Dynamic Longwall Equipment Location Model

This report describes the outcomes of a 12 month ACARP-funded research project that explored the relationship between longwall equipment position requirements and the associated sensing specifications. This project was motivated by the fundamental need to achieve safer and more effective autonomous longwall operations. The research was guided by a systems-level approach which advocated an engineering requirements-driven methodology. The project explored three main objectives, namely:

· To capture required longwall positioning performance;

· To assess current longwall positioning performance;

· To develop a model for longwall positioning using state-of-the-art engineering methods.

A set of longwall equipment position requirements were captured through interactions with longwall OEMs, minesite personnel and CSIRO personnel. This investigative stage provided a useful set of requirements for the major dynamic longwall components. This stage also provided insights regarding the availability of accurate current longwall positioning performance. This proved to be a challenging task as much of the relevant information is retained in confidence as business intelligence; furthermore some of the key components are not being routinely measured. This situation was mitigated by using known position data sources which included CSIRO's shearer position measurement system and roof support measurements from inclinometers. Coordinate reference systems were formally introduced to provide a common conceptual framework for a consistent integrated model approach.

A reference model was developed to indicate how available sensor input sources could be mapped into a common coordinate space.  The complexities encountered by developing the fundamental mathematical model highlighted the need for a more systematic approach. To this end, a formal model-based approach was introduced based on the use of a kinematic approach. The flexibility of this model-based approach for sensor-position requirements analysis was demonstrated by the ease with which an existing longwall CAD model could be used to generate a kinematic model, complete with hierarchical mechanical equipment elements. Inclinometer data previously collected from roof supports from Ulan West was then used to create an equivalent sensor in the model framework, providing a virtual longwall minibuild to compare between required and actual positioning outcomes for an identified drum-to-tip use case.

Specific outcomes and innovations delivered in this research project include:

· New insights regarding the longwall position requirements and sensing performance;

· Advocacy of a common coordinate reference frame for longwall equipment;

· New systematic modelling approach that allows user performance requirements to drive the sensor specification, avoiding the current try-and-see approach.

The report concludes with recommendations that identify ongoing activities to further improve model-based technology. It is hoped that this report will provide a reference document for further longwall operational refinements and specifications to accelerate more comprehensive longwall automation solutions for industry.