Richard Olamide Olanite
Back to all work

BeamTwin: a digital twin that finds damage in beams from their vibration

When a beam cracks it gets less stiff, and its natural frequencies drop. BeamTwin works backwards from those frequencies to say where the damage is and how bad it is.

0.001%FEM vs textbook frequency error, first four modes
55×lower surrogate error with physics built in
0.3%median location error, low-noise sensors
0.03 sper damage estimate

The problem

Vibration-based structural health monitoring rests on a simple idea: damage lowers stiffness, and lower stiffness lowers natural frequencies. Turning a few measured frequencies back into a damage location is an inverse problem, and solving it needs thousands of simulations per estimate.

A finite element model is accurate but too slow to run thousands of times on demand. The question was whether a neural network could stand in for it without losing the small frequency shifts that damage causes (about 1.5% on average).

What I did

  1. Wrote an Euler-Bernoulli beam finite element solver from scratch and validated it against textbook formulas: tip deflection matches PL³/3EI exactly, and the first four natural frequencies agree to within 0.001%.
  2. Trained two identical networks on the solver's results. The plain one learns everything from 7 inputs. The hybrid one gets the healthy frequencies from the exact cantilever formula and learns only the damage-induced ratio, which dimensional analysis shows depends on just 2 inputs.
  3. Built a Bayesian detector that scores 18,281 candidate damage states in one batch with the surrogate, and reports the most likely location and severity with 90% ranges and the probability that the beam is damaged at all.
  4. Served it with FastAPI and Docker, tested with pytest in GitHub Actions, and deployed a web demo where you hide damage in a beam and watch the twin find it.

Results

Predicted vs true damage-induced frequency drop on 600 unseen beams. The hybrid surrogate (physics built in) sits on the diagonal.
Predicted vs true damage-induced frequency drop on 600 unseen beams. The hybrid surrogate (physics built in) sits on the diagonal.
Probability maps of damage location and severity from noisy vibration data: an easy case near the clamp, and a harder one near the free tip, where the estimate is close but slightly off.
Probability maps of damage location and severity from noisy vibration data: an easy case near the clamp, and a harder one near the free tip, where the estimate is close but slightly off.
Plain networkHybrid (physics + network)
Mean frequency error0.38–0.42%0.006–0.008%
Worst frequency error2.1–2.8%0.04–0.06%
Signal-to-error ratio3–4190–240

Test set: 600 unseen beams. Damage lowers frequencies by about 1.5% on average, so the plain network's error is too large to see it reliably.

Sensor noiseLowMediumHigh
Location error, median0.3% of length0.5%1.3%
Damaged beams detected100%98%100%
False alarms on healthy beams0%4%32%

Detector accuracy on 200 damaged and 50 healthy simulated beams per noise level.

Limitations

  • All results so far use simulated data. Validation on a physical clamped steel strip is the next step.
  • Uniform beams and a single damage zone only. The detector assumes geometry and material are known exactly.
  • With noisy sensors (±0.5% frequencies), about a third of healthy beams raise a false alarm.
  • The surrogate is only valid inside its training ranges (0.5–2 m long, metals, up to 50% stiffness loss).