Longwall roof control through a fundamental understanding of shield–strata interaction

Abstract

Real-time and non-real-time software has been developed that encapsulates a fundamental understanding of the interaction between a longwall powered support (shield) and the surrounding strata based on the interpretation of the leg pressure data within each load cycle. A load cycle is the change in support pressure with time from setting the shield against the roof to the next release and movement of the support. It is now possible to automatically identify when a shield has too low a set pressure, and when a shield is faulty and/or has an inadequate capacity for the conditions. The use of this software has the potential to significantly reduce or even eliminate roof control problems on a longwall face with significant benefits to both productivity and safety. By automatically identifying potential causes of roof control problems and offering solutions, the software has the potential to aid longwall automation. A Beta test version of the real-time software has been successfully working at BMA's Broadmeadow Mine in Australia for some time and several Australian mines have benefited from expert off-line analyses using the software. The software can also be used to isolate the many interconnected factors affecting roof control on a longwall face, which will enable their quantification, and is therefore a powerful research tool.

Introduction

A number of geotechnical models have been developed that claim to be able to relate how a longwall powered support interacts with the surrounding strata. The methods include the detached block theory [1], the yielding foundation theory [2], the empirical nomograph-based method [3], the load cycle analysis [4], [5], neural networks [6], various numerical models [7], [8], and ground response curves [9]. All models in the public domain literature were reviewed by the authors of this paper [10]. It was concluded that the above models, while offering important contributions towards understanding support–strata interactions, did not offer effective means of interpreting how supports interacted with the strata.

In recent years, longwall roof supports (shields) have been equipped with pressure sensors and hydraulic leg pressures can be displayed in real time. Despite the vast quantity of monitoring data captured on a daily basis from the majority of modern longwall faces, few geotechnical analyses are undertaken. The collected data are largely unused for roof control or for other purposes such as maintenance. One of the major reasons for this was that it was unclear what the signals meant in the context of support–strata interaction.

The authors of this paper have developed a fundamental understanding of what the pressure signals mean in the context of support–strata interaction. The interpretations are based on the recognition of characteristic load cycles and critical load cycle features. A load cycle is the change in support pressure with time from setting the support against the roof to the next release and movement of the support. The recognition of the load cycles depends on the accurate delineation of roof support pressure behaviour and on the extraction of key features such as average pressure throughout the cycle, the number of yield events, the set pressure, the rate of loading in part or all of the cycle, and the cycle length using purposely designed software.

Using the software developed by the authors, it is now possible to automatically identify when a shield has too low a set pressure, and when a shield is faulty and/or has an inadequate capacity for the conditions. It has been found that once set conditions deteriorate and shields are set manually, it is very common for set pressures to be too low for the conditions, resulting in roof control problems. The software can automatically identify set pressures that are too low, which will enable the auditing of shield operation and corrections to be made. Up to 10% of shield legs have faults on a typical Australian longwall and these periodically result in localised roof control problems. Faulty support components are automatically identified, enabling timely repairs to be made. On some longwalls the shields become overloaded at the peak of the periodic weighting cycle or in the vicinity of faulty supports. It is now possible to identify the difference between a heavily loaded support and one that is overloaded using the software. By minimising the cycle time and making sure that set pressures are adequate in cycles following the overloading event, it is quite possible to successfully mine through an overloading event if the event is correctly identified.