
Fault Diagnostic of Drain Pump Based on AI SVM
The ceiling cassette have drain pump that operated 24 hours per day. It drains water from indoor to outside. In this reason, it makes a little bit higher noise compared to normal operating indoor unit. To get a good sound quality, the AC motor was changed to BLDC motor to reduce the
2nd harmonic noise. Abnormal noise caused by process problems in the production line of the BLDC motor is generated when installed in the ceiling cassette. The manufacturing line of the BLDC motor is slightly different in the process of the rotor. But the pump is same. This difference makes
noise, which is caused by poor assembly or magnetization. This process has many uncertainties because abnormal noise occurs only when assembled in a ceiling cassette. So this is difficult to divided normal and abnormal drain pump case. To prevent these abnormal sounds, diagnostic systems using
vibration and noise were considered, but the classification was poor in conventional methods. So the system is developed to analyse data using the basic AI machine learning model, SVM(Support Vector Machine), to classify normal and abnormal drain pumps. In this paper, a nonlinear classification
technique, the RBF(Radial Basis Function) kernel SVM model is used to classify the fault diagnosis model with 98% accuracy.
The requested document is freely available to subscribers. Users without a subscription can purchase this article.
- Sign in below if you have already registered for online access
Sign in
Document Type: Research Article
Affiliations: LG Electronics
Publication date: 12 October 2020
The Noise-Con conference proceedings are sponsored by INCE/USA and the Inter-Noise proceedings by I-INCE. NOVEM (Noise and Vibration Emerging Methods) conference proceedings are included. All NoiseCon Proceedings one year or older are free to download. InterNoise proceedings from outside the USA older than 10 years are free to download. Others are free to INCE/USA members and member societies of I-INCE.
- Membership Information
- INCE Subject Classification
- Ingenta Connect is not responsible for the content or availability of external websites
- Access Key
- Free content
- Partial Free content
- New content
- Open access content
- Partial Open access content
- Subscribed content
- Partial Subscribed content
- Free trial content