Engmodel Tutorial
Location of Data Files
The sample files used in this tutorial are in /advisor/extras/engmodel/sample_files.  You will use two types of files:
  1. *.txt files:  Text files (generated by engine tests) that are processed by engmodel.  The first row is the headings for the columns, and the data is in columns.
  2. *.mat files:  The test data from the text files saved as a Matlab mat file with specific names for the variables to be processed.
Overview
In this tutorial, you will process raw data files and plot the results for an Mercedes 1.7L CI engine and then generate an ADVISOR engine file.  This engine is available in the ADVISOR library as FC_CI60_emis.m.  Note that the data set processed in this tutorial includes only a few of the data points used to generate the full ADVISOR map.

Text to be entered in edit boxes are indicate by bold typeface.  Button pushes on the Engmodel screens are indicated by red.  Filenames are italicized.

Process Raw Data Files
First, change directories to /advisor/extras/engmodel/sample_files.  If it is not already open, type >> engmodel at the Matlab command prompt.

Read in text file
Select Read in text file.  Select Select file to convert and choose the file named Mercedes1pt7_sample.txt.  The variables in the header automatically appear in the drop down list.  You need to change them to have the correct names recognized by engmodel.  With the first variable (fc_spd_RPM) selected, Modify name.  Type in fc_spd and hit OK.  Note that if you have extra columns in your text file, you can simply 'Delete from list.'

Either modify all the names in the header list to be the correct names, or Select file to convert and choose the file named Mercedes1pt7_good_headings.txt, which already has the correct header names.  Select *.mat file to save to and enter mercedes_17.mat.

Set Parameters/Process
Select Set Parameters/Process.  Select Name of mat file and choose the file named mercedes_17.mat.  The minimum/maximum speed and torque are automatically set from the limits of the test data.  You can change these if you want to extrapolate (through the chosen algorithm) beyond your data set.

Change the 'Speed increments (rpm)' to 200.  Change the 'Lower Heating Value (J/g)' to 42937.  Ensure that 'Polynomial Fit' is selected and the 'Plot results?' box is checked, and select Process.

Seven plots will appear:




Now try processing the file by extrapolating points based on the 'Nearest Neighbor' algorithm.  Select Set Parameters/Process.  Change the 'Speed increments (rpm)' to 200.  Change the 'Lower Heating Value (J/g)' to 42937.  Select 'Nearest Neighbor' and select Process.  Compare the efficiency map (below) to the previous map (above).  Note that within the range of the data, the map is unchanged, but in the areas of extrapolation, the contours have changed.

Generate an ADVISOR Model
Once you have processed the data, you can automatically create an ADVISOR engine file.

The data files used and accuracy/error data from the processing is automatically saved in the header.  You'll need to enter thermal parameters, including mass of the various engine components.  Default scalings based on the power of the engine are assumed if the information is unknown.

Select Create ADVISOR File.  Fill in the edit boxes as below:

Select Make ADVISOR File and you will create a new file FC_CI60_sample.m.  Now you're done and you can add this to your ADVISOR engine list!

Last revised: 1/26/01, vhj