Performance model for parabolic trough solar thermal power plants with thermal storage: Comparison to operating plant data

TL;DR: In this article, a simulation model that reproduces the performance of parabolic trough solar thermal power plants with a thermal storage system is presented to facilitate the prediction of the electricity output of these plants during the various stages of their planning, design, construction and operation.

About: This article is published in Solar Energy.The article was published on 2011-10-01 and is currently open access. It has received 265 citations till now. The article focuses on the topics: Peaking power plant & Parabolic trough.

Summary

1. Introduction

• Solar power technology has seen great advances over the past decade.
• The world’s first commercial solar tower,PS10, was finished in Seville, Spain, in 2007 and has a capacity of 11 MWe.
• From the available CSP technologies, parabolic trough is the most widespread today, with around 29 plants in operation and around 1220 MWe of installed power in the world, corresponding to 96.3 % of the total operational concentrating solar power as of the beginning of 2011 (see Fig. 1).the authors.the authors.
• The good agreement obtained validates the model and confirms it as an effective tool for predicting the electricity output of these plants.

2. Parabolic trough power plant with thermal storage

• A simplified schematic for a parabolic trough solar thermal power plant with thermal storage is shown in Fig.
• The HTF and water-steam circuits and the HTF and TES circuits can exchange energy at the corresponding heat exchangers.
• The Solar Field consists of a number of parabolic trough collector loops connected in parallel to each other.
• Solar energy can be stored as sensible heat in the thermal storage medium during the hours of high insolation, while it is possible to extract that heat during the hours of low or zero insolation for the production of electricity.
• In this paper the authors will assume that the TES system is a two-tank molten-salt system.

3.1. Algorithm structure

• Fig. 3 shows a simplified schematic with the general structure of the information flow within the simulation algorithm.
• Details of the calculations are given in the next section 3.2.
• The box on the left hand side of the figure summarises the inputs required by the simulation.
• The graph shows the input solar irradiance and results for the useful thermal power delivered by the Solar Field, thermal power sent to the Power Block, stored energy, wasted solar power (see section 3.2.15) and gross electric power generated.
• The algorithm within the main function for a single day divides the calculations into several blocks: a night-time period before sunrise, a start-up period during which the HTF in the Solar Field is warmed up and the Power Block is started, a full-operation period during daylight hours, and a second night-time period after sunset.

3.2. Detailed description of calculations

• The following subsections explain the details of the relevant calculations within the model as outlined in the previous Fig. 3.
• Most of the time, these calculations have been particularised to the actual 50 MWe plant simulated in this paper.the authors.the authors.

3.2.1. Geographical data

• Geographical data for a given site location are required as input to the performance model.
• These include the latitude,ϕ, and longitude,λ, at the site, the longitude of the corresponding time-zone meridian,λTZM, the collector’s orientation angle,γcol (0 rad for NorthSouth orientation andπ/2 rad for East-West orientation) and the collector’s tilt angle,βtilt (angle of the collector’s axis to the local horizontal, 0 rad for horizontal tracking axis).

3.2.2. Meteorological data

• Meteorological data are usually provided in the form of a Typical Meteorological Year (TMY) representative of the chosen location.
• The authors input format is based on the standard TMY3 format (Wilcox and Marion, 2008) and adjusted to any given time step between data points.
• Values of date and time, direct normal solar irradiance (Eb, in W/m2), dry-bulb or ambient temperature (Tamb, in ◦C), wind speed (vwind, in m/s), relative humidity (rh, in %) and atmospheric pressure (pamb, in mbar) can be extracted from the TMY data.
• Typical uncertainties are 10− 20 % for solar irradiation and smaller than variations measured between one year and another.
• The simulation code has been designed to function correctly for any time frequency of the input meteorological data (e.g., data given for every hour, every ten or fifteen minutes, etc.), since the time interval between consecutive data points is automatically detected in the first instances of the simulation.

3.2.3. Calculation of solar time

• The procedure followed to calculate the solar time from the local time provided in the TMY data follows the steps in Stine and Geyer (2001).
• Daylight savings can be taken into consideration by providing a list of specific daylight savings clock change times for a given location and for the years under study.
• Therefore, for any time between these two, the value ofDS(hour,day,month, year) in Eq.(1) would be 1, and 0 for any other time that year.

3.2.4. Calculation of the angle of incidence

• The angle of incidence,θ of the solar radiation on the parabolic trough collectors in the Solar Field is the angle between the direction of the incident radiation and the normal to the collector’s aperture.
• This angle determines the intensity of the radiation incident on the collector’s mirror aperture area.
• Since the directnormal solar irradiance,Eb is by definition measured on a surface normal to its propagation direction, a factor cos(θ) must be considered in order to calculate the total energy that can be focused and concentrated by the collectors onto the receiver tubes.
• The angle of incidence is calculated following the explanations in Stine and Geyer (2001) for a single-axis tracking collector.

3.2.5. Solar Field characteristics

• Calculations in this paper simulate a 50 MWe plant for which the Solar Field consists of 156 collector loops (Nloops = 156), a loop consists of 4 SCAs in series, forming two rows with two SCAs each, and a single SCA consists of 12 SCEs.the authors.the authors.
• The peak optical efficiency is defined as the optical efficiency at zero angle of incidence.
• Several expressions can be found in the literature for this factor.
• The end loss factor,fendLoss, accounts for the reduction of effective mirror aperture area caused by the spacing between SCAs and SCEs within a row of collectors in the Solar Field.
• The Solar Field heat transfer fluid considered in this paper isTherminol VP-1, for which the density, specific heat and specific enthalpy are known functions of the fluid temperature.

3.2.7. Receiver heat losses

• Receiver thermal losses have a relevant impact in the final electricity production of a parabolic trough power plant and therefore need to be modelled as accurately as possible.
• Whereas it is possible to model these losses from first principles using a physical model, in this work the authors have chosen to use the empirical results obtained by NREL through heat-loss testing of several specific commercial receiver tubes.
• Alternatively, the empirical expression found in Burkholder and Kutscher (2008a) is used whenSchott’s 2008 PTR70receiver tubes are part of the Solar Field design.
• Receiver heat losses are given as a function of the ambient temperature, wind speed, direct normal solar irradiance, angle of incidence and tube working condition.

3.2.8. Piping thermal losses

• Solar Field piping heat losses are calculated making use of an empirical equation (Patnode, 2006) derived per unit Solar Field aperture area for the SEGS plants in the USA.
• The thermal loss in Watts as a function of the difference between ambient temperature and average HTF temperature is calculated as: whereNloops is the number of collector loops in the Solar Field, Ac,gross is the gross loop aperture area and T f l,pipes is the average HTF temperature in the insulated pipes.

3.2.9. Calculation of HTF temperatures

• Termine the decision to operate the plant in one configuration or another, as detailed in the following section 3.2.15.
• This pipe circuit model would apply during the initial stage of the HTF warm-up period, when the fluid circulates through the Solar Field bypassing the Power Block heat exchangers.
• The authors can simplify the problem by assuming a linear and discrete approximation for Eq.(19) where they have dT(t)/dt ≈ ∆T/∆t = (T − T0)/∆t, whereT0 andT indicate the HTF temperatures in a given pipe portion at the beginning and end of a time interval∆t, respectively.
• The simulation makes use of three different predefined functions which follow this methodology to find solutions for the HTF temperatures iteratively, using a calculation step of 10 seconds.
• Since the receiver tubes are not insulated, the temperatures at the loop inlets and outlets,T1 andT4, respectively, fall much faster than the HTF temperature in the insulated pipes.

3.2.10. Solar Field useful thermal power

• This limit guarantees that the HTF temperature at the loop outlets does not surpass its maximum design value.
• In an actual plant, control of the HTF loop-outlet temperatures under high solar radiation conditions is realised through the deliberate de-focusing of a number of collectors within a loop, in order to avoid damage to the HTF.
• The fraction of collectors that remain in focus is calculated as the ratio of the thermal power absorbed by a loop after the limitation is applied to the same power before the limit is applied.
• The total HCE losses in a loop are subtracted from the thermal power absorbed by a loop, and the result is then multiplied by the number of loops in the Solar Field.

3.2.11. HTF freeze-protection system

• A freeze-protection system for the HTF in the Solar Field is implemented by considering the thermal power provided by the HTF heaters.
• This system is triggered when the HTF loop outlet temperatures fall below a given value,Tanti f reeze−on, and functions until it reaches a second higher value,Tanti f reeze−o f f .

3.2.12. Power Block model

• A certain amount of thermal power is delivered to the Power Block from the Solar Field or the TES system (or both), for the production of electricity in the turbine and generator.
• For the results presented in this paper, only the dependence on the former has been considered.
• MWt, obtained from a fit of the turbine’s specification data provided by a supplier.
• The actual gross electric power,Pe,gross, generated from a given thermal power,Pt,in,ht f on the HTF side (ηexchPt,in,ht f on the steam side), sent to the Power Block, is therefore given byPe,gross = ηexchPt,in,ht f × ηPB(ηexchPt,in,ht f ) in solar-only mode, and similarly with the corresponding efficiencies, for the storage and mixed operating modes.
• An update of this simplified model is currently under development to include the influence of both ambient temperature and relative humidity on the electricity output from the Power Block.

3.2.13. Thermal Energy Storage (TES) system

• In this paper the authors assume that the TES system is a two-tank molten-salt system (our simulation can include up to three pairs of molten-salt tanks).
• The salts are a mixture of Sodium and Potassium Nitrates with known properties of density and specific heat as a function of temperature.
• During a typical storage charge, the excess thermal power delivered by the Solar Field is sent to the TES circuit, so that part of the HTF passes through the HTFTES heat exchangers in order to transfer and store heat in the TES fluid.
• The model for the TES system performs all calculations in terms of thermal power delivered to or extracted from it.
• This mass flow rate is in turned used to calculate the parasitic electricity consumption of the salt pumps.

3.2.14. Parameter limitations

• A maximum limit of 1100 kg/s for the HTF mass flow rate in the Solar Field as a whole is also enforced, as given by the specifications of the main HTF pumps.
• The maximum thermal power (HTF side) which can be sent to the Power Block from the Solar Field during solar-only operation is considered to bePmaxCS toPB' 140 MWt, which leads to a maximum gross electric power generated of∼ 52.6 MWe, at a maximum Power Block efficiency (ηPB) of 39.5%.
• This leads to a minimum allowable HTF thermal power of∼.
• 21 MWt sent to the TES system during storage charge, calculated using the previous minimum HTF flow rate and the HTF temperatures during storage charge under design conditions (see section 3.2.13).

3.2.15. Simulated plant’s operation strategy

• As outlined in section 3.1, the algorithm that calculates a single day is divided into several blocks: a nighttime period before sunrise, a period for warm-up of the HTF in the Solar Field and start-up of the Power Block, a full-operation period during daylight hours and a second night-time period after sunset.
• During any of thenight-time periods, the algorithm first checks if TES discharge is possible taking into account the state of the TES system and the limitations explained in section 3.2.14.
• The HTF warm-up and turbine start-up periods begin with sunrise and last until the HTF reaches its design Solar Field outlet temperature and the turbine reaches 100% steam input load.
• During theHTF warm-up period, the useful thermal power collected by the HCEs in the Solar Field heats up the HTF in the loops and the electric power output of the plant is zero.
• Once an operating HTF mass flow rate through the Solar Field has been chosen according to the diagram in Fig. 8, the HTF temperatures (in each SCA and in the Solar Field insulated pipes) can be calculated as detailed in section 3.2.9.

3.2.16. Parasitics

• A detailed calculation of parasitic electric consumption can be carried out based on the characteristics of the specific equipment for a given plant.
• Offline parasitics are those taking place when the gross electric power generated is zero, as opposed to online parasitics.
• Solar Field parasitics include main HTF pumps, HTF circulation pumps, tracking and communication system for the Solar Field, HTF system (lubrication, expansion, ullage, HTF heaters, electrical heating).
• Power Block parasitics include condensate pumps, feedwater pumps, water circulation pumps, closed/open loop refrigeration pumps, service water pumps, cooling tower, balance of plant consumption, water treatment plant, auxiliary heaters, compressed air system and electrical losses.
• Simulated results for parasitic consumption are not presented in this paper, since the corresponding actual data was not available for comparison.

3.3. Simulation results and comparison to actual plant data

• Actual data from the plantAndasol 2, operated by the ACS Industrial Group in Granada, Spain, was kindly made available to Initec-Energı́a.
• The actual data used for comparison with the simulation results are described as follows.
• Actual data for the temperature of the HTF into the Power Block, just after the HTF flows out of the Solar Field and TES system are combined, is also available and can at times be compared to the simulated HTF temperature at the output of the Solar Field (calculated Tpipes).
• The difference in mass flow rates and the decision to stop the circulation of HTF in the Solar Field at around 5pmexplain the observed discrepancies between simulated results and actual data towards the end of the day.
• Fig. 12 shows a comparison of the simulated and actual total daily gross electric energy generated by the plant during the 42 days for which data are available.

4. Conclusion and outlook

• The authors have presented a detailed model which simulates the performance of a parabolic trough solar thermal power plant with thermal storage.
• The simulation algorithm is flexible and can be generalised to reproduce the performance of any trough plant of choice, since component specifications and particular equations can be easily adjusted to the characteristics of a new project.
• It would be desirable to carry out a more extended comparison of the simulation results to the actual plant data considering periods within all seasons of the year.
• Only summer data were available for the purpose of this paper.
• Work is currently in progress to develop a less simplified Power Block model which takes into consideration the fact that meteorological conditions affect the Power Block performance substantially.

Figures

Figure 7: Power Block efficiency for conversion of input thermal power into electric power in solar-only mode (solid orange line) and storage mode (blue dashed line) as a function of input thermal power.

Figure 1: Solar thermal power plant projects in operation in the world (March 2011). Left : installed power by country.Right: installed power by technology.

Figure 6: Example of the HTF temperatures calculated during 24 hours.Solid blue line: calculated average HTF temperature in the insulated pipes,Tpipes(Solar Field output).Dashed red line: calculated average HTF temperature in the first SCA of a loop,T1. Dot-dashed green line: calculated average HTF temperature in the fourth SCA of a loop,T4 (loop outlet). The corresponding simulated HTF mass flow rates per loop and operating conditions are those shown in Fig. 10 for July 13th 2010.

Figure 8: Schematic for the decisions taken by the algorithm during the full-operation period on daylight hours. SF stands for Solar Field, PB, for Power Block, TES, for thermal energy storage system, ˙mcalc is the calculated HTF mass flow rate in a loop (see text), ˙mmin is the minimum HTF mass flow rate in a loop,Puse f ulFieldis the useful thermal power captured by the Solar Field after all losses have been subtracted (see section 3.2.10) andPmaxCS toPBis the maximum thermal input to the Power Block on the HTF side during solar-only operation.

Figure 2: Schematic diagram of a parabolic trough solar thermal power plant with thermal storage. In the figure, HX stands for heat exchanger, PH, SG, SH and RH for preheater, steam generator, superheater and reheater, respectively, D stands for deaerator and EV for expansion vessel.

Figure 12: Comparison of simulated (green dashed line) and actual (solid black line) total daily gross electric energy generated by the plant during 42 days, from June 26th to August 6th of 2010.

Figure 5: Heat loss in W per unit length of receiver tube as a function of average HTF temperature,T f l , for evacuated Schott’s 2008 PTR70 heat collection elements (Burkholder and Kutscher, 2008a). Solid red curves use an average HTF temperature assuming that there is a temperature jump of 100◦C between input and output HTF temperatures (Tin = T f l − 50◦C andTout = T f l + 50◦C), whereas dashed blue lines assume a constant HTF temperature throughout the receiver (Tin = T f l − 0.1◦C andTout = T f l ). Both assumptions yield similar results. Lower curves correspond to an ambient temperature of 40◦C, middle ones to 20◦C and higher curves to 0◦C. Receiver heat losses are shown to increase with increasing HTF temperature and decrease with increasing ambient temperature.

Figure 4: Example of simulation results for a summer day (July 6th) for the plant considered in this paper.

Figure 3: Schematic representation of the information flow within the performance model algorithm.

Figure 11: Comparison of actual data and simulated results for June 29th (left) and July 8th (right), 2010. The legend is shared for plots on the left and right hand sides. Note that the zero reading for the actual HTF temperature at 10amon June 29th was caused by a technical fault.

Figure 10: Comparison of actual data and simulated results for three days: 13th, 14th and 15th of July 2010. Ambient temperature during these days was between 13◦C and 22◦C, wind speed between 0 m/s and 6 m/s and relative humidity between 13 % and 47 %.

Figure 9: Comparison of actual data and simulated results for three days: 26th, 27th and 28th of June 2010. Ambient temperature during these days was between 13◦C and 26◦C, wind speed between 0 m/s and 7 m/s and relative humidity between 22 % and 73 %.

Frequently asked questions

Q1. What have the authors contributed in "Performance model for parabolic trough solar thermal power plants with thermal storage. comparison to operating plant data" ?

This paper describes a simulation model that reproduces the performance of parabolic trough solar thermal power plants with a thermal storage system. The aim of this model is to facilitate the prediction of the electricity output of these plants during the various stages of their planning, design, construction and operation. Model results for a 50 MWe power plant are presented and compared to real data from an equivalent power plant currently operated by the ACS Industrial Group in Spain. 

Q2. What are the future works in "Performance model for parabolic trough solar thermal power plants with thermal storage. comparison to operating plant data" ?

The authors have presented a detailed model which simulates the performance of a parabolic trough solar thermal power plant with thermal storage. The authors point out the need to consider the actual plant ’ s operation philosophy in order to match the decisions taken by the simulation algorithm to the real ones during operation. It would be desirable to carry out a more extended comparison of the simulation results to the actual plant data considering periods within all seasons of the year. The authors estimate that the gross electrical power generated could vary around 2 − 3 % if they consider the ranges of ambient temperature and relative humidity corresponding to the results shown in section 3. 3. The HTF temperature out of the Power Block heat-exchanger trains and into the Solar Field is also affected by ambient conditions, so that they can expect to better reproduce the actual plant data making use of a more detailed Power Block model.