Real-Time Shield Pressure Analysis for the Evaluation of Roof Behavior in Longwall Panels

Abstract

Pressure data from 146 two-legged shields of capacity 1,152 t have been continuously collected from two longwall panels. The combined panel length is approximately 5 km. Geological variations of roof strata of these two panels are well documented from the borehole sections. Roof strata of Panel 1 are relatively competent and those of Panel 2 are weak and fractured. Temporal pressure data are found to be noisy and it is difficult to extract mining cycle parameters using the existing heuristic concepts. Due to the aforementioned reasons, a probability-based algorithm is developed to identify mining cycle parameters such as setting and final pressure, and time weighted average pressure (TWAP). The proposed algorithm detects approximately 98% and 88% of mining cycles from the noisy pressure versus time data. The results of the study are further validated by statistical hypothesis testing and show that roof strata of Panel 1 and Panel 2 are different as corroborated by the geological observations. Finally, the proposed algorithm is found to be an effective tool in analyzing noisy pressure versus time data of longwall panels and can be used for future panels.