 Dear colleagues, my topic is about modeling of adiabatic shape bands formation, an attempt to describe by this method high explosive response to low velocity and mechanical insults. This is outline of my talk. First, we describe high explosive violent reaction problem, then propose basic mechanism of this phenomena, then introduce method, which is the most convenient to model and then show the results of simulation. Safety of different devices, which contain high explosive of propellant is determined in many aspects by chemical reactions initiated by impact on high explosive with different velocities. For high velocities of impactor, there are many different models that is used in engineering practice, but for low velocity of impactor or to which pressure in material is less than 1 gPa and duration of this process is about 10 to 100 microseconds. There are few of such models. Here on figure, some experimental techniques for testing high explosive response to low velocity mechanical insults are shown. Some researchers consider adiabatic shape bands formation during shear loading on high explosive as the main mechanism of the non-homogeneous high temperature heating, which leads to detonation. There are several phenomenological models for describing these processes. The main weakness of those models is the usage of a number of poorly defined parameters, high sensitivities to their variations and sensitivity to mesh size. For clarifying physical processes, we have to obtain the non-homogeneous pattern of high explosive heating explicitly. On this picture, different crystalline defects are shown which affect on elastoplastic behavior of material. These defects have a very different spatial and temporal scales and for their behavior, modeling different methods are used, such as molecular dynamics, discrete dislocation dynamics and finite element method. And there is a big problem of accounting for all these defects on macroscopic sample behavior. For accurate calculation of mutual interaction between high explosive grains, high explosive grains and plasticizer and to account the various plastic responses of individual grains, in particular dislocation dynamics inside grain, in modeling we have to choose a method which able to account those deformation processes. Three decades ago, the finite element method of crystal plasticity was proposed, with aim to model deformation process in metals and alloys, where anisotropic properties of material grains play a significant role in material response. Here we apply this method to consider deformation process in high explosive. In this method, deformation gradient decompose to factors, elastic and plastic. Plastic deformation rate is expressed in this form. Then in ILP we can use different terms for different plasticity mechanism. The dislocation and twinning terms are shown here. Crystal plasticity finite element method is usually used for modeling different metallurgical processes, where deformation rate is very low. It allows to consider thermal activation regime of dislocation motion only. But in our processes, deformation rate is from 10 power 4 to maybe 10 power 8 inverse seconds, so we have to consider for non-drug regime also. We use it in the form from this article. But for high explosive, such as HMX, it's octogen-based explosive. Dislocation motion parameters are completely lacking, and anisotropic elastic properties for HMX are different in several times in different articles. So we have to use a phenomenological constitutive model for dislocation motion, shear rate to receive at least qualitative results. We account for plastic dissipation and heating due to it, and we use the Arrhenius law for burn rate calculation. Here typical simulation setup for crystal plasticity finite element method are shown. First we mesh microscopic sample into a set of finite elements. Usually we use solid finite elements with 8 computational nodes. In each node we assign 10 randomly oriented grains. For each grain we applied a model described above. And then in order to obtain macroscopic deformation, we average a plastic response of all material grains in each computational node by simple iso strain method which assumes equal deformation of each grain. This model was implemented by using Dusseldorf advanced material simulation kit in LSDyna finite element code, and we simulated test experiment. HMX cylinder deformed under motion of its upper plane, which with some extent corresponds to experiment carried out in Russian Research Institute of Experimental Physics in Sarov. And the simulation result corresponds to experimental data. Here the temperature field and burn rate field are shown at force microsecond after applying upper plane motion. In conclusion, the crystal plasticity finite element method was used for direct mesoscopic modeling of shear bands generation in HE explosive with randomly oriented grains. We demonstrated that the process of heating and ignition in HMX is initiated by plastic deformation in shear bands, and the obtained ignition time is in good agreement with available experiment. Thank you for your attention. Mesoscopic, we use a constitutive equation like this in each grain. And in each grain then assumes two computational nodes in solid element. For grain this, it's about several microns in length. Sample is about one centimeter.