铁道车辆车轮扁疤故障检测技术综述

从铁道车辆车轮扁疤对轨道的冲击效应及其对车辆零部件造成的损伤出发，系统梳理了检测车轮扁疤的多种方案，对各类车轮扁疤故障检测方法特点进行了讨论，对比了不同检测方法的优缺点，对车轮扁疤故障检测技术体系的发展方向进行了预测。分析结果表明:车轮扁疤故障检测技术可分为车载检测法和地面检测法，其中地面检测法运用较为广泛；现阶段较为成熟的车轮扁疤检测技术按检测手段可主要分为轮轨冲击检测法、超声波检测法、噪声检测法、踏面位移法、振动加速度检测法、图像检测法、光学检测法、轨道电路中断法等；近年来，随着科学技术的发展，又涌现了如多普勒效应法、超声波回声定位法等；随着现代智能算法的进步，应用神经网络等智能算法对设备进行故障识别训练能大大简化设备开发进程和结构，智能算法或将成为车轮扁疤故障识别的主要发展方向；随着时间推移，检测设备的多故障集成化趋势越发明显，多故障检测集成化与功能多样化已是智能化检测设备发展的重要方向之一；未来，操作系统方面的提升也将主要集中于平台的人性化和智能化方面；检测体系建议由正线实时监测、车辆段入库精准检测、数据信息化平台三部分组成，未来发展方向会集中在装置简易化、算法精准化与操作智能化等方面。 

Review on detection technologies of railway vehicle wheel flat fault

From the impact effect of wheel flat of railway vehicle on track and its damage to vehicle parts, several schemes for the wheel flat detection were systematically combed out.The characteristics of various kinds of wheel flat fault detection methods were discussed, the advantages and disadvantages of different methods were compared, and the development trend of the system for wheel flat fault detection technologies was predicted. Analysis results reveal that the wheel flat fault detection technologies can be divided into the vehicle-mounted detection method and wayside detection method, among which the wayside detection method is widely used. At present, the relatively mature wheel flat detection technologies are mainly divided into the wheel and rail impact detection method, ultra-sonic detection method, noise detection method, wheel tread displacement detection method, vibration acceleration detection method, image detection method, optical detection method, track circuit interruption method and so on.In recent years, with the development of science and technology, methods such as the Doppler effect method, ultrasonic echolocation method and so on have emerged.With the progress of modern intelligent algorithms, intelligent algorithms such as the neural networks are employed to the train equipment for the fault identification, which can greatly simplify the equipment development process and device structure. Therefore, intelligent algorithms may become the main development direction of wheel flat fault identification. As time goes by, the trend of multi-fault integration of detection equipment becomes more prominent, and multi-fault detection integration and functional diversification have become one of the important directions in the development of intelligent detection equipment. In the future, the improvements in operating systems will also focus on the humanization and intelligence of platforms. Suggestions on the detection system are put forward from three aspects, namely, real-time monitoring of the operation line, accurate detection of depot entry, and information-based data platforms. Future development should emphasize simple devices, accurate algorithms, and intelligent operation. 3 tabs, 22 figs, 79 refs.