in your little example there it seems you purposefully dodge the tube. I assume because of the perspective issue, you'd have to have the bottom portion of the tube on one layer and the top portion on another, I actually can't think of an easy way to combat this.
also, what are the cell dimensions for the floors and walls? Its the only one I can't seem to find
Dimensions for floors and walls are both 32x32 pixels. The tileset itself is comprised of 37x23 cells.
Using two separate components on two distinct layers may become cumbersome in the long term. Instead, the most efficient approach I can come up with right now is employing a dynamic Z-Order system oriented around the Y-Axis and Object Origin.
Here's a quick example I cooked up:
