Abstract
In order to real time monitor the working status of a sensor, the effective index of the sensor output data is taken as an evaluation benchmark based on the performance indexes of the sensor. The time series from sensor output to be processed by the wavelet transform is proposed, and the real time method of the wavelet filter is discussed. The predicting value for the sensor output using artificial neural network is compared with the measuring value in the real field, and then the residual can be worked out. Therefore, the sensor effective index is figured out. Experimental results show that the evaluation method for sensor performance is real-time and effective, and the abrupt changing of process parameters can be online detected.
| Original language | English |
|---|---|
| Pages (from-to) | 161-165 |
| Number of pages | 5 |
| Journal | Shuju Caiji Yu Chuli/Journal of Data Acquisition and Processing |
| Volume | 22 |
| Issue number | 2 |
| Publication status | Published - Jun 2007 |
Keywords
- Neural network
- Performance evaluation
- Sensor
- Wavelet filter