The Importance of Hand Calculations in the Context of Computer Simulation Tools
Early in my career, I was taught something that always stuck with me related to computer simulations, such as finite element analysis (FEA). On my first day, one of the more experienced engineers told me, “Just because you run an FEA, doesn’t mean it’s correct. You must always verify it with an equivalent hand calculation.” Up to that point, being relatively inexperienced I remember my fellow students running simulations without this verification step. It was never given a thought, and more times than not, the simulation results were considered, “truth.”
Years later, I took a job with a startup that was making heavy industrial equipment based on newly invented technology. This newly invented technology was developed by non-engineers that knew SolidWorks CAD, and self-taught how to run the COSMOS simulations (a predecessor to SolidWorks Simulation). The assembly that they modeled was an actuated barrel cam that rotated about ninety degrees and would push a pair of round plates closer or farther away relative to another pair of rotors via four linkages with exposed bronze bushings at the linkage joints. The application for the device was in places like cement factories or paper and pulp mills. The amount of axial load that the actuator needed to work against was about 1300 lbs.
Long story short, these devices started to fail prematurely in service, less than twenty-five percent of their useful life. My job was to conduct forensic analysis on the load train and determine root cause of failure(s). In my investigation I learned that absolutely no hand calculations had been made to verify the FEA results. I knew this because I could not find a free body diagram anywhere. What I found in my visual inspections was astonishing: cam track compressive deformation, galling of the linkage joint bearing surfaces, evidence of excessive heat developed in the roller bearings (grease breakdown), and cement dust infiltration on every bearing surface (not the fault of a hand calculation). Then, I began from scratch to produce the hand calculation. Starting with the rotor plates and the link-pivot, I drew out all the free body diagrams back to the motor actuated lever on the barrel cam. Next, I wrote out all the kinematic relationships, coupled with the determined forces, and created a mathematical model that would produce a hand calculation of all need-to-know stresses at every angle of actuation.
If they had only started there, it would have been realized that the actuating system was severely undersized, the heat treating for required hardness was incorrect, and with cement dust flying around, perhaps isolating the actuator from the environment was in order!