All 1 entries tagged Arrays

TAGUCHI EXPERIMENTATION

Today was the third day of the PIUSS module. We were introduced to Taguchi loss function and experimentation. It was the time to revisit an already known fact. It is important to get the design of the process and product right because design itself can save a lot of cost and improve quality. It is something we learnt in the plane making simulation. During making the plane, we all made a major mistake and focussed on improving himan training, but the actual focus should have been the process improvement. Today was the time to make helicopters fly in the air for as long as possible. We need to make sure that we do as few tests as possible in order to save experimentation costs. We used orthogonal arrays to identify the major control factors that will keep helicopter in the air for longer. This time my focus was on improving the design of the helicopter. We started off with considering some important factors. While designing the helicopter, I thought that the idea of making a streamlined tail is going to be a major factor and it will help the helicopter to stay in the air (I was again making an assumption, how can you conclude something without having data and analyzing the process? My bad). We listed 7 factors and calculated the Signal noise ratio for eight different prototypes. We got good results. But now arises the major question. We know that we have a good model. But did we really know why that particular model was a good model? It is a very important question to answer. Because there are a lot of factors that play a role in defining the design. The same thing I ahve experienced in the industry as well. Sometimes people end up making good products, but they don't know why it was good. Everything cannot be relied on hit and trial. So using the L8 array we found out the major and most significant factors.

I was surprised seeing the results. We had the best Signal-Noise ratio. But did we really know why it was so? Before the start of the process I assumed that streamlined tail design is going to be the difference. But data showed a whole different story. It was the use of transparent tapes that was a significant factor in increasing the air time. Tail design made a difference but not much. till is the power of scientific process and data interpretation. Now I imagine how do manufacturing companies back in my city operate. i am sure they don't collect data. I am sure that they somehow know why they achieved a particular result. After today's experience I can just think of how many wrong assumptions they might be making. It is important to follow the DMAIC process. The implementation will be interesting. Looking forward to it.

Saurabh Kukreja : 21 Nov 2018 21:25 |  Tags: Arrays Experiment Orthogonal Scientific |  Comments (0) |  Close comments |  Report a problem