GO Feedback Controller Simulation with Kafka Integration

This project implements a PID controller using Kafka to simulate temperature regulation based on sensor data. The system consists of a producer that sends sensor data to Kafka and a consumer that processes the data, calculates the PID control input, and logs it to a CSV file. The simulation can run for 100 seconds, generating a controlled environment for the PID loop.

Prerequisites

Ensure the following are installed on your system:

• Docker and Docker Compose
• Go 1.18+
• Python 3.x (for the simulation script)

Installation and Setup

1. Clone the Repository

git clone <repo_name>
cd <repo_name>

2. Build and Run the Project

The Makefile automates the setup and execution of the project. To start everything (Kafka, producer, and consumer), run:

make all

This will:

• Start Kafka using Docker Compose.
• Build the Kafka producer executable.
• Build the Kafka consumer executable.

3. Run the Simulation

Kafka Producer

The producer simulates sensor data for the PID controller. It sends temperature reference, humidity, and the number of people in the room to a Kafka topic (temperature_data). The data is sent at regular intervals, enabling the consumer to process it.

To manually run the producer executable:

./producer

The producer continuously sends simulated data to Kafka until stopped.

Kafka Consumer with PID Controller

The consumer processes the sensor data, calculates the PID control input, and logs the output to a CSV file.

Customizing PID Parameters

The consumer allows customization of the PID controller parameters (Kp, Ki, Kd) using command-line flags:

./consumer --Kp <value> --Ki <value> --Kd <value>

For example:

./consumer --Kp 1200.0 --Ki 0.2 --Kd 0.05

If no flags are provided, the default values are loaded from the config.yaml file.

After running the producer and consumer for approximately 60–100 seconds, visualize the results using the Python simulation script:

make run-simulation

The script processes the CSV data generated by the consumer and plots a graph of the simulation.

4. Cleanup Resources

To stop Docker Compose and remove built executables, run:

make clean

Additional Help

To view available options and flags:

./consumer --help

Makefile Targets

The Makefile simplifies project execution. Here are the available targets:

• make all: Start Docker, build producer, and consumer.
• make docker-start: Start Docker Compose in detached mode.
• make build-producer: Build the producer executable.
• make build-consumer: Build the consumer executable.
• make run-simulation: Run the Python simulation script on the CSV data.
• make clean: Stop Docker and remove executables.
• make docker-stop: Stop Docker Compose containers.
• make help: Display the list of available commands.

PID Controller Overview

The PID controller regulates temperature using the following components:

• Kp (Proportional): Adjusts the response based on the current error.
• Ki (Integral): Accumulates past errors to adjust over time.
• Kd (Derivative): Reacts to the rate of change of the error.

The controller computes the control input based on sensor data (temperature reference, humidity, and room occupancy).

Notes

• Run the producer and consumer for at least 60 seconds to generate sufficient data for simulation.
• The simulation script visualizes up to 100 seconds of data. To simulate longer durations, rerun the script or restart the producer and consumer.

Commands -

Simulation -