| |
|
|
| |
| Case
Study - Locomotive Structural Analysis, 10 May 2006 |
|
|
|
| |
|
|
| |
Most of the mechanical
engineering design work we do includes an element of stress analysis,
since a brilliantly innovative design is no use at all if it’s
going to break!
The standards against
which we measure the structural strength of our designs depend
on the application, since requirements often vary with industry
or with type of machine. For example railway vehicles of all types
intended for main line use, are covered by either Railway Group
standards (for the UK), UIC standards (for Europe) or AAR standards
(for USA) – which are generally very prescriptive, with
“load cases” and “acceptance criteria”
covering all critical components. For non-rail applications welded
steel structures can be analysed against BS 7608:1993 ‘Fatigue
Design and Assessment of Steel Structures’. In addition
to the available standards, we test components to levels agreed
on by our team of Engineers in conjunction with information from
the customer and our own scientific findings.
Nowadays most of our
stress analysis is computer-based using SRAC “Geostar”
finite element analysis software running on one of our dual-Athlon
64 workstations; a high performance graphics based machine for
pre- and post-processing and model tidying/creation; and a powerful
dual core machine for resource intensive parallel solving. Brian
is our structures expert and having spent time working for the
Honda/BAR Formula 1 racing team during his MSc year he commands
quite a lot of respect! He has developed methods and processes
to streamline the Finite Element method as much as is possible
and utilise the processing power available twenty four hours a
day.
Here are some examples
of his recent work on our Corus steelworks locomotive project:-
The loco frame design
was developed from a merger of knowledge of existing very heavy
duty industrial locomotives and current and future safe working
practices with the health and safety of the operator as important
as the structural integrity of the frame. The locomotive had to
be suitable for both the steelworks/heavy industrial application
as well as for use on the main line. Network Rail in the UK defines
clearly the structural standards which must be met in order to
gain approval for use on their network.
|
|
| |
|
|
| |
|
|
| |
GMRT
2100 Structural Standards |
|
| |
|
|
| |
The
standard yields a list of proof load cases which must be applied
to the design, including buffing and drawing loads, jacking and
lifting loads, imposed loads from body-mounted equipment and any
combination of the above. Fatigue load criteria are also specified.
In addition to the Network Rail load cases we have our own a Steelworks
load case, based on the Network Rail standard but allowing for very
heavy shunts, poor quality track, regular derailments and awkward
lifts during re-railing. The data for defining the Steelworks load
case was collected from implicit observations of current industry
practice and derailment records as well as explicit loads and accelerations
collected from load cells, strain gauges and accelerometers positioned
on working test vehicles.
With
the load cases defined, the first stage is to build a global locomotive
frame model to have an overall look at what the frame is doing under
the differing load conditions. For such a large structure, made
up of relatively thin sections (where the thickness of the plate
is very small compared to the length of the frame 40:12500) a surface
model is created to allow a shell mesh to be used to speed up solution
times and efficiency. |
|
| |
|
|
| |
|
|
| |
Creating
the meshed component from the solid model. |
|
| |
|
|
| |
The
‘tidy’ surface geometry was created in Unigraphics ‘Solid
Edge’ and exported as an IGES file to Geostar. The geometry
must be created in this way from the onset to prevent analysis anomalies
and allow a stable mesh to be created. The various load cases can
then be run and the results analysed in order to establish further
how best to tackle any problems that arise. |
|
| |
|
|
| |
|
|
| |
Stress
Analysis results, brighter colours denoting higher concentrations
of stress. |
|
| |
|
|
| |
In the case of the locomotive
frame it was decided that the shell mesh did not yield enough detailed
information on particularly interesting parts of the geometry e.g.
weld configurations, plate lapping etc. so a secondary analysis was
carried out largely using a ‘tidied’ solid quarter and
half models where symmetry allowed. |
|
| |
|
|
| |
|
|
|
Quarter
model for detailed analysis. |
Half
model with load points applied. |
Mesh
refinement. |
|
|
| |
|
|
| |
Each
of the load cases was then run to convergence and the results pooled
ready for a design review in which modifications to eliminate areas
of highest stress were discussed by our team of Engineers. The design
review identified areas where additional strength was needed, methods
of eliminating stress-raising geometry, improved plate lapping configurations
as well as fabrication and machining economies such that the final
design could be manufactured easily and cost effectively. |
|
| |
|
|
| |
The result?
|
|
| |
|
|
| |
An
efficient frame design, production-engineered, strong and safe.
The design meets the Network Rail acceptance criteria, as well as
our derived Heavy Industrial/Steelworks criteria, is economical
to manufacture and easy to maintain. On top of all of this the design
enables the locomotive to meet all the relevant Health & Safety
at Work requirements and in particular those relating to driver/operator
safety and ergonomics. |
|
| |
|
|