See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/269511957

Nonlinear Trend Analysis of Mill Fan System Vibrations for Predictive

Maintenance and Diagnostics

Article in International Journal of Electronics and Telecommunications · December 2012

DOI: 10.2478/v10177-012-0048-9

CITATION READS

1 83

3 authors, including:

Mincho Hadjiski Lyubka Doukovska

University of Chemical Technology and Metallurgy Bulgarian Academy of Sciences
41 PUBLICATIONS 140 CITATIONS 156 PUBLICATIONS 868 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Lyubka Doukovska on 30 January 2017.

The user has requested enhancement of the downloaded file.

 INTL JOURNAL OF ELECTRONICS AND TELECOMMUNICATIONS, 2012, VOL. 58, NO. 4, PP. 351–356
Manuscript received October 1, 2012; revised November, 2012. DOI: 10.2478/v10177-012-0048-9

Nonlinear Trend Analysis of Mill Fan System

Vibrations for Predictive Maintenance and
Diagnostics
Mincho B. Hadjiski, Lyubka A. Doukovska, and Stefan L. Kojnov

Abstract—Present paper considers nonlinear trend analysis for engineering – specialist maintenance and repair, operators
diagnostics and predictive maintenance. The subject is a device and dispatchers, designers of industrial systems with great
from Maritsa East 2 thermal power plant a mill fan. The choice complexity, so as to ensure high reliability, fault tolerance and
of the given power plant is not occasional. This is the largest
thermal power plant on the Balkan Peninsula. Mill fans are availability. One of the study tasks is evaluating the probability
main part of the fuel preparation in the coal fired power plants. of failure of element, machine or center as a time function.
The possibility to predict eventual damages or wear out without This is important from engineering point of view in order to
switching off the device is significant for providing faultless and adjust the machine inspection policy.
reliable work avoiding the losses caused by planned maintenance.
This paper addresses the needs of the Maritsa East 2 Complex
According to the International Standardization Organization
aiming to improve the ecological parameters of the electro energy (ISO) ”Prognostics is time for estimation of damage and risk
production process. for one or several future damages”, [1]. Thus technological
Keywords—Technical diagnosis, fault diagnosis, predictive
diagnosis can be understood as a process of estimation of
maintenance, manufacturing execution system (MES), enterprise Remaining Useful Life (RUL) before damage occurs, which
resource planning (ERP). is estimated based on the current status of the facility and last
operating mode [2].
The current state can be considered as the degree of im-
plementation of the specified functionality. The diagnostics is
I. I NTRODUCTION not classified in only two or three classes, but is conceived as
a continuous process of degradation. Quantitative evaluation
T HE predictive maintenance includes four stages: predic-
tive diagnosis, estimation of potential looses, decision
making for device maintenance and maintenance schedule ar-
of the degree of degradation is the result of applying the
diagnostic methods on the current measurement data from sen-
rangement. Technological diagnosis as the basis for predictive sors and Supervisory Control And Data Acquisition (SCADA)
maintenance is established field of scientific and applied inves- or Decentralized Control System (DCS) measurements with
tigations. Top-managers of leading factories in the chemical, special diagnostic equipment. Thus predictive maintenance
petrochemical, silicate industry, energy, metallurgy, mining should be seen as a process of prediction and evaluation
industry in Europe and around the world have adopted in their of future behavior [3], [4]. Critical stage is forecasting the
current practice that economic success and competitiveness development of degradation of the facility. Highly debatable
depend heavily on the security of technological facilities and issue is the assessment of the accuracy of the forecast. Two
low cost of maintenance them. Predictive maintenance based approaches are outlined for formulation of the metrics for
on diagnosis, prolongs the life of machines and aggregates, evaluation [3], [5]. One of them is connected to risk assessment
reducing downtime, maintain optimal level of production, of a forecast. The other is defined based on the quality of the
ensure compliance with the precise timing of delivery of action taken for preventive or corrective maintenance.
production (raw materials, energy), allows for effective man- The second approach is based on measurements and data –
agement of maintenance of facilities, using the least staff and driven. There are numerous other approaches, but all can refer
costs, planning and repairs according to actual conditions of to two types of techniques – artificial intelligence (AI) and
machines and in the most appropriate time in accordance with the theory of probability and mathematical statistics. Between
the tasks of the subsystems of Enterprise Resource Planning the artificial intelligence techniques have become particularly
(ERP). The results of the system to predictive maintenance popular the methods of artificial neural networks and neuro-
are critical for a number of user groups from industry and fuzzy networks. These are flexible and general approaches for
predictive maintenance [6]. It is shown that their effective will
This work has been supported by the National Science Fund of Bulgaria be improved if is using different techniques such as:
under the Project no. TK-01-485/09.
Mincho B. Hadjiski is from the University of Chemical Technology and • radial-based recurrent neural networks [7],
Metallurgy Sofia, Bulv. St. Kliment Ohridski 8, 1756 Sofia, Bulgaria (e-mail:
[email protected]).
• wavelet neural networks [8],
Lyubka A. Doukovska and Stefan L. Kojnov are from the Institute • and robust neural networks [9].
of Information and Communication Technologies, Bulgarian Academy of
Sciences, Acad. G. Bonchev Str., bl. 2, 1113 Sofia, Bulgaria (e-mails: The results are illustrated by different examples using ma-
[email protected]; [email protected]). chines with rotating parts [10].

Unauthenticated
Download Date | 1/30/17 12:31 PM
352 M. B. HADJISKI, L. A. DOUKOVSKA, S. L. KOJNOV

In last few years is increasing interest in the diagnosis paper is to provide a brief overview of various methods and
based on the method of precedents (Case-Based Reasoning operations modules that would particularly share information
– CBR) [11]. This method is effective in the absence of with the fault diagnosis module and also outline the nature of
sufficient measurements and especially for analysis and re- interaction that one can expect.
trieval of precedents with their attributes and relationships.
The adaptation procedure is implemented in the base of rules II. P ROBLEM F ORMULATION
derived from experts. The results in paper [12] show that it is
appropriate to combine the diagnostic method based on the In the presented paper we have chosen to analyze a device
precedents with neural network or fuzzy logic, which will from Maritsa East 2 power plant – a mill fan. Maritsa East
mitigate the disadvantages of the method. Similar results are 2 thermal power plant (TPP) has built up eight blocks – 4
contained in [13]. × 175 MW and 4 × 210 MW. In historic plan in 1962 a
The predictive maintenance, as a complex interdisciplinary decision has been taken for building up Maritsa East 2 TPP,
field of science, technology and economy is far from com- and since 1970 the electro energy of least price cost for the
pletion. Its application in the industry faces a number of country is produced in Maritsa East 2 TPP. At the end of 1995
complexities. They are related to problems of technological 8th energy block has been connected in parallel to the energy
diagnostics (limited opportunities for application of active system of the country by which the second stage of Maritsa
diagnosis, lack of repeatability in testing objects, extremely East 2 TPP enlargement was completed. Achieved installed
large complex of facilities, equipment variability in their life capacity is 1450 MW. This turns Maritsa East 2 TPP into the
cycle, inevitably a high degree of uncertainty), procedures for biggest thermal power plant in the Balkans. After following
decision making for proactive or corrective actions to maintain, reconstructions and modernizations installed capacity at the
the methods for the scheduling of the action, integration moment reaches 1556 MW as in the end of 2009 block 6
into Manufacturing Execution System (MES) and Enterprise was cut off for the purpose of modernization and increasing
Resource Planning (ERP) systems. its capacity to 230 MW. The Maritsa East 2 TPP being the
From industrial application viewpoint, the maximal number largest thermal power plant on the Balkan Peninsula and the
of fault diagnostic applications in process industries are based choice of the given power plant is not occasional.
on process historian based approaches. This is due to the The mill fans are used to mill, dry and feed the coal to the
fact that process history-based approaches are implementing burners of the furnace chamber. They are together milling and
modeling an a priori knowledge. Further, even for processes transporting devices. Mill fans are most often used for power
for which models are available, the models are usually steady plants burning brown and lignite coal. In general these are
– state models. It would require considerable effort to de- large centrifugal fans which suck flue gases with temperature
velop dynamic models specialized towards fault diagnosis around 800-1000 °Cfrom the top of the furnace chamber. In
applications. The scope of the process history based systems the same pipe the coal is fed, thus diminishing the drying
as applied in the industry is mainly restricted to sensor agent temperature and drying the coal prior entering the fan.
failures. There are very few industrial applications in published The coal is being milled by the fast rotating rotor of the fan
literature that deal with parametric failures. Among the process and turn into coal dust. This dust is transferred to separator
history based approaches, statistical approaches seem to have which returns the bigger particles to the fan.
been well studied and applied. The reason for this might be The separator can be tuned for a desired dust granulometric
that with the current state-of-the-art in applications, detection size. One of the most important parameters to control is
seems to be a bigger concern than detailed diagnosis. Hence, the discharge temperature of the dust-air mixture. For the
statistical approaches that are easy to build and which do considered mill fan it should be between 145-195 °C. Lower
very well on fast detection of abnormal situations have been than 145 °Cmay cause clogging of the mill and higher than
successful in industrial applications. Chemical processes are 195 °Cmay cause the dust to be fired in the ducts prior the
inherently nonlinear in nature. While the theory of linear quan- burners. This temperature is also a measurement for the load
titative model-based approaches is quite mature, the design and of the mill. The lower the temperature the higher the load is
implementation for nonlinear models is still an open issue. more coal is fed to the mill. The part which suffers the most
Fault diagnosis shares with other process operations and should be taken care of is the rotor of the mill fan. Because
the realization that with powerful knowledge representation of the abrasive effect of the coal it wears out and should be
schemes, one can capture the expertise of operators and repaired by welding to add more metal to the worn out blades.
control engineers that was gained over years of experience The mill fans (MF) are centrifugal fans of the simplest type
with process plants. Process specific knowledge can be used with flat radial blades adapted for simultaneous operation both
to improve general purpose methodologies. There is a close like fans and also like mills. The basic components of a mill
coupling between diagnosis and process operations design fan are (Fig. 1): steel body , covered by bar shield plates 2
of chemical plants. The proper design of a chemical plant thick 7080 mm, and a rotor, consisting of a mill wheel 3 with
can reduce the burden on the task of diagnosis. Also, the twelve blades 6. The rotor is fixed by a console to a shaft 4,
information from diagnosis can be used to continuously im- located upon two roll bearings [14].
prove the performance of process operations. The information The mill wheel blades 1 (Fig. 2) bear shield beats 2 thick
from fault diagnosis can be incorporated into the traditional 30-40 mm, directly beating the coal particles. The fuel entering
solution paradigms of other process operations. The aim of this the mill is sucked to the beating elements of the mill-fan

Unauthenticated
Download Date | 1/30/17 12:31 PM
NONLINEAR TREND ANALYSIS OF MILL FAN SYSTEM VIBRATIONS FOR PREDICTIVE MAINTENANCE AND DIAGNOSTICS 353

Fig. 1. Mill fan.

Fig. 2. Mill fan rotor.

rotor together with the drying agent (usually hot gases with a
temperature of 950-1100o C). The located before the mill fan As drying agent are used hot flue gases from the furnace
descending mine is the place for intensive drying of the fuel chamber with low oxygen content which makes the process
where a good deal of the external moisture is reduced; in the explosion proof for very high temperatures. This process also
milling zone, i.e. between the shield bars -blades an additional diminishes the nitrogen oxides emissions. These features make
drying takes place and also an effective milling of the dried the mill fan system suitable for boilers firing low caloricity
fuel. lignite.
The ground product with the cooled drying agent is forced For 210 MW power units the milling rotor has diameter
by the dozing rotor of the aggregate into a centrifugal or D = 3.4m, width b = 0.9m and rotation speed n = 490rpm.
inertial separator 3 from where the big particles return to the There is also a system for drying agent temperature control.
mill and the small ones are carried out by the gas stream Such mill fan is shown in Fig. 1.
through the dust pipes to the burners. At the output end of Drying agent flow depends on the amount of coal to be
the pipe for returning big particles from the separator a cone milled. Thus if the mill gets loaded the discharge temperature
valve is set to regulate the amount of recycling gas from decreases and vice versa.
the separator downwards to the bottom of the fuel-delivering Coal and flue gases enter the mill through a special duct
drying mine to induce into the mill fan. taking gases from the top of the furnace chamber. The rotor
Whenever it is necessary to reduce the intensiveness of dry- works as fan and a mill. After milling the coal dust gets into a
ing, the recirculation of the cooled drying agent is intensified separator, where the milled fractions are directed to the burners
(increased). Metallic parts and the trioxide are caught in a or back for additional milling depending on their size.
specific box 10 in the down section of the corpse. The higher The boiler which milling system is studied is a Benson
the productiveness of the mill fan, the greater its diameter; this type once-though sub-critical boiler. There are four mills per
is accompanied by a deterioration of the even fuel distribution boiler. Each mill fan system has four radial bearings two
along the blades, by an intensified local wearing-out of the in the mill and two in the motor. All the data used in the
blades, also of the trioxide disc and a decrease of the milling present research are obtained from the historian system of the
effectiveness. The drying-agent temperature after the drying Distributed Control System (DCS).
mine, i.e. before the very mill fan, must not exceed 450-500o .
The percentage of 02 in the damp mixture before the mill
can reach 2,5 % and the part of 02 together with the moist III. P ROCESS C ONTROL S YSTEM
evaporation in the drying mine is reduced to less than 15-16 ®
Experion Process Knowledge System (PKS) is a cost-
%; this ensures the installation against explosions this makes effective open control and safety system that expands the
the installation explosion-proof. role of distributed control. It addresses critical manufacturing
The total pressure developed by the mill fan equals to objectives to facilitate sharing knowledge and managing work-
1-2 kPa (100-200 mm water column) and it is directed to flow. Experion provides a safe, robust, scalable, plant-wide
overcoming the resistance of the drying path which is under system with unprecedented connectivity through all levels of
pressure and also of the resistance of the separator, the dust the plant as illustrated in the following high-level view of
piping and the burners. the architecture. The Experion unified architecture combines
Coal milling systems with mill fans are widely used in DCS functionality and a plant-wide infrastructure that unifies
the fossil fired power plants, due their possibility to simul- business, process, and asset management to: Facilitate knowl-
taneously dry, mill and transport the coal to boilers furnace edge capture; Promote knowledge sharing; Optimize work
chamber. processes; Accelerate improvement and innovation.

Unauthenticated
Download Date | 1/30/17 12:31 PM
354 M. B. HADJISKI, L. A. DOUKOVSKA, S. L. KOJNOV

The Experion platform is well suited for both small and humidity of fuel, imprecise dosing of coal, stohasticity of
large systems. It provides the power and flexibility required to temperature of the intake oven gases due to non-stationarity of
handle the full spectrum of process control and safety appli- the flame position) [15]–[19]. The symptoms in a diagnostic
cations. Experion offers state-of-the-art Decentralized Control problem are always indirect and they are of stochastic nature
System (DCS) capabilities that include Abnormal Situation (changes of temperature of the dust-air mixture (DAM) at
Management (ASM), Safety Management, and Information the separator output related to the average one for the DPS,
Management technologies. Experion interfaces with Founda- the flame position (horizontal, vertical), the relative power
tion Fieldbus, Profibus, DeviceNet, HART, LON, ControlNet consummation for grinding, corrected rotating frequency for
and Interbus. Robustness, security, compliance, control, safety, the raw fuel feeders (RFF)) [16]–[20].
and reliability are plant-wide. Its distributed control features
include a complete continuous, logic, sequential, and drive IV. E XPERIMENTAL R ESULTS
object-oriented control environment hosted on fully redundant
controllers. Experion Process Knowledge System features in- Mill fans are a part of the equipment of power units that are
clude: most often repaired due to intensive erosion of the operative
wheel blades in the process of grinding of low-calorific lignite
• Sophisticated human-machine interface. coal from the Trayanovo 1 and Trayanovo 2 mines with high
• Tightly integrated databases, engineering tools, and con- percentage of dust (28%-45% of the dry mass).
trol and safety applications. In spite of the constructive measures to use steels with high
• Operational integration of control and safety applications. resistance index in critical parts of the mill fan, their interval
• Open, deterministic, high-speed control network commu- between two successive repairs equals to 2-2.5 months. The
nications system for predictable and repeatable control mill fans are of critical importance for the automation of the
linking servers, controllers, and remote I/O. power-unit (PU) supply, so their technical state is an object
• A configurable Control Execution Environment (CEE) of strict monitoring, repair and on-line duty according to the
provides deterministic, consistent, and reliable control operative dispatcher time-table. The control range of mill fan
application execution. is small, therefore the primary power control and especially the
• A single builder tool, Configuration Studio, allows inte- secondary power control is realized by stopping and starting
grated application configuration. some of the mill fan.
• Three CEE-based controllers: The C200 Process Con- The experimental research is done in the Bulgarian national
troller is a compact and cost-effective solution located Maritsa East 2 thermal power plant. The plant has four
close to the process with direct IO connections. It is ideal double blocks with direct-current boilers 175 MW each and
for integrated regulatory, fast logic, sequential, and batch four monoblocks with drum boilers 210 MW each. The fuel
control applications. The C300 Process Controller is the for both types of blocks is one and the same: low-quality
next generation controller that builds on the reliability and Bulgarian lignite coal from the Trayanovo 1 and Trayanovo
robustness of the C200 controller to provide even more 2 mines with caloricity of 1200 – 1600 kkal/kg (5000 – 6700
versatile control integration through innovative mounting kkal/kg). The basic data are obtained from steam-generator 6
and connecting techniques. The Application Control En- of block 3. The steam generator has four mill fans. There is a
vironment (ACE) is ideally suited for supervisory control decentralized control system (DCS) mounted over the steam
solutions and integration with third party control systems. generator. All used data are recorded in the Historian system
It is hosted on a server grade computer platform. of DCS. The duration of the observations is 8 months in 2010.
• Safety Instrumented Systems (SIS). Two types of data are used for the research – vibrosignals from
• The Simulation Control Environment (SCE) supports sys- mill fan and data about the basic mode parameters related to
tem simulation on computers without requiring dedicated mill fan.
controller hardware or process connections. The mill fan vibration state may become a rather useful
• Redundancy support for servers, networks, and con- component of their diagnostics to determine their affiliation
trollers. to some zone of efficiency – S1 – the normal one; S2
• Distributed System Architecture (DSA) that integrates – partial damages; still possible exploitation with lowered
multiple servers into a single operational system. mode parameters (e.g. loading) and measures for current
• Support for internationalization/localization. maintenance (lubrication, jamming bolt joints of mill fan to
• Interfaces for wide variety of third-party controllers and the bearers, technological adjustments (angles of rotation of
protocols. valves, jalousie); S3 – zone of serious damages, requiring
• A cost-effective architecture that – makes extensive use immediate stopping at the first opportunity (stop the unit). The
of open technologies and commonality of hardware, and mill fan vibration state may become a rather useful component
is scaleable from just a few points, to thousands of points. of their diagnostics to determine their affiliation to some zone
The measurements of the necessary mode quantities for of efficiency (see Tab. 1). Where:
mill fan in the DCS or SCADA system are rather inaccu- • S1 – the normal one;
rate due to the significant changeability of the conditions • S2 – partial damages; still possible exploitation with
for measurements (wear, slagging, sensor pollution) and also lowered mode parameters (e.g. loading) and measures for
due to the great amount of external disturbances (dust and current maintenance (lubrication, jamming bolt joints of

Unauthenticated
Download Date | 1/30/17 12:31 PM
NONLINEAR TREND ANALYSIS OF MILL FAN SYSTEM VIBRATIONS FOR PREDICTIVE MAINTENANCE AND DIAGNOSTICS 355

mill fan to the ground), technological tunings (angles of

rotation of valves, jalousie);
• S3 – zone of serious damages, requiring immediate
stopping at the first chance (stop the unit).
The mill fan state is multidimensional. The basic compo-
nents are grinding productiveness B [t/h], fan productiveness
W [m3 /h] and vibration state S VM B (amplitude of vibrations
A (or velocity/acceleration of vibrations)).
Exemplary limits for the workability of the Maritsa East 2
power plant monoblocks of 8 MW each are shown in Tab. 1.

TABLE I

Component B W A
S, dimention t/h m3 /h mm Fig. 3. Distribution of reflecting objects for the computer simulation.
S1 55 200 000 6
S2 48 180 000 7
S3 <40 <140 000 >8

Analogical are the limits of workability and also of the mill

fans for the double-blocks of Maritsa East 2 power plant they
are installations with 4 mill fans each.
The basic problem for diagnostics of the mill fan technical
state is the impossibility for direct measurement neither of the
components B, W, A. Difficulties with the assessment of the
mill fan vibrostate is the object of interest below.
It is a major problem that the increased values for vibrations
may not refer to the technical state of the mill fan. They may
be due to:
• Hydrodynamic instability in the sectors between the
Fig. 4. Passive SAR raw signal.
blades of the operative wheel;
• Uneven filling of the operative wheel with coal dust;
• Personnel faults;
vibration values in millimeters as a function of the mills rotor
• Worsened quality of fuel and especially the contents of
working hours.
dust of dry mass Ad .
Presented data shows working and idle states of the device
The process of changing the mechanical properties which before and after replacing the rotor. Next figure shows vibra-
are likely to induce increasing vibrations is basically due to tion rates before rotor replacement.
uneven worn-out blades of the operative wheel, but also due Due to higher nonlinearity of the device a nonlinear trend
to traditional sources of vibrations in mechanical systems [21] analysis for diagnostics and predictive maintenance is done.
• Damaged bearings; Regression analysis is done in Matlab environment. It was
• Unbalanceness; found that a polynomial of degree 7 fits the data best. Results
• Uncoaxiness; are presented in the next figure. From industrial application
• Insufficiently robust joint of the mill fan with the bearers; viewpoint, the maximal number of fault diagnostic applica-
• Lubrication problems. tions in process industries are based on process historian based
The results from the few references for research of the approaches. This is due to the fact that process history-based
nature of vibrations in the mill fan [22], [23], and also of approaches are implementing modeling an a priori knowledge.
similar systems [24], [25], our observations included, and
show that there are enough reasons to treat these vibrations It is observed that after the replacement the vibrations with
as nonlinear. new rotor have higher amplitudes than with the worn out one.
Prognostic maintenance of a mill fan is considered in the This is because of the abrasive wear out of the rotor – the
presented paper. It is based on the vibration of the nearest to blades become thinner and the rotor becomes lighter. The
the mill rotor bearing block. In the present paper the analysis new rotor is heavier so the vibrations are more intensive even
is done using data archived by the installed on the site DCS though the rotor has been carefully balanced.
Honeywell Experion PKS R301. The observation period is After preliminary data processing including idle periods
16.12.2010 – 16.01.2011. On 31.12.2010 the rotor of the mill removal in Fig. 6 are shown the vibration rates after rotor re-
fan is changed. After the replacement it has been working for placement. It is plain to see the vibration rate is higher. Taking
378 hours. The period chosen allows for vibrations analyzes into account the observed trend before rotor replacement it is
before and after the replacement. In Fig. 3 are shown the possible to prognosticate the next planned device maintenance

Unauthenticated
Download Date | 1/30/17 12:31 PM
356 M. B. HADJISKI, L. A. DOUKOVSKA, S. L. KOJNOV

• b
R EFERENCES
[1] “Condition monitoring and diagnostics of machines-prognostics - part:
General guidelines,” ISO Standard 13381-I, Tech. Rep., 2004.
[2] A. Jardine, D. Link, and D. Banjevic, “A review on machinery diag-
nostics and prognostics, implementing condition-based maintenance,”
Mechanical Systems and Signal Processing, vol. 20, no. 7, pp. 1483–
1510, 2006.
[3] O. Dragomir, R. Gouriveau, F. Dragomir, E. Miuca, and N. Zerhoni,
“Review of prognostic problem in condition-based maintenance,” Proc.
of European Control Conference, pp. 25–30, 2009.
[4] O. Dragomir, R. Gouriveau, and N. Zerhouni, “Adaptive neuro-fuzzy
inference system for midterm prognostic error stabilization,” Inter. Jour.
of Comp. Comm. and Control, vol. 3, pp. 271–276, 2008.
[5] G. Vachtsevanos, F. Lewis, M. Roemer, A. Hess, and B. Wu, Intelligent
fault ll-ldynosis and prognosis for engineering systems. Wiley & Sons,
2006.
[6] Y. L. Dong, Y. J. Gu, K. Yang, and W. K. Zhang, “A combining
condition prediction model and its application in power plant,” Proc. of
Fig. 5. Computer simulation results for a) TDC and b) RDA imaging
International Conference on Machine Learning and Cybernetics, vol. 6,
algorithms.
pp. 3474–3478, 2004.
[7] R. Zemouri, “Recurrent radial basis function network for time-series
prediction,” Engin. Appl. of Artificial Intelligence, vol. 16, pp. 453–463,
2003.
[8] P. Wang and G. Vachtsevanos, “Fault prognostics using dynamic wavelet
neural networks,” Artificial Intelligence for Engineering Design Analysis
and Manufacturing, vol. 15, pp. 349–265, 2001.
[9] W. Q. Wang, M. F. Golnaraghi, and F. Ismail, “Prognosis of machine
health condition using neuro-fuzzy systems,” Mechanical Systems and
Signal Processing, vol. 18, pp. 813–831, 2004.
[10] S. Zhang and R. Ganesan, “Multivariable trend analysis using neural
networks for intelligent diagnostics of rotating machinery,” Journal of
Engineering for Gas Turbines and Power, vol. 119, pp. 378–384, 1997.
[11] J. L. Kolonder, Case-based reasoning. Morgan Kaufmann, 1993.
[12] M. Hadjiski and V. Boishina, “Enhancing functionality of complex plant
hybrid control system using case-based reasoning,” Proc. of 5th IEEE
International Conference on Intelligent Systems, pp. 25–30, 2010.
[13] R. Heoshmand and M. Banejad, “Application of fuzzy logic in fault diag-
nosis in transformers using dissolved gas based on different standards,”
Word Academy of Sciences, Engineering and Technology, vol. 17, no. 2,
Fig. 6. Experiment organizational scheme. pp. 157–161, 2006.
[14] http://geyz.ru/.
[15] B. Bonev, T. Totev, J. Artakov, and M. Nikolov, “Diagnostic of coal dust
prepraration systems with milling fans,” in New Trends in Automation
shutdown. of Energetic Processes98.
[16] T. Totev, B. Bonev, and J. Artakov, “Systems for determination of the
quality of coal, burning in the steamgenerator p-62 in tpp,” Maritza East
V. C ONCLUSIONS Energetics, no. 1-2, 1995.
This paper proposes a nonlinear trend analysis for diagnos- [17] M. Hadjiski, M. Nikolov, S. Dukovski, G. Drianovski, and E. Tamnishki,
“Low-rank coal fired boilers monitoring by applying hybrid models,”
tics and predictive maintenance. Vibration rates of a mill fan in 8th IEEE Mediterranean Conference on Control and Automation
system for Maritza East 2 power plant have been analyzed. MED2000, 2000.
The obtained results show that nonlinear trend analysis may [18] M. Hadjiski and V. Totev, “Hybrid modeling of milling fan of steam-
generators in tpp,” Automatics and Informatics, no. 4, 2000.
be successfully applied to estimate and predict changes of the [19] M. Hadjiski, V. Totev, and R. Yusupov, “Softsensing-based flame po-
mill rotor state. Because of the chosen device features it is sition estimation in steam boiler combustion chamber,” in Distributed
important to mention the necessity of additional information Computer and Communication Networks, 2005.
[20] M. Hadjiski, V. Petkov, and E. Mihailov, “A software environment
regarding the exploitation rates. In the considered example this for approximate model design of a low-calorithic coal combustion in
is expressed in knowledge about the planned maintenance and power plant boilers,” Modelling and Optimization of Pollutant Reduced
rotor replacement shutdowns. To increase the usability of the Industrial Furnaces, 2000.
[21] R. Eisenman, Machinery Malfunction Diagnosis and Correction. Di-
proposed diagnostic procedures it is also necessary to take into agnostyka3 (39), 2006.
account the efficiency rates of the process, which would allow [22] M. Tabaszewski, Forecasting of Residual Time of Milling Fans by Means
for a better picture about the necessity of effective and in time of Neural Networks. Prentice-Hall, 1998.
[23] C. Cempel and M. Tabaszewski, Singular Spectrum Analysis as a
prognostic maintenance of the thermal power plant devices. Smoothing Method of Load Variability. Diagnostyka4 (56), 2010.
Vibrosignals may be successfully used as a substantial [24] N. Zhang, Dynamic Characteristic of Fow Induced Vibration in a Rotor-
additional symptom for isolation and diagnostics of mill fan Seal System. IOS Press, 2010.
[25] Z. Zong and Z. Ma, Nonlinear Vibration Analysis of a Eccentric Rotor
system which is not done at present. The assessment of with Unbalanced Magnetic Pull. I25-th IAHR Symp, 2010.
the mill fan vibration state is a complex problem due to
the exceptionally big uncertainty in the measurements which
follows from the temporally re-covered changes of multimode
factors. The mill fan vibration state S VM B is a valuable integral
indicator for its efficiency.

Unauthenticated
Download Date | 1/30/17 12:31 PM

View publication stats