Case StudyExperiment

Impact Detection in Sensorised Structures

A machine learning approach to identifying impact events in sensorised structural panels using measured response data.

SHMSignal ProcessingClassificationSensors

Problem

In aerospace, civil, and mechanical structures, detecting impact or damage events can be important for safety, maintenance, and long-term monitoring.

The challenge is that structural response signals can be noisy, sparse, and sensitive to sensor placement, boundary conditions, and material behaviour.

Approach

This project explores the use of sensor data and machine learning methods to infer impact-related information from structural response signals.

Sensor data processing
Feature extraction
Model training
Validation
Performance analysis

Output

Processed dataset
Model pipeline
Evaluation results
Visualisations
Technical report

Commercial angle

This connects to structural health monitoring, predictive maintenance, aerospace safety, infrastructure monitoring, and digital twin systems.

Next steps

Improve robustness under noise
Test generalisation across structures
Explore uncertainty-aware predictions

Proof artifacts

Signal-processing pipeline

Next to publish

A documented preprocessing path from raw sensor response to model-ready features, including filtering and windowing assumptions.

Sensor placement matrix

Planned

A small table for comparing how prediction quality changes with sensor location, signal quality, and missing-channel scenarios.

Noise robustness check

Drafting

An evaluation note showing how classification confidence changes as signal noise and boundary-condition variation increase.

This is an in-progress research project. Status, methods, and results are exploratory and will be updated as the work develops.