Modifications

benchmarks/newton_solver_benchmark_set/nonlinear_channel_flow/README.md

@@ -5,11 +5,10 @@ can be used to recreate the nonlinear channel flow figures from
 {cite:t}`fraters:etal:2019`.
 Start by compiling the `simple_nonlinear` plugin in this directory as described (at the example of another benchmark)
 in {ref}`sec:benchmark-run` .
-After compiling the plugin adjust `metabash.sh` and/or `bash.sh` to reflect
-the parameter search you are interested in and run the script.
-Running the `bash.sh` script without changes will recreate the model runs of Fig. 1 and Fig. 2 in {cite:t}`fraters:etal:2019`
+After compiling the plugin you can run (or adjust) the script `run.sh`, which will run all the models necessary
+to recreate Fig. 1 and Fig. 2 of {cite:t}`fraters:etal:2019`
 (by default the script will assume your processor has 4 compute cores, but you can adjust the number in the script).
-After running the model series you can use the gnuplot files in `plotfiles/` to recreate the following figures.
+After running the model series you can use the gnuplot files `plot_v.gnuplot` and `plot_t.gnuplot` to recreate the following figures.
 
 ```{figure-md} fig:benchmark-newton-nonlinear-channel-flow-tractions
 <img src="doc/figure_t.png" />
@@ -24,3 +23,7 @@ Convergence history for several methods for a rheology with n = 3 where in- and
 ```
 
 The nonlinear convergence behavior is shown in {numref}`fig:benchmark-newton-nonlinear-channel-flow-tractions` and {numref}`fig:benchmark-newton-nonlinear-channel-flow-velocities`.
+
+Note, that the linear residual is stalling in some cases. This is related to the choice of boundary velocity (in {numref}`fig:benchmark-newton-nonlinear-channel-flow-velocities`)
+and boundary traction (in {numref}`fig:benchmark-newton-nonlinear-channel-flow-tractions`). You can adjust the base input files `input_t.prm` and `input_v.prm` to
+test what happens when varying the boundary tractions or velocities.