Graham Slater is a Principal Research Engineer in the Engineering Department. He joined the Institute in 1979 after completing a combined engineering degree course and apprenticeship with the Ministry of Defence.
After some years in the Fatigue Department, he took on responsibility for the Institute's experimental stress analysis services, and in particular the on-site strain gauging activities which he still controls.
His current activities are biased towards consultancy in the areas of design and stress analysis of welded structures.
Graham Slater takes a deeper look at the escalator testing in the January issue of Connect.
Escalators, in common with most other types of public service equipment, have to be carefully designed, manufactured and tested to minimise the possibility of mechanical failure. The consequences of structural failure in an escalator system could be very serious indeed, and there have been occasions when passengers have been injured as a direct result of such failure. In recent years the Institute has been involved in testing a number of different designs of one of the main components of a system: the step.
Design
Like the proverbial iceberg, nine-tenths of an escalator step is below the surface. Most of us are familiar with the top and front of a step, but there is a lot more hidden beneath. The ribbed tread plate is usually of aluminium although older systems, some of which are still operational in the London Underground, may be of wood. The curved front plate may be steel, aluminium or plastic. Beneath these two components is a sturdy steel frame, made up of punched, folded and welded steel sheet. Sheet thickness varies from (typically) 2 to 6mm.
Each step runs on four wheels which engage in guide channels forming part of the support structure. The upper pair of wheels is fitted to the ends of an axle shaft running more than the full width of the step: the distance between these wheels may be nearly half as much again as the width of the visible surface of the step. The lower pair of wheels runs inboard of the upper pair in separate guide channels: these wheels are normally mounted on stub axles protruding from the lower part of the main steel frame. The wheels themselves are usually made of aluminium with nylon tyres. The individual steps are connected by links on each side of the step engaging with the upper axles.
The major design requirement for an escalator step is adequate fatigue strength. Every step has to survive a very large number of loading cycles in a typical design life (20 years, for example) and the welded fabrication is naturally prone to fatigue cracking at relatively low stress ranges. Static strength and stiffness are also important. These two parameters are linked: clearly a step must not collapse under a large applied load; equally, it must not deflect sufficiently to cause fouling of any moving parts or trap passengers' feet or luggage.
BS 5656: 1983 Safety rules for the construction and installation of escalators and passenger conveyors covers the design of all the major components of an escalator system, including specification for type-testing steps by an authorised testing authority. The Institute has performed a number of these tests.
Testing
We have purpose-built facilities for testing escalator steps, including a number of test frames to accommodate the dimensional variations between different systems. Careful design of test frames is important - the upper wheels must he rigidly fixed, but the lower wheels must be free to move on an inclined plane of the same slope as the inclination of the real escalator system.
Two types of test are specified - static and dynamic. The static test is a simple application of a single 3.0kN force vertically downwards on the centre of the tread surface through a load-spreading plate 200 x 300 x 25mm. In the dynamic test the same loading arrangement is used but the force range is specified as 0.5-3.0kN for a minimum of five million cycles. In each test a fully assembled step must be used, complete with axles and wheels. In the static test, the step tread is allowed to deflect elastically by up to 4mm, but no permanent deformation is permitted. In the dynamic test. there must be no fracture or permanent deformation greater than 4mm.
Not surprisingly, the test criteria appear conservative, particularly when compared with likely service conditions. For example, the maximum applied force of 3kN, roughly equivalent to a 300kg load, is clearly in excess of all but the most exceptional loading for a pedestrian escalator. Two large men with heavy suitcases might conceivably have a combined weight of 300kg, but they would find it very difficult to squeeze on to the same escalator step! Our experience of testing pedestrian escalator steps suggests that the 4mm limitation on allowable deflection probably relates more to jamming than strength. In practice, the maximum deflections under the static load are typically in the range 0.5-1.0mm, ie not more than 25% of the maximum allowed.
The fatigue test duration of five million cycles also appears to be rather arbitrary in relation to the current approach to fatigue design. The fatigue limit in constant-amplitude fatigue testing is commonly assumed to occur at around ten million cycles. [1] Since most escalator systems will be expected to operate beyond the fatigue limit (albeit under a spectrum of loading most of which will be lower than the maximum), it would perhaps be more rational to test to ten million cycles rather than five. However. the conservatively high force range used in testing (2.5kN) gives a good measure of confidence in the fatigue performance of the step under operating conditions.
Results
Of the many steps we have tested, only one has shown signs of premature fatigue cracking, and the cause of that was diagnosed as defective welding. The step was a one-off prototype, and it was unlikely that the defect would have passed any reasonably effective production quality control inspection.
Some manufacturers have taken the opportunity to investigate the strength and stressing of their steps in more detail than the basic requirements demand. For example, we have used the standard loading arrangement to test a step to destruction by applying a single static overload. This enabled us to determine the elastic limit load, plastic collapse load and mode of failure, and gave us an indication of the margin of safety on static strength and the location of the weakest part of the step. In another case, stresses arising in a step during fatigue testing were thoroughly investigated using the Institute's SPATE 8000 system.* This proved very useful to the designer, because the complexity of the design made accurate theoretical stress analysis very difficult.
*Slater G: 'Thermoelastic stress analysis with SPATE 8000'. Welding Institute Bulletin 29 (5/6) 113-119.
Meaningful validation
Problems are experienced in the design of any component which, like an escalator step, is required to sustain severe fatigue loading in service but for which detailed load spectra are not available. As experience with escalator steps has shown, testing offers a meaningful method of validating a design, highlighting areas of potential weakness ( eg where weld quality is particularly important) and increasing confidence in the product. The extensive static and fatigue testing facilities at the Institute are available to Industrial Members for all such types of test.
