Identification of electrical appliances using non-intrusive magnetic field and Probabilistic Neural Network (PNN)
The electricity waste is severe especially in large organizational buildings where the use of air conditioners, fridges and electrical motors are rampant. Due to lack of energy saving consciousness, users may not switch off this equipment after use. Thus, it would be an advantage if there exist a sy...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | |
| Published: |
2018
|
| Online Access: | http://dspace.uniten.edu.my/jspui/handle/123456789/7939 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The electricity waste is severe especially in large organizational buildings where the use of air conditioners, fridges and electrical motors are rampant. Due to lack of energy saving consciousness, users may not switch off this equipment after use. Thus, it would be an advantage if there exist a system that will be able to identify the appliances from one place without the residence having to go and check the state of the appliance or without having to place various sensors intrusively. Since most electrical appliances emit magnetic fields, the paper proposes to use non-intrusive magnetic field signature waveforms to identify the type of appliance used. The magnetic field emitted by table fan, blender and hairdryer are chosen for this purpose. The magnetic field from these three appliances are collected from four different measurement distances i.e. (i) 0cm (ii) 10cm (iii) 30cm and (iv) 60cm. The features of the magnetic field are then extracted and trained offline using the Probabilistic Neural Network (PNN). Once trained, the PNN shows that it is able to successfully identify the appliances regardless of the measurement distance. © 2014 IEEE. |
|---|