Home / Monte Carlo / Model Building / Create Design Sheet
Create Design (IPO) Sheet¶
From Excel click...
QXL Monte Carlo Tab > Create/Modify Design Sheet > Create Design Sheet
Quantum XL supports several forms of model building including Free Form Model Building, Create Design (IPO) Sheets, Import from DOE Pro, Merge Designs, and Visual Merge. Design Sheets are also called "legacy mode" as older versions of Quantum XL did not support free form models. This topic describes the Creating Design (IPO) Sheet.
Creating a Design Sheet is a four-step process.
Step #1: Define the number of inputs and outputs.¶
Step #2. Define the distribution and parameters of the inputs.¶
You may define inputs from the Normal, Exponential, Uniform, or Empirical data distributions. For each empirical distribution you have defined in the active workbook, another entry in the "Distribution" drop down box will be included. Therefore, you must define the empirical distribution before you create the Design Sheet. If you would like to create a model using all of the distributions available to Quantum XL, then create the model using free form mode.
When you have defined all of the inputs, a new worksheet will be added to the active workbook.
Step #3: Define the equation for each of the outputs.¶
For each output, you must define an equation in the column entitled "Function" under "Process Outputs". The variables for each output must reference the "Exper" column under "Process Inputs". If you define the output using other cells, the math parser will not be able to analyze the equation. The example below shows the equation for Output1 defined as Temp * Press.
You can include constants such as =3.14*F6*F7.
You can also include mathematical functions such as =3.14*F6*cos(F7).
Quantum XL does not use Excel's calculation engine to perform the Monte Carlo analysis. A proprietary math parser is used that is 450 times faster than Excel. Without the math parser, the time to compute the number of simulations required would be excessive. While the math parser is much faster, it does have limitations. Quantum XL does not support all the mathematical functions that Excel does.
List of all the mathematical functions Quantum XL supports
Step #4: Assign specification limits to the outputs.¶
At least one specification (spec) limit is required to perform Tolerance Allocation, Parameter Design, or Sensitivity Analysis. You can perform Expected Value Analysis without a spec limit.