We will here describe how the single geological events of a Noddy history are organised within pynoddy. We will then evaluate in some more detail how aspects of events can be adapted and their effect evaluated.
In [55]:
cd /Users/flow/git/pynoddy/docs/notebooks
In [56]:
from IPython.core.display import HTML
css_file = 'pynoddy.css'
HTML(open(css_file, "r").read())
Out[56]:
In the current set-up of pynoddy, we always start with a pre-defined Noddy history loaded from a file, and then change aspects of the history and the single events. The first step is therefore to load the history file and to extract the single geological events. This is done automatically as default when loading the history file into the History object:
In [57]:
import sys, os
import matplotlib.pyplot as plt
# adjust some settings for matplotlib
from matplotlib import rcParams
# print rcParams
rcParams['font.size'] = 15
# determine path of repository to set paths corretly below
os.chdir(os.getenv('HOME') + '/git/pynoddy/docs/notebooks/')# some basic module imports
repo_path = os.path.realpath('../..')
import pynoddy
In [58]:
# Change to sandbox directory to store results
os.chdir(os.path.join(repo_path, 'sandbox'))
# Path to exmaple directory in this repository
example_directory = os.path.join(repo_path,'examples')
# Compute noddy model for history file
history = 'simple_two_faults.his'
history_ori = os.path.join(example_directory, history)
output_name = 'noddy_out'
reload(pynoddy.history)
reload(pynoddy.events)
H1 = pynoddy.history.NoddyHistory(history_ori)
# Before we do anything else, let's actually define the cube size here to
# adjust the resolution for all subsequent examples
H1.change_cube_size(100)
# compute model - note: not strictly required, here just to ensure changed cube size
H1.write_history(history)
pynoddy.compute_model(history, output_name)
Events are stored in the object dictionary "events" (who would have thought), where the key corresponds to the position in the timeline:
In [59]:
H1.events
Out[59]:
We can see here that three events are defined in the history. Events are organised as objects themselves, containing all the relevant properties and information about the events. For example, the second fault event is defined as:
In [60]:
H1.events[3].properties
Out[60]:
In [73]:
reload(pynoddy.history)
reload(pynoddy.events)
H1 = pynoddy.history.NoddyHistory(history_ori)
# get the original dip of the fault
dip_ori = H1.events[3].properties['Dip']
# dip_ori1 = H1.events[2].properties['Dip']
# add 10 degrees to dip
add_dip = -10
dip_new = dip_ori + add_dip
# dip_new1 = dip_ori1 + add_dip
# and assign back to properties dictionary:
H1.events[3].properties['Dip'] = dip_new
# H1.events[2].properties['Dip'] = dip_new1
In [74]:
pwd
Out[74]:
In [75]:
H1.events[3]
Out[75]:
What is left now is to write the model back to a new history file, to recompute the model, and then visualise the output, as before, to compare the results:
In [76]:
reload(pynoddy.output)
new_history = "dip_changed"
new_output = "dip_changed_out"
H1.write_history(new_history)
pynoddy.compute_model(new_history, new_output)
# load output from both models
NO1 = pynoddy.output.NoddyOutput(output_name)
NO2 = pynoddy.output.NoddyOutput(new_output)
# create basic figure layout
fig = plt.figure(figsize = (15,5))
ax1 = fig.add_subplot(121)
ax2 = fig.add_subplot(122)
NO1.plot_section('y', position=0, ax = ax1, colorbar=False, title="Dip = %.0f" % dip_ori)
NO2.plot_section('y', position=0, ax = ax2, colorbar=False, title="Dip = %.0f" % dip_new)
plt.show()
In [77]:
pynoddy.compute_model(new_history, new_output)
The geological history is parameterised as single events in a timeline. Changing the order of events can be performed with two basic methods:
The first method is probably the most useful to test how a simple change in the order of events will effect the final geological model. We will use it here with our example to test how the model would change if the timing of the faults is swapped.
The method to swap two geological events is defined on the level of the history object:
In [79]:
reload(pynoddy.history)
reload(pynoddy.events)
H1 = pynoddy.history.NoddyHistory(history_ori)
H1.change_cube_size(100)
# compute model - note: not strictly required, here just to ensure changed cube size
H1.write_history(history)
pynoddy.compute_model(history, output_name)
In [80]:
# The names of the two fault events defined in the history file are:
print H1.events[2].name
print H1.events[3].name
In [81]:
# Now: swap the events:
H1.swap_events(2,3)
In [82]:
# And let's check if this is correctly relfected in the events order now:
print H1.events[2].name
print H1.events[3].name
Now let's create a new history file and evaluate the effect of the changed order in a cross section view:
In [83]:
new_history = "faults_changed_order.his"
new_output = "faults_out"
H1.write_history(new_history)
pynoddy.compute_model(new_history, new_output)
In [85]:
reload(pynoddy.output)
# Load and compare both models
NO1 = pynoddy.output.NoddyOutput(output_name)
NO2 = pynoddy.output.NoddyOutput(new_output)
# create basic figure layout
fig = plt.figure(figsize = (15,5))
ax1 = fig.add_subplot(121)
ax2 = fig.add_subplot(122)
NO1.plot_section('y', ax = ax1, colorbar=False, title="Model 1")
NO2.plot_section('y', ax = ax2, colorbar=False, title="Model 2")
plt.show()
Just as another quick example of a possible application of pynoddy to evaluate aspects that are not simply possible with, for example, the GUI version of Noddy itself. In the last example with the changed order of the faults, we might be interested to determine where in space this change had an effect. We can test this quite simply using the NoddyOutput objects.
The geology data is stored in the NoddyOutput.block attribute. To evaluate the difference between two models, we can therefore simply compute:
In [86]:
diff = (NO2.block - NO1.block)
And create a simple visualisation of the difference in a slice plot with:
In [88]:
fig = plt.figure()
ax = fig.add_subplot(111)
ax.imshow(diff[:,10,:].transpose(), interpolation='nearest')
Out[88]:
(Adding a meaningful title and axis labels to the plot is left to the reader as simple excercise :-) Future versions of pynoddy might provide an automatic implementation for this step...)
In [ ]: