3
granularity, and 4. Continuous monitoring. Despite these
significant advantages, the adoption of LCDMs for hazard
monitoring faces some challenges, including: 1. Analytical
performance concerns, 2. Complexity of use, and 3. Lack
of understanding about the value of the data. Concerns
regarding analytical performance and the perceived com-
plexity of usage have, at times, distracted the professionals
from exploring the possible value. This case study show-
cases the installation, maintenance, and data visualization
techniques employed, along with the fruitful collaboration
between NIOSH researchers and the mine’s management,
employees, and industrial hygienist professionals. By delv-
ing into the case study findings, this article aims to shed
light on the benefits of real-time respirable dust monitor-
ing using LCDMs. Additionally, it emphasizes the signifi-
cance of such advancements in enhancing workplace safety,
informed decision-making, and paving the way for future
advancements in hazard monitoring technologies.
LCDMs offer significant potential advantages for real-
time data generation, and their implementation in occupa-
tional settings can enhance hazard monitoring capabilities
and strengthen the pre-existing workplace safety protocols.
Additionally, we will underscore the value proposition of
LCDMs to stakeholders, making a case for their wide-
spread adoption. Through presenting this case study, we
aim to establish LCDMs as a practical tool in the tool belt
of industrial hygienist professionals for enhancing work-
place safety, informed decision-making, and advancing haz-
ard monitoring technologies in the realm of occupational
health and safety.
METHODS
For this study, NIOSH researchers established a collabora-
tion with an industrial mineral sand mine in Wisconsin.
The company actively mines and crushes ore into screened
final size-selected sand products. The processing of the ore
involves a multi-story screen house where all the size-selec-
tive screening occurs and two dry houses that remove the
excess moisture from the products. These three buildings
were identified by the industrial hygienist on site as having
the highest likelihood for high respirable dust concentra-
tion, and the company was interested in monitoring and
identifying sources of high concentrations of dust so that
remediation strategies can be implemented. Following dis-
cussions with the mine, five units were distributed across
two buildings (screen house and dry house). Three units
were placed in various locations in the screen house, and
two units were placed in the dry house (Table 1). The par-
ticulate matter (PM) sensors used in this study were pur-
chased as off-the-shelf standalone units. The sensors record
PM1, PM2.5 PM10, temperature, and humidity data at
a 15-second resolution. The sensors use a light-scattering
methodology to monitor dust particles. Additional thought
regarding the expected humidity levels and type of aerosol
needs to be considered for proper calibration [16]. When
installing the sensors, the location of each unit was driven
by the needs of the mine and the availability of both 120v
power and Wi-Fi at each location. The sensors have a small
internal battery to maintain stability through intermit-
tent power interruptions, but they need to be plugged in
and thus cannot be located too far from a power supply.
Additionally, each sensor needs to have access to Wi-Fi or
a hardwired ethernet connection to transmit the collected
data, which also may limit sensor placement based on avail-
ability. For this application, we attached four high-strength
magnets to the back of each unit to allow for easy installa-
tion while also being able to withstand the vibrations typi-
cally seen in a screen house.
After the installation of the dust monitors, respirable
dust samples were collected for two days at each sensor
location using three 37-mm, 5-μm pore PVC filters (SKC
Inc., Eighty Four, PA) with 10-mm nylon Dorr Oliver
cyclones at a flowrate of 1.7 lpm [15]. This series of trip-
licated gravimetric samples was used to determine a cor-
rection factor for respirable dust that could be applied to
the monitors if needed. After 3 months of continuous use,
we noticed “spike events” where the sensor would imme-
diately go up to a PM10 reading above 1 million µg/m3,
persist for a short period (less than 10 minutes), and then
immediately fall back to pre-spike levels. We documented
these events, removed them from downstream analysis, and
displayed the total number of spike-related timestamps for
each monitor in Table 1. After removing the spike events,
all monitors retained a greater than 99.9% sensor uptime,
and we believe these spikes are a hardware issue on the sen-
sor side and not related to true dust levels.
Once the sensors were all set up and transmitting data,
the next task was data management and delivery of pro-
cessed data to the mine. Two data management approaches
were followed. The first approach relied on the online dash-
board through the manufacturer, like most LCDMs. The
Table 1. Sensor Metrics
Data points
Sensor name Location (Sept '22 -Sept '23) Spike events Uptime
Dry 1 Dry House 1,978,164 732 99.963%
Dry 2 Dry House 1,412,776 976 99.931%
Screen 1 Screen House 1,978,204 0 100.000%
Screen 2 Screen House 1,978,196 305 99.985%
Screen 3 Screen House 1,978,204 0 100.000%
granularity, and 4. Continuous monitoring. Despite these
significant advantages, the adoption of LCDMs for hazard
monitoring faces some challenges, including: 1. Analytical
performance concerns, 2. Complexity of use, and 3. Lack
of understanding about the value of the data. Concerns
regarding analytical performance and the perceived com-
plexity of usage have, at times, distracted the professionals
from exploring the possible value. This case study show-
cases the installation, maintenance, and data visualization
techniques employed, along with the fruitful collaboration
between NIOSH researchers and the mine’s management,
employees, and industrial hygienist professionals. By delv-
ing into the case study findings, this article aims to shed
light on the benefits of real-time respirable dust monitor-
ing using LCDMs. Additionally, it emphasizes the signifi-
cance of such advancements in enhancing workplace safety,
informed decision-making, and paving the way for future
advancements in hazard monitoring technologies.
LCDMs offer significant potential advantages for real-
time data generation, and their implementation in occupa-
tional settings can enhance hazard monitoring capabilities
and strengthen the pre-existing workplace safety protocols.
Additionally, we will underscore the value proposition of
LCDMs to stakeholders, making a case for their wide-
spread adoption. Through presenting this case study, we
aim to establish LCDMs as a practical tool in the tool belt
of industrial hygienist professionals for enhancing work-
place safety, informed decision-making, and advancing haz-
ard monitoring technologies in the realm of occupational
health and safety.
METHODS
For this study, NIOSH researchers established a collabora-
tion with an industrial mineral sand mine in Wisconsin.
The company actively mines and crushes ore into screened
final size-selected sand products. The processing of the ore
involves a multi-story screen house where all the size-selec-
tive screening occurs and two dry houses that remove the
excess moisture from the products. These three buildings
were identified by the industrial hygienist on site as having
the highest likelihood for high respirable dust concentra-
tion, and the company was interested in monitoring and
identifying sources of high concentrations of dust so that
remediation strategies can be implemented. Following dis-
cussions with the mine, five units were distributed across
two buildings (screen house and dry house). Three units
were placed in various locations in the screen house, and
two units were placed in the dry house (Table 1). The par-
ticulate matter (PM) sensors used in this study were pur-
chased as off-the-shelf standalone units. The sensors record
PM1, PM2.5 PM10, temperature, and humidity data at
a 15-second resolution. The sensors use a light-scattering
methodology to monitor dust particles. Additional thought
regarding the expected humidity levels and type of aerosol
needs to be considered for proper calibration [16]. When
installing the sensors, the location of each unit was driven
by the needs of the mine and the availability of both 120v
power and Wi-Fi at each location. The sensors have a small
internal battery to maintain stability through intermit-
tent power interruptions, but they need to be plugged in
and thus cannot be located too far from a power supply.
Additionally, each sensor needs to have access to Wi-Fi or
a hardwired ethernet connection to transmit the collected
data, which also may limit sensor placement based on avail-
ability. For this application, we attached four high-strength
magnets to the back of each unit to allow for easy installa-
tion while also being able to withstand the vibrations typi-
cally seen in a screen house.
After the installation of the dust monitors, respirable
dust samples were collected for two days at each sensor
location using three 37-mm, 5-μm pore PVC filters (SKC
Inc., Eighty Four, PA) with 10-mm nylon Dorr Oliver
cyclones at a flowrate of 1.7 lpm [15]. This series of trip-
licated gravimetric samples was used to determine a cor-
rection factor for respirable dust that could be applied to
the monitors if needed. After 3 months of continuous use,
we noticed “spike events” where the sensor would imme-
diately go up to a PM10 reading above 1 million µg/m3,
persist for a short period (less than 10 minutes), and then
immediately fall back to pre-spike levels. We documented
these events, removed them from downstream analysis, and
displayed the total number of spike-related timestamps for
each monitor in Table 1. After removing the spike events,
all monitors retained a greater than 99.9% sensor uptime,
and we believe these spikes are a hardware issue on the sen-
sor side and not related to true dust levels.
Once the sensors were all set up and transmitting data,
the next task was data management and delivery of pro-
cessed data to the mine. Two data management approaches
were followed. The first approach relied on the online dash-
board through the manufacturer, like most LCDMs. The
Table 1. Sensor Metrics
Data points
Sensor name Location (Sept '22 -Sept '23) Spike events Uptime
Dry 1 Dry House 1,978,164 732 99.963%
Dry 2 Dry House 1,412,776 976 99.931%
Screen 1 Screen House 1,978,204 0 100.000%
Screen 2 Screen House 1,978,196 305 99.985%
Screen 3 Screen House 1,978,204 0 100.000%