SlideShare a Scribd company logo
1 of 3
Practice Article
Technical note: Advantages of the self-organizing controller
for high-pressure sterilization equipments
V. Pilipovik a
, C. Riverol b,n
a
AIChemEng Research and Development Group, J.C Engineers & Partners, Av. Andres Bello, Edif. Las Rozas, Urb. La Florida, Caracas, Venezuela
b
Chemical Engineering Department, University of the West Indies, St. Augustine Campus, Trinidad, Trinidad and Tobago
a r t i c l e i n f o
Article history:
Received 9 February 2012
Received in revised form
3 July 2013
Accepted 7 July 2013
Available online 29 August 2013
This paper was recommended
for publication by Dr. A.B. Rad
Keywords:
Spores
SOM
Sterilization
Adjustment mechanism
Controller
a b s t r a c t
A study of a self-organizing controller is implemented in a way that response to controlled system
follows the desired given by the model. The self-organizing controller has proven to be a valuable tool in
sterilization equipment in order to verify the capacity of the response to any change in the pressure or
temperature. Basically, this type of controller is based on the Self-Organizing Map (SOM) that is a neural
network algorithm of unsupervised learning. The new ideas include clustering visualization, interactive
training and one-dimension arrays.
& 2013 ISA. Published by Elsevier Ltd. All rights reserved.
1. Introduction
High-pressure processing has emerged as an attractive processing
technology for preservation of foods. Since the introduction of high
pressure processing in 1990 [1,2] the range of products has gradually
expanded. The effect that this type of processing has on microorgan-
isms is similar to pasteurization while the product retains its fresh or
just-prepared appearance and nutritional quality [1,2]. High pressure
processing causes minimal changes in the ‘fresh’ characteristics of
foods thereby eliminating the processing by thermal degradation.
Compared to thermal processing, high pressure sterilization (HPS)
results in foods with fresher taste, and better appearance, texture and
nutrition. The technology is especially beneficial for heat sensitive
products.
A recent innovation in sterilization of food products using high
pressure is the complete inactivation of vegetative micro-organ-
isms, as well as spores, resulting in ambient stable products
shown in [2,3]. High pressure sterilization is a combined process
where both pressure and temperature contribute to sterilization.
A good control system is necessary for keeping the temperature
and pressure in the adequate values with less adverse effects
on product quality. In this paper a self-organizing controller is
introduced as a hierarchical structure in which the inner loop is
a table-based controller and the outer loop is the adjustment
mechanism. The Self-Organizing Map (SOM) is especially suitable
for data analysis [4] where the application in process control is
focused on training neural networks [4–6]. It creates a set of
prototype vectors representing the data set and carries out a
topology preserving projection of the prototypes from the
d-dimensional input space onto a low-dimensional grid. This
ordered grid can be used as a convenient visualization surface
for showing different features of the SOM. In its basic form, it
produces a similarity graph of input data. It converts the nonlinear
relationships into simple geometric relationships [7,8].
2. Laboratory set-up and results
Destruction of Clostridium sporogenes spores by high pressure
treatment was used as the study case. In the laboratory, a
sterilizing stainless steel tank of 4 L capacity and 1.5 m length
was used, see Fig. 1. The tank has 3 PT100 for monitoring the
temperature at the center of the tank. One of the PT100 is located
at the top of the tank, another at the middle and the last one at the
bottom. The accuracy of the sensor is 70.1 1C. The controller
received one value (the average of the three temperatures)
because the values from the top to the bottom do not differ more
than 1 1C.
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/isatrans
ISA Transactions
0019-0578/$ - see front matter & 2013 ISA. Published by Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.isatra.2013.07.007
n
Corresponding author. Tel.: +1 8686622002.
E-mail address: Carmen.Riverol@sta.uwi.edu (C. Riverol).
ISA Transactions 53 (2014) 186–188
The first step is to justify the SOM ability analytically. The first
approach was to restrict the consideration to one-dimensional
linear unit with which each scalar valued input signal is connected.
The SOM control system (SOMCS) adapts the system in accordance
with the desired response [7–9]. Each sample is recording and the
deviation of desired state is evaluated. An example of the best
match is shown in Fig. 2. According to the kinetics of the spore, the
adiabatic temperature can increase from 2.2 to 9.3 1C/100 MPa
approximately; thus the equipment was tested using three pres-
sure pulses of 100, 300 and 600 MPa for 1, 3 and 5 min each, as
reported in Table 1. Nevertheless, the importance of the minimum
product temperature for the degree of spore inactivation confirms
that the temperature should be monitored during high pressure
sterilization.
On the other hand, the SOM neural network was designed
using cluster membership for test data sets that include various
levels of data dispersion combined with outliners.
A clustering Q means partitioning a data set into a set of
clusters Q [10–12]. In crisp clustering, each data sample belongs to
exactly one cluster. Fuzzy clustering is a generation of crisp where
each sample has different degrees of membership. However, this
leaves much room for variation within and between cluster
distances and can be defined as follows:
dsl ¼ minjjxiÀxjjj ð1Þ
dcl ¼ maxjjxiÀxjjj ð2Þ
dml ¼
∑i;jjjxiÀxjjj
NiNj
ð3Þ
outliner ¼ jjciÀcjjj ð4Þ
where N is the total value inside the cluster i or j, dsl is the single
linkage, dcl is the complete linkage, dml is the average and outliner
is the distance between centers shown in [9] and [10].
An example of the cluster definition results is shown in Fig. 2
where the deviation is indicated (d). The idea behind the adapta-
tion is to let the adjustment mechanism update the values in the
control table F on the basis of the current performance of the
controller as reported in Fig. 3. The adjustable neural network
taught in previous simulation is inserted in the SOC using the
model M. This model was empirically calculated. The adjustable
neural network is tuned with each experimental result in the plant
(P). Adiabatic heating is the uniform temperature rise within the
product, which is solely caused by pressurization [7]. The control
Fig. 1. The equipment used in the laboratory.
Fig. 2. Cluster definition results.
Table 1
Experimental set-up and temperature control (SOC).
Pressure
(MPa)
Target
temp (1C)
Initial
temperature
(1C)
Mean
temperature
achieved (1C) SOM
Mean
temperature
achieved (1C) PI
100 70 40 70.8 71.2
85 55 85.6 85.9
100 60 101 99.6
300 70 41 70.1 71.2
85 58 84.9 84.6
100 62 100.6 101.3
600 70 51 71.2 71.6
85 56 85.3 85.9
100 66 100.3 100.7
V. Pilipovik, C. Riverol / ISA Transactions 53 (2014) 186–188 187
system was implemented using MATLAB 7.0, real time toolbox and
MF624 multifunction I/O card. A SOM was trained using the
sequential training algorithm [8]. All maps were linear. In the
agglomerative clustering (see Fig. 2), single, average and complete
linkages were used in the construction phase. For additional
information about SOM, readers are referred to [5]. In this article
the training time was 32 h using 10,000 values collected from
2005 to 2010. The input values were obtained during the last
5 years because this equipment worked 35 years without any
control system.
How to tune the gains does deserve attention, since they may
finally stand in the way for a successful implementation. The SOC
works ‘surprisingly well’ [13]. For example, the output gain is
lowered between two training sessions. The adjustment is com-
pensated by an F-table with numbers of larger magnitude [13–15].
Therefore it is possible to start with a linear F-table, and set the
gains loosely according to a PID tuning rule or hand-tuning. This is
a good starting point for self-organization.
The SOM network can improve the quality of decisions in
cluster analysis especially in non-uniform cluster densities
[5,6,13]; for example, temperature is a parameter that changes
very fast; thus some of empirical data can become “messy data”.
Fig. 4 depicts an example in which the SOM controller offers a
better performance than the PI controller. Basically, the overshoot
is reduced and the setting conditions are attained early. The
highest temperature with the SOMCS is about 19.5 1C lower than
that with PI as reported in Fig. 4. The same behavior was observed
24 times during the testing sections. The reduction of the
overshoot can be translated in a reduction of the steam consump-
tion; thus some hundreds of dollars can be saved. Also an over-
shoot over 10 1C is not desirable because the food can become dry
and lose its texture because of which designing of new controllers
may continue in this area. Table 1 indicates that the SOMCS offers
a better performance over PI in this system and shows that the
results are not sensitive to the initial learning coefficients at
different initial conditions. In summary, the new control system
shows a good performance and can also improve the traditional PI
controller; nevertheless, future research should focus on reducing
the training time for the SOM because its implementation on the
equipment consumes a long time.
3. Conclusion
This paper presented the development of a SOM controller and
its application to a high pressure sterilization equipment. The
algorithm presented a satisfactory performance and efficiency. The
simulation showed that it is possible to achieve a good behavior
after few steps of learning using a simple model. The implementa-
tion in the sterilization equipment improves the response capacity
of the control system although the implementation of the self-
organizing system consumes time and is complex. The adjustable
control table (F) can be increased using more data for improving
the accuracy; however a good computer should be used for
reducing the time consumption (the computer used in this article
has a 2GB RAM only).
References
[1] Matser A, Krebbers B, van den Berg R, Bartels P. Advantages of high pressure
sterilization on quality of food products. Trends in Food Science & Technology
2004;15(2):79–85.
[2] Meyer R, Cooper K, Knorr D, Lelieveld H. High pressure sterilization of foods.
Food Technology 2000;54(11):67–72.
[3] Hoogland H, De Heij W, Van Schepdael L. High pressure sterilization: novel
technology, new products, new opportunities. New Food 2001;3:21–6.
[4] Jantzen J. Foundations of fuzzy controller. 1st ed. NY, USA: Wiley; 117–39.
[5] Bloch G, Denoeux T. Neural networks for process control and optimization:
two industrial applications. ISA Transactions 2003;42(1):39–51.
[6] Das S, Saha S, Das S, Gupta A. On the selection of tuning methodology of FOPID
controllers for the control of higher order processes. ISA Transactions 2011;50
(3):376–88.
[7] Dovžan D, Škrjanc I. Recursive fuzzy c-means clustering for recursive fuzzy
identification of time-varying processes. ISA Transactions 2011;50(2):159–69.
[8] Lokriti A, Salhi I, Doubabi S, Zidani Y. Induction motor speed drive improve-
ment using fuzzy IP-self-tuning controller. A real time implementation. ISA
Transactions 2013;52(3):406–17.
[9] Tisan A, Cirstea M. SOM neural network design—a new Simulink library based
approach targeting FPGA implementation. Mathematics and Computers in
Simulation 2013;91:134–49.
[10] Rizal M, Ghani Jaharah A, Nuawi M, Hassan C, Haron C. Online tool wear
prediction system in the turning process using an adaptive neuro-fuzzy
inference system. Applied Soft Computing 2013;13(4):1960–8.
[11] Yamazaki T Mamdani TE. On the performance of a rule-based self-organizing
controller. In: Proceedings of the IEEE conference on applications of adaptive
and multivariable control. 19–21 July, Hull; 1982. p. 121–32.
[12] Vesanto J, Himberg J, Alhoniemi E, Parhankangas J.. Self-organizing map in
Matlab: the SOM Toolbox. In: Proceedings of the Matlab DSP conference.
Espoo, Finland; 1999. p. 35–40.
[13] Nauck D, Klawonn F, Kruse R. Foundations of neuro-fuzzy systems. NY, USA:
John Wiley & Sons; 37–56.
[14] Kohonen T. Self organizing maps. 3rd ed. NY, USA: Springer; 71–99.
[15] Ghaseminezhad M, Karami A. A novel self-organizing map (SOM) neural
networks discrete gropus of data clustering. Applied Soft Computing 2011;11
(4):3771–8.
Fig. 3. SOC control system. The solid line indicates the field data and the dash lines
indicate the calculated values.
0 5 10 15 20 25 30 35 40 45 50
0
20
40
60
80
100
120
Time (min)
Temperature(oC)
PI
SOM-NN
Fig. 4. Behavior of the temperature using different controllers. P¼600 MPa.
V. Pilipovik, C. Riverol / ISA Transactions 53 (2014) 186–188188

More Related Content

What's hot

Download-manuals-water quality-technicalpapers-inter-labfindingsaqc1stround
 Download-manuals-water quality-technicalpapers-inter-labfindingsaqc1stround Download-manuals-water quality-technicalpapers-inter-labfindingsaqc1stround
Download-manuals-water quality-technicalpapers-inter-labfindingsaqc1stround
hydrologywebsite1
 
Bulletin : Thermal Analysis (TGA) using Perkin Elmer & Mettler Toledo
Bulletin : Thermal Analysis (TGA) using Perkin Elmer & Mettler ToledoBulletin : Thermal Analysis (TGA) using Perkin Elmer & Mettler Toledo
Bulletin : Thermal Analysis (TGA) using Perkin Elmer & Mettler Toledo
Noor Fatihah Suhaimi
 
Report_EBACProject_BBP
Report_EBACProject_BBPReport_EBACProject_BBP
Report_EBACProject_BBP
Arun Sankar
 

What's hot (19)

30120140507008
3012014050700830120140507008
30120140507008
 
2013 sk~1
2013 sk~12013 sk~1
2013 sk~1
 
30120130405019 2
30120130405019 230120130405019 2
30120130405019 2
 
Dsc application 2
Dsc application 2Dsc application 2
Dsc application 2
 
Anilkumar2007
Anilkumar2007Anilkumar2007
Anilkumar2007
 
Thermal method of analysis
Thermal method of analysis Thermal method of analysis
Thermal method of analysis
 
EXPERIMENTAL INVESTIGATION ON BOILING HEAT TRANSFER USING R134A
EXPERIMENTAL INVESTIGATION ON BOILING HEAT TRANSFER USING R134AEXPERIMENTAL INVESTIGATION ON BOILING HEAT TRANSFER USING R134A
EXPERIMENTAL INVESTIGATION ON BOILING HEAT TRANSFER USING R134A
 
Calibration of A Five-Hole Probe in Null and Non-Null Technique
Calibration of A Five-Hole Probe in Null and Non-Null Technique Calibration of A Five-Hole Probe in Null and Non-Null Technique
Calibration of A Five-Hole Probe in Null and Non-Null Technique
 
W04405127131
W04405127131W04405127131
W04405127131
 
13 c analyses of calcium carbonate comparison between gb and ea
13 c analyses of calcium carbonate comparison between gb and ea13 c analyses of calcium carbonate comparison between gb and ea
13 c analyses of calcium carbonate comparison between gb and ea
 
Ao044265268
Ao044265268Ao044265268
Ao044265268
 
E012252736
E012252736E012252736
E012252736
 
Bi4301333340
Bi4301333340Bi4301333340
Bi4301333340
 
Download-manuals-water quality-technicalpapers-inter-labfindingsaqc1stround
 Download-manuals-water quality-technicalpapers-inter-labfindingsaqc1stround Download-manuals-water quality-technicalpapers-inter-labfindingsaqc1stround
Download-manuals-water quality-technicalpapers-inter-labfindingsaqc1stround
 
Bulletin : Thermal Analysis (TGA) using Perkin Elmer & Mettler Toledo
Bulletin : Thermal Analysis (TGA) using Perkin Elmer & Mettler ToledoBulletin : Thermal Analysis (TGA) using Perkin Elmer & Mettler Toledo
Bulletin : Thermal Analysis (TGA) using Perkin Elmer & Mettler Toledo
 
Thermogravimetric Analysis (TGA)
Thermogravimetric Analysis (TGA)Thermogravimetric Analysis (TGA)
Thermogravimetric Analysis (TGA)
 
Umbral Anaeróicbo
Umbral AnaeróicboUmbral Anaeróicbo
Umbral Anaeróicbo
 
OPTIMIZATION OF CONVECTIVE HEAT TRANSFER MODEL OF COLD STORAGE USING TAGUCHI ...
OPTIMIZATION OF CONVECTIVE HEAT TRANSFER MODEL OF COLD STORAGE USING TAGUCHI ...OPTIMIZATION OF CONVECTIVE HEAT TRANSFER MODEL OF COLD STORAGE USING TAGUCHI ...
OPTIMIZATION OF CONVECTIVE HEAT TRANSFER MODEL OF COLD STORAGE USING TAGUCHI ...
 
Report_EBACProject_BBP
Report_EBACProject_BBPReport_EBACProject_BBP
Report_EBACProject_BBP
 

Viewers also liked

Digital redesign of analog Smith predictor for systems with input time delays
Digital redesign of analog Smith predictor for systems with input time delaysDigital redesign of analog Smith predictor for systems with input time delays
Digital redesign of analog Smith predictor for systems with input time delays
ISA Interchange
 
Guidelines for the tuning and the evaluation of decentralized and decoupling ...
Guidelines for the tuning and the evaluation of decentralized and decoupling ...Guidelines for the tuning and the evaluation of decentralized and decoupling ...
Guidelines for the tuning and the evaluation of decentralized and decoupling ...
ISA Interchange
 
Self tuning adaptive control for an industrial weigh belt feeder
Self tuning adaptive control for an industrial weigh belt feederSelf tuning adaptive control for an industrial weigh belt feeder
Self tuning adaptive control for an industrial weigh belt feeder
ISA Interchange
 
101 Tips for a Successful Automation Career Appendix F
101 Tips for a Successful Automation Career Appendix F101 Tips for a Successful Automation Career Appendix F
101 Tips for a Successful Automation Career Appendix F
ISA Interchange
 
Adaptive backstepping sliding mode control with fuzzy monitoring strategy for...
Adaptive backstepping sliding mode control with fuzzy monitoring strategy for...Adaptive backstepping sliding mode control with fuzzy monitoring strategy for...
Adaptive backstepping sliding mode control with fuzzy monitoring strategy for...
ISA Interchange
 
Design PID controllers for desired time domain or frequency domain response
Design PID controllers for desired time domain or frequency domain responseDesign PID controllers for desired time domain or frequency domain response
Design PID controllers for desired time domain or frequency domain response
ISA Interchange
 
Improving performance using cascade control and a Smith predictor
Improving performance using cascade control and a Smith predictorImproving performance using cascade control and a Smith predictor
Improving performance using cascade control and a Smith predictor
ISA Interchange
 
Integrating Analyzers with Automation Systems: Oil and Gas by David Schihabel
Integrating Analyzers with Automation Systems: Oil and Gas by David SchihabelIntegrating Analyzers with Automation Systems: Oil and Gas by David Schihabel
Integrating Analyzers with Automation Systems: Oil and Gas by David Schihabel
ISA Interchange
 
Gas Detectors & Detectability
Gas Detectors & DetectabilityGas Detectors & Detectability
Gas Detectors & Detectability
ISA Interchange
 
HMI Design: The Good, the Bad, and the Ugly
HMI Design: The Good, the Bad, and the UglyHMI Design: The Good, the Bad, and the Ugly
HMI Design: The Good, the Bad, and the Ugly
ISA Interchange
 

Viewers also liked (19)

Influence of time and length size feature selections for human activity seque...
Influence of time and length size feature selections for human activity seque...Influence of time and length size feature selections for human activity seque...
Influence of time and length size feature selections for human activity seque...
 
Global adaptive output feedback control for a class of nonlinear time delay s...
Global adaptive output feedback control for a class of nonlinear time delay s...Global adaptive output feedback control for a class of nonlinear time delay s...
Global adaptive output feedback control for a class of nonlinear time delay s...
 
Robust fault detection for switched positive linear systems with time varying...
Robust fault detection for switched positive linear systems with time varying...Robust fault detection for switched positive linear systems with time varying...
Robust fault detection for switched positive linear systems with time varying...
 
Optimized sensor selection for control and fault tolerance of electromagnetic...
Optimized sensor selection for control and fault tolerance of electromagnetic...Optimized sensor selection for control and fault tolerance of electromagnetic...
Optimized sensor selection for control and fault tolerance of electromagnetic...
 
Calibration workbook
Calibration workbookCalibration workbook
Calibration workbook
 
Digital redesign of analog Smith predictor for systems with input time delays
Digital redesign of analog Smith predictor for systems with input time delaysDigital redesign of analog Smith predictor for systems with input time delays
Digital redesign of analog Smith predictor for systems with input time delays
 
Guidelines for the tuning and the evaluation of decentralized and decoupling ...
Guidelines for the tuning and the evaluation of decentralized and decoupling ...Guidelines for the tuning and the evaluation of decentralized and decoupling ...
Guidelines for the tuning and the evaluation of decentralized and decoupling ...
 
Self tuning adaptive control for an industrial weigh belt feeder
Self tuning adaptive control for an industrial weigh belt feederSelf tuning adaptive control for an industrial weigh belt feeder
Self tuning adaptive control for an industrial weigh belt feeder
 
Finite time control for nonlinear spacecraft attitude based on terminal slidi...
Finite time control for nonlinear spacecraft attitude based on terminal slidi...Finite time control for nonlinear spacecraft attitude based on terminal slidi...
Finite time control for nonlinear spacecraft attitude based on terminal slidi...
 
101 Tips for a Successful Automation Career Appendix F
101 Tips for a Successful Automation Career Appendix F101 Tips for a Successful Automation Career Appendix F
101 Tips for a Successful Automation Career Appendix F
 
Adaptive backstepping sliding mode control with fuzzy monitoring strategy for...
Adaptive backstepping sliding mode control with fuzzy monitoring strategy for...Adaptive backstepping sliding mode control with fuzzy monitoring strategy for...
Adaptive backstepping sliding mode control with fuzzy monitoring strategy for...
 
Design PID controllers for desired time domain or frequency domain response
Design PID controllers for desired time domain or frequency domain responseDesign PID controllers for desired time domain or frequency domain response
Design PID controllers for desired time domain or frequency domain response
 
Real time simulation of nonlinear generalized predictive control for wind ene...
Real time simulation of nonlinear generalized predictive control for wind ene...Real time simulation of nonlinear generalized predictive control for wind ene...
Real time simulation of nonlinear generalized predictive control for wind ene...
 
Improving performance using cascade control and a Smith predictor
Improving performance using cascade control and a Smith predictorImproving performance using cascade control and a Smith predictor
Improving performance using cascade control and a Smith predictor
 
Real Time Optimization of Air Separation Plants
Real Time Optimization of Air Separation PlantsReal Time Optimization of Air Separation Plants
Real Time Optimization of Air Separation Plants
 
Integrating Analyzers with Automation Systems: Oil and Gas by David Schihabel
Integrating Analyzers with Automation Systems: Oil and Gas by David SchihabelIntegrating Analyzers with Automation Systems: Oil and Gas by David Schihabel
Integrating Analyzers with Automation Systems: Oil and Gas by David Schihabel
 
Gas Detectors & Detectability
Gas Detectors & DetectabilityGas Detectors & Detectability
Gas Detectors & Detectability
 
ENERGY MODELING OF THE PYROPROCESSING OF CLINKER IN A ROTARY CEMENT KILN
ENERGY MODELING OF THE PYROPROCESSING OF CLINKER IN A ROTARY CEMENT KILNENERGY MODELING OF THE PYROPROCESSING OF CLINKER IN A ROTARY CEMENT KILN
ENERGY MODELING OF THE PYROPROCESSING OF CLINKER IN A ROTARY CEMENT KILN
 
HMI Design: The Good, the Bad, and the Ugly
HMI Design: The Good, the Bad, and the UglyHMI Design: The Good, the Bad, and the Ugly
HMI Design: The Good, the Bad, and the Ugly
 

Similar to Advantages of the self organizing controller for high-pressure sterilization equipments

Ijmer 46067782
Ijmer 46067782Ijmer 46067782
Ijmer 46067782
IJMER
 
CFD Simulation of Air Conditioning System of the Classroom
CFD Simulation of Air Conditioning System of the ClassroomCFD Simulation of Air Conditioning System of the Classroom
CFD Simulation of Air Conditioning System of the Classroom
ijtsrd
 

Similar to Advantages of the self organizing controller for high-pressure sterilization equipments (20)

Optimization of Convective Heat Transfer Model of Cold Storage with Cylindric...
Optimization of Convective Heat Transfer Model of Cold Storage with Cylindric...Optimization of Convective Heat Transfer Model of Cold Storage with Cylindric...
Optimization of Convective Heat Transfer Model of Cold Storage with Cylindric...
 
V01 i030602
V01 i030602V01 i030602
V01 i030602
 
Theoretical heat conduction model development of a Cold storage using Taguch...
Theoretical heat conduction model development of a Cold storage  using Taguch...Theoretical heat conduction model development of a Cold storage  using Taguch...
Theoretical heat conduction model development of a Cold storage using Taguch...
 
Ijmer 46067782
Ijmer 46067782Ijmer 46067782
Ijmer 46067782
 
Climatology Applied To Architecture: An Experimental Investigation about Inte...
Climatology Applied To Architecture: An Experimental Investigation about Inte...Climatology Applied To Architecture: An Experimental Investigation about Inte...
Climatology Applied To Architecture: An Experimental Investigation about Inte...
 
Laboratory Proposal for Studies on Poultry Environment
Laboratory Proposal for Studies on Poultry EnvironmentLaboratory Proposal for Studies on Poultry Environment
Laboratory Proposal for Studies on Poultry Environment
 
CFD Simulation of Air Conditioning System of the Classroom
CFD Simulation of Air Conditioning System of the ClassroomCFD Simulation of Air Conditioning System of the Classroom
CFD Simulation of Air Conditioning System of the Classroom
 
THERMAL KINETICS OF THIN LAYER DRYING OF INDIAN GOOGEBERRY OR ANOLA FLAKS (PH...
THERMAL KINETICS OF THIN LAYER DRYING OF INDIAN GOOGEBERRY OR ANOLA FLAKS (PH...THERMAL KINETICS OF THIN LAYER DRYING OF INDIAN GOOGEBERRY OR ANOLA FLAKS (PH...
THERMAL KINETICS OF THIN LAYER DRYING OF INDIAN GOOGEBERRY OR ANOLA FLAKS (PH...
 
Constrained discrete model predictive control of a greenhouse system temperature
Constrained discrete model predictive control of a greenhouse system temperatureConstrained discrete model predictive control of a greenhouse system temperature
Constrained discrete model predictive control of a greenhouse system temperature
 
Dynamic indoor thermal comfort model identification based on neural computing...
Dynamic indoor thermal comfort model identification based on neural computing...Dynamic indoor thermal comfort model identification based on neural computing...
Dynamic indoor thermal comfort model identification based on neural computing...
 
30120140507008
3012014050700830120140507008
30120140507008
 
MODEL BASED ANALYSIS OF TEMPERATURE PROCESS UNDER VARIOUS CONTROL STRATEGIES ...
MODEL BASED ANALYSIS OF TEMPERATURE PROCESS UNDER VARIOUS CONTROL STRATEGIES ...MODEL BASED ANALYSIS OF TEMPERATURE PROCESS UNDER VARIOUS CONTROL STRATEGIES ...
MODEL BASED ANALYSIS OF TEMPERATURE PROCESS UNDER VARIOUS CONTROL STRATEGIES ...
 
C04821220
C04821220C04821220
C04821220
 
MEP Design Project
MEP Design ProjectMEP Design Project
MEP Design Project
 
Analytical approach of thermosyphon solar domestic hot
Analytical approach of thermosyphon solar domestic hotAnalytical approach of thermosyphon solar domestic hot
Analytical approach of thermosyphon solar domestic hot
 
Development of a State-Space Thermal Model for High Precision Temperature Con...
Development of a State-Space Thermal Model for High Precision Temperature Con...Development of a State-Space Thermal Model for High Precision Temperature Con...
Development of a State-Space Thermal Model for High Precision Temperature Con...
 
Modeling of heat and moisture transfer in building using rlf method
Modeling of heat and moisture transfer in building using rlf methodModeling of heat and moisture transfer in building using rlf method
Modeling of heat and moisture transfer in building using rlf method
 
1 s2.0-s0309170811002351-main
1 s2.0-s0309170811002351-main1 s2.0-s0309170811002351-main
1 s2.0-s0309170811002351-main
 
Design of Controllers for Continuous Stirred Tank Reactor
Design of Controllers for Continuous Stirred Tank ReactorDesign of Controllers for Continuous Stirred Tank Reactor
Design of Controllers for Continuous Stirred Tank Reactor
 
IRJET- Thermal Analysis and Management for an Autonomous Underwater Vehicle
IRJET- Thermal Analysis and Management for an Autonomous Underwater VehicleIRJET- Thermal Analysis and Management for an Autonomous Underwater Vehicle
IRJET- Thermal Analysis and Management for an Autonomous Underwater Vehicle
 

More from ISA Interchange

Fractional order PID for tracking control of a parallel robotic manipulator t...
Fractional order PID for tracking control of a parallel robotic manipulator t...Fractional order PID for tracking control of a parallel robotic manipulator t...
Fractional order PID for tracking control of a parallel robotic manipulator t...
ISA Interchange
 
Model based PI power system stabilizer design for damping low frequency oscil...
Model based PI power system stabilizer design for damping low frequency oscil...Model based PI power system stabilizer design for damping low frequency oscil...
Model based PI power system stabilizer design for damping low frequency oscil...
ISA Interchange
 
Fault detection of feed water treatment process using PCA-WD with parameter o...
Fault detection of feed water treatment process using PCA-WD with parameter o...Fault detection of feed water treatment process using PCA-WD with parameter o...
Fault detection of feed water treatment process using PCA-WD with parameter o...
ISA Interchange
 
Effects of Wireless Packet Loss in Industrial Process Control Systems
Effects of Wireless Packet Loss in Industrial Process Control SystemsEffects of Wireless Packet Loss in Industrial Process Control Systems
Effects of Wireless Packet Loss in Industrial Process Control Systems
ISA Interchange
 
An adaptive PID like controller using mix locally recurrent neural network fo...
An adaptive PID like controller using mix locally recurrent neural network fo...An adaptive PID like controller using mix locally recurrent neural network fo...
An adaptive PID like controller using mix locally recurrent neural network fo...
ISA Interchange
 

More from ISA Interchange (20)

An optimal general type-2 fuzzy controller for Urban Traffic Network
An optimal general type-2 fuzzy controller for Urban Traffic NetworkAn optimal general type-2 fuzzy controller for Urban Traffic Network
An optimal general type-2 fuzzy controller for Urban Traffic Network
 
Embedded intelligent adaptive PI controller for an electromechanical system
Embedded intelligent adaptive PI controller for an electromechanical  systemEmbedded intelligent adaptive PI controller for an electromechanical  system
Embedded intelligent adaptive PI controller for an electromechanical system
 
State of charge estimation of lithium-ion batteries using fractional order sl...
State of charge estimation of lithium-ion batteries using fractional order sl...State of charge estimation of lithium-ion batteries using fractional order sl...
State of charge estimation of lithium-ion batteries using fractional order sl...
 
Fractional order PID for tracking control of a parallel robotic manipulator t...
Fractional order PID for tracking control of a parallel robotic manipulator t...Fractional order PID for tracking control of a parallel robotic manipulator t...
Fractional order PID for tracking control of a parallel robotic manipulator t...
 
Fuzzy logic for plant-wide control of biological wastewater treatment process...
Fuzzy logic for plant-wide control of biological wastewater treatment process...Fuzzy logic for plant-wide control of biological wastewater treatment process...
Fuzzy logic for plant-wide control of biological wastewater treatment process...
 
Design and implementation of a control structure for quality products in a cr...
Design and implementation of a control structure for quality products in a cr...Design and implementation of a control structure for quality products in a cr...
Design and implementation of a control structure for quality products in a cr...
 
Model based PI power system stabilizer design for damping low frequency oscil...
Model based PI power system stabilizer design for damping low frequency oscil...Model based PI power system stabilizer design for damping low frequency oscil...
Model based PI power system stabilizer design for damping low frequency oscil...
 
A comparison of a novel robust decentralized control strategy and MPC for ind...
A comparison of a novel robust decentralized control strategy and MPC for ind...A comparison of a novel robust decentralized control strategy and MPC for ind...
A comparison of a novel robust decentralized control strategy and MPC for ind...
 
Fault detection of feed water treatment process using PCA-WD with parameter o...
Fault detection of feed water treatment process using PCA-WD with parameter o...Fault detection of feed water treatment process using PCA-WD with parameter o...
Fault detection of feed water treatment process using PCA-WD with parameter o...
 
Model-based adaptive sliding mode control of the subcritical boiler-turbine s...
Model-based adaptive sliding mode control of the subcritical boiler-turbine s...Model-based adaptive sliding mode control of the subcritical boiler-turbine s...
Model-based adaptive sliding mode control of the subcritical boiler-turbine s...
 
A Proportional Integral Estimator-Based Clock Synchronization Protocol for Wi...
A Proportional Integral Estimator-Based Clock Synchronization Protocol for Wi...A Proportional Integral Estimator-Based Clock Synchronization Protocol for Wi...
A Proportional Integral Estimator-Based Clock Synchronization Protocol for Wi...
 
An artificial intelligence based improved classification of two-phase flow patte...
An artificial intelligence based improved classification of two-phase flow patte...An artificial intelligence based improved classification of two-phase flow patte...
An artificial intelligence based improved classification of two-phase flow patte...
 
New Method for Tuning PID Controllers Using a Symmetric Send-On-Delta Samplin...
New Method for Tuning PID Controllers Using a Symmetric Send-On-Delta Samplin...New Method for Tuning PID Controllers Using a Symmetric Send-On-Delta Samplin...
New Method for Tuning PID Controllers Using a Symmetric Send-On-Delta Samplin...
 
Load estimator-based hybrid controller design for two-interleaved boost conve...
Load estimator-based hybrid controller design for two-interleaved boost conve...Load estimator-based hybrid controller design for two-interleaved boost conve...
Load estimator-based hybrid controller design for two-interleaved boost conve...
 
Effects of Wireless Packet Loss in Industrial Process Control Systems
Effects of Wireless Packet Loss in Industrial Process Control SystemsEffects of Wireless Packet Loss in Industrial Process Control Systems
Effects of Wireless Packet Loss in Industrial Process Control Systems
 
Fault Detection in the Distillation Column Process
Fault Detection in the Distillation Column ProcessFault Detection in the Distillation Column Process
Fault Detection in the Distillation Column Process
 
Neural Network-Based Actuator Fault Diagnosis for a Non-Linear Multi-Tank System
Neural Network-Based Actuator Fault Diagnosis for a Non-Linear Multi-Tank SystemNeural Network-Based Actuator Fault Diagnosis for a Non-Linear Multi-Tank System
Neural Network-Based Actuator Fault Diagnosis for a Non-Linear Multi-Tank System
 
A KPI-based process monitoring and fault detection framework for large-scale ...
A KPI-based process monitoring and fault detection framework for large-scale ...A KPI-based process monitoring and fault detection framework for large-scale ...
A KPI-based process monitoring and fault detection framework for large-scale ...
 
An adaptive PID like controller using mix locally recurrent neural network fo...
An adaptive PID like controller using mix locally recurrent neural network fo...An adaptive PID like controller using mix locally recurrent neural network fo...
An adaptive PID like controller using mix locally recurrent neural network fo...
 
A method to remove chattering alarms using median filters
A method to remove chattering alarms using median filtersA method to remove chattering alarms using median filters
A method to remove chattering alarms using median filters
 

Recently uploaded

+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
?#DUbAI#??##{{(☎️+971_581248768%)**%*]'#abortion pills for sale in dubai@
 
Modular Monolith - a Practical Alternative to Microservices @ Devoxx UK 2024
Modular Monolith - a Practical Alternative to Microservices @ Devoxx UK 2024Modular Monolith - a Practical Alternative to Microservices @ Devoxx UK 2024
Modular Monolith - a Practical Alternative to Microservices @ Devoxx UK 2024
Victor Rentea
 
Finding Java's Hidden Performance Traps @ DevoxxUK 2024
Finding Java's Hidden Performance Traps @ DevoxxUK 2024Finding Java's Hidden Performance Traps @ DevoxxUK 2024
Finding Java's Hidden Performance Traps @ DevoxxUK 2024
Victor Rentea
 
Why Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire businessWhy Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire business
panagenda
 

Recently uploaded (20)

Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024
 
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
 
CNIC Information System with Pakdata Cf In Pakistan
CNIC Information System with Pakdata Cf In PakistanCNIC Information System with Pakdata Cf In Pakistan
CNIC Information System with Pakdata Cf In Pakistan
 
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerHow to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected Worker
 
Artificial Intelligence Chap.5 : Uncertainty
Artificial Intelligence Chap.5 : UncertaintyArtificial Intelligence Chap.5 : Uncertainty
Artificial Intelligence Chap.5 : Uncertainty
 
Modular Monolith - a Practical Alternative to Microservices @ Devoxx UK 2024
Modular Monolith - a Practical Alternative to Microservices @ Devoxx UK 2024Modular Monolith - a Practical Alternative to Microservices @ Devoxx UK 2024
Modular Monolith - a Practical Alternative to Microservices @ Devoxx UK 2024
 
"I see eyes in my soup": How Delivery Hero implemented the safety system for ...
"I see eyes in my soup": How Delivery Hero implemented the safety system for ..."I see eyes in my soup": How Delivery Hero implemented the safety system for ...
"I see eyes in my soup": How Delivery Hero implemented the safety system for ...
 
Finding Java's Hidden Performance Traps @ DevoxxUK 2024
Finding Java's Hidden Performance Traps @ DevoxxUK 2024Finding Java's Hidden Performance Traps @ DevoxxUK 2024
Finding Java's Hidden Performance Traps @ DevoxxUK 2024
 
MS Copilot expands with MS Graph connectors
MS Copilot expands with MS Graph connectorsMS Copilot expands with MS Graph connectors
MS Copilot expands with MS Graph connectors
 
DBX First Quarter 2024 Investor Presentation
DBX First Quarter 2024 Investor PresentationDBX First Quarter 2024 Investor Presentation
DBX First Quarter 2024 Investor Presentation
 
Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...
Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...
Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...
 
Strategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a FresherStrategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a Fresher
 
DEV meet-up UiPath Document Understanding May 7 2024 Amsterdam
DEV meet-up UiPath Document Understanding May 7 2024 AmsterdamDEV meet-up UiPath Document Understanding May 7 2024 Amsterdam
DEV meet-up UiPath Document Understanding May 7 2024 Amsterdam
 
TrustArc Webinar - Unlock the Power of AI-Driven Data Discovery
TrustArc Webinar - Unlock the Power of AI-Driven Data DiscoveryTrustArc Webinar - Unlock the Power of AI-Driven Data Discovery
TrustArc Webinar - Unlock the Power of AI-Driven Data Discovery
 
Apidays New York 2024 - The value of a flexible API Management solution for O...
Apidays New York 2024 - The value of a flexible API Management solution for O...Apidays New York 2024 - The value of a flexible API Management solution for O...
Apidays New York 2024 - The value of a flexible API Management solution for O...
 
Ransomware_Q4_2023. The report. [EN].pdf
Ransomware_Q4_2023. The report. [EN].pdfRansomware_Q4_2023. The report. [EN].pdf
Ransomware_Q4_2023. The report. [EN].pdf
 
Why Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire businessWhy Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire business
 
Navigating the Deluge_ Dubai Floods and the Resilience of Dubai International...
Navigating the Deluge_ Dubai Floods and the Resilience of Dubai International...Navigating the Deluge_ Dubai Floods and the Resilience of Dubai International...
Navigating the Deluge_ Dubai Floods and the Resilience of Dubai International...
 
[BuildWithAI] Introduction to Gemini.pdf
[BuildWithAI] Introduction to Gemini.pdf[BuildWithAI] Introduction to Gemini.pdf
[BuildWithAI] Introduction to Gemini.pdf
 
FWD Group - Insurer Innovation Award 2024
FWD Group - Insurer Innovation Award 2024FWD Group - Insurer Innovation Award 2024
FWD Group - Insurer Innovation Award 2024
 

Advantages of the self organizing controller for high-pressure sterilization equipments

  • 1. Practice Article Technical note: Advantages of the self-organizing controller for high-pressure sterilization equipments V. Pilipovik a , C. Riverol b,n a AIChemEng Research and Development Group, J.C Engineers & Partners, Av. Andres Bello, Edif. Las Rozas, Urb. La Florida, Caracas, Venezuela b Chemical Engineering Department, University of the West Indies, St. Augustine Campus, Trinidad, Trinidad and Tobago a r t i c l e i n f o Article history: Received 9 February 2012 Received in revised form 3 July 2013 Accepted 7 July 2013 Available online 29 August 2013 This paper was recommended for publication by Dr. A.B. Rad Keywords: Spores SOM Sterilization Adjustment mechanism Controller a b s t r a c t A study of a self-organizing controller is implemented in a way that response to controlled system follows the desired given by the model. The self-organizing controller has proven to be a valuable tool in sterilization equipment in order to verify the capacity of the response to any change in the pressure or temperature. Basically, this type of controller is based on the Self-Organizing Map (SOM) that is a neural network algorithm of unsupervised learning. The new ideas include clustering visualization, interactive training and one-dimension arrays. & 2013 ISA. Published by Elsevier Ltd. All rights reserved. 1. Introduction High-pressure processing has emerged as an attractive processing technology for preservation of foods. Since the introduction of high pressure processing in 1990 [1,2] the range of products has gradually expanded. The effect that this type of processing has on microorgan- isms is similar to pasteurization while the product retains its fresh or just-prepared appearance and nutritional quality [1,2]. High pressure processing causes minimal changes in the ‘fresh’ characteristics of foods thereby eliminating the processing by thermal degradation. Compared to thermal processing, high pressure sterilization (HPS) results in foods with fresher taste, and better appearance, texture and nutrition. The technology is especially beneficial for heat sensitive products. A recent innovation in sterilization of food products using high pressure is the complete inactivation of vegetative micro-organ- isms, as well as spores, resulting in ambient stable products shown in [2,3]. High pressure sterilization is a combined process where both pressure and temperature contribute to sterilization. A good control system is necessary for keeping the temperature and pressure in the adequate values with less adverse effects on product quality. In this paper a self-organizing controller is introduced as a hierarchical structure in which the inner loop is a table-based controller and the outer loop is the adjustment mechanism. The Self-Organizing Map (SOM) is especially suitable for data analysis [4] where the application in process control is focused on training neural networks [4–6]. It creates a set of prototype vectors representing the data set and carries out a topology preserving projection of the prototypes from the d-dimensional input space onto a low-dimensional grid. This ordered grid can be used as a convenient visualization surface for showing different features of the SOM. In its basic form, it produces a similarity graph of input data. It converts the nonlinear relationships into simple geometric relationships [7,8]. 2. Laboratory set-up and results Destruction of Clostridium sporogenes spores by high pressure treatment was used as the study case. In the laboratory, a sterilizing stainless steel tank of 4 L capacity and 1.5 m length was used, see Fig. 1. The tank has 3 PT100 for monitoring the temperature at the center of the tank. One of the PT100 is located at the top of the tank, another at the middle and the last one at the bottom. The accuracy of the sensor is 70.1 1C. The controller received one value (the average of the three temperatures) because the values from the top to the bottom do not differ more than 1 1C. Contents lists available at ScienceDirect journal homepage: www.elsevier.com/locate/isatrans ISA Transactions 0019-0578/$ - see front matter & 2013 ISA. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.isatra.2013.07.007 n Corresponding author. Tel.: +1 8686622002. E-mail address: Carmen.Riverol@sta.uwi.edu (C. Riverol). ISA Transactions 53 (2014) 186–188
  • 2. The first step is to justify the SOM ability analytically. The first approach was to restrict the consideration to one-dimensional linear unit with which each scalar valued input signal is connected. The SOM control system (SOMCS) adapts the system in accordance with the desired response [7–9]. Each sample is recording and the deviation of desired state is evaluated. An example of the best match is shown in Fig. 2. According to the kinetics of the spore, the adiabatic temperature can increase from 2.2 to 9.3 1C/100 MPa approximately; thus the equipment was tested using three pres- sure pulses of 100, 300 and 600 MPa for 1, 3 and 5 min each, as reported in Table 1. Nevertheless, the importance of the minimum product temperature for the degree of spore inactivation confirms that the temperature should be monitored during high pressure sterilization. On the other hand, the SOM neural network was designed using cluster membership for test data sets that include various levels of data dispersion combined with outliners. A clustering Q means partitioning a data set into a set of clusters Q [10–12]. In crisp clustering, each data sample belongs to exactly one cluster. Fuzzy clustering is a generation of crisp where each sample has different degrees of membership. However, this leaves much room for variation within and between cluster distances and can be defined as follows: dsl ¼ minjjxiÀxjjj ð1Þ dcl ¼ maxjjxiÀxjjj ð2Þ dml ¼ ∑i;jjjxiÀxjjj NiNj ð3Þ outliner ¼ jjciÀcjjj ð4Þ where N is the total value inside the cluster i or j, dsl is the single linkage, dcl is the complete linkage, dml is the average and outliner is the distance between centers shown in [9] and [10]. An example of the cluster definition results is shown in Fig. 2 where the deviation is indicated (d). The idea behind the adapta- tion is to let the adjustment mechanism update the values in the control table F on the basis of the current performance of the controller as reported in Fig. 3. The adjustable neural network taught in previous simulation is inserted in the SOC using the model M. This model was empirically calculated. The adjustable neural network is tuned with each experimental result in the plant (P). Adiabatic heating is the uniform temperature rise within the product, which is solely caused by pressurization [7]. The control Fig. 1. The equipment used in the laboratory. Fig. 2. Cluster definition results. Table 1 Experimental set-up and temperature control (SOC). Pressure (MPa) Target temp (1C) Initial temperature (1C) Mean temperature achieved (1C) SOM Mean temperature achieved (1C) PI 100 70 40 70.8 71.2 85 55 85.6 85.9 100 60 101 99.6 300 70 41 70.1 71.2 85 58 84.9 84.6 100 62 100.6 101.3 600 70 51 71.2 71.6 85 56 85.3 85.9 100 66 100.3 100.7 V. Pilipovik, C. Riverol / ISA Transactions 53 (2014) 186–188 187
  • 3. system was implemented using MATLAB 7.0, real time toolbox and MF624 multifunction I/O card. A SOM was trained using the sequential training algorithm [8]. All maps were linear. In the agglomerative clustering (see Fig. 2), single, average and complete linkages were used in the construction phase. For additional information about SOM, readers are referred to [5]. In this article the training time was 32 h using 10,000 values collected from 2005 to 2010. The input values were obtained during the last 5 years because this equipment worked 35 years without any control system. How to tune the gains does deserve attention, since they may finally stand in the way for a successful implementation. The SOC works ‘surprisingly well’ [13]. For example, the output gain is lowered between two training sessions. The adjustment is com- pensated by an F-table with numbers of larger magnitude [13–15]. Therefore it is possible to start with a linear F-table, and set the gains loosely according to a PID tuning rule or hand-tuning. This is a good starting point for self-organization. The SOM network can improve the quality of decisions in cluster analysis especially in non-uniform cluster densities [5,6,13]; for example, temperature is a parameter that changes very fast; thus some of empirical data can become “messy data”. Fig. 4 depicts an example in which the SOM controller offers a better performance than the PI controller. Basically, the overshoot is reduced and the setting conditions are attained early. The highest temperature with the SOMCS is about 19.5 1C lower than that with PI as reported in Fig. 4. The same behavior was observed 24 times during the testing sections. The reduction of the overshoot can be translated in a reduction of the steam consump- tion; thus some hundreds of dollars can be saved. Also an over- shoot over 10 1C is not desirable because the food can become dry and lose its texture because of which designing of new controllers may continue in this area. Table 1 indicates that the SOMCS offers a better performance over PI in this system and shows that the results are not sensitive to the initial learning coefficients at different initial conditions. In summary, the new control system shows a good performance and can also improve the traditional PI controller; nevertheless, future research should focus on reducing the training time for the SOM because its implementation on the equipment consumes a long time. 3. Conclusion This paper presented the development of a SOM controller and its application to a high pressure sterilization equipment. The algorithm presented a satisfactory performance and efficiency. The simulation showed that it is possible to achieve a good behavior after few steps of learning using a simple model. The implementa- tion in the sterilization equipment improves the response capacity of the control system although the implementation of the self- organizing system consumes time and is complex. The adjustable control table (F) can be increased using more data for improving the accuracy; however a good computer should be used for reducing the time consumption (the computer used in this article has a 2GB RAM only). References [1] Matser A, Krebbers B, van den Berg R, Bartels P. Advantages of high pressure sterilization on quality of food products. Trends in Food Science & Technology 2004;15(2):79–85. [2] Meyer R, Cooper K, Knorr D, Lelieveld H. High pressure sterilization of foods. Food Technology 2000;54(11):67–72. [3] Hoogland H, De Heij W, Van Schepdael L. High pressure sterilization: novel technology, new products, new opportunities. New Food 2001;3:21–6. [4] Jantzen J. Foundations of fuzzy controller. 1st ed. NY, USA: Wiley; 117–39. [5] Bloch G, Denoeux T. Neural networks for process control and optimization: two industrial applications. ISA Transactions 2003;42(1):39–51. [6] Das S, Saha S, Das S, Gupta A. On the selection of tuning methodology of FOPID controllers for the control of higher order processes. ISA Transactions 2011;50 (3):376–88. [7] Dovžan D, Škrjanc I. Recursive fuzzy c-means clustering for recursive fuzzy identification of time-varying processes. ISA Transactions 2011;50(2):159–69. [8] Lokriti A, Salhi I, Doubabi S, Zidani Y. Induction motor speed drive improve- ment using fuzzy IP-self-tuning controller. A real time implementation. ISA Transactions 2013;52(3):406–17. [9] Tisan A, Cirstea M. SOM neural network design—a new Simulink library based approach targeting FPGA implementation. Mathematics and Computers in Simulation 2013;91:134–49. [10] Rizal M, Ghani Jaharah A, Nuawi M, Hassan C, Haron C. Online tool wear prediction system in the turning process using an adaptive neuro-fuzzy inference system. Applied Soft Computing 2013;13(4):1960–8. [11] Yamazaki T Mamdani TE. On the performance of a rule-based self-organizing controller. In: Proceedings of the IEEE conference on applications of adaptive and multivariable control. 19–21 July, Hull; 1982. p. 121–32. [12] Vesanto J, Himberg J, Alhoniemi E, Parhankangas J.. Self-organizing map in Matlab: the SOM Toolbox. In: Proceedings of the Matlab DSP conference. Espoo, Finland; 1999. p. 35–40. [13] Nauck D, Klawonn F, Kruse R. Foundations of neuro-fuzzy systems. NY, USA: John Wiley & Sons; 37–56. [14] Kohonen T. Self organizing maps. 3rd ed. NY, USA: Springer; 71–99. [15] Ghaseminezhad M, Karami A. A novel self-organizing map (SOM) neural networks discrete gropus of data clustering. Applied Soft Computing 2011;11 (4):3771–8. Fig. 3. SOC control system. The solid line indicates the field data and the dash lines indicate the calculated values. 0 5 10 15 20 25 30 35 40 45 50 0 20 40 60 80 100 120 Time (min) Temperature(oC) PI SOM-NN Fig. 4. Behavior of the temperature using different controllers. P¼600 MPa. V. Pilipovik, C. Riverol / ISA Transactions 53 (2014) 186–188188