Ncode GlyphWorks Software Use For Test Data Processing
Ncode GlyphWorks Software Use For Test Data Processing
Abstract
Some technique and results of experimental data processing on the stress-and-strain state
study of wide-spray field sprayer boom parts have been described in the paper under
discussion. The test data have been corrected to remove the data drift effect. Two approaches
have been applied to correct the test data, namely the use of high frequency filter and the
removal of the data array average values. The results of test data correction by the above-
mentioned methods have intercommunicated very well. The data have been processed by
nCode GlyphWorks software aimed at preparation for the further use of the obtained results
to estimate the structure durability.
Keywords 1
Experiment, test data, signal, processing, software, glyph, data correction
1. Introduction
Measurement, i.e. experimental determination of the values of physical quantities by special
hardware, namely measuring apparatuses, is the basic way to obtain some information about the
environment and the processes occurring in it. The measurements are the only way to prove or reject
any statements or conclusions of theoretical models describing the real objects behavior.
Any experimental research has always involved some complex and time-consuming procedures
dealing with data collecting and processing, analysis of the obtained information and construction of
processes models and fields of various nature. Two approaches have been known to provide the
analogue measuring signals recording aimed at further processing by digital methods [1].
One of the approaches is based on specialized complex systems use involving the equipment of
analogue-digital conversion, microprocessors of digital processing and devices of information
transformation. Another approach is based on the use of interface devices of data collection and
universal computer systems.
As for the advantages of the second approach, based on the use of some additional interface
modules and signals digital processors as parts of a personal computer, we can list the following ones:
flexibility of a measurement system regarding the implementation of different algorithms of
processing; functional completeness of the system (the tasks of data input, processing, management,
analysis, visualization, measurements data and analysis results storage are being solved); good
metrological characteristics.
In various spheres of reality, in particular, in economics, biology, medicine, technology etc are
widespread conditional cyclic random processes and phenomena. The study of cyclical processes
involving modern information systems requires the preliminary development of adequate
mathematical models for them. Many different mathematical models of cyclic processes are known
today, including harmonic, periodic and almost periodic deterministic functions, periodically
correlated and periodically distributed random processes, linear periodic random process, almost
periodically correlated random process, cyclic random process [2-6].
ITTAP’2021: 1nd International Workshop on Information Technologies: Theoretical and Applied Problems, November 16–18, 2021,
Ternopil, Ukraine
EMAIL: stashkiv@tntu.edu.ua (M. Stashkiv); oleksandr.matsiuk@mail.com (O. Matsiuk);
ORCID: 0000-0002-7325-8016 (M. Stashkiv); 0000-0003-0204-3971 (O. Matsiuk)
©️ 2021 Copyright for this paper by its authors.
Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR Workshop Proceedings (CEUR-WS.org)
In particular, the work [7] is devoted to the development of mathematical modeling of digital
cyclic signals with double stochasticity, namely, the construction of their mathematical model in the
form of a conditional cyclic random process of discrete argument.
In the paper [8] considers the unified approach to the modeling and processing signals of electrical,
magnetic and acoustic (mechanical) nature based on the model of the theory of cyclic random
functions, namely, using cyclic random process and vector of cyclic rhythmically related random
processes.
In the paper [9] the new methods of statistical analysis of heart rhythm based on its generalized
mathematical model in a form of random rhythm function, that allows to increase the informativeness
and detailed analysis of heart rhythm in cardiovascular information systems are offered.
Nowadays, there is a great number of hardware and software for signals obtaining, segmentation,
statistical analysis, processing and modeling. Although, many scientists have developed their own
software to solve certain problems. It can be either a single-purpose software [10, 11, 12] or the whole
program complexes [13].
Among the whole range of various software for the work with experimental data the software
nCode made by the company HBM Prenscia [14, 15] has been of great interest, which was designed
to process the signals, to work with large volumes of test data and calculation of the fatigue durability
of the products.
nCode GlyphWorks software has a convenient object-oriented graphical interface and it is
optimized for the complex work with large volumes of multi-channel data. The software functional
includes a wide range of tools for temporary, frequency and statistical analysis of signals. Moreover,
some tools are available to assess both the resource and fatigue durability of the products and
synchronized reproduction of GPS signals, video and other data obtained during the tests, and also a
convenient mechanism for automated creation of reports has been implemented.
nCode GlyphWorks software has a module system which enables us to create the required feature
set either by means of large embedded libraries or by the use of language Python. Working templates
in nCode GlyphWorks can be prepared in advance and can be found in the specified library to provide
reliability and high speed of calculation. Thus, the users do not need to create the required feature set
by themselves.
For effective work with such software it is necessary to have large sets of experimental data. The
formation of such test data sets is an important phase of any scientific research, especially in modeling
the dynamic load of parts and components of large mobile agricultural machinery.
Carrying out full-scale field tests of agricultural machinery involves significant costs and certain
organizational problems, especially during a COVID-19 pandemic.
One of the ways to solve this problem is to conduct laboratory tests. However, such studies do not
allow for a full set dynamic loading research of the large mobile agricultural machinery bearing
structures.
To solve this problem the authors [16] developed a test bed for stationary semi-natural studies of
the dynamic loading of wide-spray field sprayer booms. The test bed design is protected by a utility
model patent [16]. Description of the design of the test bed for the study of the wide-spray field
sprayer booms load dynamics and methods of semi-natural research are presented in [17].
The results of any measurements, despite the accuracy of the procedure, are likely to have some
errors. It may be caused by the impact of different external factors on the measurement process result
in an offset on the sensor output. This effect should be taken onto account as it may cause some errors
in the results. In this case, the analysis of obtained results and errors of measurement has been an
essential part of any scientific experiment, so a researcher must be able to apply some methods of test
data correction.
The results of measurements are recorded as a 12-bit code and are given by the universal
measuring system as DATA.DAT file where the data of each channel is recorded in a row as digital
codes separated by a separating character whose form is prescribed in previous settings of the system
in MODE.DAT file. Thus, in the obtained data array the information from each specific channel of the
universal measuring system is recorded in every eighth log. These data flow is a set of different digital
codes which should be given in more convenient for calculations form. To achieve this goal there is a
special technique of preliminary processing of test data prior to their possible use in further
calculations.
A special software providing the data sorting of the obtained array to certain channels and reducing
the digital codes of each channel to the real physical quantities with dimensional units (such as, for
example, deformation (mm) or stress (МРА)) has been developed [12] for test data preliminary
processing. The program interface view illustration (interface is in Ukrainian) is shown on fig. 3.
Figure 3: Program interface to converting test data array obtained from UMS
Converting of the digital codes of data array to the real physical quantities with dimensional units
has been provided by formula (see the program interface on fig. 3) by means of taking into account a
number of coefficients: zero level of data (Нуль =); tensosensors calibration test coefficient,
amplification factor, and others (К1–К4); coefficient of static stress level (К5). The level of static
stress should be taken into account when tensosensors are installed on preliminary loaded parts of a
bearing structure. For analytical determination of static stress level in the components of the boom
structure under discussion a mathematical model has been developed which describes the distribution
of internal efforts in the sprayer boom parts on the basis of the potential energy of deformation
minimizing method [18].
Due to the test data array processing by the developed program [12] we have obtained eight files
ChannelN.txt (where N = 1 – 8 matches the channel order number) with the data prepared for the
signal processing from each channel. The data from the first four separate tensometric channels of the
universal measuring system (fig. 4) is a function of time, their amplitude and level are constantly
varying.
Channel 1 Channel 2
Channel 3 Channel 4
Figure 4: Test data of tensometric channels of UMS at test of the sprayer boom
Despite the curves start from the nominal zero, a so-called drift (offset) can be observed here
which varies between the channels and non-linearly over time. It has been caused by the impact of
different external factors (change in temperature, screen grid guidance, contacts reduction due to the
vibration action, humidity etc.) on the measurement process resulted in an offset on the sensor output.
This caused a final value displacement at the end of the test that is no longer the expected zero
value. This effect should be taken onto account as it may cause some errors in the results which are
important in assessment of the structure fatigue durability.
2
4
3
Figure 5: The main windows of the GlyphWorks interface
GlyphWorks glyphs are classified according to their functionality. The glyphs are available in the
glyph palette (the window on the right side на рис. 5). There are sections in the glyph palette for
glyphs that input data, perform basic digital signal processing (DSP), display results, and so on.
Glyphs are organized into the following categories (palettes), according to their functionality:
Input; Function; BasicDSP;Signal; DesignLife; Frequency; Fatigue; AcceleratedTesting; Optimized
Testing; GlyphBuilder; SuperGlyph; Display and Output.
Glyphs and input files, etc., can be dragged onto the workspace from their respective palettes.
In GlyphWorks, a process is defined as a combination of glyphs that define a data flow. A process
typically starts with an input glyph to define the data to be processed. Additional glyphs define
subsequent steps in the process for calculation, display, or writing output.
To implement the procedure of test data correction a detailed design has been developed
(fig. 6) containing the following structural elements (glyphs): Exel Input, Multi-Column To
Time Series, Running Statistics, Arithmetic, Time Series Output, XY Display. Functional
purpose, parameters settings of these glyphs and structural relations between glyphs are
described lower.
Structural elements (glyphs) with relations can be combined conditionally into functional
blocks (fig.6) where the successive steps of test data correction procedure are taking place
and the obtained results are represented as curves:
І – a block of incoming data input,
ІІ – a block of incoming data statistical processing,
ІІІ – a block of mathematical operations on incoming data,
ІV – a block of results analysis and comparison,
V – a block of results saving.
II
III
V IV
Figure 6: The process flow that removes running mean of the sprayer boom test data (completed)
To input some test data into the detailed design an executable file .xls has been formed where the
data columns from binary files were added including the results of primary processing of four
tensometric channels (fig.4). Above each data column the lines are located with channel name
(Channel1- Channel4) and units of measurements (МРА).
The obtained executable file .xls was uploaded to the glyph of multicolumn data ExelInput1 (fig.6,
block І). For correct representation of the uploaded data in glyph ExelInput1 settings it is necessary
(fig. 7):
– in section Column Title Cells the tape cells with channels name should be specified;
– in section Column Units Cells the tape cells with dimensional values should be specified;
– in section Value Cells the tape cells with the initial data of the columns should be specified;
– Process rows up to first empty row should be specified.
MultiColumnToTimeSeries glyph (fig.6, block І) is used for the conversion of multi-column data
into the data of temporal series. The glyph is connected to the orange output pad on ExcelInput1.
The following parameters should be specified in the properties of this glyph (fig. 8):
Method → Resample;
SampleRate → 500 (fig. 2);
Interpolation → Linear.
A special glyph Running Statistics has been used for incoming data statistical processing enabling
us to process the data by methods of mathematical statistics using the special methods nCode. This
glyph input is connected with the glyph MultiColumnToTimeSeries output.
In the settings of calculation template RunningStats1 in Tab General (fig. 10) we have chosen a
type of data statistical processing Мean (by average value). The parameter ProcessingMethod
determines how the current data will be processed. We have chosen the method PointByPoint where
each value of test data will be processed in a series resulted in a set of sequential measurements and
the general statistics of the experiment will be formed on their basis. Such method guarantees that the
source file with processed data will have the same amount of data that the initial original file contains.
In the tab WindowLength we have specified the number of displayed points in the parameters of
visualization – 1000.
The results of statistical processing of test data from Channel 3 by their average value are shown
on fig. 11. On fig. 6 this glyph is renamed as Running Mean Calculated. The first input of this glyph
is connected with the output of MultiColumnToTimeSeries and the second input is connected with
RunningStats1. Thus, the upper curve shows the data incoming array from Channel 3, and the lower
one represents their average values. It has made possible to follow the data offset on the curves and to
estimate the drift of the most important values of the results.
Glyph Arithmetic1 makes possible to perform mathematical operations on the data prepared in
advance. The first stroke of this glyph is connected with the output MultiColumnToTimeSeriesand the
second output is connected with RunningStats1.
In settings of the glyph Arithmetic1 (fig. 12) in the tab General in section Operator we have chosen
the type of arithmetical operation Equation. In section EquationDefinition we write the parameter
Equation as «Test 1 – Test 2» (difference in incoming and processed in the module RunningStats1
data array).
Figure 12: Arithmetic glyph properties
The results of test data processing by glyph Arithmetic are shown on fig. 13. On fig. 6 this glyph is
renamed as Running Mean Removed. This glyph input is connected with the glyph Arithmetic1
output. The obtained data have not included “spurious” current data yet which can alter the general
operating characteristics of the unit under discussion.
After calculations in the graphic glyph XY Display (entitled Data Comparison on fig. 6) two
curves will be displayed (fig. 15, а) obtained due to the incoming data array processing by two
different methods – by average values subtraction of the initial data array and using the filter of high
frequencies (the curves only from the first channel are shown on the figure). To draw the curves the
first input of the glyph Data Comparison was connected with the glyph Arithmetic1, the second one –
with the glyph ButterworthFilter2 output.
a) b)
Figure 15: Comparison of test data array correction by two methods
To make the comparison more convenient these curves have been superimposed (fig. 15, b).
As we can see on fig. 15 the results of test data correction by two methods intercommunicate very
well. Moreover, both methods from the engineering point of view correctly describe the conversion
procedure of the results obtained by the method of signals digital capture from the tensosensors
installed on the real structure. In these corrected data the drift of average values of the process indices
has already taken into account and we can be sure that the amplitude will vary correctly enabling to
determine the real number of cycles of the structure loading according to the techniques and
theoretical substantiation of the above-mentioned results [19].
The following ways are mostly used in calculations: of maximums, of ranges (amplitudes), of
augmented ranges, of complete cycles, "rain-flow" method. The main advantage of these ways is that
they do not restrict the schematized process, i.e. any mode of loading can be processed by these
methods [20].
By means of the glyph TSOutput2 the obtained results, which are the function in time, are
recorded in a separate file of the pattern *.s3t with the prefix _out. The obtained file with the
processed and prepared test data can be further used to calculate the fatigue durability of the structure.
3. Conclusion
The results of any experimental research dealing with data collecting and processing, are likely to
have some errors. The analysis of obtained results and errors of measurement is an essential part of
any scientific experiment, so a researcher must be able to apply some methods of test data correction.
At present, there is a great number of software for signals obtaining, segmentation, statistical
analysis and processing. The software nCode, which was designed to work with large volumes of test
data, makes possible to carry out processing of experimental signals, both by standard methods and by
the developed alternative techniques of data processing.
In the paper under discussion the drift of experimental data has been corrected so that to reduce the
test data curves to the nominal zero level. Two approaches have been applied to test data correction to
remove the drift effect, namely the use of high frequency filter and the removal of the data array
average values. The results of test data correction by two methods have intercommunicated very well.
Both methods from the engineering point of view correctly describe the conversion procedure of
the results obtained by the method of signals digital capture from the tensosensors installed on the real
structure. The obtained results can be used in carrying further out research of the structure durability.
4. Acknowledgements
I am very grateful to the company HBM Prenscia and the team nCode for the possibility to use
their software and for the information support. My special thanks to Lukasz Pieniak – Account
Manager Prenscia.
I highly appreciate the assistance of the managers of the company «Bohuslav agricultural
machinery» Havrylenko P.М. and Havrylenko М.P. in their cooperation with the department of
technical mechanics and agricultural machines of Ternopil Ivan Puluj national technical university
during the experimental studies of the dynamic loading of wide-spray field sprayer boom.
5. References
[1] V.V. Shiriaev, Computer measuring instruments (KIS): Textbook, Tomsk, 2008. [in Russian].
[2] H. Pishro-Nik, Introduction to Probability, Statistics, and Random Processes, USA, 2016, 744.
ISBN 978-0-9906372-0-2.
[3] A. C. Micheas, Theory of Stochastic Objects Probability, Stochastic Processes and Inference,
CRC Press, 2018, 408. ISBN: 9781466515215.
[4] P. Olofsson, M. Andersson, Probability, Statistics, and Stochastic Processes, 2nd Edition,
Canada, 2014, 576. ISBN-13: 978-0470889749.
[5] S. Miller, D. Childers, Probability and Random Processes: With Applications to Signal
Processing and Communications, 2nd Edition, Academic Press, 2012. ISBN: 9780123869814.
[6] H. Kobayashi, B. L. Mark, W. Turin, Probability, Random Processes, and Statistical Analysis,
Cambridge University Press, 2011, 812. DOI: 10.1017/CBO9780511977770.
[7] S. Lupenko, I. Lytvynenko, N. Stadnyk, A. Zozulia, Model of signals with double stochasticity in
the form of a conditional cyclic random process (2020), 2762, 201-208. 229366904.
[8] S. Lupenko, O. Orobchuk, N. Stadnik, A. Zozulya, Modeling and signals processing using cyclic
random functions. 13th IEEE International Scientific and Technical Conference on Computer
Sciences and Information Technologies (CSIT), Ukraine, Lviv, 2018, T. 1, pp. 360-363.
ISBN 978-1-5386-6463-6.
[9] S. Lupenko, N. Lutsyk, O. Yasniy, L. Sobaszek, Statistical analysis of human heart rhythm with
increased informativeness. Acta Mechanica et Automatica. Vol. 12: Issue 4 (2018) pp. 311-315.
DOI: 10.2478/ama-2018-0047.
[10] I. V. Lytvynenko, P. O. Maruschak, S. A. Lupenko, Yu. I. Hats, A. Menou and S. V. Panin.
Software for segmentation, statistical analysis and modeling of surface ordered structures. AIP
Conference Proceedings 1785, 030012 (2016). DOI: 10.1063/1.4967033.
[11] S. Lupenko, I. Lytvynenko, A. Sverstiuk, A. Horkunenko, B. Shelestovskyi, Software for
statistical processing and modeling of a set of synchronously registered cardio signals of different
physical nature, 2021, 2864, рр. 194-205. 234753381.
[12] T. Rybak, Yu. Gladyo, M. Stashkiv, P. Palamarchuk, Software "File converter for experimental
data RGSP-v 1.0" / Certificate of registration of copyright to the work № 63257. 29.12.2015.
[in Ukrainian].
[13] I. Lytvynenko, P. Onyskiv, S.Lupenko, A. Zozulia, Processing of Vector Rhythmocardiosignal
Based on Software Complex. International Scientific and Technical Conference on Computer
Sciences and Information Technologies, 2020, 1, pp. 411-414. 9322020.
[14] https://www.hbmprenscia.com/
[15] https://www.ncode.com/
[16] M. Stashkiv, P. Havrilenko, M. Havrilenko, P. Palamarchuk, Ih. Bortnyk, Stand for dynamic
loads testing of sprayers boom. Utility model patent № 118740 (u 2017 01793). 08/28/2017. №
16. [in Ukrainian].
[17] P. Palamarchuk, M. Havrylenko, M. Stashkiv and Ih. Bortnyk, Stand tests of crop sprayer booms.
National interdepartmental scientific and technical collection "Design, manufacture and
operation of agricultural machinery", Kirovograd: KNTU, 2015, Issue 45, Part II, pp. 98 - 102.
[in Ukrainian].
[18] R. Hevko, M. Stashkiv, O. Lyashuk, Y. Vovk, V. Oleksyuk, O. Tson, I. Bortnyk, Investigation of
internal efforts in the components of the crop sprayer boom section. Journal of Achievements in
Materials and Manufacturing Engineering. Volume 105, Issue 1 (2021), pp. 33 – 41.
DOI: 10.5604/01.3001.0014.8743.
[19] T. Rybak, M. Pidgurskyi, M. Stashkiv Prediction of mobile agricultural machines load-bearing
systems durability based on the results of operational tests. Motorization and power industry in
agriculture (MOTROL), Lublin, 2007, V. 9A, 86–92. [in Ukrainian].
[20] Bogdan Ligaj. An analysis of the influence of cycle counting methods on fatigue life calculations
of steel. Scientific problems of machines operation and maintenance, 4 (168) 2011, 25-43.
122489329.