=Paper=
{{Paper
|id=Vol-2762/paper8
|storemode=property
|title=Intelligent Polymetric Systems Industrial Applications
|pdfUrl=https://ceur-ws.org/Vol-2762/paper8.pdf
|volume=Vol-2762
|authors=Yuriy Zhukov,Oleksii Zivenko
|dblpUrl=https://dblp.org/rec/conf/ictes/ZhukovZ20
}}
==Intelligent Polymetric Systems Industrial Applications==
https://ceur-ws.org/Vol-2762/paper8.pdf
Intelligent Polymetric Systems Industrial
Applications
Yuriy D. Zhukova,b , Oleksii V. Zivenkoa,c
a
Admiral Makarov National University of Shipbuilding, Central Avenue 3, Mykolaiv, 54000, Ukraine
b
“C-Job Mykolaiv” LLC, Shoseina Street 5/1, 54029, Mykolaiv, Ukraine
c
“AMICO Digital” LLC, Spasska Street 1/7, 54001, Mykolaiv, Ukraine
Abstract
In this article, authors present the latest developments of polymetric sensory systems (PSS) theoretical
background and their integration with intelligent industrial robotics and control systems. New results
in the axiomatic theory of polymetric measurements development are described. Introduction of the
concept of hyper signals of the third genus allowed authors to develop a new set of consequences of
the axioms of this theory. The influence of these theoretical developments on further diversification
of PSS application in information technologies and embedded smart systems in different industries are
analyzed.
Keywords
Polymetric sensory system, smart embedded system, axiomatic theory, hyper signal
1. Introduction and Related Works Review
The concept of the polymetric signal [1], development of the axiomatic theory of polymetric
measurement [2], and experience in the application of polymetric sensors for the embedded ship
safety control systems [2, 3] engendered strong incentive to apply similar sensing subsystems
in intelligent and robotic systems in different industries. Smart decisions are based on the
knowledge of the system as well as ambient conditions and influence factors provided with
high accuracy by sensors. The sensor serves as the source of information for all subsequent
applications and decisions. Sensing subsystems nowadays are practically realized utilizing
embedded systems, IoT and appropriate infrastructure. In the world of Instrumentation and
process control, a "sensor" is defined as a device that detects changes in physical properties
and produces an electrical output in response to that change. Nowadays, the usage of so-
called «smart sensors» in industrial applications, their effectiveness and positive impact are
undeniable [4, 5]. The «smart sensor» concepts play an important role in Industry 4.0, where
«smart sensors» are one of the key points in establishing manufacturing intelligence [6]. Roughly
«smart sensor» is a sensor that can perform sensing of a physical property change. It also
executes data conversion and digital processing. It can communicate to external devices, do some
of self-diagnostics, self-assessment, self-calibration and even has reconfigurable hardware or/and
ICT&ES-2020: Information-Communication Technologies & Embedded Systems, November 12, 2020, Mykolaiv, Ukraine
" Yuriy.zhukov@nuos.edu.ua (Y.D. Zhukov); oleksii.zivenko@nuos.edu.ua (O.V. Zivenko)
0000-0002-7454-8007 (Y.D. Zhukov); 0000-0002-1539-8360 (O.V. Zivenko)
© 2020 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR
Workshop
Proceedings
http://ceur-ws.org
ISSN 1613-0073
CEUR Workshop Proceedings (CEUR-WS.org)
software. Such smartness became possible due to sustainable development and integration of
computer-based technologies and embedded systems [7].
In contrast to conventional, non-networked sensors, Industry 4.0 sensors deliver more than
just measurement data. Their computing power and flexible programmability are essential
characteristics for making production processes more flexible, dynamic, and efficient. Intelligent
sensory systems of the future should be energy-efficiency because of the focus on autonomy
and cost-efficient functionality.
According to [2, 3] the polymetric sensing might be effective road-map of such a development.
The polymetric sensor can be defined as a device that detects changes in more than one
parameter of physical properties of a controllable object but uses one specially designed sensory
element or a set of specially-designed sensors and transducers. Such an approach allows to
measure and estimate multiple parameters simultaneously with fewer sensory components.
Thus the essential difference between PSS and conventional measurement systems lies in their
informational and energy efficiency.
Unlike traditional multi-parameter sensors/systems, e.g. [8, 9, 10], PSS uses one sensor to
estimate the full set of required parameters or maximum available number of parameters in
uncertain or harsh conditions of measurements. Polymetric sensor (or polymetric sensory
system) is designated to give an output for the end-user wider than the conventional multi-
parameter measurement system. It is important to note that polymetric sensory systems are
designed to reduce hardware complexity (fewer sensors) and at the same time to enlarge the
output of information for an end-user and to improve the quality of this information. It also
estimates itself the quality of measurements and estimation results. The above advantage may
be achieved only by using specially designed sensors, appropriate measuring and processing
techniques [11] and corresponding data-based domain models loaded to PSS [2].
Modern control system sensors have evolved from simple transducers of physical quantities
to expert systems. Those systems can assess the measures carried out and take the appropriate
decisions in complete autonomy using, for example, the concepts of multi-agent systems [12, 13].
It has become widely recognized [14] that the future factory should be a smart facility where
design and manufacture are fully integrated into a single engineering process that enables
’right first time’ every time production of products to take place. It seems entirely natural for
use embedded monitoring or intelligent control systems [15] at such factories to provide an
online, suitable and precise assessment of not only the quantity and quality of the produced
products but also the technological, economic and ecological efficiency and safety of appropriate
processes realized. At the same time, the introduction of the concepts of holonic multi-agent
agencies, artificial intelligence and machine learning methods [16, 17, 18] has enormously
facilitated the development of intelligent monitoring and control systems in various fields of
human activities including different production, transportation and social processes. The idea
to integrate sensory PSS with holonic agencies philosophy and methods seemed to be very
fruitful. The achievements in applying such concept for building embedded PSS and appropriate
synergy were achieved in frames of projects for the processes control at the motor oil production
plant, floating docks processes automation, safety control at nuclear power stations, at grain
trans-loading, oil-products or LPG terminals, etc.
The classical control scheme is based on the use of information from various sensors or
measurement subsystems, user inputs and the use of preprogrammed logic, set points and
controllable object model, other algorithms for producing control actions. The control action
usually initiated by a programmed logical controller (PLC) or equivalent. Due to the complexity
of the hardware, complicated control algorithms, practical expediency, safety-related reasons
measurement subsystem and control subsystem are usually separated in hardware (sensors
+ data acquisition and preparation system + control system). In contra, in the case with PSS,
it seems natural to add to the structure of PSS components for operative control based on
estimated and predicted object state parameters. This feature can be defined as the next step in
the convergence of control and measurement systems. The authors claim that control signals
can be the part of an output vector of a PSS and can be defined as a derivative function of
the polymetric signal, target function (or setpoints) and preliminary information about the
controllable object (object model, data-based model, etc.). Such PSS became a controller and
can be implemented for operative control and urgent actions producing (so the control loop
can be realized on the bottom level of a controllable process). This feature can be especially
useful in safety-related applications and applications where harsh conditions are present, and
the signals can’t be operatively transmitted, but the urgent action needed.
An example can be easily illustrated based on a control system with wirelessly connected
sensors – e.g. the technique described in [10]. Such a system may have possible problems with
a wireless network and data transmission (due to electromagnetic disturbances), that lead to
the control in uncertain conditions and control algorithms complexity. The usage of PSS in this
case with an internal local control loop makes it possible to avoid some emergencies.
Let us use a biological analogue with human or animal reflexes system. PSS may in a
similar way percept (sense) the physical properties of an environment (controllable object or
process) and save obtained information, produce some conclusions using various loaded models
and knowledge (experience or heuristic) about it. Then PSS is capable of reacting (reflexing)
accordingly, making control signals for actuators in local control loops if it’s needed and transmit
the information about it to the upper level.
However, such an optimistic prospect meets in practice particular problems and restrictions.
One of the most critical issues relates to quite different axiomatic grounds of metrology theory
and control science and somewhat discrepant methods from timing constraints point of view.
Thus, we face an actual and urgent need to integrate and combine the theoretical foundations
for the above-advanced approaches, as this may be a precondition for generating real benefits
in the field.
2. Proposed technique
The latest results of the development of the axiomatic theory of polymetric measurements
engendered intensive diversification of application of polymetric sensors in different new fields
and further improvement of already industrialized monitoring and control systems. It is based
on the use of latest achievements in metrology, Artificial Intelligence or AI, Holonic Agencies
theory, Big Data and Machine Learning instruments, which engendered its development.
2.1. Basics of the axiomatic theory of polymetric measurements
In classical industrial metrology, it is presupposed that for practical measurement processes it is
necessary to have the set of different instruments with various sensor transformation principles,
sensors construction, and time constants. Furthermore, each device has a custom-made uniform
scale for the assessment of the actual value of the object-specific measure of its characteristic
under control.
Polymetric measurement, in general, is the process of getting simultaneous assessments of a
set of object physical quantities (more than two) with different time constants using one special
measuring transformer (a polymetric sensor).
In general, polymetric measurements theory is based on the following axioms (PMTA):
• Axiom 1𝑃 . Tuple 𝑍 =< 𝑧1 𝑧2 ...𝑧𝑖 ...𝑧𝑛 > of integer numbers is brought in correspondence
with every result of any measuring experiment.
• Axiom 2𝑃 . A topology of tuples of hyper signals of first genus 𝐻𝜏𝐼 ∈ 𝑅 𝑀 and of hyper signals
of second genus 𝐻𝜏𝐼 𝐼 ∈ 𝑅 𝑀 are brought in correspondence with the set of measurements
concerning the same object and held in the same conditions 𝐶𝑜𝑛𝑑 and in the period
𝜏 < 𝑚𝑖𝑛𝑇𝑖 .
• Axiom 3𝑃 . In the space of results of polymetric measuring experiments (in the space
of hypercomplex numbers 𝑍ℎ ∈ 𝑅 𝑀 ) In the space of results of polymetric measuring
experiments (in the space of hypercomplex numbers 𝐶𝑜𝑛𝑑 a generalized result of mea-
surement of the object characteristics – charge 𝑍ℎ as a finite set of values 𝑍 (𝐻 , 𝐶𝑜𝑛𝑑) =<
ℎ1 ℎ2 ...ℎ𝑗 ...ℎ𝑀 > of an assessment of the actual value of object-specific characteristics
are designated. This set of assessments is defined on 𝜎 –algebra of subsets of tuples
𝑍 =< 𝑧1 𝑧2 ...𝑧𝑛 > as real functions 𝑍ℎ = 𝐻𝜏𝐼 /𝐻𝜏𝐼 𝐼 ∈ 𝑅 𝑀 .
The comparison of the basics of classical and polymetric measurement theories shows the
vital importance of not only PMTA but also their consequences (three of them described in [2]
and three new ones presented below). It seems essential to comment on these new practical
consequences of PMTA.
2.2. Consequences of polymetric measurements theory axioms
In the general case for an industrial control system or an intelligent system and robot sensing
process, it is necessary to have 𝑁 instruments for each characteristic under monitoring at
𝑀 possible locations of the components of the object under control. The more practical the
application is, the quicker we face the curse of multidimensionality for such system from both
technical and financial points of view.
It seems rather evident that in the case when initial polymetric signal – tuple 𝑍 has only one
element 𝑍1 we are back to the set of axioms of classical metrology. And this means that poly-
metric theory is just generalization of classical metrology in the case when the initial measuring
signal contains information concerning several physical quantities and their measures.
Furthermore, any polymetric measuring experiment may be presented by the simple model
(𝑍 , 𝐻 , 𝑍ℎ ) – measurable topological space with charge, where the set of tuples 𝑍 as space-
medium is defined in Axiom 1𝑃 , 𝜎 -algebra of subsets 𝐻 as one of the options of topology 𝐻 of
open subsets of set 𝑍 – in Axiom 2𝑃 , and charge 𝑍ℎ , as real function in space 𝑍 on 𝜎 -algebra
of 𝐻 [2].
Thus, the result of the polymetric measurement is the set of uniquely defined functions of
the sequence of charge 𝑍ℎ values. On the other hand, polymetric measurement inaccuracy in
topological space with charge (𝑍 , 𝐻 , 𝑍ℎ ) is defined as the diameter of values of the charge {𝑍ℎ },
i.e. least upper bound between pair of elements of set {𝑍ℎ } [2].
After the introduction of the concept of so-called hyper signals of the third genus 𝐻𝜏𝐼 𝐼 𝐼 ∈ 𝑅 𝑀
the novel set of definishings and consequences of the PMTA can be summarised in such a way:
• Definition 8𝑃 . Deviations of a hyper signal of first genus 𝐻𝜏𝐼 ∈ 𝑅 𝑀 is strongly correlated
with variations of the value of the physical parameter under control and is inherently the
digital assessment of the value of this particular parameter in a fixed time of measurement.
• Definition 9𝑃 . Deviations of a hyper signal of the second genus 𝐻𝜏𝐼 𝐼 ∈ 𝑅 𝑀 is strongly
correlated with variations of the value of the standard of the physical parameter under
control and thus is inherently the digital representation of this particular parameter
standard in a fixed time of measurement.
• Consequence 4𝑃 Deviations of a hyper signal of the third genus 𝐻𝜏𝐼 𝐼 𝐼 ∈ 𝑅 𝑀 are proportional
to variations of the value of the feedback control signal of the system in its critical situation
under human-machine control and thus may be used as automatic reflexive control system
response.
Thus the introduction of the principle of the simultaneous assessment of not only a physical
quantity and its standard but also the value of feedback control signal from the same polymetric
signal is one of the key conditions for developing appropriate polymetric components application
in intelligent automation and smart sensory systems.
The structure of an appropriate perceptive intelligent control and/or monitoring system
should be modified correspondingly.
2.3. Advanced structure of PSS
For multi-agent control or monitoring systems and intelligent robots to satisfactorily fulfil the
potential missions in diversified applications envisioned for them, it is necessary to incorporate
all the above-described advances in the polymetric measurements theory, industrial metrology,
AI, Big Data and Machine Learning instruments.
That is why, in contrast to the multi-sensor perceptive agency concept based on the use of
several measuring transformers each one of them being the sensing part of each particular agent
within the distributed multi-sensor control system (i.e. several perceptive agents in involved
perceptive agency), the use of one equivalent polymetric sensor as a part of perceptive hybrid
agency was proposed [3]. The concept of Polymetric Perceptive / Control Agency (PPCA) for
intelligent system and robot sensing is schematically illustrated in Fig. 1. Such a simplified
structure is designated to be used in different industries and technologies. Four layers are shown
on the Fig. 1: 1 – Analysis Layer, 2 – Experience and Memory layer, 3 – Executive Layer, 4 –
Perceptive Layer.
Figure 1: Scheme of the usage of Equivalent Polymetric Perceptive/Control Agent in frames of Hybrid
Decision-Making and Control Agency
Each layer includes several agents (or agencies) which are intended for their specialized
functions. The explanation of the scheme is reasonable to explain in the opposite order.
Perception agency is realized through the usage of a set of required sensors or/and PSS. As
it was mentioned, in contrast to the classical systems, the PSS requires, for example, only one
sensor for the measurement of all the necessary parameters of the controllable object. PPCA
includes required model to obtain all states parameters of the controllable object and their
uncertainties and can act as a controller to manage the actuators operatively autonomously
or take no action (if required). All the information about measured and estimated parameters
of controllable object are transmitted through available communications channels (commu-
nications bus) to Decision Making and Control Agency (DMCA) which is at level 3. DMCA
is usually MCU or preprogrammed PLC which realize calculations and/or produces control
actions for current control algorithm realization. The information about measured parameters,
steps taken, external to the system information is time-scaled and logged using Storage Agency
– layer 2. Usually, Storage Agency is realized as a database or distributed database in the
cloud. The stored data is a resource available for processing and analysis. Layer 1 includes
Knowledge Discovery Agency (KDA) which stands for Analysis of all gathered data and creation
of algorithms and rules for DMCA and PPCA. KDA can be used both for online and offline data
analysis. It executes several Machine Learning and Preprocessing algorithms to reach goals user
defines. Layer 0 is present but not shown. Layer 0 includes the information from users or other
agencies: initial setpoints, objectives, criteria and various input data. KDA can execute data
preparation (selection, filtering and smoothing, normalization, frequency-domain processing,
etc.), feature extraction related activity (model-based analysis, correlation/statistical parameters
calculation and modelling, Principal Component Analysis, etc.), feature selection, classification
and evaluation of obtained models/results. The results of KDA is an algorithm/knowledgebase /
neural network/classification algorithm or a set of models that are transferred to DMCA and
PPCA in a standardized form of deployable models. The solution about model applicability and
deployment is done on the basis of the evaluation and validation of the model. The essential
advantage of the proposed solution is that models can be synthesized and evaluated/optimized
for a specific application using collected data about the object and its parameters and behaviour.
Therefore, if we assume that traditional and polymetric sensory systems are equivalent in the
measurement information quality and reliability (systems are interchangeable without any loss
of measurement information quality), it is evident that polymetric system has the advantage
in the measurement channels number. The cost criterion can be used for the comparison
of the efficiency of traditional and polymetric sensory systems. And as it can be seen from
the described above LASCOS example, the polymetric sensory system efficiency is two times
greater than the efficiency of the traditional sensory system (𝐸𝑇 𝑆𝑆/𝑃𝑆𝑆 = 44/22 = 2). It is worth
mentioning that this comparison of the system efficiency is very rough, and it is used only for
the demonstration of the main advantage of the polymetric sensory system [2, 3].
It should be noted that according to PSS can be reconfigured to «sense» new parameters
(indirectly) or increase sensitivity for the required parameter with the use of «knowledge» from
KDA. Even if the polymetric sensor has a very low sensitivity to some critical parameter or it
can be used with a high level of uncertainty. It still can be used for monitoring purposes as
an additional source of information about the object, in the form of indicator or its equivalent.
The described technique also facilitates the concept of «measurement as a service» when not
a hardware is being sold but the result of measurements and estimations and uncertainty
evaluation. PSS can be realized as a sensing part installed on a controllable object and a «soft»
sensor – a model or an algorithm for processing the signal or multiple signals and estimation
of required parameters and their uncertainties. Signal or a set of signals is transmitted to the
upper level where appropriate models for parameter estimation are used. In response to such
a request, the upper-level system outputs a set of estimated parameters, their uncertainties
and required control signals, etc. In this case, the informational flows can be directed through
secure internet channels and cloud computing technologies [19, 20].
3. Results and Industrial Applications
Evident practical success had been achieved during wide deployment of SADCO™ polymetric
systems for monitoring of a variety of multiple both quantitative and qualitative indicators of
different liquid or loose ship cargoes. In these systems, single polymetric sensors for measuring
more than three of the mentioned indicators were used. Such an approach also showed positive
technical and commercial results in remote control of complex technological processes for
oil production and refinery control, storage, and commercial stock-taking. Later, the positive
experience was gained in application of PSS within another onshore technologies and entities
also was a success from efficiency, reliability, serviceability point of view.
3.1. Naval applications
One of the first polymetric sensory systems was designated for onboard ship loading and safety
control (LASCOS) for tankers, fishing, offshore, supply and research vessels. The prototypes of
these systems were developed, industrialized, and tested during full-scale sea trials in the early
90-es of the last century. These systems have an intricate hybrid polymetric sensing subsystem
(from the "topos-chronos" compatibility and accuracy points of view) rather.
That is why these systems are also successfully providing commercial control of cargo
handling operations onshore and onboard the ship. The structure of the hardware part of
LASCOS for a typical offshore supply vessel is presented in [2, 3]. It consists of the following:
operator PC with sensor screen interface; a set of sensors for ship draft monitoring; a set
of polymetric sensors for fuel-oil, ballast water and other liquid cargo quantity and quality
monitoring and control. Also, it includes radar complex with antenna, display and a keyboard;
onboard anemometer; a set of polymetric sensors for liquefied LPG or LNG cargo quantity
and quality monitoring and control; switchboards of the subsystem for actuating devices and
operating mechanisms control; a basic electronic block of the subsystem for liquid, liquefied or
friable cargo monitoring and management and a block with the gyroscopic sensors for real-time
monitoring of dynamic parameters of the ship.
The structure of the hardware part of LASCOS for a typical offshore supply vessel is presented
in Figure 2.
Figure 2: The general structure of LASCOSS hardware elements
It consists of the following components: operator workplace (1); a radar antenna (2); an
onboard anemometer (3); the radar display and a keyboard (4); a set of sensors for ship draft
monitoring (5); a set of polymetric sensors for fuel-oil, ballast water and other liquid cargo
quantity and quality monitoring and control (6); a set of polymetric sensors for liquefied LPG
or LNG cargo quantity and quality monitoring and control (7); switchboards of the subsystem
for actuating devices and operating mechanisms control (8); a basic electronic block of the
subsystem for liquid, liquefied and loose cargo monitoring and control (9); specialized electronic
block with the sensors for real-time monitoring of parameters of ship dynamics (10).
The structure of the software part of the typical sensory intelligent LASCOS of the above
kind is also described in [3]. It consists of three main elements (see Fig. 3): a sensory monitoring
agency (SMA) which includes three other sensory monitoring agencies – SSM (sea state, e.g.
wind and wave model parameters), SPM (ship parameters) and NEM (navigation environment
parameters); an information environment agency (INE) including fuzzy holonic models of ship
state (VSM) and weather conditions (WCM), and also data (DB) and knowledge (KB) bases; and
last but not least – an operator interface agency (OPIA) which provides the decision-making
person (DMP) with necessary visual and digital information.
The listed above sensory monitoring subsystems are vital in a ship operations control con-
text, as ship safety refers to the inherently distributed and stochastic perturbation of its state
parameters and external weather excitations. Holonic agents are welcome in the onboard
LASCOS system design because they provide properties such as autonomy, responsiveness,
distributiveness, openness, and redundancy. They can be designed to deal with uncertain and/or
incomplete information and knowledge, and this is exceptionally topical for any fuzzy control
system. On the other hand, the problem of their sensing has never been under systematic
consideration yet. That is why one may assume that the challenges mentioned above have to
be combined in an integrated model of sensing of all agents involved, which may enable the
complex application of cutting-edge and novel methods for efficient solving of practical tasks.
Sensory agent-based computations are adaptive to information changes, and disruptions, exhibit
intelligence and are inherently distributed. Thus, polymetric sensory holonic systems inherently
may help operational control processes in self-recovery reaction to real-time environmental
perturbations. And in such a way polymetric sensor could be efficiently used in much more
diversified industrial applications.
As mentioned above, all initial real-time information for the sensory systems is provided
to DMP by the holonic agent-based sensory monitoring agency (SMA) which includes three
other sensory monitoring agencies – SSM, SPM and NEM. But the hardware part of LASCOS
is somewhat redundant, and that’s why too expensive and barely acceptable in non-critical
operational situations from accuracy and reliability point of view, thus limiting the application
of LASCOS mainly for rough sea safety monitoring and also loading-uploading operations
in the port conditions. In the first scenario, the functions of the system are to predict safe
combinations of ship speed and course in rough sea actual weather conditions. On the safe
storming diagram, the seafarer may see three colored zones of "speed-course" combinations:
the green zone of completely safe ship speed and course combinations; the dangerous red zone
with "speed-course" combinations leading to capsized of the ship with up to more than 95%
probability; and the yellow zone with "speed-course" combinations with intensive ship motions,
preventing standard navigation and other operational activities including bunkering or similar
tasks.
Figure 3: The general structure of LASCOSS software elements and functions
3.2. Maritime end energy infrastructure applications
The first example of the application of PSS for maritime infrastructure technologies is a computer-
aided floating dock ballasting process control and monitoring system (SADCO™ Dock) presented
in [2]. These systems were designated to ensure effective and safe management of docking
operations in the automatic remote mode. The ballast system is one of the most critical systems
on the floating dock [10]. The operation must be finely coordinated with the requirements
of the marine industry. The mentioned PSS enables to take high-tech solutions to support of
safe control and monitoring of the dock ballast system and connected systems. This solution
provides the safe and reliable operations of dock facilities in harsh real-life conditions, reduces
risks for vessels and generates both operational and financial benefits to customers.
The main interface window of the system contains the technological equipment layout and
displays the actual status of valves, pumps, etc. This interface also includes the toolbar, the
information panel and the specialized interface containing control elements. So, it is possible
to start, change parameters or stop technological processes by clicking a particular control
element. The user interface enables the operator to supervise and control every detail of any
technology process efficiently. All information concerning the dock and docked ship safety,
other operations monitoring and control, event and alarm management, database management
or message control are structured in functional windows.
The key advantage of the system is that it can be programmed for various operations scenarios
and operation patterns for each dock with its specificity. In ideal conditions, dockmaster can
control all the data and only select preprogrammed operations patterns with some tunable data.
The data collected during the operations is stored locally and can be used for future training
process (not only for dockmasters training but for system training too – VSM for the dock) or
operations pattern generation. This feature allows users to save time on typical operations and
optimize hardware usage.
Latest versions of the dock-system continuously get additional functionality and updates
of the corresponding software. E.g., for correct operation planning, it is necessary to obtain
information about the current state of pipelines along with the ballast tanks parameters, valves
and pumps parameters. This information gives more flexibility to predict the state of the
controllable object and plan the operation cycle more accurately and reliably using internal
models: flow and pipeline models included as subsystems into complete dock state model.
Additionally, the functionality of dry compartments control was added for leakage control and
periodically flooded compartments. This feature gives more information for the dockmaster. It
can be used to precise the deformation models, especially in non-standard situations, when the
object trajectory must be calculated in the uncertain conditions.
Another example of the successful polymetric sensing of computer-aided control and monitor-
ing systems is connected with different marine/shore terminals (crude oil, light and low-volatility
fuel, diesel, liquefied petroleum gas, grain) and even with bilge water cleaning shops [2, 12]. Such
polymetric sensing allows simultaneous monitoring of practically all the necessary parameters
using one sensor for each tank, silo, or other reservoirs. Especially these systems are useful if the
objects are spatially distributed or their structure/parameters can be dynamically transformed,
and the user needs the current information about all the operations during transportation and
storage. Collected by the PSS data can be processed locally, and the result generated can be
presented for the local users. Still, the whole balance and «process owner view» the distributed
data collection and processing system may also be used. The critical moment of such PSS usage
is that during storage and transportation process, the characteristics of cargo can be changed,
and it must be controlled on each stage of the process – see Fig. 4 for description.
Figure 4: Generalized scheme of data collection and processing of the current information from the
liquid’s storage and transportation hubs
During transportation from manufacturer to the end-user liquid fuel can be transported by
2-3 types of transport and can change its storage several times, can be diluted, etc. On each stage,
the parameters of the controllable fuel or liquid (e.g. palm oil) must be controlled and checked
by the process holder to ensure that the cargo is in safe and operation is going according to
plan. Cloud-based platform and corresponding applications give the possibility to collect the
data from various sources, process collected data and form views and features needed by each
class of clients. Another critical aspect of the use of a cloud-based platform is in the possible
continuous learning of the system used and related to «big data» instruments. As the data
collected can be processed by various methods - various models of processing collected data
for multiple properties calculation can be generated. All the developed models can be verified
on a control sample. If results of a verification are acceptable – the model and appropriate
applications can be deployed into the process, enhancing the PSS usage effect many times.
Namely, a single PSS can be used for control of a set of the following characteristics:
• The level, separation level of the non-mixed media and the online control of upper/lower
critical levels and volume of the corresponding cargo in each tank.
• The temperature field in the media, the temperature of the product at fixed points.
• Density, octane/cetane number of fuels, the propane-butane concentration in LPG, cor-
responding octane number, presence, and percentage of water (including aggressive
chemicals – acids, alkalis, etc.).
The main progress in diversification of PSS application was achieved in very topical tasks
of operational control of quantitative and qualitative parameters of aviation fuels, Liquefied
Petroleum Gas (LPG) terminals operation and safety control systems (SADCO™ GNS) and
borated water parameters control in a spent fuel pool at the nuclear power plant under harsh
conditions (increased radioactivity, high temperatures and pressure values). The main advantage
of the developed polymetric system is that the sought-for fuel quality parameter is functionally
related with several operatively evaluated from polymetric sensor hyper signals such as complex
dielectric permittivity, amplitudes of pulse components of polymetric signal, time delays between
this pulses, etc.). The combined use of operatively evaluated spectral and temporal parameters
allows correct calculating of essential fuel quality parameter 𝑄 [21].
LPG is an important energy carrier and must be controlled on each stage of the path from
the gas-producing plant to the end-users. LPG is a specific product that significantly changes
his properties during its storage and transportation, and those properties must be measured
[22]. Main controllable parameters during custody transfer are volume and mass of the LPG
transferred and related quality parameters – as the density of the phases and chemical composi-
tion. The typical approach of operative LPG parameters estimation in pressurized reservoirs is
to use level sensors, density sensors for liquid and gas phase and the level sensor for water (if
present), pressure sensor, etc. It’s also necessary to control the temperature of the liquid and
gas phase for temperature volume correction. The proposed method was used to collect the
measurement data from various sensors and user inputs (data, related to the parameters that
were not directly measured during custody operations – such as chemical composition or the
source of the product, data about density of the product). This data was combined with the
data from polymetric sensors and corresponding features and models where generated. Via a
comparison of calculated and level sensors data, automatic correction functions for level sensors
were generated. Proposed models were evaluated on the historical data about errors collected
during regular system operation. PSS was improved by applying new knowledge mined from
collected historical data. Another example of the results obtained from is an equation that
connects the dielectric constant of the vapour phase of the LPG 𝜀𝜈 with its density 𝜌𝜈 :
𝜀𝜈 = 1 + 0.00109747 ⋅ 𝜌𝜈 . (1)
This data can be used as an alternative path for the LPG vapour phase density estimation.
Comparison of the results of the model with the data from other sources showed a relative
error of less than 0.7%. It should be noted that if LPG is not clear, includes water and it is
vaporized/condensated in the reservoir – the equation (1) cannot be applied because of the low
accuracy reasons.
As a result, a considerable increase of the measuring performance factor 𝐸𝑇 𝑆𝑆/𝑃𝑆𝑆 was achieved
in each application of PSS up to 4-6 times [2]) with the essential concurrent reduction of the
number of instruments and measuring channels of the monitoring system. All the customers
(more than 50 objects in Ukraine, Qatar, Russia, Uzbekistan, etc.) reported safety increase and
commercial benefits after intensive SADCO™ polymetric monitoring systems deployment.
Unique PSS systems «UNITEK-NPP» for Nuclear Power Plants (NPP) safety control via
measuring the cooling water level in spent nuclear fuel storage basins are designated to operate
in normal conditions, disturbance and accident conditions, including «heavy» accidents, are
successfully working on all 15 NPP units of Ukraine. The essential advantage of such a system
is in its high reliability and self-diagnostics features. The PSS design is based on best practices
of the guided wave radar technology - as preferred technology from the market leaders for
spent fuel pool level measurement [23]. Fig.5 shows the structure of PSS for usage in harsh
conditions.
Figure 5: The structure of the PSS channel used in nuclear power plants
Electronics block (EB) of PSS is in a «safe» zone – outside the containment (marked as C
in Fig. 5). The sensing part of the system is placed in the spent fuel pool (SFP) – where harsh
conditions and various accidents can occur. EB produces signals for Information Control System
(ICS) and can be used as a source of a control signal for level controlled valve (LCV) and Distance
Controlled Pumps (DCP) for pumping of the borated water from Borated Water Reservoirs
(BWR) into SFP. Unlike standard time-domain reflectometry methods (e.g. [24]), the advanced
technology and technique, based on multiple reflections are used. This technique allows to
reduce the error occurred when the harsh conditions are present and to detect a change in the
state of some sensor parts – as a self-diagnostics method.
3.3. Prospective industrial PSS applications
Some special PSS were developed, tested, and integrated into control and monitoring systems
before 2000-2010 in limited numbers for the following industrial applications:
• real-time remote monitoring of the production process of motor lubrication oil (quantita-
tive and qualitative control of the components depending on time and real temperature
deviation during the production process).
• real-time remote control of aggressive chemicals at a water purification shop of the
nuclear power stations (quantitative and qualitative control of the components depending
on time and real temperature deviation during the production process), etc.
Further diversification of PSS application in different new fields and further improvement of
already industrialized systems could be achieved via intensive development of the axiomatic
theory of polymetric measurements and digitalization of all PSS structural components.
Other prospects are connected to utilization of neoteric high precision and reliable radars
which allow finding out unique hyper signals of the third genus directly correlated to optimal
steering command for ship autopilot in parallel assessing of spectral parameters of wave and
wind perturbations. Then the structure of both hardware and software parts of the above
described onboard LASCOS system must be drastically changed, making it much simpler
and more effective from practically all points of view. It will be reasonable to consider some
latest developments in techniques, algorithms and hardware that can be applied to extend PSS
informativeness and efficiency including a possible adaption of technologies for the development
of non-trivial polymetric sensors, e.g. [25, 26, 27].
4. Conclusions
This paper has been intended as a general presentation of new achievements in the development
of the axiomatic theory of polymetric measurements and their influence on potential diversifi-
cation of application of newly developed and already industrialized polymetric components
(subsystems) in intelligent robotics and embedded control systems.
Rapidly developing and very topical area of building up various intelligent robotics and
embedded control systems needs more and more innovations in the theory, software and
hardware development and optimization. For solving such a topical objective, one needs to
implement the leading-edge solutions for each component of such systems, including their
embedded sensory subsystems.
Despite intensive investments and investigations carried out in the described field last decade,
many aspects of the theory and practice of PSS application in different technologies from
a metrological point of view and certification regulations requirements remain to be better
defined.
After the enthusiasm that characterized the first decade of the discovery that more than two
parameters of media may be measured by one single instrument (polymetric sensor), it is now
time to start a serious and coordinated action to define how to diversify such sensors application
for more and more industrial problems, especially for embedding PSS for intelligent robotics
and smart automation. Actual progress in such issue can be achieved only in frames of shared
international research programs involving investigators from all over the world and making
the results widely available.
References
[1] Y. Zhukov, Solitonic models of polymetric monitoring systems, Proc. of International
Conf. "Shipbuilding problems: state, ideas, solutions" (1997) 251–252.
[2] Y. Zhukov, Polymetric Systems: Theory and Practice, Atoll, 2012.
[3] Y. P. Kondratenko, R. J. Duro, Advances in intelligent robotics and collaborative au-
tomation, in: Advances in Intelligent Robotics and Collaborative Automation, River Pub-
lisher, 2015, pp. 1–364. URL: https://doi.org/10.13052/rp-9788793237049. doi:10.13052/
rp-9788793237049.
[4] A. Schütze, N. Helwig, T. Schneider, Sensors 4.0 – smart sensors and measurement
technology enable industry 4.0, Journal of Sensors and Sensor Systems 7 (2018) 359–371.
URL: https://doi.org/10.5194/jsss-7-359-2018. doi:10.5194/jsss-7-359-2018.
[5] D. G. Pascual, P. Daponte, U. Kumar, Handbook of Industry 4.0 and SMART Systems, CRC
Press, 2019.
[6] R. Frank, Understanding smart sensors, Artech House, 2013.
[7] R. Taymanov, K. Sapozhnikova, What makes sensor devices and microsystems “intelligent”
or “smart”?, in: Smart Sensors and MEMs, Elsevier, 2018, pp. 1–22. URL: https://doi.org/10.
1016/b978-0-08-102055-5.00001-2. doi:10.1016/b978-0-08-102055-5.00001-2.
[8] Y. P. Kondratenko, V. V. Korobko, O. V. Korobko, Multisensor data acquisition system for
thermoacoustic processes analysis, in: Proceedings of the 6th IEEE International Confer-
ence on Intelligent Data Acquisition and Advanced Computing Systems, IEEE, 2011. URL:
https://doi.org/10.1109/idaacs.2011.6072710. doi:10.1109/idaacs.2011.6072710.
[9] Y. Kondratenko, O. Korobko, O. Kozlov, O. Gerasin, A. Topalov, PLC based system for
remote liquids level control with radar sensor, in: 2015 IEEE 8th International Confer-
ence on Intelligent Data Acquisition and Advanced Computing Systems: Technology
and Applications (IDAACS), IEEE, 2015. URL: https://doi.org/10.1109/idaacs.2015.7340699.
doi:10.1109/idaacs.2015.7340699.
[10] A. Topalov, O. Kozlov, Y. Kondratenko, Control processes of floating docks based on SCADA
systems with wireless data transmission, in: 2016 XII International Conference on Perspec-
tive Technologies and Methods in MEMS Design (MEMSTECH), IEEE, 2016. URL: https:
//doi.org/10.1109/memstech.2016.7507520. doi:10.1109/memstech.2016.7507520.
[11] N. V. Kirianaki, S. Y. Yurish, N. O. Shpak, V. P. Deynega, Data Acquisition and Signal
Processing for Smart Sensors, John Wiley & Sons, Ltd, 2002. URL: https://doi.org/10.1002/
0470846100. doi:10.1002/0470846100.
[12] S. Cavalieri, L. Fasanotti, S. Ierace, C. E. Pereira, M. Zuccolotto, A biomimetic approach to
distributed maintenance management based on a multi-agent system, in: Industrial Agents,
Elsevier, 2015, pp. 361–379. URL: https://doi.org/10.1016/b978-0-12-800341-1.00020-6.
doi:10.1016/b978-0-12-800341-1.00020-6.
[13] P. Leitão, Agent-based distributed manufacturing control: A state-of-the-art survey,
Engineering Applications of Artificial Intelligence 22 (2009) 979–991. URL: https://doi.org/
10.1016/j.engappai.2008.09.005. doi:10.1016/j.engappai.2008.09.005.
[14] K. Hossain, M. Milton, B. Nimmo, Metrology for the 2020s, 2013.
[15] S. Russell, P. Norvig, Artificial intelligence: a modern approach (2002).
[16] L. Monostori, J. Váncza, S. Kumara, Agent-based systems for manufacturing, CIRP Annals
55 (2006) 697–720. URL: https://doi.org/10.1016/j.cirp.2006.10.004. doi:10.1016/j.cirp.
2006.10.004.
[17] F. Amigoni, A. Brandolini, G. D'Antona, R. Ottoboni, M. Somalvico, Artificial intelligence
in science of measurements: From measurement instruments to perceptive agencies,
IEEE Transactions on Instrumentation and Measurement 52 (2003) 716–723. URL: https:
//doi.org/10.1109/tim.2003.814694. doi:10.1109/tim.2003.814694.
[18] E. Alpaydin, Introduction to machine learning, MIT press, 2020.
[19] M. N. Birje, P. S. Challagidad, R. Goudar, M. T. Tapale, Cloud computing review: concepts,
technology, challenges and security, International Journal of Cloud Computing 6 (2017)
32. URL: https://doi.org/10.1504/ijcc.2017.083905. doi:10.1504/ijcc.2017.083905.
[20] G. Kumar, A review on data protection of cloud computing security, benefits, risks and
suggestions, United International Journal for Research and Technology (2019) 26–34.
[21] Y. Zhukov, B. Gordeev, A. Zivenko, Polymetric method of operational control of qualitative
and quantitative characteristics of aviation fuels, in: Proc. of 5th World Congress: Aviation
in the XXI century, , 2012.
[22] O. Zivenko, LPG ACCOUNTING SPECIFICITY DURING ITS STORAGE AND TRANS-
PORTATION, Measuring Equipment and Metrology 80 (2019) 21–27. URL: https://doi.org/
10.23939/istcmtm2019.03.021. doi:10.23939/istcmtm2019.03.021.
[23] S. Ansari, Building Computer Vision Applications Using Artificial Neural Net-
works, Apress, 2020. URL: https://doi.org/10.1007/978-1-4842-5887-3. doi:10.1007/
978-1-4842-5887-3.
[24] V. Melnikov, V. Ivanov, I. Teplyashin, M. Timonin, Development and study of a microwave
reflex-radar level gauge of the nuclear reactor coolant, Nuclear Energy and Technol-
ogy 4 (2018) 185–190. URL: https://doi.org/10.3897/nucet.4.31857. doi:10.3897/nucet.4.
31857.
[25] B. Zhang, Y. Pi, J. Li, Terahertz imaging radar with inverse aperture synthesis techniques:
System structure, signal processing, and experiment results, IEEE Sensors Journal 15
(2015) 290–299. URL: https://doi.org/10.1109/jsen.2014.2342495. doi:10.1109/jsen.2014.
2342495.
[26] J. A. Cantoral-Ceballos, N. Nurgiyatna, P. Wright, J. Vaughan, C. Brown-Wilson, P. J. Scully,
K. B. Ozanyan, Intelligent carpet system, based on photonic guided-path tomography, for
gait and balance monitoring in home environments, IEEE Sensors Journal 15 (2015) 279–289.
URL: https://doi.org/10.1109/jsen.2014.2341455. doi:10.1109/jsen.2014.2341455.
[27] S. Wang, J. Song, J. Lien, I. Poupyrev, O. Hilliges, Interacting with soli, in: Proceedings of
the 29th Annual Symposium on User Interface Software and Technology, ACM, 2016. URL:
https://doi.org/10.1145/2984511.2984565. doi:10.1145/2984511.2984565.