This paper addresses the necessity of refining standard triaxial testing methods for characterizing the mechanical behaviour of salt rocks. Triaxial testing is a key tool for determining the strength and deformation characteristics of rocks; however, existing standards often fail to account for the unique features of salts, such as their highly plastic behaviour, creep, temperature sensitivity, and defect-healing capability. The work highlights the critical importance of considering large strains and volumetric changes of specimens during testing, as this enables a more accurate representation of the behaviour of salt rocks, as this enables a more accurate representation of the behaviour of salt rocks. It is proposed that current standards be updated by incorporating well-established correction equations for geometry evolution and volumetric strain, as well as by adopting the Hencky strain measure. Experimental results obtained on natural salt rock specimens and salt-based geomaterials demonstrate significant errors in the evaluation of the stress-strain state when traditional data-processing methods are applied without accounting for the specific properties of salts. The analysis underscores the need to revise existing triaxial testing standards in line with the proposed approaches, thereby improving the accuracy and reproducibility of data that underpin geomechanical modelling and engineering design.

The article suggests using a combination of the modified Burgers model and the Mohr – Coulomb model with the degradation of the adhesion coefficient and the increase in the friction coefficient to determine the parameters of salt rocks. A comparative analysis of long-term laboratory tests and field observations in underground mine workings with the results obtained using a calculated model with certain parameters is carried out. The parameters of the Mohr – Coulomb model with the degradation of the adhesion coefficient and the increase in the friction coefficient were obtained from the statistically processed data of laboratory tests, and the parameters of the modified Burgers model were determined. Using numerical methods, virtual (computer) axisymmetric triaxial tests, both instantaneous and long-term, were performed on the basis of the proposed model with selected parameters. A model problem is solved for comparing the behavior of the model with the data of observation stations in underground mine workings obtained from borehole rod extensometers and contour deformation marks. The analytically obtained coefficients of the nonlinear viscous element of the modified Burgers model for all the analyzed salt rocks did not need to be corrected based on the monitoring results. At the same time, optimization was required for the viscoelastic element coefficients for all the considered rocks. The analysis of the model studies showed a satisfactory convergence with the data on the observation stations. The comparative analysis carried out on the models based on laboratory tests and observations in the workings indicates the correct determination of the parameters for salt rocks and the verification of the model in general.