simulation of low-intensity, low-temperature solar arrays with software … · 2015. 12. 11. ·...

Aerotecnica Missili & Spazio, The Journal of Aerospace Science, Technology and Systems

Simulation of Low-Intensity, Low-Temperature Solar Arrays withSoftware and Hardware Tools∗

F. Topputoa and F. Bernelli-Zazzeraa

aPolitecnico di Milano, Dept. of Aerospace Science and TechnologyVia La Masa, 34 – 20156, Milano, Italy

Abstract

In this work we discuss issues related to the simulation of low power systems with hardware means. Simulating lowpower systems is a challenging task as the possible low-intensity, low-temperature environment, together with possibledust deposition and ice condensation, worsen not only the production of power but also make it difficult to predict it. Toovercome these problems, we have developed solutions in terms of software and hardware tools for power estimation andsimulation. The developed low power, hardware solar array simulator system is briefly discussed in this paper. Althoughthis solution is reported for the case of Rosetta lander Philae, it applies also to possible low power future missions aimedto perform in-situ operations on comets and asteroids.

1. Introduction

Rosetta is the third of ESA’s cornerstone missionswithin the science program Horizon 2000. The ambi-tious goal of this mission is the injection into an orbitaround a comet and the delivery of a lander that willperform a detailed in-situ investigation of a comet nu-cleus for the first time. The prime scientific objectiveis to study the comet 67P/Churyumov-Gerasimenkoto help understand the origin and the evolution of theSolar System. The comet will be reached in 2014 af-ter a journey involving several close encounters withminor and major bodies. After a phase of close cometinvestigation, a safe and scientifically important sitewill be selected for in-situ investigations. The landerdelivery is foreseen in November 2014 at a distance ofabout 3 AU from the Sun.

The Rosetta probe is made up by two spacecraft:an orbiter, Rosetta, with 11 scientific instruments on-board, which will orbit the comet 67P/Churyumov-Gerasimenko, and a lander, Philae, with 10 scientificinstruments, which will land on the comet to performin-situ analysis. This large array of instruments willperform the most extensive study of a comet to date.Philae can be considered as an independent spacecraft,although it is the main payload of Rosetta. Philaehas a mass of 97.9 Kg including 26.7 Kg for scientificpayload.

At the time this paper is being written, Rosetta isapproaching the comet 67P, at about 5 AU from the

∗Based on paper presented at the XXII CongressoNazionale AIDAA, Settembre 2013 Napoli, Italia1 c©AIDAA, Associazione Italiana di Aeronautica e Astronautica

Sun. Both Rosetta and Philae have been placed intohibernation due to the weakness of the sunlight, whichprevents from producing enough power to fully operatethe probe. The wake-up is foreseen for January 2014,six months before the rendez-vous with the comet.

Politecnico di Milano is involved in the Rosetta mis-sion through the activities on the lander Philae. Inparticular, the Aerospace Engineering Department isresponsible for the activities of the solar generator andthe instrument SD2 (Sampler Drill and Distributionsubsystem [1]).

2. Philae Power Subsystem

The Philae Power Subsystem (PSS) manages all theelectrical power needed by the lander during its entirelifetime. The main source of power is represented bythe primary and secondary batteries, with capacity of1000 Wh and 130 Wh at comet arrival (1200 Wh and150 Wh at launch, respectively). The primary batterywill allow the operations of Philae during the mainscientific phase (about 5 days) just after the landingon the comet, during which all instruments will beoperated at least once. The secondary battery will berecharged with the power produced by the solar arrays.

The power distribution is basically performed usingthe Philae Primary Bus. The main subsystems, theCommand Data Management System (CDMS) andthe Thermal Control Units (TCU), are directly con-nected to the primary bus via dedicated DC-DC con-verters. The other subsystems and all the experimentsare connected through switches to the Primary Bus ordirectly to the Secondary Bus, which is stabilized. The

100

Simulation of Low-Intensity, Low-Temperature Solar Arrays with Software and Hardware Tools 101

Wake-Up System provides for the exact and safe startof the Philae operations. It monitors if the tempera-ture and the available power are in the expected rangesbefore switching power to start the system.

Rosetta is the first deep-space mission that will gobeyond the main asteroid belt relying only on solarcells for power generation. During the comet in-situinvestigations, the solar generator of Philae, made upby 2 m2 of solar arrays, will produce about 8 W of peakpower (Figure 1). To optimize the in-situ power pro-duction, the solar generator was built using new Low-Intensity Low-Temperature (LILT) solar cells that areable to produce energy in the very tough environmentin which Rosetta and Philae will survive and work [2].In total, there are 1224 silicon solar cells with dimen-sion 32.4 mm × 33.7 mm, 200 mm thick (Figure 1).

Figure 1. Philae’s solar generator

Excluding the bottom, all but one side (i.e., the Lid)of Philae are covered with solar cells to maximize theproduced power. With reference to Figure 2, the solargenerator is made up by 8 panels: Wall 1 to Wall 5,Balcony 1 and 2, and the Lid. The two balcony panelsare connected to Wall 1 and Wall 5, respectively. Thus,the solar generator is made up by six separate electricalsections, or Solar Arrays (SA). It can be seen thateach array is orientated towards another direction, toenable the Lander to generate power whenever possiblewithout solar arrays movements [3, 4].

The six solar panels are connected to five individualMaximum Power Point Trackers (MPPT), with SA1and SA5 connected to the same MPPT as they are op-posite and never exposed to the Sun simultaneously.Politecnico di Milano is in charge of developing modelsand software tools to estimate the produced power, aswell as of developing ad hoc hardware simulators to op-erate the Philae’s Ground Reference Module (GRM).These are described in the remainder.

Figure 2. Philae’s solar arrays

3. Predicting Solar Arrays Performances

During its nominal lifetime, Philae will experiencelarge variations of operating conditions, which stronglyaffects the performances of solar arrays. As an ex-ample, the Sun intensity will vary from 0.11 SC (at3 AU) to 0.69 SC (at 1.2 AU) in case the extendedmission will be confirmed. The characteristic temper-atures will vary accordingly, and the accumulated ra-diation dose will increase as the comet approaches theperihelion. In order to correctly simulate the powerproduced by solar arrays in these conditions, a soft-ware simulator has been implemented [5]. This simu-lator computes the I-V curves of each of the six SA asa function of

• the Sun aspect angles (azimuth and elevation)

• the Sun distance

• the solar arrays temperature

• the fluence (1014 MeV equivalent rad. dose).

The characteristic I-V curve is described by four mainparameters (see Figure 3):

• the short-circuit current (Isc)

• the open-circuit voltage (Voc)

• the maximum power point current (Imp)

• the maximum power point voltage (Vmp).

Given the shape of the I-V curve, it is possible to cal-culate the current as a function of the voltage (i.e.,I=I(V)), or vice-versa, once the four parameters arecomputed as a function of the environmental condi-tions. This is done by using a simple model. A screen-shot of the SW Solar Array Simulator is reported inFigure 4. In the first two columns the six I-V curvesare reported; the third column reports the power pro-file, the Sun aspect angles, and the Lander geometrywith respect to the Sun. These are updated at eachsimulation step.

Aerotecnica Vol.88, No.1/2, January-June 2009

102 F. Topputo and F. Bernelli-Zazzera

Figure 3. I-V characteristic curve

Figure 4. Philae’s solar array simulator software

4. Simulating Solar Arrays Performances

Testing on-comet operations is not trivial from thepower production point of view. In this case, the on-comet I-V curves profile produced by the solar gener-ator has to be mimicked using on-ground equipment.It is known that solar cells currents scale with the co-sine of the Sun incidence angles. Thus, the low solarintensity coupled with high values of Sun incidenceangles produce very low currents (below 10 mA) thathave to be simulated. The variations of SA tempera-ture during the comet day produce changes of the I-Vcurve, too (for increasing temperature, Voc increasesand Isc decreases). The effect of temperature varia-tions on SA1 is shown in Figure 5, whereas Figure 6reports the variations due to varying Sun distance onthe same SA [6].

Other problems arise due to the presence of theMPPT. In order to track the maximum power pointon the solar panels IV-curve, Philae’s five MPPT movearound the knee-point with a tracking frequency of ap-

Figure 5. I-V curves of SA1 for varying temperatures(Sun perpendicular to SA1, 0.11 Solar Constants)

Figure 6. I-V curves of SA1 for varying solar constants(Sun perpendicular to SA1, T = -100C)

proximately 30–70 Hz. This tracking frequency mustnot couple with the frequency at which the on-groundequipment provides power to the MPPT (i.e., the fre-quency at which the current is given for varying volt-age).

Previous experiments [7] have shown that using ex-isting, off-the-shelf components to simulate the SA be-haviour causes two major problems:

• it is not possible to simulate currents below 100mA due to hardware limitations of the simulator;

• operating the simulator and the MPPT simulta-neously causes instabilities due to the couplingbetween the MPPT tracking frequency and thefrequency at which the I-V curve is represented.



5. The Solar Array Simulator System

In order to both simulate low currents and allow theoperation with the GRM, a dedicated Solar Array Sim-ulator (SAS) has been designed and developed. TheSAS requirements have been agreed between Politec-nico di Milano and the Lander Control Centre (LCC)at Deutsche zentrum fur Luft- und Raumfahrt (DLR).The operational architecture has been conceived byPolitecnico di Milano (PoliMi), whereas the SAS hasbeen designed and produced by CBL Electronics Srlsupported by PoliMi.

5.1. Minimal RequirementsThe minimal hardware requirements of the SAS

have been agreed by DLR and PoliMi [8,9]. The mainhardware requirements were:

• six independent channels

• voltage in the range 30–120 V

• current in the range 0–200 mA

• fully analog system

• bandwidth higher than MPPT tracking fre-quency

The six channels were asked to simulate the six SA ofPhilae independently; the current and voltage rangeswere assessed through parametric analyses [6]; an ana-log simulator was required in order to stabilize theoperations with the GRM. The main software require-ments were:

• static/dynamic simulation scenarios

• I-V curves imported with text files

• log files with 10 ms minimum sampling rate.

The first requirement is related to the possibility ofsimulating frozen I-V curves as well as time-varyingcurves (i.e., to mimic the SA behaviour during a cometday); the I-V curve data had to be imported with textfile in order to use the SAS SW (Figure 4); the re-quirement on the log file was wanted to sample theoutput I-V data at a frequency higher than that of theMPPT.

5.2. Operational ArchitectureThe whole simulation system’s architecture is

sketched in Figure 7. Beside the SW SAS, two HWSAS have been developed: Fast Loop SAS (FL-SAS)and Diode SAS (D-SAS). These two modules are com-plementary and can be used independently. For rea-sons that will be explained later, the FL-SAS is thenominal simulator and the D-SAS is used as backup.Input file. This file contains the values of the environ-mental parameters (Sun aspect angles, Sun distance,SA temperatures, fluence) as a function of time.

Figure 7. Sketch of the operational architecture

SAS SW. This software calculates the I-V curvesdata as a function of time. For each time label, thecurves are discretized into 15 (I,V) coordinate pointsand stored [11].Output file. This file contains the data on the dis-cretized I-V curves as well as summary informationon the simulation (e.g., the estimation on the powerproduced by each SA).Control SW. This software imports the data to setthe two HW SAS, and it is also used to operate them.Log File. This files contains the information on thesimulation; i.e., the sampled values of the output cur-rent and voltage. The sampling rate is specified in thecontrol SW.FL-SAS. This is the HW Fast Loop SAS. It repro-duces the I-V curve as a union of 14 straight lineswith varying slope.D-SAS. The Diode SAS reproduces the I-V curve us-ing a chain of diodes. There is no control on the shapeof the curve.GRM. This is the Philae GRM. It is equipped with allPhilae?s subsystems and payloads but the solar arrays.The SAS is connected to the GRM MPPT.Figure 8 shows the Philae’s GRM operated with theD-SAS.

5.3. The Fast Loop SASThe FL-SAS is made up by six independent elec-

trical boards, each one representing the I-V curve ofthe SA of Philae (Figure 9). The boards consist of anumber of programmable potentiometers used to dis-cretized the I-V curve into a set of 14 lines with varyingslope (15 discretization points with non-uniform dis-tribution are used). The potentiometers are summedup so to construct the whole I-V curve. The flexibilitywith which the I-V curve is constructed allows us toreproduce cases with varying maximum power point,beside the mere variation of Voc and Isc. A sample,low current curve with this feature is shown in Figure10 [10].



Figure 8. D-SAS (bottom left) and GRM (top right)

Figure 9. Overview of FL-SAS

5.4. The Diode SASThe D-SAS is a device with six independent, electri-

cal channels, too. The I-V curves in this case are rep-resented by a chain of diodes connected in series. TheIsc is set by controlling the diodes current, whereas theVoc is achieved by varying the number of diodes. Itis straightforward that the D-SAS suffers from severallimitations:

• the setting on Voc is poor and depends on the

Figure 10. Three I-V curves (Isc = 10 mA, Voc = 50V) with varying MPP produced by the FL-SAS SA2;red: theoretical curve, blue: real curve

diode forward voltage drop (about 0.7 V);

• the diode’s I-V characteristic varies with diodestemperature;

• the shape of the I-V curve depends on the shapeof the I-V characteristic of the diodes;

• there is no control on the maximum power point.

For the reasons above the D-SAS has been chosen asa back-up simulator, although it guarantees stable op-eration when connected to the MPPT.

6. Operation of the FL-SAS with MPPT

The SAS system described in this paper has beenmainly developed to overcome the instability arosewhen another SAS was used to feed the GRM [7].Thus, the proper working of FL-SAS connected to theMPPT is crucial. To assess this, a dedicated test hasbeen performed. The experiment, shown in Figure 11,consists in connecting the SAS to the MPPT board,whose output is represented by the bus, kept at a fixedvoltage by the power sink. A 660 µF capacitor is in-troduced to simulate Philae’s bus capacitors.

The FL-SAS operates at variable current and volt-age (I,V); these values are logged by the FL-SAS built-in logging system. This allows us to report the (I,V)points at which the FL-SAS operates and the MPPTis fed. The output of the MPPT board (Ibus, Vbus)is seen on the scope, and eventually recorded.

In Figure 12 a sample test case is reported. In thiscase the bus voltage is set to 28 V, whereas the I-Vcurve is specified by Isc = 80 mA, Voc = 80 V, Imp



Figure 11. System used to tests the FL-SAS withMPPT

= 70 mA, Vmp = 70 V. The red line is the I-V curve;the blue points are the (I, V) values sampled by theFL-SAS. These represent the output of the FL-SASand therefore the values at which the MPPT is fed. Asmooth, stable operation has been observed, so indi-cating the effectiveness of the FL-SAS when connectedto the MPPT. It can be seen that the MPPT works inthe region about the “knee” of the curve, which is theregion where the maximum power is extracted fromthe SAS. As the blue points move away from the MPP(green point), the efficiency of the MPPT decreases. Arough estimation of the efficiency can be achieved bymeasuring Ibus (an 80% efficiency has been observedin the test).

Figure 12. FL-SAS SA3 with MPPT (Isc = 100 mA,Voc = 60 V); red: theoretical curve, blue: real (I,V)points

7. Scenario Simulation

An on-comet scenario has been tested. From theFL-SAS point of view, simulating a scenario meansreproducing given I-V curves as a function of time.A few-minute scenario, which is representative of atypical six hours comet day, has been prepared andloaded on the FL-SAS. Simulating a longer scenariodoes not add any other qualitative feature that canbe obtained from the short scenario. The scenario hasbeen sampled using the built-in sampling feature with10 ms of sampling rate. From the settings it is possibleto reconstruct the theoretical I-V curves, whereas fromthe sampled (I,V) points it is possible to check thevalues at which the MPPT works.

Figure 13 reports the theoretical and the sampled(I,V) points of the scenario simulation for SA5. Itcan be seen that the MPPT works smoothly in theregion about the MPP, whose width depends on thecurrent. For low currents (e.g., currents below 10 mA)the MPPT stops tracking and works in “pass-through”mode. It can be concluded that the FL-SAS operationin conjunction with the MPPT has been successful.

Figure 13. Theoretical (red) and real (blue) I-V pointsof SA5 recorded during the scenario simulation

Both the FL-SAS and the D-SAS performed nomi-nally when connected to the GRM through its built-inMPPT (Figure 14). In particular, a stable Lander op-eration has been observed by analysing the telemetrydata. This allows us to use the two SAS to execute andassess GRM operations before commanding Philae forits real, on-comet operation.



Figure 14. FL-SAS and D-SAS (top-right) connectedto the GRM (centre-left) at LCC

8. Conclusions

In this paper we have described the Solar ArraySimulator tools that have been developed to work inconjunction with the Philae Ground Reference Mod-ule. The simulator is made up by software and hard-ware pieces. The software SAS is used to predict theproduced power and to generate the theoretical I-Vcurves of the six SA once some environmental param-eters are provided. Another task of the SW SAS isto control and set the HW SAS. The two HW SAS,i.e., the FL-SAS and the D-SAS, are conceived to re-place the Philae six SA in terms of electrical behaviour.Dedicated tests have proved that low currents can besimulated, and that the two SAS operate in stable con-ditions when connected to the MPPT. This is a greatdeparture from existing, off-the-shelf Solar Array Sim-ulators. The case of the Rosetta lander Philae is dis-cussed, although the same concepts apply in similarcontext.

Acknowledgments

The authors would like to acknowledge the ItalianSpace Agency (ASI) that supported the activities de-scribed in this paper through the contract I/062/08/0.The authors are also grateful to the colleagues at LCCthat supported this project with useful feedback andvaluable suggestions.

REFERENCES

1. A. Ercoli–Finzi, F. Bernelli-Zazzera, C. Dainese, F. Mal-nati, P.G. Magnani, E. Re, P. Bologna, S. Espinasse, andA. Olivieri, “SD2 – How To Sample A Comet”, Space Sci-ence Reviews, Vol. 128, 2007, pp. 281-299.

2. G. D’Accolti, G. Beltrame, E. Ferrando, L. Bram-billa, R. Contini, L. Vallini, R. Mugnuolo, C. Signorini,H. Fiebrich, and A. Caon, “The Solar Array PhotovoltaicAssembly for the Rosetta Orbiter and Lander Spacecraft’s”,6th European Space Power Conference, Porto, Portugal, 6–11 May, 2002.

3. M. Pesco and A. Ercoli–Finzi, “Rosetta Space Mission:The Solar Array Photovoltaic Assembly”, XIX CongressoNazionale AIDAA, Forlı, Italy, 17–21 September, 2007.

4. VV. AA., “Lander User Manual, 2.2-9b Solar Arrays”, RO-LAN-UM-3100-SA, Deutsche Zentrum fur Luft- und Raum-fahrt, 2007.

5. F. Topputo, F. Bernelli–Zazzera, and A. Ercoli–Finzi, “On-Comet Power Production: the Case of Rosetta Lander Phi-lae”, EPSC2010-96, European Planetary Science Congress2011, Rome, Italy, 19–24 September, 2010.

6. F. Topputo, “Sample I-V Curves to Reproduce with theNew Solar Array Simulator”, PHILAE-SA-SAS-005, Po-litecnico di Milano, 2009.

7. K. Geurts, “Rosetta Lander Solar Array Simulator& MPPT Operation Instability”, RO-LAN-TN-3224,Deutsche Zentrum fur Luft- und Raumfahrt, 2009.

8. K. Geurts, “Rosetta Lander MPPT Operation”, RO-LAN-TN-3225, Deutsche Zentrum fur Luft- und Raumfahrt,2009.

9. F. Topputo, “New Philae Solar Array Simulator: TechnicalRequirements”, PHILAE-SA-SAS-004, Politecnico di Mi-lano, 2009.

10. K. Geurts, “Rosetta Lander Solar Array Simulator Require-ments”, RO-LAN-TN-3304, Deutsche Zentrum fur Luft-und Raumfahrt, 2009.

11. F. Topputo, “Philae Solar array Simulator 1.0 User’s Man-ual”, PHILAE-SA-SAS-006, Politecnico di Milano, 2010.

12. F. Topputo, “Acceptance and Test of CBL Fast Loop So-lar Array Simulator”, PHILAE-SA-SAS-007, Politecnico diMilano, 2011.


simulation of low-intensity, low-temperature solar arrays with software … · 2015. 12. 11. ·...

Documents