Load Introduction in Structural Tests

Load introduction is part of the test design

How load is introduced into a specimen is not a rig detail; it is part of the test design. The load introduction method determines the distribution of load over the specimen surface, the local stress field at the introduction point, and the load path by which the load enters the structure. A test can apply the correct total load — the load cell reads the right number — but introduce it through a path that does not represent service, producing local stresses that do not exist in service and altering the global response the test was designed to measure. The objective of load introduction design is to apply the intended load distribution through the intended load path, with local effects at the introduction point that are either representative of service or sufficiently remote from the region of interest that they do not contaminate the measurement.

A TEST CAN APPLY THE CORRECT TOTAL LOAD THROUGH THE WRONG LOAD PATH.

Load introduction methods and their character

Several methods are used to introduce load into a structural specimen, and each produces a different distribution with different local effects. A hydraulic actuator applies load through a single point or a small interface; it is versatile and precise in magnitude but concentrated. A spreader beam distributes a single actuator load across two or more points; it converts a point load into a line or multi-point load but introduces the stiffness of the beam. A whiffletree distributes a single load across many points through a tree of beams and links; it produces a close approximation of a distributed load but is complex to design and only works for a specific load distribution. Pressure loading applies load over an area through pneumatic or hydraulic pressure pads; it produces a true distributed load but is limited in magnitude and can introduce friction or adhesion. Distributed loading via weights, sandbags or pressure pads applies a body-force-like load. Point loading applies load at a single point; it is simple but creates a local stress concentration. Straps or reaction fittings introduce load through flexible or contoured interfaces that conform to the surface.

Load introduction methods visualised

The diagram shows four common load introduction methods and the load paths they create into the specimen. A point load concentrates force at a single point; a spreader beam distributes one actuator load across two points; a whiffletree distributes a single load across many points through a tree of beams; pressure loading applies force continuously over an area. The choice depends on what the service load distribution is, and how closely the introduction must approximate it.

LOAD INTRODUCTION METHODS

  POINT LOAD
        ↓ F
      ──●──        (concentrated at one point;
     specimen       high local stress)

  SPREADER BEAM
       ┌───┐
       │ F │
       └─┬─┘
      ┌──┴──┐
     ──●───●──     (one actuator, two points;
      specimen      beam stiffness affects distribution)

  WHIFFLETREE
         │ F
     ┌───┴───┐
    ┌─┴─┐   ┌─┴─┐
   ┌┴┐ ┌┴┐ ┌┴┐ ┌┴┐
  ──●─●─●─●──      (many points;
   specimen         approximates a distributed load)

  PRESSURE LOADING
     ↓↓↓↓↓↓↓↓↓↓
    ┌──────────┐
    │ pressure │    (continuous over an area;
    │   pad    │     true distributed load)
    └──────────┘
     specimen

  The method must represent the intended load path.
  A point load where service load is distributed
  creates local stresses that do not exist in service.

Avoiding unrealistic local concentration

The most common load introduction error is introducing a distributed service load through a single point. This creates a local stress concentration at the introduction point that does not exist in service, and this concentration can dominate the local response — causing local yielding, local buckling or local failure — while masking the global response the test was designed to measure. The defence is to match the introduction method to the service load distribution: if the service load is distributed, use a spreader beam, a whiffletree or pressure loading; if the service load is introduced through a fitting, use a representative fitting; if the service load is a pressure, use a pressure pad. Where a perfect match is not possible, the local effect of the introduction should be assessed and shown to be remote from the region of interest, or included in the analysis so that the test and the model share the same load introduction.

INTRODUCING LOAD THROUGH A SINGLE POINT WHERE THE REAL STRUCTURE CARRIES DISTRIBUTED LOAD CAN CREATE LOCAL STRESSES THAT DO NOT EXIST IN SERVICE AND MASK THE GLOBAL RESPONSE THE TEST WAS DESIGNED TO MEASURE. The load introduction must represent the intended load path.

Load alignment and eccentricity

Load alignment is the second axis of load introduction design. The load must be applied along the intended line of action; any eccentricity between the actuator axis and the intended load line introduces a moment that the specimen was not designed to carry. A small eccentricity on a large load can introduce a significant moment: an eccentricity of a few millimetres on a load of hundreds of kilonewtons produces a moment that can change the stress distribution in a fitting or a joint by a non-trivial amount. Alignment is checked mechanically — by measuring the actuator position relative to the load line — and instrumentally — by monitoring the load cell for off-axis components or by measuring the specimen response for asymmetry that indicates an unintended moment. For tests where alignment is critical, the rig should allow adjustment of the actuator position and the load line before the test begins.

Load rate and load control

The rate at which load is applied and the mode of control — load control, displacement control or strain control — are part of load introduction design. A load-controlled test applies a defined force regardless of specimen displacement; it is appropriate for stiff specimens and for proof or qualification tests, but it can be unstable at the point of specimen softening, where the actuator races to maintain load. A displacement-controlled test applies a defined actuator displacement regardless of load; it is stable beyond peak load and is appropriate for tests to failure and for post-peak behaviour. A strain-controlled test uses a measured strain as the feedback signal; it is used when the strain at a specific location is the controlling quantity. The choice affects what the test can show: a load-controlled test to failure may end in sudden collapse, while a displacement-controlled test to failure may show the post-peak softening that characterises the failure mode.

Whiffletree design and its limits

A whiffletree is the most accurate way to approximate a specific distributed load with a finite number of point loads. It works by distributing a single actuator load through a tree of beams and links, with each beam splitting the load equally between its branches. The number of points determines how closely the discrete load approximates the continuous distribution: more points give a closer approximation but a more complex tree. The design is specific to one load distribution — a whiffletree designed for a uniformly distributed load does not produce a triangular load — so it is most appropriate when the same load distribution is applied repeatedly. The links must be free to articulate, so that each point applies a normal load without introducing parasitic moments, and the beam stiffnesses must be high enough that the load sharing is not affected by beam deflection.

Interaction with the fixture

Load introduction and fixture design are coupled. The load introduction method determines the local load at the specimen surface; the fixture determines the boundary condition at the supported edges. If the load introduction creates a local stress at the introduction point and the fixture bridges across a nearby region, the two effects can interact to produce a stress distribution that represents neither the service load nor the intended test condition. The load introduction and the fixture should be designed together, and the combined system — specimen, fixture, load introduction — should be checked against the analysis assumption before the test begins.