Due to wireless sensor networks（ WSNs） technology route is less and simple

short distance wireless sensing technology can solve problems of complex line difficulty of maintain caused by the underground cable channel. Through establishment of underground cable channel system based on WSNs

use temperature and humidity sensor for monitoring. The real-time monitoring data is transmitted back by WSNs technology

and through BP neural network model for grid training on monitoring data

set up optimization model and get weights before and after optimization and error curve of threshold training

simulation error of training samples and test samples before optimization are 0. 19 and 0. 30

respectively

these two values are 0. 17 and 0. 024 after optimization. Select 20 set of monitoring data of Chu Xiong transformer substation randomly to BP model

compare the obtained results with cable channel temperature and humidity security levels

accuracy up to 93. 3 %.