CRANK IT UP The test we set up had to overcome a major challenge in that we would not have the ability to use a laser to measure a consistent metal level or any type of a molten metal flow meter and thus we would not have the normal feedback loop that we generally design into our systems. This placed the emphasis on being able to establish an initial consistent metal level using other methods that focused on the system design, the system components and the system programming that would all be “new” to us. We could not allow there to be any “leaks” in the system as this would make the metal level control much more difficult. In many of our launder transfer systems, the customer prefers to have the pump mate up with a dividing wall in the furnace, and this inherently allows for some metal to leak back through that port. In this case we designed a new riser that would completely eliminate any potential leak in the system. Over the past many years, we have been working to combine both graphite and refractory materials in a way that now enables us to design pumps that can incorporate both in a way that enables us to achieve new outcomes. Effectively, a new tool in the toolbox. We also had to look at our impellor design along the same lines so that it would generate a more consistent flow given the new system requirement of 1% variation or less. Here again, we made some changes to some of our core impeller designs so that we could test how each would perform. On the other end of the system where we deliver the metal through a launder system, we also made modifications to the shape, slope and size of the launder so that we could better control bias and work towards a much more consistent dose size (with water we used weight to measure, in the aluminum application it would be volume). Lastly, we had to develop new programming parameters to control the pump speeds so that we could improve the consistency of the metal delivery. Leveraging some of our SMART technology, we were able to find improvements that ultimately contributed to the success of the system. RESULTS The data we gathered for each of the rotor designs focused on pump speed (idle and dosing), time (initial goal set at 4 seconds) and weight (average over 25 trials). The chart included shows the results for one such trial. We took the standard deviation and the extreme spread to be the two most critical outputs for our testing. We knew we needed to deliver metal consistently within 1% of the target weight/volume in as little time as possible and avoid outliers as this would result in a major issue at the press with shot weight. Short shots needed to be eliminated from the realm of possibility. As we did the 24
RkJQdWJsaXNoZXIy NDI4Njg=