This is a repository copy of Low-cost IoT-based sensor system: A case study on harsh

environmental monitoring.

White Rose Research Online URL for this paper:

http://eprints.whiterose.ac.uk/171502/

Version: Published Version

Article:
Sunny, A.I., Zhao, A., Li, L. et al. (1 more author) (2021) Low-cost IoT-based sensor
system: A case study on harsh environmental monitoring. Sensors, 21 (1). 214. ISSN
1424-8220

https://doi.org/10.3390/s21010214

Reuse
This article is distributed under the terms of the Creative Commons Attribution (CC BY) licence. This licence
allows you to distribute, remix, tweak, and build upon the work, even commercially, as long as you credit the
authors for the original work. More information and the full terms of the licence here:
https://creativecommons.org/licenses/

Takedown
If you consider content in White Rose Research Online to be in breach of UK law, please notify us by
emailing [email protected] including the URL of the record and the reason for the withdrawal request.

[email protected]
https://eprints.whiterose.ac.uk/
 sensors
Communication
Low-Cost IoT-Based Sensor System: A Case Study on Harsh
Environmental Monitoring
Ali Imam Sunny 1 , Aobo Zhao 1 , Li Li 1, * and Sambu Kanteh Sakiliba 2

1 Nuclear AMRC, Advanced Manufacturing Park, Brunel Way, Rotherham S60 5WG, UK;
[email protected] (A.I.S.); [email protected] (A.Z.)
2 Solartech-UK Ltd., 11 Llansannor Drive, Cardiff, Wales CF10 4BW, UK; [email protected]
* Correspondence: [email protected]

Abstract: Wireless Sensor Networks (WSNs) are promising technologies for exploiting in harsh envi-
ronments such as can be found in the nuclear industry. Nuclear storage facilities can be considered
harsh environments in that, amongst other variables, they can be dark, congested, and have high
gamma radiation levels, which preclude operator access. These conditions represent significant
challenges to sensor reliability, data acquisition and communications, power supplies, and longevity.
Installed monitoring of parameters such as temperature, pressure, radiation, humidity, and hydrogen
content within a nuclear facility may offer significant advantages over current baseline measurement
options. This paper explores Commercial Off-The-Shelf (COTS) components to comprise an installed
Internet of Things (IoT)-based multipurpose monitoring system for a specific nuclear storage situation
measuring hydrogen concentration and temperature. This work addresses two major challenges of
developing an installed remote sensing monitor for a typical nuclear storage scenario to detect both
hydrogen concentrations and temperature: (1) development of a compact, cost-effective, and robust
multisensor system from COTS components, and (2) validation of the sensor system for detecting
temperature and hydrogen gas release. The proof of concept system developed in this study not only
demonstrates the cost reduction of regular monitoring but also enables intelligent data management



through the IoT by using ThingSpeak in a harsh environment.



Citation: Sunny, A.I.; Zhao, A.; Li, L.; Keywords: low-cost sensor; energy harvesting; wireless sensor network; IoT; harsh environment;
Kanteh Sakiliba, S. Low-Cost
condition monitoring
IoT-Based Sensor System: A Case
Study on Harsh Environmental
Monitoring. Sensors 2021, 21, 214.
https://doi.org/10.3390/s21010214
1. Introduction
Received: 10 November 2020 According to the Nuclear Decommissioning Authority (NDA), on 1 April 2016 the
Accepted: 23 December 2020 total amount of radioactive waste is estimated to be 4.77 million m3 [1]. This increasing
Published: 31 December 2020 amount of waste material, which needs to be stored, treated, and disposed of in a proper
manner, presents a great number of technical and temporal challenges for the nuclear
Publisher’s Note: MDPI stays neu- industry [2]. In the decommissioning site at Cumbria, United Kingdom, the majority of
tral with regard to jurisdictional clai- the country’s nuclear wastes are currently stored on site. Depending on the type and
ms in published maps and institutio- radioactivity, varying storage strategies are applied within this site. One of the legacy
nal affiliations. storage facilities at Sellafield requires extraction of the historic Magnox Swarf, which is
followed by packaging the material for interim storage before the final processing of the
geological disposal. Over many decades of interim storage, a monitoring system needs
Copyright: © 2020 by the authors. Li-
to be implemented in order to predict the correct chemical evolution of the waste, which
censee MDPI, Basel, Switzerland.
is mainly affected by the release of hydrogen gases and heat dissipation. This requires
This article is an open access article
an assurance monitoring scheme in order to make sure that the hydrogen emissions and
distributed under the terms and con- temperature of radioactive waste are well within the accepted parameter ranges.
ditions of the Creative Commons At- The Internet of Things (IoT) is composed of numerous inter-related and interconnected
tribution (CC BY) license (https:// devices, machines, and objects sharing data over a network aimed at reaching a common
creativecommons.org/licenses/by/ goal [3]. Being an enabling technology of Industrial Revolution 4.0, the goal of the IoT is to
4.0/). allow things and objects to be connected anytime and anywhere with anyone using any

Sensors 2021, 21, 214. https://doi.org/10.3390/s21010214 https://www.mdpi.com/journal/sensors

Sensors 2021, 21, 214 2 of 12

network, path, and service. Nowadays, the IoT is offering innovative solutions through
advanced sensing technologies for major industries, e.g., healthcare services, the food
supply chain, mining production, transportation and logistics, and firefighting [4]. The
nuclear industry is no exception. One of the crucial aspects of the IoT is data management.
Intelligent data management and data processing within the IoT system through the help
of advanced Artificial Intelligence (AI) techniques [5,6] will further enhance the nuclear
waste management productivity, safety and durability.
The primary parameters of interest in the condition monitoring of legacy nuclear
waste containers are temperature variation and hydrogen gas release rate. Innovative ways
to efficiently monitor these parameters from every container utilizing wireless data transfer
could provide a greater level of safety assurance over a longer period of time. Numerous
research projects have been carried out on Wireless Sensor Networks (WSNs) for varying
applications, but the main difference with respect to the gaseous sensors used in WSNs
is the power consumption. Some investigations are focusing on the implementation of a
WSN-based air monitoring system. Each reported solution varies in the type of sensors
tailored for the individual case for the purpose of optimizing power consumptions. To
the best of the authors’ knowledge, there are no existing cost-effective devices that have
been developed for the nuclear waste storage application. Different energy harvesting
techniques have also been applied to develop such systems such as shown in Wang et al. [7].
They developed a wireless sensor network system where a solar panel is mounted on the
sensor node, harvesting constant energy and providing adequate power to the sensors. A
study on low-power Surface Acoustic Wave (SAW) sensors has been carried out in order to
address the energy issues [8]. Furthermore, a reduction in power consumption has been
adopted based on the duty cycle measurements using off-the-shelf sensors [9,10]. In addi-
tion, other low-power sensor systems have shown huge potential in harsh environmental
conditions [11–15]. However, none of these sensors have taken into consideration the
extreme environmental effects when deployed in real scenarios, such as high temperatures,
release of hazardous gases, or changes in atmospheric conditions. Moreover, all of the
aforementioned systems require bulky equipment to be connected in order to extract data.
The purpose of this paper is to develop a low-cost, compact, multimodal characteri-
zation sensor system with the current state-of-the-art, Commercial Off-The-Shelf (COTS)
components. The IoT could provide an installed remote measurement system on each
legacy nuclear waste container, thus providing a real-time understanding of the condition
of the entire nuclear waste storage content deployed for in-situ or ex-situ measurements. In
the paper, we have focused on the characterization of the sensors’ behaviors and firmware
optimization. The outcome is a qualitative estimation of the gas concentration, which is
essential to discriminate the small changes that will be presented as a result of changes in
chemical reaction rates.
The paper is organized as follows: Section 2 describes the sensing principles, the
sensor selection, and the proposed integration, followed by a description of the sensor
calibration. Section 3 outlines the experimental setup. Section 4 presents the results and
discussions. Finally, in Section 5, the conclusions and the future development perspectives
are described.

2. Methods and Measurement Principles

This section presents the measurement principles and approach for implementing
the robust sensing system. As low power consumption, along with cost effectiveness, is
the main target, this section will delve more into how the COTS components have been
selected in building a low-cost prototype sensor to demonstrate the concept for the nuclear
waste storage application.

2.1. Sensing Principle and Circuitry

The sensor system is implemented by integrating multiple sensors with supplementary
modules, such as energy harvesting, battery, charging circuit, and wireless communication
Sensors 2021, 21, 214 3 of 12

for an Internet Gateway, to achieve real-time monitoring of the environmental parameters,

e.g., H2 , pressure, temperature, and humidity, as shown in Figure 1. The sensor system is
constructed with harsh-environment-resistive sensors and provides accurate measurement
for essential parameters in the environment. The energy harvesting module supplies the
system with the necessary power, which extends the lifetime of the system and reduces the
need for maintenance in long-term operation over decades. The wireless communication
module provides a connection to the cloud server via the Internet gateway, enabling
real-time monitoring and decision-making with the aid of Artificial Intelligence (AI).

Figure 1. System block diagram for the sensing setup.

The sensors used in the project are in the MQ family along with the Bosch Sensortec
BME680. To achieve the sensing and monitoring function, the system is constructed using
two solar panels for energy harvesting with a chargeable battery and a power management
circuit, a hydrogen sensor MQ-8, a multifunction environment sensor BME680, a Wi-Fi
Module ESP8266EX, and a Wemos organic light-emitting diode (OLED) display. The circuit
connectivity can be seen in Figure 2. The power lines deliver the electricity from the energy
harvesting and the power management module to the sensors and the communication
module, which is shown as blue lines in Figure 2. The environment data are collected
in two forms: analogue data and digital data. The digital data are transmitted via I2 C
buses, shown as brown lines. The Serial Clock Line (SCL) is sent by ESP8266EX as a clock
signal and the Serial Data Line (SDA) as a bidirectional data signal. The analogue data are
collected directly using the ADC (Analog to Digital) port of the microcontroller, which is
shown as pink lines. Then, the ESP8266EX module can send the data to the base station via
self-contained Wi-Fi protocol. The specifications of the components are introduced in the
following section.

Figure 2. The sensor circuit wiring diagram.

Sensors 2021, 21, 214 4 of 12

2.2. Sensor Integration

The power management is carried out by the solar panels and stored in the recharge-
able battery for the operation of the entire system. The environmental data is collected by
the Bosch Sensortec BME680, which is an environment sensor for gas, humidity, temper-
ature, and pressure detection, and by MQ-8 sensors, and is transmitted to the local area
network via Wi-Fi by ESP8266EX.

2.2.1. Sensor Selection

Using a single sensor not only simplifies the system design but also reduces the
energy consumption for environmental monitoring. However, there is no integrated
sensor commercially available for sensing hydrogen gas together with other environmental
parameters. Thus, the two sensors used in this system, namely MQ-8 and BME680, are
suitable for monitoring the temperature and hydrogen release rate to prove the concept.
MQ-8 is a H2 detection sensor selected for its high sensitivity and is also able to
detect a variety of other hydrogen-containing gases (e.g., nitrogen or air) [16]. The sensor
uses stannic oxide (SnO2 ) as the sensing material, which has lower electrical conductivity
within clean air. When the concentration of hydrogen gas increases, the conductivity of the
material increases. Then, the variation in gas concentration within the environment can be
converted from the conductivity change using a simple circuit, as shown in Figure 3. A
heater voltage (VH ) is introduced to supply a DC or an AC current to heat the sensor to
working temperature. VC is a DC power source for converting the conductivity change
to a voltage signal with the presence of a load resistance RL , which is in series with the
sensor. VL is the voltage of load resistance RL and can be connected to an amplifier circuit
as shown in Figure 3.

Figure 3. MQ-8 circuit diagram.

BME680 is selected as a 4-in-1-sensor that can measure barometric pressure, rel-

ative humidity, ambient temperature, and gas concentration. The package is just
3.0 × 3.0 × 0.93 mm3 , enabling a compact design. Another advantage of this sensor is
its low temperature coefficient offset (TCO) [17]. The sensor provides digital data via
I2 C or a Serial Peripheral Interface (SPI) and operates with low current consumption
(microamps) at a sampling rate of 1 Hz.

2.2.2. Power Management

The sensor circuit can be operated at a very small current (approximately 5–20 mA)
and 5 V DC, however, most of the current in this case is consumed by the heater circuit
(around 45 mA). The used photovoltaics panel is able to deliver about 6.2 V and up to
300 mA, which is connected as an unregulated power source, through the battery to the
ground and voltage input. Hence, a built-in regulator is employed to supply the sensor
Sensors 2021, 21, 214 5 of 12

with constant 5 V. The DC–DC converter that is already embedded in the ESP8266EX
Board converts the circuit voltage to 3.3 V. The power of the circuit (without the Hydrogen
sensor) is 0.1 W. Therefore, the energy for one day is about 2.4 Wh/day, assuming that
the power is on for 24 h. The MQ-8 sensor uses 0.5 W of power due to the embedded
heater; hence, the energy for one day is 12 Wh/day. In short, the total power of the circuit
including the hydrogen sensor MQ-8 is 0.6 W, and the total maximum energy consumption
is 14.4 Wh/day.

2.2.3. Communication Method and Display

There are several wireless communication protocols supporting wireless sensing and
monitoring, e.g., Bluetooth and ZigBee. The advantage of Wi-Fi is that the sensor can be
directly connected to a LAN and update the data to either a local control center or a remote
one through the Internet. It is also supported by ThingSpeakTM [18] to aggregate, visualize,
and analyze live data streams in the cloud. The Wi-Fi Module based on an ESP8266EX chip
offers a complete Wi-Fi networking solution and is also used as a controller for the sensors
and the OLED display, which is the Wemos Mini D1 OLED. This module was used as a
display module with the Wi-Fi Module based on the ESP8266EX, providing a user-friendly
setup and debugging interface.
A serial communication protocol with a baud rate of 115,200 is set and the device is
connected to the Internet with a pre-set Wi-Fi Service Set Identifier (SSID) and password.
Rather than implementing a classical PC-based user interface, our aim is to let the user
access the data everywhere and, at the same time, increase the portability of the complete
system. With a ThingSpeak private account and an Application Programming Interface
(API) address, a Transmission Control Protocol connection is then established between the
ESP8266EX and ThingSpeak cloud and the data can be monitored on the ThingSpeak API
website, as shown in Figure 4. The sensor data shown are prior to calibration.

Figure 4. Example of data representation through the Application Programming Interface (API)
provided by the open application platform Thingspeak.com.

2.3. Sensor Calibration

The sensitivities of most sensors degrade over long-term operation when meeting the
actual performance criterion. Hence, calibration is required to make the acquired sensor
dataset more accurate. Generally, the result is ambiguous when the sample exerts more
than one gas if it can sense both of the gases. To differentiate or to sense in a more efficient
and better way, a new and advanced sensor is needed. The MQ-8 sensor is designed in such
Sensors 2021, 21, 214 6 of 12

a way that it can recognize the number of gases simultaneously. The MQ family sensors
are capable of measuring the concentration and substances that co-exist in a mixture. In
the calibration process, we used a voltage sensor to adapt a voltage result of the gas
sensor [19]. The first formula is used in [20], showing it is nonlinear for the sensor with gas
concentration (1):

− β
R/R0 = 1 + k gas Cgas (1)
where R is the sensor resistance, R0 is the sensor resistance, Cgas the concentration of the
used gas, β is the law of the characteristic power of the sensor, and kgas is the gas constant.
The formula shows a power function with a negative exponent as:

y = mxn , n < 0 (2)

According to [20], the formula for measuring the clean air resistance with a known
supply voltage VCC and a load resistance R L of 10 kΩ can be found as:

R L (VCC − Vout )
Rclean_air = (3)
Vout

where
R L = Load resistance
VCC = Sensor supply voltage
Vout = Output voltage
Therefore, the sensor resistance R0 can be determined by the ratio of clean air resistance
and air ratio as stated as below:
R
R0 = clean_air (4)
airratio
Rclean_air is the sensor reference resistance for clean air. airratio = 9.56 and represents a
constant of the MQ-8 sensor.
If ppm is the gas concentration in parts per million, then according to nonlinear
regression, the output equation of the sensor is:
n
R gas

ppm = m (5)
R0

R gas
 
log( ppm) = logm + nlog (6)
R0
R gas
logm+nlog( )
ppm = 10 R0
(7)
In the equations, Rgas is the sensor resistance in the presence of gas. The subscript gas
stands for certain gases where H2 is hydrogen. From the datasheet, there is no formula pro-
vided for each gas type of MQ sensor. Using the datasheet’s graphical representations [16],
we extracted the formula of gas and this is shown in Figure 5.

We used the MQ-8 sensor to extract the points on the graph (log ppm H2 = logm +
RH
nlog( R02 )) for H2. A set of points was extracted using WebPlotDigitizer to get a mathematical
model that matches the data. Figure 6a shows the sensitivity curve, which shows the VRL in
hydrogen with different concentrations and that the resistance load RL is 10 kΩ, and Figure
6b shows the long-term stability curve. The response graphs of the sensor provided by the
manufacturer are plotted under standard conditions [16]. It provides the baseline of the
characterizations for different practical applications, in this case, nuclear waste storage.
Sensors 2021, 21, 214 7 of 12

𝑜𝑔(𝑝𝑝𝑚𝐻2 ) 𝑙𝑜𝑔𝑚 +
𝑅𝐻2
𝑛𝑙𝑜𝑔( ))
𝑅0

10 kΩ, and

Figure 5. WebPlotDigitizer plotting for the MQ-8 sensor to extract the point for H2.

𝑜𝑔 𝑝𝑝𝑚 𝑙𝑜𝑔𝑚
𝑛𝑙𝑜𝑔

(a) (b)

Figure 6. Sensor characteristics: (a) the sensitivity curve; (b) the long-term stability curve [16].

Similarly, the formulas, shown in Table 1, can be found experimentally by properly

calibrating the MQ-8 sensor for a 1000 ppm H2 concentration in air and a value of RL of
about 10 KΩ (5 KΩ to 33 KΩ).
about 10 KΩ (5 KΩ to 33 KΩ).
Table 1. Formula obtained for the MQ-8 sensors for measuring different gases [16].

Gas log( ppm𝑹)𝒄𝒍𝒆𝒂𝒏−𝒂𝒊𝒓

= logm + nlog( Rclean − air
)
𝐥𝐨𝐠(𝒑𝒑𝒎) = 𝒍𝒐𝒈𝒎 + 𝒏𝒍𝒐𝒈( ) R0
H2 logm = 𝑹 −𝟎4.23 and n = 3.09
𝑯𝟐 LPG 𝑙𝑜𝑔𝑚 = −4.23 𝑎𝑛𝑑 logm
𝑛 = 3.09
= −2.53 and n = 1.88
𝑳𝑷𝑮 CH 4 𝑙𝑜𝑔𝑚 = −2.53 𝑎𝑛𝑑 logm
𝑛 = 1.88
= −2.55 and n = 2.26
𝑪𝑯𝟒 CO 𝑙𝑜𝑔𝑚 = −2.55 𝑎𝑛𝑑 logm
𝑛 = 2.26
= −7.36 and n = 6.48
𝑪𝑶 Alcohol 𝑙𝑜𝑔𝑚 = −7.36 𝑎𝑛𝑑 𝑛 = 6.48
logm = −4.5 and n = 4.82
Ω Ω Ω
𝑨𝒍𝒄𝒐𝒉𝒐𝒍 𝑙𝑜𝑔𝑚 = −4.5 𝑎𝑛𝑑 𝑛 = 4.82
3. Experimental Setup
The system is designed to perform under harsh environmental conditions 𝑹𝒄𝒍𝒆𝒂𝒏 𝒂𝒊𝒓 and, in
particular, the system is supposed to be deployed𝐥𝐨𝐠 𝒑𝒑𝒎 𝒍𝒐𝒈𝒎
within 𝒏𝒍𝒐𝒈 site. The sensors
the nuclear
𝑹𝟎
have been designed 𝑯𝟐 to be attached to ◦legacy nuclear
𝑙𝑜𝑔𝑚 waste
− containers,
𝑎𝑛𝑑 𝑛 which have an
inner temperature between ~22 and 55 C at normal
𝑳𝑷𝑮 room−
𝑙𝑜𝑔𝑚 temperature.
𝑎𝑛𝑑 𝑛 Additionally, the
containers have a𝑪𝑯 relatively low external contact radiation
𝑙𝑜𝑔𝑚 − dose rate that
𝑎𝑛𝑑 𝑛 does not cause any
𝟒
signal or data transmission
𝑪𝑶 interference [21]. However,
𝑙𝑜𝑔𝑚the system
− at this
𝑎𝑛𝑑 𝑛 stage has not been
designed to be 𝑨𝒍𝒄𝒐𝒉𝒐𝒍 𝑙𝑜𝑔𝑚 − design
exposed to radiation, as it is in the conceptual 𝑎𝑛𝑑 𝑛phase. Therefore, no
Sensors 2021, 21, 214 8 of 12

shielding is required, and neither is a communication protocol with loss detection and re-
transmission. A 3D model rendering image is shown in Figure 7, which depicts the stacking
up of the legacy nuclear waste containers within the storage facilities. The containers have
four filters on the lid to vent hydrogen gas to the ambient. The sensors can be fixed in the
vicinity of the filters. Therefore, to replicate the scenario in a laboratory-based experiment,
a small stainless steel cubic box was used for the validation of the multidetector sensor’s
performance within the metal’s proximity.

Figure 7. Concept design for legacy nuclear waste containers containing the four filters. The filters
are recessed and screwed into the box lid, which provides two pathways to each filter.

Two scenarios are investigated:

• Scenario 1: Hot air flow test—pressure, gas, humidity, and temperature change inside
and outside the stainless steel box under a hot air flow to monitor the temperature
variation;
• Scenario 2: Hydrogen flow test—pressure, gas, humidity, and temperature change
when a hydrogen flow is introduced into the stainless steel box to sense the hydrogen
concentration.
In scenario 1, the temperature change is undertaken with a hot air flow passing
through the surface of the sensor. Two conditions are tested individually: outside and
inside the stainless steel box, and on the cap of the stainless steel bottle. The hot air flow
is induced by a DEWORX Original 2000 W Hot Air Gun, which has two heating levels
(600 ◦ C and 300 ◦ C) and two flow rate levels (300 L/min and 500 L/min). For both the
inside and outside testing, the hot air gun is set to a fixed flow rate and held at a fixed
distance above the sensor with the hot air flow directly pointing to the sensor. This can be
seen in Figure 8a.

(a) (b)
Figure 8. Experimental validation setup: (a) scenario 1 and (b) scenario 2.
Sensors 2021, 21, 214 9 of 12

In scenario 2, a hydrogen gas bottle is used to generate a hydrogen stream in the metal
container and increase the hydrogen concentration level inside. The stream is produced
at ambient temperature and guided into the metal box. This is shown in Figure 8b. The
environmental parameters are monitored during the process for analyzing the influence of
the hydrogen concentration increase.
Figure 9 below shows the overall sensor system prototype. The total size of the
prototype is only 3 cm high, 6 cm long, and 1.5 cm wide with the PTFE casing, which has a
temperature tolerance of up to 250 ◦ C, used to protect the electrical components.

Figure 9. Compact robust multisensor system prototype in a thermally resistant PTFE case.

4. Results and Discussion

This section evaluates the performance of the proposed system for analyzing the
selected parameters under two different scenarios, which are explained below. The results
obtained were processed in MATLAB and the varying characteristics of the sensor system
are presented.

4.1. Scenario 1: Hot Air Flow Test

The results were selected from 100 s before switching on the heater gun and then 280 s
after, to give a steady reading. Ten individual tests were carried out and the results were
averaged statistically. The divisions of different tests were calculated as error bars. To
compare the trend of different parameters, these results were normalized by subtracting
the average and then dividing by the standard division of the data set. The results are
shown in Figure 10.
The results show that the temperature increases when the hot air gun switches on
and starts to fall after the gun is switched off. The pressure and gas results show the
same trend as the temperature change, as the pressure drops more quickly as the gun
switches off, followed by the gas sensor results. It takes a longer time for the temperature
to reset to a normal status due to the residual heat compared with pressure and gas. The
humidity decreases during the increase in temperature and increases when the temperature
goes down. This is because the humidity sensor measures the relative humidity of the
surrounding environment. When the temperature increases, the saturate water vapor
pressure increases, which decreases the relative humidity when the water vapor mass in
the air remains the same.
The results also suggest that the recovery of the humidity is slower than that of the
temperature. This is because the humidity sensor has a response time of 8 s, which means
it takes 8 s to reach 63% of the total humidity change.
For the test results inside the stainless steel box, the hot air gun is switched on from
0 and switched off at 100 s. The trend in temperature, humidity, and pressure changes
during the heating and cooling process are similar to the outside results, but in a slower
path. This is reasonable because the interaction between the inner environments is either
Sensors 2021, 21, 214 10 of 12

through the small window or via the heat interchange through the metal wall. Notably, the
error bar between different tests is also smaller than the outside results for temperature
and pressure measurement. It suggests that the inside measurement is more stable.

(a) (b)

(c) (d)
Figure 10. Scenario 1 test results for both inside and outside the stainless steel box: (a) Temperature; (b) Humidity;
(c) Pressure; and (d) Air quality index.

4.2. Scenario 2: Hydrogen Flow Test

This test shows the hydrogen detection characteristics in comparison with the other
parameters. In the test, the hydrogen stream started to blow inside the stainless steel from
150 s and the parameter results were recorded to 240 s. The results for the hydrogen flow
test are shown in Figure 11, which is used for comparing the parameter change during
the increase in hydrogen concentration with other parameters. As the parameters are in
different scales, they must be normalized to fit in one figure for comparison as shown in
Equation (8) below:
P−P
Pnorm = (8)
σ
where P stands for the parameter value, P is the mean of the parameter value in one test,
and σ is the standard division of the parameter value in one test.
 𝑃 − 𝑃̅
𝑃𝑛𝑜𝑟𝑚 =
𝜎
Sensors 2021, 21, 214 11 of 12
𝑃̅
𝜎

Figure 11. Scenario 2 test results for the detection of H2 components in comparison with the other
environmental parameters.

It is shown that the hydrogen concentration level starts to increase as the stream starts,
which shows that the hydrogen sensor is working properly. At the same time, it shows
a decrease in all other parameters. The decrease in temperature can be explained by the
temperature decrease, which happens when the pressurized gas is released from the bottle.
As there is no water content in the hydrogen stream, the humidity also decreased during
the process. The pressure seems to decrease but is actually increasing and decreasing due
to environmental influence.

5. Conclusions and Future Work

In this paper, we have presented a conceptual multisensor system that is able to
monitor the environmental parameters for nuclear waste storage containers or other harsh
environmental applications. The proposed system architecture breaks down the IoT struc-
ture into functional blocks and provides simple means of modularity between the layers.
The approach is cost-effective, consumes less power, and is faster with respect to develop-
ing a new silicon sensor device. The research also describes the calibration methods for
the MQ family of gas sensors for detecting leakage of H2 along with the radiation-tolerant
environmental monitoring sensor BME680. The formulae for calibration of this type of
sensor have been extracted from the data representation of the sensitivity curve. The
sensor performs with higher sensitivity towards the desired parameter characterization
in a preset environment. For both the experiments of temperature monitoring and H2
flow testing, the measured parameters from the proposed sensor demonstrate the expected
results. The current system is self-powered by a solar cell from which the energy is stored
in a researchable battery. Other energy-harvesting techniques for indoor storage facilities,
e.g., from gamma radiation, ambient and transmitted radio frequency, conducted heat
from nearby containers, or a resonant magnetic field, should be investigated in the future.
In addition, the signal interference and electronics shielding in a radiation environment
should also be studied.
The robust sensing system provides an alternative way for in-situ online monitoring
of gaseous changes and future work will include the design of a radiation-shielded case
for sensors that can be readily deployed in a harsh environment. The work will enable
the installation and routine operation of a permanent monitoring solution for legacy
nuclear waste containers or any other radioactive waste containers, in order to support the
decommissioning strategy in the country. This paper has demonstrated the possibility of
using the IoT technology for inventory management in nuclear industries.

Author Contributions: Conceptualization, A.I.S. and S.K.S.; Investigation, A.I.S. and A.Z.; Method-
ology, A.I.S.; Validation, S.K.S.; Visualization, A.I.S. and A.Z.; Writing—original draft, A.I.S. and
Sensors 2021, 21, 214 12 of 12

A.Z.; Writing—review and editing, A.I.S., A.Z., and L.L., Supervision, L.L. All authors have read and
agreed to the published version of the manuscript.
Funding: This research received no external funding.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Data sharing not applicable.
Acknowledgments: The authors would like to thank the Design and Prototyping Group of AMRC
for providing CAD design support.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Nuclear Decommissioning Authority: Annual Report and Accounts 2019 to 2020. Available online: https://www.gov.
uk/government/publications/nuclear-decommissioning-authority-annual-report-and-accounts-2019-to-2020 (accessed on
16 September 2020).
2. Ewing, R.C.; Von Hippel, F.N. Nuclear Waste Management in the United States—Starting Over. Science 2009, 325, 151–152.
[CrossRef] [PubMed]
3. Atzori, L.; Lera, A.; Morabito, G. The Internet of Things: A Survey. Comput. Netw. 2010, 54, 2787–2805. [CrossRef]
4. Da Xu, L.; He, W.; Li, S. Internet of Things in Industries: A Survey. IEEE Trans. Ind. Inform. 2014, 10, 2233–2243.
5. Maitra, S.; Abdelgawad, A.; Yelamarthi, K. Lab in a Box: A Rapidly Deployable Environmental Monitoring IoT System.
In Proceedings of the 2019 IEEE 62nd International Midwest Symposium on Circuits and Systems (MWSCAS), Dallas, TX, USA,
4–7 August 2019.
6. Maurizio, R.; Brunelli, D. Ultra Low Power Wireless Gas Sensor Network for Environmental Monitoring Applications. In
Proceedings of the 2012 IEEE Workshop on Environmental Energy and Structural Monitoring Systems (EESMS), Perugia, Italy,
28 September 2012.
7. Wang, D.; Agrawal, D.P.; Toruksa, W.; Chaiwatpongsakorn, C.; Lu, M.; Keener, T.C. Monitoring Ambient Air Quality with Carbon
Monoxide Sensor-Based Wireless Network. Commun. ACM 2010, 53, 138–141. [CrossRef]
8. Tang, K.-T.; Chiu, S.-W.; Hao, H.-C.; Wei, S.-C.; Lin, T.-H.; Yang, C.-M.; Yao, D.-J.; Yeh, W.-C. An Electronic-Nose Sensor Node
Based on Polymer-Coated Surface Acoustic Wave Array for Environmental Monitoring. In Proceedings of the 2010 International
Conference on Green Circuits and Systems, Shanghai, China, 21–23 June 2010.
9. Vana, J.; Magno, M.; Paci, G.; Brunelli, D.; Benini, L. Design, Characterization and Management of a Wireless Sensor Network
for Smart Gas Monitoring. In Proceedings of the 2011 4th IEEE International Workshop on Advances in Sensors and Interfaces
(IWASI), Savelletri di Fasano, Italy, 28–29 June 2011.
10. Vana, J.; Magno, M.; Brunelli, D.; Paci, G.; Benini, L. Context-Adaptive Multimodal Wireless Sensor Network for Energy-Efficient
Gas Monitoring. IEEE Sens. J. 2012, 13, 328–338.
11. Sunny, A.I.; Zhang, J.; Tian, G.Y.; Tang, C.; Rafique, W.; Zhao, A.; Fan, M. Temperature Independent Defect Monitoring Using
Passive Wireless RFID Sensing System. IEEE Sens. J. 2019, 19, 1525–1532. [CrossRef]
12. Sunny, A.I.; Tian, G.Y. Enhanced Sensitivity of Low Frequency (LF) RFID Sensor Signal for Structural Health Monitoring (SHM)
in High Temperature Environment. In Proceedings of the 19 World Conference on Non-Destructive Testing, Munich, Germany,
13–17 June 2016.
13. Tang, C.; Rashvand, H.F.; Tian, G.Y.; Hu, P.; Sunny, A.I.; Wang, H. Structural Health Monitoring with WSNs. In Wireless Sensor
Systems for Extreme Environments: Space, Underwater, Underground, and Industrial; Wiley: London, UK, 2017; pp. 383–408.
14. Sunny, A.I.; Tian, G.Y.; Zhang, J.; Pal, M. Low frequency (LF) RFID sensors and selective transient feature extraction for corrosion
characterisation. Sens. Actuators A Phys. 2016, 241, 34–43. [CrossRef]
15. Jun, Z.; Tian, G.Y.; Marindra, A.M.J.; Sunny, A.I.; Zhao, A.B. A Review of Passive Rfid Tag Antenna-Based Sensors and Systems
for Structural Health Monitoring Applications. Sensors 2017, 17, 265.
16. Flammable Gas Sensor (Model: MQ-8) Manual. Available online: https://cdn.sparkfun.com/datasheets/Sensors/Biometric/
MQ-8%20Ver1.3%20-%20Manual.pdf (accessed on 13 September 2020).
17. BME680-Low Power Gas, Pressure, Temperature & Humidity Sensor. Available online: https://www.bosch-sensortec.com/
media/boschsensortec/downloads/datasheets/bst-bme680-ds001.pdf (accessed on 13 September 2020).
18. Thingspeak for Iot Projects. Available online: https://thingspeak.com/ (accessed on 14 September 2020).
19. Ghee, O.K.; Zeng, K.; Grimes, C.A. A Wireless, Passive Carbon Nanotube-Based Gas Sensor. IEEE Sens. J. 2002, 2, 82–88.
20. Daniel, D.; Hnatiuc, M.; Lazar, I. Acquisition and Calibration Interface for Gas Sensors. In Proceedings of the 2018 IEEE 24th
International Symposium for Design and Technology in Electronic Packaging (SIITME), Iasi, Romania, 25–28 October 2018.
21. Banos, A.; Hallam, K.R.; McEwan, C.; Whittaker, D.; Hughes, P.; Scott, T. Proof of concept trials for in-situ testing of filter
performance on Sellafield Self Shielded boxes. Prog. Nucl. Energy 2019, 116, 10–20. [CrossRef]