MODEL STRUCTURE AND PARAMETER IDENTIFICATION FOR CUTTING HORIZON PREDICTION OF SHEARER LOADERS

Abstract

To reach autonomous control with good performance in shearer operated longwall mines, control algorithms have to be adapted to the system behaviour. Thus, one of the challenging tasks in this area is the identification of the behaviour, which is declared as complex and specific to each mining area. Therefore, it is necessary to determine a suitable model that reproduces the cut behaviour of a shearer loader and can be adapted to a real mining system. This paper shows and analyses possible model structures to predict the cutting path of a shearer loader in the direction of face advance, combined with the least squares method for parameter identification. Single-step, as well as multi-step prediction are considered. Especially results of multi-step prediction of the cutting path, compared to the reference path shows whether the model structure and the identified parameters are valid or not. Data source for the parameter identification and cutting path reference is a complex and iteratively working simulation tool developed in cooperation with a German mining company for the training of operators. Finally, a suitable, analytical model for parameter identification and prediction of the cutting horizon is pointed out.