Flow Lab

Syllabus / Module 12: Applications: Flight & Aerospace

Lesson 12.4

6-DOF Aircraft Equations of Motion

Body vs Earth frames, forces (lift, drag, thrust, weight), moments, aerodynamic coefficients and the full nonlinear 6-DOF model.

Written byClaude Opus

🌊
Why this matters for simulation: The core of a flight simulator.

By the end of this lesson you can…

  • Define the earth, body and wind frames, compute $V$, $\alpha$, $\beta$, and list the twelve (or thirteen) states.
  • Derive $m(\dot u + qw - rv) = X$ and its partners from Newton's law in a rotating frame, and pair them with Euler's equations and the kinematics $\dot{\mathbf x} = R(q)\mathbf v_b$.
  • Build forces and moments from weight, thrust and aerodynamic coefficients, and read the sign of each stability derivative.
  • Write the derivative function and a fixed-step RK4 loop with renormalisation, choose the step, and explain the accumulator.
  • Find trim by Newton's method, verify that the derivative vector vanishes, and describe the short-period and phugoid response to a pitch kick.
On this page
·
  1. Frames and states
  2. Newton in a rotating frame
  3. Forces
  4. The coefficients
  5. Reconciling the mass
  6. Moments and stability derivatives
  7. The derivative function
  8. Integrating it
  9. Finding trim
  10. Worked example: trim by hand, then by Newton
  11. A disturbance: two modes
  12. Computing it
  13. Concept checks
  14. Problems
  15. Key Takeaways

12.3 Rigid-Body Dynamics & Rotations ended with rotation in hand: quaternions that never lock, Euler’s equations, and the light aircraft’s inertia tensor. 12.2 Thin Airfoil Theory & Panel Methods gave it a wing with a lift slope and induced drag. Now we join rotation to forces and moments in twelve states, position, velocity, attitude and body rates, whose derivatives one function returns. Hand it to RK4 at a fixed step, read the stick every step, draw the result: that is the core of every flight simulator.

Tools: 4.3 Runge–Kutta Methods & Butcher Tableaux, 4.5 Stability Regions & Stiffness, 7.7 Nonlinear Systems & Linearization, 12.1 Airfoils, Circulation & Lift.

▶ Video Three frames, twelve states, Newton in a rotating frame, forces from coefficients, and the fixed-step loop that flies themcaptions only
Transcript
  1. Lesson 12.3 gave us rotation, and 12.2 gave us a wing. Twelve states: position, body velocity, attitude and body rates. Our lab carries thirteen, because attitude is a quaternion. One function returns all thirteen derivatives.
  2. Three frames: earth, north east down; body, glued to the airframe; and wind, along the velocity. The body velocity gives the airspeed V, the angle of attack alpha, how far the air comes from below the nose, and the sideslip beta, from the right.
  3. Newton holds in the earth frame, forces are known in the body, so the acceleration gains a term: m times v dot plus omega cross v equals F. It is the same omega-cross term as in Euler's equations. And it is real: in a steady pull-up, Z equals minus m q u, the centripetal force.
  4. Weight, 1100 kilograms times g, is 10 791 newtons. Dynamic pressure times wing area is 25 265.6. So level flight at 50 metres per second needs a lift coefficient of 0.4271. Lessons 12.1 and 12.2 used 12 633 newtons: that was a C L of 0.5, a choice, and would mean 1287.8 kilograms.
  5. Lift: C L zero is 0.1993 from the panel solution, plus a slope of 5.313 per radian. Drag adds an induced term with factor 0.05426. Moments use assumed stability derivatives: a negative C m alpha puts the neutral point 0.1675 chord behind the centre of gravity.
  6. The derivative function takes thirteen numbers in and returns thirteen out: air data, coefficients, forces, moments. Runge Kutta four calls it four times a step, then the quaternion is renormalised.
  7. Trim: three unknowns, three equations. Newton converges in 3 iterations: alpha 2.6606 degrees, elevator minus 2.7452, throttle 0.3880, thrust 931.2 newtons. Check the balance: thrust cosine alpha is 930.2, equal to drag; lift plus thrust sine alpha is 10 747.8 plus 43.2, equal to the weight.
  8. Is the step small enough? Kick the trimmed aircraft and fly 30 seconds. Altitude errors at 0.05, 0.02 and 0.01 seconds: 5.36 times 10 to the minus 6, 1.39 times 10 to the minus 7, 8.72 times 10 to the minus 9 metres. Ratios 38.5 and 15.9: fourth order. Accuracy does not set the step; stability, 0.296 seconds, and input sampling do.
  9. Now kick the pitch rate by 0.2 radians per second. This is the short period: pitch rises from 2.66 degrees to 4.71 at 0.41 seconds, then settles, restored by C m alpha and damped by C m q.
  10. Then the slow motion: the flight path points up, the aircraft climbs and slows, down to 49.05 metres per second at 6.73 seconds, and peaks at 1005.10 metres at 7.26. The next peak is 1003.64 metres at 35.24 seconds: a phugoid, period 27.98 seconds, kinetic and potential energy trading places.
  11. A 30 degree level turn. Lift must carry the weight: load factor 1.1547, so C L is 0.4932. Its horizontal part curves the path at 0.11328 radians per second. In body axes that is pitch rate 0.0566 and yaw rate 0.0981: the omega cross v terms are the centripetal acceleration.
  12. Real time and physics time stay apart. Each frame adds the elapsed time to an accumulator, takes fixed steps of 0.01 seconds while one fits, and carries the remainder. Physics runs at 100 Hz whatever the screen does, 60 Hz, or 144, so a flight repeats bit for bit.
  13. Honest limits. The derivatives are typical values from memory. There is no stall: C L is capped at 1.4, a label, not physics. Flat, non-rotating Earth, still air, no engine dynamics.
  14. Next, 12.5: linearise these equations about this trim. The eigenvalues of the matrix A give the short period and the phugoid we just watched, read straight off the complex plane.

Frames and states

  • Earth frame (flat, non-rotating, inertial): xx north, yy east, zz down (NED). Altitude is h=−zh = -z.
  • Body frame, glued to the airframe at the centre of gravity: xbx_b out of the nose, yby_b out of the right wing, zbz_b out of the belly. Inertia and aerodynamic tables are constant here.
  • Wind frame: xwx_w along the velocity relative to the air; lift and drag are perpendicular and parallel to it.

With body velocity vb=(u,v,w)\mathbf v_b = (u, v, w), the airspeed and the two flow angles are

V=∣vb∣,α=arctan⁡wu,β=arcsin⁡vV.V = |\mathbf v_b|, \qquad \alpha = \arctan\frac{w}{u}, \qquad \beta = \arcsin\frac{v}{V}.

Angle of attack α\alpha: how far the air comes from below the nose; sideslip β\beta: from the right. In the sim the solid line is xbx_b and the dashed line is the velocity: seen from the side, the angle between them is α\alpha.

The states. Six degrees of freedom, each with a position and a rate:

groupsymbolsframecount
positionx,y,zx, y, zearth3
velocityu,v,wu, v, wbody3
attitudeϕ,θ,ψ\phi, \theta, \psi or q0,q1,q2,q3q_0, q_1, q_2, q_3body to earth3 or 4
angular velocityp,q,rp, q, rbody3

Textbooks count twelve with Euler angles. The lab carries thirteen, with a unit quaternion, for the reason 12.3 measured: Euler-angle rates contain 1/cos⁡θ1/\cos\theta, and a loop puts the nose through θ=90∘\theta = 90^\circ. Quaternion kinematics are linear and singularity-free, for one extra number and a renormalisation. Euler angles are computed only for display. (The letter qq now does double duty, pitch rate and quaternion; context decides.)

Newton in a rotating frame

Newton’s law holds in the earth frame, m ddtv=Fm\,\frac{d}{dt}\mathbf v = \mathbf F, but the forces are known in body components. 12.3 showed that for any vector stored by its body components b\mathbf b, ddt(Rb)=R(b˙+ω×b)\frac{d}{dt}(R\mathbf b) = R(\dot{\mathbf b} + \boldsymbol\omega\times\mathbf b): the change seen from the body plus the turning of the frame. Put b=vb\mathbf b = \mathbf v_b:

m(v˙b+ω×vb)=Fb.m\big(\dot{\mathbf v}_b + \boldsymbol\omega\times\mathbf v_b\big) = \mathbf F_b .

The same ω×\boldsymbol\omega\times term as Euler’s ω×Iω\boldsymbol\omega\times I\boldsymbol\omega, acting on momentum. In components, with ω=(p,q,r)\boldsymbol\omega = (p, q, r) and Fb=(X,Y,Z)\mathbf F_b = (X, Y, Z),

m(u˙+qw−rv)=X,m(v˙+ru−pw)=Y,m(w˙+pv−qu)=Z.m(\dot u + qw - rv) = X, \qquad m(\dot v + ru - pw) = Y, \qquad m(\dot w + pv - qu) = Z .

Why the extra terms are real. In a steady pull-up at pitch rate qq, w≈0w \approx 0, the velocity turns with the body, so u˙=w˙=0\dot u = \dot w = 0, yet Z=−mquZ = -mqu: an upward force mqumqu, the centripetal force for a path of radius u/qu/q.

The rotational equations are 12.3’s, with the full tensor (including IxzI_{xz}):

Iω˙+ω×(Iω)=Mb,I=[1527.30−107.602462.00−107.603613.1] kg m2.I\dot{\boldsymbol\omega} + \boldsymbol\omega\times(I\boldsymbol\omega) = \mathbf M_b, \qquad I = \begin{bmatrix} 1527.3 & 0 & -107.6 \\ 0 & 2462.0 & 0 \\ -107.6 & 0 & 3613.1\end{bmatrix}\ \text{kg m}^2 .

The code solves this 3×33\times3 system for ω˙\dot{\boldsymbol\omega}; through IxzI_{xz} a rolling moment also starts a little yaw. The kinematics close the system:

x˙=R(q) vb,q˙=12 q⊗(0,ω).\dot{\mathbf x} = R(q)\,\mathbf v_b, \qquad \dot q = \tfrac12\,q\otimes(0, \boldsymbol\omega).

Thirteen nonlinear ODEs, one function.

Forces

Weight is mgmg straight down in the earth frame, so in body axes it is RT(0,0,mg)R^{\mathsf T}(0, 0, mg). With 3-2-1 angles that is mg(−sin⁡θ, sin⁡ϕcos⁡θ, cos⁡ϕcos⁡θ)mg(-\sin\theta,\ \sin\phi\cos\theta,\ \cos\phi\cos\theta) (Problem 3); the code uses the bottom row of R(q)R(q) and never needs the angles.

Thrust is T=δt Tmax⁡T = \delta_t\,T_{\max} along xbx_b through the centre of gravity, throttle δt∈[0,1]\delta_t \in [0, 1], Tmax⁡=2400T_{\max} = 2400 N, constant with speed (an assumption: a real propeller’s thrust falls with airspeed).

Aerodynamics use the dynamic pressure qˉ=12ρV2\bar q = \tfrac12\rho V^2 (sea-level ρ=1.225\rho = 1.225 kg/m3^3) and the wing area S=11×1.5=16.5S = 11\times1.5 = 16.5 m2^2:

L=qˉSCL,D=qˉSCD,Ya=qˉSCY.L = \bar qSC_L, \quad D = \bar qSC_D, \quad Y_a = \bar qSC_Y .

Rotating lift and drag by α\alpha into body axes gives

Xa=qˉS(CLsin⁡α−CDcos⁡α),Za=−qˉS(CLcos⁡α+CDsin⁡α),X_a = \bar qS(C_L\sin\alpha - C_D\cos\alpha), \qquad Z_a = -\bar qS(C_L\cos\alpha + C_D\sin\alpha),

resolved with α\alpha only, which is accurate while β\beta is small.

The coefficients

Lift. From 12.2’s panel solution for the NACA 2412 wing, with the lifting-line correction: slope a=5.313a = 5.313 per radian and zero-lift angle αL0=−2.149∘\alpha_{L0} = -2.149^\circ. So

CL=CL0+CLαα+CLδeδe,CL0=−a αL0=0.1993,CLα=5.313.C_L = C_{L0} + C_{L\alpha}\alpha + C_{L\delta_e}\delta_e, \qquad C_{L0} = -a\,\alpha_{L0} = 0.1993, \quad C_{L\alpha} = 5.313 .

The elevator adds CLδe=0.43C_{L\delta_e} = 0.43 per radian (from memory); the tail’s own lift slope is ignored. There is no stall model: CLC_L is capped at 1.41.4, and the HUD says so in red when the cap acts.

Drag polar. CD=CD0+CL2/(πe AR)C_D = C_{D0} + C_L^2/(\pi e\,AR) with AR=112/16.5=7.333AR = 11^2/16.5 = 7.333. 12.2 used the elliptic e=1e = 1; a rectangular planform does worse, so we take e=0.8e = 0.8 and CD0=0.027C_{D0} = 0.027 (typical values, from memory).

Reconciling the mass

12.1 and 12.2 used a lift of 12 63312\,633 N at CL=0.5C_L = 0.5: a choice of CLC_L, not a weight (it would mean 1287.81287.8 kg). 12.3’s inertia assumed 1100 kg, and the sim uses that mass, because the tensor above belongs to it. Then

W=mg=10 791 N,qˉS=12(1.225)(502)(16.5)=25 265.6 N,CL=WqˉS=0.4271W = mg = 10\,791\ \text{N}, \qquad \bar qS = \tfrac12(1.225)(50^2)(16.5) = 25\,265.6\ \text{N}, \qquad C_L = \frac{W}{\bar qS} = 0.4271

at 5050 m/s. Ignoring the elevator, α=(0.4271−0.1993)/5.313\alpha = (0.4271 - 0.1993)/5.313 rad =2.457∘= 2.457^\circ.

Moments and stability derivatives

Moments use a length too: span b=11b = 11 m for roll and yaw, chord c=1.5c = 1.5 m for pitch. Rates enter non-dimensionally, p^=pb/2V\hat p = pb/2V, q^=qc/2V\hat q = qc/2V, r^=rb/2V\hat r = rb/2V:

L=qˉSb (Clββ+Clpp^+Clrr^+Clδaδa),M=qˉSc (Cm0+Cmαα+Cmqq^+Cmδeδe),N=qˉSb (Cnββ+Cnrr^+Cnδrδr).\begin{aligned} \mathcal L &= \bar qSb\,\big(C_{l\beta}\beta + C_{lp}\hat p + C_{lr}\hat r + C_{l\delta_a}\delta_a\big),\\ \mathcal M &= \bar qSc\,\big(C_{m0} + C_{m\alpha}\alpha + C_{mq}\hat q + C_{m\delta_e}\delta_e\big),\\ \mathcal N &= \bar qSb\,\big(C_{n\beta}\beta + C_{nr}\hat r + C_{n\delta_r}\delta_r\big). \end{aligned}

Each stability derivative is a slope of a moment coefficient, the first term of a Taylor series about level flight (7.7 Nonlinear Systems & Linearization). The values are stated assumptions typical of a light aircraft, roughly Cessna-172 class and quoted from memory, not derived; Cm0C_{m0} is chosen here.

derivativevaluesign means
CmαC_{m\alpha}−0.89-0.89nose-up disturbance, nose-down moment: statically stable
CmqC_{mq}−12.4-12.4pitching swings the tail, whose changed angle of attack opposes the motion
CmδeC_{m\delta_e}−1.28-1.28trailing edge down pushes the nose down
Cm0C_{m0}−0.02-0.02cambered wing’s nose-down moment, partly offset by the tail
ClpC_{lp}−0.47-0.47the down-going wing meets a higher α\alpha: roll damping
ClβC_{l\beta}−0.089-0.089dihedral: a slip to the right rolls left
ClrC_{lr}, ClδaC_{l\delta_a}0.0960.096, 0.1780.178yawing right speeds the left wing; positive aileron rolls right
CnβC_{n\beta}0.0650.065weathercock: the fin swings the nose into the wind
CnrC_{nr}, CnδrC_{n\delta_r}−0.099-0.099, 0.06570.0657yaw damping; right pedal yaws right
CYβC_{Y\beta}, CYδrC_{Y\delta_r}−0.31-0.31, −0.187-0.187side force against the slip; the rudder’s push

Static margin. Extra lift is CLαΔαC_{L\alpha}\Delta\alpha and its moment CmαΔαC_{m\alpha}\Delta\alpha, so it acts −Cmα/CLα=0.89/5.313=0.1675-C_{m\alpha}/C_{L\alpha} = 0.89/5.313 = 0.1675 chord behind the centre of gravity, at the neutral point: the static margin, studied in 12.5.

The derivative function

Thirteen numbers in, thirteen out:

  1. Unpack vb\mathbf v_b, qq, ω\boldsymbol\omega; form R(q)R(q) from the normalised quaternion.
  2. Air data: VV, α\alpha, β\beta, qˉ\bar q, and the non-dimensional rates.
  3. Coefficients, then forces: aerodynamic plus thrust plus RT(0,0,mg)R^{\mathsf T}(0, 0, mg).
  4. Moments from the coefficients.
  5. Return (Rvb, F/m−ω×vb, 12q⊗(0,ω), I−1(M−ω×Iω))\big(R\mathbf v_b,\ \mathbf F/m - \boldsymbol\omega\times\mathbf v_b,\ \tfrac12q\otimes(0, \boldsymbol\omega),\ I^{-1}(\mathbf M - \boldsymbol\omega\times I\boldsymbol\omega)\big).

The controls (δe,δa,δr,δt)(\delta_e, \delta_a, \delta_r, \delta_t) are a second argument, constant within a step. The exercise asks for exactly this.

Integrating it

RK4 at a fixed step hh, then q←q/∣q∣q \leftarrow q/|q|. The sim’s step check kicks the trimmed aircraft (q=0.2q = 0.2 rad/s), flies 30 s and compares altitude with h=0.002h = 0.002:

hh (s)altitude error at 30 s (m)
0.055.36×10−65.36\times10^{-6}
0.021.39×10−71.39\times10^{-7}
0.018.72×10−98.72\times10^{-9}

Ratios 38.538.5, 15.915.9 against 2.54=392.5^4 = 39 and 24=162^4 = 16: fourth order. Micrometres at h=0.05h = 0.05, so accuracy does not set the step here. Two other things do.

Stability. The fastest mode is roll damping: IxxI_{xx} is the smallest moment and ClpC_{lp} acts across the whole span. Alone, Ixxp˙=qˉSb Clp pb/2VI_{xx}\dot p = \bar qSb\,C_{lp}\,pb/2V, so p˙=λp\dot p = \lambda p with λ=ClpqˉSb2/(2VIxx)=−9.41\lambda = C_{lp}\bar qSb^2/(2VI_{xx}) = -9.41 s−1^{-1} (lab cell). RK4’s stability region reaches −2.785-2.785 on the real axis (4.5 Stability Regions & Stiffness), so h<0.296h \lt 0.296 s. Faster aircraft and stiff landing-gear springs push this down.

Inputs and events. The stick is read, and the ground tested, once per step; at h=0.01h = 0.01 s (100 Hz) that is finer than a hand moves.

Finding trim

Trim is a state whose derivatives vanish, apart from steady motion. Level flight at 5050 m/s: wings level, ω=0\boldsymbol\omega = 0, θ=α\theta = \alpha so the velocity is horizontal. Three unknowns (α,δe,δt)(\alpha, \delta_e, \delta_t), three equations u˙=w˙=q˙=0\dot u = \dot w = \dot q = 0; the rest vanish by symmetry, except x˙=50\dot x = 50.

Worked example: trim by hand, then by Newton

Pitching moment. Cm=0C_m = 0 gives δe=−(Cm0+Cmαα)/Cmδe\delta_e = -(C_{m0} + C_{m\alpha}\alpha)/C_{m\delta_e}. At α=2.457∘\alpha = 2.457^\circ: δe=−(−0.02−0.89⋅0.04288)/(−1.28)=−2.604∘\delta_e = -(-0.02 - 0.89\cdot0.04288)/(-1.28) = -2.604^\circ, trailing edge up.

Lift. Up-elevator removes 0.43⋅0.04544=0.01950.43\cdot0.04544 = 0.0195 of CLC_L, so α\alpha must rise to 2.668∘2.668^\circ. Repeat both: the lab cell’s fixed-point loop settles at α=2.680∘\alpha = 2.680^\circ, δe=−2.758∘\delta_e = -2.758^\circ.

Drag and thrust. CD=0.027+0.05426⋅0.42712=0.0369C_D = 0.027 + 0.05426\cdot0.4271^2 = 0.0369, so D=932.2D = 932.2 N and δt≈0.388\delta_t \approx 0.388.

Newton. The hand estimate ignores that thrust, tilted up by α\alpha, carries Tsin⁡αT\sin\alpha of the weight. Newton on the three residuals, with a finite-difference Jacobian, converges in 3 iterations in the sim:

α=θ=2.6606∘,δe=−2.7452∘,δt=0.3880 (T=931.2 N).\alpha = \theta = 2.6606^\circ, \qquad \delta_e = -2.7452^\circ, \qquad \delta_t = 0.3880\ (T = 931.2\ \text{N}).

Check the balance in wind axes: Tcos⁡α=930.2T\cos\alpha = 930.2 N =D= D, and L+Tsin⁡α=10 747.8+43.2=10 791L + T\sin\alpha = 10\,747.8 + 43.2 = 10\,791 N =W= W. So CL=0.4254C_L = 0.4254. The largest state derivative at trim is of order 10−1510^{-15} (sim readout), and a 30 s hold changes the altitude by exactly 00 m. Press trim: the HUD shows 50.0050.00 m/s, α=2.66∘\alpha = 2.66^\circ and load factor n=1.00n = 1.00 (aerodynamic −Z-Z over WW; just under 1, as thrust carries part of the weight).

◆ Interactive The light aircraft flown by a 13-state RK4 loop: stick, rudder and throttle by keyboard or sliders, HUD air data, trim, pitch kick, and altitude and airspeed charts

Loading simulation…

A disturbance: two modes

Press pitch kick: qq jumps by 0.20.2 rad/s, controls at trim. The readout and charts show two motions on very different clocks.

  • Short period. The pitch rises from 2.66∘2.66^\circ to a peak of 4.71∘4.71^\circ at 0.410.41 s and settles back to about 3.9∘3.9^\circ within a second and a half. That is CmαC_{m\alpha} restoring α\alpha, damped hard by CmqC_{mq}.
  • Phugoid. The kick has bent the flight path upward. With α\alpha back near trim the aircraft climbs and slows (slowest 49.0549.05 m/s at 6.736.73 s, altitude peak 1005.101005.10 m at 7.267.26 s), then dives and speeds up; the next peak is 1003.641003.64 m at 35.2435.24 s, a period of 27.9827.98 s, decaying slowly. Kinetic and potential energy trade back and forth.

Time warp 4x shows several phugoid cycles; the step stays 0.010.01 s. Positive aileron rolls right at a steady rate set by ClδaC_{l\delta_a} against ClpC_{lp}. Centre it: the bank decays slowly and the heading swings, but tilted lift no longer carries the weight, so the nose drops and the aircraft dives unless you pull back (Problem 1).

Computing it

🐍 PythonLab: mass reconciliation, trim by hand, static margin and the RK4 step limit
⌘/Ctrl + Enter
🧪 Coding exerciseExercise: the 13-state derivative function, trim by Newton, and the quaternion norm
⌘/Ctrl + Enter

Graded: Newton on your ff lands on the sim’s trim (2.6606∘2.6606^\circ, −2.7452∘-2.7452^\circ, 0.38800.3880); every derivative except x˙=50\dot x = 50 vanishes there to 10−910^{-9}; 30 s of RK4 holds the altitude to a micrometre; and a 5 s tumble at h=0.05h = 0.05 keeps ∣q∣=1|q| = 1 to 10−1210^{-12}, which needs the renormalisation line.

Concept checks

✅ Concept check

The 1100 kg light aircraft ($S = 16.5$ m$^2$) flies level at $50$ m/s in sea-level air. Ignoring thrust's share, what lift coefficient does it need?

✅ Concept check

In $m(\dot w + pv - qu) = Z$, what is the $-qu$ term?

✅ Concept check

Why does the lab carry 13 states with a quaternion instead of 12 with Euler angles?

✅ Concept check

With $C_{m\alpha} = -0.89$ and $C_{L\alpha} = 5.313$, how far behind the centre of gravity, in chords, does extra lift act?

✅ Concept check

Why does the sim loop use a fixed step with an accumulator instead of stepping by each frame's elapsed time?

Problems

✏️ Pen & paperA level turn●●○

The aircraft holds a level turn at bank ϕ=30∘\phi = 30^\circ, V=50V = 50 m/s, no sideslip. (a) Show the load factor is n=1/cos⁡ϕn = 1/\cos\phi and find the CLC_L needed (thrust’s share ignored). (b) Find the turn rate Ω\Omega and the body rates qq and rr. (c) Which terms of the body-axis equations carry the centripetal acceleration?

Show worked solution

(a) Lift tilts with the wings; its vertical part must carry the weight, Lcos⁡ϕ=WL\cos\phi = W, so n=L/W=1/cos⁡30∘=1.1547n = L/W = 1/\cos 30^\circ = 1.1547 and CL=0.4271⋅1.1547=0.4932C_L = 0.4271\cdot1.1547 = 0.4932.

(b) The horizontal part Lsin⁡ϕ=Wtan⁡ϕL\sin\phi = W\tan\phi curves the path: mVΩ=mgtan⁡ϕmV\Omega = mg\tan\phi, so Ω=9.81⋅0.57735/50=0.11328\Omega = 9.81\cdot0.57735/50 = 0.11328 rad/s, a full circle in 55.555.5 s. The turn is about earth zz; in body axes, tilted by ϕ\phi, that is q=Ωsin⁡ϕ=0.0566q = \Omega\sin\phi = 0.0566 rad/s and r=Ωcos⁡ϕ=0.0981r = \Omega\cos\phi = 0.0981 rad/s.

(c) With u≈Vu \approx V, v˙b=0\dot{\mathbf v}_b = 0: the −qu-qu in the ww equation and the ruru in the vv equation. Together they are ω×vb\boldsymbol\omega\times\mathbf v_b, of size ΩV\Omega V, pointing toward the centre of the turn. The induced drag rises 1.15472=1.3331.1547^2 = 1.333 times, so the throttle must go up.

✏️ Pen & paperTrim by hand●●○

Using CL0=0.1993C_{L0} = 0.1993, CLα=5.313C_{L\alpha} = 5.313, CLδe=0.43C_{L\delta_e} = 0.43, Cm0=−0.02C_{m0} = -0.02, Cmα=−0.89C_{m\alpha} = -0.89, Cmδe=−1.28C_{m\delta_e} = -1.28 and CL=0.4271C_L = 0.4271: (a) solve the two linear equations CL=C_L = required and Cm=0C_m = 0 for α\alpha and δe\delta_e directly. (b) Newton found α=2.6606∘\alpha = 2.6606^\circ, δe=−2.7452∘\delta_e = -2.7452^\circ. Explain the difference.

Show worked solution

(a) From Cm=0C_m = 0: δe=−(Cm0+Cmαα)/Cmδe=−0.015625−0.6953α\delta_e = -(C_{m0} + C_{m\alpha}\alpha)/C_{m\delta_e} = -0.015625 - 0.6953\alpha. Into the lift equation: 0.1993+5.313α+0.43(−0.015625−0.6953α)=0.42710.1993 + 5.313\alpha + 0.43(-0.015625 - 0.6953\alpha) = 0.4271, so 5.014α=0.23455.014\alpha = 0.2345, α=0.04678\alpha = 0.04678 rad =2.680∘= 2.680^\circ and δe=−0.04814\delta_e = -0.04814 rad =−2.758∘= -2.758^\circ, as the lab cell’s loop found.

(b) The tilted thrust carries Tsin⁡α=43.2T\sin\alpha = 43.2 N of the weight, so the wing needs only CL=0.4254C_L = 0.4254 and a little less α\alpha; the elevator follows through Cm=0C_m = 0. Newton solves the full nonlinear equations.

✏️ Pen & paperWeight in body axes●●●

(a) With R=Rz(ψ)Ry(θ)Rx(ϕ)R = R_z(\psi)R_y(\theta)R_x(\phi), show that RT(0,0,g)=g(−sin⁡θ, sin⁡ϕcos⁡θ, cos⁡ϕcos⁡θ)R^{\mathsf T}(0, 0, g) = g(-\sin\theta,\ \sin\phi\cos\theta,\ \cos\phi\cos\theta). (b) At trim (θ=α=2.6606∘\theta = \alpha = 2.6606^\circ, ϕ=0\phi = 0) the body XX force (aero plus thrust) is 500.9500.9 N. Show it balances the weight’s xx component, and explain why u˙=0\dot u = 0 even though the aircraft is pointing slightly nose-up.

Show worked solution

(a) RT(0,0,1)R^{\mathsf T}(0, 0, 1) is the third row of RR, the earth zz axis in body components. RzR_z leaves zz alone; Ry(θ)TR_y(\theta)^{\mathsf T} takes it to (−sin⁡θ,0,cos⁡θ)(-\sin\theta, 0, \cos\theta); Rx(ϕ)TR_x(\phi)^{\mathsf T} then mixes the last two into (sin⁡ϕcos⁡θ,cos⁡ϕcos⁡θ)(\sin\phi\cos\theta, \cos\phi\cos\theta).

(b) X=qˉS(CLsin⁡α−CDcos⁡α)+T=498.9−929.2+931.2=500.9X = \bar qS(C_L\sin\alpha - C_D\cos\alpha) + T = 498.9 - 929.2 + 931.2 = 500.9 N, and mgsin⁡θ=10 791⋅0.04642=500.9mg\sin\theta = 10\,791\cdot0.04642 = 500.9 N. Weight pulls back along the nose by mgsin⁡θmg\sin\theta; the forward-tilted lift and the thrust push forward by the same amount (in wind axes, Tcos⁡α=DT\cos\alpha = D). With ω=0\boldsymbol\omega = 0 there is no rotating-frame term, so u˙=0\dot u = 0.

Key Takeaways

  • Earth, body and wind frames; V=∣vb∣V = |\mathbf v_b|, α=arctan⁡(w/u)\alpha = \arctan(w/u), β=arcsin⁡(v/V)\beta = \arcsin(v/V). Twelve states, thirteen with the quaternion the lab uses to avoid gimbal lock.
  • Newton in the rotating body frame: m(v˙b+ω×vb)=Fbm(\dot{\mathbf v}_b + \boldsymbol\omega\times\mathbf v_b) = \mathbf F_b, beside Euler’s Iω˙+ω×Iω=MbI\dot{\boldsymbol\omega} + \boldsymbol\omega\times I\boldsymbol\omega = \mathbf M_b and x˙=R(q)vb\dot{\mathbf x} = R(q)\mathbf v_b.
  • Forces from coefficients, CL=0.1993+5.313α+0.43δeC_L = 0.1993 + 5.313\alpha + 0.43\delta_e and a drag polar with AR=7.333AR = 7.333; moments from assumed stability derivatives (Cmα=−0.89C_{m\alpha} = -0.89: static margin 0.16750.1675 chord).
  • The 1100 kg aircraft trims at 5050 m/s with α=2.6606∘\alpha = 2.6606^\circ, δe=−2.7452∘\delta_e = -2.7452^\circ, T=931.2T = 931.2 N; the derivatives vanish to about 10−1510^{-15}.
  • RK4 at h=0.01h = 0.01 s is accurate to 10−810^{-8} m over 30 s; stability (h<0.296h \lt 0.296 s for roll damping) and input sampling set the step. A fixed step with an accumulator makes flight deterministic.
  • A pitch kick excites a fast, well-damped short period (pitch peak at 0.410.41 s) and a slow phugoid (period 27.9827.98 s).
  • Next, 12.5: linearise these equations about this trim, form the longitudinal matrix AA, and its eigenvalues give the short period and the phugoid we just watched, periods and damping read straight off the complex plane.
Finished this lesson?

Mark it complete to track your progress (saved in this browser). Reading every section does this for you.

Where this leads

These later lessons build directly on this one: