Blog 4: Design Of Experiment (DOE) Blog Entryš©āšØšÆ
- Valarie Goh
- Jan 21, 2023
- 8 min read
Updated: Feb 4, 2023
Yo sup everyone! š¤ I am back with another blog for CPDD woohoo!š„³ So today I will be sharingš my journey in practical 4 about the Design of Experiments with you all. So let's kick off with the case study of the Design of Experiment session experienceš¤ and I hope yall will learn something from it woohoo! š
There are 2 tasks that I will be explaining on this blog pageš§:
Determine the effect of single factors and their ranking.
Determine the interaction effects.
Include all tables and graphs both as pictures and as excel files (hyperlink to google drive or OneDrive)
Include the conclusion of the data analysis for full factorial data analysis
Determine the effect of single factors and their ranking.
Include all tables and graphs both as pictures and as excel files (hyperlink to google drive or OneDrive)
Include the conclusion of the data analysis for fractional factorial data analysis
Personal learning reflection on this Design Of Experiment learning experiences.
Tutorial Lesson
Practical Lesson
CASE STUDY
What could be simpler than making microwave popcorn? Unfortunately, as everyone who has ever made popcorn knows, itās nearly impossible to get every kernel of corn to pop. Often a considerable number of inedible ābulletsā (un-popped kernels) remain at the bottom of the bag.
What causes this loss of popcorn yield? In this case study, three factors were identified:
1. Diameter of bowls to contain the corn, 10 cm, and 15 cm
2. Microwaving time, 4 minutes and 6 minutes
3. Power setting of microwave, 75%, and 100%
8 runs were performed with 100 grams of corn used in every experiment and the measured variable is the amount of ābulletsā formed in grams and data collected are shown below:
Factor A= diameter
Factor B= microwaving time
Factor C= power
Run Order | A | B | C | Bullets (grams) |
|---|---|---|---|---|
1 | + | - | - | 3.61 |
2 | - | + | - | 2.61 |
3 | -- | - | + | 0.74 |
4 | + | + | - | 1.61 |
5 | + | - | + | 0.95 |
6 | + | + | + | 0.32 |
7 | - | + | + | 0.61 |
8 | - | - | - | 3.12 |
Performed the FULL FACTORIAL data analysis to solve the case study.
Determine the effect of single factors and their ranking.
The quantity of unpopped popcorn kernels, or "bullets," that stay at the bottom of the bag of popcorn is most significantly influenced by component C, followed by factor B and factor A.
Based on the gradients of each of the factor's linear graphs, the factor with a gradient of a higher magnitude translates to their higher significance in impact regarding the experimental results. Therefore, these gradients can be obtained from the linear equations displayed on the graph or the calculation from the total average difference for each factor from the excel table
The gradient of factor A: 0.435
The gradient of factor B: -0.515
The gradient of factor C: -0.89
It can be observed the gradients of 2 out of 3 factors are negative. This can mean that for all 3 of the factors, a HIGH(+) value will result in a lower amount of bullet which leads to a higher yield of edible popcorn.
Determine the interaction effects.
Interaction Effect (A x B)
At LOW B,
Average of LOW A = (0.74 + 3.12) / 2 = 1.93
Average of HIGH A = (3.61 + 0.95) / 2 = 2.28
Total effect of A = 2.28 - 1.93 = 0.35 (Increase)
At HIGH B,
Average of LOW A = (2.61+ 0.61) / 2 = 1.61
Average of HIGH A = (1.61 + 0.32) / 2 = 0.965
Total effect of A = 0.965 - 1.61 = - 0.645 (Decrease)
Interaction Effect (A x C)
At LOW C,
Average of LOW A = (2.61 + 3.12) / 2 = 2.865
Average of HIGH A = (3.61 + 1.61) / 2 = 2.61
Total effect of A = 2.61 - 2.865 = - 0.255 (Decrease)
At HIGH C,
Average of LOW A = (0.74 + 0.61) / 2 = 0.675
Average of HIGH A = (0.95 + 0.32) / 2 = 0.635
Total effect of A = 0.635 - 0.675 = -0.04 (Decrease)
Interaction Effect (B x C)
At LOW C,
Average of LOW B = (3.61 + 3.12) / 2 = 3.365
Average of HIGH B = (2.61 + 1.61) / 2 = 2.11
Total effect of B = 2.11 - 3.365 = -1.255 (Decrease)
At HIGH C,
Average of LOW B = (0.74 + 0.95) / 2 = 0.845
Average of HIGH B = (0.32 + 0.61) / 2 = 0.465
Total effect of B = 0.465 - 0.845 = - 0.38 (Decrease)
Include all tables and graphs both as pictures and as excel files (hyperlink to google drive or OneDrive)
Tables:
Graphs:
Interaction Effect (A x B)
Interaction Effect (A x C)
Interaction Effect (B x C)
Full Factorial
Include the conclusion of the data analysis for full factorial data analysis
The number of bullets is most significantly impacted by factor C (microwave power setting). Factor A (diameter of the bowl used to hold the corn) has the least significant effect on the number of bullets, whereas factor B (microwaving duration) has a substantial effect.
Nevertheless, it is established that an increase in each element will lead to a decrease in the number of bullets and, as a result, an increase in popcorn yield, despite the fact that each factor's influence on the number of bullets varies.
As a result, all the factors should be on the + (HIGH) side to achieve an optimal popcorn yield. Furthermore, there is a strong interaction between factors A and B. The same applies to factors A and C. Nevertheless, a less significant interaction exists between variables B and C.
Performed the FRACTIONAL FACTORIAL data analysis by selecting 4 experiments from the full factorial data that are orthogonal to solve the case study.
Determine the effect of single factors and their ranking.
Four experimental runs with an equal amount of - (LOW) and + (HIGH) data for each factor are to be chosen to begin the fractional factorial data analysis. As a result, the data have a balanced design and are orthogonal, giving them acceptable statistical features. Runs 2, 3, 4, and 5 were chosen as a result.
The quantity of unpopped popcorn kernels, or "bullets," that stay at the bottom of the bag of popcorn is significantly influenced by factors B and C, with factor A coming in third. This may be deduced from the gradients of each factor's linear graphs because a factor with a higher gradient is more significant in terms of its influence on the experimental outcomes. The linear equations shown on the graph can be used to derive these gradients.
The gradient of factor A: -0.1975
The gradient of factor B: 0.635
The gradient of factor C: -0.64
Because the gradients of factors B and C are of equal size, their effects on the number of bullets are equally important. Factor A has the lowest degree of importance since it has the shortest gradient. The gradients of components A and C are then shown to be negative. This indicates that a + (HIGH) number will produce fewer bullets and, as a result, a larger output of edible popcorn. Contrarily, factor B has a positive gradient. This indicates that a + (HIGH) number will produce more bullets and, as a result, a lesser yield of edible popcorn.
Include all tables and graphs both as pictures and as excel files (hyperlink to google drive or OneDrive)
Tables:
Graphs:
Include the conclusion of the data analysis for fractional factorial data analysis.
Factor A (Diameter of the bowl used to contain the corn) has the least significant effect on the number of bullets, whereas factors B (Microwaving duration) and C (Power setting of the microwave) have the most effects.
The decrease in bullets and yield of edible popcorn can be seen to follow an increase in factors A and C. An increase in component B will lead to more bullets and a decrease in the amount of edible popcorn.
Therefore, factor A and C should be on the + (HIGH) side, while factor B should be on the - (LOW) side, to produce the best output of popcorn. This goes against the result of the entire factorial data analysis, which states that for a higher yield of edible popcorn, all factors should be on the plus (HIGH) side.
This is due to the "less than full" nature of fractional factorial data analysis. There is a chance of missing information, yet it is more effective and efficient with resources. In order to still offer sufficient data to assess the factor effect, fewer treatments than all those that may be used are selected.
Personal learning reflection on this Design Of Experiment learning experiences.
Tutorial Lesson(Week 13)
During the tutorial session, I did not come for the lesson as I was not feeling well on that day. Regardless, I went on to bright space to check out this topic on the Design of Experiments I read through the whole slides and I understand how every experiment has its own experimental parameters and runs planned through DOE. One of the similar DOE we encounter was back when we did our experiment on the parameters affecting the leaching of coffee solubles in CP5202, Lab and Process Skills 2.
To carry out DOE, you have to use a lot of excel like inputting the values, and parameters and plotting the required graphs. This does not have many difficulties for most of my classmates as we have been through the Excel Day during year 1 taught by our seniors on how to plot and navigate the shortcut system.
Thus, it is clear that the knowledge and abilities we have acquired in the past will be of tremendous assistance to and a wonderful guide for us in our future academic endeavors. Without these abilities, it would have been extremely difficult for us to comprehend DOE.
Learning about DOE also enables us to comprehend how to design an experiment for research with changing parameter values. I had always believed that settings were changed at random to achieve the best experimental outcome. Without understanding DOE, I would never have understood how such experiments are planned or how to vary parameters so that an orthogonal balanced design is carried out with equal - (LOW) and + (HIGH) values of each parameter.
While we develop tests in the future, especially when we are prototyping our CA2 product (my group's invention is an exhaust fan for soldering), this newly acquired understanding of DOE will be quite helpful.
Practical Lesson(Week 13)
During the practical lesson, my classmates and I were given a job to do during the practical session: conduct an experiment. In this experiment, a catapult and ball were used to measure the ball's travel distance using three different parameters at - (LOW) and + (HIGH) values. The three variables are the ball's weight, the catapult arm's length, and the angle at which the arm stops. We performed 8 runs, using a combination of - (LOW) and + (HIGH) parameters in each run. The precision of the experimental results was increased by replicating each run eight times and averaging the results. Following the collection of this information, we used DOE to establish the order of the most important factors affecting the experimental results by plotting the required graphs.
Afterward, we had a fun and interesting activity for our practical session, The task of the activity was to knock out 3 standees of our lecturers that were placed at different locations. Each standee with 3 tries per round which total has 4 rounds each.
At first, my group was thought to depend on luck to win, and guess what we won second place as we saw how the first team's strategies and see how far the ball depend on the distance we press and the angle as well. We used their method and it turns out it works we were surprised and shocked that it works. Most of us shot the first target which is Mr. Ting first as he has the biggest shape among the rest and is easier to shoot. At the second trial, my teammate Asraf manage to shoot 3 lecturers in one round which shocks most of us. Nevertheless, we have fun and were awesome in doing this shooting.
Through this activity, we discovered that our ability to win or lose is not predetermined and can be altered based on our work and preparation. We would have little chance of winning, therefore we should not overlook the preparation of the activity.
Therefore, as you can see I reach the end of my blogš¤£ I hope you enjoy my blogš and remember to likeā¤ļøļø, shareš¤ and subscribeš¤³. Watchš my other cool blogsš± categoriesš¤© have a great weekend aheadš and all the best for your TST 1 CEDC to my SP DCHE classmates and friendsš¤š¤. I will be back strongeršŖ and better blog š¤see ya on my next blogš!!!
Comments