# Surface result plan

The resoning behind the construction of this plan is the following:

1.	Define the factors and the levels
2.	Build the design of experiment
3.	Input the test results
4.	Analyze the plan
5.	Optimize the different plan results if required

Three buttons are available in this window:
- Delete line: Eliminate a factor
- New: Reset the sheet
- Create a plan: Write the constructed plan in the spreadsheet in order to enter the test results

## Define the factors and the levels
The definition of the factors consists of defining for all factors:

![Plan en surface de réponse2](baseUrl/planexp/surfaceRep/Plan-en-surface-de-reponse2.jpg)


- a long name;
- a short name (3 characters) the system automatically uses A, B, C… But we recommend you create meaningful short names (example Time => tim);
- the levels being studied, the factors should be quantitative

## Construction of design of experiments
Once the factors have been defined, Ellistat will suggest different types of design of experiments allowing you to study the effects of these factors:

![Plan en surface de réponse3](baseUrl/planexp/surfaceRep/Plan-en-surface-de-reponse3.jpg)


Here are the advantages and disadvantages of these various plans:

|||||
|-|-|-|-|
|**Plans**|**Advantages**|**Disadvantages**|**Use example**|
|Centered composite|Most accurate coefficients Uniform variance over the field of study|A few more tests Factors must be quantitative|The goal is a highly predictive model with few coefficient errors|
|Hybrid|Accurate coefficients Fewer tests than the centered composite|Factors must be quantitative|The goal is to carry out fair coefficient prediction while still diminishing the number of tests|
|Box Behnken|Comprised of only -1, 0, and 1, so it can possibly be used with other qualitative factors|Low accuracy with quadratic coefficients|The goal is to obtain a fair coefficient prediction, there is also at least one qualitative factor at 3 levels. |
|Hoke|Plans almost saturated Minimal number of tests Comprised of only -1, 0, and 1, so it can possibly be used with qualitative factors|Low accuracy over all coefficients|We want to obtain an idea of the surface result but prioritize a low number of tests|
|Doehlert networks|Capable of evolving strategies: - either by moving within the field of study - or progressively adding factors Relatively low number of tests |Low accuracy over all coefficients|A lower number of tests is prioritized, we also want to be able to develop the field of study during the course of testing|


Once the plan has been constructed, click on "Write plan" and the tests relative to the defined experimental plan will be input in the spreadsheet. All you will need to do from then on is carry out the tests before being able to analyze them.

![Plan en surface de réponse4](baseUrl/planexp/surfaceRep/Plan-en-surface-de-reponse4.jpg)

User Guide

Surface result plan

Define the factors and the levels

Construction of design of experiments

A complete solution for Industry 4.0


Plans	Advantages	Disadvantages	Use example
Centered composite	Most accurate coefficients Uniform variance over the field of study	A few more tests Factors must be quantitative	The goal is a highly predictive model with few coefficient errors
Hybrid	Accurate coefficients Fewer tests than the centered composite	Factors must be quantitative	The goal is to carry out fair coefficient prediction while still diminishing the number of tests
Box Behnken	Comprised of only -1, 0, and 1, so it can possibly be used with other qualitative factors	Low accuracy with quadratic coefficients	The goal is to obtain a fair coefficient prediction, there is also at least one qualitative factor at 3 levels.
Hoke	Plans almost saturated Minimal number of tests Comprised of only -1, 0, and 1, so it can possibly be used with qualitative factors	Low accuracy over all coefficients	We want to obtain an idea of the surface result but prioritize a low number of tests
Doehlert networks	Capable of evolving strategies: - either by moving within the field of study - or progressively adding factors Relatively low number of tests	Low accuracy over all coefficients	A lower number of tests is prioritized, we also want to be able to develop the field of study during the course of testing