASPSH is the annual power required to compensate the power loss because of the lesser efficiency of the PSH unit; AShydro is the annual power required to compensate the decreased power output during a non- peakpower period,

CostR = LCr × AWr + LCBS × AEBS + LCNGCC × APNGCC

(18)

in which CostR is the annual cost for the reference case; LCr is the levelized cost of the new reservoir and AWr is the annual water supplied by the new reservoir; LCBS is the LCOE of the battery-storage system; AEBS is the annual power discharged by the battery-storage system; LCNGCC is the LCOE of additional NGCC units required for the peakpower period and APNGCC is the annual power produced by these units during periods of peak-power demand. 4. Results and discussion In this section, the feasibility of water supply and power supply are examined first; comparisons of benefits between ISRWR and the reference case under the same input profiles follow. 4.1. Analysis of the water supply The water supply of ISRWR can fulfill the requirement in all cases; the water released for generation and the rate of water storage of Techi reservoir are shown in Fig. 7. The profiles shown in Fig. 7 are similar to those in the previous study [20]; the most limiting case is Case A, in which the least rate of water storage is below 10%. On comparison with the actual rate of water storage, the operation of the hydroelectric units during period T2 in drought years, such as in Case A, second-half year of Case D and first-half year of Case E, makes the rate of water storage decrease rapidly, but, with the operating strategies during periods T1 and T3, the help of the desalination plant and the water support from Liyutan reservoir, the rate of water storage can be recovered to its safe range, or can be maintained at a more slowly decreasing rate in Case E. The domestic water demand in the ISRWR has been raised to 573.78 Mm3 per year for every case, which is the expected domestic water demand of Taichung city in 2031; the ISRWR entails no water

Fig. 7. Water released for generation and rate of water storage of Techi reservoir.

450

Energy Conversion and Management 160 (2018) 439–454

Y.-C. Tsai et al.

Table 6 Results of power supply for each case. Item

Content

1

Maximun power of Occurrence time Occurrence time Maximun power of Occurrence time Occurrence time Maximun power of Occurrence time Occurrence time Maximun power of Occurrence time Occurrence time Maximun power of Occurrence time Occurrence time

2

3

4

5

Ed

Ebalance*

Ebalance − Edajia** + Edes

Ebalance − Ehydro,PSH + Epump

Ebalance − Edajia + Edes − Ehydro,PSH + Epump

Unit

Case A

Case B

Case C

Case D

Case E

MW h hh:mm MW h hh:mm MW h hh:mm MW h hh:mm MW h hh:mm

33,788 5510 14:00 32,027 5468 20:00 30,784 5468 20:00 30,983 5468 20:00 29,741 5468 20:00

33,081 4623 15:00 31,504 4844 20:00 30,237 4844 20:00 30,456 4844 20:00 29,246 4822 22:00

33,957 5295 15:00 32,019 5228 20:00 30,751 5228 20:00 30,970 5228 20:00 29,779 5229 21:00

34,821 4695 15:00 33,014 4700 20:00 31,747 4700 20:00 31,974 6211 19:00 30,760 4702 22:00

34,821 4695 15:00 33,014 4700 20:00 31,752 4700 20:00 31,974 6211 19:00 30,744 6186 18:00

* Ebalance = Ed − wind power − solar power. ** Edajia: output of all hydropower units, excluding PSH.

the period of peak-power demand in all cases. Fig. 8 shows how the hydroelectric units work during a week. In most conditions, the output strategy shown in Table 4 works satisfactorily to respond to the peakpower demand, but the delay of peak hours appearing in Case D can disturb the efficacy of the hydroelectric units. Understand the capability of peak-power sharing of the PSH units alone requires a combination of the power generated (Ehydro,PSH) and power consumed (Epump) per PSH unit to Ebalance, as shown in item 5. The peak value is 31,974 MW and occurs at 19 h. The minimum difference between items 2 and 4 is 1040 MW, which demonstrates that the installed PSH unit (1099 MW) can effectively share the power output during the period of peak-power demand in all cases. Fig. 9 shows how the PSH unit works during a week. In most conditions, the output strategy shown in Table 4 works satisfactorily to respond to the peak-power demand. The blue1 and green lines represent the generation and pumping powers, respectively. The efficacy of peak-power sharing is similar to that of the hydroelectric units, but the effort to fill the valley of power demand through the pumping operation is not evident. The principal reason lies in the large installed capacity of solar power and its large fluctuation during the valley of power demand, which is beyond the ability of the PSH unit. On consideration of the combined effect of the hydroelectric units and the PSH unit (item 5), the peak value of power demand is 30,760 MW and occurs at 22 h; the minimum capability of peak-power sharing among all cases is 2240 MW. Fig. 10 shows the overall effect of the hydroelectric units and the PSH unit. The sharp profile of the actual power demand is obviously rounded to a smoother one. 4.3. Analysis of benefits Table 7 shows the results of the benefit analysis; the annual costs of additional water production, peak-power sharing and energy storage are listed as items (A), (B) and (C), respectively. Because of the higher levelized cost of water production for a desalination plant (US$1.022 per m3), the ISRWR must spend US$97 million for the additional water production per year, which is US$41 million more than the reference case. According to the simulation results of ISRWR, with only a slight increase of the hydroelectric units, 853 MW of NGCC units can be eliminated to handle the peak-power demand. Through the decrease of these NGCC units, the cost of the peak-power generation of ISRWR is decreased from US$89 million to US$44 million per year, which equals

Fig. 8. Power sharing of hydroelectric units during a peak-power period. *Case E uses the same power demand curve as Case D.

1 For interpretation of color in Fig. 9, the reader is referred to the web version of this article.

451

Energy Conversion and Management 160 (2018) 439–454

Y.-C. Tsai et al.

Fig. 9. Power sharing of PSH units during a peak-power period. *Case E uses the same power demand curve as Case D.

Fig. 10. Power sharing of hydroelectric units and PSH unit during a peak-power period. * Case E uses the same power demand curve as Case D.

a saving of US$45 million per year. On comparing the cost of energy storage between the PSH unit and the battery storage, the PSH unit shows a great advantage over the battery system. The LCOE of the PSH unit and the battery system are US $0.226 and US$0.341 per kWh, which results in the cost of energy storage per year to be US$375 million and US$541 million for the ISRWR and the reference case, respectively. This condition also represents a saving of US$166 million in the ISRWR. On combining the above benefits, the ISRWR can save US$171 million per year relative to the reference case, which is about a 24.9% saving.

utilization of solar power within a day. The demand of water for Taichung city in 2031 can be satisfied with the operation of the ISRWR; the water safety can be ensured even during a drought year. The development targets solar power 20 GW and wind power 4.2 GW are considered in this study, which makes great changes in the profile of the power demand. Because of their intermittent nature, these solar and wind powers can decrease the peak of the power demand by only 1807 MW. Through the cooperation of the hydroelectric units and the PSH unit, the ISRWR can further decrease the peak of the power demand by 2240 MW and smooth the actual sharp profile. To achieve the same capability, a 853-MW NGCC unit and a 1099-MW battery storage system are required for the reference case. From the point of view of benefit, even the increased cost of the desalination water can cause an extra cost US$41 million per year; the benefits generated from the power-generation sector, US$45 million, and the energy-storage sector, US$166 million, can together readily compensate for this expenditure. To have the same capability of water production, power generation and energy storage, the ISRWR can save US$171 million per year relative to the reference case. The results of the analysis in this study indicate also the following.

5. Conclusion An integrated system for renewable energy sources and water resources (ISRWR) with an operating model that integrates the utilization of RES and water resources is established in this study. Through the proposed strategies of operation, the impact of intermittent characteristics of each resource can be alleviated. The ISRWR converts the existing hydroelectric units into the character of peak-power support with only a slight increase of capacity, 102 MW; the derivative water issues are covered through the addition of a desalination plant and the mechanism of water transport between Shigang and Liyutan reservoirs. Furthermore, a 1099-MW PSH unit is included to yield an improved

• Even with the large amount of solar power and wind power in-

volved, the concept of the ISRWR remains applicable. It is timely for a policy maker to re-evaluate the water resources on hand; a highly

452

Energy Conversion and Management 160 (2018) 439–454

Y.-C. Tsai et al.

Table 7 Results of the benefit analysis. Item

Content

ISRWR

Case Ref.

(A) Additional water production

Type of additional water supply units Annual water production by additional water supply units, Mm3 Levelized cost of water produced by new system, US$ m−3 Additional cost for water per year/million US$ an−1 Type of added unit, MW Peak power production by added units during T2, GWh Annual power requirement from NGCC plants during non-peak period, GWh Cost recovery period, an Levelized cost of electricity produced by added units, US$ kWh−1 Additional cost for peak-power generation per year, million US$ an−1 Type of additional energy storage unit, MW Discharging power by additional energy storage units, GWh Charging power by additional energy storage units, GWh Annual power requirement from NGCC plants during charging period, GWh Levelized cost of electricity produced by added units, US$ kWh−1 Additional cost for power storage per year, million US$ an−1 ISRWR(A) – Case Ref.(A) ISRWR(B) – Case Ref.(B) ISRWR(C) – Case Ref.(C) ISRWR((A) + (B) + (C)) − Case Ref. ((A) + (B) + (C))

Desalination plants 95 1.022 97 Hydropower 54 466 50 0.227 44 PSH 1574 2119 262 0.226 375 -41 45 166 171

Reservoir 95 0.594 56 NGCC 292 0 30 0.306 89 Battery System 1574 1857 0 0.344 541

(B) Peak-power sharing

(C) Power storage

Summary

•

•

•

operation.

integrated system among solar power, wind power and hydropower is possible to harmonize their respective intermittencies. Under the global climate changes, the desalination water provides a water source better ensured than rainfall, but the greater cost of water production makes desalination practical for only those areas suffering a serious water shortage. A new business model for the large scale of a desalination plant is demonstrated in this study, which expands the applicable area and climate condition of this technology. On considering the sharing of peak-power demand, the solar power typically works well for the first peak in the afternoon, and for those days on which solar power is weak because of cloudy or rainy weather, the power demands are generally less than on sunny days. This phenomenon makes the second peak occurring at night become a greater issue when the rate of penetration of solar power is large. A concentrated operation of hydropower as shown in this study might be a possible solution. As the public awareness of environmental protection is great in recent years, building a new reservoir becomes difficult. If PSH units might be added to existing reservoirs, these issues about environmental protection might be greatly diminished. The ISRWR demonstrates also that the PSH unit and the existing hydroelectric units can be operated together to create increased benefits in a reservoir.

Acknowledgment Ministry of Science and Technology of Taiwan (ROC, Taiwan) partially supported this work under contract MOST 106-3113-F-002-004. References [1] IEA. The power of transformation - wind, sun and the economics of flexible power systems. Paris: OECD Publishing; 2014. [2] Ali B. The cost of conserved water for power generation from renewable energy technologies in Alberta, Canada. Energy Convers Manage 2017;150:201–13. [3] Foley AM, Leahy PG, Li K, McKeogh EJ, Morrison AP. A long-term analysis of pumped hydro storage to firm wind power. Appl Energy 2015. [4] Anagnostopoulos JS, Papantonis DE. Study of pumped storage schemes to support high RES penetration in the electric power system of Greece. Energy 2012;45:416–23. [5] Tuohy A, O’Malley M. Pumped storage in systems with very high wind penetration. Energy Policy 2011;39:1965–74. [6] Chen CL, Chen HC, Lee JY. Application of a generic superstructure-based formulation to the design of wind-pumped-storage hybrid systems on remote islands. Energy Convers Manage 2016;111:339–51. [7] Portero U, Velázquez S, Carta JA. Sizing of a wind-hydro system using a reversible hydraulic facility with seawater. A case study in the Canary Islands. Energy Convers Manage 2015;106:1251–63. [8] Ma T, Yang H, Lu L, Peng J. Pumped storage-based standalone photovoltaic power generation system: modeling and techno-economic optimization. Appl Energy 2015;137:649–59. [9] Georgiou D, Mohammed ES, Rozakis S. Multi-criteria decision making on the energy supply configuration of autonomous desalination units. Renew Energy 2015;75:459–67. [10] Mentis D, Karalis G, Zervos A, Howells M, Taliotis C, Bazilian M, et al. Desalination using renewable energy sources on the arid islands of South Aegean Sea. Energy 2016;94:262–72. [11] Ismail TM, Azab AK, Elkady MA, Abo Elnasr MM. Theoretical investigation of the performance of integrated seawater desalination plant utilizing renewable energy. Energy Convers Manage 2016;126:811–25. [12] Leiva-Illanes R, Escobar R, Cardemil JM, Alarcón-Padilla D-C. Thermoeconomic assessment of a solar polygeneration plant for electricity, water, cooling and heating in high direct normal irradiation conditions. Energy Convers Manage 2017;151:538–52. [13] Spyrou ID, Anagnostopoulos JS. Design study of a stand-alone desalination system powered by renewable energy sources and a pumped storage unit. Desalination 2010;257:137–49. [14] Segurado R, Costa M, Duić N, Carvalho MG. Integrated analysis of energy and water supply in islands. Case study of S. Vicente, Cape Verde. Energy 2015;92(Part 3):639–48. [15] Novosel T, Ćosić B, Krajačić G, Duić N, Pukšec T, Mohsen MS, et al. The influence of reverse osmosis desalination in a combination with pump storage on the penetration of wind and PV energy: a case study for Jordan. Energy 2014;76:73–81. [16] Perković L, Novosel T, Pukšec T, Ćosić B, Mustafa M, Krajačić G, et al. Modeling of optimal energy flows for systems with close integration of sea water desalination and renewable energy sources: case study for Jordan. Energy Convers Manage

In this research, most input data and system configuration are based on real world data or official reports; the benefit that it can extract makes the ISRWR readily practical for Taiwan. With some modifications of the system’s configuration, which is specific for the different countries, the philosophy used in this research is also amenable to other countries with hydropower systems. For those countries having abundant water resources, the desalination plant in the ISRWR might be eliminated or decreased to extract increased total benefits; for those countries without much water resource, ISRWR can be a small-scale system depending on the available water resources. The result of this study also demonstrates that, with an appropriate combination of intermittent RES, desalination plants and hydropower, a regular operating pattern is possible to alleviate the impact from intermittent RES. Even without a large penetration of solar or wind power, ISRWR remains applicable to the existing power system [20]. Without awaiting the maturity of new technologies, this research taps the great potentials, which hide behind the nexus between power and water resources, through the integration of mature technologies and a novel strategy of 453

Energy Conversion and Management 160 (2018) 439–454

Y.-C. Tsai et al.

comefrom=lp#8890 > ; 2016 [accessed 12.01.16]. [23] Bee LJ. Reviewable report on feasibility study of Tien Hwa Hu Reservoir. Taichung: Water Resources Planning Institute, WRA; 2011. [24] Water Resource Agency. Monitoring data of the reservoirs. Taiwan. Available online: < http://fhy.wra.gov.tw/fhy/Monitor/Reservoir > ; 2015 [accessed 12. 01.16]. [25] Central Region Water Resources Office WRA. Statistics of Shigang reservoir and Liyutan reservoir. Available online: < http://hydro2.wracb.gov.tw/ WracbHydrology/ResoirReports/ShgaYearRpt.aspx > ; 2016 [accessed 01.12.16]. [26] EIA. Addendum: Capital Cost Estimates for Utility Scale Electricity Generating plants. Available 2017 [accessed 20.04.17]. [27] EIA. Capital Cost Estimates for Utility Scale Electricity Generating plants. Available 2016 [accessed 20.04.17]. [28] SINOTECH engineering consultants. Investigation and planning project of pumped hydroelectric potential sites in Taiwan. Environmental protective administration (Taiwan). Taipei. Available 2013 [accessed 20.04.17].

2016;110:249–59. [17] Al-Nory M, El-Beltagy M. An energy management approach for renewable energy integration with power generation and water desalination. Renew Energy 2014;72:377–85. [18] Saif Y, Almansoori A. A capacity expansion planning model for integrated water desalination and power supply chain problem. Energy Convers Manage 2016;122:462–76. [19] Kyriakarakos G, Dounis AI, Rozakis S, Arvanitis KG, Papadakis G. Polygeneration microgrids: a viable solution in remote areas for supplying power, potable water and hydrogen as transportation fuel. Appl Energy 2011;88:4517–26. [20] Tsai YC, Chiu CP, Ko FK, Chen TC, Yang JT. Desalination plants and renewables combined to solve power and water issues. Energy 2016;113:1018–30. [21] Bureau of Energy. Energy Statistics Handbook 2016. Available online: < http:// web3.moeaboe.gov.tw/ecw/populace/content/SubMenu.aspx?menu_id=5427 > ; 2017 [accessed 30.09.17]. [22] Water Resource Agency. Current Status of the Reservoirs (2015). Taiwan. Available online: < http://www.wra.gov.tw/ct.asp?xItem=20063&ctNode=8890&

454