2235 Figure 13 – Events identification The model incorporates variables such as liner condition, pressure relief actions, grinding media changes, grate configurations, neural network implementation, and the GIO Center. This regression framework allows ANOVA-based validation to confirm whether observed TPH variations are truly attributable to implemented projects rather than natural process variability. Figure 14 – Multivariable modelling To properly handle stacked or overlapping effects, an iterative calculation approach is applied. A base scenario is first defined without project implementation. Then, projects are activated sequentially according to their real implementation timeline, recalculating the estimated tonnage at each step while holding other variables constant. The incremental difference between iterations represents the isolated contribution of each project. This methodology enables accurate decomposition of total TPH improvement into individual project contributions, even when multiple initiatives are implemented simultaneously or progressively.
RkJQdWJsaXNoZXIy MTM0Mzk2