Correction of Non-linear Dynamic Properties of Temperature Sensors by the Use of ANN
The paper presents a new method for correction of dynamic errors by means of Artificial Neural Networks (ANNs), in which an inverse dynamic model of the sensor is realised by a neural corrector. Feedforward multilayer ANNs and a moving window method were applied. The proposed correction technique has been evaluated experimentally for small platinum RTD immersed in oil. The obtained results were compared to classical, linear correction method. Different multilayer perceptron networks were applied as neural correctors. The ANN approach significantly improved correction performance for the sensor exhibiting non-linear behaviour.
KeywordsArtificial Neural Network Step Response Dynamic Error Classical Correction Artificial Neural Network Approach
Unable to display preview. Download preview PDF.
- Haykin S. Neural Networks: a comprehensive foundation - 2nd ed. Prentice Hall, 1999.Google Scholar
- Jackowska-Strumillo L.: Computerised System for RTD Dynamics Identification. Int. IMEKO Seminar on Low Temperature Thermometry and Dynamic Temperature Measurement. Wroclaw, (1997), pp. D-(25–30).Google Scholar
- Jackowska-Strumillo L.: ANN based correction of dynamic errors in temperature measurements. Fifth Conference „Neural Networks and Soft Computing“ - Zakopane, June 6–10, 2000, pp. 681–686.Google Scholar
- Michalski L., Eckersdorf K., McGhee J.: Temperature Measurement. John Wiley and Sons Ltd., Chichester, 1991.Google Scholar
- Sydenham P.: Handbook of Measurement Science. John Wiley andSons Ltd., 1982.Google Scholar