What is DOE and what are the types of DOE

In the CDM project, the DOE's function is to carry out qualitative "validation" and quantitative "verification certification" of the CDM (Clean Development Mechanism) project.

DOE (Design of Experiments) plays a very important role in the whole process of quality control, which is an important guarantee for the improvement of our product quality and process flow. Through the quantitative analysis of product quality and process parameters, we can find the key factors and control the factors related to them.

In fact, DOE is very critical in the operation of a CDM project, as it directly determines whether a CDM project can be successfully registered, and whether and how much greenhouse gas emission reductions can be issued.

What is Design of Experiments (DOE):

Since Fisher used the Design of Experiment (DOE) method in agricultural production in the 20s of the 20th century, the experimental design method has been widely used and developed in agriculture, biology, genetics, engineering and other fields.

The experimental design mainly applies the basic knowledge of physical statistics, discusses how to reasonably arrange the experiment, obtain data, and then conduct comprehensive scientific analysis, so as to obtain the optimal combination scheme as soon as possible. In product design, the use of design of experiments can effectively design and verify the performance of products in the lowest test cost and in the shortest time; In the manufacturing process, design of experiments can be used to quickly find and optimize the process parameters that have a significant impact on the process output metrics from many influencing factors.

Uses of the experimental design:

1. Factorial analysis to identify which variables x have a significant impact on the response y;

2. Parameter optimization, when determining where the significant impact of x is set, so that y is almost always close to the expected value;

3. Reduce the variation and determine where the influential x is set, so that the variation of y can be minimized;

4. Robust design, when determining where the influential x is set, can minimize the effect of the uncontrollable variable u.

What are the types of Design of Experiments (DOEs)?

1) Factor design.

The purpose of factor design is to determine which independent variable x significantly affects the response variable y through experiments: for those independent variable x that does not significantly affect y, it should be deleted when establishing the correlation between y=f(x); Which independent variables, x, significantly affect y, should be retained.

According to its purpose, it is called "factor screening design". Since the purpose of this type of test is to be factor-specific, this type of experimental design is a "factorial design" or "factorial design".

In factor design, it can be divided into two levels of factor design, three level factor design and mixed level factor design according to the number of factor levels. It has been proved that in the factor design, the two-level orthogonal test method is the simplest and most effective.

In factor design, it can be divided into two categories: full factor experimental design and partial factor design.

2) Return to design.

The purpose of the regression design is to find out the relationship between the response variable y and the independent variable x, so as to further find out what value the independent variable x will make y reach the optimal value. Since the purpose of such trials is to address regression relationships, this experimental design is referred to as "regression design". The response surface method is commonly used in regression design, which is a method that takes the establishment of quadratic regression equations as the main research method.

3) Robust design.

Robustness refers to the ability of the process to resist interference, i.e., when the process is severely affected by factors that are difficult to control (or "noise"), the fluctuation or variability of the process output y should be small enough. In order to do this, try to choose a combination of control factors that make the system insensitive to noise changes, which is called "robust parametric design". In Japan, this type of design is often referred to as the "Taguchi parametric design method".

4) Mixing design.

If the question of formulation is discussed, then the question of the ratio of the individual components in the whole product is being studied, and it is clear that the sum of these ratios should be 100%. A design of experiments that investigates this type of problem is called a "mixture design".

5) Tune the operation.

The existing process conditions have basically met the requirements, but if you want to obtain better results, you can find a solution by slightly adjusting the original conditions, which is the problem solved by the "tuning operation" method.

[Experimental design]:

Design of Experiments (DOE) is also known as Design of Experiments. Since the 20s of the 20th century, Fisher (using the experimental design method in agricultural production), the experimental design method has been widely developed, and statisticians have found many very effective experimental design techniques. In the 50s of the 20th century, the Japanese statistician Kenichi Taguchi made the most widely used orthogonal design in experimental design, and made a well-known contribution to the wider use of experimental design in terms of method explanation.

The role of DOE]:

Design of experiments can play an important role in industrial production and engineering design, mainly including:

1. Increase yield;

2. Reduce the fluctuation of quality and improve the level of product quality;

3. Greatly shorten the test cycle of new products;

4. Reduce costs;

5. Experimental design to extend product life.

In industrial and agricultural production and scientific research, it is often necessary to do experiments in order to achieve the desired purpose. For example, in industrial and agricultural production, it is hoped that high quality, high yield, and low consumption can be achieved through experiments, especially new product tests, and there are many unknown things, so it is necessary to explore the process conditions or formulas through experiments. How to do experiments, there is a lot of knowledge.

If the experiment is well designed, it will get twice the result with half the effort, and vice versa, it will do more with half the effort, or even work without success.

Scientific experiments must be designed using the scientific method if they are to be carried out most effectively. The so-called statistical design of the experiment is the process of designing the experiment so that the collected data is suitable for statistical analysis and can draw valid and objective conclusions. If meaningful conclusions are to be drawn from the data, statistical methods are necessary for experimental design.

When the problem involves data that is affected by test errors, only statistical methods are objective methods of analysis. In this way, there are two sides to any trial problem: the design of the trial and the statistical analysis of the data.

The two topics are closely linked, as the analytical method is directly dependent on the design used.

There are 1 types of Design of Experiments (DOE).Full Factor DOE

2.Divisional DOE

Common DOE methods can be divided into two categories, one is orthogonal design of experiments and the other is factorial.

1) Orthogonal experimental design method.

Orthogonal experimental design is a scientific method to study and deal with multifactorial experiments. It makes use of a normalized ** - orthogonal tables.

Select the test conditions, arrange the test plan and the anti-field pants for testing, and find out the better production conditions, that is, the best or better test plan through a small number of tests. It is mainly used to investigate the influence of certain characteristics or multiple factors of a complex system (product, process) on some characteristics of a system (product, process), to identify the more influential factors in the system, the magnitude of their impact, and the possible interrelationship between the factors, so as to promote the design and development of the product and the optimization of the process, control or improvement of the existing product (or system).

2) Factorial method.

The cause method is also known as factorial test design, factorial test, etc. It is an effective way to study the effects of two or more factors that are changing. Many trials require the effects of two or more variables.

For example, several factors: the effect on product quality; effects on a certain machine; effects on the properties of a material; the effect on the combustion consumption of a certain process, and so on. The factors under study are tested one by one according to all combinations of all levels (levels) of all factors, which is called a factorial test, or a complete factorial test, referred to as the factorial method.

It is used for new product development, product or process improvement, and installation services, through a relatively small number of tests, to find a combination of high-quality, high-yield, and low-consumption factors to achieve the purpose of improvement.

What are the prerequisites that need to be in place for a DOE to be successful before a DOE can be designed?

It is necessary to ensure that the process of experimental research is stable and realistic. If the conditions are limited, if this cannot be done, you may wish to use randomization, blocking, replication and other methods to avoid it as much as possible.

Measuring systems must be reliably repeatable.

and reproducibility. Otherwise, none of the measured data can be trusted. The results of natural trials cannot be trusted.

Hello, there are four methods: 1. Factorial analysis, to identify which variables x have a significant impact on the response to the amount of tung grasp y;

2. Parameter optimization, when determining where the significant impact of the x setting is repentant, y can almost always be close to the expected value;

3. Reduce the variation and determine where the influential x is set, so that the variation of y can be minimized;

4. Robust design, when determining where the X setting that has an impact on local infiltration Qing, the effect of the uncontrollable variable U can be minimized.

Our traditional approach is to change only one variable in the same experiment with many input variables that affect the output, and the other variables are fixed.

Disadvantages of traditional methods: long test cycle, waste of time, high test cost; The experimental method is crude and cannot effectively assess the interaction between inputs.

The test method that can effectively overcome the above shortcomings is: DOE

What the DOE has achieved is a breakthrough improvement.

When planning a trial, it is important to study how to arrange the experiment in the most efficient way, which can effectively identify the impact of multiple input factors on the output.

During the experiment, the changes in the output were observed by making precise and systematic artificial adjustments to the selected input factors;

After the experiment, the results of the experiment were analyzed to obtain the most information, and the conclusion was drawn that "which independent variables x significantly affect the output y, and what values of these x's will make y reach the optimal value".

We used the regression analysis method to analyze the historical data in the first mu analysis stage, and obtained the regression equation with similar regressions, and obtained the relationship between y and each x. But the acquisition of this relationship is "passive", because we use existing and ready-made data, and we have almost no control over the scope of application, we cannot control the precision of the equation, and we can only be in a situation of "what is what".

We use the DOE method, in which the independent variables are often taken from values that have not been taken before, and are precisely controlled, and the problem to be studied is explored more broadly, with the aim of achieving breakthrough improvements.

The role of Design of Experiments (DOE) is to study and analyze the system by controlling and optimizing the experimental conditions. The purpose of the experimental design is to improve the efficiency of the experiment, to identify the key factors affecting the destruction of the cedarinium system, and to improve the performance of the control system. Through experimental design, researchers can make efficient use of limited resources, gather more valuable information, and can better understand and control the performance of the system.