Temperature prediction method based on improved GA_BP neural network

Accurate temperature measurement is very important in industrial production and household electric control temperature. In practical applications, it is crucial to precisely know the temperature of target object to be measured, Bat it is not convenient to directly measure the temperature of a certain part. This kind of temperature is often predicted by measuring the surrounding temperature of a certain part of the measured object and condacting experimental modeling. However, the existing linear or nonlinear models have difficulties such as big temperature measurement error and poor robustness. In order to address the aforementioned issues, this study proposes an improved genetic back propagation (GA_BP) neural network algorithm to increase the model’s robustness and temperature measurement accuracy. Then this study designs a non-contact temperature measurement system, selects the electromagnetic stove as the object to carry out experiments, analyzes and compares the advantages and disadvantages of each method, and finds the improved GA_BP neural network algorithm has the best error compensation effect, and can achieve the purpose of noncontact, strong robustness and accurate prediction of the temperature of the target part.