Prediction of Cement Strengths to Improve the Bottom Line

摘要

28 day cement strength is probably the main quality parameter on which cement sales are based. Yet this is not readily quantified, requiring a representative cement sample to undergo a laboratory test procedure and then a break test 28 days later to determine its compressive breaking strength. The problem for the cement manufacturer is that 28 days is a long time to wait for a test result for making any adjustments to the process, and in fact in many cases the cement has already been dispatched and used in concrete before the test result is available. Furthermore, the integrity of the result can easily be compromised in many ways: ldr Was the test sample representative of the production run ldr Was the laboratory procedure done correctly (different laboratory operators often produce different results for the same sample) ldr Was the breaking machine operating correctly Furthermore, even if the result was correct 28 days ago, how does this relate to what the manufacturing process is doing today. The general response by the cement manufacturer is that as he/she does not know the cement strength being produced at any time, the targets are set conservatively high so as not to fall below the acceptable minimum. This means manufacturing costs are higher than necessary in order to have a safety margin. As a result, many attempts have been made to get a more timely cement strength value in order to get useful information for control of quality and the process. This paper reviews some of these approaches, and in particular discusses one technique possible for real time continuous prediction of cement strength for automated quality control at the cement mill.