25-Nav-B2 Marine Engineering and Vibrations · May 2017
Question 7 of 7: First Torsional Natural Frequency by Holzer's Method
Nivaar worked solution (AI-drafted; not reviewed by a licensed engineer)
Notes on this paper
Paper format: National Exams, May 2017; closed book, 3 hours (Casio/Sharp approved calculator only); seven numbered problems of equal value (20 marks each), of which any five constitute a complete paper (only the first five appearing in the answer book are marked). All seven are solved below so the set is complete for study.
Reference texts. Shigley, Shigley's Mechanical Engineering Design (11th) – fatigue & shaft design; ABS, Rules for Building and Classing Steel Vessels (2009) – propulsion shafting; Hibbeler, Structural Analysis (10th) – three-moment equation; Fox & McDonald, Introduction to Fluid Mechanics (10th) – pump & pipe systems; Incropera, Fundamentals of Heat and Mass Transfer (8th) – LMTD heat exchangers; Wilson & Sadler, Kinematics and Dynamics of Machinery (3rd) – reciprocating balance; Rao, Mechanical Vibrations (6th) – Holzer's method.
Question 7: First Torsional Natural Frequency by Holzer's Method (20 marks)
Given. A free–free, six-inertia torsional chain: propeller inertia $J_P=0.14$ kg·m² and gearbox/flywheel inertia $J_0=1.6$ kg·m² coupled by stiffness $K_1=2.6\times10^5$ N·m/rad, then four equal engine-cylinder inertias $J=0.75$ kg·m² each, each pair coupled by stiffness $K=4.2\times10^5$ N·m/rad.
Given data
Symbol
Value
JP
0.14 kg·m²
J0
1.6 kg·m²
J (×4)
0.75 kg·m² each
K1
2.6×10⁵ N·m/rad
K (×4 segments)
4.2×10⁵ N·m/rad
Find. The first (lowest non-zero) natural torsional frequency of the chain, expected between 300 and 400 rad/s.
Free–free torsional lumped-inertia model: propeller (JP) and gearbox (J0) coupled through K1, then four engine cylinders (J) each coupled through K.
Approach. Assume unit angular amplitude at JP, march station-by-station through Holzer's recursion – at each inertia accumulate the running inertia torque and use it to find the next station's amplitude across the connecting stiffness – and search for the trial frequency at which the residual torque left over after the last (free) inertia vanishes; that frequency is a natural frequency of the free–free system.
Holzer recursion. With $\theta_1=1$ (at $J_P$) and running torque sum $\Sigma J_k\omega^2\theta_k$, each subsequent amplitude is $$\theta_{i+1}=\theta_i-\frac{\sum_{k\le i}J_k\omega^2\theta_k}{K_i}$$ where $K_i$ is the stiffness connecting station $i$ to $i+1$ ($K_1$ first, then $K$ four times). A trial $\omega$ is a natural frequency of the free–free chain when the FULL running torque sum (after including the last, free-end inertia $J_4$) comes out to zero – that end has no stiffness left to react a residual torque against.
Scanning the stated range. Evaluating the residual torque at trial frequencies swept from 300 to 400 rad/s (the range the question itself suggests) shows exactly one sign change, between 388.0 and 388.5 rad/s.
Bisecting to the root. Refining by bisection inside that bracket converges to $$\boxed{\omega_1 \approx 388.4\ \text{rad/s}}$$ with residual torque $<10^{-9}$ N·m at convergence, comfortably inside the suspected 300–400 rad/s band. The corresponding (unnormalised) mode shape is $\theta = [1.000,\ 0.919,\ 0.340,\ -0.330,\ -0.911,\ -1.247]$ for $[J_P, J_0, J_1, J_2, J_3, J_4]$ – the sign change partway along the engine's four cylinders (between $J_2$ and $J_3$) marks the node of this first torsional mode.