Study On Temperature Compensation Of Roadbed Settlement Monitoring System And Settlement Prediction

With the continuous development of high-speed railway,the issue of subgrade settlement has become increasingly important,and traditional subgrade settlement measurement methods are increasingly difficult to meet the needs of engineering construction and maintenance.The research team,relying on the project "Chain Image Monitoring Method and Model for Subgrade Surface Settlement of Long and Long Railway Tunnels",proposed and designed an image based subgrade settlement monitoring system.Compared to traditional measurement methods,this system has non-contact,low cost There are many advantages such as sustainable observation,high degree of automation,and high measurement accuracy.However,at the application level,there are still some areas where the system needs to be optimized and improved,such as suppressing the impact of temperature changes on measurement accuracy,and achieving settlement trend prediction through monitoring system data.Based on the technical background of the application of camera measurement system in the settlement measurement of ballastless track,this thesis studies related issues with the goal of suppressing measurement errors caused by temperature changes and achieving settlement prediction.First,the basic theory and operating principle of the image-based roadbed settlement monitoring system are studied.The system uses the direct aiming characteristics of laser,combined with the centering algorithm and coordinate transformation,and then realizes the measurement of roadbed settlement.The coordinate systems involved include the world coordinate system,camera coordinate system,image coordinate system,and pixel coordinate system.By deriving the basic theory of spatial coordinate transformation,the mutual conversion relationship between the world coordinate system and the pixel coordinate system can be obtained,and then the settlement value can be calculated.For the monitoring system,one monitoring unit can only reflect the settlement condition in a small area,and the measurement range can be effectively expanded through the chain transfer structure.According to the monitoring system design principle,the monitoring system terminal model is designed.The monitoring system model includes imaging module,data processing module,display module,data communication module,main controller module,and power supply module by function.After completing the design and selection of each module of the system,functional tests were conducted,and the tests found that the system terminal model could achieve the required functions.Secondly,in order to reduce the influence of temperature change on the measurement results of image-based roadbed settlement monitoring system,experiments related to the temperature drift phenomenon of its core equipment-industrial camera are conducted.The experimental device consists of an imaging system,a temperature acquisition system,and a temperature control part,and the change of corner point coordinates is studied by controlling the change of temperature.It is found that the temperature change causes a slight change in the coordinates of the corner point,and this change is related to the temperature variable and the position of the corner point.Based on the theory and analysis,we reveal the causes of the above phenomenon and the inner law,and establish a drift parameter compensation model.The correctness and practicality of the model are verified by fitting calculation and model comparison.The results show that the model can improve the measurement accuracy of the monitoring system,which is of great significance for the practical application of image-based roadbed settlement monitoring system.Finally,roadbed settlement prediction algorithms are investigated.The first is an autoregressive difference moving average(ARIMA)model and the second is a variant of a neural network,the long and short term memory neural network(LSTM).In this thesis,we combine the two and propose a combined prediction model,which is based on the principle of minimising the sum of squares of errors to find the optimal weight coefficients,which can make the combined model achieve the optimal prediction effect.Comparing the root-mean-square error and other indicators of the three algorithms,it is concluded that the combined model outperforms the single prediction model and therefore the combined ARIMA-LSTM model is chosen as the final subsidence prediction model.