Damage detection in complex mechanical structures is important for cost-effective and safe operation. Conveyor belts with steel cords are used for bulk material transport in mining companies. Due to harsh environmental conditions, both covers and cords are subjected to damage. As lengths of conveyors may vary from dozens of meters to kilometers, a belt loop consists of many connected belt pieces. Thus, the condition of splices between belt pieces is also critical. For both steel cord damage/wear detection and splice condition evaluations the NDT techniques based on magnetic field measurement and variability analysis are used. To obtain appropriate resolution, multi-channel data are collected. Here we propose a pre-processing technique developed for signal synchronization for biased splices data. The biased splices mean a phase shift between signals from a multi-channel sensor due to the design technology of the splice. As the quality of the splice is related to the appropriate precision of splice production, splice evaluation is defined as a similarity analysis of each signal with respect to the estimated pattern. Due to the mentioned phase shift, signals should be "synchronized" first, before final analysis. In industrial conditions, many factors may influence the signal shape. Thus, the problem of automated synchronization by shifting the signals may be defined as a multidimensional optimization problem. Here, we proposed to use a genetic algorithm with an algorithmically simple cost function for that purpose. In this paper, the authors propose an automated procedure applied to real measurement data and final results. A multidimensional optimization has been compared to simple signal shifting according to several criteria, and GA-based results were the best.

In copper ore deep mines of KGHM Polska Miedၺ S.A., the first stage of the transport department is to deliver the spoil to the screen. This stage is carried out by self-propelled mining machines. Transport can be done in two ways: by loaders or loaders cooperating with dump trucks. The amount of dump truck cycles during a single work shift can be used as indicators of a machine and operator performance evaluation. In order to increase the mining potential, it is necessary to continuously acquire and analyze information on work efficiency. The ability to synthesize the obtained data and case-study the relationship between them allows deepening knowledge of industrial processes and supports decision-making. Dump trucks used in the KGHM copper ore mines are equipped with a self-unloading mechanism. The paper describes a developed method and described an algorithm for segmenting signals coming from a dump truck with a self-unloading crate. The operation of the algorithm has been verified on data from a machine operated by several operators, taking into account various operational events.