The simulations compared in this paper (called henceforth ”Limitations”) are a substantial contribution to the research on constitutive modelling. The cyclic behaviour of sand is intricate and, therefore, a challenge to constitutive modelling. Here I present simulations obtained with a new version of barodesy published in 2021 [1]. Barodesy is derived from the fading memory of soil, and this version holds equally for sand and clay. The critical void ratio \(e_c\) is here given by an evolution equation, see Table 7. Its initial value, \(e_{c0}\), depends on the method of sample preparation. Developed in the time of corona shutdown, this new version, and in particular its calibration, can be further worked out. The version in Table 7 is slightly different than in [1].

As in previous constitutive models proposed by the author [2], the stress is here denoted by \(\mathbf {T}\), the mean effective stress is denoted as p, and the arrow on a tensor, \(\vec{\mathbf{A}}\), is replaced by the exponent 0: \(\mathbf {A}^0\).

The simulations of tests are obtained with the parameters for Karlsruhe sand given in Table 8, and the results are shown in Figs. 13, 14, 15, 16 and 17. Stresses are given in kPa. The initial void ratio \(e_{c0}\) depends on the sample preparation. A procedure to capture this dependence is still missing and, therefore, its values were determined by trial and error for each simulation and are shown in the related plots. \(e_{c0}\) has a considerable influence on the results.

The obtained results can be commented as follows:

Overshooting (”Limitation 1”)::

See Fig. 13. The strain cycles produce undershooting, and the curve approaches asymptotically the one for the monotonic experiment.

Cyclic mobility (”Limitation 2”)::

See Fig. 14. The butterfly-like part of the curve is, admittedly, reached too early.

Liquefaction with strain cycles (”Limitation 4”)::

See Fig. 15. The initial stress amplitude is smaller than in the experiment. In accordance with the experimental result, the stress amplitude approaches zero (i.e. almost complete liquefaction) with increasing number of cycles.

Oedometric cycles (”Limitation 6”)::

See Fig. 16. Unloading–reloading cycles lead nearly to the monotonic curve. The final settlement is realistic.

CU after CD triaxial test (”Limitation 7”)::

See Fig. 17. It is somehow striking that the experimental curves for drained unloading and undrained reloading more or less coincide. The lack of any deformation after the hydrostatic loading–unloading cycle is remarkable.

The coincidence of experimental results with simulations is considered as the only criterion for the quality of the latter. However, experimental results are also prone to enormous scatter, as known by comparing undrained test results obtained with pluviated and tampered samples. With respect to constitutive relations, this emphasizes the importance of capturing several observed patterns without focussing on exact values. Simplicity and easy implementation are crucial.

No datasets were produced during the current study.

Table 7 Equations of barodesy, cf. Table 5 of ”Limitations”
Full size table
Table 8 Parameters of barodesy, cf. Tables 1–4 of ”Limitations”
Full size table
Fig. 13
figure 1

Simulation of undrained cycles following drained compression with barodesy for Karlsruhe sand, cf. Fig. 2 of ”Limitations”. The monotonic test is shown with a thick line

Full size image
Fig. 14
figure 2

Simulation of undrained stress cycles obtained with barodesy for Karlsruhe sand, cf. Fig. 4 of ”Limitations”

Full size image
Fig. 15
figure 3

Simulation of undrained strain cycles obtained with barodesy for Karlsruhe sand, cf. Fig. 7 of ”Limitations”

Full size image
Fig. 16
figure 4

Simulation of oedometric cycles obtained with barodesy for Karlsruhe sand, cf. Fig. 11 of ”Limitations”. The monotonic test is plotted with a thick line

Full size image
Fig. 17
figure 5

Simulation of undrained triaxial tests following drained deformation, cf. Fig. 12 of ”Limitations”. Obtained with barodesy for Karlsruhe sand

Full size image