AI-Driven Fault Detection in Biquad Filter Circuits

Project Overview

This project implements an intelligent diagnostic system using Artificial Neural Networks (ANN) and Support Vector Machines (SVM) to detect component failures in electronic circuits. The core of the study is a Tow-Thomas Biquad Filter, where we analyze how deviations in passive components (like $C_9$) alter the system's transfer function.

---

Circuit Design & Technical Description

The design is based on a second-order active Biquad filter topology, which is highly valued in analog signal processing for its low sensitivity to component tolerances. It utilizes three ADTL082 Op-Amps:

1. Integrator with Loss ($U_1$): Functions as a summing stage and lossy integrator ($R_{31}$, $C_{10}$).
2. Pure Integrator ($U_2$): Executing the second integration step ($C_9$).
3. Inverting Amplifier ($U_3$): Ensures the $180^\circ$ phase shift required for stable negative feedback.

Mathematical Model

The system behavior is governed by the following Transfer Function $H(s)$:

$$H(s) = \frac{V_{out}(s)}{V_{in}(s)} = -\frac{\frac{1}{R_{33} C_{10}} s}{s^2 + \frac{1}{R_{31} C_{10}} s + \frac{1}{R_{32} R_{29} C_{10} C_9}}$$

---

Machine Learning Performance

The models were trained and validated on a synthetic dataset of 510 samples, categorized into 51 distinct fault classes.

Comparison Table

Machine Learning Model	Test Accuracy	Status
Artificial Neural Network (ANN)	99.34%	✅ Superior
Support Vector Machine (SVM)	91.50%	Standard

---

Results & Output Analysis

Understanding the Results:

The visualization above demonstrates the high precision of the ANN model in classifying circuit states.

• Signal Signature: The AI analyzes the amplitude and phase response shifts caused by component aging or failure (e.g., $C_9$ deviation).
• Fault Mapping: Even subtle changes that are hard to detect with traditional multimeters are identified by the neural network through non-linear pattern recognition.
• Success Rate: Achieving 99.34% accuracy confirms that the system can reliably automate hardware diagnostics in critical analog systems.

---

Project Structure

• Biquad high-pass filter.asc: LTspice schematic for circuit simulation.
• faulty_detection_model.py: Python firmware for ANN/SVM training and evaluation.
• output.png: High-resolution output graph of the model's performance.
• biquad_fault_dataset.csv: The raw feature dataset (Amplitude, Phase, Frequency).
• model_comparison.csv: Statistical logs of the training results.
• requirements.txt: Necessary Python packages (numpy, pandas, scikit-learn).

---

How to Run

1. Hardware Simulation: Open Biquad high-pass filter.asc in LTspice XVII.
2. AI Training:
   • Install libraries: pip install -r requirements.txt.
   • Execute the script: python faulty_detection_model.py.