Research On Nondestructive Monitoring System For Corrosion Of Reinforced Concrete

With the development of Marine economy,more and more Marine engineering projects are put into construction or operation.At present most of the reinforced concrete structure directly exposed in the environment or an under-voltage protection status and construction of reinforced concrete engineering in harsh Marine environment are susceptible to corrosion and degradation,in the long standing part of coastal engineering construction operations is less than several decades even produced corrosion damage,to the people’s life and property safety brought great potential safety hazard.Therefore,it is particularly important to develop a non-destructive monitoring system for reinforced concrete corrosion based on the mechanism of reinforcement corrosion in concrete in the Marine environment,grasp the durability and service state of buildings in real time,improve the detection accuracy of the damage state of reinforcement in service,and make timely and accurate prediction for dangerous projects.In this paper,non-destructive monitoring and evaluation methods of reinforcement corrosion in concrete are studied from the following aspects:(1)According to the present reports of X-ray,CT scan and ground penetrating radar and other high precision steel corrosion monitoring equipment is expensive,complex operation,large in actual engineering and some equipment data acquisition need regular monitoring the inconvenience problem such as artificial inspection,developed the anode ladder sensor and strain gauge sensor based on the combination of embedded remote steel corrosion monitoring system,The monitoring system is small in size and low in cost,and can simultaneously monitor various types of reinforcement corrosion gradient signals(such as corrosion voltage,corrosion current,strain voltage,concrete impedance and other corrosion information)at different depths.The Internet of Things and cloud platform system are used to upload,display and save corrosion data,and artificial intelligence technology is used to analyze,excavate and integrate corrosion characteristic information,so as to realize the whole process monitoring of gradual depth corrosion information of steel bars.(2)Mobile network coverage is poor and signal occlusion is serious in some areas of cross-sea bridge and undersea tunnel.Aiming at the data transmission problem in signal blind area of reinforcement corrosion monitoring device,a wireless gateway device integrating Lora and 4G is developed.The Lora wireless communication scheme can be used for remote local wireless data transmission with low power consumption.It effectively makes up for the defect that data cannot be transmitted in the blind area of local signal in 4G communication scheme.The reinforcement corrosion monitoring device in the signal blind area can transparent transmit data to the Lora wireless gateway device in the mobile signal coverage area through the extended Lora module.The gateway device will collate the received data into corresponding data frames and upload them to the Ali cloud server for display and preservation through the 4G module.Or directly connect to the upper computer through bus 485,and use Modscan master station software to read gateway data in accordance with MODBUS-RTU protocol polling.(3)According to steel corrosion monitoring system measured data of different stages of reinforcement corrosion characteristics,establish steel corrosion data set,the data set after artificial screening remove outliers normalization preprocessing,design new convolution residual neural network topology model to test the steel corrosion state,ADAM algorithm to optimize training convolution residual neural network topology model,The piecewise learning rate attenuation strategy is introduced to suppress the learning rate oscillation at the later stage of network training,and the gradient change of second-moment estimation is adjusted to improve the iterative convergence efficiency.The establishment of a data-driven intelligent reinforced concrete corrosion judgment model with high recognition accuracy is helpful to get rid of the dependence on artificial interpretation of expert knowledge.