Development of an automated monitoring system for operational parameters of mobile crushing and screening plants

Abstract

Mobile crushing and screening plants operate under conditions for which stationary technical diagnostics schemes were never designed: autonomous diesel-generator power supply, an intermittent communication channel, relocation every few weeks and the absence of permanent engineering personnel at the site. Scheduled time-based maintenance under such operation diverges from the actual wear trajectory, and the direction of this divergence is not known in advance. The aim of the work is to substantiate the architecture and parameter composition of an automated monitoring system suitable for a self-propelled plant with an interrupted measurement history. Computational modelling of the information load, Monte Carlo modelling (10 000 iterations) of false alarm rates and failure warning time, factor decomposition of annual output by chain substitution and discounting of cash flows by the cumulative method were applied. A composition of 34 measuring channels is substantiated with computation distributed between the onboard and the remote circuits: transmitting aggregated features instead of digitised waveforms cuts the daily traffic from 221.2 MB to 92,2 kB. The adaptive boundary of 3,91 mm/s, reconstructed from the natural variability of vibration velocity, raises the mean warning time from 406 to 478 h, yet when triggered by a single measurement it increases the false alarm rate from 2,6 to 27,5 per year; the «2 out of 3» confirmation rule reduces it to 0,11 per year. The technical utilisation ratio of the model plant grows from 0,940 to 0,968, annual output increases by 32,58 thousand tonnes, and the net present value over a five-year cycle amounts to 1,63 million roubles at a discount rate of 18,5%. The ratio between the gain in warning time and the growth of the false alarm rate shows that the decisive element of a monitoring system for a self-propelled plant is the confirmation rule rather than the threshold level, while the economic result remains positive only if emergency downtime is reduced by more than 62,4 h per year.