Level 3 — ProductionSpacecraft Structures & Systems Engineering

Spacecraft Structures & Systems Engineering

45 minutes60 marksprintable — key stays hidden on paper

Difficulty: Level 3 — From-scratch derivations, code-from-memory, explain-out-loud Time limit: 45 minutes Total marks: 60

Answer all questions. Show all working. Where code is requested, pseudocode or clear language equivalents are acceptable if syntactically consistent.


Question 1 — Euler Column Buckling (from scratch) [12 marks]

(a) Starting from the beam-bending relation EId2wdx2=M(x)EI\,\dfrac{d^2 w}{dx^2} = M(x), derive the critical buckling load PcrP_{cr} for a pin-pinned column of length LL, flexural rigidity EIEI. State the governing ODE, apply boundary conditions, and obtain the eigenvalue result. [8]

(b) A hollow aluminium strut (E=70 GPaE = 70\ \text{GPa}) has outer diameter 40 mm, wall thickness 2 mm, and length L=1.5L = 1.5 m, pin-pinned. Compute PcrP_{cr}. [4]


Question 2 — Random Vibration & Miner's Rule [12 marks]

(a) For a single-DOF system with natural frequency fnf_n and quality factor QQ, subjected to a flat acceleration PSD of value W0W_0 (in g2/Hzg^2/\text{Hz}) at fnf_n, derive the Miles' equation for the RMS response acceleration GRMSG_{RMS}. Show the key integral. [6]

(b) Given W0=0.04 g2/HzW_0 = 0.04\ g^2/\text{Hz}, fn=200f_n = 200 Hz, Q=20Q = 20, compute GRMSG_{RMS} (in gg) and state the 3σ3\sigma peak acceleration. [3]

(c) State Miner's rule and explain what a cumulative damage value of D=1D = 1 signifies and why designers use D<1D < 1 as an allowable. [3]


Question 3 — Fracture Mechanics & Factor of Safety [10 marks]

(a) Define the stress intensity factor K=YσπaK = Y\sigma\sqrt{\pi a} and fracture toughness KICK_{IC}. A pressure-vessel wall of a titanium alloy has KIC=55 MPamK_{IC} = 55\ \text{MPa}\sqrt{\text{m}}, a geometry factor Y=1.12Y = 1.12, and operates at hoop stress σ=400\sigma = 400 MPa. Find the critical crack length aca_c. [5]

(b) If NDI (non-destructive inspection) can reliably detect cracks of 1.0 mm, compute the margin (ratio ac/adetecta_c/a_{detect}) and comment on whether the design is damage-tolerant. [3]

(c) Explain the difference between yield-based FOS and fracture-based (damage-tolerance) design philosophy in one or two sentences. [2]


Question 4 — Reliability & Redundancy (code-from-memory) [12 marks]

(a) For an exponential failure model, derive the reliability R(t)R(t) from a constant failure rate λ\lambda, and show that MTTF=1/λMTTF = 1/\lambda. [4]

(b) Two identical units each with λ=5×105 hr1\lambda = 5\times10^{-5}\ \text{hr}^{-1} are placed in active (hot) parallel redundancy. Write the system reliability Rsys(t)R_{sys}(t) and evaluate it at t=8760t = 8760 hr (1 year). [4]

(c) Write a short function (any language / pseudocode) system_reliability(lam, t, n) returning the active-parallel reliability of nn identical units. [4]


Question 5 — Link Budget [8 marks]

A LEO downlink has: transmit power Pt=5P_t = 5 W, transmit antenna gain Gt=6G_t = 6 dBi, free-space path loss 165 dB, receive antenna gain Gr=40G_r = 40 dBi, and system losses 3 dB.

(a) Compute the EIRP in dBW. [3] (b) Compute the received power PrP_r in dBW. [3] (c) Define G/TG/T in one sentence and state its units. [2]


Question 6 — Explain Out Loud [6 marks]

In your own words (3–4 sentences each):

(a) Explain the systems-engineering V-model and where verification maps onto it. [3] (b) Explain the difference between TID and SEE radiation effects and give one mitigation for each. [3]

Answer keyMark scheme & solutions

Question 1 — Euler Buckling [12]

(a) Consider a buckled pin-pinned column under axial load PP. Bending moment at position xx for deflection ww: M(x)=PwM(x) = -P\,w (compressive load produces restoring moment). [1]

Substitute into EIw=MEI\,w'' = M: EId2wdx2=Pw    w+k2w=0,k2=PEIEI\,\frac{d^2w}{dx^2} = -Pw \implies w'' + k^2 w = 0,\quad k^2 = \frac{P}{EI} [2]

General solution w(x)=Asinkx+Bcoskxw(x) = A\sin kx + B\cos kx. [1] BCs: w(0)=0B=0w(0)=0 \Rightarrow B=0; w(L)=0AsinkL=0w(L)=0 \Rightarrow A\sin kL = 0. [2] Non-trivial (A0A\neq0) requires sinkL=0kL=nπ\sin kL = 0 \Rightarrow kL = n\pi. [1] Lowest mode n=1n=1: k=π/Lk = \pi/L, so Pcr=k2EI=π2EIL2P_{cr} = k^2 EI = \dfrac{\pi^2 EI}{L^2}. [1]

(b) Second moment of area for thin annulus, Do=0.040D_o = 0.040 m, t=0.002t=0.002 m, Di=0.036D_i = 0.036 m: I=π64(Do4Di4)=π64(0.04040.0364)I = \frac{\pi}{64}(D_o^4 - D_i^4) = \frac{\pi}{64}(0.040^4 - 0.036^4) 0.0404=2.560×1060.040^4 = 2.560\times10^{-6}, 0.0364=1.6796×1060.036^4 = 1.6796\times10^{-6}; difference =8.804×107=8.804\times10^{-7}. I=0.049087×8.804×107=4.322×108 m4I = 0.049087 \times 8.804\times10^{-7} = 4.322\times10^{-8}\ \text{m}^4. [2] Pcr=π2(70×109)(4.322×108)1.52=2.987×1042.251.33×104 N13.3 kNP_{cr} = \frac{\pi^2 (70\times10^9)(4.322\times10^{-8})}{1.5^2} = \frac{2.987\times10^{4}}{2.25} \approx 1.33\times10^{4}\ \text{N} \approx 13.3\ \text{kN} [2]

Question 2 — Random Vibration & Miner [12]

(a) The response PSD of a SDOF to white-noise base input is amplified by the transmissibility. The mean-square response is GRMS2=0W0H(f)2dfG_{RMS}^2 = \int_0^\infty W_0 |H(f)|^2\,df For a lightly damped SDOF the integral of H2|H|^2 over frequency evaluates to π2fnQ\dfrac{\pi}{2} f_n Q. [3] Hence Miles' equation: GRMS=π2fnQW0G_{RMS} = \sqrt{\frac{\pi}{2} f_n Q\, W_0} [3]

(b) GRMS=π2(200)(20)(0.04)=π2×160=251.3=15.85 gG_{RMS} = \sqrt{\frac{\pi}{2}(200)(20)(0.04)} = \sqrt{\frac{\pi}{2}\times160} = \sqrt{251.3} = 15.85\ g. [2] 3σ3\sigma peak =3×15.85=47.5 g= 3 \times 15.85 = 47.5\ g. [1]

(c) Miner's rule: cumulative damage D=iniNiD = \sum_i \dfrac{n_i}{N_i} where nin_i = applied cycles at stress level ii, NiN_i = cycles to failure at that level (from S–N curve). [1] Failure predicted when D=1D = 1. [1] Designers adopt D<1D < 1 (e.g. 0.3–0.5) because Miner's rule ignores load-sequence effects and scatter, so a margin guards against premature fatigue. [1]

Question 3 — Fracture Mechanics [10]

(a) KK measures the crack-tip stress-field intensity; failure occurs when K=KICK = K_{IC}. Set Yσπac=KICY\sigma\sqrt{\pi a_c}=K_{IC}: ac=1π(KICYσ)2=1π(551.12×400)2a_c = \frac{1}{\pi}\left(\frac{K_{IC}}{Y\sigma}\right)^2 = \frac{1}{\pi}\left(\frac{55}{1.12\times400}\right)^2 [2] 55448=0.12277\frac{55}{448} = 0.12277; squared =0.015078=0.015078; /π=4.799×103/\pi = 4.799\times10^{-3} m 4.80\approx 4.80 mm. [3]

(b) Margin =ac/adetect=4.80/1.0=4.8= a_c/a_{detect} = 4.80/1.0 = 4.8. [2] Since detectable flaws are ~5× smaller than critical, the design is damage-tolerant (cracks are caught by NDI well before reaching critical size). [1]

(c) Yield-based FOS keeps operating stress below yield/ultimate by a factor; fracture/damage-tolerance design instead assumes pre-existing cracks and ensures they stay sub-critical and grow slowly enough between inspections. [2]

Question 4 — Reliability & Redundancy [12]

(a) Constant hazard rate: dRdt=λR\dfrac{dR}{dt} = -\lambda R. Solve: R(t)=eλtR(t) = e^{-\lambda t} (with R(0)=1R(0)=1). [2] MTTF=0R(t)dt=0eλtdt=1/λMTTF = \int_0^\infty R(t)\,dt = \int_0^\infty e^{-\lambda t}dt = 1/\lambda. [2]

(b) Active parallel: system fails only if both fail. Unit unreliability F=1eλtF = 1-e^{-\lambda t}. Rsys(t)=1(1eλt)2R_{sys}(t) = 1 - (1-e^{-\lambda t})^2 [2] λt=5×105×8760=0.438\lambda t = 5\times10^{-5}\times8760 = 0.438; e0.438=0.6453e^{-0.438}=0.6453; F=0.3547F = 0.3547; F2=0.12581F^2 = 0.12581. Rsys=10.12581=0.8742R_{sys} = 1 - 0.12581 = 0.8742. [2] (single unit alone would be 0.6453.)

(c)

def system_reliability(lam, t, n):
    from math import exp
    F = 1 - exp(-lam * t)      # single-unit unreliability
    return 1 - F**n            # active parallel: all n must fail

[4] (1 for exp model, 1 for unreliability, 1 for F**n, 1 for return of complement.)

(a) Pt=5P_t = 5 W =10log105=6.99= 10\log_{10}5 = 6.99 dBW. EIRP =Pt+Gt=6.99+6=12.99= P_t + G_t = 6.99 + 6 = 12.99 dBW 13.0\approx 13.0 dBW. [3]

(b) Pr=EIRPLpath+GrLsys=13.0165+403=115 dBWP_r = \text{EIRP} - L_{path} + G_r - L_{sys} = 13.0 - 165 + 40 - 3 = -115\ \text{dBW}. [3]

(c) G/TG/T is the receiver figure of merit = antenna gain divided by system noise temperature (referenced to same point); units dB/K. [2]

Question 6 — Explain Out Loud [6]

(a) V-model: left arm = decomposition (concept → requirements → design → component build); right arm = integration & verification (unit → subsystem → system → validation). Each level on the left maps horizontally to its verification activity on the right (e.g. requirements ↔ acceptance test). Verification demonstrates the built item meets the corresponding-level requirement. [3]

(b) TID (Total Ionising Dose) is cumulative charge trapping in oxides causing gradual parametric drift/failure — mitigated by shielding and rad-hard/rad-tolerant parts. SEE (Single Event Effects) are instantaneous events from a single ion (upset, latch-up) — mitigated by EDAC/TMR, watchdog resets, and current-limiting/latch-up protection. [3]

[
{"claim":"Euler Pcr for hollow strut approx 13.3 kN","code":"import math\nE=70e9; L=1.5; Do=0.040; Di=0.036\nI=math.pi/64*(Do**4-Di**4)\nPcr=math.pi**2*E*I/L**2\nresult = abs(Pcr-13300)<400"},
{"claim":"Miles GRMS approx 15.85 g","code":"import math\nG=math.sqrt(math.pi/2*200*20*0.04)\nresult = abs(G-15.85)<0.1"},
{"claim":"critical crack length approx 4.80 mm","code":"import math\nac=(1/math.pi)*(55/(1.12*400))**2\nresult = abs(ac*1000-4.80)<0.05"},
{"claim":"active parallel reliability at 1 yr approx 0.8742","code":"import math\nlam=5e-5; t=8760\nF=1-math.exp(-lam*t)\nR=1-F**2\nresult = abs(R-0.8742)<0.001"},
{"claim":"link received power = -115 dBW","code":"import math\nPt=10*math.log10(5); EIRP=Pt+6\nPr=EIRP-165+40-3\nresult = abs(Pr-(-115))<0.1"}
]