GM/GN2642. Guidance on Wheel / Rail Low Adhesion Measurement. Railway Group Guidance Note. Uncontrolled When Printed. Published by

Transcription

1 GN Published by Rail Safety and Standards Board Evergreen House 160 Euston Road London NW1 2DX Copyright 2008 Rail Safety and Standards Board Limited GM/GN2642 Issue One: February 2008 Railway Group Guidance Note

2 Issue Record Issue Date Comments One February 2008 Original document Superseded documents Supply This Railway Group Guidance Note does not supersede any other Railway Group documents. Controlled and uncontrolled copies of this Railway Group Guidance Note may be obtained from the Corporate Communications Department, Rail Safety and Standards Board, Evergreen House, 160 Euston Road, London NW1 2DX, telephone or Railway Group Standards and associated documents can also be viewed at Page 2 of 61 RAIL SAFETY AND STANDARDS BOARD

3 Contents Uncontrolled When Printed Section Description Page Part 1 Introduction Purpose of this document Copyright Approval and authorisation of this document 5 Part Background to wheel / rail low adhesion Application of this document Low adhesion measurement methods Selection of measuring method Recording conditions 13 Part 3 Train Mounted Methods Scope Purpose Uses Tribometer train Load measuring wheels Load measuring wheels - stick-slip oscillations Strain gauged axles 22 Part 4 Vehicle Borne Tribometer Scope Purpose Uses Vehicle borne tribometer method 23 Part 5 Portable Tribometer Methods Scope Purpose Uses Portable tribometers Braked wheel tribometer Static breakout friction tribometer Skid resistance slider tribometer 32 Part 6 Laboratory Test Rigs Scope Purpose Uses Twin disc test rig Pin-on-disc machine Full scale wheel-on-rail rig 37 Part 7 Service Train Scope Purpose Uses OTMR data TMS WSP data 42 Part 8 Service / Test Train Braking Test Scope Purpose Uses Train braking method 44 RAIL SAFETY AND STANDARDS BOARD Page 3 of 61

4 Part 9 Adhesion / Speed Relationship 47 Part 10 Measurement of Contamination Properties Scope Purpose Uses Measurement of contamination 48 Appendix A Examples of Recording Adhesion Conditions 55 Definitions 59 References 61 Tables Table 1 Options for assessing adhesion levels 9 Table 2 Advantages / limitations of low adhesion measurement and assessment methods 10 Table 3 Condition matrix 13 Table 4 Example table for recording low adhesion site conditions 16 Table 5 Comparison of adhesion values: portable tribometer and train 27 Table 6 Distribution of friction coefficients 30 Table 7 Bottom disc diameters at percentage creep levels 35 Table 8 Shear torque readings at different leaf film contamination conditions 52 Table 9 Pencil hardness arbitrary range 53 Table 10 Pencil hardness values for different track conditions 54 Figures Figure 1 Load measuring wheel 19 Figure 2 Wheel / rail lateral force stick-slip oscillations 21 Figure 3 Vehicle borne tribometer 23 Figure 4 Friction profile for mitigation testing 24 Figure 5 Braked wheel tribometer 26 Figure 6 Compensation for changing adhesion conditions 28 Figure 7 Kett / Heidon handheld static breakout friction tribometer 29 Figure 8 BR research tribometer measuring wheelset 30 Figure 9 Sample adhesion measurements between Kett and Munro tribometers 31 Figure 10 Pendulum tribometer for skid resistance measurement 32 Figure 11 Relationship between skid resistance and wheel / rail adhesion levels 33 Figure 12 Amsler twin disc machine 34 Figure 13 Amsler twin discs (40mm diameter) 34 Figure 14 Amsler percentage creep levels showing peak adhesion levels 35 Figure 15 Pin-on-disc friction and wear machine 36 Figure 16 Full-scale wheel-on-rail rig 37 Figure 17 OTMR speed-time output 40 Figure 18 Mitrac map identifying events on a route 41 Figure 19 LAWS low adhesion network map 42 Figure 20 Measurement of train deceleration for estimating adhesion levels 45 Figure 21 Measurements for assessing a low adhesion mitigation method 46 Figure 22 Adhesion / speed relationship 47 Figure 23 Sonacoat thickness gauge 49 Figure 24 Test results for two treatment methods 50 Figure 25 Normalised thickness 50 Figure 26 Hand held surface roughness measuring device 51 Figure 27 Moisture and temperature compensation probes 51 Figure 28 Shear torque measurements 53 Figure 29 Pencil hardness tester 53 Page 4 of 61 RAIL SAFETY AND STANDARDS BOARD

5 Part 1 Introduction 1.1 Purpose of this document This document has been published by Rail Safety and Standards Board to give guidance on wheel / rail low adhesion measurement. It does not constitute a recommended method of meeting any set of mandatory requirements. RSSB make no endorsement for the use or identification of the equipment and its uses within the document. In association with the fundamental study of leaf contamination, this guidance note can be used in conjunction with Guidance Note GM/GN2643, Wheel / Rail Low Adhesion Simulation. This Guidance Note and GM/GN2643 have been produced to aid in the selection of adhesion measuring and monitoring methods and simulate low adhesion. Stakeholders from the industry-supported Adhesion Working Group (AWG) approved the content of this work. Further Guidance Notes may be produced related to Low Adhesion as outputs from other studies in this topic are completed. 1.2 Copyright Copyright in the Railway Group documents is owned by Rail Safety and Standards Board Limited. All rights are hereby reserved. No Railway Group document (in whole or in part) may be reproduced, stored in a retrieval system, or transmitted, in any form or means, without the prior written permission of Rail Safety and Standards Board Limited, or as expressly permitted by law. RSSB members are granted copyright licence in accordance with the Constitution Agreement relating to Rail Safety and Standards Board Limited. In circumstances where Rail Safety and Standards Board Limited has granted a particular person or organisation permission to copy extracts from Railway Group documents, Rail Safety and Standards Board Limited accepts no responsibility for, and excludes all liability in connection with, the use of such extracts, or any claims arising therefrom. This disclaimer applies to all forms of media in which extracts from Railway Group Standards may be reproduced. 1.3 Approval and authorisation of this document The content of this document was approved by: Rolling Stock Standards Committee on 06 July 2007 This document was authorised by RSSB on 05 December 2007 RAIL SAFETY AND STANDARDS BOARD Page 5 of 61

6 Part 2 Guidance on Wheel / Rail Low Adhesion Measurement 2.1 Background to wheel / rail low adhesion Over the last 30 years a significant effort has been applied to developing mitigation methods for low adhesion conditions. In the late 1970 s and 1980 s British Rail Research conducted a wide range of projects to measure adhesion levels on the network and to test the effectiveness of methods to combat leaf fall contamination At this time sandite was developed and became the standard method of treating contaminated rails. However, the application of sandite requires a special low speed train which impacts on the timetable. Additionally, sand is an insulator and therefore puts track circuit operation at risk. Hence there has been continued interest in alternative treatment methods such as water jetting, lasers, mechanical scrubbing and chemicals, each of which has its own advantages and limitations RSSB conducted a research project The Characteristics of Railhead Leaf Contamination. The objective was to provide a comprehensive overview of current and previous research on low adhesion with respect to key areas chosen by stakeholders during the consultation process. The research included a review of past research into low adhesion management and identified and recommended further work in five other areas. These were: a) Influence of traction systems on low adhesion performance b) Whole train control systems c) Creating and measuring low adhesion levels for research and test purposes d) Aerodynamic performance of rolling stock e) Characteristics of leaf contamination properties RSSB s research into The Characteristics of Railhead Leaf Contamination in relation to the requirements for measuring wheel / rail low adhesion were: a) To identify industry requirements and purposes for measuring low adhesion levels. b) To review the available devices for measuring low adhesion levels and match the capability of the devices to industry requirements The work identified that moisture content, thickness, shear torque and hardness of leaf film contamination have been measured on a few ad hoc occasions to try and understand the properties of railhead contaminants, but the data has not been acquired consistently and no procedure exists for the application of measuring devices This guidance note lists the reasons for measuring low adhesion levels between wheel / rail and considers which methods are most suited to different requirements and looks at the measuring methods currently available. A range of methods and devices for the measurement of low adhesion conditions have been included as examples An input has also been obtained from other railway administrations including German (DB), French (SNCF), Dutch (NS) and Italian (FS) so that overseas experience could be incorporated into a review of adhesion measurement methods. Page 6 of 61 RAIL SAFETY AND STANDARDS BOARD

7 2.1.8 The relationship between adhesion levels and speed is explained in Part 9. Guidance on the tools that are available for measuring contaminant properties is given in Part Application of this document Low adhesion measurements are required for a number of reasons relating to the safe operation of the railway and for improving performance. These reasons are: a) The safe operation of trains in low adhesion conditions requires the industry to assess the risks and provide mitigation methods (see GE/RT8040, GI/RT7006 and GI/GN7606). Development of successful methods is dependent upon the ability to reliably measure and in some instances reproduce low adhesion conditions. b) Post incident analysis enables the industry to determine the causes of problems that have arisen and therefore improve equipment and procedures (see GO/RT3252 and GO/GN3652). c) Simulated low adhesion levels allows driver experience to be obtained under controlled conditions. An assessment of simulated low adhesion levels enables consistent conditions to be produced for training purposes. d) Simulated low adhesion levels can also be used to test the effect of changes to existing equipment on performance and safety Note that for braking, adhesion levels of 0.14 or greater should sustain all braking rates and an adhesion level between 0.1 and 0.3 is required for reliable operation in traction. 2.3 Low adhesion measurement methods Measurement of adhesion levels can be obtained directly by use of a tribometer, which may be full sized or scaled for manual operation on-site. Scaled tribometers should only be used for comparative measurements because the scaling can influence the magnitude of the results This document provides guidance on the selection of a method for measuring low adhesion levels. It also provides details of the various measuring options; Part 3 (Train Mounted Methods), Part 4 (Vehicle Borne Tribometer) and Part 5 (Portable Tribometer Methods). Part 6 provides guidance on the use of laboratory test rigs for measuring adhesion levels. Report AEATR-VTI under project T354: The Characteristics of Railhead Leaf Contamination is a useful additional reference document on the measurements of low adhesion levels and can be found on the RSSB web site ( Where portable tribometers have been identified, practical measurements were carried out in the laboratory to assess the results obtained from the devices for measurements on a number of contaminants. These results have then been compared to historical data from the full-scale tribometer train Assessment of low adhesion levels can also be obtained indirectly by a number of methods as shown in Part 7 and Part 8. These are: a) Measurement of the train deceleration (for example as acquired from the on train monitor recorder (OTMR) - see 7.4). b) Monitoring wheel slide protection (WSP) activity using train management system (TMS) for example Mitrac - see 7.5 and / or WSP Data for example LAWS see 7.6). c) Timing or measuring the distance for a train to stop from a known braking point (Part 8). RAIL SAFETY AND STANDARDS BOARD Page 7 of 61

8 2.3.5 An adhesion management system (AMS) for monitoring adhesion conditions is under development by the industry but is not yet available and details are therefore not included within this Guidance Note. 2.4 Selection of measuring method The requirement for assessing low adhesion conditions (for example incident investigation, skidpan driver training, test purposes, product testing) will determine the method(s) that are suitable for the application. In this document test purposes refers to checking brake systems, WSP etc and product testing refers to checking products that are used on the railway that could affect the railhead adhesion levels, for example weed killers Table 1 lists industry requirements for the measurement / assessment of adhesion levels and ranks the options (1 being the preferred option) of devices / methods suitable for obtaining data for the specified purposes Choice of tribometer or assessment method will be governed by availability of the device or method and the practicality of using it at the location where the data is to be acquired. Where possible it is recommended that full scale or vehicle borne tribometer should be used in preference to portable tribometer because of scaling effects on the results from the latter The selection process in Table 1 should be used in conjunction with Table 2, which gives guidance regarding the advantages and limitations of using a particular tribometer or assessment method for a specific application As an example of using the Tables 1 and 2, take the scenario of assessing a new rail cleaning technique. Assuming that a service train and section of track are available on which to conduct the tests, Table 1 suggests that there are three options: OTMR, brake test and modified service or test train (as shown in the box where the assess low adhesion treatment methods / rail cleaning intersects with the service test train column). Assuming all three methods are available then the advantages and limitations of each can be obtained from Table 2. Each method is considered in turn and for example one possible outcome would be: a) Option 1 the OTMR, is rejected on the grounds that the staff and equipment for downloading the data are not available. b) Option 2 the brake test is considered the best option on this occasion since the method is simple and only a preliminary assessment is needed. c) Option 3 the modified test train, is rejected because of the time and expense of fitting load measuring wheels For ease of reference the sections of Table 2 are set out in the same sequence as the order of tribometers in the header row of Table 1: a) Train mounted methods b) Vehicle borne tribometers c) Portable tribometers d) Laboratory test rigs e) Service test train f) Adhesion management methods. Page 8 of 61 RAIL SAFETY AND STANDARDS BOARD

9 Train mounted methods Vehicle borne tribometers Portable tribometers Laboratory test rigs Service / test trains Adhesion management methods Incident investigation No No 1. Braked wheel tribometer 2. Static breakout friction No 1. OTMR 2. Brake test No 3. Skid resistance slider Assess low adhesion treatment methods / rail cleaning Yes Yes 1. Braked wheel tribometer 2. Static breakout friction 3. Skid resistance slider 1. Twin disc test rig 2. Pin on disc machine 3. Full-scale wheel-on-rail rig 1. OTMR 2. Brake test 3. Modified service or test train (load measuring wheels or strain gauged axles) No Network adhesion status Identify low adhesion sites Yes No No No 1. TMS / WSP data 2. Modified service train (strain gauged axles) Yes Yes No No 1. TMS / WSP data 2. Modified service train (strain gauged axles) Yes Yes Check adhesion levels for test purposes (see 2.3.2) Yes Yes 1. Braked wheel tribometer 2. Static breakout friction 3. Skid resistance slider No 1. Brake test 2. OTMR 3. Test train (load measuring wheels) No Check adhesion levels for driver training No No 1. Static breakout friction No 1. Brake test distance to stop / deceleration No Laboratory investigations No No 1. Skid resistance slider 1. Twin disc test rig No No 2. Static breakout friction 2. Pin on disc machine 3. Full-scale wheel-on-rail rig Product testing Yes Yes 1. Braked wheel tribometer 2. Skid resistance slider 1. Twin disc test rig 2. Pin on disc machine 1. Brake test 2. OTMR No 3. Static breakout friction 3. Full-scale wheel-on-rail rig Table 1 Options for assessing adhesion levels RAIL SAFETY AND STANDARDS BOARD Page 9 of 61

10 Tribometer type Outputs Advantages Limitations Train mounted methods Tribometer train see 3.4 Exact measurement Provides a direct measurement of the adhesion level. No scaling effects. Adhesion profiles can be measured. Peak adhesion levels can be measured. Could be incorporated into treatment train(s). System not readily available. Tribometer train expensive to construct and operate. Difficult to get to sites and a path is required. Data processing and analysis needed to determine the wheel / rail forces. Flange contact will affect the results Load measuring wheels see 3.5 / 3.6 Exact measurement Full-scale wheel / rail forces are measured to provide the peak adhesion levels and adhesion profiles. Could be fitted to a service vehicle for test purposes (vehicle withdrawn from service whilst modified). Fitted to a train so can measure anywhere on the system. System not readily available. Expensive and complex to build. Needs to be fitted to a train (service or departmental). Significant analysis and data processing required. Flange contact will affect the results. Strain gauged axles see 3.7 Exact measurement Uses full sized wheel / rail contact. Could be fitted to treatment train(s) or potentially to service trains. Peak adhesion levels and adhesion profiles can be obtained. System still under development. Requires fitting to train in harsh environment. Data processing and analysis needed to determine the wheel / rail forces. Flange contact will affect the results. Vehicle borne tribometers Vehicle borne tribometer see 4.4 Relative measurement Readily available and easy to use. Rolling / sliding measurement. Both rails can be measured simultaneously with adjustable wheel position across rail. Measurement speed up to 30 mile/h. Possession required for taking the measurements. More expensive than portable devices. Scaling effects (but more accurate than the portable device). Significant lengths of track measurable. More accurate than portable tribometers. Table 2 Advantages / limitations of low adhesion measurement and assessment methods Page 10 of 61 RAIL SAFETY AND STANDARDS BOARD

11 Tribometer type Outputs Advantages Limitations Portable tribometers Braked wheel tribometer see 5.5 Relative measurement Readily available. Portable and easy to use. Rolling / sliding measurements gives the peak adhesion level. Adjustable wheel position across rail. Scaling effects with no simple calibration, adhesion values obtained are generally high. Low speed measurements (walking pace). Short recording time approximately 15 to 25 seconds per adhesion value measurement. Static breakout friction tribometer see 5.6 Relative measurement Readily available Hand held, easy to use. Normally friction should be measured over a significant length of rail but this device could be used on very short sections (15 to 20 cm) of rail for example on damaged track at a derailment site. Very short recording time per measurement, approximately 5-10 seconds. Sliding friction measurements only. Flat slider will not be in full contact with curved rail. Roughness and lubricants affect the results - scaling effects. Requires modification - rubber slider - to improve resolution of results. Skid resistance slider tribometer see 5.7 Relative measurement Readily available. Easy to use. Portable. Demonstrated to give significant resolution between contaminant types. Design of pendulum frame not very suitable for use on track. Sliding friction measurements only. Slightly more complex set up than braked wheel or static breakout friction tribometers. Scaling effects. Laboratory test rig Laboratory test rigs see Part 6 Direct measurement Readily available. Controlled environment. Rig(s) can be instrumented to measure adhesion level. Contaminants easily introduced into the contact area. Multiple tests can be easily conducted to determine statistical variation. Most rigs are small size wheel-on-wheel configurations (for example Amsler) and scaling affects results. Wear of discs can be high new or turned discs are required for each test. Table 2 Advantages / limitations of low adhesion measurement and assessment methods (continued) RAIL SAFETY AND STANDARDS BOARD Page 11 of 61

12 Tribometer type Outputs Advantages Limitations Service / Test trains Service trains OTMR data see 7.4 Average adhesion levels over length of track Readily available. All trains are required to have OTMR fitted. Most OTMR systems require downloading after the incident. Measures effect of adhesion levels in a real wheel / rail contact. The railway industry are familiar with the output from OTMR since it is in regular use for incident analysis. OTMR requires search and analysis of data. Output is an average value. Only provides data when the train is braking. Experience with analysing the data is necessary. Service trains TMS for example Mitrac see 7.5 Provides estimate of adhesion levels Readily available. Measures effect of adhesion levels in a real wheel / rail contact. Mitrac can be accessed in real time. Only provides data when the train is braking. Output is an estimate of adhesion level. Experience with using the software is necessary. Provides mapping. Service trains WSP data for example LAWS see 7.6 Provides estimate of adhesion levels Readily available. Measures effect of adhesion level in a real wheel / rail contact. Operates in real time. LAWS provides mapping and statistics. Only provides data when the train is braking. Output is an estimate of adhesion level. Experience with using the software is necessary. Service / test train braking test see Part 8 Provides estimate of adhesion levels Demonstrates train performance. Very useful for incident investigations (for example SPADs or overruns) No instrumentation necessary to record data (but an accelerometer or OTMR output could be used to record / obtain train deceleration). Takes time to set up test during which adhesion conditions may change, (for example results may be significantly affected by moisture level). Train needs to be braked, and this impacts on service performance. Path required if using a non-service train. Adhesion management methods Adhesion management system Indirect measure Provides a network (route) relative adhesion status and enables identification of low adhesion sites. System under development. Not a direct measurement of adhesion level. Table 2 Advantages / limitations of low adhesion measurement and assessment methods (continued) Page 12 of 61 RAIL SAFETY AND STANDARDS BOARD

13 2.5 Recording conditions Table 3 is a matrix of the conditions that should be recorded for a range of purposes, the main reason being that these parameters can influence adhesion levels. Recording of the conditions at a site should generally be subjective. The information should be used to support the interpretation of results / findings from an investigation and monitor changes in conditions during tests or driver training. Parameter Purpose Reason Incident investigation Assess rail cleaning Check adhesion levels for test purposes Driver training Weather Affects wheel / rail adhesion levels Site features Affects wheel / rail adhesion levels Railhead contamination Affects wheel / rail adhesion levels Contaminant properties (Note: contaminant properties at incident sites are normally recorded by taking swabs). Affects test results or assessment Speed Adhesion / speed variation Adhesion measurement Selected method affects interpretation of adhesion conditions Table 3 Condition matrix An example table for recording low adhesion site conditions is given in Table 4. In general subjective assessments (see to ) should be made and the table should not take more than a few minutes to complete. The table should be completed at the start of the activity and if conditions change significantly (for example the onset of rain or a deliberate change of test parameters) a note should be made of the change Example record sheets are given in Appendix A for each purpose (that is incident investigation, driver training etc) The ambient weather conditions under which the measurements are taken should be recorded including: a) Wind speed and direction (a subjective assessment - light easterly breeze, strong south westerly gusts etc) b) Temperature (record only required for detailed studies) c) Humidity (record only required for detailed studies) d) Rainfall (a subjective assessment that is drizzle, light, moderate or heavy, showery etc) e) Other weather conditions (a subjective assessment that is sunny, dull, dry, mist or fog). RAIL SAFETY AND STANDARDS BOARD Page 13 of 61

14 2.5.5 Track formation at the measurement location should be recorded for example: a) Flat open track b) Shallow cutting c) Deep cutting d) Stepped formation - embankment one side cutting the other e) Embankment f) Gradient. Uncontrolled When Printed Other factors that may influence the adhesion levels should be recorded including: a) Vegetation for example large broad leafed trees with branches overhanging the track, open land with a few shrubs, hawthorn etc b) Line side factories, which are a potential for pollution c) Local lakes, which indicate that the area may be susceptible to mist and fog d) Proximity of airports, roads etc A subjective assessment of the condition of the railhead should be recorded for example: a) Amount of rust present for example none, low, moderate or high b) Presence of sand or sandite for example none, low, moderate or high c) Contamination for example leaf film none, low, moderate or high d) Moisture level for example dry, damp or wet e) Surface roughness if considered to be a relevant factor for example where the contaminant film thickness appears to be similar to the roughness The condition of contamination under which adhesion levels is measured should be recorded (see Part 10 for guidance) including the following: a) Hardness or shear torque of visible films with an area of 15 mm x 15 mm or greater b) Thickness (3 random sample measurements) c) Moisture content (3 random sample measurements) In the event that there is insufficient time available to take the measurements (as might arise during an incident investigation for example) then subjective notes can be made: a) Hardness - film removed / penetrated with finger nail, film difficult to scratch b) Thickness - film appears to be compressed to a very thin black layer, film appears to be relatively thick, multiple layers of leaves in green condition c) Moisture - record if the film is dry, damp or wet. Page 14 of 61 RAIL SAFETY AND STANDARDS BOARD

15 The measurement speed needs to be recorded so that, if required, the adhesion / speed relationship can be taken into account. If a portable tribometer is used at walking speed a best estimate of the speed may be used. The vehicle speedometer or timings between two fixed points may also be used to provide an estimate of speed A sufficient number of adhesion measurements should be taken in order to obtain a statistically significant sample In the following Parts 3 to 8 guidance is given on the application of each of the adhesion assessment and measurement options listed in Tables 1, 2 and Part 9 gives guidance on the adhesion / speed relationship which applies to all of the tribometer measurement options Part 10 gives guidance on the measurement of contaminantion properties to provide supporting evidence to the adhesion measurement. RAIL SAFETY AND STANDARDS BOARD Page 15 of 61

16 Location Purpose: Comments Parameter Time: Weather Sub parameter Wind speed Wind direction Temperature Humidity Sun Dry Dull Mist Fog Drizzle Rain Sleet Snow Site features Flat open Shallow cutting Deep cutting Embankment Stepped Gradient Vegetation Railhead contamination Factories Lakes River / stream Airport Busy road Rust Sand Sandite Leaf Grease Oil Other Dry Wet Damp Contaminant properties Speed Roughness Hardness Thickness Moisture Adhesion measurements Table 4 Example table for recording low adhesion site conditions (see Appendix A for examples) Page 16 of 61 RAIL SAFETY AND STANDARDS BOARD

17 Part 3 Train Mounted Methods 3.1 Scope 3.2 Purpose 3.3 Uses The methods described in this Part are intended to directly evaluate the available wheel / rail adhesion level that would be experienced by a train in service This includes measurement for each of the following parameters: a) Measurement vehicle speed and location b) Vertical force per axle or per wheel c) Lateral and / or longitudinal force per axle or per wheel This Part gives guidance on achieving accurate and repeatable adhesion measurements from a tribometer train or service train (including test trains formed from service vehicles). Measurements with a tribometer train could use load cells, load measuring wheels or axle mounted strain gauges. Measurements on service or test trains could be achieved by fitting axle mounted strain gauges or load measuring wheels These methods should only be used to obtain an accurate measurement of the adhesion level Data obtained from train mounted tribometers can be used for: 3.4 Tribometer train a) Assessing low adhesion treatment methods / rail cleaning using a tribometer train, load measuring wheels or modified service train for example fitted with axle mounted strain gauges. b) Measuring network adhesion status and using a tribometer train is possible. It would be an expensive method and it is difficult to obtain train paths. c) Identifying low adhesion sites using a tribometer train is possible. It would be an expensive method and it is difficult to obtain train paths. d) Checking adhesion levels for test purposes for example assessing brake systems using a tribometer train or load measuring wheels. e) Product testing for example checking weed killer solutions do not cause low adhesion conditions and braking problems Introduction to the tribometer train The measurements of longitudinal and vertical forces should be carried out with the suspension in its normal operating condition There are two options for including the vertical wheel / rail force in the calculation of the low adhesion levels: a) Use the static wheel load b) Measure the vertical wheel / rail force. RAIL SAFETY AND STANDARDS BOARD Page 17 of 61

18 If the approximation of the static wheel loads is not used, the test arrangement should be capable of determining the vertical wheel / rail forces. In most applications the use of the static wheel loads will be sufficient (particularly if the measurement speed is low that is less than 48 km/h (30 mile/h)) The vertical force is a combination of static and dynamic components. A measurement may be made of the resultant force or of only the dynamic component depending on the arrangement available for the measuring system. The resultant force should be determined to an accuracy of ± 5 kn irrespective of the measurement method Measurement might be achieved by use of strain gauges or a linear suspension displacement technique. Where a linear displacement technique is used, the vertical force is derived from the product of the known vertical stiffness and the displacement The test arrangement should be capable of determining the longitudinal wheel / rail force(s). If a load cell is used for this measurement and a low adhesion level is to be measured (µ down to 0.01) then it should be capable of reading to ± 0.1 kn Test conditions The data should be captured on tangent track in order to avoid flange contact which would affect the results When the measurement is taken the vehicle should be free from external forces other than the traction force. The traction force applied to the measuring vehicle should be maintained as near to constant as possible during the measurement The test vehicle brake application should be independent of the train brakes The brake force should be increased steadily until wheel slide is detected. Then the brake should be released to avoid wheel damage A WSP system that has optimised for low adhesion conditions should be used to detect wheel slide Any low adhesion mitigation system (for example sanders) should be disabled during the measurements provided that due consideration has been given to safety implications The test conditions set out in to should be recorded as appropriate for the parameters given in Table Measured values and evaluations The speed of the train should be recorded during the measurement Unless the static wheel load is being used for the calculation of the coefficient of friction, the vertical wheel forces (N) of both wheels of the axle should be measured. Evaluation should be done using the average of the arithmetic mean of the values for each wheel recorded during the measurements. If only the dynamic component of the vertical force is measured then the average from the two wheels should be added to the static load If the longitudinal wheel forces (F Long ) of both wheels of the axle are measured, the evaluation should be done using the average of the arithmetic mean of the values for each wheel recorded during the measurements The coefficient of friction (F Long / N) should be calculated from the directly measured mean longitudinal force or from the value calculated from the arithmetic mean of the values from each wheel. Page 18 of 61 RAIL SAFETY AND STANDARDS BOARD

19 3.4.4 Statistical sample A statistically significant set of measurements should be taken in to account for the variability along the track under nominally constant conditions The mean and standard deviation for the measurements obtained should be calculated for stated conditions, for example: An average adhesion level of 0.1 with standard deviation ± 0.05 on dry, leaf affected track, measured with a full-scale wheel at a measurement speed of 15 km/h Constraints and limitations At the time of issuing this document the original BR tribometer research train is obsolete. A new train would need to be built and equipped to undertake new measurements A train path is required when using tribometer trains Flange contact will affect the results Significant data processing and analysis is needed to determine the wheel / rail forces Advantages A tribometer train provides full-scale wheel / rail adhesion measurements and is therefore not subject to scaling effects Adhesion profiles can be acquired with this type of train measuring system. Each adhesion profile point is obtained from a full traction or braking curve showing the peak adhesion level The system could be built into rail head treatment trains. 3.5 Load measuring wheels Introduction to load measuring wheels Load measuring wheels can be fitted to a specially built tribometer vehicle or fitted under a service vehicle for test purposes An example of a load measuring wheel is shown in Figure 1. Figure 1 Load measuring wheel RAIL SAFETY AND STANDARDS BOARD Page 19 of 61

20 The wheels are built with strain gauges to measure the vertical, longitudinal and lateral wheel / rail forces normally by use of a wheel with spokes and gauges on each spoke The forces are reconstructed by use of software to combine the strain gauge measurements from each of the spokes Calibrating the strain gauges should be against known forces applied to the wheel The test arrangements should be capable of determining the longitudinal wheel / rail force(s) F Long. If a low adhesion level (µ down to 0.01) is to be measured then the longitudinal forces should be measured to ± 0.1 kn There are two options for including the vertical wheel / rail force in the calculation of the low adhesion level: a) Use of the static wheel loads b) Measurements of the vertical wheel / rail force If the approximation of the static wheel loads is not used, the test arrangement should be capable of determining the vertical wheel / rail forces. In most applications the use of the static wheel loads will be sufficient (particularly if the measurement speed is low that is less than 48 km/h (30 mile/h)) The vertical force is a combination of static and dynamic components. Measurements may be made of the resultant force or of only the dynamic component depending on the arrangement available for the measuring system. The resultant force should be determined to an accuracy of ± 5 kn irrespective of the measurement method Measured values and evaluations The speed of the train should be recorded during the measurements The vertical wheel forces (N) of both wheels of the axle should be measured. Evaluation should be done using the average of the arithmetic mean of the values for each wheel recorded during the measurements If the longitudinal wheel forces (F Long ) of both wheels of the axle are measured, the evaluation should be done using the average of the arithmetic mean of the values for each wheel recorded during the measurements The coefficient of friction (F Long ) can be calculated for each wheel / rail contact from the ratio of the measured longitudinal force to the vertical force (either static or measured force) If a coefficient of friction (F Long / N) is required for the wheelset then it should be calculated from the arithmetic mean of the values from each wheel Statistical sample A statistically significant set of measurements should be taken in to account for the variability along the track under nominally constant conditions The mean and standard deviation for the measurements obtained should be calculated for stated conditions, for example: An average adhesion level of 0.1 with standard deviation ± 0.05 on dry, leaf affected track, from longitudinal forces measured with a full-scale load measuring wheels at a measurement speed of 15 km/h. Page 20 of 61 RAIL SAFETY AND STANDARDS BOARD

21 3.6 Load measuring wheels - stick-slip oscillations Introduction to Stick slip oscillations The capability of changing the angle of attack between the load measuring wheels and rails is required for this measurement. A record of wheel / rail lateral force with time will show stick-slip oscillations (Figure 2, Note the data in this graph is only indicative and should not be used for purposes other than illustration) Force measurements from the wheel whose angle of attack is towards the rail in the direction of travel should not be used for calculation of the coefficient of friction since flange contact may occur and affect the magnitude of the lateral force. Dry rail with 75 kn vertical load and 20 m Rads angle of attack Peak Traction Coefficient = Breakout Frction Traction coefficient (T/N) Slip phase Stick phase Time (Seconds) Figure 2 Wheel / rail lateral force stick-slip oscillations (hypothetical data) The coefficient of friction (F Lat /N) should be calculated from the peak lateral force prior to the slip phase Statistical sample A statistically significant set of measurements should be taken in to account for the variability along the track under nominally constant conditions The mean and standard deviation for the measurements obtained should be calculated for stated conditions, for example: An average adhesion level of 0.1 with standard deviation ± 0.05 on dry, leaf affected track, from lateral forces measured with a full-scale load measuring wheels at a measurement speed of 15 km/h Constraints and limitations Load measuring wheels are not readily available and are complex and expensive to build Compensation is required for heat input to the strain gauges from braking the wheel(s) A train path is required when using load measuring wheels Flange contact will affect the results. RAIL SAFETY AND STANDARDS BOARD Page 21 of 61

22 Significant data processing and analysis is required to determine the wheel / rail forces Advantages Load measuring wheels provide full-scale wheel / rail adhesion measurements and are therefore not subject to scaling effects A full traction and / or braking curve can be measured showing the peak adhesion level Adhesion profiles can be acquired with a train measuring system Load measuring wheels could be incorporated into a rail head treatment train or be fitted to a service vehicle to be used for test purposes. 3.7 Strain gauged axles Introduction to strain gauge axles The application of strain gauges for measuring forces generated in the axle, from force variations in the wheel / rail contact, is in the process of being investigated and might become available for measuring wheel / rail adhesion level. The guidance note will be updated at an appropriate time to include this method should it become available Constraints and limitations Strain gauged axles are not readily available, they are still under development Strain gauged axles require fitting to a train where the trains need to operate in a harsh environment A train path is required if using a non-service train Flange contact will affect the results Strain gauge methods require significant data processing and analysis to determine the wheel / rail forces Advantages Strain gauged axles record forces for full-scale wheel / rail conditions and are therefore not subject to scaling effects Adhesion profiles could be acquired with this type of train measuring system They could be incorporated into a rail head treatment train or potentially be fitted to service trains. Page 22 of 61 RAIL SAFETY AND STANDARDS BOARD

23 Part 4 Vehicle Borne Tribometer 4.1 Scope 4.2 Purpose 4.3 Uses The methods described in this Part are intended to provide a measure of the available wheel / rail adhesion level from a vehicle borne tribometer These methods should only be used to provide a direct assessment of relative adhesion levels from a rail mounted vehicle Data obtained from a vehicle borne tribometer can be used for: a) Assessing low adhesion treatment methods / rail cleaning b) Identifying low adhesion sites (subject to track access being available) c) Checking adhesion levels for test purposes for example testing brake systems d) Product testing for example checking weed killer solutions do not cause low adhesion conditions and braking problems. 4.4 Vehicle borne tribometer method Introduction to the vehicle borne tribometer Vehicle borne tribometers measure significant lengths of track, which is achieved by use of a road / rail vehicle to propel the tribometer at speeds up to 48 km/h (30 mile/h). An example of this is the Portec (formerly Salient Systems) High Speed Tribometer (Figure 3). This type of tribometer is based on similar principles to the scaled braked wheel portable tribometer. Figure 3 Vehicle borne tribometer The vehicle borne tribometer permits measurement of both rails and the friction on the gauge face as well as the railhead (four simultaneous measurements) The vehicle borne tribometer should be calibrated and operated according to the manufacturer s instructions The test conditions set out in to should be recorded as appropriate for the parameters given in Table Friction measurements can be made over a significant length of track (up to 200 miles can be measured per day according to the suppliers, based on usage in the USA). Output from the device gives a friction profile (Figure 4) for the route or test section. RAIL SAFETY AND STANDARDS BOARD Page 23 of 61

24 4.4.2 Statistical sample Uncontrolled When Printed A statistically significant set of measurements should be taken in to account for the variability along the track under nominally constant conditions. 0.6 High Speed Tribometer track adhesion profile for low adhesion mitigation Coefficient of friction Treatment Section Distance miles Before treatment After treatment Figure 4 Friction profile for mitigation testing (hypothetical data) Figure 4 shows how friction profiles could be superimposed to demonstrate the effectiveness (or otherwise) of low adhesion treatment methods. (Note, the data in this graph is only indicative and should not be used for purposes other than illustration) Constraints and limitations A track possession is required for the measurements to be obtained There will be some gauging issues and the vehicle borne tribometer would not be useable on third and fourth rail electrified lines (but the manufacturers might produce a customised version to overcome electrical issues) The vehicle borne tribometer still uses small-scale wheels (but larger than the portable tribometer) so some scaling effects remain compared to the full-scale railway vehicle wheel / rail conditions The vehicle borne tribometer is more expensive than portable tribometer Advantages Vehicle borne tribometers are available, easy to use and both rails can be measured simultaneously. The wheel position across the rail can be adjusted The measurement is with a braked wheel and the result is the peak adhesion levels. Significant lengths of track are measurable The vehicle borne tribometer is more accurate than the portable tribometers and can be used up to speeds of 48 km/h (30 mile/h). Page 24 of 61 RAIL SAFETY AND STANDARDS BOARD

25 Part 5 Portable Tribometer Methods 5.1 Scope 5.2 Purpose 5.3 Uses The methods described in this Part are intended to evaluate the available wheel / rail adhesion levels using scaled portable tribometers This includes measurement using one of the following techniques: a) Braked wheel tribometer b) Skid resistance slider tribometer c) Static breakout friction measurement using a slider This Part gives guidance on the use of the above portable tribometers with advice on the limitations relating to the measurements obtained from each tribometer This methods should only be used to obtain a relative measure of adhesion levels because of the scaling factor compared to full size tribometers Data obtained from the portable tribometer can be used for: a) Incident investigation 5.4 Portable tribometers b) Assessing low adhesion treatment methods / rail cleaning c) Track testing purposes for example testing of braking systems etc d) Checking adhesion levels for driver training e) Laboratory measurement of adhesion levels for research purposes (skid resistance and static breakout tribometers only) f) Product testing for example checking weed killer solutions do not cause low adhesion conditions and braking problems When taking adhesion measurements it is recommended that: a) Reference measurements should be made on conditions of a control section of clean dry or clean wet rail. b) Sufficient measurements (a minimum of 10 on each rail) should be taken for each railhead condition. c) Measurements should be taken over a sufficient length of track to be representative of the site and the site should be suitable for the purpose for which the measurements are being taken. d) If repeat measurements are required on the same length of track, then the second series of measurements should be obtained on a new contact point. e) If investigating a possible low adhesion incident the adhesion levels should be measured in the as found condition and then repeat measurements made with the rail lightly dampened since moisture activates low adhesion conditions. RAIL SAFETY AND STANDARDS BOARD Page 25 of 61

26 f) The measuring wheel should be cleaned before making measurements on a new site to avoid contamination from previous measurements affecting the results. g) The test conditions as specified in to should be recorded. h) Repeat a series of measurements on the control section and compare with the initial control section measurements to monitor if conditions have remained constant during the measurement period. i) Avoid stepping onto the rail surface that is to be measured. 5.5 Braked wheel tribometer Introduction to the braked wheel tribometer An example of a braked wheel tribometer is shown in Figure 5. Currently this type of tribometer is supplied only by Portec (formerly Salient Systems). Figure 5 Braked wheel tribometer The measuring wheel can be orientated to obtain results from the gauge corner or from a number of positions across the railhead There are some general actions that should be observed when using the braked wheel tribometer: a) The method for using the braked wheel tribometer is set out in the user s manual provided by the supplier. Follow the instructions for set-up and the calibration procedure. b) Ensure that the guide wheels are properly located against the rails. c) Check that the tribometer does not foul any part of the rail for example welds. d) Prevent the measuring wheel fouling the braked wheel tribometers magnetic clutch meters. e) Ensure the measuring wheel is clean, avoid contamination, do not touch the measuring surface with fingers. Page 26 of 61 RAIL SAFETY AND STANDARDS BOARD

27 5.5.2 Test conditions When taking adhesion measurements it is recommended that the procedure outlined in 5.4 should be followed Measured values and evaluations It is generally accepted that the adhesion values obtained from the braked wheel tribometer are high compared to those obtained from a full-scale tribometer. Also, the type of contamination on the railhead affects the magnitude of the results. This means that there is no single scaling factor that can be applied to the results. Table 5 gives guidance on interpretation of the portable tribometer readings relative to adhesion levels that might be expected from service trains Where measurements are measured on wet rail the average value is marginally less than that measured on wet leaf. However, the minimum value on wet leaf is much less than that on wet rail. The scatter of results should therefore be taken into account when evaluating results Calibration measurements were made on simulated leaf (paper tape) in dry and damp conditions when the portable tribometer was calibrated against the tribometer train because real leaf film was not available. Tribometer train adhesion measurements on real dry leaf film along with artificially dampened (water mist sprays from the vehicle) real leaf film were investigated and the results are provided in Table 5. Rail condition Train adhesion values Portable tribometer values Maximum Minimum Average Maximum Minimum Average Clean dry Clean wet Dry simulated leaf Dry leaf Wet simulated leaf Dampened Leaf Oil Oil and water Table 5 Comparison of adhesion values: portable tribometer and train Where the control conditions have changed, compensate the assessment measurements accordingly as follows: a) Plot the measurements in time sequence / measurement order and determine the average value for the initial control measurements. Determine the increase in control section adhesion value relative to the initial control average for each test measurement (for example at point A on Figure 6, note the data in this graph is indicative and should not be used other than for illustrative purposes). Calculate the ratio of the test measurement to the final control measurement and multiply by the increase in control adhesion value. b) Increase in control adhesion value = µ c c) Ratio of test adhesion value to final control adhesion value = µ T / µ FC d) Compensation = µ T ± µc x µ T / µ FC where the sign of the compensation is dependent upon whether the control adhesion value increased (+) or decreased (-). RAIL SAFETY AND STANDARDS BOARD Page 27 of 61

28 Compensation for changing conditions Initial control section measurements Average initial control section adhesion level Change in control adhesion for point A Coefficient of friction A Time Control section Test or incident section Linear (Control section) Linear (Test or incident section) Figure 6 Compensation for changing adhesion conditions (hypothetical data) Not all conditions will require compensation for example grease and oil contamination may not be greatly influenced by changes in weather conditions or even the passing of train wheels. A subjective assessment should be made of the conditions and contamination at the time of the measurements to determine if compensation is warranted Statistical sample A statistically significant set of measurements should be taken to account for the variability along the track under nominally constant conditions. For example, at an incident site a 200 m length of track could be divided into ten, 20 m zones and between five and ten measurements taken on both rails in each zone The mean and standard deviation for the measurements obtained should be calculated for stated conditions, for example: An average adhesion level of 0.1 with standard deviation ± 0.05 on dry, leaf affected track, measured with a 100 mm braked wheel tribometer at a measurement speed of 5 km/h Constraints and limitations The adhesion values obtained from portable tribometers or braked wheel tribometers are high and are influenced by rail contamination. There are no simple scaling factors that can be applied to the results The operation of the braked wheel tribometer has on an isolated occasion been affected by electric traction taking power from the overhead line. The effect was to switch the device off Operated at low speed measurement, such as walking pace Advantages Braked wheel tribometers are readily available, easy to use and are portable Recording time is approximately 15 to 25 seconds per single measurement of adhesion level once the braked wheel tribometer is on the track. Page 28 of 61 RAIL SAFETY AND STANDARDS BOARD

29 The braked wheel tribometer can be adjusted to measure at different positions across the rail The measurement is with a braked wheel tribometer and the result is the peak adhesion level (subject to scaling effects as described in ). 5.6 Static breakout friction tribometer Introduction to static breakout friction tribometer using a slider This method normally uses a small hand held tribometer as shown in Figure 7. This type of tribometer has a flat disc shaped chromium slider. This type of tribometer should only be used for a very coarse assessment of friction conditions and one example is supplied by Kett. Figure 7 Kett / Heidon hand held static breakout friction tribometer There are some general actions that should be observed when using this type of tribometer: a) The method for using this type of tribometer is set out in the user s manual provided by the supplier. Follow the instructions for the set-up procedure. b) Ensure that the slider is clean to avoid contamination and the counter value is set to zero, do not touch the measuring surface (that is to avoid grease from fingers). c) Only measure on the central portion of the railhead (± 20 mm) to avoid further inaccuracy caused by tilting the tribometer. d) Remove any excess viscous lubricant (grease or thick detergent etc) that might be contaminating the railhead, the tribometer will read high if there are significant amounts of lubricant because of viscous drag forces on the sides of the slider (under normal conditions a wheel will squeeze most of the excess grease from the contact). e) Record the individual readings rather than the average readings from the tribometer so that an assessment of the consistency of the results can be made during the measurements, the average can be subsequently calculated. f) Keep the tribometer dry, shelter from rain and keep the base facing downwards in wet conditions to avoid moisture entering the tribometer Test conditions When taking adhesion measurements it is recommended that the procedure outlined in 5.4 should be followed. Where recommendations to clean or adjust the position of the wheel are made this will apply to the slider in the case of the Kett tribometer Measured values and evaluations Friction coefficients (µ) measured by the BR research tribometer train (Figure 8), were found on the mainline to have a mean value of 0.23 and a standard deviation (σ) of 0.05 (see Table 6). RAIL SAFETY AND STANDARDS BOARD Page 29 of 61

30 Figure 8 BR research tribometer measuring wheelset (obsolete) Friction coefficient (µ) Normal distribution (up to % of population) < % < % < % < % < % Table 6 Distribution of friction coefficients Assuming a normal distribution of the above friction coefficients; Kett values of µ below 0.18 would indicate low adhesion conditions, 0.18 to 0.28 normal adhesion levels and above 0.28 high adhesion Using the above figures for guidance, the control section will be equivalent to a friction coefficient greater than Compare the Kett measured results with those for the control clean dry section to assess the friction level. Under lubricated conditions the results can be artificially high with this tribometer Statistical sample A statistically significant set of measurements should be taken to account for the variability along the track under nominally constant conditions. For example, at an incident site a 200 m length of track could be divided into ten 20 m zones and between five and ten measurements taken on both rails in each zone The mean and standard deviation for the measurements obtained should be calculated for stated conditions, for example: An average adhesion level of 0.1 with standard deviation ± 0.05 on dry, leaf affected track, measured from static breakout using a tribometer with a chromium slider Alternative option The resolution obtained from this type of tribometer can be improved by replacing the chromium slider with a rubber slider. Even with a rubber slider the resolution of the Kett type tribometer is significantly less than with a Munro pendulum tribometer (Figure 9) The results obtained from the hand held static breakout friction tribometer are very dependent upon the properties of the contamination. Figure 9 gives guidance on the values that can be expected for some common railhead conditions when using a rubber slider. Page 30 of 61 RAIL SAFETY AND STANDARDS BOARD

31 Comparison of friction measuring devices Kett tribometer with rubber slider Kett friction level Munro skid resistance Kett Static breakout friction tribometer Munro Skid resistance slider tribometer (see 5.7) 0.00 Dry Wet Detergent Dr Lignin y / floc We Lignin t / floc 0 Rail contamination condition Kett tribometer Munro skid resistance Figure 9 Sample adhesion measurements between Kett and Munro tribometers Before using this type of tribometer in the field with a rubber slider, a series of calibration measurements should be carried out on common contaminants in the laboratory to produce a table of expected results for that particular tribometer (see Figure 9 for an indication of adhesion values with a rubber slider on several contaminants). A note should also be made of the surface roughness of the test plate and slider during this calibration, since roughness affects results from this type of tribometer. Calibration contaminants that should be used are: a) None (for example a clean dry surface) b) Wet clean surface c) Detergent (for example a typical high quality detergent for dishwashing) d) Standard lubricator grease (only use a thin film since excess grease will affect the results) Other contaminants that could be used to calibrate the device are: e) Dry simulated leaf film for example lignin f) Damp simulated leaf film for example lignin g) Any other as required for a specific application Note that between each contaminant sample the slider should be cleaned with a degreaser to prevent cross-contamination affecting the calibration For field measurements, the procedure as described in 5.4, should be used in combination with the above calibration Constraints and limitations The hand held static breakout friction tribometer provides for a very coarse assessment of sliding friction only The flat chromium slider will not be in full contact with the curved rail making it difficult to maintain a steady condition Surface roughness of the slider and lubrication affect the results and the scaling effects The resolution of results on different contaminants is poor when using the chromium slider - a rubber slider potentially offers an improvement in resolution. RAIL SAFETY AND STANDARDS BOARD Page 31 of 61

32 5.6.7 Advantages The hand held static breakout friction tribometer is readily available and very easy to use The measuring time is short at 5 seconds The tribometer can be used on very short sections of rail. 5.7 Skid resistance slider tribometer Introduction to skid resistance slider tribometer measurement An example of a skid resistance slider tribometer is a pendulum tribometer as shown in Figure 10. Munro is one supplier of this type of tribometer which is based on a pendulum arm with rubber skid. Figure 10 Pendulum tribometer for skid resistance measurement The pendulum tribometer is designed for use in the road industry and requires a flat surface on which to operate and is therefore more suited to laboratory investigations. Field measurements on the railway could be obtained if the pendulum tribometer was modified or a measuring platform could be provided to fit over the rail There are some general actions that should be observed when using the pendulum tribometer: a) The method for using the pendulum tribometer is set out in the user s manual provided by the supplier. Follow the instructions for the set-up procedure. b) Ensure that the pendulum is not fouled during the swing whilst measuring. c) Check that the pendulum counter is set to zero. d) Ensure that the pendulum does not alter the peak displacement value on its return swing. e) Ensure the slider is clean and avoid contaminating, do not touch the measuring surface (that is to avoid grease from fingers) Test conditions When taking adhesion measurements it is recommended that the procedure outlined in 5.4 is followed. Where recommendations to clean or adjust the position of the wheel are made this will apply to the skid resistance slider in the case of the pendulum tribometer. Page 32 of 61 RAIL SAFETY AND STANDARDS BOARD

33 5.7.3 Measured values and evaluations The measured skid resistance values should be compared to adhesion levels measured with the BR research tribometer train (Figure 11). An estimate of the wheel / rail adhesion level µ, can be obtained from a linear fit between skid resistance R, and adhesion level at the 90% confidence level. Calibration of skid resistance against wheel / rail adhesion levels y = x R 2 = Skid resistance BR research tribometer train adhesion levels Rubber skid Linear (rubber skid) Figure 11 Relationship between skid resistance and wheel / rail adhesion levels Where the control conditions have changed, these should be taken account of in the assessment of the measurements as described in and Statistical sample A statistically significant set of measurements should be taken to account for the variability along the track under nominally constant conditions. For example, at an incident site a 200 m length of track could be divided into ten 20 m zones and between five and ten measurements taken on both rails in each zone The mean and standard deviation for the measurements obtained should be calculated for stated conditions, for example: An average adhesion level of 0.1 with standard deviation ± 0.05 on dry, leaf affected track, measured with a pendulum type tribometer with rubber slider Constraints and limitations The design of the pendulum tribometer is not suitable for use on the track without provision of a frame for standing the device on during measurements The pendulum measures skid resistance rather than a rolling / sliding peak of adhesion levels. The value of skid resistance will be subject to scaling effects The pendulum tribometer has slightly more complex set up procedures than other portable tribometers Advantages The pendulum tribometer is readily available, portable and relatively easy to use The pendulum tribometer is known to give significant resolution between the skid resistance values for a range of contaminants. RAIL SAFETY AND STANDARDS BOARD Page 33 of 61

34 Part 6 Laboratory Test Rigs 6.1 Scope 6.2 Purpose 6.3 Uses The methods described in this Part are intended to provide a measure of friction using scaled or full sized laboratory test rigs To provide a facility for measuring friction in a controlled environment and to enable accelerated testing (that is many wheel passes over a single point on the rail within a short timescale) Laboratory test rigs may be used for: 6.4 Twin disc test rig a) Assessing low adhesion treatment methods / rail cleaning b) Laboratory investigation of fundamental low adhesion factors c) Product testing for example checking weed killer solutions do not cause low adhesion conditions and braking problems Introduction to the twin disc test rig An example of a laboratory twin disc test rig is shown in Figure 12 and Figure 13. The most common laboratory test rig is supplied by Amsler. Figure 12 Amsler twin disc machine Figure 13 Amsler Twin discs (40 mm diameter) Page 34 of 61 RAIL SAFETY AND STANDARDS BOARD

35 The load applied to the discs should be adjusted so that the peak contact pressure q o, is the same as for a wheel / rail contact. This is given by: Load / unit contact width = π 2 q o 2 R t R b / (E (R t + R b )) Where R t = radius of the top disc, R b = radius of the bottom disc and E is the modulus of elasticity for the disc steel The disc diameter should be selected to generate the required percentage creep: Percentage creep, c = 100 * (1.104R b - R t ) / (1.104R b ) % As an example assume that the top disc has a diameter of 40 mm in all cases then the diameter of the bottom disc can be calculated for a range of creep values from: Diameter of bottom disc = (2R t ) / (1.104 * (1-c / 100)) Table 7 gives the results of these calculations. Percentage creep Diameter of bottom disc mm Table 7 Bottom disc diameters at percentage creep levels In practice the diameter of the discs will vary because of wear (particularly at high creep values, 5% or greater), measurements should be taken of the diameter at appropriate points during the tests and the percentage creep levels calculated To determine the peak adhesion levels a series of measurements of friction should be acquired for a range of percentage creep levels as shown in Figure 14. (Note the data in this graph is only indicative and should not be used for purposes other than illustration) Coefficient of friction Peak adhesion Creep percentage levels Figure 14 Amsler percentage creep levels showing peak adhesion levels (hypothetical data) RAIL SAFETY AND STANDARDS BOARD Page 35 of 61

36 6.4.2 Constraints and limitations Contact between discs is much smaller than full-scale wheel / rail contact and is therefore subject to scaling effects. Twin disc machines are more suited to testing fluid type friction modifiers or weed killer solution than particulate friction modifiers Wear of the discs can be high and change the percentage creep levels New or re-turned discs must be used for each percentage creep level to be tested Advantages The Amsler twin disc test rig is a standard type of machine for friction measurement that has been widely used for wheel / rail research. The Amsler twin disc test rig is readily available in university research and consultancy laboratories Test conditions can be controlled Multiple tests can be easily conducted to determine statistical variation Contaminants can be easily introduced into the contact. 6.5 Pin-on-disc machine Introduction to the pin-on-disc friction and wear machine Pin-on-disc machines (Figure 15) are commonly used in friction and wear research and are designed to provide a measure of the friction between a pin loaded against a rotating disc. Plint is one supplier of this type of machine Universities / consultancy firms offer the use of this type of facility. Figure 15 Pin-on-disc friction and wear machine Rail steel should be used for the manufacture of a disc for use in this type of machine whilst the pin should be produced from wheel steel The maximum normal load for a Plint pin-on-disc machine is 1 kn and should be used for wheel / rail contact test purposes with a pin diameter of 5.5 mm. Page 36 of 61 RAIL SAFETY AND STANDARDS BOARD

37 6.5.2 Constraints and limitations The pin-on-disc machine records friction under sliding conditions The maximum contact pressure is a factor of 10 less than the maximum pressure in a wheel / rail contact Maximum sliding speed for the Plint pin-on-disc machine is 4 m/s (9 mile/h) The contact area is a factor of 8 less than the wheel / rail contact hence there will be scaling effects. The pin-on-disc machine is more suited to testing fluid type friction modifiers or weed killer solutions than particulate friction modifiers Advantages Conditions can be controlled that is the environment and loads Contaminants can be easily introduced into the contact Repeat measurements can be made in a short period of time. 6.6 Full-scale wheel-on-rail rig Introduction to the full-scale wheel-on-rail rig The test rig shown in Figure 16 is designed so that traction or braking forces can be applied and the wheel can be rotated to produce an angle of attack relative to the rail. This design allows two methods of measuring wheel / rail friction: a) Forcing the wheel into lateral stick-slip oscillations through application of a large angle of attack (10 to 30 milliradians) b) Application of tractive force to induce wheel spin c) Application of a braking force to induce wheel slide A vertical load is applied equivalent to a rail vehicle wheel / rail load (for example 75 kn is typical of a modern diesel multiple unit). Figure 16 Full-scale wheel-on-rail rig RAIL SAFETY AND STANDARDS BOARD Page 37 of 61

38 If lateral stick-slip oscillations are used to measure friction then the peak force at the point of slip is used to calculate the coefficient of friction (peak adhesion level) by simply dividing by the static wheel / rail contact load Each stick slip oscillation (see Figure 2 and ) can be used to obtain a value for the coefficient of friction and these can then be averaged If traction or braking forces are applied then the force at which the wheel spins (traction) or slides (braking) is used to calculate the coefficient of friction by simply dividing by the static wheel / rail contact load Constraints and limitations There are very few of full-scale wheel-on-rail rigs available Full-scale wheel-on-rail test rigs measurements are at low speed (less than 5 mile/h) Advantages Conditions can be controlled that is the environment and loads The contact is full-scale Contaminants can be easily introduced into the contact Repeat measurements can be made in a short period of time. Page 38 of 61 RAIL SAFETY AND STANDARDS BOARD

39 Part 7 Service Train 7.1 Scope 7.2 Purpose 7.3 Uses The methods described in this Part are intended to provide an indirect evaluation of the available wheel / rail adhesion level from service train on-board equipment This includes adhesion level assessment using data from OTMR, TMS or WSP systems This Part gives guidance on the use of data from the above sources. The data could be obtained from a train that is in normal service or from a train that has been taken out of service and used in trials OTMR, WSP and TMS provide an indirect assessment of adhesion levels from service trains The uses for data from OTMR, WSP and TMS include: 7.4 OTMR data a) Incident investigations b) Assessing low adhesion treatment methods / rail cleaning c) Determination of the network adhesion status d) Identification of low adhesion locations e) Assessing low adhesion conditions for test purposes, for example brake system test f) Product testing for example checking weed killer solutions do not cause low adhesion conditions and braking problems Introduction to OTMR data OTMR data is regularly used by the industry to investigate incidents. The speed-time output (Figure 17) can be obtained from the OTMR data and analysed to determine brake rate and distances and times between events. The achieved brake rates can also be used to provide an indication of average adhesion level over the stopping distance. RAIL SAFETY AND STANDARDS BOARD Page 39 of 61

40 OTMR Speed-time data m AWS 1064 m Emergency brake sanding 1328 m Train stopped 30 Braking rate = 0.4 m/s^2 25 Speed (m/s) Braking period brake steps 1, 2 and 3 used. 10 Braking rate = 0.92 m/s^ Time (seconds) Figure 17 OTMR speed-time output In low adhesion conditions the speed signal can become corrupted by sliding and thus produce a low reading. On the graph of Figure 17, the speed appears to reduce on two occasions (minimum at 50 seconds and again at 57 seconds). An estimate of the braking rate can be obtained from the slope of a line tangent to the peak speed recorded during the period 40 to 60 seconds. The speed after 40 seconds was 31.5 m/s falling to 23.3 m/s after 60 seconds. This gives a braking rate of: ( ) / (60-40) = 0.41 m/s The estimated adhesion level is simply the brake rate divided by the acceleration due to gravity in units of m/s 2. In the example above the adhesion level is therefore 0.41 / 9.81 = and indicates a very low level. In carrying out the calculation care should be taken to use consistent units and in the example metric units have been used that is metres and seconds Account should be taken of any track gradient on the measured value so that level track can be used as the reference condition for the results. For example, if the output was from an incident on a 1 : x gradient then the deceleration due to the gradient is 9.81 / x m/s 2. Hence if the measurements are made on a rising gradient the achieved deceleration on level track will be the measured deceleration plus 9.81 / x and if on a falling gradient the measured deceleration minus 9.81 / x The improved brake rate (between approximately 60 to 80 seconds) of 0.92 m/s 2 indicates that the sander was operating and was effective in mitigating the low adhesion conditions (Figure 17) In addition to the direct OTMR data, supporting information can be obtained from some modern stock (for example class 375) by downloading from the brake control unit (BCU). This provides brake pressure, and brake demand. Speed data and variations in the brake pressure give an indication of the adhesion level. The BCU information is stored on an as used basis and not elapsed time so some effort is required to match the information to the OTMR data. The quantity of data from the BCU is also limited by the storage capacity of the system. Page 40 of 61 RAIL SAFETY AND STANDARDS BOARD

41 7.4.2 Constraints and limitations OTMR provides an indirect assessment of the adhesion level through achieved deceleration or acceleration data. This information has to be obtained by a search and analysis process OTMR data normally must be downloaded manually The system only provides data when the train is braking or accelerating Experience of analysing the data is necessary Advantages Readily available since all trains are required to have OTMR fitted. 7.5 TMS The information is for full-scale wheel / rail contact, that is no scaling effects The railway industry is familiar with the use of data from OTMR outputs since they are commonly used in incident analysis Introduction to TMS Modern stock can provide adhesion information via the TMS, an example of this is the Bombardier Mitrac system. This type of system can relay real time data (including positional information) from the train to a central location. The data includes slip and slide events and a train speed profile from which deceleration can be calculated. Mitrac has been developed as a general diagnostic tool but provides similar data to OTMRs There is also a mapping facility available from Mitrac (Figure 18) that allows events to be shown on a route map. Figure 18 Mitrac map identifying events on a route RAIL SAFETY AND STANDARDS BOARD Page 41 of 61

42 An estimate of adhesion conditions can be obtained from Mitrac by following the procedure described in to Constraints and limitations Mitrac only provides an indirect assessment of the adhesion level through achieved acceleration or deceleration data This information has to be obtained by a search and analysis process. Experience of using the associated software is necessary The system only provides data when the train is braking or accelerating Advantages Mitrac is available and provides information for the full-scale wheel / rail contact that has no scaling effects The system provides mapping of slip and slide events Mitrac can provide real time data. 7.6 WSP data Introduction to WSP data LAWS TM is propriety equipment developed for use on service vehicles. By taking output from existing WSP systems it reports to a central database the occurrence and location of wheel slide and spin events caused by low adhesion conditions. The adhesion status of the network is then displayed as a map on a control room monitor (Figure 19) The colour coded arrows on the map indicate the adhesion conditions at specific locations, where green represents acceptable adhesion conditions and red represents low adhesion conditions. Yellow and orange indicators represent intermediate levels of adhesion. A change from green to yellow indicates the onset of low adhesion conditions The LAWS indicators can be used to assist in the deployment of treatment trains or track teams to target low adhesion sites. Figure 19 LAWS low adhesion network map Page 42 of 61 RAIL SAFETY AND STANDARDS BOARD

43 Since the data is being updated in real time the information can be used to inform drivers of changes in adhesion conditions (either improvement or deterioration) LAWS holds the collected information in a database so that statistical reviews can be made to aid adhesion management for example vegetation control. The inset graph in Figure 19 shows the variation of WSP activity on the down line on the Hounslow loop at 7 miles 53 chains for the period between early April and late May Constraints and limitations LAWS only provides an indirect assessment of the adhesion level through achieved acceleration or deceleration data The information has to be obtained by a search and analysis process. Experience of using the associated software is necessary The system only provides data when the train is braking or accelerating Advantages LAWS is readily available and provides information for the full-scale wheel / rail contact that is no scaling effects The system provides mapping and statistics Real time data is provided by the system. RAIL SAFETY AND STANDARDS BOARD Page 43 of 61

44 Part 8 Service / Test Train Braking Test 8.1 Scope 8.2 Purpose 8.3 Uses The methods described in this Part are intended to evaluate the available wheel / rail adhesion level from the achieved braking rate or stopping distance of a service or test train This includes adhesion assessment using one of the following options: a) Measured train deceleration b) Timing to stop from a known speed at brake application c) Measured stopping distance This Part gives guidance on the use of data from the above sources. The data could be obtained from trains that are in normal service or the train could be taken out of service and used in trials Braking tests provide an indirect assessment of adhesion levels from service or test trains There are a range of uses for data from braking tests including investigation of: 8.4 Train braking method a) Incident investigations and checking that the adhesion level has been restored after an incident. b) Assessing low adhesion treatment methods / rail cleaning. c) Testing purposes (for example assessing brake systems (to check blending), WSP, magnetic track brakes etc). d) Check adhesion levels for driver training. e) Product testing for example checking weed killer solutions do not cause low adhesion conditions and braking problems Introduction to train braking method When using brake tests to assess low adhesion levels the maximum stopping distance needs to be established before carrying out the brake test so that due allowance can be made for safety. In very low adhesion conditions it is normally assumed that there is a worst case deceleration of 1.5% g and 10% error on the speedometer. The distance to stop from v m/s (+ 10%), assuming a 2.5 second brake build up time will be: Distance, d = (v x 2.5 / 2) + ((v 2 - u 2 ) / 2ä) Where u = 0 m/s (train comes to a stand), v is the speed at the point of brake application and ä is the deceleration m/s An accelerometer capable of measuring ± 1.0 g with a resolution of 0.05 g can be used to measure the longitudinal deceleration of the train. The accelerometer should be mounted on a rigid structure on the floor of the test vehicle to reduce any vehicle body dynamic accelerations / decelerations in the measurement. Page 44 of 61 RAIL SAFETY AND STANDARDS BOARD

45 Account should be taken of any track gradient on the measured value so that level track can be used as the reference condition for the results. For example, if the tests are on 1 : x gradient then the deceleration due to the gradient is 9.81 / x m/s 2. Hence if the measurements are made on a rising gradient the achieved deceleration on level track will be the measured deceleration plus 9.81 / x and if on a falling gradient the measured deceleration minus 9.81 / x Time to stop from a known speed can be used as an alternative to the measurement of deceleration provided that there is a constant brake application. The deceleration is calculated from v / t where v is the speed in m/s at the braking point and t the time in seconds to stop A pedometer can be used to measure the distance to stop in a braking test. The deceleration is calculated from v 2 / d where v is the speed in m/s at the point of brake application and d is the distance to stop in m For any of the three methods of measuring deceleration, the estimated adhesion level is simply the brake rate divided by the acceleration due to gravity in units of m/s Figure 20 shows a comparison of the three methods of measuring deceleration and indicates that each of the methods will give a similar estimate of the adhesion level. 3 Measurement of Train Deceleration 2.5 Deceleration (%g) Deceleration from time to stop (%g) Deceleration from distance to stop (%g) Longitudinal Deceleration (%g) Test Number Figure 20 Measurement of train deceleration for estimating adhesion levels Where the above braking test measurements are used in a controlled assessment of a mitigation method a consistent low adhesion level should be established before the mitigation is applied The achieved deceleration should be recorded for each test run and may be plotted as shown in Figure 21. In this example the mitigation method restores the adhesion levels to about the pre-test level The test conditions set out in to should be recorded as appropriate for the parameters given in Table 3. RAIL SAFETY AND STANDARDS BOARD Page 45 of 61

46 Assessment of mitigation 6 Deceleration (%g) Mitigation applied Deceleration from time to stop (%g) Longitudinal deceleration (%g) 1 Low adhesion 0.5 condition Test number Figure 21 Measurements for assessing a low adhesion mitigation method (hypothetical data) Constraints and limitations The service / test train brake takes time to set up the test during which the adhesion conditions may change The test train needs to be braked, so will impact upon service performance. This can be avoided in some circumstances (for example trials to assess rail cleaning methods or low adhesion mitigation techniques) by conducting the tests on dedicated track Results will be affected by any changes in moisture level during the time taken to set up and conduct the tests A constant and sufficient brake application is required to ensure that conditions are maintained over the stopping distance A train path is required if using a non-service train Advantages The service / test train brake does not require any specialist equipment (but OTMR output or an accelerometer could be used to obtain train deceleration) and an unmodified service train can be used for the test Very useful for incident investigations (for example SPADs or overruns) Demonstrates train performance with no scaling effects. Page 46 of 61 RAIL SAFETY AND STANDARDS BOARD

47 Part 9 Adhesion / Speed Relationship 9.1 Measurement of adhesion levels using full sized railway wheels on rails has shown that adhesion levels decrease with increasing speed as shown in Figure Mean adhesion on UK railways at 32 km/h Dry rail (UIC) Damp rail (UIC) Damp leaf film (UIC) Adhesion Wet rail (Japan) X X Speed (km/hr) Figure 22 Adhesion / speed relationship 9.2 If a small-scale contact portable tribometer is used to measure adhesion levels then scaling factors between the results and full size must be taken into account before the above relationship can be used to allow for speed variations. 9.3 The adhesion / speed relationships shown in Figure 22 apply only to the contamination conditions given in the plot (damp leaf film, wet, damp and dry rail). Other contaminants (for example a significant amount of grease mixed with leaf) may result in a different relationship. In these cases measurement of the variability of adhesion levels with speed using a full-scale tribometer would be necessary. 9.4 Example: suppose a portable tribometer has been used to measure the adhesion levels on dry leaf film and an average value of 0.46 is obtained at a measuring speed of 5 km/h and an operator requires an estimate of the likely adhesion level for a train at 150 km/h. Table 5 in indicates that the portable tribometer readings equate to a full-scale wheel-on-rail average adhesion level of Assuming that no other full-scale data is available, the closest adhesion / speed data to this value ( marked as point x 1 on Figure 22) is the curve relating to wet rail as produced by Japan (reference. Ohyama Dr. T, 4th RTRI Symposium Tokyo Nov 7th 1991). Assuming that the adhesion / speed data for dry leaf film follows the same falling characteristic as the other conditions shown in Figure 22, then the adhesion level at 150 km/h will be close to 0.05 (point marked x 2 on the graph). RAIL SAFETY AND STANDARDS BOARD Page 47 of 61

48 Part 10 Measurement of Contamination Properties 10.1 Scope 10.2 Purpose 10.3 Uses The measurements described in this Part are intended to evaluate the properties of visible railhead contamination This includes measurement of the following properties: a) Thickness gauge b) Surface roughness c) Moisture content d) Shear torque e) Pencil hardness This Part gives guidance on the use of data from the above measurements. The data could be obtained for natural railhead contamination or for simulated low adhesion contaminants To provide a consistent and complete set of information relating to contamination in low adhesion conditions. Provides data to support the interpretation of measured results and enables comparisons to be made between sets of data obtained under similar conditions Measurement of contaminant properties support: a) Identification of conditions consistent with a low adhesion incident. b) Ensuring that consistent conditions are produced for testing purposes. c) Providing additional information on the effectiveness of mitigation methods. d) Monitoring the build-up of contaminant films Measurement of contamination Thickness gauge Thickness measurements should be made using a ferromagnetic probe on the Sonacoat F-Gauge or equivalent instrument (Figure 23). The probe measures substrates / coatings or contaminants above a steel (ferrous) base layer. Page 48 of 61 RAIL SAFETY AND STANDARDS BOARD

49 Figure 23 Sonacoat thickness gauge The calibration is checked relative to a reference ferromagnetic base plate using reference thickness foils supplied with the instrument prior to taking readings. These gauges allow the range of the instrument to be adjusted according to the thickness of the film to be measured The Sonacoat instrument measures the average thickness over the area of the probe. A sufficient number of measurements should be made to provide a statistically significant sample size, but in any case not less than 3 measurements should be taken The average, minimum and maximum values of thickness should be quoted and where sufficient numbers of measurements have been made the standard deviation should also be calculated Where there is a black contamination film with thickness in the range of 10 to 100 µm with an average of 42 to 44 µm, then this indicates that leaf film is present. Other property measurements should be taken to confirm that the contaminant is leaf film for example moisture, hardness, adhesion level under damp conditions If tests are carried out on an artificially generated contaminant layers the results might be affected by variations in the thickness of the layers. For example, a comparison of a range of treatment methods using the measurement of thickness would have to take this into account A normalisation process could be used to take account of thickness variations. This process could use one set of results as a reference Figures 24 and 25 show some invented data to illustrate the normalisation process. Thickness values are recorded at a number of times throughout the tests and plotted (Figure 24) for two treatment methods T1 and T In Figure 25 the results for treatment T2 are shifted so that the thickness prior to the application of the treatment overlies the thickness of the layer used to test treatment T The results indicate that treatment T2 removes more of the contaminant layer than treatment T1. These results should be checked for repeatability. RAIL SAFETY AND STANDARDS BOARD Page 49 of 61

50 Normalisation for varying contaminant thickness (Hypothetical data to illustrate the normalisation method) Formation of contaminant layer Treatments applied Thickness Time (hours) Treatment T1 Treatment T2 Average T1 Average T2 Figure 24 Test results for two treatment methods Contaminant thickness normalised (Hypothetical data to illustrate the normalisation method) Formation of contamination layer Treatments applied Thickness Time (hours) Treatment T1 Compensated T2 Average T1 Figure 25 Normalised thickness Surface roughness The surface roughness of the leaf film or other contamination can be measured using a Talysurf / Hobson Surtronic 10 hand held device or equivalent instrument (Figure 26) This enables a comparison to be made between the contaminant film thickness and surface roughness to indicate if there is likely to be metal-to-metal contact between the wheel and rail. The measurements should be calibrated against a known surface roughness supplied with the device. Page 50 of 61 RAIL SAFETY AND STANDARDS BOARD

51 Figure 26 Hand held surface roughness measuring device A worst case rail surface roughness has been measured on rails ground with aggressive stones giving 20 µm whilst for wheels the maximum measured roughness is 3.5 µm. If the measured film thickness is less than approximately 25 µm then the probability of metal-to-metal contact will increase rapidly. This thickness will reduce for smoother wheels and rails Measurements should be taken along the length of the rail. A minimum of 3 measurements should be obtained Wheel surface roughness measurements should be taken across the tread since measurements around the circumference will be affected by the curvature Moisture content The level of moisture in the leaf film can be measured using the Protimeter Timbermaster or equivalent instrument (Figure 27). This device uses a set of external twin probes that can be rested on the surface or used to penetrate the contamination being assessed. Moisture probes Temperature compensation probe Figure 27 Moisture and temperature compensation probes In wood, the instruments measure the material s actual percentage moisture content (percentage of H 2 O). When testing material other than wood, the meter measures the wood moisture equivalent (percentage WME) value of the material. WME is the moisture level that would be attained by a piece of wood in equilibrium with the material being tested. RAIL SAFETY AND STANDARDS BOARD Page 51 of 61

52 Moisture level will vary according to temperature and therefore the temperature compensation probe should be placed in contact with the surface on which the contamination is bonded A sufficient number of measurements should be made to provide a statistically significant sample size, but in any case not less than 3 measurements should be taken Conditions may vary considerably during the measurement due to the effects of wind, sun, rain, drizzle or fog. A note should be made of the weather conditions at the time of the measurement If conditions are changing then measurements can be made every 15 to 30 seconds in order to monitor the drying or moisture uptake of the contaminant. If %WME values are measured that approach 100 then this indicates that the contaminant is becoming saturated with water and that the adhesion level is likely to be low Shear torque Consideration should be given to safety requirements before using this technique since at least 5 minutes is required to obtain the measurement and the method involves gluing a stud to the railhead. An alternative (see below), but less accurate method (the pencil hardness test) is available for situations where the shear torque method is inappropriate A torque test can be used to measure the shear torque of contaminant films. A stud of 15 mm diameter should be adhered to the rail head / contaminant with Loctite 401 (super glue / cyanoacrylate) and allowed to cure (normally 3 minutes but subject to atmospheric conditions). The torque should then be applied to the stud with a torque screwdriver (Figure 28) and increased to the point at which the film is sheared from the surface Leaf film shear torque is in the range 0.3 to greater than 1.0 Nm. Table 8 lists the range of shear torque that relates to different leaf contamination conditions. Leaf film contamination conditions Shear torque (Nm) range (Note 10 lbf. in. approximately 1 Nm) Moist to saturated 0.0 to 0.5 Lightly dampened 0.5 to 1.0 Dry > 1.0 Table 8 Shear torque readings at different leaf film contamination conditions The maximum shear torque measurable with the screwdriver should be 2.5 Nm if slightly moist (damp) to wet (saturated) contaminant films are to be measured The maximum shear torque measurable with the screwdriver should be 5.0 Nm if dry contaminant films are to be measured. Page 52 of 61 RAIL SAFETY AND STANDARDS BOARD

53 Figure 28 Shear torque measurements A minimum of 3 measurements should be made under each rail condition. Dry contaminant may be lightly dampened and a shear torque measurement acquired under these simulated low adhesion conditions Pencil hardness test The pencil hardness tester (Figure 29) Q1001 supplied by Paint Test Equipment Ltd or equivalent, can be used to measure the relative hardness of leaf films in a range of conditions. Figure 29 Pencil hardness tester The following standards BS 3900-E19 and ISO 15184, provide instructions for the use of pencil hardness testers The test method uses pencils of a known hardness grade and are moved over the surface. The pencil holder provides a fixed angle and pressure to the test surface. An arbitrary scale for the pencil hardness should be used for plotting against other parameters as shown in Table 9. Soft Pencil Hardness Range (lead type) Hard 5B 4B 3B 2B B HB F H 2H 3H 4H 5H Table 9 Pencil hardness arbitrary range Track measurements on leaf film indicate that the values in Table 10 will be obtained under dry leaf and damp leaf conditions. RAIL SAFETY AND STANDARDS BOARD Page 53 of 61