鐵軌光學檢測平臺設計含6張CAD圖.zip,鐵軌,光學,檢測,平臺,設計,CAD www.elsevier.com/locate/ndteint NDT&E International 37 (2004) 111–118 Rail flaw detection: overview and needs for future developments Robin Clark* Sperry Rail, Inc., Danbury, CT, USA Received 10 February 2003; accepted 20 June 2003 Abstract Rail Flaw Detection has an important part to play in ensuring the safety of the world’s railroads. Recent accidents caused by broken rails have focused attention on the technologies that enable the detection of flaws in railroad rail. This paper reviews the technologies currently employed, along with examples of recent field applications. Some of the ongoing advancements and options for the future are also discussed. q 2003 Elsevier Ltd. All rights reserved. Keywords: Rail inspection; Ultrasonic testing; Electromagnetic testing; Fatigue defects; Roller search unit 1. Introduction Rail Flaw Detection (RFD) is very much in the spotlight at present. If we look at rail testing today we see a very different picture than that presented five years ago. This sector of the industry has undergone many changes in all parts of the world. The Hatfield accident of October 2000 has played a major role in this. That is a terrible statement to make, but how often does that happen? It took a tragedy for people to understand the real value of rail testing. This paper will discuss the rail testing industry from the early days to what we see today. The progress of the work being performed for Railtrack (now Network Rail) will be presented along with some thoughts for the future. 2. Rail testing: from the early days Rail Testing really became a serious entity in the late 1920s when Dr Elmer Sperry, driven by the needs of the US railroad industry, developed the induction method for testing railroad rail [1]. Over the years this technique was refined in the US and then in the 1950s ultrasonic testing emerged and started to become the mainstay of rail testing globally. This is still true today, even as more companies have entered the rail-testing arena. Some exceptions to E-mail address: r.p.clark@aston.ac.uk (R. Clark). * Present address: School of Engineering and Applied Science, Aston University, Birmingham, UK. this have been Sperry in the US, where the idea of ‘complementary testing techniques’ has been developed, and in Russia where there are still many cars that use the magnetic induction technique only. The report prepared by the Transportation Technology Center, Inc. (TTCI) for the Office of the Rail Regulator in October 2000 provides much useful background on the global rail testing industry today [2]. The basic physical principles of the two major techniques are as follows. In the case of ultrasonics, we are sending a beam of ultrasonic energy into the rail and looking for the return of reflected or scattered energy using a collection of transducers. The amplitude of any reflections together with when they occur in time can tell us about the integrity of the rail. Since defects are not totally predictable, we send in energy at several different incident angles in order to ensure that we maximize our chances of finding any detrimental features [3]. The refracted angles generally used are 0, 37 or 45 and 708. In addition, transducers are also positioned to look across the rail head for longitudinal defects such as vertical split heads and shear defects (Fig. 1a). The induction technique is based on the physics of electromag- netic induction. An high amperage current is injected into the rail via brushes that make contact with the rail head. In effect the rail becomes part of an electrical circuit. If the current encounters a defect, the current will travel around the defect (Fig. 1b). This ‘distortion’ of the current flow is detected via a block of sensors that detect disturbances of the magnetic field associated with the current flow [4]. 0963-8695/$ - see front matter q 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.ndteint.2003.06.002 R. Clark / NDT&E International 37 (2004) 111–118 112 Fig. 1. (a) Ultrasonic roller search unit; (b) induction principles. 3. Response to an accident 115 On March 18, 2001, the westbound California Zephyr, an Amtrak service from Chicago to the San Francisco Bay Area derailed in rural Iowa. The cause—a broken rail. One person was killed and 96 were injured. A detector car had been over the track a few weeks before the accident. A defect had been found and the railroad concerned had replaced the rail as determined by the Federal Railroad Administration (FRA) regulations. The problem was that the replacement rail used also had a defect in it. Within one week from the accident, the US rail testing companies had been mobilized to provide a manual testing capability to test the replacement rail on every subdivision of the railroad concerned. This incident was the first major use of manual rail testing on the North American freight railroads. The initial report related to this accident was published by the National Transportation Safety Board (NTSB) in March 2002 [5]. Following the preparation of a procedure [6] and adequate test trials on a 0.4 km test track in Danbury, CT, up to six Roller Search Unit (RSU) based portable rail detectors or ‘walking sticks’ were put to work by Sperry. Other contractors contributed equipment and personnel as well. Subsequently, a more controlled schedule has been developed in co-operation with the railroad concerned and they now have a number of rail testers working year round on a regional basis. The Portable Rail Detector (PRD) is shown in Fig. 2. The PRD makes use of the same RSU as that employed on the test cars operating across North America. As the PRD is moved across the rail surface, nine transducers operating at 2.25 MHz are simultaneously inspecting the volume of the rail. To ensure adequate coupling, a water tank is fitted to the PRD and a copper tube directs the water flow 1 cm or so  Fig. 2. Portable rail detector (North America). in front of the moving RSU. Although water usage is not excessive, the tank may need filling once or twice in a typical 8-h day. The display is in the form of an A-scan on a conventional ultrasonic test instrument. The NTSB report recommended that the replacement rail testing should continue. Recommendation R-02-5 to the FRA states ‘railroads should conduct ultrasonic or other appropriate inspections to ensure that rail used to replace defective segments of existing rail is free from internal defects’. An extension of this process would be to incorporate ultrasonic rail testing stations at rail recycling plants similar to those employed at new rail production plants. The response to Hatfield, although it evolved over a longer time period, saw the development of a PRD for the UK rail system (Fig. 3). This unit has recently started to become a more common feature on the UK rail network. Ergonomically, the unit is more attractive and lighter than the first model. The electronics package has been developed to enable flexibility and use with different ultrasonic testing instruments. The main unchanged feature though, is that of the RSU. The next logical step R. Clark / NDT&E International 37 (2004) 111–118 113 Fig. 3. Portable rail detector (Europe). is to develop a method of data storage that allows a complete record of the testing work performed to be available for post test review and archive. At present the testing record is a filled out report based on the technician’s interpretation of what is seen on the instrument screen and visually on the rail. The RSU is a more appropriate front end unit than the alternative slider probe arrangement, whether considering manual inspection or high speed inspection. The main benefits are the robustness, ability to conform to the rail geometry and most importantly a reliable and efficient transfer of energy into the rail being inspected. These features have encouraged the majority of the ultrasonic rail testing companies around the world to employ this technology. In the PRD case, the RSU houses nine ultrasonic transducers at three different angles, the aim being to ‘fill the rail’ with as much energy as possible at the optimum incident angles. The unit has been developed with Railtrack’s support to particularly address the detection of gauge corner cracking (rolling contact fatigue) that contributed to the Hatfield accident [7].  4. Existing technology Fig. 4 shows the most recent technology to be deployed on the US railroads. The system brings together the complementary ultrasonic and induction testing techniques on a hi-rail platform. This provides the railroad with a high quality test and increased flexibility of deployment. In the past, induction has only really been possible on a railbound vehicle because of the size of the plant needed to generate the high currents injected into the rails. With developments in power supply technology, the production of a hi-rail based vehicle has proven feasible [8]. To date, seven of these cars have been built and released since January 2001. They operate at speeds of up to 32 km/h, although with the ‘stop and confirm’ testing requirements in North America, there is always an operational trade-off between going forward faster and the risk of longer reversing moves when a confirmation is required. In addition to the above mentioned features, the vehicles have also seen the transition from a strip chart based display to a B-scan based display. The B-scan is a more intuitive representation of the ultrasonic interactions within the rail and is easier for the operator to interpret. Within the B-scan based system, algorithms work to reduce the data presented for interpretation. ‘Data overload’ for the operator can be a significant problem with rail testing, but on the flip side, mathematics is not always good at dealing with the unusual features encountered on the railroad. There has to be a balance. Gates are employed throughout the rail cross- section and the data is analyzed based on algorithms developed from both lab and field tests, and many years of testing experience. Both the ultrasonic and induction data are presented on the same display in the correct alignment. Much of the work done in the US is focused on the heavy haul railroad environment. Although the train speeds are less, the rail conditions encountered on the close to 200,000 miles of track that make up the North American railroads are much more severe than on most of the passenger biased railroads of Europe. Thoughts on the challenges to the rail testing engineer have previously been presented and they Fig. 4. Hi-rail ultrasonic/induction vehicle. R. Clark / NDT&E International 37 (2004) 111–118 114 continue to form the driving force for the development work that is performed all over the world [9]. The International Heavy Haul Association (IHHA) has acknowledged the demanding environment of the heavy haul railroads and they have collected together much of that wisdom in a book published in 2001 and launched at the IHHA meeting in Brisbane, Australia in that year [10]. The International Railway Union (UIC) also has an ongoing initiative that is looking at rail testing globally under the title ‘Rail Defect Management’. A few documents have been released so far [11,12] and a good overview was given by Lundgren et al. [13] at the Brisbane IHHA meeting. The reason for mentioning these initiatives is to highlight the global interest in rail testing today and to start to set the scene for the discussion of new technology that follows. In North America, by far the most common and problematical defects are transverse defects, weld defects and vertical split head defects. These defects constitute around 55% of the yearly detected defects by Sperry. They also constitute 75% of the notified failures received. A notified failure is an instance where a rail has broken and the company has been informed of the occurrence. In many cases an investigation will be performed to try and identify the cause of the failure. The possible causes are many—each situation presenting a source of further learning. On many occasions the defect may be classified as undetectable at the time of test because it may have been too small or the surface condition of the rail may have presented additional ‘noise’ that may have masked the defect. Also, the cause of the broken rail may have been something such as a wheel flat. In these instances a latent defect likely to be found at the next test may become a catastrophe when subjected to the impact of a train wheel with flats. The hi-rail ultrasonic/induction equipment has been carefully monitored as it has entered service. To date the positives have far outweighed the negatives. Being more sophisticated in its totality, the early days saw many teething troubles from generators to test carriage manipu- lation. The engineers had many challenges presented to them. Now, though the problems seem to be bottoming out as the trucks and the design have stabilized. One railroad has analyzed the performance of the vehicles as part of a Six Sigma Quality Project. The cars have been put on the demanding ‘coal territory’ of Appalachia—mountains, curves, severely worn rail and very variable weather conditions. The analysis has shown that since the introduc- tion of the new vehicles, the instances of rail head defect rail failures have dropped significantly (an over 50% reduction). The test vehicles have been finding more defects and at an earlier point in the defect’s growth cycle. This helps to emphasize the power of the complementary techniques on a ‘tough to test’ territory [14]. The system has also resulted in fewer stops to do manual verification and a 60% increase in testing speed in most cases. An analysis of the defects detected on one of the main North American freight  railroads over the course of a year showed that 20% of the defects marked indicated only with induction. This helps reinforce the value of complementary techniques. 5. Work for Railtrack (now Network Rail) In the latter half of 2001, Sperry was awarded a contract by Railtrack to build and operate an ultrasonic test vehicle on the UK rail network. To that end, 2002 saw much activity aimed at achieving that target. The instrumented carriage (UTU2) was fitted out in Derby in the UK and commis- sioned at the Old Dalby test track in late 2002 (Fig. 5). The ultrasonic test system build and mechanical design and build work for the deployment of the RSUs were performed in the US. The test system is B-scan based and much the same as that used on the hi-rail ultrasonic/induc- tion cars. The mechanical design proved to be a challenge. Conscious of the need to comply with the vehicle acceptance rules in the UK and the tighter clearance demands, the design has been developed from that used on a high-speed vehicle already operating in Europe, most specifically in Sweden, Norway and Germany. The RSU housing has been redesigned to meet the clearance requirements. A model of one quadrant of the RSU deployment arrangement mounted under the vehicle is shown in Fig. 6. The B-scan display is similar to that shown in Fig. 7 without the windows for the induction data. These windows were removed and the ultrasonic display was re-configured to make use of the available space. For ease of training and use the software has been developed in a Windows environment. The vertical dimension of each window relates to the vertical position of the indication source in the rail. The horizontal dimension of each window relates to the longitudinal position of the indication in the rail. The third part of the vehicle work has been the opportunity to design a new calibration rail for Railtrack. Drawing on experience and the requirements/suggestions of other railroads, a new design has been produced for Railtrack. The rails have been installed in the Old Dalby Fig. 5. Network Rail test vehicle. R. Clark / NDT&E International 37 (2004) 111–118 115 Fig. 6. Mechanical arrangement on UTU2.  test track and have been used in the commissioning work. It is hoped that the design will be adopted by other railroads such that the rail testing community can work towards some form of international standard. The next step in the UK will be the introduction of a new all ultrasonic hi-rail vehicle similar to one recently developed for use in Germany. This vehicle uses the same B-scan based test system and a carriage derived from that used for the hi-rail ultrasonic/induction vehicle in the USA (Fig. 8). Again the stringent vehicle acceptance criteria have dictated the design steps taken. This carriage has room for both ultrasonic RSUs and eddy current sensors. 6. Technology development Rail testing has never seen large amounts of develop- ment funding. It seems that it is only when an accident occurs that money becomes more readily available. When this occurs though, it is perhaps even more important that we demand careful use of that money and co-ordinate Fig. 7. B-Scan display. Fig. 8. European ultrasonic hi-rail vehicle. the development efforts in a coherent fashion. We often jump on the ‘new is obviously better’ bandwagon without thinking. Yes it is good to invite new ideas and people to the table, but we must always be careful that they do truly have something to offer. Also, there is often a tendency to forget the people who have made rail testing their life. These people work for companies that have invested time and money in moving the industry forward when the need for improvements has appeared less of a priority. Different countries have different approaches to development, but the accident scenario I have just described seems common in the majority of cases. The majority of the technology development in the US is currently performed by the rail testing service suppliers. The railroad challenge has recently received a higher profile in the wider arena [15,16], but many of the new investigators are in the early stages of projects. One of the main problems that few researchers grasp at the outset is how the industry works. The variables that have to be dealt with are more demanding than they realize and the path from the lab to the field is difficult. Rail does not particularly lend itself to being used in exhaustive trials ahead of the service implementation, so developments often reach the field with much work still to be done. The Rail Defect Test Facility (RDTF) at the TTCI in Pueblo, CO is an attempt to help reduce this hurdle. 9 鋼軌探傷：概述和未來發(fā)展的需要 Robin Clark * 斯佩里鐵路公司，丹伯里，CT，美國 2003 年 2 月 10 日收到；2003 年 6 月 20 日接受 摘要 鋼軌探傷在保證世界鐵路安全方面發(fā)揮著重要作用。最近由破碎鋼軌引起的事故集中在能夠檢測鐵路鋼軌缺陷的技術上。本文回顧了目前采用的技術，以及最近的現(xiàn)場應用的例子。一些正在進行的進步和未來的選擇也進行了討論。Q 2003 愛思唯爾有限公司保留所有權利。 關鍵詞：鋼軌檢測；超聲波檢測；電磁檢測；疲勞缺陷；滾輪搜索單元 1、引言 鋼軌探傷是目前備受關注的熱點問題。如果我們看看今天的鐵路測試，我們看到的景象與五年前的情況截然不同。這個行業(yè)在世界各地都發(fā)生了許多變化。2000 年 10 月的哈特菲爾德事故在其中起了重要作用。這是一個可怕的聲明，但這種情況多久發(fā)生一次？人們很難理解鐵路測試的真正價值。 本文將討論鐵路測試行業(yè)從早期到我們今天看到的。正在進行的工作正在進行的 Railtrack（現(xiàn)在網(wǎng)絡鐵路）將提出一些想法，為未來。 2、軌道測試：從早期開始 在 20 世紀 20 年代末，當 Elmer Sperry 博士在美國鐵路行業(yè)的需求下，開發(fā)了用于測試鐵路鋼軌的感應法（1）時，鐵路測試真正成為一個嚴肅的實體。多年來，這種技術在美國得到了細化，然后在 20 世紀 50 年代出現(xiàn)了超聲波測試，并開始成為全球鐵路測試的主流。盡管更多的公司已經進入鐵路測試領域，但這一點仍然適用。一些例外的是斯佩里在美國，那里的“互補測試技術”的想法已經開發(fā)，在俄羅斯，仍然有許多汽車只使用磁感應技術。運輸技術中心（TTCI）為 2000 年 10 月鐵路調度員辦公室準備的報告為當今全球鐵路測試行業(yè)提供了許多有用的背景（2）。 這兩項主要技術的基本物理原理如下。在超聲波的情況下，我們將一束超聲波能量送入軌道，并利用傳感器的集合來尋找反射或散射能量的返回。隨著時間的推移，任何反射的振幅都能告訴我們軌道的完整性。由于缺陷不是完全可預測的，所以我們發(fā)送能量在幾個不同的入射角，以確保我們最大限度地發(fā)現(xiàn)任何不利特征的機會[3 ]。通常使用的折射角 是 0, 37 或 45 和 708。此外，換能器也被定位為感應技術是以電磁感