527
Depends on the machine type and basic or advanced version of the monitoring system, the
list may contain different parameters. During the experiment, we were able to acquire 19 variables listed in Table 1. They are related to the various component, have different dynamics
(low-frequency processes as temperature, high-frequency processes as pressure, etc.) and have
different robustness. All data are sampled with 1 Hz.
2.3 Exemplary data presentation
In Fig. 2 one might see raw data acquired by the system from a regular shift in an underground mine. As mentioned, the most interesting and clear signal is in sub-plot 11 (variable called Hyd_Oil_P). The usage of alternative variables is not intuitive. The preliminary
selection was done by mining engineers based on the physics of the machines and processes
(operations). However, we look for information-based criteria for selection of variables.
3 INFORMATIVE VARIABLES SELECTION METHODOLOGY
3.1 General description of the approach
Extraction of information from multidimensional data from multi-channel monitoring system is a well-known problem in data analysis [11,14,15]. It could be done by selection or
transformation [14–15]. In the case of selection, the most critical issue is the criterion of
selection. From a signal processing perspective, searching for indicators describing the effectiveness of the haulage process is the discovery of a cyclic pattern in acquired signals. The
Haul Truck (HT) is used for transport of copper ore from the vicinity of mining face (ore
is provided by loader) to mining screen (unloading point, next ore is transported by the belt
conveyors to the shaft). This operation is cyclic, the daily scenario may be as follow: on the
Figure 2. A raw data from 1 shift—an example.
f':r __r----:=-===: ""Y'J
2000
4000
6000
8000
10000
12000
14000
16000
18000
~ ' ,--------,-----~____;= EN ~ GHOURS "'------r-----~---,--,
'o.of - - - - - -
~ _,
,,~f l l : : : ENGO>LP : : I I :ll
2000
4000
6000
8000
10000
12000
\.WOO
16000
18000
~
ENGAPM
Samples
.g
HVDOIL c"------ P
~~
f':f : : :! I: I I I : j t' I : l
- - - - ·- - - -UL =:c~= L~
- - - - ·- - - - -
.g
~NTAKEP
t'~f 1: ~~!diruMmrm l . l
- - - - ·- - - - -
~
INTAKET
f:t :;; ~w-<~~,.._~Ld
2000
4000
6000
8000
10000
12000
14000
16000
18000
Samples
t~m:-almuiliJirmnmm- ~ 'l
- - - - - - - - -
~~f _ f-IIU~IIdDiOOIMIIm I : l
- - - - - - - - -
r~f ll
" : FI• : ol ~:~n:'
: JD J
- - - - - - - - -
~~~ ::;;=~ = : g
- - - - - - - - ,_
i,:f 1: 1 1: l [[]fflnJDI I l
- - - - - - - - -
-·~
Depends on the machine type and basic or advanced version of the monitoring system, the
list may contain different parameters. During the experiment, we were able to acquire 19 variables listed in Table 1. They are related to the various component, have different dynamics
(low-frequency processes as temperature, high-frequency processes as pressure, etc.) and have
different robustness. All data are sampled with 1 Hz.
2.3 Exemplary data presentation
In Fig. 2 one might see raw data acquired by the system from a regular shift in an underground mine. As mentioned, the most interesting and clear signal is in sub-plot 11 (variable called Hyd_Oil_P). The usage of alternative variables is not intuitive. The preliminary
selection was done by mining engineers based on the physics of the machines and processes
(operations). However, we look for information-based criteria for selection of variables.
3 INFORMATIVE VARIABLES SELECTION METHODOLOGY
3.1 General description of the approach
Extraction of information from multidimensional data from multi-channel monitoring system is a well-known problem in data analysis [11,14,15]. It could be done by selection or
transformation [14–15]. In the case of selection, the most critical issue is the criterion of
selection. From a signal processing perspective, searching for indicators describing the effectiveness of the haulage process is the discovery of a cyclic pattern in acquired signals. The
Haul Truck (HT) is used for transport of copper ore from the vicinity of mining face (ore
is provided by loader) to mining screen (unloading point, next ore is transported by the belt
conveyors to the shaft). This operation is cyclic, the daily scenario may be as follow: on the
Figure 2. A raw data from 1 shift—an example.
f':r __r----:=-===: ""Y'J
2000
4000
6000
8000
10000
12000
14000
16000
18000
~ ' ,--------,-----~____;= EN ~ GHOURS "'------r-----~---,--,
'o.of - - - - - -
~ _,
,,~f l l : : : ENGO>LP : : I I :ll
2000
4000
6000
8000
10000
12000
\.WOO
16000
18000
~
ENGAPM
Samples
.g
HVDOIL c"------ P
~~
f':f : : :! I: I I I : j t' I : l
- - - - ·- - - -UL =:c~= L~
- - - - ·- - - - -
.g
~NTAKEP
t'~f 1: ~~!diruMmrm l . l
- - - - ·- - - - -
~
INTAKET
f:t :;; ~w-<~~,.._~Ld
2000
4000
6000
8000
10000
12000
14000
16000
18000
Samples
t~m:-almuiliJirmnmm- ~ 'l
- - - - - - - - -
~~f _ f-IIU~IIdDiOOIMIIm I : l
- - - - - - - - -
r~f ll
" : FI• : ol ~:~n:'
: JD J
- - - - - - - - -
~~~ ::;;=~ = : g
- - - - - - - - ,_
i,:f 1: 1 1: l [[]fflnJDI I l
- - - - - - - - -
-·~