derailment in Ledsgård, N county, 28 February 2005

Transcription

1 ORGANISATION INTERGOUVERNEMENTALE POUR LES TRANSPORTS INTERNATIONAUX FERROVIAIRES OTIF ZWISCHENSTAATLICHE ORGANISATION FÜR DEN INTERNATIONALEN EISENBAHNVERKEHR INTERGOVERNMENTAL ORGANISATION FOR INTER- NATIONAL CARRIAGE BY RAIL INF.2 5 Mai 2008 (English only) RID: 9 th Session of the Working Group on Tank and Vehicle Technology (Berne, 14 and 15 May 2008) Subject: Report involving train 5525 collision with buffer stop assembly, with subsequent derailment in Ledsgård, N county, 28 February 2005 Note from Sweden Excerpt from: ISSN Report RJ 2007:2 Report involving train 5525 collision with buffer stop assembly, with subsequent derailment in Ledsgård, N county, 28 February 2005 Case no. J-01/05 (Translation: Swedish Rescue Services Agency) The Swedish Accident Investigation Board (SHK) investigates accidents and near-accidents from a safety perspective. The purpose of these investigations is to prevent similar incidents in future. On the other hand, the investigations of the Swedish Accident Investigation Board do not aim to assign guilt or liability. All material in this report is free to all for publication or use for any other purpose, as long as the source is stipulated. This report (Swedish) is also available on our website: For reasons of cost, only a limited number of copies of this document have been made. Delegates are asked to bring their own copies of documents to meetings. OTIF only has a small number of copies available. Tel. (+41) Fax (+41) info@otif.org Gryphenhübeliweg 30 CH 3006 Berne/Bern

2 Summary At 12:46 on Monday, 28 February 2005, the locomotive and the first few wagons in freight train 5525 were derailed after having passed a main signal at stop-position (at danger) and then driven through the buffer stop after the derailing point on the extension of track 1 at Ledsgård station. Train 5525 was made up of an electric locomotive and 12 wagons loaded with chlorine. The wagons had been loaded in Bohus and then, following a functional inspection (safety inspection and brake test), driven to the marshalling yard in Sävenäs using a diesel locomotive as train At Sävenäs, the locomotive was replaced and a new brake test was performed before the train departed. When the train was put in order prior to its departure from Bohus, the position of the wagons empty/loaded handles were not checked. These are used to control an empty/loaded changeover device, which affects the braking capacity in relation to the load of the wagon. The braking capacity of the train was established according the conductor's report (list of wagons) which provided weight details which were based on the situation that the wagons should be loaded and that the handles should be in the correct position. The driver did not have access to this list of wagons during the functional inspection and brake test. However, the empty/loaded handles were in the EMPTY position, which resulted in a considerable reduction of the actual braking capacity of the train. Both the driver of train 6505 on the way to Sävenäs marshalling yard and the driver of train 5525 following the exchange of locomotive both carried out running brake tests, but did not notice that the braking capacity of the train set was insufficient. At Ledsgård station, train 5525 was to enter and stop on track 1, so that another train could pass. At the station approach, the driver braked slightly to reduce the speed. When he later saw that the entry points were positioned to track 1 (curved position) and that intermediate signal 33 was set to stop (at danger), he braked for stop. Just after that, the ATC (automatic train protection system) initiated a braking intervention. There is a downhill gradient at the station yard in Ledsgård in the direction of the train. There is a gradient of 9 adjacent to the intended stop position. This gradient was not coded in the ATC data. Had this coding been applied, the ATC brake intervention would have been initiated earlier. The position of the empty/loaded handles and the gradient resulted in the actual deceleration during braking being approximately half of what the driver was anticipating and what was entered in the ATC. Intermediate signal Lgd 33, where the train was supposed to stop, was passed at 43 km/h and the train collided with the buffer stop after the derailing point at 39 km/h. The locomotive and almost three wagons passed the end of the protective track and stopped in the mud in the field beyond the buffer stop. The locomotive and first four wagons sustained extensive damage. When the rescue services arrived at the scene of the accident, it was established that no chlorine gas was leaking from the wagons. The damaged wagons were emptied before being towed away. The work of the rescue services was time consuming and took place over 16 days. During the rescue operation, the rescue services director was of the opinion that, following the initial stage of the rescue operation, not all criteria for the rescue services were being met in accordance with Civil protection act, a legislation on protection against accidents and on rescue services. (Lag om skydd mot olyckor, LSO, SFS 2003:778). The direct cause of train 5525 being unable to stop before the signal at stop-position on track 1 in Ledsgård was the fact that the actual braking capacity of the train was greatly reduced; due to the fact that all the empty/loaded changeover devices were in the incorrect position EMPTY instead of LADEN. This was not discovered when the braking capacity of the train was established in Bohus. In Ledsgård, the intervention of the ATC was delayed when the speed was not reduced as expected due to the fact that there were no gradients coded in the ATC data. This would not have 2

3 prevented the incident, but it would have reduced the consequences. The same would be true if the buffer stop on the protective track had been set up correctly. Both of these conditions arise from inadequate planning and procedures. The shortcomings in the execution of shunting, brake test and establishment of the braking capacity prior to departure from Bohus may have come about on the basis of inadequate procedures and training, as well as the relatively rough system applied for audits of the operational workforce. There are no barriers which would prevent a train from departing before the position of the changeover empty/loaded handles had been checked to correspond to the braking weights used when calculating the brake percentage. The system involving running brake test and retardation calculation is not sufficient to optimally trace deviations from the braking percentage established in theory for the train composition. 1.4 Deaths, personal injury and damage to property Damage to rolling stock The locomotive sustained extensive damage. The first four wagons sustained extensive damage to the under frame area. In addition, the tank on wagon 1 sustained minor but distinct damage to the end of the tank. The tank on wagon 4 sustained a fairly large, soft dent in the end of the tank. The damage was not noticed until the protective plate had been removed: this was done after the wagon had been transported away from the scene of the accident. The valve arrangements at the tops of the wagons were undamaged. The tank on wagon 2 had to be cut free from the sub frame when the wagons were towed away. Other wagons sustained minor damage. The energy absorbing buffers on the wagons were deformed to a certain extent. This is described in greater detail in section Above (on top): Wagon 1 without bogie and parts of the sub frame from wagon 2 after towing the wagons away. Above (left): The locomotive and wagon 1 at the scene of accident. Above (right): Wagons 2 and 3 at the scene of accident Damage to the railway infrastructure The derailing points, the protective track up to the buffer stop, the buffer stop itself and an overhead contact line were damaged in the accident. The buffer stop assembly was pushed down into the ground and has not been recovered during clearance or restoration work. 3

4 1.4.4 Damage to the surroundings and environment There was no permanent damage to the surrounding area as a result of the accident or the towing work. There were no emissions of dangerous goods, either at the time of the accident or during the rescue operation. 1.5 The incident environment The train and its cargo Excerpt from timetable (only a selection of locations included here) Train Sävenäs marshalling yard Bohus Sävenäs marshalling yard dep. 09:38 Bohus arr. 09:59 Train 6505 Bohus Sävenäs marshalling yard Bohus dep. 11:18 Agnesberg 11:26 Gothenburg Marieholm 11:31 Gothenburg Sävenäs 11:34 Sävenäs marshalling yard arr. 11:39 Train 5525 Sävenäs marshalling yard Malmö Sävenäs marshalling yard dep. 12:21 Gubbero 12:30 Almedal 12:34 Mölndal lower 12:36 Lindome 12:43 Ledsgård 12:50 12:55 (scheduled stop, awaiting available train route) Kungsbacka 13:00 Composition of the train Train 5525 was made up of an electric locomotive, type Rc4 no and twelve loaded tank wagons, type Zagns. The train weighed 1070 tonnes and the braking weight was 696 tonnes. The train was 169 m long (excluding the locomotive). Entered ATC data, see section At the time of the accident, the wagons were running in the following order (viewed from the locomotive): No. Wagon number The wagons were owned by VTG Lehnkering AG in Hamburg and were on hire to Akzo in the Netherlands for use for the transportation of chlorine from, inter alia, Bohus to Rotterdam. The wagons were provided with orange coloured plates (ID plate) with hazard identification number 268 (toxic gas, corrosive) and UN no (Chlorine), and with placards no. 2.3 (toxic gases) and no. 8 (corrosive substances), as well as marshalling label no. 13 (to be shunted with care). 4

5 Tank load The train set was loaded at Eka Chemicals in Bohus. Each wagon was loaded with approx. 65 tonnes of chlorine. According to the regulations for dangerous goods, chlorine is classified as a toxic and corrosive gas, Class 2, with UN number Under normal pressure and temperature, chlorine is an extremely toxic and reactive gas. Chlorine gas is perceptible even at very low concentrations and has an extremely irritant effect on the respiratory organs. In the liquid phase, chlorine is a red coloured liquid with a density equivalent to 1.5 times that of water. This liquid is reactive (corrosive) on organic material and causes frostbite (boiling point -34 C). The potential risks of chlorine are in many ways comparable with other toxic, condensed gases such as sulphur dioxide, ammonia, formaldehyde and vinyl chloride. 2.8 Condition and function of the railway vehicles The empty/loaded change over device for braking capacity The wagons are equipped with a manual empty/loaded change over device for the braking capacity with a handle for the loaded and empty position. In the empty position the braked weight is 30 tons and in the loaded position the braked weight is 58 tons. The brake changeover weight is 48 tons. On all 12 wagons the changeover handle was in the empty position in the investigation at the scene of accident. The goods-passenger changeover device (G-P-R) on the locomotive was in the P position together with all the corresponding handles on the wagons. The brake was active on all wagons. The empty/loaded change over device for braking capacity in empty position Wagons The tank wagons are essentially traditional, 4 axle bogie wagons fitted with frames. The wagons were manufactured in 2002 and 2003 at Waggonbau Brüninghaus on behalf of VTG. This type of wagon has been provided with protective shields at the ends of the tanks in order to reduce damage of overriding buffers as a consequence of derailment or collision. This type of 5

6 wagon is also provided with energy absorbing buffers in order to be able to absorb collision energy in excess of those encountered in normal conditions of rail transport. The manufacturer states that this type of wagon has a reinforced chassis structure. Approval, etc. The Zagns wagons involved in the accident were fitted with energy absorbing buffers. Energy absorbing buffers were not yet mentioned in the design requirements laid down in UIC leaflet 573 at the time of the accident. As a result, the wagons had no RIV marking. Instead, they were provided with the exchange sign GC for Sweden, along with the corresponding sign for most of the European freight transport administrations. The wagons had German type approval, and Green Cargo had notified its approval in accordance with RIV 23 in a letter to VTG dated 9 July Energy absorbing buffers Besides normal elastic shock absorption, energy absorbing buffers must also have an energy absorbing capability through plastic deformation. Such energy absorption must not take place under normal operation and consequently be able to influence rail traffic safety. Energy absorption through plastic deformation may only take place in the event of impacts at speeds higher than 12 km/h. The buffers must be able to absorb 800 kj per wagon end without affecting the tank shell. The energy absorbing buffers were dismantled when the wagons were towed without the buffers being measured at all at the scene of accident. The total plastic deformation of the buffers was later measured by Green Cargo at 1665 mm. The deformation force of the buffers is specified as 1550 kn, which means that the buffers absorbed 2.5 MJ through plastic deformation, which is only a small proportion compared with the total kinetic energy of the train. On the wagons which were at the curve of the switch at the moment of impact, the buffers in some cases have overridden or overlapped so that no energy absorption could take place in the buffer. The wagon frames and tanks deformed instead. Further back in the train, the longitudinal forces did not exceed the buffers plastic deformation force of 1550 kn. If all buffers had been able to absorb energy, this would have been equivalent to 70% of the kinetic energy of the train at the moment of impact. 6

7 Above (top): Drawing of energy absorbing buffer. If the impact force exceeds the threshold force the buffer box (C) is forced towards the deformation zone (D) behind the frame end (B). Above (left): A non activated energy absorbing buffer. Above (right): An energy absorbing buffer which has been activated by the collision force and a deformed index marker Technical investigation of the strength of the tanks The Swedish Accident Investigation Board has had a special investigation and assessment carried out of some of the damage which occurred to the wagons at the moment of impact with the buffer stop and the subsequent derailment. The objective is to estimate the forces and energies to which the energy absorbing buffers, protective shields and tanks were subject at the time of the accident, and to estimate the corresponding resultant damage in the case of a wagon without a protective shield fitted. This investigation was carried out by Interfleet Technology AB. Below is a brief summary. The report in its entirety can be found in the archive of the Swedish Accident Investigation Board, designated TS RES, case J-01/05, document annex 96a. Deformation of the tank on wagon 4 When the accident occurred, the end of the tank on wagon 4 was deformed by a buffer head on wagon 3. Pictures show that no significant plastic deformation has taken place on the buffer which hit the end of the tank on wagon 4. The protective place shows minor damage in the form of cuts from the buffer head. The frame on the wagon also shows minor deformation which may also have absorbed energy. By measuring the deformation of both the protective shield and the end of the tank, it is possible to estimate the forces absorbed by the protective shield and the end of the tank when the train collided with the buffer stop. Dent in the protective shield on front end of wagon 4. 7

8 Measurement method The extent of the deformation in the protective plate and tank was measured using a laser distance meter in a defined coordinate system. Measurement data has been translated to a mathematical model of the protective plate and tank, which formed the basis of FEA-calculations. Displacement in the tank shield (mm) Displacement in the tank end (mm) 8

9 Calculation method The FEA calculation determines the forces and energies which were required to bring about the damage which occurred. Once these are known, corresponding calculations are carried out on a FEA model without a protective plate in order to estimate what damage would have occurred in the case of a wagon with no protective shield. Results Throughout the entire sequence of events, the forces required to deform the protective shield and tank are less than the force specified for plastic deformation of an energy absorbing buffer. At a stress of 500 kn, the attachment joints of the protective shield were torn off. The stress in the tank is estimated to have amounted to 1150 kn, which led to a dent approx. 200 mm deep in the end of the tank (large radius). The protective plate absorbed 36.7 kj before it came loose and touched the tank. The continuing deformation of the protective plate and tank absorbed a further kj. In the present case, the protective plate and tank absorbed a total force of kj. Further calculations show that for a tank without a protective plate, an anticipated fracture of the tank will take place following energy absorption equivalent to almost 500 kj Force [kn] Tank with shield Tank without shield 8 mm tank Displacement [mm] The diagram above shows the forces which affected the protective shield and tank. The blue curve represents the actual configuration, with a 6 mm protective shield and a tank with a wall thickness of 11 mm. The red curve represents a tank of this type without a protective shield. The green curve is a tank with a wall thickness of 8 mm and without a protective shield. The red and green lines are drawn to fractures for the respective tank types. 9

10 Calulation of absorbed energy 600 Total absorbed energyi [kj] Tank with shield 300 tank without shield 8 mm tank The diagram above shows the energy which is absorbed during the deformation process. The blue line shows the tank in question with the protective plate. The red and green lines are drawn to fractures for the respective tank types as shown in the diagram above. Conclusions The protective plate has protected the end of the tank from possible cuts from the buffer head. The energy absorption of the protective shield is relatively low. A different design of the attachment arrangements of the protective shield could give it more energy absorbing qualities. There is a good margin between the energy absorption in the present case compared with the anticipated fracture of the tank, even for a tank with no protective shield. Discussion Displacement [mm] In the present case, the force which acted upon the end of the tank did not amount to 1550 kn and no energy absorption took place through deformation of the buffer. The energy absorption buffers' primary way of acting should be assumed to take place via energy absorption in the event of a buffer to buffer collision instead of a buffer to tank end collision. Future requirements for energy absorbing buffers specify that a wagon fitted with energy absorbing buffers must have an energy absorbing capacity of 800 kj per wagon end without energy absorption taking place by deformation of the tank. Available buffers which meet this requirement require according to specifications a force exceeding 1550 kn to be pressed together (deformed) and hence absorb collision energy. This force thus exceeds the force which can be expected to be sufficient to puncture a tank-wagon, for example, designed for flammable liquids. Conditions arising when wagons fitted with conventional buffers are placed alongside tank-wagons should also be noted. 3.6 Consequential analysis Derailing points, protective track and buffer stop assemblies The principles for frontal, flank and lateral protection of a train route and for the protective track beyond a derailing point have changed over the years. However, the basic concept of the derailing point and a protective track to lead vehicles in motion away from other train routes remains. 10

11 The design of the protective track, and primarily the buffer stop which is normally located there, also varies depending on what is present directly beyond the end of the track. In practice, the end of the track can consist of a heap of coarse aggregate or a strong concrete buffer stop. Between the two are a range of variants designed so as not to cause a sudden stop, but rather to brake a vehicle in motion. It is not part of this investigation to assess the subsequent damage had the train passed the stop signal at a different location along the route travelled by train However, the fact that there is a muddy field beyond the end of the track must be considered to be a factor which mitigated the damage. A derailment point and a protective track in accordance with regulations together with a buffer stop correctly fastened in position, could have brought the vehicle to a standstill better. However, in the current situation this would not have been sufficient to stop the train before the end of the track. The special energy absorption buffers on the chlorine wagons had a certain effect, but not to the extent they could have. See below. The Swedish Accident Investigation Board is of the opinion that there is a reason to review the requirements relating to the formulation of how the end of the track should be designed; both after a derailment point and in locations where a train route ends in a buffer stop. However, these requirements must take into account both the resulting damage to the vehicles which risk colliding with the buffer stop as well as the consequences of any structures or equipment located beyond the end of the track. The latter may affect waiting rooms or other public areas or affect terrain conditions, buildings, etc Assessment of stress on the train upon collision with the buffer stop The speed of the train at the moment of impact, approx. 40 km/h, is equivalent to a kinetic energy of approx. 66 MJ. This energy was released and absorbed in the form of deformations of the vehicle, the buffer stop and other infrastructure. The total plastic deformation of the buffers has been measured to be 1665 mm. The deformation force of a buffer of this type is specified to 1550 kn. The buffers absorbed 2.5 MJ through plastic deformation, which is only a small fraction of the total kinetic energy of the train. On the wagons which were at the curve of the switch at the moment of impact, the buffers in some cases have overridden or overlapped so that no energy absorption could take place in the buffer. The wagon frames and tanks deformed instead. Further back in the train, the longitudinal forces did not exceed the buffers plastic deformation force of 1550 kn. If all buffers had been able to absorb energy, this would have been equivalent to 70% of the kinetic energy of the train at the moment of impact. The standard dimension of a buffer stop normally refers to a train weight of 1000 tonnes at 15 km/h, which corresponds to a kinetic energy of approx. 8.7 MJ to be absorbed. In the present case, the buffer stop assembly was designed for 300 tonnes at 5 km/h. This is equivalent to just 0.3 MJ. There is uncertainty as to how the buffer stop was fastened to the track prior to the collision. The actual energy absorbing effect at the time of the accident cannot be assessed. It is stated that a buffer stop with a high degree of energy absorption capacity, which is not flexible in the event of a collision, will cause stresses in the train which can greatly exceed the forces generated in the present situation Wagon measures to alleviate the consequences This accident illustrated the function of the technical measures which would have reduced the stresses on the tanks in the train. The modern wagon structure, with energy absorbing buffers and protective shields on the ends of the tanks, ended up alleviating the consequences or at least reducing the risks involved in the accident. It is possible to state that further safety benefits are pos- 11

12 sible to implement by developing modern wagon structures. It is our opinion that the energy absorbing buffers in the event of accidents occurring on straight track and where the stresses on the train are greater would have a better effect. It is our opinion that the function of the protective plates could be improved with minor changes to the technique used to fasten them. It is not part of this investigation to assess the consequences if chlorine gas has leaked out. However, it is obvious that extensive emissions of chlorine in densely populated areas will involve an extremely extensive disaster scenario. 12