- Email: [email protected]
Demisability of critical spacecraft components during atmospheric re-entry
Journal of Space Safety Engineering 6 (2019) 181–187
Contents lists available at ScienceDirect
Journal of Space Safety Engineering journal homepage: www.elsevier.com/locate/jsse
Demisability of critical spacecraft components during atmospheric re-entry Patrik Kärräng a,∗, Tobias Lips a, Tiago Soares b a b
Hyperschall Technologie Göttingen GmbH, Am Handweisergraben 13, Bovenden 37120, Germany European Space Agency, Keplerlaan 1, AZ Noordwijk 2201, Netherlands
a r t i c l e Keywords: Spacecraft re-entry Design-for-demise Space debris Casualty risk reduction
i n f o
a b s t r a c t According to international safety guidelines, the on-ground casualty risk for a re-entering object shall not exceed 1 in 10,000. The casualty expectancy can be reduced in two ways (1) by selecting a suitable impact area and population density within, or (2) by reducing the casualty area from the surviving fragments. Due to the cost associated with a controlled targeted re-entry, the latter option has attracted a lot of attention. To achieve the requirement of reducing the casualty area, the number, size and kinetic energy of the surviving fragments have to be limited. The fragments which survive re-entry are often from recurring spacecraft components (e.g. propellant tanks, reaction wheels, solar array drive mechanisms, magnetic torquers, etc.), therefore the interest of applying designs which increase the demisability of these components is high. Understanding the demise process during re-entry helps in identifying feasible design-for-demise options. For this study, we conducted re-entry risk analysis of two critical spacecraft components, a solar array drive mechanism, and a reaction wheel using a spacecraftoriented re-entry tool, in order to assess the break-up and demise behaviour of the components. Detailed models of the components were created using design input from the manufacturers and initial conditions for the simulations were selected within a release window (58–98 km) along a reference trajectory. We have investigated the casualty risk metrics for the components, derived the most-probable casualty area over release altitude and investigated its uncertainties. Together with the manufacturer, we identified feasible design-for-demise options for the components and evaluated their impact on the casualty risk.
1. Introduction Since the beginning of the space age humanity have been launching objects into orbit, steadily growing the number of operational satellites. This has enabled great benefit on Earth by furthering human technical capabilities. When reflecting over these achievements, few think about what was left behind after their operational lifetime. Rocket-stages and spacecraft are often left in an orbit where natural perturbations will over time bring them back to Earth, ending in an uncontrolled destructive re-entry. This has led to the situation where today the number of operational satellites is only a small fraction of the total number of objects in orbit, and several tons of debris re-enter every month. Space debris mitigation standards, which have been implemented by many space agencies around the world, specify an upper limit for the acceptable casualty expectancy for a re-entry as 10−4 . During re-entry most of a spacecraft will burn up but some components (e.g. propellant tanks, reaction wheels, solar array drive mechanisms, magnetic torquers) tend to survive. If the casualty risk during the development of a satellite is
∗
expected to be above the casualty threshold, a controlled re-entry is required, where any surviving fragments are targeted into a designated impact zone. The cost of developing, launching and ensuring reliability of a system for controlled re-entry is high, therefore, for most satellites in low-Earth orbit, it is more compelling to perform an uncontrolled re-entry at the end-of-life disposal. During an uncontrolled re-entry, the way to limit the on-ground casualty risk is to reduce the casualty area of the surviving fragments. The casualty area is the effective cross-section for a casualty, where the projected area of an impacting fragment is enlarged by the human cross-section to make a casualty cross-section. The theory of intentionally designing spacecraft to reduce the casualty risk during re-entry is called design-for-demise [1–3]. To evaluate the effectiveness of different design-for-demise strategies on system and component level, re-entry analysis tools are used to simulate the conditions experienced by the object. In this paper we describe a general method of comparing design-for-demise strategies and applied it to two selected common spacecraft components. We also discuss a general theory of demisability analysis and casualty expectancy calculations.
Abbreviations: D4D, Design-for-demise; DOF, Degrees-of-freedom; RWL, Reaction Wheel; SADM, Solar array drive mechanism. Corresponding author. E-mail addresses: [email protected] (P. Kärräng), [email protected] (T. Lips), [email protected] (T. Soares).
https://doi.org/10.1016/j.jsse.2019.08.003 Received 22 March 2019; Received in revised form 22 July 2019; Accepted 30 August 2019 Available online 23 September 2019 2468-8967/© 2019 Published by Elsevier Ltd on behalf of International Association for the Advancement of Space Safety.
P. Kärräng, T. Lips and T. Soares
Journal of Space Safety Engineering 6 (2019) 181–187
commonly computed using the Schaaf–Chambre method [6]. The aerodynamic coefficients in transitional flow (0.01 < Kn < 10) are computed based on a bridging function between free-molecular and continuum flow. The hypersonic heating, like the aerodynamic forces, makes a distinction between the flow regimes. The heat flux per unit area, 𝑞̇ , can be expressed as a function of the free-stream density, 𝜌∞ , the free-stream velocity, v∞ , and the Stanton number, St, as shown in Eq. (1).
Nomenclature Ac Afrag Ah Aref AS B Clam cp Ec h H0 hm kav
total debris casualty area surviving fragment cross-section projected human cross-section (𝐴ℎ = 0.36𝑚2 ) reference area surface area ballistic coefficient (kg/m2 ) laminar hypersonic heating constant (𝐶𝑙𝑎𝑚 = 1.23 × 10−4 kg0.5 /m) specific heat capacity casualty expectancy altitude density scale height (m) latent heat of melting ( ) 𝑞̇ shape dependent scaling factor 𝑘𝑎𝑣 = 𝐴 1 ∫𝑆 𝑞̇𝑙𝑜𝑐𝑎𝑙 𝑑𝑆
Kn L m 𝑚̇ Q 𝑄̇ 𝑞̇ RN Re2 St T Tm 𝑇̇ u v vE v∞ 𝜖 𝜃 𝜌 𝜌0 𝜌P 𝜌∞ 𝜎
Knudsen number characteristic length mass mass rate of change total heat absorbed heat flux heat flux per unit area nose radius Reynolds number behind the shock Stanton number temperature melting temperature rate of change in temperature ratio between release and circular orbit velocity velocity entry velocity free-stream velocity emissivity flight path angle atmospheric density atmospheric density at sea-level population density free-stream density Stefan–Boltzmann constant
𝑟𝑒𝑓
𝑞̇ =
1 𝜌 𝑣3 𝑆𝑡 2 ∞ ∞
(1)
In the free-molecular flow, St can be assumed to be close to 1 and for continuum flow the St can be determined by the modified Lees theory [7] as expressed in Eq. (2). 2.1 𝑆𝑡 = √ 𝑅𝑒2
(2)
The Stanton number for the transitional flow is computed using a bridging formula between the free-molecular and continuum flow. Initially the heat flux, 𝑄̇ , experienced by the body is converted into rise in body temperature, T, of the exposed surface. The exposed surface starts to ablate, resulting in a change of mass, 𝑚̇ , of the body once the melting temperature, Tm , is reached. This is described in the following heat balance equations, where cp is the specific heat capacity, hm is the latent heat of melting, 𝜀 is the emissivity, 𝜎 is the Stefan–Boltzmann constant and AS is the surface area.
𝑠𝑡𝑎𝑔
𝑄̇ = 𝑐𝑝 𝑚𝑇̇ + 𝜀𝜎𝐴𝑆 𝑇 4 𝑄̇ = −ℎ𝑚 𝑚̇ + 𝜀𝜎𝐴𝑆 𝑇 4
(𝑇 < 𝑇𝑚 ) (𝑇 = 𝑇𝑚 )
(3)
(4)
The equations of motion and heating are solved for a re-entering object during its decent in the atmosphere until it completely demises, or reaches the ground if part of its mass is surviving. 2.2. Analytical demise criteria Even though numerically solving the behaviour during re-entry is more popular, there have been analytical methods developed for assessing the demisability of simple shaped objects during re-entry [8–10]. These analytical solutions are based on simplifying assumptions which render them not exact for uncontrolled re-entries, but they provide good insight into the demisability of components during re-entry. The work done for ballistic trajectories has been described by Allen-Eggers [11]. Herein, the along-track velocity, v(h), for a straight entry flight path is given as a function of altitude, h, for a entry velocity, vE , using a constant flight path angle, 𝜃, ballistic coefficient, B, and assuming an exponential atmosphere with sea-level density, 𝜌0 , and sea-level density scale height, H0 . [ ] [ ] 𝜌0 𝐻 0 𝑣(ℎ) = 𝑣𝐸 exp − exp −ℎ∕𝐻0 (5) 2𝐵 sin(|𝜃|)
2. Re-entry demisability analysis 2.1. Re-entry dynamics and heating When identifying spacecraft components which are likely to survive re-entry, the trajectory, the heating and the fragmentation of a re-entering object have to be analyzed. This is usually done with computational methods, propagating the re-entry trajectory by 3 or 6-DOF flight dynamics (e.g. orbital state vector in Cartesian coordinates) with aerodynamics, aerothermodynamics, thermal, and structural loads evaluated along the trajectory, from its orbital state until impact. Aerodynamic forces acting on the body are by far the dominant force influencing the re-entry. Computation of the aerodynamic force and moment coefficients of a body can be done analytically for simple shapes [4] or by integrating the local force (as a function of the local flow inclination) over the surface of the body. During re-entry the object passes through different flow regimes, which are characterized by their Knudsen number. Knudsen number is the ratio of the mean-free path in the atmosphere and the size of the object (characteristic length). The computation of the aerodynamic coefficients changes with the flow regime. In the hypersonic continuum flow regime (Kn < 0.01) the coefficients are usually approximated using the modified Newtonian flow theory [5], while in free-molecular flow conditions (Kn > 10) the coefficients are
The attitude dependent stagnation point heat flux as derived by Fritsche and Lips [10] is using Lees theory [7] for laminar convective heat flux together with Eq. (5). The equation make use of the laminar hypersonic heating constant, Clam , and nose radius, RN , to solve the heat flux as a function of altitude. ] [ √ [ ] 𝜌0 3 3𝑝0 𝐻0 ℎ 𝑞̇ (ℎ) = 𝐶𝑙𝑎𝑚 𝑣𝐸 exp − − exp −ℎ∕𝐻0 (6) 𝑅𝑁 2𝐻0 2𝐵 sin(|𝜃|) Chapman [12] used laminar stagnation point heat transfer laws, and the ballistic trajectory in an exponential atmosphere to achieve an expression for total heat absorbed, Q(u), during re-entry. Assuming a non/ dimensional initial release velocity (𝑢 = 𝑣𝑟𝑒𝑙𝑒𝑎𝑠𝑒 𝑣(𝑐 𝑖𝑟𝑐 ,𝑜𝑟𝑏𝑖𝑡) ) and an ending velocity of zero (𝑢0 = 0), the solution can be expressed as √ 𝐵 1 1 𝑄(𝑢) = 1.82 ⋅ 106 𝑘𝑎𝑣 𝐴𝑟𝑒𝑓 𝑢2 (7) √ 𝑅𝑁 cos2 (𝜃) sin(𝜃) 182
P. Kärräng, T. Lips and T. Soares
Journal of Space Safety Engineering 6 (2019) 181–187
where Aref is the reference area of the body and kav is the ratio between the local and stagnation point heat flux integrated over the surface. Eqs. (6) and (7) together with Eqs. (3) and (4) reveal the two main surviving criteria for spacecraft components. A component can either survive re-entry by having a high heat storage capacity (see Eq. (8)), meaning that it cannot absorb enough heat during re-entry to demise. ( ) 𝑐𝑝 𝑚 𝑇𝑚 − 𝑇𝑖𝑛𝑖𝑡𝑖𝑎𝑙 + 𝑚ℎ𝑚 > 𝑄(𝑢) (8) A re-entry object can also survive by having an effective re-radiation (see Eq. (9)). In this case the second term of the heat balance equations is dominant, therefore the body re-radiate more heat than it absorbed and cannot demise. 𝜀𝜎𝐴𝑆 𝑇𝑚4 > 𝐴𝑟𝑒𝑓 𝑘𝑎𝑣 𝑞̇ max
(9)
Replacing the non-dimensional release velocity, u, in Eq. (7), with the along-track velocity in Eq. (5) gives us an expression for the total heat absorbed as a function of the release altitude. Then, by modifying this relation with Eq. (8) gives us a demising altitude for a re-entry object. Note that the analytical survival criteria (Eqs. (8) and (9)) are approximate solutions. As an example, they are not taking into consideration the interaction between the absorbed and the re-radiated heat. As a consequence, because these equations generally over-estimate demisability, they can only be used as a rough estimate for the latter. Eqs. (8) and (9) can also assist in realizing which survival criteria are dominant for the component being investigated, and what can be done to increase its demisability.
Fig. 1. Fragmentation events from a typical spacecraft model using SCARAB.
2.3. Commonly used re-entry tools There are a number of re-entry tools which are used by manufacturers and space agencies to assess the demisability and on-ground casualty risk of a re-entry. The tools usually fall into two main categories, objectoriented and spacecraft-oriented. Object-oriented re-entry tools model a spacecraft as several components with basic geometric shapes and a material assigned to it. Usually, these tools assume that the spacecraft breaks-up at a certain altitude where the internal components are released and a separate re-entry analysis is performed for each individual component. Some tools allow multi-layer nesting of components, once a layer has demised the next layer of components are released. Popular object-oriented tools include DAS [13], ORSAT [14], DRAMA/SESAM [15] and DEBRISK [16]. Spacecraft-oriented re-entry tools use a 3D geometry model of the full spacecraft to perform re-entry analysis. The complete destructive process along the re-entry trajectory is simulated with thermal and mechanical fragmentation. An example of spacecraftoriented tool is SCARAB, which has been used to produce the results in this paper. SCARAB is a high-fidelity software tool developed to simulate the thermal destruction of atmospheric re-entry of spacecraft. The 6-DOF equations of motion are solved deterministically with a Runge– Kutta integrator, with variable time step size and error control. Aerodynamic forces, torques and aerothermodynamic heat loads are calculated at each time step for the current attitude and geometry. The fragmentation analysis is performed along the trajectory and the fragmentation occurs when the integrity of an object is lost due to melting or mechanical failure (see Fig. 1). SCARAB has been applied in many studies to evaluate the effectiveness of different design-for-demise strategies.
Fig. 2. Illustration of casualty area for one fragment (the full grey circle area).
re-entry, targeting an area far away from any landmasses and traffic routes (e.g. South Pacific Ocean Uninhabited Area). Another alternative is to perform a semi-controlled re-entry [17], where the impact zone is phased in such a way that the track on ground is limiting the population effected by the debris. This phasing can lower the average population density within the possible impact zone. For an individual impact the casualty area is computed by the cross-section of the impacting fragment and a human cross-section, giving the “effective” cross-section for a casualty (see Fig. 2). The total casualty area is the sum of the casualty cross-section for all the hazardous fragments (see Eq. (11)) [18]. Fragments are considered to be hazardous if their kinetic impact energy is above 15 J. 𝐴𝑐 =
𝑁 ( ∑ √ 𝑖
𝐴ℎ +
√ )2 𝐴𝑓 𝑟𝑎𝑔,𝑖
(11)
(10)
It is clear that the total casualty area can be reduced in several ways: by limiting the number of surviving fragments; by limiting the size of fragments; and by reducing the kinetic energy of the surviving fragments. Considering this, it can be said that the total casualty area is driven by the number of hazardous fragments reaching ground. For reducing the number of fragments surviving down to ground, design-for-demise techniques have been developed to increase the demisability of fragments. Simulations using common re-entry tools show that D4D works, but, because few spacecraft designs are the similar, one solution might not apply on another type of spacecraft. D4D can be seen as a philosophy to be carried by the engineers during the design process. There is unfortunately no universal solution when it comes to design-for-demise, but design approaches on component level fall into one of four major categories:
For an uncontrolled re-entry the average population density under the satellite ground track is used, which depends on the orbital inclination. The population density can be limited by performing a controlled
• Reduction of heat required for complete demise (e.g. mass reduction, change of material); • Increase of heat absorbed (e.g. early exposure, early release);
3. On-ground casualty risk reduction The expected number of human casualties from a re-entry, Ec , is the product of the total casualty area of the surviving fragments, Ac , and the average population density within the re-entry impact zone, 𝜌P . 𝐸 𝑐 = 𝜌𝑃 𝐴 𝑐
183
P. Kärräng, T. Lips and T. Soares
Journal of Space Safety Engineering 6 (2019) 181–187
a bearing unit of stainless steel; and PCB is modelled with a low-density electronic material. The total mass of the reaction wheel is 8.72 kg and is composed of 75% stainless steel, 23% aluminium and 2% electronic material. The RWL needs to absorb around 9 MJ to achieve complete demise. Using the heat required for complete demise together with the surviving criteria in Eq. (8) and simplifying the component to an equivalent sphere with constant ballistic coefficient gives the result that the RWL could be demisable if released above 67 km. For the RWL two design modifications to increase its demisability were selected: Al68 and Al45. The Al68 design relies on changing the flywheel from stainless steel to aluminium and keeping the momentum capability. Therefore, having similar mass as before but the material has been replaced with a more demisable one. The Al45 design is achieved by keeping the material change in the flywheel (from stainless steel to aluminium) and reducing the size/mass of the wheel (reducing momentum capability). The Al45 design gives the possible benefit of combination of material change and mass reduction of the component.
Fig. 3. Solar array drive mechanism. Actuator side (left) and solar array interface side (right).
5. Re-entry analysis of critical components 5.1. Method of comparison Fig. 4. Reaction wheel, with housing (left) and without housing (right).
A SCARAB re-entry analysis begins with the definition of initial flight state vector of the re-entry object.
• Reduction of possible fragment creating components (e.g. containment); • Reduction of impact energy (e.g. partitioning).
5.1.1. Release conditions During uncontrolled re-entry, spacecraft disintegration (break-up) usually occurs in an altitude window between 100 and 60 km. This is the altitude range where internal components can be released. Due to the high computational requirements of spacecraft-oriented re-entry tools only a limited number of simulation cases can be done. To ensure representative and reproducible results from a low number of simulations, the input parameter space has to be limited. For trajectory, release conditions were selected along a reference trajectory, which has been generated using the DRAMA/SESAM tool assuming a sphere re-entering with a cross-sectional area of 1 m2 (diameter of 1.13 m) and a mass of 150 kg (mass-to-area ratio of 150 kg/m2 ). Five main release altitudes have been used for this study: 60 km, 69 km, 78 km, 87 km and 96 km. It is important to investigate the sensitivity of the casualty risk metrics with release altitude variation. The release conditions of each main altitude have been changed by ± 2km along the reference trajectory. Within each release altitude a number of attitude input variations have been used for further analysing the sensitivity of the input conditions. The release attitudes are obtained from a uniform equidistance distribution of attitude pointing directions. This to ensure that no bias towards any particular initial attitude was created by randomly generating the pointing directions. The main release altitudes (i.e. 60, 69, 78, 87 and 96 km) has 25 initial attitudes, while the altitude variation cases have 10 initial attitudes each. In total, 225 cases have been simulated for each component.
4. Design-for-demise of critical components 4.1. Component models To assess the demisability of a critical spacecraft component using a spacecraft-oriented re-entry tool, detailed modelling of the component is required. Detailed component design descriptions are provided by the manufacturers to ensure realistic modelling. 4.1.1. Solar array drive mechanism The SCARAB model of the Baseline SADM (see Fig. 3) is divided into three main parts: Main Housing, Actuator and Power & Signal Transfer Unit. The total mass is 6.55 kg. The model has around 64% aluminium, 23% stainless steel, 12% titanium and 1% other materials. Using material data, it can be computed that the energy required by this SADM to achieve complete demise is around 6 MJ (energy required to bring unit from release temperature to melting temperature plus the latent heat). By simplifying the component into a mass equivalent sphere with constant ballistic properties (not taking ablation into consideration) together with surviving criteria in Eq. 8, show that the SADM could be demisable above a release altitude of 61 km. Following discussion with the manufacturer, two D4D options were considered feasible for the SADM and selected for further investigations, the open-SADM design and the open-SADM with aluminium actuator. The main idea behind the open SADM design was to expose the interior of the component to the flow earlier by removing parts of the main housing assembly. Early exposure to the aerodynamic environment might have a positive effect on the ablation process. The open SADM with aluminium actuator solution would lower the energy required for a complete demise, by changing some materials with high specific heat (e.g. titanium and stainless steel) to materials with lower specific heat (e.g. aluminium). It was decided to keep the open SADM design and in addition to change some actuator sub-components material from titanium to aluminium, thereby reducing the heat required for complete demise.
5.1.2. Comparing casualty risk metrics The most important casualty risk metrics of any re-entry are surviving mass, casualty area and kinetic energy. In this study we focus on comparing surviving mass and casualty area from each simulation. We have ignored any fragment with a kinetic energy less than 15 J. Surviving mass over release altitude is a good metric for predicting the minimum demisable altitude for a component, it follows a predictable pattern (usually S-curve shaped) and also can reveal successful design changes, which decreased the heat required for melting of component. The casualty area over release altitude does not follow a predicable pattern as surviving mass does. This is due to the discrete nature of Eq. 11, where the sum of surviving fragments cross-sections is augmented by the human cross-section. Meaning that for fragments significantly smaller than a human, the individual casualty area will be a little larger than a human cross-section. This results into clear discrete jumps when looking at the total casualty area, these jumps can be categorized into regions
4.1.2. Reaction wheel The reaction wheel model (see Fig. 4) is divided into: housing modelled as aluminium and titanium; a flywheel modelled as stainless steel; 184
P. Kärräng, T. Lips and T. Soares
Journal of Space Safety Engineering 6 (2019) 181–187
Fig. 5. Surviving mass over release altitude for SADM Baseline model.
Fig. 7. Surviving mass over release altitude for all SADM models with dashed line is the mean surviving mass for each model.
Fig. 6. Most-probable casualty area curve fit over release altitude for the SADM baseline model with box-and-whiskers showing the distribution and spread of results. Fig. 8. Most-probable casualty area curve over release altitude for all SADM models.
defined by the number of fragments and separated roughly by human sized cross-section. To be able to easily compare different components the casualty area result as a function of altitude, a most-probable casualty area curve was fitted though each main release altitude. Because of the discrete nature of casualty area the curve is fitted though the mean casualty area of the most-probable number of ground fragments to better represent the behaviour of the component. The most-probable number of ground fragments was determined by using a kernel density estimation function (estimation of the probability density function) for each case and finding the discrete number of fragments with the highest value from the estimated probability distribution. The range and spread of the casualty area within a release altitude (main release altitude including altitude variation) has been visualized with a boxplot, where the box contains the middle 50% of results and the whiskers extend to the maximum and minimum casualty area.
scatter plot of the total casualty area for each simulation with boxplot indicating the spread of the results and the derived most-probable casualty area curve. The most-probable casualty area for the baseline model, increases with release altitude until 78 ± 2 km, above this release altitude the baseline model starts to demise completely. This behaviour is due to fragmentation: when released from higher altitudes the component get more heat, which enables fragmentation into more and more pieces until a certain release altitude where the fragments get enough heat to demise. The most-probable casualty area, as explained in method of comparison (Section 5.1), is the mean casualty area of the most-probable number of surviving fragments. When comparing all the SADM design, it can be seen that the surviving mass is similar for all models (see Fig. 7), and the curves are slightly shifted to the left (less mass surviving). This is most likely the result of the reduction of heat required for complete demise. The casualty area of the three designs can be seen in Fig. 8 The open SADM design generates more fragments for release altitudes above 60 km than the baseline model, while still having a very similar mostprobable casualty area distribution as for the baseline model. Because it generates more fragments at higher release altitudes, the open SADM has a larger average casualty area than the baseline. The open SADM with actuator material changed to aluminium has very different results than the other models. The open SADM with aluminium actuator generates more fragments than the baseline for release altitudes below 69 km. At 78 ± 2 km most of the simulations experience complete demise, indicating that our strategy of lowering the heat required for total demise was successful.
5.2. Numerical simulation results 5.2.1. Solar array drive mechanism As can be seen in Fig. 5, the surviving mass of the SADM baseline model decreases with increased release altitudes. The mass surviving varies from 76% (5kg) of the initial mass at the release of 58 km altitude and the baseline model achieves complete demise, for the first time, at a release altitude of 85 km (no hazardous fragments are surviving). Because, a majority of simulations demise at this altitude, 85 km can be said to be the lowest demisable release altitude for the baseline model. The variability (the range of scatter) of surviving mass results for all release altitudes is in the order of hundreds of grams. Fig. 6 show the 185
P. Kärräng, T. Lips and T. Soares
Journal of Space Safety Engineering 6 (2019) 181–187
Fig. 11. Most-probable casualty area curve fit over release altitude for the RWL baseline model with box-and-whiskers showing the distribution and spread of results.
Fig. 9. Survivability of critical sub-components for all SADM models.
Fig. 12. Surviving mass over release altitude for all RWL models with dashed line is the mean surviving mass for each model.
Fig. 10. Surviving mass over release altitude for RWL baseline model.
There is an extreme peak in the casualty area (fragments created) around 78 km release altitude. The peak indicates that there might have been a break-up of a sub-component and its newly created fragments have not had time enough to completely demise. Above 78 km release altitude these additional fragments have demised there is no case for the baseline model. In Fig. 12 the surviving mass of the all the RWL models are compared. The surviving mass curve for the two aluminium wheels are similar for all release altitudes while the baseline model has significantly higher surviving mass for release altitudes below 78 km. The aluminium RWL designs have a similar most-probable casualty area distribution to each other, as can be seen in Fig. 13. They also show an improvement in demisability compared to the baseline, the most probable release altitude for complete demise is 96 km for the baseline model while the two aluminium wheels demise at 87 km release altitude. This shows that the strategy of changing flywheel material from stainless steel to aluminium provided an improvement in demisability, but it should be noted that the reaction wheel still survives for release altitudes below 78km. The sub-component survivability has been compared in Fig. 14 for the last release altitude where the baseline model never demised (87 ± 2 km release altitude). The aluminium reaction wheels have similar survivability which is less than half of that of the baseline model. For the baseline model the sub-components which often survive have been identified to be the bearing assembly, flywheel and parts of the motor.
The impact of the D4D modifications on the sub-component survivability has been compared in Fig. 9 for 78 ± 2 km release altitude (the highest release altitude where the baseline model always survived). The open SADM and baseline model have similar survivability, but for the open SADM with aluminum actuator the survivability drastically drops. The most critical sub-components in the SADM baseline are the actuator bearing, potentiometer shaft and harmonic drive, and by changing actuator material to aluminium these sub-components can be made demisable. 5.2.2. Reaction wheel The mass surviving for the reaction wheel, just as for the SADM, decrease with increased release altitude (see Fig. 10). About 80% of the initial mass (around 7 kg) is surviving at 58 km release altitude, and the complete demise occurs for the first time at 94 km release altitude. There is mass surviving for all release altitudes up to 96 km. The casualty area results for the baseline model is shown in Fig. 11 with the boxplots and related spread (the minimum and maximum casualty area). It is clear from the figure that, unlike the SADM, the RWL has very little variation in the casualty area on ground. This is due to the lower complexity of the component with fewer sub-components which are able to create ground fragments. 186
P. Kärräng, T. Lips and T. Soares
Journal of Space Safety Engineering 6 (2019) 181–187
mechanism). In cooperation with the manufactures feasible design-fordemise strategies were implemented. Numerical simulations show that design-for-demise on component level clearly works and has the potential to significantly lower the casualty expectancy of a re-entry, but the solutions has to be individually tailored and tested for the specific component. Acknowledgements The work presented in this paper was partially funded by the ESA project ‘High-Fidelity Re-entry Simulation on Critical Spacecraft Platform Equipment’, ESA Contract No. 5000121149/17/NL/GLC/as. References [1] R. Kelley, Using the Design for Demise Philosophy to Reduce Casualty Risk Due to Reentering Spacecraft, 63rd International Astronautical Congress, NASA Johnson Space Center; Houston, TX, United States, 2012. https://ntrs.nasa.gov/search.jsp?R=20120002794. [2] P.M. Waswa, M. Elliot, J.A. Hoffman, Spacecraft design-for-demise implementation strategy & decision-making methodology for low earth orbit missions, Advances in Space Research 51 (9) (2013) 1627–1637. [3] S. Lemmens, Q. Funke, H. Krag, On-ground casualty risk reduction by structural design for demise, Advances in Space Research 55 (11) (2015) 2592–2606. [4] B. Fritsche, Aerodynamic categorization of spacecraft in low earth orbits, in: Proceedings of the 6th international conference on astrodynamics tools and techniques, 2016 https://indico.esa.int/event/111/contributions/285/. [5] J.D. Anderson, Hypersonic and high-temperature gas dynamics, McGraw-Hill Series in aeronautical and aerospace engineering, 1989. [6] P.A. Chambre, S.A. Schaaf, Flow of rarefied gases, Princeton University Press, 1961. [7] L. Lees, Laminar heat transfer over blunt-nosed bodies at hypersonic flight speeds, Journal of Jet Propulsion 26 (4) (1956) 259–269. [8] R. Baker, M. Weaver, Orbital spacecraft reentry breakup, in: Proceedings of the 50th International Astronautical Congress, Amsterdam, The Netherlands, 1999 IAA-99-IAA.6.7.04. [9] G. Koppenwallner, B. Fritsche, T. Lips, Survivability and ground risk potential of screws and bolts of disintegrating spacecraft during uncontrolled re-entry, in: Space Debris, volume 473, 2001, pp. 533–539. [10] B. Fritsche, T. Lips, G. Koppenwallner, Analytical and numerical re-entry analysis of simple-shaped objects, Acta Astronautica 60 (8-9) (2007) 737–751. [11] H.J. Allen, A.J. Eggers, A study of the motion and aerodynamic heating of missiles entering the earth’s atmosphere at high supersonic speeds, Technical Report, National Advisory Committee for Aeronautics, 1953. https://ntrs.nasa.gov/search.jsp?R=19930091020. [12] D.R. Chapman, An approximate analytical method for studying entry into planetary atmospheres, Technical Report, National Advisory Committee for Aeronautics, 1958. https://ntrs.nasa.gov/search.jsp?R=19930085059. [13] E. Stansbery, J. Opiela, A. Vavrin, B. Draeger, P. Anz-Meador, Debris assessment software version 2.1 user’s guide, Technical Report, 2016. [14] J. Dobarco-Otero, R. Smith, K. Bledsoe, R. Delaune, W. Rochelle, N. Johnson, The object reentry survival analysis tool (ORSAT)-version 6.0 and its application to spacecraft entry, in: Proceedings of the 56th Congress of the International Astronautical Federation, the International Academy of Astronautics, and International Institute of Space Law, IAC-05-B6, 2005, pp. 17–21. 3 [15] C. Martin, C. Brandmueller, K. Bunte, J. Cheese, B. Fritsche, H. Klinkrad, T. Lips, N. Sanchez, A debris risk assessment tool supporting mitigation guidelines, in: 4th European Conference on Space Debris, volume 587, 2005, p. 345. [16] P. Omaly, M. Spel, Debrisk, a tool for re-entry risk analysis, in: A Safer Space for Safer World, volume 699, 2012. [17] T. Lips, P. Kärräng, Casualty risk reduction by semi-controlled re-entry, in: Proceedings of the 9th international association for the advancement of space saefty conference, 2017, pp. 221–226. [18] H. Klinkrad, Space debris - Models and Risk Analysis, Springer-Verlag, Berlin Heidelberg, 2010.
Fig. 13. Most-probable casualty area curve over release altitude for all RWL models.
Fig. 14. Survivability of critical sub-components for all RWL models.
6. Conclusion In this paper we have described a method for quantifying the demisability of spacecraft components by using a most-probable casualty area curve. The curve can also be used in evaluating the effectiveness of design-for-demise modifications. The method has been applied to two critical spacecraft components (reaction wheel and solar array drive
187
Contents lists available at ScienceDirect
Journal of Space Safety Engineering journal homepage: www.elsevier.com/locate/jsse
Demisability of critical spacecraft components during atmospheric re-entry Patrik Kärräng a,∗, Tobias Lips a, Tiago Soares b a b
Hyperschall Technologie Göttingen GmbH, Am Handweisergraben 13, Bovenden 37120, Germany European Space Agency, Keplerlaan 1, AZ Noordwijk 2201, Netherlands
a r t i c l e Keywords: Spacecraft re-entry Design-for-demise Space debris Casualty risk reduction
i n f o
a b s t r a c t According to international safety guidelines, the on-ground casualty risk for a re-entering object shall not exceed 1 in 10,000. The casualty expectancy can be reduced in two ways (1) by selecting a suitable impact area and population density within, or (2) by reducing the casualty area from the surviving fragments. Due to the cost associated with a controlled targeted re-entry, the latter option has attracted a lot of attention. To achieve the requirement of reducing the casualty area, the number, size and kinetic energy of the surviving fragments have to be limited. The fragments which survive re-entry are often from recurring spacecraft components (e.g. propellant tanks, reaction wheels, solar array drive mechanisms, magnetic torquers, etc.), therefore the interest of applying designs which increase the demisability of these components is high. Understanding the demise process during re-entry helps in identifying feasible design-for-demise options. For this study, we conducted re-entry risk analysis of two critical spacecraft components, a solar array drive mechanism, and a reaction wheel using a spacecraftoriented re-entry tool, in order to assess the break-up and demise behaviour of the components. Detailed models of the components were created using design input from the manufacturers and initial conditions for the simulations were selected within a release window (58–98 km) along a reference trajectory. We have investigated the casualty risk metrics for the components, derived the most-probable casualty area over release altitude and investigated its uncertainties. Together with the manufacturer, we identified feasible design-for-demise options for the components and evaluated their impact on the casualty risk.
1. Introduction Since the beginning of the space age humanity have been launching objects into orbit, steadily growing the number of operational satellites. This has enabled great benefit on Earth by furthering human technical capabilities. When reflecting over these achievements, few think about what was left behind after their operational lifetime. Rocket-stages and spacecraft are often left in an orbit where natural perturbations will over time bring them back to Earth, ending in an uncontrolled destructive re-entry. This has led to the situation where today the number of operational satellites is only a small fraction of the total number of objects in orbit, and several tons of debris re-enter every month. Space debris mitigation standards, which have been implemented by many space agencies around the world, specify an upper limit for the acceptable casualty expectancy for a re-entry as 10−4 . During re-entry most of a spacecraft will burn up but some components (e.g. propellant tanks, reaction wheels, solar array drive mechanisms, magnetic torquers) tend to survive. If the casualty risk during the development of a satellite is
∗
expected to be above the casualty threshold, a controlled re-entry is required, where any surviving fragments are targeted into a designated impact zone. The cost of developing, launching and ensuring reliability of a system for controlled re-entry is high, therefore, for most satellites in low-Earth orbit, it is more compelling to perform an uncontrolled re-entry at the end-of-life disposal. During an uncontrolled re-entry, the way to limit the on-ground casualty risk is to reduce the casualty area of the surviving fragments. The casualty area is the effective cross-section for a casualty, where the projected area of an impacting fragment is enlarged by the human cross-section to make a casualty cross-section. The theory of intentionally designing spacecraft to reduce the casualty risk during re-entry is called design-for-demise [1–3]. To evaluate the effectiveness of different design-for-demise strategies on system and component level, re-entry analysis tools are used to simulate the conditions experienced by the object. In this paper we describe a general method of comparing design-for-demise strategies and applied it to two selected common spacecraft components. We also discuss a general theory of demisability analysis and casualty expectancy calculations.
Abbreviations: D4D, Design-for-demise; DOF, Degrees-of-freedom; RWL, Reaction Wheel; SADM, Solar array drive mechanism. Corresponding author. E-mail addresses: [email protected] (P. Kärräng), [email protected] (T. Lips), [email protected] (T. Soares).
https://doi.org/10.1016/j.jsse.2019.08.003 Received 22 March 2019; Received in revised form 22 July 2019; Accepted 30 August 2019 Available online 23 September 2019 2468-8967/© 2019 Published by Elsevier Ltd on behalf of International Association for the Advancement of Space Safety.
P. Kärräng, T. Lips and T. Soares
Journal of Space Safety Engineering 6 (2019) 181–187
commonly computed using the Schaaf–Chambre method [6]. The aerodynamic coefficients in transitional flow (0.01 < Kn < 10) are computed based on a bridging function between free-molecular and continuum flow. The hypersonic heating, like the aerodynamic forces, makes a distinction between the flow regimes. The heat flux per unit area, 𝑞̇ , can be expressed as a function of the free-stream density, 𝜌∞ , the free-stream velocity, v∞ , and the Stanton number, St, as shown in Eq. (1).
Nomenclature Ac Afrag Ah Aref AS B Clam cp Ec h H0 hm kav
total debris casualty area surviving fragment cross-section projected human cross-section (𝐴ℎ = 0.36𝑚2 ) reference area surface area ballistic coefficient (kg/m2 ) laminar hypersonic heating constant (𝐶𝑙𝑎𝑚 = 1.23 × 10−4 kg0.5 /m) specific heat capacity casualty expectancy altitude density scale height (m) latent heat of melting ( ) 𝑞̇ shape dependent scaling factor 𝑘𝑎𝑣 = 𝐴 1 ∫𝑆 𝑞̇𝑙𝑜𝑐𝑎𝑙 𝑑𝑆
Kn L m 𝑚̇ Q 𝑄̇ 𝑞̇ RN Re2 St T Tm 𝑇̇ u v vE v∞ 𝜖 𝜃 𝜌 𝜌0 𝜌P 𝜌∞ 𝜎
Knudsen number characteristic length mass mass rate of change total heat absorbed heat flux heat flux per unit area nose radius Reynolds number behind the shock Stanton number temperature melting temperature rate of change in temperature ratio between release and circular orbit velocity velocity entry velocity free-stream velocity emissivity flight path angle atmospheric density atmospheric density at sea-level population density free-stream density Stefan–Boltzmann constant
𝑟𝑒𝑓
𝑞̇ =
1 𝜌 𝑣3 𝑆𝑡 2 ∞ ∞
(1)
In the free-molecular flow, St can be assumed to be close to 1 and for continuum flow the St can be determined by the modified Lees theory [7] as expressed in Eq. (2). 2.1 𝑆𝑡 = √ 𝑅𝑒2
(2)
The Stanton number for the transitional flow is computed using a bridging formula between the free-molecular and continuum flow. Initially the heat flux, 𝑄̇ , experienced by the body is converted into rise in body temperature, T, of the exposed surface. The exposed surface starts to ablate, resulting in a change of mass, 𝑚̇ , of the body once the melting temperature, Tm , is reached. This is described in the following heat balance equations, where cp is the specific heat capacity, hm is the latent heat of melting, 𝜀 is the emissivity, 𝜎 is the Stefan–Boltzmann constant and AS is the surface area.
𝑠𝑡𝑎𝑔
𝑄̇ = 𝑐𝑝 𝑚𝑇̇ + 𝜀𝜎𝐴𝑆 𝑇 4 𝑄̇ = −ℎ𝑚 𝑚̇ + 𝜀𝜎𝐴𝑆 𝑇 4
(𝑇 < 𝑇𝑚 ) (𝑇 = 𝑇𝑚 )
(3)
(4)
The equations of motion and heating are solved for a re-entering object during its decent in the atmosphere until it completely demises, or reaches the ground if part of its mass is surviving. 2.2. Analytical demise criteria Even though numerically solving the behaviour during re-entry is more popular, there have been analytical methods developed for assessing the demisability of simple shaped objects during re-entry [8–10]. These analytical solutions are based on simplifying assumptions which render them not exact for uncontrolled re-entries, but they provide good insight into the demisability of components during re-entry. The work done for ballistic trajectories has been described by Allen-Eggers [11]. Herein, the along-track velocity, v(h), for a straight entry flight path is given as a function of altitude, h, for a entry velocity, vE , using a constant flight path angle, 𝜃, ballistic coefficient, B, and assuming an exponential atmosphere with sea-level density, 𝜌0 , and sea-level density scale height, H0 . [ ] [ ] 𝜌0 𝐻 0 𝑣(ℎ) = 𝑣𝐸 exp − exp −ℎ∕𝐻0 (5) 2𝐵 sin(|𝜃|)
2. Re-entry demisability analysis 2.1. Re-entry dynamics and heating When identifying spacecraft components which are likely to survive re-entry, the trajectory, the heating and the fragmentation of a re-entering object have to be analyzed. This is usually done with computational methods, propagating the re-entry trajectory by 3 or 6-DOF flight dynamics (e.g. orbital state vector in Cartesian coordinates) with aerodynamics, aerothermodynamics, thermal, and structural loads evaluated along the trajectory, from its orbital state until impact. Aerodynamic forces acting on the body are by far the dominant force influencing the re-entry. Computation of the aerodynamic force and moment coefficients of a body can be done analytically for simple shapes [4] or by integrating the local force (as a function of the local flow inclination) over the surface of the body. During re-entry the object passes through different flow regimes, which are characterized by their Knudsen number. Knudsen number is the ratio of the mean-free path in the atmosphere and the size of the object (characteristic length). The computation of the aerodynamic coefficients changes with the flow regime. In the hypersonic continuum flow regime (Kn < 0.01) the coefficients are usually approximated using the modified Newtonian flow theory [5], while in free-molecular flow conditions (Kn > 10) the coefficients are
The attitude dependent stagnation point heat flux as derived by Fritsche and Lips [10] is using Lees theory [7] for laminar convective heat flux together with Eq. (5). The equation make use of the laminar hypersonic heating constant, Clam , and nose radius, RN , to solve the heat flux as a function of altitude. ] [ √ [ ] 𝜌0 3 3𝑝0 𝐻0 ℎ 𝑞̇ (ℎ) = 𝐶𝑙𝑎𝑚 𝑣𝐸 exp − − exp −ℎ∕𝐻0 (6) 𝑅𝑁 2𝐻0 2𝐵 sin(|𝜃|) Chapman [12] used laminar stagnation point heat transfer laws, and the ballistic trajectory in an exponential atmosphere to achieve an expression for total heat absorbed, Q(u), during re-entry. Assuming a non/ dimensional initial release velocity (𝑢 = 𝑣𝑟𝑒𝑙𝑒𝑎𝑠𝑒 𝑣(𝑐 𝑖𝑟𝑐 ,𝑜𝑟𝑏𝑖𝑡) ) and an ending velocity of zero (𝑢0 = 0), the solution can be expressed as √ 𝐵 1 1 𝑄(𝑢) = 1.82 ⋅ 106 𝑘𝑎𝑣 𝐴𝑟𝑒𝑓 𝑢2 (7) √ 𝑅𝑁 cos2 (𝜃) sin(𝜃) 182
P. Kärräng, T. Lips and T. Soares
Journal of Space Safety Engineering 6 (2019) 181–187
where Aref is the reference area of the body and kav is the ratio between the local and stagnation point heat flux integrated over the surface. Eqs. (6) and (7) together with Eqs. (3) and (4) reveal the two main surviving criteria for spacecraft components. A component can either survive re-entry by having a high heat storage capacity (see Eq. (8)), meaning that it cannot absorb enough heat during re-entry to demise. ( ) 𝑐𝑝 𝑚 𝑇𝑚 − 𝑇𝑖𝑛𝑖𝑡𝑖𝑎𝑙 + 𝑚ℎ𝑚 > 𝑄(𝑢) (8) A re-entry object can also survive by having an effective re-radiation (see Eq. (9)). In this case the second term of the heat balance equations is dominant, therefore the body re-radiate more heat than it absorbed and cannot demise. 𝜀𝜎𝐴𝑆 𝑇𝑚4 > 𝐴𝑟𝑒𝑓 𝑘𝑎𝑣 𝑞̇ max
(9)
Replacing the non-dimensional release velocity, u, in Eq. (7), with the along-track velocity in Eq. (5) gives us an expression for the total heat absorbed as a function of the release altitude. Then, by modifying this relation with Eq. (8) gives us a demising altitude for a re-entry object. Note that the analytical survival criteria (Eqs. (8) and (9)) are approximate solutions. As an example, they are not taking into consideration the interaction between the absorbed and the re-radiated heat. As a consequence, because these equations generally over-estimate demisability, they can only be used as a rough estimate for the latter. Eqs. (8) and (9) can also assist in realizing which survival criteria are dominant for the component being investigated, and what can be done to increase its demisability.
Fig. 1. Fragmentation events from a typical spacecraft model using SCARAB.
2.3. Commonly used re-entry tools There are a number of re-entry tools which are used by manufacturers and space agencies to assess the demisability and on-ground casualty risk of a re-entry. The tools usually fall into two main categories, objectoriented and spacecraft-oriented. Object-oriented re-entry tools model a spacecraft as several components with basic geometric shapes and a material assigned to it. Usually, these tools assume that the spacecraft breaks-up at a certain altitude where the internal components are released and a separate re-entry analysis is performed for each individual component. Some tools allow multi-layer nesting of components, once a layer has demised the next layer of components are released. Popular object-oriented tools include DAS [13], ORSAT [14], DRAMA/SESAM [15] and DEBRISK [16]. Spacecraft-oriented re-entry tools use a 3D geometry model of the full spacecraft to perform re-entry analysis. The complete destructive process along the re-entry trajectory is simulated with thermal and mechanical fragmentation. An example of spacecraftoriented tool is SCARAB, which has been used to produce the results in this paper. SCARAB is a high-fidelity software tool developed to simulate the thermal destruction of atmospheric re-entry of spacecraft. The 6-DOF equations of motion are solved deterministically with a Runge– Kutta integrator, with variable time step size and error control. Aerodynamic forces, torques and aerothermodynamic heat loads are calculated at each time step for the current attitude and geometry. The fragmentation analysis is performed along the trajectory and the fragmentation occurs when the integrity of an object is lost due to melting or mechanical failure (see Fig. 1). SCARAB has been applied in many studies to evaluate the effectiveness of different design-for-demise strategies.
Fig. 2. Illustration of casualty area for one fragment (the full grey circle area).
re-entry, targeting an area far away from any landmasses and traffic routes (e.g. South Pacific Ocean Uninhabited Area). Another alternative is to perform a semi-controlled re-entry [17], where the impact zone is phased in such a way that the track on ground is limiting the population effected by the debris. This phasing can lower the average population density within the possible impact zone. For an individual impact the casualty area is computed by the cross-section of the impacting fragment and a human cross-section, giving the “effective” cross-section for a casualty (see Fig. 2). The total casualty area is the sum of the casualty cross-section for all the hazardous fragments (see Eq. (11)) [18]. Fragments are considered to be hazardous if their kinetic impact energy is above 15 J. 𝐴𝑐 =
𝑁 ( ∑ √ 𝑖
𝐴ℎ +
√ )2 𝐴𝑓 𝑟𝑎𝑔,𝑖
(11)
(10)
It is clear that the total casualty area can be reduced in several ways: by limiting the number of surviving fragments; by limiting the size of fragments; and by reducing the kinetic energy of the surviving fragments. Considering this, it can be said that the total casualty area is driven by the number of hazardous fragments reaching ground. For reducing the number of fragments surviving down to ground, design-for-demise techniques have been developed to increase the demisability of fragments. Simulations using common re-entry tools show that D4D works, but, because few spacecraft designs are the similar, one solution might not apply on another type of spacecraft. D4D can be seen as a philosophy to be carried by the engineers during the design process. There is unfortunately no universal solution when it comes to design-for-demise, but design approaches on component level fall into one of four major categories:
For an uncontrolled re-entry the average population density under the satellite ground track is used, which depends on the orbital inclination. The population density can be limited by performing a controlled
• Reduction of heat required for complete demise (e.g. mass reduction, change of material); • Increase of heat absorbed (e.g. early exposure, early release);
3. On-ground casualty risk reduction The expected number of human casualties from a re-entry, Ec , is the product of the total casualty area of the surviving fragments, Ac , and the average population density within the re-entry impact zone, 𝜌P . 𝐸 𝑐 = 𝜌𝑃 𝐴 𝑐
183
P. Kärräng, T. Lips and T. Soares
Journal of Space Safety Engineering 6 (2019) 181–187
a bearing unit of stainless steel; and PCB is modelled with a low-density electronic material. The total mass of the reaction wheel is 8.72 kg and is composed of 75% stainless steel, 23% aluminium and 2% electronic material. The RWL needs to absorb around 9 MJ to achieve complete demise. Using the heat required for complete demise together with the surviving criteria in Eq. (8) and simplifying the component to an equivalent sphere with constant ballistic coefficient gives the result that the RWL could be demisable if released above 67 km. For the RWL two design modifications to increase its demisability were selected: Al68 and Al45. The Al68 design relies on changing the flywheel from stainless steel to aluminium and keeping the momentum capability. Therefore, having similar mass as before but the material has been replaced with a more demisable one. The Al45 design is achieved by keeping the material change in the flywheel (from stainless steel to aluminium) and reducing the size/mass of the wheel (reducing momentum capability). The Al45 design gives the possible benefit of combination of material change and mass reduction of the component.
Fig. 3. Solar array drive mechanism. Actuator side (left) and solar array interface side (right).
5. Re-entry analysis of critical components 5.1. Method of comparison Fig. 4. Reaction wheel, with housing (left) and without housing (right).
A SCARAB re-entry analysis begins with the definition of initial flight state vector of the re-entry object.
• Reduction of possible fragment creating components (e.g. containment); • Reduction of impact energy (e.g. partitioning).
5.1.1. Release conditions During uncontrolled re-entry, spacecraft disintegration (break-up) usually occurs in an altitude window between 100 and 60 km. This is the altitude range where internal components can be released. Due to the high computational requirements of spacecraft-oriented re-entry tools only a limited number of simulation cases can be done. To ensure representative and reproducible results from a low number of simulations, the input parameter space has to be limited. For trajectory, release conditions were selected along a reference trajectory, which has been generated using the DRAMA/SESAM tool assuming a sphere re-entering with a cross-sectional area of 1 m2 (diameter of 1.13 m) and a mass of 150 kg (mass-to-area ratio of 150 kg/m2 ). Five main release altitudes have been used for this study: 60 km, 69 km, 78 km, 87 km and 96 km. It is important to investigate the sensitivity of the casualty risk metrics with release altitude variation. The release conditions of each main altitude have been changed by ± 2km along the reference trajectory. Within each release altitude a number of attitude input variations have been used for further analysing the sensitivity of the input conditions. The release attitudes are obtained from a uniform equidistance distribution of attitude pointing directions. This to ensure that no bias towards any particular initial attitude was created by randomly generating the pointing directions. The main release altitudes (i.e. 60, 69, 78, 87 and 96 km) has 25 initial attitudes, while the altitude variation cases have 10 initial attitudes each. In total, 225 cases have been simulated for each component.
4. Design-for-demise of critical components 4.1. Component models To assess the demisability of a critical spacecraft component using a spacecraft-oriented re-entry tool, detailed modelling of the component is required. Detailed component design descriptions are provided by the manufacturers to ensure realistic modelling. 4.1.1. Solar array drive mechanism The SCARAB model of the Baseline SADM (see Fig. 3) is divided into three main parts: Main Housing, Actuator and Power & Signal Transfer Unit. The total mass is 6.55 kg. The model has around 64% aluminium, 23% stainless steel, 12% titanium and 1% other materials. Using material data, it can be computed that the energy required by this SADM to achieve complete demise is around 6 MJ (energy required to bring unit from release temperature to melting temperature plus the latent heat). By simplifying the component into a mass equivalent sphere with constant ballistic properties (not taking ablation into consideration) together with surviving criteria in Eq. 8, show that the SADM could be demisable above a release altitude of 61 km. Following discussion with the manufacturer, two D4D options were considered feasible for the SADM and selected for further investigations, the open-SADM design and the open-SADM with aluminium actuator. The main idea behind the open SADM design was to expose the interior of the component to the flow earlier by removing parts of the main housing assembly. Early exposure to the aerodynamic environment might have a positive effect on the ablation process. The open SADM with aluminium actuator solution would lower the energy required for a complete demise, by changing some materials with high specific heat (e.g. titanium and stainless steel) to materials with lower specific heat (e.g. aluminium). It was decided to keep the open SADM design and in addition to change some actuator sub-components material from titanium to aluminium, thereby reducing the heat required for complete demise.
5.1.2. Comparing casualty risk metrics The most important casualty risk metrics of any re-entry are surviving mass, casualty area and kinetic energy. In this study we focus on comparing surviving mass and casualty area from each simulation. We have ignored any fragment with a kinetic energy less than 15 J. Surviving mass over release altitude is a good metric for predicting the minimum demisable altitude for a component, it follows a predictable pattern (usually S-curve shaped) and also can reveal successful design changes, which decreased the heat required for melting of component. The casualty area over release altitude does not follow a predicable pattern as surviving mass does. This is due to the discrete nature of Eq. 11, where the sum of surviving fragments cross-sections is augmented by the human cross-section. Meaning that for fragments significantly smaller than a human, the individual casualty area will be a little larger than a human cross-section. This results into clear discrete jumps when looking at the total casualty area, these jumps can be categorized into regions
4.1.2. Reaction wheel The reaction wheel model (see Fig. 4) is divided into: housing modelled as aluminium and titanium; a flywheel modelled as stainless steel; 184
P. Kärräng, T. Lips and T. Soares
Journal of Space Safety Engineering 6 (2019) 181–187
Fig. 5. Surviving mass over release altitude for SADM Baseline model.
Fig. 7. Surviving mass over release altitude for all SADM models with dashed line is the mean surviving mass for each model.
Fig. 6. Most-probable casualty area curve fit over release altitude for the SADM baseline model with box-and-whiskers showing the distribution and spread of results. Fig. 8. Most-probable casualty area curve over release altitude for all SADM models.
defined by the number of fragments and separated roughly by human sized cross-section. To be able to easily compare different components the casualty area result as a function of altitude, a most-probable casualty area curve was fitted though each main release altitude. Because of the discrete nature of casualty area the curve is fitted though the mean casualty area of the most-probable number of ground fragments to better represent the behaviour of the component. The most-probable number of ground fragments was determined by using a kernel density estimation function (estimation of the probability density function) for each case and finding the discrete number of fragments with the highest value from the estimated probability distribution. The range and spread of the casualty area within a release altitude (main release altitude including altitude variation) has been visualized with a boxplot, where the box contains the middle 50% of results and the whiskers extend to the maximum and minimum casualty area.
scatter plot of the total casualty area for each simulation with boxplot indicating the spread of the results and the derived most-probable casualty area curve. The most-probable casualty area for the baseline model, increases with release altitude until 78 ± 2 km, above this release altitude the baseline model starts to demise completely. This behaviour is due to fragmentation: when released from higher altitudes the component get more heat, which enables fragmentation into more and more pieces until a certain release altitude where the fragments get enough heat to demise. The most-probable casualty area, as explained in method of comparison (Section 5.1), is the mean casualty area of the most-probable number of surviving fragments. When comparing all the SADM design, it can be seen that the surviving mass is similar for all models (see Fig. 7), and the curves are slightly shifted to the left (less mass surviving). This is most likely the result of the reduction of heat required for complete demise. The casualty area of the three designs can be seen in Fig. 8 The open SADM design generates more fragments for release altitudes above 60 km than the baseline model, while still having a very similar mostprobable casualty area distribution as for the baseline model. Because it generates more fragments at higher release altitudes, the open SADM has a larger average casualty area than the baseline. The open SADM with actuator material changed to aluminium has very different results than the other models. The open SADM with aluminium actuator generates more fragments than the baseline for release altitudes below 69 km. At 78 ± 2 km most of the simulations experience complete demise, indicating that our strategy of lowering the heat required for total demise was successful.
5.2. Numerical simulation results 5.2.1. Solar array drive mechanism As can be seen in Fig. 5, the surviving mass of the SADM baseline model decreases with increased release altitudes. The mass surviving varies from 76% (5kg) of the initial mass at the release of 58 km altitude and the baseline model achieves complete demise, for the first time, at a release altitude of 85 km (no hazardous fragments are surviving). Because, a majority of simulations demise at this altitude, 85 km can be said to be the lowest demisable release altitude for the baseline model. The variability (the range of scatter) of surviving mass results for all release altitudes is in the order of hundreds of grams. Fig. 6 show the 185
P. Kärräng, T. Lips and T. Soares
Journal of Space Safety Engineering 6 (2019) 181–187
Fig. 11. Most-probable casualty area curve fit over release altitude for the RWL baseline model with box-and-whiskers showing the distribution and spread of results.
Fig. 9. Survivability of critical sub-components for all SADM models.
Fig. 12. Surviving mass over release altitude for all RWL models with dashed line is the mean surviving mass for each model.
Fig. 10. Surviving mass over release altitude for RWL baseline model.
There is an extreme peak in the casualty area (fragments created) around 78 km release altitude. The peak indicates that there might have been a break-up of a sub-component and its newly created fragments have not had time enough to completely demise. Above 78 km release altitude these additional fragments have demised there is no case for the baseline model. In Fig. 12 the surviving mass of the all the RWL models are compared. The surviving mass curve for the two aluminium wheels are similar for all release altitudes while the baseline model has significantly higher surviving mass for release altitudes below 78 km. The aluminium RWL designs have a similar most-probable casualty area distribution to each other, as can be seen in Fig. 13. They also show an improvement in demisability compared to the baseline, the most probable release altitude for complete demise is 96 km for the baseline model while the two aluminium wheels demise at 87 km release altitude. This shows that the strategy of changing flywheel material from stainless steel to aluminium provided an improvement in demisability, but it should be noted that the reaction wheel still survives for release altitudes below 78km. The sub-component survivability has been compared in Fig. 14 for the last release altitude where the baseline model never demised (87 ± 2 km release altitude). The aluminium reaction wheels have similar survivability which is less than half of that of the baseline model. For the baseline model the sub-components which often survive have been identified to be the bearing assembly, flywheel and parts of the motor.
The impact of the D4D modifications on the sub-component survivability has been compared in Fig. 9 for 78 ± 2 km release altitude (the highest release altitude where the baseline model always survived). The open SADM and baseline model have similar survivability, but for the open SADM with aluminum actuator the survivability drastically drops. The most critical sub-components in the SADM baseline are the actuator bearing, potentiometer shaft and harmonic drive, and by changing actuator material to aluminium these sub-components can be made demisable. 5.2.2. Reaction wheel The mass surviving for the reaction wheel, just as for the SADM, decrease with increased release altitude (see Fig. 10). About 80% of the initial mass (around 7 kg) is surviving at 58 km release altitude, and the complete demise occurs for the first time at 94 km release altitude. There is mass surviving for all release altitudes up to 96 km. The casualty area results for the baseline model is shown in Fig. 11 with the boxplots and related spread (the minimum and maximum casualty area). It is clear from the figure that, unlike the SADM, the RWL has very little variation in the casualty area on ground. This is due to the lower complexity of the component with fewer sub-components which are able to create ground fragments. 186
P. Kärräng, T. Lips and T. Soares
Journal of Space Safety Engineering 6 (2019) 181–187
mechanism). In cooperation with the manufactures feasible design-fordemise strategies were implemented. Numerical simulations show that design-for-demise on component level clearly works and has the potential to significantly lower the casualty expectancy of a re-entry, but the solutions has to be individually tailored and tested for the specific component. Acknowledgements The work presented in this paper was partially funded by the ESA project ‘High-Fidelity Re-entry Simulation on Critical Spacecraft Platform Equipment’, ESA Contract No. 5000121149/17/NL/GLC/as. References [1] R. Kelley, Using the Design for Demise Philosophy to Reduce Casualty Risk Due to Reentering Spacecraft, 63rd International Astronautical Congress, NASA Johnson Space Center; Houston, TX, United States, 2012. https://ntrs.nasa.gov/search.jsp?R=20120002794. [2] P.M. Waswa, M. Elliot, J.A. Hoffman, Spacecraft design-for-demise implementation strategy & decision-making methodology for low earth orbit missions, Advances in Space Research 51 (9) (2013) 1627–1637. [3] S. Lemmens, Q. Funke, H. Krag, On-ground casualty risk reduction by structural design for demise, Advances in Space Research 55 (11) (2015) 2592–2606. [4] B. Fritsche, Aerodynamic categorization of spacecraft in low earth orbits, in: Proceedings of the 6th international conference on astrodynamics tools and techniques, 2016 https://indico.esa.int/event/111/contributions/285/. [5] J.D. Anderson, Hypersonic and high-temperature gas dynamics, McGraw-Hill Series in aeronautical and aerospace engineering, 1989. [6] P.A. Chambre, S.A. Schaaf, Flow of rarefied gases, Princeton University Press, 1961. [7] L. Lees, Laminar heat transfer over blunt-nosed bodies at hypersonic flight speeds, Journal of Jet Propulsion 26 (4) (1956) 259–269. [8] R. Baker, M. Weaver, Orbital spacecraft reentry breakup, in: Proceedings of the 50th International Astronautical Congress, Amsterdam, The Netherlands, 1999 IAA-99-IAA.6.7.04. [9] G. Koppenwallner, B. Fritsche, T. Lips, Survivability and ground risk potential of screws and bolts of disintegrating spacecraft during uncontrolled re-entry, in: Space Debris, volume 473, 2001, pp. 533–539. [10] B. Fritsche, T. Lips, G. Koppenwallner, Analytical and numerical re-entry analysis of simple-shaped objects, Acta Astronautica 60 (8-9) (2007) 737–751. [11] H.J. Allen, A.J. Eggers, A study of the motion and aerodynamic heating of missiles entering the earth’s atmosphere at high supersonic speeds, Technical Report, National Advisory Committee for Aeronautics, 1953. https://ntrs.nasa.gov/search.jsp?R=19930091020. [12] D.R. Chapman, An approximate analytical method for studying entry into planetary atmospheres, Technical Report, National Advisory Committee for Aeronautics, 1958. https://ntrs.nasa.gov/search.jsp?R=19930085059. [13] E. Stansbery, J. Opiela, A. Vavrin, B. Draeger, P. Anz-Meador, Debris assessment software version 2.1 user’s guide, Technical Report, 2016. [14] J. Dobarco-Otero, R. Smith, K. Bledsoe, R. Delaune, W. Rochelle, N. Johnson, The object reentry survival analysis tool (ORSAT)-version 6.0 and its application to spacecraft entry, in: Proceedings of the 56th Congress of the International Astronautical Federation, the International Academy of Astronautics, and International Institute of Space Law, IAC-05-B6, 2005, pp. 17–21. 3 [15] C. Martin, C. Brandmueller, K. Bunte, J. Cheese, B. Fritsche, H. Klinkrad, T. Lips, N. Sanchez, A debris risk assessment tool supporting mitigation guidelines, in: 4th European Conference on Space Debris, volume 587, 2005, p. 345. [16] P. Omaly, M. Spel, Debrisk, a tool for re-entry risk analysis, in: A Safer Space for Safer World, volume 699, 2012. [17] T. Lips, P. Kärräng, Casualty risk reduction by semi-controlled re-entry, in: Proceedings of the 9th international association for the advancement of space saefty conference, 2017, pp. 221–226. [18] H. Klinkrad, Space debris - Models and Risk Analysis, Springer-Verlag, Berlin Heidelberg, 2010.
Fig. 13. Most-probable casualty area curve over release altitude for all RWL models.
Fig. 14. Survivability of critical sub-components for all RWL models.
6. Conclusion In this paper we have described a method for quantifying the demisability of spacecraft components by using a most-probable casualty area curve. The curve can also be used in evaluating the effectiveness of design-for-demise modifications. The method has been applied to two critical spacecraft components (reaction wheel and solar array drive
187