Temperature Optimization of A Naphtha Splitter Unit
Temperature Optimization of A Naphtha Splitter Unit
Temperature Optimization of A Naphtha Splitter Unit
Temperature Optimization of
a Naphtha Splitter Unit
Antonio NEVADO1, Lourdes Martn OATE2,
Javier SANZ3, Ral ALCALDE4, Ken SLAVEN5
1,2,5
UNED (Spain) ([email protected],[email protected]), ADEX S.L.
([email protected] )
3,4
Repsol-YPF (Puertollano-Spain) ([email protected], [email protected])
AbstractA fully reliable and efficient adaptive control
methodology has been long awaited in industry due to the
time-varying nature of industrial plants. This paper
demonstrates that this kind of adaptive solution is now
available and simple to apply by presenting the first
application of a methodology called Adaptive Predictive
Expert (ADEX) Control in a petrochemical production unit.
A description of the plant and the ADEX solution is followed
by a comparative analysis of the results obtained with those
of the existing conventional PID control. The objectives of
the application involving the naphtha splitter at the
Puertollano Refinery of Repsol were to establish the
viability of ADEX in this environment, increase the quality
of the naphtha products by achieving closer compliance with
desired specification and maximizing the economic yield. To
achieve these objectives it was necessary to enhance the
stability of the naphtha splitter by improving control of the
principal variables and eliminating a degree of interaction
between them which was causing a resonance problem in the
column. The application of the ADEX system confirm its
viability, demonstrated a marked increase in column
stability, significant improvements in levels of control and
the elimination of the resonance problems.
Index TermsAdaptive Predictive Expert, Control,
optimization, splitter unit
I.
INTRODUCTION
[Downloaded from www.aece.ro on Tuesday, February 16, 2016 at 07:23:50 (UTC) by 117.211.2.119. Redistribution subject to AECE license or copyright.]
Flare
Light naphtha
Condenser
Stabilizer
Naphtha
Splitter
C30
Light Naphtha
Reboiler
Atmospheric
Distillation
Heavy Naphtha
[Downloaded from www.aece.ro on Tuesday, February 16, 2016 at 07:23:50 (UTC) by 117.211.2.119. Redistribution subject to AECE license or copyright.]
SET POINT
DRIVER
BLOCK
DESIRED
OUTPUT
CONTROL
BLOCK
CONTROL
SIGNAL
PROCESS
PROCESS
OUTPUT
ADAPTIVE
MECHANISM
Set Point
AP
DOMAIN
EXPERT
DOMAIN
Initial
Value
10 O C
Control
Signal
EXPERT
AP
CONTROL
CONTROL
10
15
20
25
30
35
40
45
[Downloaded from www.aece.ro on Tuesday, February 16, 2016 at 07:23:50 (UTC) by 117.211.2.119. Redistribution subject to AECE license or copyright.]
ADEX
INPUTS
CONTROLLERS
ADEX
OUTPUTS
Control &
Optimisation
Logic
CONTROL
VECTOR
DATA
ACQUISITION
Control &
Optimisation
Logic
LOCKING
LOGIC
PROCESS
PID Regulation
Pressure
Level
Temperature
Range of measurement
0 - 1 bar
0 - 100 %
0 - 140 C
0.9
55.0
133.0
Mean
0.899976
54.977456
133.016447
Variance
0.000047
14.422047
0.465088
Standard Deviation
0.006883
3.797637
0.681973
Maximum Value
0.942405
87.512817
135.319901
Minimum Value
0.856463
33.600918
130.128250
Maximum Deviation
0.043537
32.512817
2.8715000
Set Point
10
[Downloaded from www.aece.ro on Tuesday, February 16, 2016 at 07:23:50 (UTC) by 117.211.2.119. Redistribution subject to AECE license or copyright.]
bars
(a)
Tn/h
HN flow rate
HN level
(b)
Steam valve
HN Temperature
(c)
Figure 2. Control with conventional method: (a) Pressure, (b) Level, and (c) Temperature.
11
[Downloaded from www.aece.ro on Tuesday, February 16, 2016 at 07:23:50 (UTC) by 117.211.2.119. Redistribution subject to AECE license or copyright.]
V.
12
Temperature Control
[Downloaded from www.aece.ro on Tuesday, February 16, 2016 at 07:23:50 (UTC) by 117.211.2.119. Redistribution subject to AECE license or copyright.]
bars
(a)
Tn/h
HN flow rate
HN level
(b)
Steam valve
HN Temperature
(c)
Figure 3. Control under ADEX: (a) Pressure, (b) Level, and (c) Temperature.
13
[Downloaded from www.aece.ro on Tuesday, February 16, 2016 at 07:23:50 (UTC) by 117.211.2.119. Redistribution subject to AECE license or copyright.]
ADEX CONTROL
Set Point
Mean
Variance
Standard Deviation
Maximum Value
Minimum Value
Maximum Deviation
Pressure
0.900000
0.900197
0.000002
0.001446
0.911395
0.880593
0.019407
Level
50.000000
50.028710
10.394441
3.224041
64.882935
24.486973
25.513027
Temperature
133.0000
132.998559
0.145025
0.380821
134.896301
131.654877
1.896301
undergo excursions so far from the set point, and the band of
oscillations is considerably reduced.
It was proved experimentally that the intensity of pressure
oscillations significantly affected the level and temperature
variables. When ADEX was applied to pressure control, the
effect of the large pressure oscillations on the level and
temperature control was also reduced as a consequence and this
in turn avoided the resonance phenomena between the
variables.
In Fig. 8, the histograms of both control methods are shown.
It can be seen how they are centered around the set point, and
how the profiles of the graphs in the case of ADEX control are
much more peaked around the set point, with smaller tails at
the ends. This demonstrates more precise control with fewer
excursions away from the set point.
B. Level Control
Once the pressure was stabilized by applying ADEX, the
level control also achieved a reduction in maximum deviations,
this time by 20% and a reduction in standard deviation by 15%.
The variable oscillated 66% of the time between +/- 3.2 %. In
addition to improving the level control, the set point of the
level was reduced to 50% from 55% which was the previous
normal level.
C. Temperature Control
The control of temperature improved by reducing the
maximum deviation by 37% from 5.2C to 3.2C and a
reduction in standard deviation to almost half i.e.+/- 0.68C to
+/-0.38C.
Variable
Control method
14
Pressure
Level
Temperature
PID
ADEX
PID
ADEX
PID
ADEX
Standard Deviation
0.006883
0.001446
3.797637
3.224041
0.681973
0.380821
Maximum Deviation
0.043537
0.019407
32.512817
25.513027
2.8715000
1.896301
[Downloaded from www.aece.ro on Tuesday, February 16, 2016 at 07:23:50 (UTC) by 117.211.2.119. Redistribution subject to AECE license or copyright.]
PID Control
ADEX Control
Fig.ure 8. Frequency Histogram of pressure control under PID (left) and ADEX (right) control. Frequencies expressed as % of total number of values.
PID Control
ADEX Control
Fig.ure 9. Frequency Histogram of level control under PID (left) and ADEX (right) control. Frequencies expressed as % of total number of values.
PID Control
ADEX Control
Fig.ure 10. Frequency Histogram of temperature control under PID (left) and ADEX (right) control. Frequencies expressed as % of total number of values.
15
[Downloaded from www.aece.ro on Tuesday, February 16, 2016 at 07:23:50 (UTC) by 117.211.2.119. Redistribution subject to AECE license or copyright.]
VI.
CONCLUSIONS
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
REFERENCES
[1]
[2]
[3]
16
[21]
[22]
[23]
[24]