A specific device for enhanced measurement of mechanical dissipation in specimens subjected to long‐term tensile tests in fatigue

Abstract

This paper presents a calorimetric approach to the measurement of mechanical dissipation in specimens subjected to cyclic tensile tests. Mechanical dissipation, that is, the heat power produced by the material due to mechanical irreversibility, can potentially be deduced from the temperature changes captured on the specimen surface by infrared thermography. However, a difficulty arises for long-term cyclic tests: Results are easily skewed by any change in the specimen's environment. The problem is amplified by the fact that mechanical dissipation is in general small compared to the heat sources associated with thermomechanical couplings, making its estimation difficult. The paper proposes a simple procedure to extract a well-resolved estimation of mechanical dissipation by solving two key points specific to long-term cyclic tests: (a) the reduction of the parasitic effects associated with changes in the specimen's environment by using a specific device based on two references samples and (b) the choice of relevant thermal data acquisition parameters. A test is performed on a copper-based shape-memory alloy whose calorific response comprises three origins of heat sources: thermoelastic coupling, phase transformation, and mechanical irreversibility. The results obtained demonstrate the relevancy of the approach in extracting mechanical dissipation from the thermal response of the specimen subjected to long-term tensile tests in fatigue.

Comparison between 0D and 1D approaches for mechanical dissipation measurement during fatigue tests

“…As the heat due to thermo‐elastic coupling is null over an integer number of mechanical cycles, the obtained heat source s at frequency f R = 1 Hz is equal to the mechanical dissipation averaged over n R = 60 mechanical cycles for copper and over n R = 2 mechanical cycles for PMMA. As demonstrated in Delpueyo et al, this property is actually true only if the temperature fluctuation is cyclic, that is to say, in the “steady‐state thermal regime.” Note, however, that the difference between the obtained heat source s and mechanical dissipation in the transient regime is low compared to the amplitude of the thermo‐elastic coupling heat sources. The next section presents the experimental assessment of the mechanical dissipation averaged over n R mechanical cycles, still named d 1 in the following for the sake of simplicity, for the two materials considered.…”

“…Let us consider a longitudinal thin plane specimen with constant cross section, placed in the jaws of a uniaxial testing machine (see Figure ). Two reference samples are used to track variations in the specimen's environment . Each of the references is clamped in one jaw of the machine to follow the temperature variations in the jaws.…”

“…See section 2 of Delpueyo et al for some properties of the heat exchange constants. It can be noted that τ 0D < τ 1D by definition.…”

“…The temperature acquisition conditions were such that they could enable us to directly extract the mechanical dissipation d 1 from the total heat source s . The camera acquisition frequency f A was set to 100 Hz, but each image recorded was the average of 100 successive frames, leading to a recording frequency f R of 1 Hz.…”

“…This uses an averaged temperature over the whole specimen's measurement zone. The reader is referred to Delpueyo et al for a list of references in which the 0D approach is employed. This list can be updated with Wang et al dedicated to pure copper, Le Cam dedicated to rubber, and Corvec et al dedicated to different types of leathers.…”