PIDetect - A PID-Based Process Control & Anomaly Detection

This project implements a PID-based process control system with anomaly detection, leveraging Python, Streamlit, and Scikit-learn. The system allows users to adjust PID parameters, visualize system behavior, and detect anomalies using Isolation Forest. It simulates a first-order system and provides real-time insights into control performance.

---

Table of Contents

• Features
• Installation
• Usage
• How It Works
• Performance Metrics
• Anomaly Detection
• Future Enhancements
• Contributing
• License
• Connect With Me

---

Features

✅ PID Controller Implementation: Adjustable proportional, integral, and derivative gains.
✅ First-Order System Simulation: Uses ODE integration for realistic process modeling.
✅ Real-Time Performance Metrics: Computes rise time, settling time, overshoot, and IAE.
✅ Anomaly Detection: Detects irregular behavior using Isolation Forest.
✅ Interactive Visualization: Streamlit dashboard for intuitive tuning and monitoring.
✅ Customizable Setpoint: Define your desired process control target.

---

Installation

1. Clone the Repository

git clone https://github.com/v1v3x/PIDetect.git
cd PIdetect

2. Install Dependencies

pip install -r requirements.txt

3. Run the Streamlit App

streamlit run app.py

---

🎯 Usage

• Open the Streamlit web interface.
• Adjust Kp, Ki, and Kd values in the sidebar.
• Set a desired process target (setpoint).
• Monitor real-time graphs of process behavior.
• View performance metrics to evaluate control efficiency.
• Detect anomalies and analyze control instability.

---

🔍 How It Works

1. PID Controller: Computes control outputs using error-based feedback.
2. Process Model: Simulates a first-order system using ODE integration.
3. Performance Metrics: Analyzes system response based on control theory principles.
4. Anomaly Detection: Identifies system deviations with Isolation Forest.
5. Interactive Visualization: Plots process variables and control signals in real time.

📌 Control System Simulation Details

• The system follows a first-order process model with a time constant (τ) of 10.
• PID tuning parameters (Kp, Ki, Kd) influence system behavior.
• The simulation runs for 200 seconds with a step size of 0.1 seconds.

---

📊 Performance Metrics

Metric	Description
Rise Time	Time taken to reach 90% of the setpoint
Settling Time	Time required for the system to stabilize within 5% of the setpoint
Overshoot	Maximum deviation beyond the setpoint (%)
IAE (Integral Absolute Error)	Accumulated error over time

---

⚠️ Anomaly Detection

Anomalies in system behavior are detected using Isolation Forest:

• Outlier Detection: Identifies deviations in process variable and control signal.
• Visualization: Highlights anomalies as red markers in the graph.
• Threshold-Based Classification: 10% of data points are considered anomalies.

Anomaly Detection Implementation

from sklearn.ensemble import IsolationForest
from sklearn.preprocessing import StandardScaler

def detect_anomalies(data):
    scaler = StandardScaler()
    data_scaled = scaler.fit_transform(data)
    iso_forest = IsolationForest(contamination=0.1, random_state=42)
    anomalies = iso_forest.fit_predict(data_scaled)
    return anomalies == -1  # Returns True for anomalies

---

🚀 Future Enhancements

• ✅ Adaptive PID tuning using machine learning.
• ✅ Real-time cloud deployment with IoT integration.
• ✅ Advanced anomaly detection with deep learning techniques.
• ✅ Dynamic process modeling for different industries.

---

🤝 Contributing

Contributions are welcome! Follow these steps:

1. Fork the repository.
2. Create a new branch (feature-name).
3. Commit your changes.
4. Push to the branch.
5. Open a pull request.

---

📜 License

This project is licensed under the Apache 2.0 License.

---