Concentrated solar power components toolbox in an object oriented environment

Simulation Modelling Practice and Theory 70 (2017) 21–35

Contents lists available at ScienceDirect

Simulation Modelling Practice and Theory journal homepage: www.elsevier.com/locate/simpat

Concentrated solar power components toolbox in an object oriented environment Christos Kalathakis a,∗, Nikolaos Aretakis a, Ioannis Roumeliotis b, Alexios Alexiou a, Konstantinos Mathioudakis a a

Laboratory of Thermal Turbomachines, School of Mechanical Engineering National Technical University of Athens, Iroon Polytechniou 9, Zografou, Athens 15780, Greece b Section of Naval Architecture & Marine Engineering, Hellenic Naval Academy, Piraeus, Greece, Hadjikyriakou Avenue, Piraeus, Athens 18539, Greece

a r t i c l e

i n f o

Article history: Received 15 June 2016 Revised 7 September 2016 Accepted 10 October 2016

Keywords: CSP Solar STPP Solar hybrid

a b s t r a c t A toolbox for modeling solar components for gas and steam turbine Solar Thermal Power Plants (STPPs) is presented. It has been created in order to supplement the PROOSIS modeling environment that covers the power production parts. Solar and power production parts are both represented at a similar level of fidelity. The toolbox contains components for simulating all the individual solar elements used in STPPs in order to materialize Brayton, Rankine and Combined Cycles. Functionalities for computing solar irradiation properties as well as working fluid thermodynamic properties are included. The use of the toolbox is demonstrated through simulation cases at component and plant level, while its features, capabilities and merits are discussed. The developed capabilities offer the possibility to perform plant design optimization, operational support through performance prediction at various operating modes as well as assessment of the effect of components malfunctions. © 2016 Elsevier B.V. All rights reserved.

1. Introduction The use of fossil fuels for power production is one of the main contributors to climate change at local and global level, while their finite nature highly affects their selling price and availability. In order to limit their effects, renewable energy sources, such as wind and sun power, are used for their substitution. Solar energy, already extensively exploited through photovoltaics, can be used in Solar Thermal Power Plants. Solar thermal power addition to Brayton and Rankine cycles can partially or totally substitute the fuel thermal power. Solar Thermal Power Plants have already been built, mainly using troughs for solar heat collection and oil as Heat Transfer Fluid in Rankine cycle [1]. The Direct Steam Generation from troughs is under investigation as well as the use of other materials as HTF [2–4]. In the Brayton cycle, solar thermal power has been used to preheat the air before it enters the combustion chamber. EU funded projects [5–7] and other research studies examine this option [8–11]. The prediction of the behavior of such plants and the accurate simulation of their performance is a necessary task in order to estimate economic aspects of an investment and evaluate new concepts. There are many in-house and commercial ∗

Corresponding author. E-mail addresses: [email protected], [email protected] (C. Kalathakis), [email protected] (N. Aretakis), [email protected] (I. Roumeliotis), [email protected] (A. Alexiou). http://dx.doi.org/10.1016/j.simpat.2016.10.002 1569-190X/© 2016 Elsevier B.V. All rights reserved.

22

C. Kalathakis et al. / Simulation Modelling Practice and Theory 70 (2017) 21–35

Fig. 1. Solar thermal power and Brayton cycle.

Fig. 2. Solar thermal power and Rankine cycle.

simulations tools available based on or linked with two widely used simulation programs, namely SAM [12] and TRNSYS [13]. SAM is a powerful package limited to simulating only Rankine cycles [12], thus it cannot be used for the evaluation of Combined Cycle and Gas Turbine type of STPPs. TRNSYS [13] utilizes semi-empirical relations for simulating gas and steam turbine operation. This approach does not offer flexibility for performing various types of performance and optimization studies, such as the optimization of a gas turbine IGVs schedule for hybrid operation. Additionally, it does not allow efficient component faults modeling and engine health monitoring. PROOSIS is a powerful tool used for simulating the performance of thermal turbomachines in an object oriented environment and is currently used throughout the aeronautical community [14]. It contains the TURBO library [15] which allows the accurate simulation of gas turbines [16]. For modeling Rankine cycles and Combined cycles, in conjunction with the TURBO library, the authors developed the WAST library [17]. For modeling STPPs a suitable library has been developed, named SOLAR. It contains components for simulating all the individual solar elements used in STPPs (e.g. Figs. 1 and 2). The libraries have been used by the authors for investigating the effect of gas turbine configuration on the performance of a hybrid Brayton cycle and to evaluate approaches for exploiting the rejected Sun power [18,19]. In the present work, the development of SOLAR library is discussed. The components mathematical modeling along with validation examples are presented. For demonstrating the cooperation of the different libraries the model of a hybrid gas turbine based STPP is built, its operation is simulated for various scenarios and relevant results are discussed. 2. Component modeling The library has been developed in PROOSIS [14], an object-oriented environment. It uses libraries (e.g. TURBO, ELECTRICAL, CONTROL) which contain components. A library component is the mathematical model of real world component exhibiting the following features: -

providing reliable modeling of physical reality, using appropriate mathematical formulations structured with variables allowing easy adaptation to practical cases (eg handling different working media) ability to communicate with other components, so that entire installation models can be produced be user friendly and flexible to accommodate alternative configurations

C. Kalathakis et al. / Simulation Modelling Practice and Theory 70 (2017) 21–35

23

The components of SOLAR simulate the operation of flat and parabolic mirrors, tower receivers, heat exchangers, thermal storage tanks and Sun position and radiation. A detailed description of component modeling is given below. The icons used in the modeling environment for handling the components and their inputs and outputs are given in the Appendix. 2.1. General features of components 2.1.1. Component structure A component computes variable values for certain given inputs. Input data may be flow state or component attributes (e.g. geometrical aspects), the latter being represented by component Data Variables. The variables computed by a component may be input to another component. The transfer of variable information between components can be achieved by using “ports”, which transfer predefined information according to their type. The transferred information concerns the thermodynamic state of the working fluid, energy quantities (eg reflected sun thermal energy from mirrors to the receiver) or other information in order to compute the overall performance of a cycle. Communication is also performed through Global Variables which are variables computed by a certain component and their values are visible by any other component that uses them. Multi-disciplinary and inter-library simulations are possible through the use of common type of ports. SOLAR library components use ports that are defined in other libraries (inter-library connection), which transfer flow information about the thermodynamic state and the mass flow for air/gas and water/steam flow. 2.1.2. Fluid properties functions Appropriate functions are used to compute fluid properties such as thermal capacity, density, conductivity, viscosity and enthalpy of the HTF for the desired thermodynamic state. These functions use HTF data available in manufactures catalogues (like DOW Chemical Company [20]). Reverse functions are also implemented i.e. to compute temperature from known enthalpy. Functions for computing air/gas and water/steam properties are contained in TURBO and WAST libraries respectively. 2.1.3. Pressure drop Pressure drop calculations are encountered in a number of components such as tower receivers and heat exchangers. A first type of calculation is through pressure loss coefficient (K), given input value (K = KDP ):

Pout =1−K Pin

(1)

Alternatively, K is computed using fluid mass flow rate and properties and the design point values:

⎛ K = KDP · ⎝ 

m˙ in · m˙ in ·





⎞2 Tin /Pin

Tin /Pin

 ⎠

(2)

DP

For components with several input and output flows, pressure drop calculations are performed for each flow. 2.2. Individual component description The physical principles and equations governing the processes in individual solar components, used for their modeling in SOLAR, are described below. 2.2.1. Solar angles and irradiation This component computates solar zenith and azimuth angles (and irradiation, if not available) for the desired place and time. It performs the calculation of Sun position in the celestial dome during its relative movement through the computation of zenith and azimuth angles. Knowledge of Sun position permits the calculation of solar beam vector and consequently the calculation of the total available thermal energy on a mirror given the irradiation amount. The computation of these angles is based on the calculation of quantities (hour angle, day angle, etc.) depending on the time, the day and the location [21–23]. The zenith angle () and azimuth angle (z) are computed by:

cos() = sin(L ) sin(δ ) + cos(L ) cos(δ ) cos(h )



sin(z ) = sign(h ) · cos−1



cos() · sin(L ) − sin(δ ) sin() · cos(L )



(3)

Where: L is the latitude, δ is the declination angle, h is the hour angle. If solar irradiation data is not available, SUN component estimates the irradiation amount (direct, diffuse and total) for clear sky for any time and any location. This estimation is based on calculating the interaction between Sun beams heading to Earth and the atmosphere layers. In the calculation, thickness of atmospheric layers and the spectrum of solar irradiation

24

C. Kalathakis et al. / Simulation Modelling Practice and Theory 70 (2017) 21–35

are accounted. For example, the total direct irradiation (Id ) is the sum of the direct irradiation for each wavelength (Id,λ ) computed by Eq. (4). A detailed analysis concerning relevant equations and coefficient values can be found in [24,25].

Id,λ = HOλ · D · Trλ · Taλ · Twλ · Toλ · Tuλ

(4)

Where: HOλ is the irradiation at the mean Sun–Earth distance, D is the correction factor for the Sun–Earth distance, Trλ is the molecular atmospheric scattering coefficient, Taλ is the aerosol permeability coefficient, Twλ is the absorptivity coefficient due to humidity, Toλ is the absorptivity coefficient due to ozone, Tuλ is the absorptivity coefficient due to other gasses. 2.2.2. Tower receiver (air, HTF and water/steam) This component is used to calculate the outlet air flow state, given the inlet state, the amount of the reflected solar thermal power from the heliostat field, ambient conditions and receiver characteristics. It is assumed to be a pressurized volumetric receiver which is directly irradiated by the reflected beams of the heliostat field [26,27]. It is modeled as a black body with constant absorptivity, emissivity and optical efficiencies. Due to the high temperatures and insulation, radiation losses are assumed to be the main loss mechanism from the window and therefore losses due to convection are neglected. The radiation losses are computed for the mean temperature between the inlet and outlet of the receiver. In order to determine outlet air state, the thermal power given to air is computed. This quantity is computed as the incoming one from the heliostat field, reduced by the absorptivity and optical efficiencies and by the radiation losses [13]. Furthermore, the aperture area of receiver is chosen to satisfy the flux limitation imposed by the glass [28]. The optical efficiency of the receiver is input parameter and is in accordance to the minimum glass thickness imposed by the air pressure [6,26]. Therefore, the thermal power (Q) given to the working fluid is:



4 Q = ηopt arec Qin − Arec εrec σ Tm4 − Tamb = mw f (ht,out − ht,in )

(5)

Where: ηopt is the optical efficiency coefficient, α rec is the absorptivity coefficient, Qin is the incoming thermal power from the mirrors, Arec is the aperture of the window, ε rec is the emissivity coefficient, σ is the Stefan-Boltzmann constant, Tm is the mean working fluid temperature, Tamb is the ambient temperature, mwf is the working fluid mass flow, ht,out and ht,in is the working fluid’s outlet and inlet enthalpy respectively. Furthermore, there is the ability to set a limit for the outlet temperature. If the outlet temperature tends to surpass this limit due to high irradiation, it takes the limit value, while the used solar power is decreased simulating the defocusing of mirrors at the heliostat field. Finally, pressure losses are computed on the basis of actual fluid mass flow and density. The solar HTF or water heating and evaporation carried out by a tower receiver, is simulated by appropriate Tower Receiver components (Tower Receiver HTF and Tower Receiver Water/Steam respectively), in the same way as the component for air heating. A glass window is not used in this component and therefore there is no pressure limitation and no optical efficiency. The absorptivity efficiency is an input parameter and concerns the piping material in which the water/steam or HTF flows.

2.2.3. Tower receiver (dual fluid receiver) This component is a combination of two receiver components, one with working fluid air and one with water/steam. It simulates air heating and steam production from the excess solar thermal energy from defocused mirrors. They are supposed to be focused on a secondary receiver on top of the primary one. The whole amount of excess thermal power (minus the thermal losses) is now available on the secondary receiver for steam production.

2.2.4. Heliostat field With the use of this component, the reflected irradiation amount from the heliostat field to the receiver is calculated given the amount of Sun irradiation and mirrors’ and receiver’s aspects. The sun–mirror and the mirror–receiver vectors are computed for every mirror in the field in order to compute the reflectance angle and consequently, the cosine losses. The reflected power from the mirrors is the incoming one reduced by the cosine, reflectance and shading–blocking efficiencies, the mirror heat interception coefficient and by the atmospheric attenuation efficiency. The cosine and the atmospheric attenuation efficiencies are computed for each mirror for the desired hour of the day [29]. For simplicity, the reflectance, the heat interception and the shading–blocking efficiencies take a constant annualized value [30]. The total reflected power is the sum of the reflected power from each mirror (Qmir ), which is:

Qmir = Id · Amir · ηcos · ηsb · ηit · ηaa · ηre f

(6)

Where: Id is the incoming direct irradiation, Amir is the aperture of the mirror, ηcos is the cosine loss coefficient, ηsb is the shading-blocking loss coefficient, ηit is the heat interception loss coefficient, ηaa is the atmospheric attenuation loss coefficient, ηref is the reflectance efficiency. If the mirror coordinates are not available, the component has the ability to produce a heliostat field assuming it has been built according to the procedure described in [29,31] using tower and mirror dimensions.

C. Kalathakis et al. / Simulation Modelling Practice and Theory 70 (2017) 21–35

25

Fig. 3. Cross-section of trough receiver.

2.2.5. Troughs (HTF and water/steam) The task of this component is the calculation of the thermodynamic state at the outlet of troughs given the inlet state, troughs’ aspects and ambient conditions. The operation of parabolic mirrors (troughs) with N-S axis orientation and E-W rotation is simulated. Also, the option is implemented for static mirrors or fully tracking with their aperture continuously perpendicular to sun beams. Troughs with vacuumed receiver are used. Heat Transfer Fluid (oil) flows through a metallic pipe coated with absorptive material. This pipe is positioned in the center of a vacuumed glass pipe (Fig. 3). For the simulation, the entire length of each trough is divided in user defined segments of same length. For each segment, the same temperature is supposed for any point of each receiver’s part (glass envelope, absorber etc.) and heat exchange is supposed to be only at radial direction. For every part, heat balance is performed accounting for losses (cosine, conduction, convection and radiation) and efficiency factors (optical and absorptivity) to compute the temperature of each layer and consequently the thermal gain and pressure losses of the working fluid. Then, with thermal balance between the input and output of each segment and using the thermal gain, the outlet temperature, enthalpy and velocity can be computed. Furthermore, end losses and bracket losses for each trough row are fully accounted. A detailed analysis can be found in [32]. The thermal balance for each segment is:

q˙ 12conv = q˙ 23cond q˙ 3SolAbs = q˙ 34conv + q˙ 34rad + q˙ 23cond q˙ 34conv + q˙ 34rad = q˙ 45cond q˙ 45cond + q˙ 5SolAbs = q˙ 56conv + q˙ 56rad

(7)

Where: q˙ 5SolAbs is the absorbed energy from the glass envelope, q˙ 3SolAbs is the absorbed energy from the absorber,  23 q˙ cond is the thermal transfer due to conduction at the absorber, q˙ 34conv is the thermal transfer due to convection at the annulus, q˙ 34rad is the thermal transfer due to radiation at the annulus, q˙ 12conv is the thermal transfer to the working due to convection, q˙ 45cond is the thermal transfer due to conduction at the glass envelope, q˙ 56conv is the thermal transfer due to convection from the glass envelope to the environment, q˙ 56rad is the thermal transfer due to radiation from the glass envelope to the environment. The simulation of Direct Steam Generation (DSG) from parabolic mirrors is performed by the troughs component which uses water/steam as working fluid. This component is modeled in the same way as the HTF (oil) component, while appropriate equations are implemented in order to compute the thermal transfer coefficient [33,34] and the pressure losses [35] for the water/steam two phase flow 2.2.6. Heat exchangers This component computes the thermodynamic state of outlet flows of a heat exchanger given its design characteristics and the inlet state.

26

C. Kalathakis et al. / Simulation Modelling Practice and Theory 70 (2017) 21–35

The operation of counter flow heat exchangers is simulated. Two types are simulated, air/gas and HTF/water. Heat exchangers are modeled using the NTU method with the effectiveness (ε ) in the appropriate form, depending on the way thermal transfer involving sensible or latent heat [36]:

NT U =

ε=

UA Cmin Q˙

Q˙ max = min(m˙ hC ph , m˙ cC pc )

Cmin Q˙ = m˙ h (ht,h,in − ht,h,out ) = m˙ c (ht,c,out − ht,c,in ) Q˙ max = Cmin (Th,in − Tc,in )

ε = f (Cmin , NT U )

(8)

Where: Q is the transferred thermal power, Qmax is the ideal power that can be transferred, m is the stream mass flow, Cp is the stream thermal capacity, h is the enthalpy, T is the temperature, c,h denote the cold and hot stream respectively, in, out denote the inlet and outlet of the stream. At the design point (DP), the quantity UA is computed so that the desired outlet flow conditions are met. For every other operation point, the quantity UA can take the design value or it can be computed again (Eq. (9)) taking into account its design value and the difference of mass flow for the two streams from their design values [37]. This option is selected by an appropriate switch.



UA/U ADP =

.8 m0c,DP

m0c .8 + m0h.8

+

.8 m0h,DP

m0c .8 + m0h.8



·

mc · mh mc,DP · mh,DP



(9)

Pressure losses are computed using data for the actual fluid mass flow and density as described. 2.2.7. Thermal energy storage (TES) The amount and thermodynamic state of the stored mass is computed, given the inlet and outlet flows and information about thermal losses and thermal addition. The performance of the direct thermal storage is simulated, where hot HTF is stored in a tank in order to be used at low or zero irradiation conditions. The HTF mass flow is chosen in a way that the desired temperature is achieved at the outlet of solar heating component (e.g. troughs). If the mass flow is greater than the required from the steam generation component, then the excess mass flow is directed to the storage tank. Furthermore, with the use of a heat exchanger, the working fluid in the storage tank can be different from the one that is used by the solar heating component. In the storage tank mass and energy balance is:

dMstor = m˙ in − m˙ out dt

(10)

dQstor = m˙ in · hin − m˙ out · hstor − qloss + qadd dt

(11)

Where: Mstor is the stored mass, min , mout is the inlet and outlet mass flow to the tank respectively, Qstor is the stored thermal energy, hin , hstor is the enthalpy of the inlet and the stored mass respectively, qloss is the thermal loss, qadd is the added thermal power to the storage tank via heaters. Thermal loss at design point (qloss,DP ) is an input to the model. For any other operating point, thermal loss can be set to the desired value or can be computed taking into account the design value and the difference of the ambient temperature and the stored mass between the actual and the design point [38]:

qloss = qloss,DP · fT · fM

(12)

Where: fT is the correction coefficient due to ambient temperature; fM is the correction coefficient due to stored mass amount. 2.2.8. Performance monitor It is used in order to visualize variable and parameter values for monitoring the performance of a modeled installation. It receives values from other components, via appropriate ports, and computes quantities such as thermal efficiency, fuel thermal efficiency, solar share etc, which are useful for assessing the overall performance of solar hybrid cycles. This component may also be used for interfacing with monitoring or control schemes of a solar power plant. 3. Simulation test cases - validation For validating the component modeling, suitable test cases are simulated utilizing public available data.

C. Kalathakis et al. / Simulation Modelling Practice and Theory 70 (2017) 21–35

0.20

1200

Value

800

Data Predictions % Difference

0.15 0.10 0.05

600

0.00

400

-0.05

200

-0.10

0

Data Predicons % Difference

Zenith Angle [deg] 30.25 30.22 -0.10

Azimuth Angle [deg] 65.67 65.6 -0.11

Direct Diffuse Total Irradiance Irradiance Irradiance [W/m2] [W/m2] [W/m2] 843.3 193.12 1009.4 844.8 193.2 1011 0.18 0.04 0.16

% Difference

1000

27

-0.15

1000 900 800 700 600 500 400 300 200 100 0

Data Predicons % Difference

0.60 Data Predictions

% Difference

0.40 0.20 0.00

-0.20 -0.40

% Difference

Value

Fig. 4. Difference between data and computed zenith and azimuth angles and direct, diffuse and total irradiation.

-0.60 Tout [oC] 871 875.3 0.49

Pout [kPa] 270.3 270.3 0.00

eff [%] 89.25 88.75 -0.56

-0.80

Fig. 5. Difference between data and computed outlet temperature and pressure and efficiency.

3.1. SUN test case The operation of the SUN component is demonstrated by simulating two cases, one for computing azimuth and zenith angles and one for computing direct, diffuse and total irradiation. The first case is the Example 2.6 in [23] and concerns the computation of solar azimuth and zenith angles for a city located at 40 °N at 14:00 of June 15. The second case is the NRELs Excel worksheet [39] and concerns the computation of direct, diffuse and total irradiation for a place with longitude and latitude 30° and 38° respectively, for 12:00 of June 21st. The mentioned data is input for the model and the results of the examples and the model are shown in Fig. 4. Differences are found to be lower than 0.2%. 3.2. TOWER RECEIVER (air) test case For the RECEIVER example, the simulation of the experiment reported in [40] is performed, the experimental operation of a hybrid microturbine. Input parameters are the measured/computed values of incoming thermal power, inlet mass flow, temperature and pressure, ambient temperature and pressure loss. Fig. 5 depicts data and computed values and their difference for the outlet air temperature (Tout ) and pressure (Pout ) and for the receiver’s efficiency (eff). The differences of about 0.5% are observed. 3.3. HELIOSTATS test case The operation of the HELIOSTATS component is demonstrated by simulating two cases, one textbook example for computing incidence angle and one qualitative example for computing the reflected power during a day for two reflectance coefficient values.

28

C. Kalathakis et al. / Simulation Modelling Practice and Theory 70 (2017) 21–35 Table 1 Difference between example and computed incidence angle.

Incidence angle [deg]

[29]

Model

18

18.002

900 Qrefl @100%

Reflected Power [W]

800

Qrefl @60%

700

Id

600

800 700

600

500

500

400

400

300

300

200

200

100

100

Id [W/m2]

900

0

0 0

5

10 15 Time of the Day [hours]

20

400

0.90

350

0.80

300

0.70

250

0.60 0.50

200

Data

150

0.40

Predictions % Difference

0.30

100

Data Predicons % Difference

0.20 0.10

50 0

% Difference

Value

Fig. 6. Heliostat reflected thermal power over a period of one day.

ECO Tout [oC] 297.78 297.85 0.02

SH Tout [oC] 377.22 377.97 0.20

m [kg/s] 38.969 39.3 0.85

0.00

Fig. 7. Difference between data and computed outlet temperature and water mass flow.

The first case concerns the simulation of the Example in [29] for computing the incidence angle (the dominant loss mechanism) of a heliostat positioned 50 m N and 100 m E of the tower with aiming point 200 m above its height and for azimuth and solar altitude angles of 180° and 50° respectively. Computed and reference incidence angle values are shown in Table 1. The second case is a qualitative validation of the performance, during the summer solstice, of a mirror of 1m2 positioned at distance of 75% of tower height on the North axis for reflectance coefficient values of 100% and 60%. The reflected thermal power (Qrefl ) for the two reflectance coefficient values and the direct irradiation (Id ) are shown in Fig. 6. As it was expected, reflected power decreases as the reflectance coefficient decreases.

3.4. HEAT EXCHANGERS test case The operation of the HEAT EXCHANGERS component is demonstrated by simulating the performance of HTF/water heat exchangers for steam production of the SEGS plant as reported in [41]. Scope of this simulation is the comparison between the computed and reported values of the water/steam mass flow and of the outlet HTF temperature (Tout) of each heat exchanger (Economizer – ECO, Evaporator – EVA and Super Heater – SH). Inlet water and HTF states, the pressure drop for each heat exchanger and the temperature of the produced steam were used as inputs to the model. The reported value of pinch point temperature was used (the Evaporator outlet HTF temperature) while, due to the lack of reported value of approach temperature, the economizer was considered to produce saturated water. Fig. 7 depicts the aforementioned data and computed values and their difference.

400 350

Value

300

Data Predictions % Difference

250 200 150 100 50 0

Tout Tout Tout Tout Tout Tout [oC] [oC] [oC] [oC] [oC] [oC] Data 91.26 121.98 171.61 222.62 270.84 369.67 Predicons 91.18 121.43 171.26 223.14 271.25 370.69 % Difference -0.09 -0.45 -0.20 0.23 0.15 0.28

0.40 0.30 0.20 0.10 0.00 -0.10 -0.20 -0.30 -0.40 -0.50

29

% Difference

C. Kalathakis et al. / Simulation Modelling Practice and Theory 70 (2017) 21–35

Fig. 8. Difference between data and computed HTF outlet temperature.

C Β Α

Fig. 9. Experimental installation of DSG by troughs [43].

3.5. TROUGHS test case The operation of the TROUGHS component is demonstrated by simulating two test experiments, one for HTF heating and one for Direct Steam Generation. The first one has been performed by Sandia National Laboratories on a LS-2 collector on a full tracking platform [42]. The appropriate geometrical data for the collector, the ambient conditions and temperatures of HTF were taken from [42]. Full tracking collector mode was set on the component and receiver’s length division in ten segments was chosen. HTF and ambient conditions are used as inputs and the model outlet temperature for the HTF (Syltherm® 800) is compared with the measured one. Deviations below 0.5% were found for the entire temperature range of the experiment. Indicative results are presented in Fig. 8. The second one concerns the Direct Steam Generation (DSG) by troughs [43]. The appropriate geometrical data for the collector, the ambient conditions and inlet water characteristics were taken from [43]. Water, pipe roughness and ambient conditions are used as inputs and the model values for temperature (T), pressure (P) and mass flow (m) for each position (A, B and C as noted in Fig. 9), are compared with the measured ones. The results are presented in Fig. 10. Differences are below 0.4%. 3.6. TES test case The operation of the TES component is demonstrated by simulating the performance of a storage tank with molten salts as reported in [44]. Their analysis concerns several days, which differ only on the values of inlet and outlet flow and thermal losses. Thus, an indicative time period between 188 and 189 day is chosen for simulation. Input variables to the model are the inlet temperature, the charging and discharging mass flows and the thermal losses for the case of 0.5 m insulation. Zero time is chosen as the time when the charging starts for the aforementioned period. Fig. 11 shows the computed and data values of the stored mass temperature in accordance with time. Because the data values are extracted

C. Kalathakis et al. / Simulation Modelling Practice and Theory 70 (2017) 21–35

0.40

400 Data

350

Value

300 250

Predictions

0.30

% Difference

0.20

200

0.10

150

0.00

100

% Difference

30

-0.10

50 0

PA PB TA [oC] [bar] [bar] Data 75 290 73 Predicons 74.92 290.5 72.97 % Difference -0.11 0.17 -0.04

mB [kg/s] 1.1 1.101 0.09

PC TC [oC] [bar] 71.7 362 71.9 363.33 0.28 0.37

-0.20

Fig. 10. Difference between data and computed temperature, pressure and water mass flow.

844 Reference

Model

-0.2%_Ref

+0.2%_Ref

Temperature [K]

842 840

838 836 834 832 830 0

20000

40000 Time [sec]

60000

Fig. 11. Computed and data storage temperature in accordance with time.

Sun Heliostat Field

Receiver

Combustor Compressor

Turbine

Fig. 12. Solar hybrid gas turbine scheme.

by digitization and are subjected to error, their values for ± 0.2% deviation are also presented. As it can be seen, reference and model curves have the same behavior and their values difference is below 0.2%. 4. Solar hybrid gas turbine installation In order to demonstrate the cooperation of library components for building the model of an STPP in PROOSIS, the installation of Fig. 1 (a solar Brayton cycle) is materialized. Components from TURBO and SOLAR libraries are used, to build the plant’s schematic (Fig. 12), and connected with each other through their ports simulating the real world working fluid flow direction. After having the schematic built, performance simulation for various operating scenarios can be performed.

C. Kalathakis et al. / Simulation Modelling Practice and Theory 70 (2017) 21–35

31

50

Percentage Difference [%]

40 30 20 10 0

ΔE

-10

ΔF

Δeff_f

-20 -30 -40

Fig. 13. Annual difference on performance between solar hybrid and fuel only engine.

Percentage Difference [%]

20 0 -20

ΔP -40

ΔWF

-60 -80 -100 -120

0

5

10

15

20

Time of the day [hours] Fig. 14. Percentage difference of fuel consumption and power production for one day between solar hybrid and fuel only engine.

In the demonstrated case, a single shaft gas turbine designed for fuel only operation has been modified in order to add a solar receiver before the combustion chamber to preheat the air by solar thermal power. The modified combustion chamber is assumed to have the same aerothermodynamic behavior, in terms of efficiency versus combustor loading and pressure losses, as the non-modified one. For the design point of the fuel only engine certain restrictions are taken into account. The maximum allowable temperature at the receiver outlet is 10 0 0 °C [28]. This limit is chosen as the design TIT of the fuel only engine in order to achieve high solar share. It is assumed that the engine is designed for maximum thermal efficiency. Compressor pressure ratio at design point is chosen to respect the maximum receiver’s allowed pressure (about 10 bar [28]) and to satisfy the assumption of maximum efficiency. For the chosen TIT and respecting the pressure limit imposed by the receiver, the maximum efficiency is shown for pressure ratio 10. Furthermore, an added pressure drop of 2% is assumed due to receiver addition. It is assumed that the solar hybrid engine is installed at the top of the receiver tower, so no piping is needed for linking the combustion chamber with the receiver. Engines of 5 MW nominal power are considered to be of a suitable size for this purpose and thus chosen for our study. The number of mirrors (600 of 64 m2 aperture each) in the heliostat field is chosen such that the receiver can heat the air at the maximum temperature for direct irradiance of 600 W/m2 at noon of a summer day. For such conditions no fuel is added to the cycle and the engine operates solely on solar thermal power. Environmental conditions are chosen to represent an area with suitable radiance for CSP implementation. For our test case, we have chosen a location in the south of the island of Crete, in Greece. The chosen operation scenario is to constantly request the maximum allowed power production and thus, the TIT is chosen to be the maximum allowed (10 0 0 °C). The annual percentage difference of produced energy ( E), consumed fuel ( F) and fuel efficiency ( eff_f) between the solar hybrid and the fuel only engines are depicted in Fig. 13 while in Fig. 14 the percentage difference of produced power ( P) and fuel consumption ( WF) are depicted for a single day.

32

C. Kalathakis et al. / Simulation Modelling Practice and Theory 70 (2017) 21–35

The use of solar thermal power for air preheating reduces the amount of fuel required for obtaining the desired TIT. At high irradiation, the engine can even operate solely on solar power. This is reflected on fuel consumption, which becomes zero in such a case. Furthermore, a reduction of produced energy is noticed even when TIT is kept at its maximum value. This is caused by two factors: First, the added receiver pressure losses lead to reduction in produced power and thus the annual produced energy. Second, the composition of the working fluid changes. For a fuel only engine, after the combustion chamber the working fluid is gas with a fuel to air ratio of 2%. In a solar hybrid gas turbine, this ratio is lower since the fuel consumption is lower. It becomes zero for high irradiation. This composition change results to lower enthalpy at the turbine entrance and consequently reduction of produced power. For more information about the use of the Toolbox for performing studies of Solar Power Plants, the reader is referred to the already mentioned authors’ publications [18,19].

5. Toolbox features and functionalities The Toolbox presented here allows the construction of a solar power plant model by combining the appropriate components. Complex configurations, consisting of a large number of different components can be handled fast, in a user friendly manner. The fact that it operates within the PROOSIS environment allows the implementation of all features available for thermal power plants, making it very flexible and suitable for both design and performance studies. A model can be configured as an input-output box that can be placed under an optimization algorithm. Optimization algorithms are available within the environment itself, while the user has the option to produce object libraries for use with other optimization packages, if desired. Appropriate selection of input and output variables as well as objective functions, allows automated studies for optimizing the design to obtain certain targets. For example maximum efficiency, maximum output or combinations thereof can be sought. Furthermore, easiness of layout construction and quick performance assessment permit the easy comparison of alternative candidate solar configurations. An example of such a case was presented in [19], where the authors compared trough and tower technologies for Direct Steam Generation in a STIG Brayton cycle. It was found that the tower configuration shows better annual performance while less affected by seasonal changes. An existing plant can be modeled for operating scenarios at various conditions. Optimum performance studies can then be performed, with functionalities similar to the abovementioned design optimization. Additionally, such performance models are indispensable in supporting the plant operator, when bidding in electricity markets. A plant model is a key tool for assessing the effect of system malfunctions or alteration of the condition of its parts. Component models are built in manner that reflects the physical condition of the corresponding parts. Alteration caused by faults in individual plant parts can thus be simulated. For example, mirror soiling effect can be simulated by changing the reflectivity coefficient’s value, heat exchanger fouling by change loss and heat transfer coefficients etc. It should be noticed that this ability is already contained in the existing libraries of PROOSIS, for the thermal power plant components. Models can thus be used either directly for plant monitoring or for producing information that couples them to general fault assessment algorithms, allowing the constitution of plant condition and fault diagnostic schemes. A useful feature of the toolbox is that it allows the user to implement component models different from those already built-in. This can be done either by intervening in the existing component software or by building new components, added to the existing ones. The user can thus improve model accuracy, if needed. Furthermore, a higher fidelity level modeling of a component can be implemented easily in an existing configuration, by appropriately replacing the desired component (the authors have demonstrated higher fidelity implementation is this environment, for turbine engine applications [e.g. [45]). The Toolbox environment operates presently under WindowsTM . Current day personal computers allow fast execution times (for example, the execution time for one operating point of the hybrid gas turbine configuration of Section 4, is 10 ms on a PC with a dual core, 2.5 GHz processor).

6. Conclusions Asolar power components toolbox has been developed in an object oriented environment (PROOSIS). It consists of a library of components, which can interact with each other and with components from other libraries, as for example those for modeling conventional thermal power plants, for building Solar Thermal Power Plant (STPP) models. The nature of PROOSIS and its libraries allows performance simulation of any practical STPP, existing or conceptual. Implementation of individual component modeling to cases obtained in the literature was used to demonstrate the suitability for reliable modeling of the corresponding physical components. An example case of Solar Hybrid Gas Turbine Installation has been demonstrated and sample information that can be generated from such models has been presented. The results highlight the benefits that may accrue, with respect to fuel consumption and efficiency, by implementing hybridization to existing gas turbines. The ability of the models built though the toolbox to support plant design studies and optimization was discussed, as well as the capabilities they offer to the plant operator for performance management, condition monitoring and fault diagnosis.

Appendix Name

Image

Input variables Ports

Output variables Global var.

SUN

Data

Ports

day, time, site coordinates, standard meridian

Global var.

Vars.

TOWER RECEIVER (AIR)

Fluid, HeliPwr

ambient temperature

absorptance, optical efficiency, emissivity, glass Fluid window aperture, outlet temperature limit

TOWER RECEIVER (WATER/STEAM)

WaSt, HeliPwr

ambient temperature

absorptance, emissivity, window aperture, outlet temperature limit

WaSt

TOWER RECEIVER (HTF)

HTF, HeliPwr

ambient temperature

absorptance, emissivity, window aperture, outlet temperature limit, HTF type

HTF

TOWER RECEIVER (DFR)

Fluid, WaSt, HeliPwr

ambient temperature

absorptance, optical efficiency, emissivity, window aperture, outlet temperature limit

Fluid, WaSt

C. Kalathakis et al. / Simulation Modelling Practice and Theory 70 (2017) 21–35

solar zenith and azimuth, day, declination and hour angles, (DNI if not available)

solar azimuth and zenith reflectivity, heat interception, shading and HeliPwr angles, DNI blocking coefficients, mirror aperture and height, tower aiming point height, number of mirrors, (coordinates of each mirror)

HELIOSTATS

HEAT EXCHANGERS (AIR/GAS)

Fluid, Fluid

UA and streams mass flows at design

Fluid, Fluid

HEAT EXCHANGERS (HTF/WATER)

HTF, WaSt

UA and streams mass flows at design, HTF type HTF, WaSt (continued on next page)

33

34

Appendix (continued) Name

Image

Input variables

Output variables Global var.

Data

Ports

TROUGHS (HTF)

HTF

solar zenith, declination and hour angles, DNI, ambient pressure, temperature and humidity, latitude, wind speed

Trough: number of trough rows, number of HTF troughs in each row, number of parts to divide the trough length, HTF type Mirror: shadowing, tracking error, geometry error, unaccounted error and reflectance coefficients, width and length, slope if static mirror Envelope: absorptance, emissivity, thermal conductivity coefficient, annulus pressure, inner and outer diameter Absorber: material, absorptance, transmittance, emissivity, inner and outer diameter Bracket: perimeter, effective diameter, minimum cross-sectional area, conduction coefficient

TROUGHS (WATER/STEAM)

WaSt

solar zenith, declination and hour angles, DNI, ambient pressure, temperature and humidity, latitude, wind speed

Trough: number of trough rows, number of WaSt troughs in each row, number of parts to divide the trough length Mirror: shadowing, tracking error, geometry error, unaccounted error and reflectance coefficients, width and length, slope if static mirror Envelope: absorptance, emissivity, thermal conductivity coefficient, annulus pressure, inner and outer diameter Absorber: material, absorptance, transmittance, emissivity, inner and outer diameter Bracket: perimeter, effective diameter, minimum cross-sectional area, conduction coefficient

TES

HTF

ambient temperature

Design stored mass, design thermal loss, added HTF thermal power, HTF type

MONITOR

Info

Global var.

Vars.

C. Kalathakis et al. / Simulation Modelling Practice and Theory 70 (2017) 21–35

Ports

thermal efficiency, fuel thermal efficiency, solar share, specific power

C. Kalathakis et al. / Simulation Modelling Practice and Theory 70 (2017) 21–35

35

References [1] D. Kearney, Solar electric generating stations (SEGS), IEEE Power Eng. Rev. 9 (1989) 4–8, doi:10.1109/MPER.1989.4310850. [2] S.D. Odeh, G.L. Morrison, M. Behnia, Modelling of parabolic trough direct steam generation solar collectors, Sol. Energy. 62 (1998) 395–406, doi:10. 1016/S0 038-092X(98)0 0 031-0. ´ ´ J.M. Martinez-Val, Thermofluidynamic model and comparative analysis of parabolic trough collectors using oil, water/steam, [3] M.J. Montes, A. Abanades, or molten salt as heat transfer fluids, J. Sol. Energy Eng. 132 (2010) 21001, doi:10.1115/1.4001399. [4] H. Yamaguchi, X.R. Zhang, K. Fujima, M. Enomoto, N. Sawada, Solar energy powered Rankine cycle using supercritical CO2, Appl. Therm. Eng. 26 (2006) 2345–2354, doi:10.1016/j.applthermaleng.2006.02.029. [5] P. Heller, SOLHYCO_Final Public Report, 2011. [6] EU Commission, SOLGATE: Solar Hybrid Gas Turbine Electric Power System, EU Commission, Directorate-General for Research and Innovation, Luxembourg, 2005. [7] M. Quero, R. Korzynietz, M. Ebert, a.a. Jiménez, a. del Río, J.a. Brioso, Solugas – operation experience of the first solar hybrid gas turbine system at MW scale, Energy Procedia 49 (2014) 1820–1830, doi:10.1016/j.egypro.2014.03.193. [8] L. Aichmayer, J. Spelling, B. Laumert, T. Fransson, Micro gas-turbine design for small-scale hybrid solar power plants, J. Eng. Gas Turbines Power. 135 (2013) 113001, doi:10.1115/1.4025077. [9] J. Spelling, B. Laumert, T. Fransson, A comparative thermoeconomic study of hybrid solar gas-turbine power plants, J. Eng. Gas Turbines Power. 136 (2013) 11801, doi:10.1115/1.4024964. [10] K. Kitzmiller, F. Miller, Effect of variable guide vanes and natural gas hybridization for accommodating fluctuations in solar input to a gas turbine, J. Sol. Energy Eng. 134 (2012) 41008, doi:10.1115/1.4006894. [11] G. Barigozzi, G. Bonetti, G. Franchini, A. Perdichizzi, S. Ravelli, Thermal performance prediction of a solar hybrid gas turbine, Sol. Energy. 86 (2012) 2116–2127, doi:10.1016/j.solener.2012.04.014. [12] N. Blair, A.P. Dobos, J. Freeman, T. Neises, M. Wagner, T. Ferguson, P. Gilman, S. Janzou, System advisor model, SAM 2014.1.14: general description NREL/TP-6A20-61019, Golden, USA. [13] P. Schwarzbözl, A TRNSYS model library for solar thermal electric components ( STEC ) reference manual Köln, Germany. [14] Empresarios Agrupados Internacional, EcosimPro/PROOSIS system modelling and simulation toolkits and services, (n.d.) accessed November 5 accessed November 5. [15] A. Alexiou, Introduction to Gas Turbine Modelling with PROOSIS, 2nd Edition, Empresarios Agrupados Internacional, Madrid, Spain, 2014. [16] N. Aretakis, I. Roumeliotis, A. Alexiou, C. Romesis, K. Mathioudakis, Turbofan engine health assessment from flight data, J. Eng. Gas Turbines Power. 137 (2014) 41203, doi:10.1115/1.4028566. [17] I. Roumeliotis, N. Aretakis, A. Alexiou, Industrial gas turbine health and performance assessment with field data, in: Ceram. Control. Diagnostics Instrumentation; Educ. Manuf. Mater. Metall., 6, ASME, Seoul, South Korea, 2016 p. V006T05A022, doi:10.1115/GT2016-57722. [18] C. Kalathakis, N. Aretakis, I. Roumeliotis, A. Alexiou, K. Mathioudakis, Investigation of different solar hybrid gas turbines and exploitation of rejected sun power, in: Coal, Biomass Altern. Fuels; Cycle Innov. Electr. Power; Ind. Cogener. Org. Rank. Cycle Power Syst., 3, ASME, Seoul, South Korea, 2016 p. V003T06A017, doi:10.1115/GT2016-57700. [19] C. Kalathakis, N. Aretakis, I. Roumeliotis, A. Alexiou, K. Mathioudakis, Assessment of solar steam injection in gas turbines, in: Coal, Biomass Altern. Fuels; Cycle Innov. Electr. Power; Ind. Cogener. Org. Rank. Cycle Power Syst., 3, ASME, Seoul, South Korea, 2016 p. V003T06A012, doi:10.1115/ GT2016-57272. [20] DOW, DOW chemical company, (n.d.) accessed June 10 accessed June 10. [21] J. Duffie, W. Beckman, Solar Engineering Of Thermal Processes, Fourth, John Wiley & Sons, New Jersey, 2013. [22] Z. Sen, Solar Energy Fundamentals and Modeling Techniques, Springer, London, London, 2008, doi:10.1007/978- 1- 84800- 134- 3. [23] S.A. Kalogirou, Solar Energy Engineering: Processes and Systems, Elsevier Inc., 2009. [24] R.E. Bird, R. Hulstrom, A simplified clear sky model for direct and diffuse insolation on horizontal surfaces, Colorado, USA, 1981. [25] R.E. Bird, C. Riordan, Simple solar spectral model for direct and diffuse irradiance on horizontal and tilted planes at the earth’s surface for cloudless atmospheres, J. Clim. Appl. Meteorol. 25 (1986) 87–97 doi:10.1175/1520-0450(1986)025<0087:SSSMFD>2.0.CO;2. [26] P. Heller, M. Pfänder, T. Denk, F. Tellez, A. Valverde, J. Fernandez, A. Ring, Test and evaluation of a solar powered gas turbine system, Sol. Energy. 80 (2006) 1225–1230, doi:10.1016/j.solener.2005.04.020. [27] J. Spelling, B. Laumert, T. Fransson, Optimal gas-turbine design for hybrid solar power plant operation, J. Eng. Gas Turbines Power. 134 (2012) 92301, doi:10.1115/1.4006986. [28] A.L. Ávila-Marín, Volumetric receivers in solar thermal power plants with central receiver system technology: a review, Sol. Energy. 85 (2011) 891–910, doi:10.1016/j.solener.2011.02.002. [29] W. Stine, M. Geyer, Power from the sun, (n.d.) accessed January 23 accessed January 23. [30] F.J. Collado, Quick evaluation of the annual heliostat field efficiency, Sol. Energy. 82 (2008) 379–384, doi:10.1016/j.solener.2007.10.007. [31] B. Kistler, A user’s manual for DELSOL3, 1986. [32] R. Forristall, Heat Transfer analysis and modeling of a parabolic trough solar receiver implemented in engineering equation solver, 2003. [33] V.V. Klimenko, A generalized correlation for two-phase forced flow heat transfer, Int. J. Heat Mass Transf. 31 (1988) 541–552, doi:10.1016/ 0 017-9310(88)90 035-X. [34] V.V. Klimenko, A generalized correlation for two-phase forced flow heat transfer—second assessment, Int. J. Heat Mass Transf. 33 (1990) 2073–2088, doi:10.1016/0017-9310(90)90110-G. [35] R. Balasubramaniam, E. Ramé, J. Kizito, M. Kassemi, Two phase flow modeling: summary of flow regimes and pressure drop correlations in reduced and partial gravity, Cleveland, Ohio, US, 2006. [36] A. Mills, Heat and Mass Transfer, CRC Press, 1995. [37] S.M. Flueckiger, B.D. Iverson, S.V. Garimella, J.E. Pacheco, System-level simulation of a solar power tower plant with thermocline thermal energy storage, Appl. Energy. 113 (2014) 86–96, doi:10.1016/j.apenergy.2013.07.004. [38] NREL, SAM manual-help, (n.d.). https://www.nrel.gov/analysis/sam/help/html-php/index.html. [39] NREL, SPCTRAL2, (n.d.). [40] B. Dickey, Test results from a concentrated solar microturbine brayton cycle integration, in: Vol. 3 Control. Diagnostics Instrumentation; Educ. Electr. Power; Microturbines Small Turbomachinery; Sol. Brayt. Rank. Cycle, ASME, 2011: pp. 1031–1036. doi:10.1115/GT2011-45918. [41] A. Patnode, Simulation and Performance Evaluation of Parabolic Trough Solar Power Plants, UNIVERSITY OF WISCONSIN-MADISON, 2006. [42] T.A. Moss, D.A. Brosseau, Final test results for the Schott HCE on a LS-2 collector., Albuquerque, NM, and Livermore, CA (United States), 2005. doi:10.2172/958386. [43] E. Zarza, M.E. Rojas, L. González, J.M. Caballero, F. Rueda, INDITEP: the first pre-commercial DSG solar power plant, Sol. Energy. 80 (2006) 1270–1276, doi:10.1016/j.solener.2005.04.019. [44] S. Torras, C.D. Pérez-Segarra, I. Rodríguez, J. Rigola, A. Oliva, Parametric study of two-tank TES systems for CSP plants, Energy Procedia 69 (2015) 1049–1058, doi:10.1016/j.egypro.2015.03.206. [45] I. Roumeliotis, A. Alexiou, N. Aretakis, G. Sieros, K. Mathioudakis, Development and integration of rain ingestion effects in engine performance simulations, J. Eng. Gas Turbines Power. 137 (2014) 41202, doi:10.1115/1.4028545.