lable at ScienceDirect

Renewable Energy 62 (2014) 424e431

Contents lists avai

Renewable Energy

journal homepage: www.elsevier .com/locate/renene

Technical note

Automatic control strategies for hybrid solar-fossil fuel power plants

Armando Fontalvo, Jesus Garcia, Marco Sanjuan, Ricardo Vasquez Padilla*

Department of Mechanical Engineering, Universidad del Norte, Barranquilla, Colombia

a r t i c l e i n f o

Article history:Received 7 March 2013Accepted 17 July 2013Available online

Keywords:PID controlCascade controlFeedforward controlSolar power plantsHybrid solar power plantsParabolic trough collector

* Corresponding author. Department of MechanicalNorte, Km 5 Via Antigua Pto Colombia, Barranquilla, Cofax: þ57 5 3509255.

E-mail addresses: aefontalvo@uninorte.edu.couninorte.edu.co (J. Garcia), msanjuan@uninorte.eduuninorte.edu.co (R.V. Padilla).

0960-1481/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.renene.2013.07.034

a b s t r a c t

Solar electrical generating systems are a class of solar energy systems which use parabolic trough col-lectors (PTC) to produce electricity from sunlight. In order to provide power production, one of the majorchallenges is to held the collector outlet temperature or steam temperature around of a specified setpoint by adjusting the flow rate of the heat transfer fluid (HTF) within upper and lower bounds. In somecases, an auxiliary heater can be used to provide heat in absence of solar radiation or during cloudy days.This paper presents a comprehensive study of three control schemes proposed to keep the steamtemperature around its set point by adjusting the fuel (propane) and air mass flow rate of the auxiliaryfossil fuel-fired heater. A non-linear dynamic model was developed in SIMULINK� to study the perfor-mance of each control scheme. Variation of controlled and manipulated variables along with the valvesignals is presented for a period of a cloudy day. The results showed that the combination of feedforwardand three level cascade control is the best alternative to track the temperature set point. It was also foundthat a single three level cascade control without feedforward had less oscillations and low fuel con-sumption compared to the others control strategies.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Solar electric generating systems (SEGS) use parabolic troughcollectors technology (PTC) to produce electricity from sunlight.Parabolic trough collectors are long rows of mirrors with a para-bolic shape that concentrates solar energy on an absorber pipe,typically stainless steel tube with a selective absorber surface, thatpasses trough the focus of the parabola. These collectors have asystem that allows them to track the sun by rotating around anorthesouth axis. A thermal oil is used as a heat transfer fluid(HTF), which circulates trough the absorber pipes and is used totransfer the energy gained from the solar radiation to the thermalpower cycle. The HTF leaves the PTCs at a specified outlet tem-perature and after that is pumped to several heat exchangers (SeeFig. 1) where the heat gained is transferred to the working fluid,water or steam, which is used to drive a steam turbine coupled witha generator to produce electricity [1].

In order to provide stable power production one of the majorchallenges is to keep the collector outlet temperature or steam

Engineering, Universidad dellombia. Tel.: þ57 5 3509272;

(A. Fontalvo), jesusmg@.co (M. Sanjuan), rvasquez@

All rights reserved.

temperature near to a specified set point, by adjusting the flow rateof the HTF. The HTF temperature at collector outlet is affected byvariations in the sun intensity, the HTF temperature at collectorinlet and its volumetric flow rate. The ambient conditions, speciallyvariations in ambient temperature and wind speed also influencethe outlet temperature but their influence is small [1].

During recent years, many control methods have been employedin Concentrated Solar Power applications to overcome the prob-lems caused by the intermittent nature of solar radiation [2]. Theuse of PID controllers has been studied but the results have showedthat the traditional ZieglereNichols tuning method for PID pro-duces an unstable closed-loop system [3] and PID controllers withfixed tuning parameters have been usually restricted as backupcontrollers [2]. On the other hand, feedforward controllers are usedin industrial applications to correct the effect caused by externaland measurable disturbances, in fact new generation solar plantswith direct steam generation are implementing PID controllerscombined with feedforward controllers [4,5]. Cascade control,another traditional control technique, is used in solar power plantsbecause it splits the control problem in two scales and two or morecontrol loops, employing a master control loop which controls theprocess output and slaves control loops that measure intermediatevariables and cancel the effects of the disturbances before thecontrolled variable is affected [2,6].

Advanced control schemes have also been implemented [7].Johansen et al. [8] implemented a gain-scheduled pole placement

Nomenclature

Q$

heat transferred [kW]V$

volumetric flow rate [m3/s]Aabs, surf surface area per unit length [m]Cp specific heat capacity at constant pressure

[kJ/kg � K]D diameter [m]Lcol length [m]n number of collectorsT temperature [K]V volume [m3]r HTF density [kg/m3]abs absorbed, absorberamb ambientcol collectorExp expansion vesselfurn furnaceHTF heat transfer fluidin collector inletLoop collector loopout collector outletsurf surface

Fig. 2. Heat transfer in HCE for the simplified model. Adapted from Ref. [12].

A. Fontalvo et al. / Renewable Energy 62 (2014) 424e431 425

control strategy on a pilot-scale solar power plant, the resultsshowed that the gain-scheduled control strategy performs verywell because it can effectively handle the nonlinearities of theplant. Farkas and Vajk [9] presented an internal model-basedcontroller (IMC) designed for the operation of a solar power plantand found that IMC met the quality requirement of the plant con-trol under clear radiation. Fuzzy logic CONTROL (FLC) was firstapplied by Rubio et al. [10], who performed the application of Fuzzylogic control on a distributed collector field (DSC); the controlsystem showed high degree of robustness and performance despiteof the variation of its operating conditions. FLC shows a high per-formance when there is a certain level of uncertainty or when theknowledge of the process operation can be translated into a control

Fig. 1. Schematic diagram of the SEGS VI Solar Thermal Powe

strategy that improves the results achieved by other classicalstrategies [7].

As it was mentioned, all previously described techniques werefocused on controlling the solar collector outlet temperature byvarying the heat transfer fluid (HTF) flow rate (the manipulatedvariable) through the collector field [11] but no auxiliary backupwas considered. On the other hand, modeling the solar collectorfield with an auxiliary fossil fuel-fired heater creates an additionalmanipulated variable that could increase the performance oftraditional control schemes. This paper presents a comprehensivestudy of the performance of three different control schemes tocontrol the steam temperature that leaves the boiler heatexchanger by adjusting the fuel and air mass flow rate of theauxiliary fossil fuel-fired heater instead of the HTF flow rate, whichwas considered as a disturbance variable. These control techniqueswere evaluated in order to determine the system robustness andthe best scheme to track the temperature set point when fossil fuel-fired auxiliary heater is used in PTCs solar plants.

2. Solar plant description

Solar thermal power plants are systems for power and elec-tricity generation by employing solar radiation as a thermal source.

r Plant with fossil-fired backup. Adapted from Ref. [12].

A. Fontalvo et al. / Renewable Energy 62 (2014) 424e431426

The solar radiation is absorbed by a heat transfer fluid (HTF), whichis usually thermal oil. Four types of concentrated solar power sys-tems (CSP) can be used: parabolic trough collectors (PTC), linearFresnel reflector system (LF), power tower or central receiver sys-tem (CRS), and dish/engine system (DE) [12].

For the current article, the solar electricity generation system(SEGS) VI was taken as reference. SEGS VI is located at MojaveDesert in California and works with therminol VP1 as HTF and LS-2parabolic trough collectors. The LS-2 solar collector is 50 m longand has an aperture width of 5 m and a concentration ratio of 70:1.PTCs are distributed in a solar field, which is composed of severalrows of single axis tracking that form a loop where flows a fractionof the HTF mass flow rate. SEGS is composed by 50 loops and eachloop has 16 PTCs. Finally, the total area covered by the PTCs is188,000 km2 [13]. A summary of the SEGS Power Plant is presentedin Table 1. A schematic diagram of the solar thermal power plantwith fossil backup is shown in Fig. 1.

As it was mentioned above, the solar field provides the heat forvapor generation. The HTF flows through a heat collector element(HCE), where the incoming solar radiation is concentrated andabsorbed by the thermal oil. HCE typically is composed by astainless steel tube surrounded by a glass envelope. The steel tubeis coated with a selective surface with high absorptivity and lowemissivity (see Fig. 2). The glass envelope is a glass pipe with anti-reflective properties and between the absorber and the glass en-velope, a vacuum enclosure is used to reduce heat losses betweenthe absorber and its surroundings and protect the absorber surfacefrom oxidation [14]. Conventional glass to metal seals and metalbellows is used to obtain the necessary vacuum enclosure andflexibility for the simultaneous thermal expansion between glassenvelope and absorber [15]. After the HTF is heated, the HTF leavesthe collectors and the energy gained is transferred to the powercycle, which is a Regenerative Rankine Cycle [16]. HTF flowsthrough several heat exchangers (HX): preheater, where com-pressed water coming from a closed feedwater heater is heated upuntil saturated liquid condition is reached; boiler where the satu-rated liquid that comes from the preheater is heated until a changeof phase from liquid to vapor occurs; finally, superheater whereadditional energy is added to the steam, bringing it to a super-heated vapor condition. The superheated steam is then expandedthrough the high pressure turbine. Other important component ofthe solar field system is the expansion vessel, whichworks as a heatstorage element, especially at night and cloudy days and also pro-vides space for expansion of the HTF due to the change in volumewhen the HTF is heated up in the solar field to the operatingtemperature [13].

3. Dynamic model of the PTC solar power plant

A simplified model of the power plant with auxiliary heater isshown in Fig. 1. The system is composed by five elements: the solar

Table 1Characteristics of SEGS VI Power Plant Ref. [13]. Thermal gross output, net power andnet electric output were obtained for radiation values of 20 June of 1998 Ref. [1].

Location Mojave desert (California)

HTF Therminol VP1Collector technology LS-2Solar field size 188.000 km2

Startup year 1988Capacity (net) 30 MWDesign solar field supply temperature 390 �CThermal gross output 100e110 MWNet power output 30e35 MWNet electric output 29e30 MW

field, which is an arrangement of 50 loops of 16 PTCs, an auxiliaryheater that supplies the amount of energy to keep the steamtemperature at the desirable value, an expansion vessel that worksas a storage energy system, a heat exchanger that represents thesteam generation system, where water is heated and boiled, andfinally a HTF pump. In order to simulate the solar system, asimplified model based on energy balances combined with heattransfer analysis is performed and differential equations are ob-tained. The model is based on the equations presented by Stuetzleobtained for the SEGS VI located at the Mojave desert in California.A detailed description of the model can be found in Ref. [1].

For the auxiliary heater propane (C3H8) is used as fuel. Propaneis an energy-rich gas which is one of the liquefied petroleum gases(LP-Gas or LPGs) that are found mixed with natural gas and oil.Propane and other liquefied gases, including ethane and butane, areseparated from natural gas at natural gas processing plants, or fromcrude oil at refineries. Although propane accounts for less than 2percent of all energy used in the USA and Canada, it has some veryimportant uses: propane is the most common source of energy inrural areas that do not have natural gas service. This is especiallyuseful in developing countries where there is insufficient infra-structure to transport natural gas, and therefore the effect of usingpropane as fuel in the auxiliary heater need to be studied andevaluated.

3.1. Solar field

After an energy balance on the solar field the following differ-ential equation for the HTF collector outlet temperature isobtained:

rCpVcoldToutðtÞ

dt¼ QgainðtÞ þ Lcol½qabsðtÞ � qambðtÞ� (1)

The energy gained by the HTF is given by the followingexpression:

QgainðtÞ ¼ rCp _VHTFTinðtÞ � rCp _VHTFToutðtÞ (2)

Eq. (1) can be written in terms of HTF between the entrance andthe exit temperature of the solar field:

rCpVcoldToutðtÞ

dt¼ rCp _VHTFTinðtÞ � rCp _VHTFToutðtÞ

þ Lcol�q0absðtÞ � q0ambðtÞ

� (3)

The solar field is composed by several loops, each one of themcomposed by several parabolic trough solar collectors. The totalenergy gained by the HTF is proportional to the volume containedin the solar field (Vcol), which is calculated by Eq. (4):

Vcol ¼p4D2absLloop$nloop (4)

where Dabs is the diameter of the absorber pipe, Lloop is the totallength in one loop, and nloop is the number of loops in the solarfield. The overall heat transfer loss to the environment per unitlength, q0ambðtÞ is given by:

q0ambðtÞ ¼ hambA0abs;surf ½ToutðtÞ � TambðtÞ� (5)

In this equation a constant overall heat transfer coefficient ofhamb ¼ 2.5 W/m2 K is used and the surface area per unit length isA0abs;surf ¼ p$Dabs. It is important to point out that this overall

convective coefficient, hamb, includes radiation and convection [1].

A. Fontalvo et al. / Renewable Energy 62 (2014) 424e431 427

3.2. Expansion vessel

The expansion vessel temperature is determined by the nextexpression [1]:

rCpVexpdTexpðtÞ

dt¼ rCp _VHTFTfurnðtÞ � rCp _VHTFTexpðtÞ (6)

For SEGS VI, the expansion vessel volume is Vexp ¼ 287.7 m3.

3.3. Heat exchanger

The energy balance on the heat exchanger is as follows [1]:

rCpVHEdTinðtÞ

dt¼ rCpðaAþaBÞ _VHTFTexpðtÞ�rCpðaAþaBÞ _VHTFTinðtÞ�LHEqtransf ðtÞ

(7)

where aA ¼ aB ¼ 0.4375 [13] and:

qtransf ðtÞ ¼ hHEðtÞAHE;surf ½ThðtÞ � TcðtÞ� (8)

Th(t) and Tc(t) are the mean temperatures of the HTF and thesteam respectively in the heat exchanger. The temperatures pre-viously mentioned can be calculated from the following equations[1]:

ThðtÞ ¼ 12�TexpðtÞ þ TinðtÞ

�(9)

TcðtÞ ¼ 12½TsteamðtÞ þ TwaterðtÞ� (10)

The heat transfer coefficient, hHE(t), depends on the two flowrate measurements available from SEGS VI [1]:

hHEðtÞ ¼ 74;0002

,

"_VHTFðtÞ_VHTF;0

þ _mðtÞ_m0

#W

m2$K(11)

where _m is themass flow rate of theworking fluid (water or steam).For SEGS VI, the reference flow rate of the HTF is_VHTF;0 ¼ 0:624 m3=s and the reference mass flow rate of theworking fluid is _m0 ¼ 39:9 kg=s [1]. The surface area per unitlength is given by:

AHE;surf ¼ pDHE (12)

The volume of the heat exchanger is calculated by:

VHE ¼ p4D2HELHE (13)

with an assumed diameter of DHE ¼ 1 m and length of LHE ¼ 10 m.The temperature of the steam is calculated from the heat exchangereffectiveness as follows:

εðtÞ ¼ TsteamðtÞ � TwaterðtÞTexpðtÞ � TwaterðtÞ (14)

The Eq. (14) can be formulated as a differential equation in orderto consider a dynamic behavior and assume the heat exchangereffectiveness, ε(t), to be flow rate dependent. The following differ-ential equation is obtained:

dTsteamðtÞdt

¼ 1sε

�TwaterðtÞ � TsteamðtÞ þ εðtÞ�TexpðtÞ � TwaterðtÞ

��(15)

Stuetzle [1] introduced a factor, sε¼ 100 s, to adjust the time

constant of this differential equation to the time constant range of

the entire system in order to avoid a stiff differential equationsystem. The heat exchanger effectiveness is calculated by Ref. [1]:

εðtÞ ¼ 1:025� 0:12

"_VHTFðtÞ_VHTF;0

þ _mðtÞ_m0

#(16)

3.4. Auxiliary fossil fuel system

Originally the auxiliary system was not modeled by Stuezle [1].In this paper a simplified model is used to introduce the dynamicbehavior of the backup system. The new energy balance on the HTFis as follows:

rCpVtubesdTfurnðtÞ

dt¼ rCp _VHTFToutðtÞ � rCp _VHTFTfurnðtÞ � qfurnðtÞ

(17)

where qfurn(t) is the heat transferred to the HTF from the flue gasescoming from the fossil-fuel-fired furnace. Zhang et al. [17] devel-oped a linear dynamic model for the furnace. They used full-scalecomputational fluid dynamics (CFD) simulations to generate therequired small signal input and output data sets, and a least squaresbased system identification technique to obtain the linear dynamicmodel. The transfer function for the heat as a function of the fuelmass flow rate is:

GQ_mfuel

ðsÞ ¼ 21524sþ 0:6363

(18)

where _mfuel is the mass flow rate of fuel, propane (C3H8) in thiscase, in kg/s and the heat is in kW. On the other hand, the transferfunction for the heat as a function of the air mass flow rate is:

GQ_mair

ðsÞ ¼ �12:92ðsþ 0:0701Þðsþ 0:42Þðsþ 0:0819Þ (19)

where _mair is the mass flow rate of air, in kg/s and the heat is in kW.For the power plant system, the mass flow rate of fuel was calcu-lated for the mean absorbed radiation of 1667 W/m, which corre-sponds to the maximum value of absorbed radiation measured inDecember 20 of 1998 for the SEGS VI solar power plant [1]. Theinformation of solar irradiation and ambient temperature areshown in Fig. 3.

4. Proposed control strategies

The solar energy power plants have to deal with the problem ofthe intermittent solar radiation, especially during cloudy days.Automatic control strategies are employed to keep the steamtemperature near its set point or desired temperature. Further-more, the collector outlet temperature should not exceed 85 �Cabove the temperature at collector inlet in order to keep an opti-mum power requirement in the solar field [18]. The HTF is exposedto several disturbances: ambient temperature fluctuations, windspeed and, finally, the variation of solar radiation, due to thepresence of clouds and different seasons throughout the year.

Fig. 4 shows the traditional control scheme using a PID-feedbackcontroller. Although several strategies have been employed foradvanced control, they have in common that the HTFmass flow rateis the manipulated variable, which is used to control the steamtemperature or the HTF temperature after passing trough the PTCs.An alternative strategy to control these two temperatures is

Fig. 3. a) Ambient temperature and b) Solar radiation absorbed for December 16th of1998. Adapted from Ref. [1].

A. Fontalvo et al. / Renewable Energy 62 (2014) 424e431428

introducing an auxiliary heater after the PTC outlet and a two levelcascade (alternative 1): two slaves PID controllers which controland manipulate the fossil fuel flow rate and the air flow rate and amaster controller which controls the steam temperature byadjusting the air flow and fuel flow set point. This alternativecontrol strategy is shown in Fig. 5.

Fig. 4. Traditional control strategy with PID controller without the auxiliary heater.

Twomore control schemes are considered for the current study:a three level cascade control (alternative 2) and a three levelcascade control plus a feedforward controller (alternative 3). Thethree level cascade control scheme arises from the fact that the HTFat the furnace outlet changes before the steam temperature and itcan be measured and gives feedback to a PID controller that adjustthe set point of the two air flow and fuel flow controllers. Themaster control receives the steam temperature measurement fromthe sensor-transmitter and if there is a deviation from the set point,it adjusts the set point of the furnace temperature controller, thatadjust the air flow and fuel flow set point and change the air andfuel flow rate. This alternative is shown in Fig. 6.

The last control scheme proposed is shown in Fig. 7. The threelevel cascade control plus a feedforward controller improves theperformance of the controller by the compensation of the radiationdisturbance before they affect the controlled variable by measuringthe incoming solar radiation.

5. Simulation details

The dynamic behavior of the solar power plant was performedin Simulink� tool under the MATLAB� environment. Simulink� is asimulation tool based on a functional approach. The principle isbased on the connection of several subsystems represented bytransfer functions. This approach allows access to various inter-mediate parameters with tools for displaying and storing pre-defined in the standard library. For the simulation of the simplifiedmodel, a block diagram model was built. The tuning parameters ofthe PID controllers employed in the control schemes mentionedabove were calculated to produce a specific closed loop responsewith 0% overshoot for set point changes [19]. Table 2 summarizesthe parameters considered for the simulation.

6. Results and discussion

Fig. 8 shows the different strategies previously proposed. Ac-cording to its ability to keep the process near the set point tem-perature. Alternative 3 (three level cascade control plus afeedforward controller) has the best performance since the steamtemperature was between 638.2 and 638.7 K. The main concernabout this strategy is that oscillations around set point are pre-sented when solar radiation and climate conditions changes,because of the feedforward controller strategy.

The second best strategy is achieved when a three level cascadecontrol scheme (alternative 2) is considered with PID controllers,that was also considered in alternative 2. This strategy does notperform as well as the one with a feedforward controller because ithas not knowledge of the solar radiation effect before it enters tothe process. Finally, alternative 1, where a two level cascade controlwith PID controllers is applied with fossil backup, had the lowestperformance since the controlled temperature varies between622.3 and 644.1 K. All alternatives presented have a better perfor-mance that the traditional control scheme, which manipulates theHTF mass flow rate, so the use of fossil fuel-fired auxiliary heaterincreases the robustness of the system. However, the traditionalcontrol scheme allows to save cost because it does not involves theuse of fossil fuel.

It can be inferred from the previous results that feedforwardcontrol combined with cascade control improves the performanceof the system and ensure robustness. The single use of feedback tocontrol steam temperature requires that error is present before thatthe controller works. This explains the overshot presented on bothalternatives 1 and 2. However, the use of feedforward and a threelevel cascade control in alternatives 2 and 3 involves more pa-rameters to adjust than the alternative 1, which is more simple.

Fig. 6. Alternative 2: three level cascade control scheme employing PID controllers.

Fig. 5. Alternative 1: two level cascade control scheme employing PID controllers.

Fig. 7. Alternative 3: three level cascade plus FFC control scheme employing PID controllers.

A. Fontalvo et al. / Renewable Energy 62 (2014) 424e431 429

Table 2Simulation parameters.

Parameter Value

Average ambient temperature (K) 258nloop 50Lloop (m) 753.6Dabs (m) 0.066Steam set point (K) 638.5Water inlet temperature (K) 496.6Water mass flow rate (kg/s) 39.9Design fuel mass flow rate (kg/s) 1.024Percent of excess of air 15HTF volumetric flow rate (m3/s) 0.624

A. Fontalvo et al. / Renewable Energy 62 (2014) 424e431430

Therefore, if the overshoot obtained in alternative 1 is acceptable,this control could be the most appropriated due to its lack ofcomplexity.

Fig. 9 shows the signals sent to the air and fuel valves. Based onthe signal output the controller behavior, for all alternatives, is soft.

Fig. 8. Behavior of the controlled variable, the steam temperature, for different controlschemes: a) All control schemes considered, b) Control schemes with fossil-fuelbackup.

Fig. 9. Valve signal for the manipulated variables.

This behavior is because the tuning parameters of the PID con-trollers were calculated to produce a response with 0% overshoot,which it is in essence a soft response. An interesting result is thatthe controller output for each control scheme is almost the same,which it agrees to the response time by the controller. The responsetime is influenced by the period of time considered (17 h, betweenthe 5th and the 22nd hour of the day), the change of the radiationand the process dead times which are in the range of 5e10 minaccording to the differential equations presented below. Fig. 9 alsoreveals that the three level cascade control response was veryclosed to the feedforward controller, despite the fact that betweenthe 9e10th hour, and between the 17e18th hour an overshoot waspresented. Themeasure of the HTF temperature at the heater outlethelps to obtain a fast response, but the effect of solar radiation hasto enter to the process to produce this response because thestrategy considered only involves feedback control.

Fig. 10 showed that the fuel consumption is almost the same forall the alternatives considered. This result is logical because of the

Fig. 10. Variation of manipulated variables, air and fuel flow rates, for the three al-ternatives that considers fossil backup. a) Air mass flow rate b) Fuel flow rate.

A. Fontalvo et al. / Renewable Energy 62 (2014) 424e431 431

distinct behavior of the controller output described below andtypical control valves are very fast compared to the whole systemconsidered.

7. Conclusions

A simplified dynamic model of a solar power plant with anauxiliary heater was presented for the SEGS VI Power Plant. Thesimplified model was used to evaluate the performance of threedifferent control schemes which manipulate the fuel and air flowrate in order to control the steam temperature that leaves the boilerheat exchanger. These schemes combined the use of feedforwardand cascade control. After the simulations were performed thefollowing results were obtained:

� The use of feedforward and a three level cascade controlimproved the performance of the system and ensured robust-ness because of the direct measure of the disturbances: solarradiation for this case. Themeasure of an intermediate variable,the HTF temperature at the heater outlet, helped to obtain afast response.

� The fuel consumption was almost the same for all the controlstrategies proposed and the use of fossil-fuel-fired auxiliaryheater increased the robustness of the system.

References

[1] Stuetzle T. Automatic control of a 30 mwe segs vi parabolic trough plant.Master’s thesis. University of Wisconsin-Madison; 2002.

[2] Camacho E, Rubio F, Berenguel M, Valenzuela L. A survey on control schemesfor distributed solar collector fields. part i: modeling and basic control ap-proaches. Solar Energy 2007;81(10):1240e51.

[3] Meaburn A, Hughes F. Pre-scheduled PID control of a solar thermal powerplant. Transactions of the Institute of Measurement and Control 1995;17(3):132e42.

[4] Zarza E, Valenzuela L, Leon J, Hennecke K, Eck M, Weyers H, et al. Direct steamgeneration in parabolic troughs: final results and conclusions of the DISSproject. Energy 2004;29(5):635e44.

[5] Valenzuela L, Zarza E, Berenguel M, Camacho E. Control concepts for directsteam generation in parabolic troughs. Solar Energy 2005;78(2):301e11.

[6] Silva R, Rato L, Lemos J, Coito F. Cascade control of a distributed collector solarfield. Journal of Process Control 1997;7(2):111e7.

[7] Camacho E, Rubio F, Berenguel M, Valenzuela L. A survey on control schemesfor distributed solar collector fields. part ii: advanced control approaches.Solar Energy 2007;81(10):1252e72.

[8] Johansen T, Hunt K, Petersen I. Gain-scheduled control of a solar power plant.Control Engineering Practice 2000;8(9):1011e22.

[9] Farkas I, Vajk I. Internal model-based controller for a solar plant. In: Preprints15th IFAC triennial world congress, Barcelona, Spain, July 2002. p. 21e6.

[10] Rubio F, Berenguel M, Camacho E. Fuzzy logic control of a solar power plant.Fuzzy Systems, IEEE Transactions on 1995;3(4):459e68.

[11] Silva R, Filatov N, Lemos J, Unbehauen H. Feedback/feedforward dual adaptivecontrol of a solar collector field. In: Control applications, 1998Proceedings ofthe 1998 IEEE International Conference on; vol. 1. IEEE 1998. p. 309e13.

[12] Padilla R. Simplified methodology for designing parabolic trough solar powerplants. Ph.D. thesis. University of South Florida; 2011.

[13] Patnode A. Simulation and performance evaluation of parabolic trough solarpower plants. Master’s thesis. University of Wisconsin-Madison; 2006.

[14] Padilla RV, Demirkaya G, Goswami DY, Stefanakos E, Rahman MM. Heattransfer analysis of parabolic trough solar receiver. Applied Energy2011;88(12):5097e110.

[15] Price H, Lüpfert E, Kearney D, Zarza E, Cohen G, Gee R, et al. Advances inparabolic trough solar power technology. Journal of Solar Energy Engineering,Transactions of the ASME 2002;124(2):109e25.

[16] Montes M, Abánades A, Martínez-Val J, Valdés M. Solar multiple optimizationfor a solar-only thermal power plant, using oil as heat transfer fluid in theparabolic trough collectors. Solar Energy 2009;83(12):2165e76.

[17] Zhang R, Zhang C, Jiang J. A new approach to design a control system for a FGRfurnace using the combination of the CFD and linear system identificationtechniques. Combustion Theory and Modelling 2011;15(2):183e204.

[18] Stine WB, Harrigan RW. Solar energy fundamentals and design. New York:John Wiley and Sons; 1985.

[19] Smith C, Corripio A. In: Principles and practice of automatic process control,vol. 111. N. Y: John Wiley & Sons^eN. Y; 1997.