🚰 Water Level Control Module

📖 Overview

This project focuses on developing a educational system for automatic water level control.
The setup allows experiments related to PID control, signal conditioning, and embedded systems using Arduino.

The system regulates the tank water level through a PWM-controlled pump driven by MOSFETs, while a manual/automatic switch enables user interaction and comparison between open-loop and closed-loop operation.

📂 Contents

• /control_lvl → C code for Arduino.

📊 Project Status

Component	Status
Ultrasonic Sensor (HC-SR04) Integration	✅ Completed
PID Control Loop Implementation	✅ Completed
Power Console PCB (AOD4184 MOSFETs)	✅ Completed
Manual / Automatic Mode Switch	✅ Completed
Processing GUI for Monitoring	✅ Completed
Data Logging Functionality	✅ Completed
Final System Validation	✅ Completed

⚙️ System Description

• Controller: Arduino Uno
• Sensor: HC-SR04 ultrasonic distance sensor
• Actuator: DC water pump
• Power Stage: Custom PCB with two AOD4184 MOSFETs for PWM control
• Modes:
  • 🧭 Manual mode: Level controlled via potentiometer
  • 🔁 Automatic mode: PID-based level regulation
• Sampling period: 0.1 seconds
• Data transmission: Serial output for GUI visualization and logging

⚠️ Serial Communication Issue on Arduino UNO

🧩 Overview

The Arduino UNO can freeze when the serial connection is interrupted — for example, if you close the Serial Monitor or stop the Processing app while data is being transmitted.
This issue does not affect the Arduino Leonardo.

⚙️ Technical Cause

• UNO: uses a separate USB–Serial chip (ATmega16U2).
  → When the PC closes the port, the main MCU (ATmega328P) keeps sending data to a disconnected USB bridge, blocking the serial buffer.
• Leonardo: uses an integrated USB controller (ATmega32u4) that detects disconnections and avoids blocking.

🧠 Additional factor:
UNO’s transmit buffer is only 64 bytes. If Serial.print() is used too often or the port is closed, the buffer fills and freezes the loop

💡 Recommendation

• Prefer Arduino Leonardo or Teensy for real-time communication.
• If using UNO, add checks before sending data:

if (Serial && Serial.availableForWrite() > 16) {
  Serial.println(sensorValue);
}

or

if (Serial) {
  Serial.println(sensorValue);
}

🔄 Control Loop

Controlled Variables

• Level → Level value control (stabilization)

📐 Digital PID Control

The PID controllers implemented in this project are incremental (velocity form) and use trapezoidal integration for the integral term.
This ensures:

• Accurate discrete-time implementation suitable for microcontrollers.
• Consistency between simulation and embedded hardware behavior.
• Avoidance of integral windup when combined with actuator saturation or anti-windup mechanisms.

The Module uses a discrete PI controller implemented on a Arduino microcontroller.
The control law in the digital domain is expressed as:

$$ u(n) = u(n-1) + K_0 e(n) + K_1 e(n-1) $$

Digital PI controller implemented for level measured,

$$ V_{PWM}(n) = V_{PWM}(n-1) + K_0 e(n) + K_1 e(n-1) $$

Parameters:

The parameters are adjusted for temperature measured,

$$ K_0 = K_p + \frac{K_p}{2T_i} T_s $$

$$ K_1 = -K_p + \frac{K_p}{2T_i} T_s $$

🔉 Signal Processing: Low-Pass IIR Filter (1st Order)

To reduce measurement noise, a first-order IIR low-pass filter was applied to the level signal before feeding it to the controller and the model.

🔹 Filter Equation

$$ y(k) =(1 - \alpha) \cdot x(k) + \alpha \cdot y(k-1) $$

Where:

• $$x(k)$$: raw sensor measurement at time step $$k$$
• $$y(k)$$: filtered output
• $$\alpha$$: smoothing factor, $$(0<\alpha<1)$$

📐 Connection Diagram

📊 PI Control Test and First-Order Model Identification

A closed-loop experiment was conducted using a PI controller applied to the water level control system.
From the acquired input–output data, a First-Order model was identified using a non-parametric step-response method.

🔹 Identified Model

The identified model follows the general first-order transfer function:

$$G(s)=\frac{K}{\tau s + 1}$$

Where:

Symbol	Meaning
$$K$$	Steady-state gain
$$\tau$$	Time constant of the process

This model structure was obtained from the experimental step response without assuming a parametric structure beforehand (non-parametric identification).

🔎 Results Overview

• ✅ The PI controller achieved stable regulation of water level
• 🌊 The system dynamics were dominated by a single time constant, consistent with first-order behavior
• 🧩 The absence of transport delay simplifies control tuning
• 🎯 The identified model accurately reproduced the experimental response
• 🧾 This model serves as the basis for controller tuning and validation

📈 Experimental Plots

🖥️ GUI — Monitoring and Data Logging

The graphical user interface (GUI) shown in the photograph is developed using Processing 4.
It is designed only for monitoring the level module in real time and for recording experimental data.

GUI preview closed loop

GUI preview open loop

Key features:

• Real-time plot of level and setpoint.
• Display of control output (PWM or equivalent).
• Logging of measurements to files for offline analysis.
• Simple visualization of system behavior during experiments.

⚠️ Note: The GUI is for observation and data recording only; it does not modify the control system or send commands to the hardware.

⚡ Power Console PCB

The custom PCB integrates:

• Dual AOD4184 MOSFETs for high-efficiency PWM control
• Manual/automatic mode switch
• Potentiometer for manual reference adjustment
• Connector headers for Arduino and sensor interface

🖼️ Render 3D PCB

⚡Physical Prototype

PCB assembled

PCB assembled in process

Setup overview

🔗 Related Repositories

For additional tutorials and examples related to digital control simulations, visit:

• Digital Control — Anti-Windup (Positional PI)
• Digital Control Simulation — First Order System + Saturation

🤝 Support projects

Support me on Patreon https://www.patreon.com/c/CrissCCL

📜 License

MIT License