Shaft Erosion

   
Click a thumbnail to view at full resolution...

Shaft Erosion, Wear, Tribology:
A shaft design with sliding components wore out prematurely. Analyze that design and compare its performance to a new design with less sliding.

Special features:
3D finite element model with shells, bricks, beams, and time-based, load curves.

Output:
Reaction forces, translation and rotation displacements.

Comments:
This analysis was very complex and the models were huge. The details must remain proprietary. However, some generalizations can be revealed.

This was a large piece of rotating machinery. The exterior components were rigidly fastened to the shaft at the ends. In the middle, the components can slide along the shaft in the axial and circumferential directions; the amount of relative motion depends on the stiffness of the machine and the loading, hence the suitability of using the FEA approach.

Where the two major pieces contact, erosion, or wear occurs, which is a function of contact pressure and sliding. This resulted in early mortality for the baseline design. From this analysis the client wanted a prediction of extended life based on a new design that limited relative motion, hence wear. Problem: the customer had only one data point. Solution, since the erosion data was for different positions on the shaft at one point in time (the end of the machine's life) and the different positions represented different relative motions, the data could be viewed as a histogram of wear life. Using FEA it was possible to gather the initial displacement of the failed machine for which the customer had empirical data. Since the wear progression through time was "path-dependent," it was possible to find a unique solution for the wear rate that matched the data. The wear rate turned out to be the combined function of three, FEA analyst derived factors: wear threshold, wear acceleration rate, and a background wear rate (this is one of those engineering insights that is not a menu item in any software). No other combination of factors would allow a match of the hypothetical wear curves and the data. So, the wear phenomenon in the failed machine had been characterized. In other words, whatever the operating conditions were that existed in the baseline machine were now captured by the unique wear factors. Fortunately, the new design operated in the same way, so it became possible to predict its life by application of the derived wear factors to the initial conditions discovered through the FEA.

Above the reader will see four charts showing the "bottom line." Note that the design life for "Design 2" degrades with an increase in clearance, because more initial relative motion can occur, hence more wear, or erosion, sooner in the life of the machine, which limits its longevity.