The “LookupTable3D” block approximates a non-linear function using interpolation on data provided as a "3D-table" with 3 inputs and an output.
The three-dimensional table is then defined by specifying three vectors (x, y, z) of size m, n, and k, and a "3D-table", a vector of size m*n*k. Output of the block is the approximation of the block input using the bilinear interpolation method.
Table data, a vector of size m*n*k is a full factorial DOE vector corresponding with 3 input vectors.
For example:
The attached 'sample_spreadsheet.xlsx' file has 3 input vectors (column1, column2, column3) and output in column4.
Input 1 (i1) has 3 unique input values
Input 2 (i2) has 2 unique input values
Input 3 (i3) has 3 unique input values
ini1 | 5 | 10 | 15 |
ini2 | 100 | 120 |
|
ini3 | 20 | 30 | 40 |
Then the full factorial DOE vector should look like:
ini1 | Ini2 | Ini3 |
5 | 100 | 20 |
10 | 100 | 20 |
15 | 100 | 20 |
5 | 120 | 20 |
10 | 120 | 20 |
15 | 120 | 20 |
5 | 100 | 30 |
10 | 100 | 30 |
15 | 100 | 30 |
5 | 120 | 30 |
10 | 120 | 30 |
15 | 120 | 30 |
5 | 100 | 40 |
10 | 100 | 40 |
15 | 100 | 40 |
5 | 120 | 40 |
10 | 120 | 40 |
15 | 120 | 40 |
To generate Full factorial DOE vector using OML functions in model initialization block in activate..
clc, clear, close all
%find the path of the scm file
modelpath = fileparts(bdeGetModelFilePath(bdeGetCurrentModel()))
%read xlsx file stored in the same location as activate file
excelfile = 'sample_spreadsheet.xlsx';
xlsmat = xlsread([modelpath '/' excelfile]);
% read table without headerlines
xlsmat = xlsmat(2:end,:);
%find unique values in the 1st,2nd,3rd column
ini1 = unique(xlsmat(:,1)); % input_1
ini2 = unique(xlsmat(:,2)); % input_2
ini3 = unique(xlsmat(:,3)); % input_3
% create full factorial DOE based on number of unique values in the inputs
de = fullfact([length(ini1),length(ini2),length(ini3)]);
%initialize ouput: of vector of length = length(itp) * length(is) * length(it) with zeros
output_table = zeros(length(de),1);
%create table data based on Full factorial DOE
for i = 1:length(de)
% finds row number of xlsmat which corresponds to input values
idx = find(ismember(xlsmat(:,1:3),[ini1(de(i,1)),ini2(de(i,2)),ini3(de(i,3))],'rows'));
%if empty fills with 0 if not takes the output from spreadsheet
if isempty(idx) == 1
output_table(i,1) = 0;
else
output_table(i,1) = xlsmat(idx,4);
end
end