Wireless Temperature Measurement Method of High-Voltage Switchgear in Nuclear Power Plant Based on Multi-Level Neural Network And Multi-Point
Under the influence of factors such as lighting conditions,viewing angle,and shooting distance,the quality of temperature sensing imaging is poor,and there is a high risk of data leakage or tampering.The temperature information of high-voltage switchgear has deviations or unstable linearity,resulting in a low temperature fitting gain coefficient in the wireless temperature measurement process,.To this end,a wireless temperature measurement method for high-voltage switchgear in nuclear power plants based on multi-level neural networks and multi-point positioning is proposed.Obtain temperature sensing images through the placement of shooting points for radioactive structures,and perform grayscale threshold segmentation processing on them;Using domain averaging method and combining with multi-point requirements,the obtained binary temperature sensing image is smoothed,and a multi-level neural network is constructed.The binary temperature sensing image is used as input,and its nonlinear mapping ability is enhanced through convolutional layers and activation functions.The wireless temperature measurement results with smooth expected values,namely the predicted surface temperature values of the device,are analyzed and calculated.The experimental results show that after the application of this method,the temperature fitting gain coefficient of the temperature measurement results is higher,and it has more ideal temperature measurement accuracy,providing accurate and reliable technical support for wireless temperature measurement of high-voltage switchgear in nuclear power plants.
neural networknuclear power plantswitchgearwireless temperature measurementtemperature sensing image
NETL
NSTL
万方数据
