I'm currently working as the technical owner of four hover motors on the Zipline aircraft. I am working on mechanical design and validation in a continuing effort to make the motors cheaper and more reliable. This project page will outline the root causing and mitigations for a structural failure found during testing.
(This project is under a NDA with Zipline Inc.)
Sector: Aerospace
Software: NX, FEMAP
Manufacturing: CNC Mill, Die Cast, Injection Molding
Project Time: Ongoing
A failure occured on the housing of the motor due to fatigue loading. Cyclic loads generated from the propeller led to crack formation and evantually failure. An overview of the major loads are shown on the left.
Shaft FBD
Shaft + Housing FBD
The root cause investigation started by considering all possible contributing factors and then working to eliminate or prove those factors through testing as quickly as possible. Root causes that were considered included:
Poor coating specification. Did the crack initate at a coating failure location?
Poor part quality. Were all the parts in spec?
Unrealistic test conditions. Why wasn't this caught earlier in component level testing?
Poor mounting or mount variation. Is the mounting in spec?
Poor analytics/simulation. Does the FEA capture the loading?
Closing out these items took varying amounts of effort from checking the inspection reports to rallying test engineers to instrument motors with strain gauges and run a sweep of tests. Working with our structural engineering team we root caused this issues to a mismatch in FEA to reality. It also highlighted our testing at the component level wasn't doing a good job accelerating this kind of failure.
By iterating through different geometries that added stiffness, minimized weight, and maximized manufacturablity, using NX and FEMAP, we were able to bring peak stresses down to a region where we could achieve our life targets. Calculated using the fatigue strength of our material and the cyclic loading condition.
Finally to verify and validate the solution the part, motor, and aircraft must pass a series of tests. Work is ongoing to develop a test setup at the part level to validate this change which will allow for the fastest iteration loop.