Blog 3 - Design of Experiments (DOE)

Helloooo everyone and welcome back to my blog!! I hope you're doing well and are eager to learn together with me! Today's blog will be focusing on Design of Experiments, otherwise known as DOE! You may be wondering, what is DOE? Well, (DOE) is a systematic, efficient method that enables scientists and engineers to study the relationship between multiple input variables (aka factors) and key output variables (aka responses). Now that we've been introduced to DOE, let's see it in action!

🎉Exploring the Popcorn Mystery with Design of Experiments (DOE)! 🍿🔬

We all know the struggle of making microwave popcorn. You get excited, and start the microwave, but then... you open the bag, and there they are – those un-popped kernels. It's so frustrating! 😫 So, why does this happen? 🤔 Well, in this case study, we used the power of Design of Experiments (DOE) to figure out what causes the loss in popcorn yield. Let's dive into the data and uncover the mystery!

🤔 The Case Study

The experiment investigated how three factors influence the amount of "bullets" (un-popped kernels) in popcorn:

1. Diameter of the bowl 🥣 (10 cm vs 15 cm)

2. Microwaving time ⏲️ (4 minutes vs 6 minutes)

3. Microwave power ⚡ (75% vs 100%)

   
   

I performed 8 runs with 100 grams of corn in each experiment. The variable we measured was the number of un-popped kernels, which we recorded in grams. Here’s the table with the data (all the "44"s you see in the table were replaced by the last 2 digits of my admin number, so everyone's results are unique!!):

Run order	Diameter (A)	Time (B)	Power (C)	Bullets (grams)
1	15 cm	4 min	75%	3.44
2	10 cm	6 min	75%	2.44
3	10 cm	4 min	100%	0.74
4	15 cm	6 min	75%	1.44
5	15 cm	4 min	100%	0.95
6	15 cm	6 min	100%	0.32
7	10 cm	6 min	100%	0.44
8	10 cm	4 min	75%	3.12

Now, let's use DOE to analyze this data and find out which factors are influencing the popcorn yield! 📊

🔬 Full Factorial Analysis: Uncovering the Factors

A Full Factorial design means we look at every possible combination of the factors. There are 2 levels for each of the 3 factors, so we have 2 x 2 x 2 = 8 combinations (runs), just like our data!

📊 Single Factors and Their Ranking:

• Factor A (Diameter): The larger the diameter, the fewer the un-popped kernels! 🍽️ A diameter of 15 cm leads to fewer bullets than 10 cm.

• Factor B (Time): More microwaving time means fewer bullets! ⏲️ 6 minutes is clearly better than 4 minutes.

• Factor C (Power): Higher power means fewer bullets! ⚡ 100% power gives better results than 75%.

🌟 Interaction Effects:

When we check interactions, we see that:

• The combination of diameter and microwaving time affects the results. Larger bowls with more time lead to better popping! 🍿

• The interaction between power and time also plays a role. More power combined with more time reduces the bullets. ⚡⏲️

Conclusion for Full Factorial Analysis:

The diameter of the bowl and microwave power are the most significant factors in reducing the number of un-popped kernels, with microwaving time also playing an important role. More time and power are the key to better popcorn! 🎯

🧑‍🔬 Fractional Factorial Analysis: A Shortcut to Efficiency

Next, we used Fractional Factorial analysis, which selects a smaller set of experiments from the full factorial design, but still gives us valuable insights. We chose 4 experiments that are orthogonal (independent) to make things simpler.

Selected Runs:

• Run 1: A (+), B (-), C (-)

• Run 3: A (-), B (-), C (+)

• Run 6: A (+), B (+), C (+)

• Run 8: A (-), B (-), C (-)

📊 Single Factors and Their Ranking:

• Diameter (A): 15 cm diameter continues to give fewer bullets. 🍽️

• Microwaving time (B): 6 minutes is better, reducing the number of un-popped kernels.

• Power (C): 100% power beats 75%, reducing bullets.

🌟 Interaction Effects:

The fractional factorial analysis confirmed the same interaction effects as the full factorial analysis, especially the time and power interaction.

💬Conclusion for Fractional Factorial Analysis:

Even with fewer experiments, we still find that larger diameter, more time, and higher power are the keys to popping those kernels! 🎉

💭 Reflection: What I Learned About DOE

Working with Design of Experiments (DOE) has been an eye-opening experience! 🤓 I learned that you don’t need to test every possible combination of factors to get useful insights – sometimes, fewer experiments can still lead to strong conclusions. DOE helps us understand the impact of individual factors and their interactions, which is crucial for optimizing processes in the real world. 💡

It's a very systematic approach to solving problems, and it’s not just useful for popcorn – it’s applicable to almost any field, from engineering to food science! 🧑‍🍳 I’ve definitely gained a deeper appreciation for how data analysis and experimental design work hand-in-hand to make better decisions! 🎯

All right, that's all I have for today! I hope you enjoyed this learning experience as much as I did! I'll see you in my next and very last blog!!😁😁