Scribd
Upload
628 views

Uncertainty and Optimization Workflow On Imported Case

1) Petrel 2010.1 allows users to perform uncertainty, sensitivity, optimization, and proxy workflows directly on imported cases without converting them to Petrel cases. 2) A sensitivity study example is shown where an Eclipse deck is loaded and an objective function is defined, then uncertainty and optimization is run on the imported case while varying parameters. 3) Results can be analyzed using tornado plots to show the relative influence of parameters on the objective function, and variable by variable plots to show relationships between objectives and parameters.

Uploaded by

Muhamad Afiq Rosnan
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
628 views

Uncertainty and Optimization Workflow On Imported Case

1) Petrel 2010.1 allows users to perform uncertainty, sensitivity, optimization, and proxy workflows directly on imported cases without converting them to Petrel cases. 2) A sensitivity study example is shown where an Eclipse deck is loaded and an objective function is defined, then uncertainty and optimization is run on the imported case while varying parameters. 3) Results can be analyzed using tornado plots to show the relative influence of parameters on the objective function, and variable by variable plots to show relationships between objectives and parameters.

Uploaded by

Muhamad Afiq Rosnan
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Schlumberger Information Solutions

Petrel 2010.1: Uncertainty and Optimization


Workflow on Imported Case
In Petrel 2010.1 users can perform Uncertainty, Sensitivity, Optimization and Proxy workflows directly on an imported
case without having to convert to a Petrel case. Below is a sensitivity study workflow example of an imported Thermal
case.
Step 1:
Load ECLIPSE deck to Petrel. You can also define objective function on the imported case by using the new “Define
objective function” process. For example, if you want to determine the relative influence of the uncertain parameters on
your objective function(s), such as Oil recovery factor or Net present value.

Open the “Define objective function” process from the


processes tree, drop in the imported case and select the
objective function of interest.

www.slb.com/sis
Schlumberger Information Solutions

Step 2: Open the Uncertainty and Optimization process and drop in the imported case. This example shows sensitivity
by variable task, you can also do the same for the other available tasks in the U &O process.

Note that if you did not define an objective


jective function on the case, you will
only have the first two commands in the workflow, that is, “Run
simulation case” and “Wait for case command”
If the Copy option is not ticked on, you will get an error message. This is because if it is not
ticked on the $variable in the keyword deck will not be substituted.

Step 3: Double click on Run simulation case >> Open the Editor >> Define your variables of interest by opening the
keywords containing the parameters.

www.slb.com/sis
Schlumberger Information Solutions

4: After clicking okay or Apply on the Run Simulation case dialogue, all the variables defined via the keyword editor will
show up on the Base case panel next to the “Run simulation case” command. You can then open the “Variables” tab to
specify the base case values and the arguments. Specify the other parameters as required, including parameters on the
sensitivity tab. Test and Run the workflow.

Note: The “free memory every” option is also new in 2010.1; this automatically frees unused memory after every nth
runs as specified by the user.

Step 5: Run the workflow and analyze your results


Petrel 2010.1 also provides new ways of viewing sensitivity analysis results.
• Viewing results on Tornado plot

www.slb.com/sis
Schlumberger Information Solutions

The Tornado plot shows the relative influence of each of the variables to the chosen summary vector, objective
function or volumetric results. In this case, the Total Oil Production is chosen as the summary
summar vector and the
Tornado plot shows that the Oil Viscosity of heaviest component (HC21B) is the most influential parameter.

Cumulative Tornado plot


shows the stack of the
sensitivity of each variable

• Variable by Variable plot


On the function window, you can
also plot variables against variable
values or objective functions
against variable values. This plot
shows the relationship between two
t
objective functions (Recovery factor
and Total oil Production)
roduction) against
the maximum temperature of the
wellbore heater.

www.slb.com/sis

You might also like