Predict Maintenance Solar Heater Controller

1 Introduction

Solar Heater is eco-friend facilities to harvest sustainable power for household purpose. The solar heater can be frozen in winter and overheating to vapor , so as to break the containers. With predict maintenance, to control the temperature of water in tubes by the circulation of water and or to empty the tube in cold weather.

2 Software and Hardware

2.1 Nordic Thingy:53 have provided enough environmental sensors to detect the ambient environment. With ML toolset, the solar heater can be adjusted to best performance. Existing solutions are simple fix temperature control without intelligence. There are limited function in such control solutions.

The main features of this controller can predict the amount of solar heat harvested and control the water circulation pump and valves for better performance and fault prevention for extremity condition.

2.2 Edge Impulse for Thingy:53 can help create the training model for this design.

3 How does your solution work? What are the main features? Please specify how you will use the Nordic Thingy:53 in your solution.

3 Step by step

3.1 Development Environment shall be made for nRF connect Desktop v3.1.11

With nRF connect SDK v2.1 from Tool Manager , Install chocolately, python 3, nodejs, visualstudio-buildtools 2017 and Edge Impulse CLI as follows

Then start Edge Impulse CLI to connect Device Thingy 53

edge-impulse-daemon

Now it is ready for Edge Impluse on Device Name PMsolarHeater

3.2 Then collect environment data including temperature and humidity

Edit Label with Normal , Abnormal in corresponsive with data characteristics

3.3 After data collected, Add Impulse as model structure

Training the data and classification model

3.4 Build the model and export the library

4 Download the firmware and run the codes

After build completed, the library is downloaded as zip files as follows.

The model can be integrated into the project as AI library in Thingy 53 project in VS code for nRF connected SDK projects.