# Generalized Anisotropic Material

## 1.Introduction

This tutorial describes how to simulate an anisotropic material in Slide2. There are actually five different ways to do this, but the emphasis of this tutorial will be on using the Generalized Anisotropic option, which allows you to specify different material types in different directions. The tutorial will also explain how to perform a probabilistic analysis with this type of material.

The finished product of this tutorial can be found in the *Tutorial 20 Generalized Anisotropic.slmd* data file. All tutorial files installed with Slide2 can be accessed by selecting **File > Recent Folders > Tutorials Folder** from the Slide2 main menu.

## 2.Model

For this tutorial, we will read in a file.

Select **File > Recent Folders > Tutorials Folder** from the Slide2 main menu, and open the *Tutorial 20 Generalized Anisotropic start.slmd* file.

**MATERIAL PROPERTIES**

For the slope material, we will use the **Generalized Anisotropic** option. This allows you to specify different material types in different directions. (Note: there are other anisotropic material models available in Slide2, see the next section for information).

Before we set up the Generalized Anisotropic material, let’s look at the sub-materials that make up the generalized material. For this example, we will assume the soil has weak, subhorizontal bedding. Two materials have been set up: one that represents the direction parallel to bedding and one that represents all other orientations.

Select **Define Materials** from the **Properties** menu.

Look at the **Soil Mass** and **Bedding materials**. So in this case, both sub-materials use **Mohr-Coulomb** strength, but have different values of cohesion and friction angle. These two materials will be assigned to a Generalized Anisotropic material as described below.

Click on **Material 3.** Change the **Strength Type** to **Generalized Anisotropic.**

Under Strength Parameters, click the New button. You will see a dialog in which you can specify different materials over different angular ranges. We want the Bedding material to be active within plus/minus 10° of horizontal and the Soil Mass to be active in all other directions. Angles in the dialog are measured from horizontal, so 90° represents vertical.

Enter −10 in the Angle To column. Leave the Material as Soil Mass. In the next row, enter 10 in the **Angle To** column and change the **Material **to Bedding. Finally, in the next column, set the **Angle To = 90** and leave the **Material** as **Soil Mass**. Select **OK**.

Material 3 is now assigned the Generalized Function ‘User Defined 1’.

Click **OK** to close the Define Material Properties Dialog.

**ANISOTROPIC MATERIAL MODELS IN SLIDE2**

There are actually five different anisotropic material strength models available in Slide2. These are:

• Anisotropic Strength

• Anisotropic Function

• Anisotropic Linear

• Generalized Anisotropic

• Modified Anisotropic Linear

For this tutorial, you could create the same model using the **Anisotropic Function** strength type, which allows you to define discrete angular ranges with Mohr-Coulomb properties. However, the **Generalized Anisotropic** strength type offers the following advantages:

• Any material type can be assigned to each angular range. All of the materials do not have to have a Mohr-Coulomb failure criterion. You can mix and match any material types you wish (e.g. Hoek-Brown and Mohr-Coulomb).

• The Generalized Anisotropic strength type allows for probabilistic analysis whereas the Anisotropic Function strength type does not. This will be explored further later in the tutorial.

See the Slide2 Help system for details about the different anisotropic strength models which are available.

### ASSIGN MATERIALS

By default, the slope is assigned **Material 1 (Soil Mass)**. We need to set the slope material to the **Generalized Anisotropic** material that you defined above (Material 3). Right click inside the model and select **Assign Material → Generalized Anisotropic**. The model will look like this.

## 3.Compute

Save the model using the **Save As** option in the **File** menu. Choose **Compute** from the **Analysis** menu to perform the analysis and choose Interpret from the Analysis menu to view the results.

## 4.Interpret

The Interpret program shows the results of the Bishop Simplified analysis by default. You can see that the factor of safety is about 1.48.

The limit equilibrium method calculates the stability of each possible failure surface by dividing up the circular area into slices and comparing the shear stress and strength on the base of each slice. With the Generalized Anisotropic material model, the angle of the base of each slice determines which material is used to calculate the strength.

Go to **Query → Query Slice Data**. Click on a slice about half way down the slope. In the Slice Data dialog, scroll down to the bottom so that you can see the **Base Material**. For this slice it should be 1 (Soil Mass).

Now click on a slice near the toe of the slope. In this region, the base of each slice is almost horizontal, so the base material is Material 2 (Bedding). This demonstrates how the applied strength model depends on the orientation of the slice base, for the Generalized Anisotropic model.

If you scroll back up in the Slice Data dialog, you will see that the base friction and cohesion reflect the values entered for Material 2 (c = 0 and φ = 20°).

Now right click on the slip surface and select Graph Query from the popup menu. In the Graph Slice Data dialog choose Primary Data = Base Cohesion. You should see the following plot, which shows the Bedding cohesion (=0) for the low base angle slices near the toe, and the soil mass cohesion (=5) for the higher base angle slices.

## 5.Non-Circular Failure Surface

Because of the weak bedding plane, it is likely that portions of the failure surface would tend to follow the bedding in a sub-horizontal direction. By forcing a circular failure surface, we are probably over-estimating the factor of safety. We can easily test this by specifying a non-circular failure surface and observing the results.

Go back to the Slide2 Model program. Select **Surfaces → Surface Options**. Under **Surface Type**, choose **Non-Circular**. Under **Search Method**, choose **Auto Refine Search**. This method will automatically search for the critical non-circular surface. See the Help system for more information.

Click **OK** to close the dialog.

## 6.Compute

Save the model using the **Save As** option in the **File** menu. Choose **Compute** from the **Analysis** menu to perform the analysis. This computation will be longer than the previous one since the program is searching for more potential surfaces and optimizing each surface.

Choose **Interpret** from the **Analysis** menu to view the results.

## 7.Interpret

With non-circular surfaces, all of the centers of rotation are displayed by default. You will need to zoom in on the slope to see the critical failure surface.

You can see that the factor of safety is about 1.26. Notably less than the value of 1.48 calculated assuming a circular surface. It is also interesting to observe the shape of the critical surface – a section of sub-horizontal slip connected to the ground surface by a steep incline. If you query the slices that make up the sub-horizontal section, you will see that the base material for each slice is Material 2 (Bedding).

This shows the importance of using a non-circular failure surface in anisotropic models since the failure surface ‘seeks out’ the weak bedding orientation to yield a lower factor of safety.

## 8.Probabilistic Analysis

Go back to the Slide2 Model program. We now assume that the friction angle of the bedding orientation is not well known, and we will determine the probability of failure for a given distribution of friction angles for the bedding.

Open the **Project Settings** dialog from the **Analysis** menu. On the left side click on **Statistics**. Check the box for **Probabilistic Analysis**. Leave the analysis type as **Global Minimum**. This will find a critical surface deterministically and then will calculate the probability of failure using this surface with varying material properties. To re-compute the critical failure surface for each randomization of material properties you could choose Overall Slope. Since this takes a longer time to compute, it will be left as an additional exercise

Click OK to close the dialog.

We are now going to define the statistical distribution of the strength of the bedding layer. Go to Statistics → Materials. We are going to vary the Bedding strength so click on Bedding, then click on the Add button. The cohesion of the bedding is 0, so we will only alter the friction angle. Check the box for Phi.

Click Next for **Statistical Distribution**, choose **Normal.**

Now click Finish. You now need to enter the MEan and STandard Deviation for the distribution. The Mean is automatically set to deterministic value from the previous analysis (20o), so we only need to set the standard deciation. Enter 5. Now you can automatically set the maximum and minimum to 3 standard deviations by clicking the 3x Std Deviation button on the right. The dialog should look like this:

Click OK to close the dialog.

## 9.Compute

Save the model using the Save option in the File menu.

Choose Compute from the Analysis menu to perform the analysis and choose Interpret from the Analysis menu to view the results.

10.Interpret

You will now see the deterministic centers of rotation and global minimum failure surface along with some statistical data (you may need to zoom in).

You can see that the mean factor of safety (1.283) and deterministic factor of safety (1.280) are nearly the same, and that the probability of failure (PF) is 8.3%.

You can look at the distribution of safety factors by going to Statistics → Histogram Plot. Set the Data to Plot = Factor of Safety – spencer. Select the Highlight Data checkbox. As the highlight criterion, select “Factor of Safety – spencer” and set the criterion to < 1.

Select the Plot button, and the **Histogram** will be generated as shown.

You can see a normal distribution of safety factors with about 10% of the area shown in red (safety factor less than 1). Because the Latin Hypercube method samples the input distributions smoothly (compared to Monte Carlo), the output in this case is also a relatively smooth normal distribution.

Right-click on the plot and select Change Plot Data. Plot the Bedding friction angle.

As you can see from this plot, the low safety factors (< 1) correspond to low sampled values of the bedding friction angle. This concludes the tutorial.