Final Report Summary - AXLEINSPECT (DEVELOPMENT OF NOVEL INSPECTION TECHNIQUES FOR TRAIN AXLES)
Executive Summary:
Technological advances in train design during the last 5 years period have enabled high-speed trains to become more commonly used. The increasing trend for the industry’s business is forecast to continue in the next ten years, since rail transportation is steadily becoming a more attractive option over other means of transport for the public. This is because train travel is generally cheaper than using a car, and usually the fastest option to reach a destination. It is also inherently safer and far more environmentally friendly in comparison to car travel, without compromising passenger convenience. As a result, today’s rail networks across Europe are getting busier, with trains travelling at higher speeds and carrying more passengers and heavier axle loads than ever before. The loads and stress that axles experience are often inadequately defined. Rolling stock, including passenger and freight wagons, are designed to last for thirty years, but are often used for much longer. The combination of these factors has put considerable pressure on the existing infrastructure, leading to increased demands in inspection and maintenance of rail assets. The expenditure on inspection and maintenance has grown steadily over the last few years without being followed by a significant improvement in the industry’s safety records. As a direct consequence, one of the immediate key challenges faced by the rail industry is improvement in the safety of the railway systems of EU member states, in both environmental and financial terms, by delivering further efficiencies and exploiting technological innovation.
The structural integrity of wheelsets used in rolling stock is of great importance to the rail industry and its customers. A number of rail accidents have been directly related to the failure of axles, leading to increased demands for the inspection and maintenance of such components. The risk of axle failure relates to both distance covered and time in use. Therefore, it is preferable for more frequently scheduled inspections to occur during the lifetime of the axle, rather than waiting for the less frequent overhaul periods in order to detect any growing fatigue cracks.
Indeed, strict rail standards such as the Railway Group Standards, exist to ensure that axles are inspected at regular and frequent periods to ensure safety is maintained. Depending on the train service operator, it is not unusual for axles to be checked by in-service inspection every 30,000km. To be cost effective and to cause minimal disruption to train services, the NDT inspection should not require the disassembly of the wheelsets and supporting bogies from the vehicles. In addition, there is also evidence that full disassembly can be counterproductive and lead to future maintenance problems due to the required assembly/disassembly of the parts during the service overhaul.
Although railway axles have been designed and used for nearly 200 years, and knowledge about fatigue prevention can be used to assist in the timing of crack inspection intervals, the severity of the mechanical environment in which axles are used means a number of failures, although rare, still occur.
The European rail network is targeting a considerable expansion of passenger and freight traffic by 2020. In order to achieve this, increased reliability and availability of rolling stock is necessary whilst maintaining the same, or better, level of safety. The axle life is a crucial part of both the safety and economic performance of the vehicles, and the axle deteriorates through its lifetime by means of fatigue and corrosion mechanisms. Periodic inspection is used to ensure that these mechanisms have not compromised the axle safety; however, inspection that takes a vehicle out of service impacts on the economic aspects of train operation. It is clear that errors are made in both aspects: Recent serious incidents due to axle failure, in Cologne (2008) and Viareggio (2009), were due to the failure of the inspection regime, and continue to affect the inspection periodicity. On the other hand axles with suspected cracks, on passenger and freight vehicles, are being taken out of service after only 3 years, when their lifetime should be at least 20 years, because of fears that cracks may initiate at corrosion pits.
Usually the inspection interval is such that an inspection is required between major overhauls, so inspection of axles while the train is in-service is still required. This is currently inconvenient and costly for the train operators. If withdrawals take place while a train is in-service, there is a consequent significant cost associated while the train is out of service for additional maintenance. There is also the financial and energy cost in making and supplying a new axle. This has become a specific problem for the highly detailed hollow axle designs such as those used in high-speed passenger trains.
Project Context and Objectives:
The AxleInspect project aimed to develop new inspection techniques based on phased array ultrasonic testing (PAUT), conventional ultrasonic testing (UT) and electromagnetic (EM) testing suitable for the inspection of both solid and hollow axles. For solid axles, techniques have been developed that inspect from the end face of the axle using new and novel phased array inspection technology. For hollow axles, probes based on ultrasonic and electromagnetic inspection have been developed, enabling detection of surface breaking defects that cannot be found by existing inspection techniques.These new techniques will allow inspection of axles whilst they are still attached to their supporting bogie, allowing minimal disassembly from the train.
AxleInspect is a collaborative project comprising the following organisations: TWI Ltd, Ideko, ZUT, Vermon, Phoenix Inspection Systems, DANOBAT Railway Systems and Balfour Beatty Rail. The project is co-ordinated and managed by TWI Ltd and is partly funded by the EC under the Research for the benefits of SMEs Project Ref: FP7-SME-2011-286573.
The main technical objectives of the AxleInspect project are:
• Develop and produce an instrument to detect transverse cracking on the critical areas such as cross-sectional area changes (curvatures), wheelseat/brake disc/gear set regions) of a solid axle whilst the axle is assembled in its wheelset and bogie. This has been achieved by:
• Development of a phased array probe to inspect the axles from the end face.
• Development of a scanning jig to carry out the inspection.
• Development of software to acquire, display and analyse the PAUT data.
• Develop and produce an instrument to inspect hollow axles using a combined rotating UT and EM probe. The instrument should be capable of easy installation in an inspection station. This has been achieved by:
• Development of a suitable electromagnetic probe for inspection of inner surface breaking cracks from inside the bore.
• Development of ultrasonic probes for the inspection of the outer diameter for outer surface breaking defects from inside the bore.
• Development of an automated scanner to deploy the UT and EM probes.
• Development of data fusion software to display the results from both inspection techniques.
Project Results:
The AxleInspect project has achieved significant technical and scientific results. A number of different solid axles have been obtained from the AxleInspect partners, and also from previous TWI completed projects, to allow the development of project technologies. A wide range of solid axles, with different end face diameters and geometry, were available for use in the experimental work. Furthermore, a hollow axle from a high speed train has been provided by the partners to allow development of the hollow axle inspection techniques. A number of simulated defects have been introduced into the solid and hollow axles to enable development of the ultrasonic and electromagnetic techniques.
Extensive ultrasonic modelling has been carried out to design a new phased array probe configuration that will be suitable for inspecting the critical areas of solid axles from the end face. A number of different phased array probe configurations, with different shapes and numbers of elements, have been considered and the most suitable design was selected for further modelling and optimisation. A new innovative phased array probe has been designed that will meet the inspection requirements from the axle end face. The main parameters that the phased array probe will fulfil are: adaptable to different end face diameters and configurations, steering and skewing abilities, large focal area and small focal spot size and ability to cover the critical areas of different solid axle geometries. The new phased array design was simulated to determine the interaction of the generated ultrasonic beam with simulated defects at the critical areas. The simulation results show that the defects in the axle critical areas can be identified with enhanced detection capabilities. The developed phased array design was manufactured and tested on the axle samples. Excellent correlation was achieved between the modelling and experimental results, and the probe performance is as predicted by the modelling.
A comprehensive numerical analysis was carried out to determine the most suitable electromagnetic method for inspection of hollow axles to detect inner surface breaking cracks. A number of different methods have been selected for evaluation, such as eddy current testing (ECT), magnetic flux leakage (MFL), alternating current field measurement (ACFM) and remanent flux leakage (RFL). After preliminary analysis of the problem, several arrangements of the electromagnetic transducers were selected for detailed numerical analysis. A numerical modelling (finite element method - FEM) is used as an effective way of evaluation and design. Two different numerical models were prepared: 3D models for general evaluation and 2D models for detailed analysis and optimisation. The electromagnetic numerical analysis concluded that the best method for this application is the RFL. Two innovative RFL transducers were designed, manufactured and tested. The results obtained from the final optimised electromagnetic transducers are promising. The developed technique has the capability to accurately size the defects using the different magnetic flux components (x, y and z).
Furthermore, ultrasonic modelling was carried out to determine the most suitable parameters for the ultrasonic probes in order to detect outer surface breaking defects from the inside bore of the hollow axle. The beam profile, and the interaction with defects at critical areas, have been simulated during the extensive modelling programme.
A semi –automatic solid axle scanner was designed and manufactured that allows the inspection of solid axles from the end face using the developed phased array probe. The developed scanner has a modular design that allows for ease of mounting as well as providing adaptability to different train axle boxes and different end face diameters. The solid scanner combined with the developed phased array software allows the acquisition of encoded data that can be saved for further analysis. The scanner allows the accurate positioning of the probe and provides repeatable and recordable inspections to be carried out.
An automated hollow axle scanner was manufactured that will be used to deploy the developed ultrasonic and electromagnetic probes. Data fusion software was developed that can control the scanner motion, acquire the ultrasonic and electromagnetic data and finally analyse the data. The software can compare the acquired results with the results obtained from an axle without any defects, which will help the operator to identify any defects. Furthermore, the developed data fusion software allows the end user to input various parameters in order to identify the asset inspected (ie axle number, train number, bogie number etc). All the data is saved in a database for future comparisons.
More details of the work carried out in the project are provided below.
The system specification document has been completed and the requirements of the inspection systems have been defined. More specifically, the defects’ requirements for their sizes and location, axle geometry and inspection requirements have been defined. The requirements for the phased array probe and technical constraints have also been defined. The inspection requirements have been selected by using railway inspection standards that define the acceptable frequencies to be used and the inspection methods. These inspection standards are providing guidance only since they are used for conventional ultrasonic techniques and are not directly applicable to phased array inspections.
A number of axle samples have been obtained from previous TWI projects and also provided by the AxleInspect partners. The samples acquired for the project include seven different designs of solid axles and one design of hollow axle. The selected axle designs are representative of the axles commonly used. The acquired axles provide a good range of different parameters such as the number of critical areas and their distance with respect to the end face, curvature radii and outer diameters. The available solid axles have different end face diameters which will allow the adaptability of the developed solid axle inspection tool to be tested.
Artificial defects have been introduced, at critical areas, into one of the solid axles being used to test the new phased array probe. Furthermore, artificial defects have been introduced into the inner surface of the bore in the hollow axle being used for the electromagnetic technique development. More defects were introduced into the remaining solid axles after the initial work with the newly developed probe was completed. Defects could be introduced at TWI’s facilities, giving the flexibility to add artificial defects as necessary during the experimental work.
One of the main aims of the project was to design and develop a new and novel 2D phased array ultrasonic probe for inspection of solid axles from the end face. An extensive simulation programme was carried out to design the new phased array probe. The new design provided the following functionalities: adaptability to different axle geometries’ beam steering and skewing abilities, large focal length and small focal spot. A number of different configurations for the element size and probe shape were simulated to find the optimum probe design. The probe to be used, both in pulse echo and in pitch–catch modes, will consist of a maximum 256 elements, thus all the simulated probe designs had this element number as their limit. The reason that this number of elements was chosen was mainly due to the availability of a suitable phased array pulser-receiver that will be used to excite the probe and inspect the axle samples. Furthermore, a higher number of elements in the probe will result in a very expensive probe, both in terms of development and as a commercial proposition. Additionally, the electronic systems available can only handle 256 elements simultaneously. In total, 26 different probe configurations have been simulated in order to determine the most suitable probe for the application. Due to the number of elements involved, the simulations were very computing intensive and each model took more than one day to be completed. The final design of the probe has 248 elements with different element sizes within the same array. The simulation work of the developed PAUT probe was carried out to measure and evaluate the defect response of the ultrasonic beam. In order to achieve this, a number of simulated defects were placed at several positions on the axle. Investigation of the defect response is important so that it can provide a reference, and prove that the proposed phased array probe can give the desired results and meet the inspection requirements. For the solid axles, the target is to inspect from the end face, which is a challenging procedure. For that reason the developed LM (Linear Matrix) probe has to operate in two different inspection modes:
• pulse-echo (sectorial scanning on one plane) when the middle of the probe is not positioned over a bolt hole and
• pitch-catch (skewing and steering scanning) that will be used when the middle of the probe is positioned over a bolt hole.
A number of defects with different dimensions were placed in critical positions, and simulated both separately and simultaneously. A number of different simulation parameters have been tested in order to acquire the maximum amount of simulated data. The results show that the developed phased array probe can meet the inspection requirements and performance. The design and manufacturing of the new phased array probe has been carried out by Vermon. Phoenix ISL was responsible for the design, manufacture and assembly of the solid axle scanner and probe holder that will be used to carry out the inspections from the axle end face. The inspection requirements to be satisfied by the scanner have been defined in the project and are listed below:
• Mount onto a wide range of axles and wheel bearing housings.
• Ability to position the probe to end face diameters from 110-165mm.
• Perform 360° encoded circumferential scan.
• Rigidly hold the large bespoke phased array transducer and load it onto the inspection surface.
• Provide axial movement to position the transducer as required.
The developed scanner can be used for:
• Inspection of dismantled wheel sets via the use of switchable electromagnets
• Inspection of fully assembled wheel sets via a mounting plate.
An elemental functional reference block has been designed and manufactured that allowed checking of the performance of the developed phased array probe. Because of the complexity of the probe element configuration, the block has been designed to check the performance of each element, the beam steering angles and the beam skewing angles. Using the elemental functional reference block and the solid axle sample, the performance of the probe was tested. An excellent correlation was found between the modelling and experimental results. The probe performed as was predicted in the modelling.
In order to develop an optimum electromagnetic technique for hollow axle inspection a comprehensive programme of numerical analysis and simulation work was carried out to select the most promising and suitable electromagnetic technique for hollow axle inspection. Electromagnetic methods such as eddy current testing (ECT), magnetic flux leakage (MFL), alternating current field measurement (ACFM) and remanent flux leakage (RFL) were explored. After preliminary analysis of the inspection problem, several arrangements of the transducers were selected for detailed analysis. Numerical modelling (finite element method - FEM) is used as an effective way of evaluation and designing. Two different numerical models were prepared: 3D models for general evaluation and 2D models for detailed analysis and optimisation. This procedure was selected because analysis of the 3D models requires lengthy calculations, and the achieved precision of the results is limited by the size of the mesh elements. The analysis allows achievement of the magnetic flux and eddy currents spatial distribution. Based on these results, signals from the sensors (magnetic field detectors) are calculated. Results of the numerical modelling can also be utilised in order to optimise construction of the transducers (size and shape of the excitation elements, location, orientation and required sensitivity of the magnetic field detectors). This comprehensive study concluded that the most suitable electromagnetic method for the hollow axle inspection is the RFL. In summary:
• The remanent flux leakage transducers are the most promising. The AC and DC excitation allows high sensitivity and simple instrumentation to be achieved
• The experiments show that the remanent flux leakage transducer offers very good sensitivity.
As part of the electromagnetic technique developments several configurations of electromagnetic transducers were evaluated during preliminary investigations utilizing results of numerical analysis and simulations. Taking into account the sensitivity, spatial resolution, speed of inspection, simplicity of the transducer and the whole measuring system, the RFL method was selected. An electromagnetic transducer was designed and manufactured based on this method. The developed electromagnetic transducer will be named RFL3AAM (Remanent Flux Leakage 3 Axis AMR Matrix). It is similar in construction to the RFLDGM transducer (Remanent Flux Leakage Differential GMR Matrix). The RFL3AAM transducer consists of a permanent magnet which magnetizes the ferromagnetic axle sample. The residual magnetization of the sample is measured by a matrix of magnetic field sensors. In order to obtain a more complete understanding of the flux distribution, the GMR (Giant MagnetoResistive) single axis sensor was replaced with an AMR (Anisotropic MagnetoResistive) three axis sensor. The AMR sensor enables information to be obtained regarding the direction of the magnetic field without any additional biasing. Simultaneous monitoring of all three magnetic field components enhances the capability to detect defects having a different orientation with a similar sensitivity. There are several magnetic field sensors available on the market. These allow measuring all three magnetic field components (x, y and z) simultaneously. After analysis of the sensor parameters (range of field magnitude, resolution, sampling speed, size of the housing, type of the interface) the sensor HMC5883L from Honeywell was selected. Using the developed technique, the intended scenario of the defect’s identification and characterization can be as follows:
1. the length of the defect based on component Bz,
2. the position and width of the defect based on component By,
3. the depth of the defect based on component Bx, taking into account information achieved in step 1 and 2.
In order to determine the parameters of the ultrasonic transducers, modelling studies were carried out to design the ultrasonic transducers that will meet the inspection requirements. The beam profile of different simulated transducers was studied, and the probe/wedge design was selected based on the beam formation, intensity and focusing depth. Following selection of the ultrasonic transducers, further simulations were carried out to determine the defect response performance of the transducers. A large number of simulations were required to cover all the different critical areas of the axle, and to ensure that the selected transducers meet the inspection requirements. The simulation work was divided in two parts:
• Simulations for transverse defects, in which their surface is vertical to the axle’s axis
• Simulations for longitudinal defects, in which their surface is parallel to the axle’s axis.
The response of transverse, as well as the longitudinal defects, was simulated in seven critical positions in respect to the distance from the axle’s end face. Scanning along the axle’s axis and circumferential in the borehole was performed for the detection of transverse defects. Transverse defects are more likely to occur in-service, and the most likely to cause axle failure. For that reason, the majority of studies are related to transverse defects, and few simulation attempts were made to address the longitudinal defects. Different angle wedges are required for each scanning method because the probe is steered in different directions. Different angle wedges will be required for all the different angles of inspection identified at the end of the simulation study. Following completion of the simulations for the ultrasonic inspection of hollow axle, it was concluded that three ultrasonic transducers are required with a crystal diameter of 10mm, and a 40mm curvature radius, mounted on angle wedges to generate incidence angles in axle of 45°, 60° and 70° The ultrasonic probe head was designed based on the findings of this study.
An automated hollow axle scanner has been designed to allow deployment of the ultrasonic and electromagnetic probes into the hollow axle bore. The axle bores are approximately 1.99m long and 30mm diameter. The scanner includes the circumferential and axial drive mechanisms, and a suitable mounting mechanism to attach the scanner onto the hollow axle. The unit is designed to operate with any control/flaw detection systems which utilises DC Servo Drives, and has capacity for 2A 24V motors. The manipulator will be combined with a motion controller and flaw detector to complete the inspection system. The flange adaptor, collar or clamping assembly is set in position depending on the inspection being performed. The manipulator body is attached to the mounting device and secured in position with the location spigot inserted at a suitable position, depending on the mounting mechanism. The required probe assembly is fitted to the rotary drive and, when all the services are connected, the reference positions can be set.
An overall concept for the data fusion software has been developed by Ideko and the developments were carried out based on this concept. The data fusion software has the following functionalities: control the scanner, acquire the ultrasonic and electromagnetic data, compare the results with the results obtained from an axle without defects, input the asset identification numbers such as bogie, axle, train and store the results in database for future comparisons.
The ultrasonic and electromagnetic systems developed in the project have been integrated and tested with the automated hollow axle scanner in order to ensure the correct functionality. The final system has performed well and met the detection requirements set in the specifications.
Because of the complexity of the element configuration of the developed phased array probe, a bespoke focal law calculator has been developed in order to allow setting the inspection parameters. The focal law calculator has been integrated with the PAUT data acquisition and analysis software. The developed software allows the focal laws to be easily changed (ie steering angles) depending upon the geometry of the solid axle. The software provides flexibility to change the inspection parameters in order to optimise the testing results. The developed phased array probe has been integrated with the solid axle scanner, the PAUT software and tested of the solid axle samples. The integration went smoothly and the system is performing well and meets the main design and inspection requirements.
The solid and hollow axle systems developed have been tested in a laboratory environment. A number of different solid axles with different geometries have been tested using the system and good detection capabilities and sensitivity have been achieved. The ultrasonic and electromagnetic techniques have been tested on hollow axle samples and the data fusion software has been finalised during these trials. The field trials of the systems were carried out on 16-18 September 2013 in Eastleigh at Arlington Fleet Services. The access to the maintenance depot and the availability of a train to test the solid axle system was arranged by BBRail. The system was tested in a real environment and the solid axle selected was very complicated. However, the detailed analysis of the phased array data showed that all the geometrical features of the axle have been identified, and a good correlation between the AxleInspect system’s results and the results from the conventional ultrasonic approved procedure was achieved. The field trial’s testing demonstrated that all the critical areas of the axle can be covered with the developed system. This demonstrates the capability of the developed system. Furthermore, the final hollow axle system and data fusion software was tested on hollow axle samples with and without defects, and both techniques (UT and EM) detected all the defects introduced into the axles with high accuracy.
Potential Impact:
The AxleInspect project will create significant impact and economic value for the consortium SMEs through development of an innovative inspection technology for inspecting the structural integrity of train axles, which can improve the safety of trains. The 2D PAUT based novel design will give the consortium SMEs a competitive edge in the rail industry’s inspection business. Europe relies heavily on its advanced and widely used rail infrastructure, and the annual rail ridership across Europe is over 380 billion passenger kilometres. Trains serve as a vital mean of transport for Europe, conveying commuters nationally within cities and internationally across borders.
The current inspection process requires removal of the wheelset from the wagons/locomotive bogies. Only complete disassembly of the wheelset facilitates access to allow inspecting axles using ultrasonic testing, visual inspection or magnetic particle inspection (MPI). It takes, on average, four hours to remove an axle from a bogie; which involves removing the bogie from the wagon. Axles on wheelsets are connected to a number of ancillary components, such as brakes, bearings and supporting structures, making disassembly very time consuming and expensive. It is also evident that constant disassembly and assembly of the wheelsets can result in long-term axle reliability problems. Railway axles undergo MPI every-time they are sent to a workshop, unless they are brand new axles. This process involves three qualified personnel - two performing the job and one safety lookout. It takes an average of four hours to MPI the axle, which involves stripping the paint, performing MPI and re-painting. AxleInspect will only take approximately one hour to inspect, reducing the overall time by three hours.
The AxleInspect developed technologies will offer the following benefits:
• Innovative PAUT probe that will be adaptable to different end face and axle geometries.
• Inspection of solid axles that require minimum disassembly of the bogie and, therefore, will allow more frequent inspections to be undertaken.
• Solid axle inspection system will offer testing accuracy, repeatability and recording ability of the results
• Excellent comparison to the results achieved by existing and approved techniques.
• Innovative hollow axle inspection system that allows inspection of the critical areas simultaneously using both ultrasonic and electromagnetic techniques.
• Data fusion software that allows the operator to assess the condition of the hollow axle using two complimentary NDT methods.
List of Websites:
A project website was set up at the start of the project to facilitate and act as a communication tool for the consortium. The website consists of two main areas: one accessible to the public, and one only accessible by the members of the consortium. The website is used also for dissemination of the project results.
The project website address is: www.axleinspect.eu.
Technological advances in train design during the last 5 years period have enabled high-speed trains to become more commonly used. The increasing trend for the industry’s business is forecast to continue in the next ten years, since rail transportation is steadily becoming a more attractive option over other means of transport for the public. This is because train travel is generally cheaper than using a car, and usually the fastest option to reach a destination. It is also inherently safer and far more environmentally friendly in comparison to car travel, without compromising passenger convenience. As a result, today’s rail networks across Europe are getting busier, with trains travelling at higher speeds and carrying more passengers and heavier axle loads than ever before. The loads and stress that axles experience are often inadequately defined. Rolling stock, including passenger and freight wagons, are designed to last for thirty years, but are often used for much longer. The combination of these factors has put considerable pressure on the existing infrastructure, leading to increased demands in inspection and maintenance of rail assets. The expenditure on inspection and maintenance has grown steadily over the last few years without being followed by a significant improvement in the industry’s safety records. As a direct consequence, one of the immediate key challenges faced by the rail industry is improvement in the safety of the railway systems of EU member states, in both environmental and financial terms, by delivering further efficiencies and exploiting technological innovation.
The structural integrity of wheelsets used in rolling stock is of great importance to the rail industry and its customers. A number of rail accidents have been directly related to the failure of axles, leading to increased demands for the inspection and maintenance of such components. The risk of axle failure relates to both distance covered and time in use. Therefore, it is preferable for more frequently scheduled inspections to occur during the lifetime of the axle, rather than waiting for the less frequent overhaul periods in order to detect any growing fatigue cracks.
Indeed, strict rail standards such as the Railway Group Standards, exist to ensure that axles are inspected at regular and frequent periods to ensure safety is maintained. Depending on the train service operator, it is not unusual for axles to be checked by in-service inspection every 30,000km. To be cost effective and to cause minimal disruption to train services, the NDT inspection should not require the disassembly of the wheelsets and supporting bogies from the vehicles. In addition, there is also evidence that full disassembly can be counterproductive and lead to future maintenance problems due to the required assembly/disassembly of the parts during the service overhaul.
Although railway axles have been designed and used for nearly 200 years, and knowledge about fatigue prevention can be used to assist in the timing of crack inspection intervals, the severity of the mechanical environment in which axles are used means a number of failures, although rare, still occur.
The European rail network is targeting a considerable expansion of passenger and freight traffic by 2020. In order to achieve this, increased reliability and availability of rolling stock is necessary whilst maintaining the same, or better, level of safety. The axle life is a crucial part of both the safety and economic performance of the vehicles, and the axle deteriorates through its lifetime by means of fatigue and corrosion mechanisms. Periodic inspection is used to ensure that these mechanisms have not compromised the axle safety; however, inspection that takes a vehicle out of service impacts on the economic aspects of train operation. It is clear that errors are made in both aspects: Recent serious incidents due to axle failure, in Cologne (2008) and Viareggio (2009), were due to the failure of the inspection regime, and continue to affect the inspection periodicity. On the other hand axles with suspected cracks, on passenger and freight vehicles, are being taken out of service after only 3 years, when their lifetime should be at least 20 years, because of fears that cracks may initiate at corrosion pits.
Usually the inspection interval is such that an inspection is required between major overhauls, so inspection of axles while the train is in-service is still required. This is currently inconvenient and costly for the train operators. If withdrawals take place while a train is in-service, there is a consequent significant cost associated while the train is out of service for additional maintenance. There is also the financial and energy cost in making and supplying a new axle. This has become a specific problem for the highly detailed hollow axle designs such as those used in high-speed passenger trains.
Project Context and Objectives:
The AxleInspect project aimed to develop new inspection techniques based on phased array ultrasonic testing (PAUT), conventional ultrasonic testing (UT) and electromagnetic (EM) testing suitable for the inspection of both solid and hollow axles. For solid axles, techniques have been developed that inspect from the end face of the axle using new and novel phased array inspection technology. For hollow axles, probes based on ultrasonic and electromagnetic inspection have been developed, enabling detection of surface breaking defects that cannot be found by existing inspection techniques.These new techniques will allow inspection of axles whilst they are still attached to their supporting bogie, allowing minimal disassembly from the train.
AxleInspect is a collaborative project comprising the following organisations: TWI Ltd, Ideko, ZUT, Vermon, Phoenix Inspection Systems, DANOBAT Railway Systems and Balfour Beatty Rail. The project is co-ordinated and managed by TWI Ltd and is partly funded by the EC under the Research for the benefits of SMEs Project Ref: FP7-SME-2011-286573.
The main technical objectives of the AxleInspect project are:
• Develop and produce an instrument to detect transverse cracking on the critical areas such as cross-sectional area changes (curvatures), wheelseat/brake disc/gear set regions) of a solid axle whilst the axle is assembled in its wheelset and bogie. This has been achieved by:
• Development of a phased array probe to inspect the axles from the end face.
• Development of a scanning jig to carry out the inspection.
• Development of software to acquire, display and analyse the PAUT data.
• Develop and produce an instrument to inspect hollow axles using a combined rotating UT and EM probe. The instrument should be capable of easy installation in an inspection station. This has been achieved by:
• Development of a suitable electromagnetic probe for inspection of inner surface breaking cracks from inside the bore.
• Development of ultrasonic probes for the inspection of the outer diameter for outer surface breaking defects from inside the bore.
• Development of an automated scanner to deploy the UT and EM probes.
• Development of data fusion software to display the results from both inspection techniques.
Project Results:
The AxleInspect project has achieved significant technical and scientific results. A number of different solid axles have been obtained from the AxleInspect partners, and also from previous TWI completed projects, to allow the development of project technologies. A wide range of solid axles, with different end face diameters and geometry, were available for use in the experimental work. Furthermore, a hollow axle from a high speed train has been provided by the partners to allow development of the hollow axle inspection techniques. A number of simulated defects have been introduced into the solid and hollow axles to enable development of the ultrasonic and electromagnetic techniques.
Extensive ultrasonic modelling has been carried out to design a new phased array probe configuration that will be suitable for inspecting the critical areas of solid axles from the end face. A number of different phased array probe configurations, with different shapes and numbers of elements, have been considered and the most suitable design was selected for further modelling and optimisation. A new innovative phased array probe has been designed that will meet the inspection requirements from the axle end face. The main parameters that the phased array probe will fulfil are: adaptable to different end face diameters and configurations, steering and skewing abilities, large focal area and small focal spot size and ability to cover the critical areas of different solid axle geometries. The new phased array design was simulated to determine the interaction of the generated ultrasonic beam with simulated defects at the critical areas. The simulation results show that the defects in the axle critical areas can be identified with enhanced detection capabilities. The developed phased array design was manufactured and tested on the axle samples. Excellent correlation was achieved between the modelling and experimental results, and the probe performance is as predicted by the modelling.
A comprehensive numerical analysis was carried out to determine the most suitable electromagnetic method for inspection of hollow axles to detect inner surface breaking cracks. A number of different methods have been selected for evaluation, such as eddy current testing (ECT), magnetic flux leakage (MFL), alternating current field measurement (ACFM) and remanent flux leakage (RFL). After preliminary analysis of the problem, several arrangements of the electromagnetic transducers were selected for detailed numerical analysis. A numerical modelling (finite element method - FEM) is used as an effective way of evaluation and design. Two different numerical models were prepared: 3D models for general evaluation and 2D models for detailed analysis and optimisation. The electromagnetic numerical analysis concluded that the best method for this application is the RFL. Two innovative RFL transducers were designed, manufactured and tested. The results obtained from the final optimised electromagnetic transducers are promising. The developed technique has the capability to accurately size the defects using the different magnetic flux components (x, y and z).
Furthermore, ultrasonic modelling was carried out to determine the most suitable parameters for the ultrasonic probes in order to detect outer surface breaking defects from the inside bore of the hollow axle. The beam profile, and the interaction with defects at critical areas, have been simulated during the extensive modelling programme.
A semi –automatic solid axle scanner was designed and manufactured that allows the inspection of solid axles from the end face using the developed phased array probe. The developed scanner has a modular design that allows for ease of mounting as well as providing adaptability to different train axle boxes and different end face diameters. The solid scanner combined with the developed phased array software allows the acquisition of encoded data that can be saved for further analysis. The scanner allows the accurate positioning of the probe and provides repeatable and recordable inspections to be carried out.
An automated hollow axle scanner was manufactured that will be used to deploy the developed ultrasonic and electromagnetic probes. Data fusion software was developed that can control the scanner motion, acquire the ultrasonic and electromagnetic data and finally analyse the data. The software can compare the acquired results with the results obtained from an axle without any defects, which will help the operator to identify any defects. Furthermore, the developed data fusion software allows the end user to input various parameters in order to identify the asset inspected (ie axle number, train number, bogie number etc). All the data is saved in a database for future comparisons.
More details of the work carried out in the project are provided below.
The system specification document has been completed and the requirements of the inspection systems have been defined. More specifically, the defects’ requirements for their sizes and location, axle geometry and inspection requirements have been defined. The requirements for the phased array probe and technical constraints have also been defined. The inspection requirements have been selected by using railway inspection standards that define the acceptable frequencies to be used and the inspection methods. These inspection standards are providing guidance only since they are used for conventional ultrasonic techniques and are not directly applicable to phased array inspections.
A number of axle samples have been obtained from previous TWI projects and also provided by the AxleInspect partners. The samples acquired for the project include seven different designs of solid axles and one design of hollow axle. The selected axle designs are representative of the axles commonly used. The acquired axles provide a good range of different parameters such as the number of critical areas and their distance with respect to the end face, curvature radii and outer diameters. The available solid axles have different end face diameters which will allow the adaptability of the developed solid axle inspection tool to be tested.
Artificial defects have been introduced, at critical areas, into one of the solid axles being used to test the new phased array probe. Furthermore, artificial defects have been introduced into the inner surface of the bore in the hollow axle being used for the electromagnetic technique development. More defects were introduced into the remaining solid axles after the initial work with the newly developed probe was completed. Defects could be introduced at TWI’s facilities, giving the flexibility to add artificial defects as necessary during the experimental work.
One of the main aims of the project was to design and develop a new and novel 2D phased array ultrasonic probe for inspection of solid axles from the end face. An extensive simulation programme was carried out to design the new phased array probe. The new design provided the following functionalities: adaptability to different axle geometries’ beam steering and skewing abilities, large focal length and small focal spot. A number of different configurations for the element size and probe shape were simulated to find the optimum probe design. The probe to be used, both in pulse echo and in pitch–catch modes, will consist of a maximum 256 elements, thus all the simulated probe designs had this element number as their limit. The reason that this number of elements was chosen was mainly due to the availability of a suitable phased array pulser-receiver that will be used to excite the probe and inspect the axle samples. Furthermore, a higher number of elements in the probe will result in a very expensive probe, both in terms of development and as a commercial proposition. Additionally, the electronic systems available can only handle 256 elements simultaneously. In total, 26 different probe configurations have been simulated in order to determine the most suitable probe for the application. Due to the number of elements involved, the simulations were very computing intensive and each model took more than one day to be completed. The final design of the probe has 248 elements with different element sizes within the same array. The simulation work of the developed PAUT probe was carried out to measure and evaluate the defect response of the ultrasonic beam. In order to achieve this, a number of simulated defects were placed at several positions on the axle. Investigation of the defect response is important so that it can provide a reference, and prove that the proposed phased array probe can give the desired results and meet the inspection requirements. For the solid axles, the target is to inspect from the end face, which is a challenging procedure. For that reason the developed LM (Linear Matrix) probe has to operate in two different inspection modes:
• pulse-echo (sectorial scanning on one plane) when the middle of the probe is not positioned over a bolt hole and
• pitch-catch (skewing and steering scanning) that will be used when the middle of the probe is positioned over a bolt hole.
A number of defects with different dimensions were placed in critical positions, and simulated both separately and simultaneously. A number of different simulation parameters have been tested in order to acquire the maximum amount of simulated data. The results show that the developed phased array probe can meet the inspection requirements and performance. The design and manufacturing of the new phased array probe has been carried out by Vermon. Phoenix ISL was responsible for the design, manufacture and assembly of the solid axle scanner and probe holder that will be used to carry out the inspections from the axle end face. The inspection requirements to be satisfied by the scanner have been defined in the project and are listed below:
• Mount onto a wide range of axles and wheel bearing housings.
• Ability to position the probe to end face diameters from 110-165mm.
• Perform 360° encoded circumferential scan.
• Rigidly hold the large bespoke phased array transducer and load it onto the inspection surface.
• Provide axial movement to position the transducer as required.
The developed scanner can be used for:
• Inspection of dismantled wheel sets via the use of switchable electromagnets
• Inspection of fully assembled wheel sets via a mounting plate.
An elemental functional reference block has been designed and manufactured that allowed checking of the performance of the developed phased array probe. Because of the complexity of the probe element configuration, the block has been designed to check the performance of each element, the beam steering angles and the beam skewing angles. Using the elemental functional reference block and the solid axle sample, the performance of the probe was tested. An excellent correlation was found between the modelling and experimental results. The probe performed as was predicted in the modelling.
In order to develop an optimum electromagnetic technique for hollow axle inspection a comprehensive programme of numerical analysis and simulation work was carried out to select the most promising and suitable electromagnetic technique for hollow axle inspection. Electromagnetic methods such as eddy current testing (ECT), magnetic flux leakage (MFL), alternating current field measurement (ACFM) and remanent flux leakage (RFL) were explored. After preliminary analysis of the inspection problem, several arrangements of the transducers were selected for detailed analysis. Numerical modelling (finite element method - FEM) is used as an effective way of evaluation and designing. Two different numerical models were prepared: 3D models for general evaluation and 2D models for detailed analysis and optimisation. This procedure was selected because analysis of the 3D models requires lengthy calculations, and the achieved precision of the results is limited by the size of the mesh elements. The analysis allows achievement of the magnetic flux and eddy currents spatial distribution. Based on these results, signals from the sensors (magnetic field detectors) are calculated. Results of the numerical modelling can also be utilised in order to optimise construction of the transducers (size and shape of the excitation elements, location, orientation and required sensitivity of the magnetic field detectors). This comprehensive study concluded that the most suitable electromagnetic method for the hollow axle inspection is the RFL. In summary:
• The remanent flux leakage transducers are the most promising. The AC and DC excitation allows high sensitivity and simple instrumentation to be achieved
• The experiments show that the remanent flux leakage transducer offers very good sensitivity.
As part of the electromagnetic technique developments several configurations of electromagnetic transducers were evaluated during preliminary investigations utilizing results of numerical analysis and simulations. Taking into account the sensitivity, spatial resolution, speed of inspection, simplicity of the transducer and the whole measuring system, the RFL method was selected. An electromagnetic transducer was designed and manufactured based on this method. The developed electromagnetic transducer will be named RFL3AAM (Remanent Flux Leakage 3 Axis AMR Matrix). It is similar in construction to the RFLDGM transducer (Remanent Flux Leakage Differential GMR Matrix). The RFL3AAM transducer consists of a permanent magnet which magnetizes the ferromagnetic axle sample. The residual magnetization of the sample is measured by a matrix of magnetic field sensors. In order to obtain a more complete understanding of the flux distribution, the GMR (Giant MagnetoResistive) single axis sensor was replaced with an AMR (Anisotropic MagnetoResistive) three axis sensor. The AMR sensor enables information to be obtained regarding the direction of the magnetic field without any additional biasing. Simultaneous monitoring of all three magnetic field components enhances the capability to detect defects having a different orientation with a similar sensitivity. There are several magnetic field sensors available on the market. These allow measuring all three magnetic field components (x, y and z) simultaneously. After analysis of the sensor parameters (range of field magnitude, resolution, sampling speed, size of the housing, type of the interface) the sensor HMC5883L from Honeywell was selected. Using the developed technique, the intended scenario of the defect’s identification and characterization can be as follows:
1. the length of the defect based on component Bz,
2. the position and width of the defect based on component By,
3. the depth of the defect based on component Bx, taking into account information achieved in step 1 and 2.
In order to determine the parameters of the ultrasonic transducers, modelling studies were carried out to design the ultrasonic transducers that will meet the inspection requirements. The beam profile of different simulated transducers was studied, and the probe/wedge design was selected based on the beam formation, intensity and focusing depth. Following selection of the ultrasonic transducers, further simulations were carried out to determine the defect response performance of the transducers. A large number of simulations were required to cover all the different critical areas of the axle, and to ensure that the selected transducers meet the inspection requirements. The simulation work was divided in two parts:
• Simulations for transverse defects, in which their surface is vertical to the axle’s axis
• Simulations for longitudinal defects, in which their surface is parallel to the axle’s axis.
The response of transverse, as well as the longitudinal defects, was simulated in seven critical positions in respect to the distance from the axle’s end face. Scanning along the axle’s axis and circumferential in the borehole was performed for the detection of transverse defects. Transverse defects are more likely to occur in-service, and the most likely to cause axle failure. For that reason, the majority of studies are related to transverse defects, and few simulation attempts were made to address the longitudinal defects. Different angle wedges are required for each scanning method because the probe is steered in different directions. Different angle wedges will be required for all the different angles of inspection identified at the end of the simulation study. Following completion of the simulations for the ultrasonic inspection of hollow axle, it was concluded that three ultrasonic transducers are required with a crystal diameter of 10mm, and a 40mm curvature radius, mounted on angle wedges to generate incidence angles in axle of 45°, 60° and 70° The ultrasonic probe head was designed based on the findings of this study.
An automated hollow axle scanner has been designed to allow deployment of the ultrasonic and electromagnetic probes into the hollow axle bore. The axle bores are approximately 1.99m long and 30mm diameter. The scanner includes the circumferential and axial drive mechanisms, and a suitable mounting mechanism to attach the scanner onto the hollow axle. The unit is designed to operate with any control/flaw detection systems which utilises DC Servo Drives, and has capacity for 2A 24V motors. The manipulator will be combined with a motion controller and flaw detector to complete the inspection system. The flange adaptor, collar or clamping assembly is set in position depending on the inspection being performed. The manipulator body is attached to the mounting device and secured in position with the location spigot inserted at a suitable position, depending on the mounting mechanism. The required probe assembly is fitted to the rotary drive and, when all the services are connected, the reference positions can be set.
An overall concept for the data fusion software has been developed by Ideko and the developments were carried out based on this concept. The data fusion software has the following functionalities: control the scanner, acquire the ultrasonic and electromagnetic data, compare the results with the results obtained from an axle without defects, input the asset identification numbers such as bogie, axle, train and store the results in database for future comparisons.
The ultrasonic and electromagnetic systems developed in the project have been integrated and tested with the automated hollow axle scanner in order to ensure the correct functionality. The final system has performed well and met the detection requirements set in the specifications.
Because of the complexity of the element configuration of the developed phased array probe, a bespoke focal law calculator has been developed in order to allow setting the inspection parameters. The focal law calculator has been integrated with the PAUT data acquisition and analysis software. The developed software allows the focal laws to be easily changed (ie steering angles) depending upon the geometry of the solid axle. The software provides flexibility to change the inspection parameters in order to optimise the testing results. The developed phased array probe has been integrated with the solid axle scanner, the PAUT software and tested of the solid axle samples. The integration went smoothly and the system is performing well and meets the main design and inspection requirements.
The solid and hollow axle systems developed have been tested in a laboratory environment. A number of different solid axles with different geometries have been tested using the system and good detection capabilities and sensitivity have been achieved. The ultrasonic and electromagnetic techniques have been tested on hollow axle samples and the data fusion software has been finalised during these trials. The field trials of the systems were carried out on 16-18 September 2013 in Eastleigh at Arlington Fleet Services. The access to the maintenance depot and the availability of a train to test the solid axle system was arranged by BBRail. The system was tested in a real environment and the solid axle selected was very complicated. However, the detailed analysis of the phased array data showed that all the geometrical features of the axle have been identified, and a good correlation between the AxleInspect system’s results and the results from the conventional ultrasonic approved procedure was achieved. The field trial’s testing demonstrated that all the critical areas of the axle can be covered with the developed system. This demonstrates the capability of the developed system. Furthermore, the final hollow axle system and data fusion software was tested on hollow axle samples with and without defects, and both techniques (UT and EM) detected all the defects introduced into the axles with high accuracy.
Potential Impact:
The AxleInspect project will create significant impact and economic value for the consortium SMEs through development of an innovative inspection technology for inspecting the structural integrity of train axles, which can improve the safety of trains. The 2D PAUT based novel design will give the consortium SMEs a competitive edge in the rail industry’s inspection business. Europe relies heavily on its advanced and widely used rail infrastructure, and the annual rail ridership across Europe is over 380 billion passenger kilometres. Trains serve as a vital mean of transport for Europe, conveying commuters nationally within cities and internationally across borders.
The current inspection process requires removal of the wheelset from the wagons/locomotive bogies. Only complete disassembly of the wheelset facilitates access to allow inspecting axles using ultrasonic testing, visual inspection or magnetic particle inspection (MPI). It takes, on average, four hours to remove an axle from a bogie; which involves removing the bogie from the wagon. Axles on wheelsets are connected to a number of ancillary components, such as brakes, bearings and supporting structures, making disassembly very time consuming and expensive. It is also evident that constant disassembly and assembly of the wheelsets can result in long-term axle reliability problems. Railway axles undergo MPI every-time they are sent to a workshop, unless they are brand new axles. This process involves three qualified personnel - two performing the job and one safety lookout. It takes an average of four hours to MPI the axle, which involves stripping the paint, performing MPI and re-painting. AxleInspect will only take approximately one hour to inspect, reducing the overall time by three hours.
The AxleInspect developed technologies will offer the following benefits:
• Innovative PAUT probe that will be adaptable to different end face and axle geometries.
• Inspection of solid axles that require minimum disassembly of the bogie and, therefore, will allow more frequent inspections to be undertaken.
• Solid axle inspection system will offer testing accuracy, repeatability and recording ability of the results
• Excellent comparison to the results achieved by existing and approved techniques.
• Innovative hollow axle inspection system that allows inspection of the critical areas simultaneously using both ultrasonic and electromagnetic techniques.
• Data fusion software that allows the operator to assess the condition of the hollow axle using two complimentary NDT methods.
List of Websites:
A project website was set up at the start of the project to facilitate and act as a communication tool for the consortium. The website consists of two main areas: one accessible to the public, and one only accessible by the members of the consortium. The website is used also for dissemination of the project results.
The project website address is: www.axleinspect.eu.