Wheel-rail impact loads generated by wheel flats - Detector measurements and simulations
Typ
Examensarbete för masterexamen
Master's Thesis
Master's Thesis
Program
Mobility engineering (MPMOB), MSc
Publicerad
2023
Författare
Mattsson, Klara
Modellbyggare
Tidskriftstitel
ISSN
Volymtitel
Utgivare
Sammanfattning
The railway system relies on trains running according to the train schedule to avoid
delays, which is why there needs to be as few interruptions as possible. This thesis
focuses on wheel tread damage in the form of wheel flats. Based on Trafikverket’s
regulations, if a flat is longer than 60 mm then the train has to be stopped and the
damaged wagon needs to be taken out of service to repair the wheel. The wagons
are then put aside on, for example, a passing siding at a station, which could have
been used for oncoming trains. This thesis aims to predict wheel-rail impact loads
caused by wheel flats, and how variables such as the flat length, axle load, train
speed, time since its formation, and unsprung mass influence the load.
The thesis is divided into two main parts: (1) an analysis of measured data from
wheel impact load detectors and (2) simulations of vertical dynamic vehicle-track
interaction and impact loads. The analysis is performed on wheels that have been
removed due to verified wheel flats, by extracting data back in time from when the
wheel started to generate a significant increase in peak load. The simulations were
carried out on wheel flat geometries that had been 3D-scanned and were performed
so that the influence of different axle loads, train speeds, and unsprung masses were
considered.
Based on 823 investigated detector passages with wheel flats of different lengths,
there was no case of peak load exceeding Trafikverket’s high alarm limit 350 kN.
This was surprising since several of the investigated cases involved wheel flats longer
than 60 mm. No clear correlation between measured impact loads and the variables
train speed and flat length was found. This could be due to reasons such as the
influence of different lateral offsets between the wheel-rail contact (rolling circle)
and the position of the wheel tread damage, and potentially the detector’s ability
and accuracy to measure high-frequency wheel-rail impact loads.
The simulations were performed for two measured wheel flats, with lengths 75 mm
and 120 mm, respectively. For axle load 25 tonnes and train speed 100 km/h, the
75 mm wheel flat with a depth of 1.4 mm reached a maximum peak load of 350 kN.
The simulations were performed so that the wheel flat hit the rail at different positions
along the sleeper bay. As expected, the most severe case was when the flat hit
on top of a sleeper where the track stiffness is higher. The 120 mm wheel flat with a
depth of 1.2 mm resulted in significantly lower peak loads, never reaching the high
alarm limit of 350 kN. In this case, the maximum calculated peak load was 238 kN
at 25 tonnes axle load and train speed 140 km/h.
A separate analysis to study the accuracy of the detectors to measure static (mean)
loads was also carried out. In this study, locomotives on Stålpendeln with known
axle loads were used. This indicated a variation in measurement accuracy between
the different detectors. It is argued that the calibration of the detectors, as well
as variations and differences track stiffness and track geometry at the detector sites
might have influenced the results. In addition, for three selected wheel flats, the
dynamic loads registered in different detectors along the same southbound or northbound
journey (in loaded or unloaded conditions) showed large variations.
The results in this thesis confirm that the depth of a wheel flat has a larger influence
on generated impact loads than its length. Further, from both the analysis of wheel
impact load detector data and the simulations, the wheel-rail impact loads for a given
wheel flat size were found to be lower than expected. Thus, from the findings of
this report, it seems more reasonable to base regulations about a wagon’s continued
operation on measured peak loads than on wheel flat size. Since, for any given wheel
flat size, unloaded wagons generally generate lower peak loads than loaded wagons,
such regulations would result in more flexibility in allowing unloaded wagons with
wheel tread damage to continue their operation to a workshop for repair as long
as the alarm limit is not exceeded. If unloaded wagons were allowed to continue
(at recommended speed) to their final destination with workshop capabilities, this
would lead to fewer traffic disruptions while maintaining safety.
Beskrivning
Ämne/nyckelord
Wheel flats , wheel tread damage , wheel-rail impact load , dynamic wheelrail interaction , wheel impact load detectors