Application of ANN on the Intelligent Temperature Sensor

There is non-linearity in the temperature sensor, most of which are rectified by consulting the tables or piecewise linearization. In this paper, the linearity principle of the thermocouple sensor with artificial neural networks (ANN) and the adjusting algorithm of ANN weights are analyzed. Meanwhile, the principle of non-linearity rectification is verified by taking platinum-rhodium 30-platinum-rhodium 6 thermocouple (B) sensor as an example, thus linearization method is simplified. After adopting ANN, the precision of the sensor system has been greatly improved compared with traditional intelligent sensors, and favorable effects have been obtained by applying it to chinaware stove