Non-destructive testing system for steel bars in concrete based on improved neural network, and control method thereof

Abstract

The invention relates to a non-destructive testing system for steel bars in concrete based on an improved neural network, and a control method thereof. The control method comprises the following steps: steel bar positioning; steel bar diameter measurement; and protective layer thickness measurement. The system consists of three parts, i.e., a coil array sensor, a lower computer and a host computer. The specially-produced coil array sensor enables the system to have the ability of accurately positioning steel bars. The lower computer comprises an embedded controller, a signal generation circuit, a signal processing circuit, an AD sampling circuit, an electronic display screen, function buttons, a WiFi module, and a raster range ranging module. The host computer comprises the improved neuralnetwork and an assistant display interface. The control method is mainly implemented via the improved neural network. The special design based on the coil array sensor can realize accurate positioning of steel bars and solve the problem that conventional steel bar positioning is not sensitive to the angular displacement of steel bars. At the same time, algorithms based on the improved neural network improve the detection accuracy of steel bars and overcome the disadvantage that steel bar diameter and protective layer thickness cannot be simultaneously detected in the prior art.

Description

technical field [0001] The invention relates to a non-destructive detection technology, a reinforced concrete detection technology and a soft measurement mathematical model of a steel bar, and belongs to the application field of building intelligent control. Background technique [0002] The most common form of building structure in my country is reinforced concrete structure, and the safety detection of reinforced concrete structure is an inevitable part of safety production. However, in real life, since the steel bar is inside the concrete, its position quality is unknown, so the detection of the steel bar inside the concrete is an urgent problem to be solved at present. At present, the method of breaking the wall and exposing the internal steel bars is mostly used for the detection of reinforced concrete, which will cause great damage to the building. [0003] At present, the coil structure used in the eddy current effect-based concrete internal reinforcement detection s...

AI Technical Summary

Problems solved by technology

Due to the limited number of detection coils and the relatively simple layout of the coils, the amount of steel bar information that can be obtained is relatively small. At present, similar products can only detect the thickness of the protective layer under the preset steel bar diameter, but the steel bar diameter is the same. It has an important impact on the quality of the building, and the diameter of the steel bar is assumed to be known, which brings great uncertainty to the accuracy of the thickness detection of the protective layer. Therefore, there are great hidden dangers in the practical application of this scheme

In terms of back-end modeling methods, linear function fitting is mostly used at present, and polynomial fitting is common. The mathematical model obtained by this scheme has large deviations, and it is not self-learning, and the effect in practical applications is poor. Difference

Images