Research On Degradation Model Of Turnout Switch Based On Neural Network

With the rapid development of railway operation and construction,railway equipment has also increased in a corresponding multiple,and the requirements for safety and reliable equipment have become higher and higher.The daily maintenance costs have also increased greatly.The turnout switch machine is an important basic equipment in the railway,and it is also a key equipment which needs a lot of energy in maintenance.The technology of health management is still in its infancy in the field of railway,but it has been widely used in other fields and achieved many achievements.If we can grasp the accurate health status of turnout switches and construct a degradation model base on the status.Then we can get the plan of maintenance and repair on the turnout switches with the degradation model,both reducing various costs and improving the efficiency and safety of railway operations.In view of the above possibilities,this article has conducted an in-depth study on the construction of a degradation model in the field of turnout switches which is one of the key technologies in health management.This article realized a neural network-based degradation model construction method,and also realized the turnout switches test platform,then the collected data combined with the field data used to complete the validation of method.This article takes ZD6 turnout switch as the research object.Firstly,it studied the construction method of the degradation model and various algorithms for preprocessing the data that may be used.After learned various models of the neural network,it used public data sets to conduct attempt and verification of the method,finally proposed the way of a Stacked Denoising Auto Encoder combined with a self-organizing mapping neural network to build a turnout switch degeneration model.On the basis of the degradation model,the remaining useful life is predicted by the indirect method with the long-short time memory neural network,the result is in line with expectation.The method provided a possibility for the research and application of the degradation model in the field of turnout switch.Secondly,on the basis of the proposed method and relying on the switch machine equipment,the test and experimental platform for the turnout switch in the laboratory environment has designed and implemented.The key components of the switch machine were introduced,and the required functions of the platform were analyzed.The platform can be logically divided into a host computer subsystem,a single-chip computer subsystem,and a switch machine device.This article discussed in detail the hardware and software design and implementation methods of each component of the platform,and successfully completed the design goal of the system.The single-chip computer can be controlled by the host computer to achieve indirect control of the switch machine,and various parameters such as the voltage and current of the switch machine during operation can be collected by various sensors and transmitted to the host computer for storage through the single-chip computer,so as to provide support for the further research work.Finally,based on the platform and the proposed method,tests and experiments were carried out to complete the construction of the degeneration model of the turnout switch,and the remaining life prediction of the switch model was completed through the degeneration model and the long-term and short-term memory neural network,which validated the method.In summary,the proposed method based on the neural network technology,in the field of turnout switch this paper has firstly completed the proposition and verification of the degradation model method,and explained the design and implementation of the test platform in detail.The research results of this article whether method or the platform can provide strong support for the subsequent research.Figure 32,Reference 72.