GopalKrishna Advanced Rural Research Foundation
GARRF Virtual Engineering & Electronics Laboratory
KALAM ZERO RESEARCH FUNDING LAB
EXPERIMENT 25
CAPSTONE EXPERIMENT
PZT SENSOR
SIGNAL PROCESSING
AI / SHM
KALAM ZERO RESEARCH FUNDING LAB
Welcome to the final experiment of the Kalam Zero Research Funding Lab.
In the previous experiments you learned the individual building blocks of electronics, sensors, signal generation, amplification, filtering and measurement.
Now we bring those ideas together into one engineering system.
Use a virtual PZT sensor to observe a structural vibration signal, separate useful information from unwanted noise, calculate signal features and make an engineering assessment of the structure.
PZT SENSOR → SIGNAL CONDITIONING → FILTERING → ADC → FEATURE EXTRACTION → AI / ENGINEERING DECISION → SHM
This is not merely a circuit experiment. It represents the complete thought process used in modern Structural Health Monitoring:
MEASURE → CLEAN → ANALYSE → INTERPRET → DECIDE
The raw signal contains the simulated structural vibration and measurement noise. The filtered signal represents the information retained inside the selected frequency band.
The dominant peak represents the principal vibration component. Changes in frequency or amplitude can provide useful information about changes in structural behaviour.
Real-world sensor measurements are rarely perfect. They may contain:
Drift, motion, environmental effects and slow structural movement.
The frequency region containing the vibration feature of interest.
Electrical interference, sensor noise and unwanted vibration.
The band-pass filter from Experiment 24 is used here to isolate the frequency region of interest.
An SHM system does not normally send an entire raw waveform directly to an engineer. Instead, useful characteristics are extracted.
Represents the effective magnitude of the vibration.
Useful for identifying strong transient events.
Indicates the strongest frequency component.
Indicates how clearly the useful signal stands above noise.
| Condition | Frequency Behaviour | Amplitude | Interpretation |
|---|---|---|---|
| Healthy | Stable | Normal | Expected structural response |
| Minor Anomaly | Small shift | Moderate change | Requires observation |
| Significant Anomaly | Noticeable shift | Large change | Requires engineering investigation |
A virtual classification is an educational demonstration. Real structural damage assessment requires calibrated sensors, validated models, appropriate sampling, environmental compensation, baseline data and engineering judgement.
Once features such as RMS, peak amplitude, dominant frequency, bandwidth and spectral energy have been extracted, machine-learning algorithms can learn patterns associated with different structural conditions.
Sensor measurements and extracted features.
The AI model learns patterns from labelled or historical data.
The model estimates the most likely structural condition.
Engineers use the prediction together with physics and inspection evidence.
This is the conceptual bridge from conventional instrumentation to intelligent Structural Health Monitoring.
Try questions such as:
Choose a structural vibration frequency.
Select a suitable band-pass region around that frequency.
Add measurement noise.
Compare healthy and abnormal conditions.
Observe how the extracted features change.
Finally explain how an AI model could distinguish the conditions.
Imagine that the PZT sensor is mounted on a bridge. The bridge has been monitored for one year. Thousands of vibration records are available.
How would you create a dataset containing:
This is where an electronics experiment becomes a research problem.
Components, circuits, voltage, current and measurement.
Waveforms, oscillators, timers and signal sources.
Amplification, filtering and frequency selection.
Sensors and measurement systems.
PZT → signal processing → features → AI → SHM.
You have completed the final virtual experiment of the KALAM ZERO RESEARCH FUNDING LAB.
You have travelled from individual electronic components and circuits to sensors, signal processing, filtering and finally intelligent Structural Health Monitoring.
LEARN → BUILD → MEASURE → ANALYSE → INNOVATE
The objective of this laboratory is not to tell students that research requires expensive infrastructure before they can begin.
The first step is curiosity. The second is understanding. The third is experimentation. The fourth is validation. The fifth is innovation.
With open-source software, virtual laboratories, simulation, low-cost electronics and intelligent tools, students can begin exploring engineering ideas before accessing sophisticated laboratory infrastructure.
"Dream, dream, dream. Dreams transform into thoughts and thoughts result in action."
— Dr. A. P. J. Abdul Kalam
This experiment is a virtual educational simulation. A real PZT-SHM system must be constructed using appropriate instrumentation, protection circuits, sensor mounting methods, grounding and calibrated measurement equipment.
Never connect experimental electronics directly to mains voltage.
Real structural decisions must never be based solely on an unvalidated virtual model or AI prediction.
Experiment 25 demonstrates the complete conceptual chain of an intelligent Structural Health Monitoring system.
PZT → AMPLIFICATION → FILTERING → DIGITISATION → FEATURE EXTRACTION → AI → ENGINEERING DECISION
The most important lesson is that a sensor by itself does not constitute an intelligent system.
The engineering value emerges when the measurement is converted into meaningful information and that information is interpreted using physics, mathematics, data analysis and, where appropriate, AI.
You have now completed the 25-experiment Kalam Zero Research Funding Laboratory.
🔬 THINK LIKE A RESEARCHER
⚙️ BUILD LIKE AN ENGINEER
🤖 EXPERIMENT WITH AI
🌱 INNOVATE FOR SOCIETY