Thermocouple sensor non-linearity compensation based on orthogonal polynomial basis functions neural network

This paper presents a new method to compensate thermocouple sensor non-linearity based on orthogonal polynomial basis functions neural network. The artificial neural network module is used to realize Chebyshev artificial neural network. The mathematical model is established based on Chebyshev artificial neural network for non-linearity compensation. The principle and training algorithms of orthogonal polynomial basis functions neural network and neural network module are introduced. The results show that non-linearity compensation method has the advantages of high precision, good real time and strong robustness, etc. The maximum non-linearity error is 0.15%.