Design And Research Of Railway Engineering Construction Safety Early Warning Algorithm And System

The railway industry has a public works department responsible for the construction,maintenance and repair of lines.Due to the complex construction site environment,the actual engineering operation is often inconsistent with the previous construction plan,resulting in potential safety hazards during operation.Therefore,through the analysis of the status quo of public works construction work,this paper proposes a grading early warning strategy and basic mathematical model of the distance between construction personnel and line vehicles;using three algorithms based on machine learning to simulate and compare the model;the early warning system for public works construction is developed And conducted a functional test.The main research contents of this article include:First,through the analysis of construction safety problems,the distance grading early warning method between line vehicles and construction personnel is proposed,and the parallel early warning strategy for construction safety is formulated;the key technologies for implementing safety early warning are elaborated in detail,which lays the foundation for the subsequent research work.Secondly,Established a mathematical distance model for railway engineering construction safety early warning,based on machine learning ideas and engineering application requirements,using"reverse direction propagation（BP）","BP neural network with improved genetic algorithm（GA-BP）"and"support vector""SVM"three algorithms to simulate the established mathematical model of the interval calculation,and with the help of the train operation data of Xiaohang Road section,the safety interval accuracy and response time of the above three early warning algorithms are evaluated.The comparison of the evaluation results shows that when the vehicle speed of the line is 49m/s（178km/h）,the minimum safety distances of the three early warning algorithms of BP,GA-BP and SVM are 56_-0⁺⁰..₆₃¹⁸m、56_-0⁺⁰..₃₇²²m、56_-0⁺⁰..₃₉²¹m,response time is 0.15s,0.09s,0.19s Considering the accuracy and real-time performance comprehensively,the GA-BP algorithm is considered to be more suitable for public works construction scenarios.Thirdly,A safety warning system is designed.The hardware part completes the structure of the system hardware and the functional design of the hardware equipment;the software part makes detailed design of the train positioning,wireless communication,early warning logic,etc.and the software host computer interface.This system is based on the STM32 core,which realizes the safety warning and response when the operation management personnel and the construction personnel and the line vehicles are running online.Finally,the applicability of the three algorithms of BP,GA-BP and SVM in the construction operation scene was tested,and the performance of the GA-BP algorithm was the best;then,the GA-BP algorithm was introduced into the early warning system to implement the function of the system The verification results show that the system realizes the functions of early warning,alarm and detection,and can provide the speed and distance of the driver and the construction personnel to meet the functional requirements of the system.Conclusion.Through the consistency analysis of the experiment and simulation results,the correctness of the safety distance model and the practicality of the early warning algorithm are verified,which can provide reference and reference for the safety and management of railway engineering.