Home Radiator Breakdown Detection using Edge Impulse

Home Radiator Breakdown Detection system using Arduino Portenta H7 and Edge Impulse software:

I am excited to present a groundbreaking prototype that revolutionizes the way we monitor home heating systems. Utilizing an advanced machine learning (ML) model, this innovative solution effectively detects the operational status of home radiators, specifically identifying when the boiler is ON or OFF. This capability not only enhances our understanding of heating efficiency but also plays a pivotal role in predicting whether your home is being adequately warmed. Deployed on the Portenta H7 and trained through Edge Impulse, this model is at the forefront of smart home technology.

To ensure the highest level of accuracy, I have integrated a real-time monitoring system that tracks the current room temperature. If the boiler remains ON for an extended period of 30 minutes—an adjustable timeframe—without any noticeable increase in temperature, the system promptly identifies a potential radiator malfunction or boiler issue, allowing for swift intervention.

Problem Statement:

During winter months, any issues with the radiator or low pressure in the boiler can hinder effective heating. For homeowners who may be away for a month, a breakdown of the boiler or radiator poses a significant risk: frozen pipes resulting from inadequate heating can lead to catastrophic damages and costly repairs.

Edge AI Solution:

Our cutting-edge tinyML model continuously monitors the boiler's operational status, comparing predicted outcomes against room temperature data. If the boiler frequently turns ON without a corresponding rise in temperature, it serves as a critical alert signal, indicating potential radiator or boiler issues. For a compelling demonstration of this state-of-the-art prototype, I invite you to watch the video linked below. This hands-on presentation will provide insights into its functionality before you delve into the detailed documentation. Don’t miss the chance to see this innovative solution in action!

https://www.youtube.com/watch?v=xh4BifHzt9A

Architecture

Model Training

I have positioned the Arduino Portenta H7 near the boiler and recorded datasets for both the boiler-on and boiler-off scenarios. To connect the Arduino Portenta H7 board to the EI account, please follow the steps outlined in the link below.

Device Connection

After installing the firmware, open the command window and type the following daemon command:

edge-impulse-daemon

Once the device is connected, navigate to the Data Acquisition section to collect the boiler sound data. Data should be collected at the specified ratio.

Model training in Edge Impulse

In the "Create Impulse" section, configure the Keras model and set the Time Series window size.

In the MFE (Micro-Feature Extraction) section, you can explore the available features.

NN Classifier

In the Neural Network section, configure the layers as indicated.

Testing

During the testing phase, the model is evaluated with new datasets that were not used during training.

Deployment

After successfully verifying the trained model, deploy it back to the Arduino Portenta H7.

However, we won't be deploying it directly to the Arduino Portenta H7 just yet, as we need to add more logic on top of the machine learning prediction.

Once the model is downloaded, follow these steps to import the library into the Arduino IDE:

Importing library in Arduino IDE:

1. Open your Arduino IDE, go to **Sketch** > **Add File**, and select the downloaded file.

2. Once imported, navigate to **Examples** > **home_Radiator_Breakdown_Inferencing** > **Portenta_H7** > **Portenta_h7_microphone**.

I have added my custom logic to this library to read temperature sensor data and analyze whether the room is heated up after the boiler has been running for more than 15 times in the last 30 minutes.

Algorithm:

The final working code with the above logic is available in my Git repository.

Link: https://github.com/Manivannan-maker/RadiatorBreakDownDetection.git

Open this code in the Arduino IDE and upload it to the Arduino Portenta H7.

Hardware connection:

(Include any necessary details regarding hardware connections here.)

Conclusion:

The ML model can identify boiler and radiator issues in a timely manner, alerting users to potential problems. This project can be enhanced by adding IoT features to provide real-time results, and the application can be utilized across various industrial machinery.