A Method for Picking Up Abnormal Forces of Stayed Cables Based on Convolutional Neural Networks

Abstract

The invention discloses a method for picking up abnormal cable force of a cable-stayed cable based on a convolutional neural network, comprising the following steps: using actual measured cable force data and artificially set abnormal cable force to establish a training data set, a verification data set and a test data set; Establish a convolutional neural network with a 12-layer structure, in which the cable force data is used as the input layer of the convolutional neural network, and the cable force anomaly is used as the output layer; the measured cable force data is input to the trained convolutional neural network, and through continuous update The cable force data realizes continuous cable force abnormal picking. When there is an abnormal cable force in the data, the output value of the convolutional neural network is the position of the abnormal data point. When there is no cable force abnormality in the cable force data, the convolutional neural network The output value is less than 0. The method of the invention can effectively pick up the abnormal cable force whose absolute value of the deviation rate is more than or equal to 1%, has good accuracy, applicability and practicability, and can be used for online picking up the abnormal cable force in the health system of the cable-stayed bridge.

Description

technical field [0001] The invention relates to a method for picking up abnormal cable forces of cable-stayed cables based on a convolutional neural network, which is mainly used for online picking up abnormal changes in cable-stayed cable forces of cable-stayed bridges. Background technique [0002] Cable-stayed bridge is a widely used long-span bridge type, mainly composed of main girder, cable-stayed cable and main tower, and the cable-stayed cable is responsible for the load transfer between the cable tower and the main beam. Cable-stayed cables are not only the key stress-bearing components of cable-stayed bridges, but also the weak components of cable-stayed bridges. During their long-term service, they are subjected to environmental erosion, traffic loads, natural disasters, etc., and corrosion, broken wires, fatigue, etc. will inevitably occur. If the damage problem is not grasped and dealt with in time, it will have a serious impact on the mechanical properties and ...

AI Technical Summary

Problems solved by technology

However, the current judgment of abnormal cable forces is mainly based on empirical thresholds. For example, my country’s "Code for Testing and Evaluation of Highway Bridge Bearing Capacity" (JTG / T J21-2011) stipulates that when the deviation rate of the measured cable force (deviation rate = (actually measured cable force - When the design cable force) / design cable force*100%) exceeds ±10%, it is determined that there is an abnormality in the cable force. Obviously, this threshold setting is too rough, which will lead to a small cable force abnormality (the absolute value of the deviation rate≤10%) For example, cable force abnormalities caused by cable breakage, cable slack, corrosion, heavy-duty vehicles, etc., these small cable force abnormalities will not immediately change the performance of the cable, but over time it may affect the Damage caused by stay cables under high stress

[0003] At present, smaller cable force anomalies are more difficult to pick up than large cable force anomalies. There are two reasons: First, the cable force of the cable is greatly affected by temperature, and the change of cable force caused by temperature will often overwhelm the smaller cable force anomalies. Abnormal picking; the second is that the cable force is constantly changing when there is no abnormality. If the threshold value of the deviation rate is set too small, it will lead to wrongly picking up the abnormal cable force.

Images