Optimization at Home
Altair EmployeeFairly often at my house there is a unique furniture or shelving need that we decide to custom make. Now we don’t have any background in wood working or furniture construction and design, but we come up with an idea and work through it with plenty of design changes in the middle of the construction. Our latest project was a laundry sorter made from 1x2 lumber to hold 8 laundry baskets. We used 16 guage nails for most of the construction and it turns out the joints were not very stiff. So much so that there was fear of the whole structure collapsing when you pulled out a heavy basket.
So, we considered putting shear panels on it to stiffen it up, but we liked the open look and feel of it. I said we could just add some stiffening triangles on the sides. I had some ideas, but then it struck me that I know the best way to determine how to maximize the stiffness, or should I say, minimize compliance. OptiStruct! Then my wife asked me the critical question: "Will it take longer than an hour?" Now I had the design constraints and a deadline. I knew with only an hour I needed to make the model as simple as possible, so I decided to just model one side of the laundry sorter with shell elements. I made 1.5” of the front and back of the rectangle of shells non-design space, but left everything between them part of the design. I chose to do a topology optimization since I wanted distinct members that I could make from more 1x2s. I constrained the bottom of the front and back posts in DOF 1-4 and then put a force at the top corner acting horizontally. The objective function was minimized compliance and I constrained the volume fraction to 10% of the total volume.
The optimization converged quickly on a design that I thought made sense and that I could make. I installed the boards in the suggested pattern, and the deflection (not measured) went from inches to not noticeable until the back corner of the whole structure came off the ground.
As you can see below, I did ignore the one small member the optimization suggested in the bottom right corner understanding that topology results usually need some engineering judgement in the interpretation.
It was fun to use my work experience at home. I was inclined to tweak more parameters to get different ideas and see different versions, but with the allotted time I was very happy with the results and so was my wife!