606
entire mine operation. Today, communication during an incident or rescue operation is often
only possible via a few permanently installed telephones. Once an overall network exists,
devices can be integrated to transmit voice or text messages. This enables workers to get help
faster, be informed about events or receive instructions quickly.
A massive increase in sensor data collection, in addition to, or even in combination with,
an intelligent evaluation algorithm, can detect increased gas concentrations or other deviations from the normal value at an early stage. Occupational safety actions can be initiated
more quickly to prevent accidents. Besides the higher frequency of automatic data acquisition in contrast to manual measurements, it is also easier to store relevant digital data for
documentation. These can be used for later assessment or as a data basis for improvements
in operation.
In addition to numerical sensor data, image data also reflect the state of the environment.
By recording point clouds and various camera and video images, mine workings which are
difficult to access can also be seen without human on site-inspection. The data can be used
for pure visualization, inspection, documentation or to create models. 3D mine models are an
excellent basis for training, e.g. for learning machine handling or for simulators. At the TU
Bergakademie Freiberg, for example, a mine rescue simulator was developed from the data
recorded in the “Mining-RoX” project from a section of “Reiche Zeche” mine (Schmieder
2017). In this simulator the students train to handle rescue operations in a team, while sitting
in front of laptops. They are using software, which is designed like a computer game and they
control virtual mine rescuers in a simulation of the ‘Reiche Zeche’. This is the preparation
for trainings in the real section in the mine. Additionally such simulations can be extended to
virtual reality environments, with the expectation of a much more higher learning effect.
Therefore, not only the health protection of the workers but also the higher efficiency and
accuracy are valuable factors for the application of these new technologies. Machines can also
work in cooperation with employees but at a constant quality, independent of break times.
5 CHALLENGES FOR REALIZATION
Despite the significant advantages the progressive automation and digitization with the longterm target of an unmanned mine are confronted with numerous difficulties. In particular,
these difficulties result from the special nature of the mine environment in contrast to surface
industrial facilities, legal issues, the handling of large amounts of data and the individual
requirements of prospective users. Furthermore, each mine must be considered separately
for each application. The following sections give an overview of possible problems and challenges, although it is not an exhaustive list.
5.1 Underground environmental conditions
Underground mines are artificial cavities created for the purpose of extracting raw materials.
The environmental conditions in a mine highly depend on the depth of the mine; the mineral
extracted, the surrounding rock and geology, the mining technology or current use of the
mine and several other factors, and are unique to each mine. Temperatures range from cool
to very high (typically from 8°C to 50°C). This defines high requirements on material durability. Materials behaviour during temperature changes should also be considered, e.g. when
entering or leaving the mine. Ore mines in particular are characterised by a very high relative
humidity, up to 100%. Often there is dripping or stagnant water on the floor, and the mine
water often has very low pH values, causing corrosion or short circuits. In general, no natural
light is available underground, hence without artificial light absolute darkness exists. Specifically, for the use of detection algorithms based on optical features, the very different lighting
conditions represent a problem. The outgassing of CH 4 in some pits can create explosive
environments, so potential ignition sources must be avoided. Depending on the extraction
method and raw material, a lot of dust is generated during driving, which can contaminate
sensors and equipment and make them unusable.
entire mine operation. Today, communication during an incident or rescue operation is often
only possible via a few permanently installed telephones. Once an overall network exists,
devices can be integrated to transmit voice or text messages. This enables workers to get help
faster, be informed about events or receive instructions quickly.
A massive increase in sensor data collection, in addition to, or even in combination with,
an intelligent evaluation algorithm, can detect increased gas concentrations or other deviations from the normal value at an early stage. Occupational safety actions can be initiated
more quickly to prevent accidents. Besides the higher frequency of automatic data acquisition in contrast to manual measurements, it is also easier to store relevant digital data for
documentation. These can be used for later assessment or as a data basis for improvements
in operation.
In addition to numerical sensor data, image data also reflect the state of the environment.
By recording point clouds and various camera and video images, mine workings which are
difficult to access can also be seen without human on site-inspection. The data can be used
for pure visualization, inspection, documentation or to create models. 3D mine models are an
excellent basis for training, e.g. for learning machine handling or for simulators. At the TU
Bergakademie Freiberg, for example, a mine rescue simulator was developed from the data
recorded in the “Mining-RoX” project from a section of “Reiche Zeche” mine (Schmieder
2017). In this simulator the students train to handle rescue operations in a team, while sitting
in front of laptops. They are using software, which is designed like a computer game and they
control virtual mine rescuers in a simulation of the ‘Reiche Zeche’. This is the preparation
for trainings in the real section in the mine. Additionally such simulations can be extended to
virtual reality environments, with the expectation of a much more higher learning effect.
Therefore, not only the health protection of the workers but also the higher efficiency and
accuracy are valuable factors for the application of these new technologies. Machines can also
work in cooperation with employees but at a constant quality, independent of break times.
5 CHALLENGES FOR REALIZATION
Despite the significant advantages the progressive automation and digitization with the longterm target of an unmanned mine are confronted with numerous difficulties. In particular,
these difficulties result from the special nature of the mine environment in contrast to surface
industrial facilities, legal issues, the handling of large amounts of data and the individual
requirements of prospective users. Furthermore, each mine must be considered separately
for each application. The following sections give an overview of possible problems and challenges, although it is not an exhaustive list.
5.1 Underground environmental conditions
Underground mines are artificial cavities created for the purpose of extracting raw materials.
The environmental conditions in a mine highly depend on the depth of the mine; the mineral
extracted, the surrounding rock and geology, the mining technology or current use of the
mine and several other factors, and are unique to each mine. Temperatures range from cool
to very high (typically from 8°C to 50°C). This defines high requirements on material durability. Materials behaviour during temperature changes should also be considered, e.g. when
entering or leaving the mine. Ore mines in particular are characterised by a very high relative
humidity, up to 100%. Often there is dripping or stagnant water on the floor, and the mine
water often has very low pH values, causing corrosion or short circuits. In general, no natural
light is available underground, hence without artificial light absolute darkness exists. Specifically, for the use of detection algorithms based on optical features, the very different lighting
conditions represent a problem. The outgassing of CH 4 in some pits can create explosive
environments, so potential ignition sources must be avoided. Depending on the extraction
method and raw material, a lot of dust is generated during driving, which can contaminate
sensors and equipment and make them unusable.