It is quite simple to assess the benefits of using digital image correlation. For test engineers, DIC makes it feasible to prospect a larger area, to better understand the physical phenomena at stake by using a non-intrusive measuring method. For engineers working on numerical simulations, it is now possible to validate their models using vaster data sets, and so to be both more efficient and predictive as shown on the « Alstom bogie case » about which we have already talked concerning technical aspects.
And, here is the point: what are the benefits of using DIC on such a structural test? Is developing skills on a completely new technique financially worth it? EikoSim guides you to highlight the main areas for improvement.
During the test: saving time and money.
The first thing every DIC user wants to do is to get rid of strain gauges. Traditional strain gauges are heavy to instrument, both in time and money (one often hears that one instrumented strain gauge costs 100€). However, these strain gauges can be replaced by virtual ones, available on the zone of interest of the DIC (with a deformation level at least superior to 0.01%).
For the bogie test case, taking into account the non-visible areas, we can estimate that at least 50% of the strain gauges can be replaced by 4 cameras disposed in only one day. This represents on this case about 5000 € of savings for the 100 gauges present on the test, without also taking into account the time gained on the definition of the instrumentation plan. The bogie would also have been available more quickly for the test, since such an instrumentation very often takes several weeks. Finally, it is also necessary to count the cost of wastage for the faulty gauges (detachment, electrical problems) for about 5% of the glued gauges for this bogie. In all, this adds up to almost 10 000 € of possible savings on this test.
On the simulation side: obtaining a predictive model more quickly.
Validation of a simulation model from test results often takes time, and mostly to extract the simulation results at the locations where the measurement was made. In addition to the uncertainties related to these operations (was the gauge where I placed it in the simulation?), they are time consuming, since they have to be repeated for all the sensors present on the test.
Having a measurement of displacements and deformations on the mesh thus makes it possible to automate and to rationalize these operations, since the extraction is made at the same points, and makes it possible to calculate the model’s error which will have to be minimized. On the bogie case, it is 2 days of engineer’s time that can be immediately saved considering only the post-processing of the measurement results.
For the adjustment of the simulation, the engineer must then find the right set of parameters for the simulation to stick to the measurement: boundary conditions, material parameters, interface parameters … Here a trial and error approach is the rule, led by the user experience.
Having the model error, moreover on a complete field, makes it possible to give a lot of interest to sensitivity calculations on each of the parameters: one will then be able to distinguish which part of the error of the model is due to a poor estimate of a material parameter or the orientation of a boundary condition. If we go on to the automatization of the identification process, it is 3 days that can be won, not to mention the quality gains related to the reduction of the model error.
Finally, having rich measurements such as those obtained by DIC makes it possible to improve the simulation more directly: measurements can be used to create measured boundary conditions, which replace the “ideal” (or rather idealized) boundary conditions of the first simulation. These elements are often the source of many questions after the test, so it is a way of removing these points of uncertainty by directly applying the reality of the field in the simulation. The gain here is difficult to estimate, say it is the price of tranquility!