Advances in Natural and Applied Sciences, 9(8) July 2015, Pages: 36-42

AENSI Journals

Advances in Natural and Applied Sciences

ISSN:1995-0772 EISSN: 1998-1090

Journal home page: www.aensiweb.com/ANAS

Hazad And Operability Study and Analysis of Safety Integrity Level Case Study:
Ammonia Refrigerant Compressor at Petrocemical Plant
Ali Musyafa, Resti Nabila, Ronny D. Noriyati

Department of Engineering Physics, Faculty of Industrial Technology, Sepuluh Nopember Institute Of Technology,Box. Surabaya,
INDONESIA

ARTICLE INFO ABSTRACT

Article history: Background: On the refrigeration system at the Petrocemical Plant there ammonia
Received 1 June 2015 refrigerant compressor 105 J that operating to raise the vapor pressure of ammonia that
Accepted 28 June 2015 comes from the refrigerant drum 120 CF flush and keep the pressure on the refrigerant
Available online 22 July 2015 flush drum 120 CF in order to maintain their liquid ammonia product temperature at -33
° C. To avoid the risks that occur in ammonia refrigerant compressor 105 J we perform
Keywords: the identification of potential Hazard that can occur in the process in ammonia
Keyword 1 (Compressor); refrigerant compressor 105 J by using Hazard and Operability Analysis (HAZOP) and
Keyword 2 (Hazop); SIL analysis. In this study determined four nodes consisting of ammonia refrigerant
Keyword 3 (SIS ); compressor stage 1 to 105 J ammonia refrigerant compressor 105 J stage 4. Based on
Keyword 4 (SIL ) ; the identification of 20 instruments are installed on all four nodes, the valve
Keyword 5 (PFD); components, namely the level of 1024 which has a high risk based standards
PT.Petrokimia Gresik. For Safety Integrity Level (SIL) value obtained from the
calculation of Safety Instrumented System (SIS) in ammonia refrigerant compressor
105 J is NO SIL. With SIS Probability Failure Demand (PFD) worth 0.3 and RRF at
3:33. The recommendations on the level of risk reduction valve 1024 level by adding
redundant generate failure rate is much lower at 4:45 x 10-10 and PFD total of 1.28 x
10-4 worth SIS SIL 2. The recommendations of the panel in the form of local turnover
based relay into Programable Logic Control (PLC) generates total PFD total =0.038
worth SIL 1.

© 2015 AENSI Publisher All rights reserved.

To Cite This Article: Ali Musyafa, Resti Nabila, Ronny D. Noriyati, Hazad And Operability Study and Analysis of Safety Integrity Level
Case Study: Ammonia Refrigerant Compressor at Petrocemical Plant. Adv. in Nat. Appl. Sci., 9(8): 36-42, 2015

INTRODUCTION When there is damage to the compressor will

pose a danger and a losses to the company because
The ammonia refrigeration system at ammonia the ammonia production will be stopped. Therefore
plant uses refrigeration cycle system consisting of perform the identification of hazards that may occur
ammonia refrigerant flush drum, ammonia in the process of ammonia refrigerant in the
refrigerant compressor, condenser refrigerant and compressor in order to keep the process still takes
expansion valve. Ammonia refrigerant compressor place safely. Hazard identification method
consists of two cases, namely low pressure and high commonly used is the Hazard and Operability
pressure. This compressor is driven by a turbine with (HAZOP). With the HAZOP study method on a
a capacity of 5.625 KW, rotation of the turbine is system that includes an evaluation and analysis of
controlled by a governor valve which includes steam process deviations know who is known by the data
which will drive the compressor at the same tubin. process with several parameters such as flow,
Aperture governor valve depends on the pressure that temperature and level. Evaluation studies conducted
is on refrigerant drum flush stage 1 because the aim at four points using process data obtained during the
is to keep the pressure on the refrigerant flush drum. first month (January 31, 2015 until February 28,
In addition it also has a compressor antisurge control 2015) with a time interval of 2 hours, then plotted to
which serves to protect the compressor from danger determine the deviation that occurs by using
backflow due to the difference in pressure Microsoft Excel. Then analysis the protection system
compressor suction and discharge. (Petrokimia of ammonia refrigerant compressor 105 J using Fault
Committee, M. T. 2010). Tree Analysis (FTA) aimed at assessing the SIL level

Corresponding Author: Ali Musyafa, Department of Engineering Physics, Faculty of Industrial Technology, Sepuluh
Nopember Institute Of Technology, Kampus ITS Keputih, Sukolilo, Surabaya – Indonesia, 60111.
Tel: 62-31-5966138/5947188 E-mail: [email protected]
37 Ali Musyafa et al, 2015
Advances in Natural and Applied Sciences, 9(8) July 2015, Pages: 36-42

of the SIS that has been installed on the ammonia to get input in the form of ammonia vapor that comes
refrigerant compressor. (Petrokimia Committee, M. from the refrigerant flush drum. Compressor consists
T. 2010,)( Adiyagsa, H., & Musyafa, A. 2012)). of four stages, namely the LP. Casing 1 st Stage, LP.
Casing 2nd Stage, HP. 1st Stage casing and HP.
1. Principle of Ammonia Refrigerant Compressor Casing 2nd Stage. (Petrokimia Committee, M. T.
The early stages of the evaluation process is to ,2010)
learn the process and retrieval of data relating to the Each stage serves to take the refrigerant vapor
object of research, namely ammonia refrigerant coming from the flush drum while keeping the
compressor. Data needed as Piping and Instrument pressure on the fourth stage flush refrigerant drum.
Diagrams (P&ID), Process Flow Diagram, Data Stage 1 is maintained the pressure about 0.05 kg /
Maintenance, and Process Data recorded by cm2 and temperature -33 ° C, stage 2 kept the
Distributed Control System (DCS). This study aims pressure around 1.8 kg / cm2 and temperature -12 °
to determine the potential hazards and risks in C, stage 3 kept the pressure around 3.4 kg / cm2 and
ammonia refrigerant compressor unit with HAZOP temperature -12 ° C. Stage 4 kept the pressure around
method and determine the value of the Safety 7.9 kg / cm2 and temperature of 13 ° C which aims to
Integrity Level (SIL) of ammonia refrigerant keep the temperature remains cool liquid ammonia.
compressor system. Ammonia refrigerant compressor

128 C 167-C

TO 127 C

HP. CASING HP. CASING LP. CASING LP. CASING TURBIN

2nd STAGE 1ST STAGE 2nd STAGE 1ST STAGE 105 JT

120 CF 120 CF 120 CF 120 CF

4TH STAGE 3RD STAGE 2ND STAGE 1ST STAGE

TK801

Fig. 1: Process of Ammonia Refrigerantm Compressor (Petrokimia Committee, M. T. 2010,)

Each stage in the ammonia refrigerant drums are 2. Guide Word with Statistical Process Control:
flush antisurge control which serves to maintain the The next step is determining guide word from
flow of input to the compressor suction still meet the data process recording on DCS, the data is processed
minimum flow to prevent surges when flow towards using a control chart Xbar-S chart using the equation:
the suction pressure becomes so low that low and
high head while the pressure on the higher discharge The X bar chart determined by equation:
causing backflow potentially damage the
UCL = x + A3 s (1)
compressor. Control surge mounted on a stage 1 to
stage 4 with a tag number and naming of this kind CL = x (2)
FIC 1012/1011/1010/1009, FT 1012/1011/1010/1009 LCL = x - A3 s (3)
and FV 1012/1011/1010/1009. In addition there is
where :
also a surge control level control that keeps the
drums flush level in order to avoid high-high level x = The average value of subgroup
and low-low level because it will cause damage to
the compressor. The process of system can be seen in
s = The average of standard deviation of
subgroup
Fig 1. (Ebeling, C. E. (1997). A3 = Constant based on the size of the subgroups
38 Ali Musyafa et al, 2015
Advances in Natural and Applied Sciences, 9(5) July 2015, Pages: 141-146

The S-Chart determined by equation be obtained which is used to determined the

UCL = B4 s (4) likelihood. Likelihood can be searched by the
equation:
CL = s (5) Τime intrererval- οf  data aquisition (7)
Likelihood 
LCL = B3 s (6) ΜΤΤF
Consequences is the impact of the incident.
Where :
Criteria consequences and likelihood in this study
B4 and B3 are constants based on the subgroup size.
using a standard PT. Petrokimia Gresik.
From the data the process can be seen
Multiplication of likelihood and consequences
irregularities are legible and can be determined guide
generate risk matrix. (Adiyagsa, H., & Musyafa, A.
his word like No, More of, Less of, as well as and
2012), (Ebeling, C. E. (1997.).
Other Than. Then, based on data maintenance MTTF

Table 1: Risk Matrix

The classification level of risk at PT. Petro qualitative methods Fault Tree Analysis (FTA). FTA
Kimia Gresik only 3 are low risk, moderate risk and can identify the basic event so that events can be the
high risk. Based on the risk matrix that has been cause of the initial event. Quantitative methods to
formed can be determined what instruments have the calculate the Probability Failure on Demand (PFD).
highest risk and then do an evaluation to reduce the PFD system is determined based on the basic event
risk of the instrument and determine actions. that has been obtained from the analysis of the FTA.
To analyze the safety instrumented system (SIS) PFD system determines the SIL value system which
in ammonia refrigerant compressor is used can be seen in table 2.

Table 2: PFD (Probability of Failure on Demand) of low demand operation and RRF for different SILs as defined in IEC EN 61508
SIL PFD RRF
4 10-5 s/d 10-4 100000 s/d 10000
3 10-4 s/d 10-3 10000 s/d 1000
2 10-3 s/d 10-2 1000 s/d 100
1 10-2 s/d 10-1 100 d 10

SI SSHH
1005A 1005
PZ
1009

TO SAHH 1005
TO LP. CASING LP. CASING Turbin
2ND STAGE 1ST STAGE 105 JT

FROM 105 J
KICKBACK
SI SSH
FV 1005
1011 1005B
PI
1646
PIC
FY
1009
1012
TO SAH 1005

FIC FT
1012 1012

PT
1009

LSHH
1217 LT
FROM 120CF STAGE 2
1024

120CF STAGE 1 LIC

1024

FY
1023

LV
1023

COLD NH3
LAHH 1217 TO TK801

Fig. 2: Low Pressure Casing 1st Stage of Turbine, UNEP, (January ,01-01-2015)
39 Ali Musyafa et al, 2015
Advances in Natural and Applied Sciences, 9(5) July 2015, Pages: 141-146

3. HAZOP Study:
In this study the ammonia refrigerant
compressor is divided into four nodes, namely, LP.
Casing 1st Stage, LP. Casing 2nd Stage, HP. 1st
Stage casing and HP. Casing 2nd Stage. This node
determination based on the sequence masukknya Daily

ammonia vapor flow toward the compressor. At each Fig. 3: Control Chart Xbar 01LT1024,Level
node to be identified based on the data contained in Transmiter
the historical destruction of data maintenance for 10
years. Then the process of data taken with variable
flow and level on each node. In the process of data
collection have 9 transmitter readings are then
processed by equation (1) to equation (6) using
Microsoft excel. Here one Xbar Chart and S-Chart on
node 1 is obtained by reading LT 1024. Hyatt, N.
(2003), (ISA. 2002)
Daily

4. Control Chart for Level Transminte 1024:

Fig. 4: Control Chart S-01LT1024 ,Level Transmiter
The deviation can be analyzed from the xbar
chart, can be seen in Fig 3 that there are points which
As analyzed one by one by reading 9 transmitter
are beyond the upper control limit (UCL) and there
located on the fourth node can be determined
are points that are outside the lower control limit
deviation and guide word shown in Table 3. Hyatt,
(UCL) so that deviations can be determined that is
N. (2003), (ISA 2002), (Musyafa, a., &
more level and less level. (Montgomery, D. C. 2009).
Kristianingsih, l. 2013).
Table 3: Guide word and deviation node 1
No. Instrument Guide word Deviation
1 Flow Valve Other Than Other Than Flow
01FV1012
2 Flow Transmitter 01FT1012 Other Than Other Than Flow
3 Level Valve More More Flow
01LV1024 Less Less Flow
4 Level Transmitter More More Level
01LT1024 Less Less Level
5 Pressure Transmitter 01PT1009 More More Pressure
Less Less Pressure
6 Pressure Controller More More Pressure
01PIC1009 Less Less Pressure

Table 4: Guide word and deviation node 2

No. Instrument Guide word Deviation
1 Flow Valve Other Than Other Than Flow
01FV1011
2 Level Transmitter More More Level
01LT1023 Less Less Level
3 Level Valve More More Flow
01LV1023 Less Less Flow
4 Intercooler High High Temperature
167 C Misdirection Misdirection Flow

Table 5: Guide word and deviation node 3

No. Instrument Guide word Deviation
1 Flow Valve Other Than Other Than Flow
01FV1010
2 Flow Transmitter Other Than Other Than Flow
01FT1010
3 Level Valve More More Flow
01LV1022 Less Less Flow
4 Level Transmitter More More Level
01LT1022 Less Less Level
5 Intercooler High High Temperature
128 C Misdirection Misdirection Flow

Table 6: Guide word and deviasation node 4

No. Instrument Guide word Deviation
1 Flow Valve Other Than Other Than Flow
01FV1009
2 Flow Transmitter Other Than Other Than Flow
40 Ali Musyafa et al, 2015
Advances in Natural and Applied Sciences, 9(5) July 2015, Pages: 141-146

01FT1009
3 Level Valve More More Flow
01LV1021 Less Less Flow
4 Level Transmitter More More Level
01LT1021 Less Less Level

Table 6: Risk Matrix Ammonia Refrigerant Compressor

Based on the risk analysis has been done, by SIS

Ammonia Refrigerant
Compressor 105 J Failure
multiplying the likelihood and consequences of the (T)

following are the results of the existing risk matrix at

node 1 to node 4. (ISA. 2002
Results of risk matrix of four nodes of ammonia Alarm Overspeed Alarm High High Alarm High High
Local Panel
Failure ESDV
105 JT Level 120 CF Vibrasi
refrigerant compressor 105 J, the risk of having a low (B) (A) (A)
(C) (D)

percentage of 21.62 %. Moderate risk has the highest SSHH

1005
ZSHH
6510
Relay
Failure
(Q)
percentage amount by 72.97% and high risk with the (P) (O)

least amount of percentage is 5:40%. Shown in Risk XV 6501

(V)
XV 6502
(X)

Matrix Annex-1.
LSHH LSHH LSHH LSHH

RESULTS AND DISCUSSION 1217

(K)
1216
(L)
1215
(M)
1214
(N)

SIL Analysis, Safety instrumented system (SIS) Fig. 5: FTA for SIS compressor (Musyafa, A. et al.
in ammonia refrigerant compressor serves as a 2013).
protection to the security system in the process of
ammonia refrigerant compressor when BPCS have Based on fault tree in fig 5 can be determined
been unable to cope with danger. SIL calculation with an equation Boolean logic in the following:
analysis on SIS in ammonia refrigerant compressor is T=A∪B∪C∪D
done by using logic Fault Tree Analysis (FTA) to see = (K ∪ L ∪ M ∪ N ) ∪ O ∪ P ∪ Q ∪ (R ∩ S)
the basic event of the cause of its failure occurred in = (K ∪ L ∪ M ∪ N ) ∪ O ∪ P ∪ Q ∪ ((V) ∩ (X))
the SIS in ammonia refrigerant compressor, and then = K + L + M + N + O + P + Q + ( (V) * (X))
calculate PFD based on logic are obtained. FTA So that PDF-SIS calculation based on basic even that
which has a top event SIS failure as fig. 5, (Hyatt, N. is acquired is:
2003), (ISA. 2002).

Table 7: Calculation of PFD_SIS.

Instrument Failure Rate PFD PFD SIS RRF
LSHH 1.74x10-6 0.00762
SSHH 0.48x10-6 0.00210 0.3 3.33
ZSHH 0.57x10-6 0.00249
Relay 0.66x10-6 0.26800
XV 6501 2.23x10-6 0.00976
XV 6502 0.66x10-6 0.28000

The value of pfd sis obtained from the done there are two deviation with the risk of high
calculation of the existing sis is 0.3, based on table 2 that was found at the nodes 1 .The high risk to exist
and the value of pfd 0.3 not included in the level of at the level of valve instrument 1024. To reduce the
the SIL 1 - 4, so worth NO SIL. (SINTEF. 2009), risk of failure caused by instrument of 1024 then give
the recommendations to add redundant on instrument
Recommendation: high risk:
Based on the evaluation hazop that has been
41 Ali Musyafa et al, 2015
Advances in Natural and Applied Sciences, 9(5) July 2015, Pages: 141-146

Existing LV 1024:
Instrument MTTF C L RR
01LV1024 12108 4 4 16 5.4 x 10-5

PFDredundant   xTi  (8.25 x10 ) x(6760 )

PFDLV102 2 2 6 2 2

T 8760
 x i  (5.4 x105 ) x  0.23, No _ SIL 3 3
2 2 = 1.74 x 10-3
Then added level valve as a redundant valve so
PFDLV1024   xTi  (5.4 x10 ) x(9760 ) =
2 2 5 2 2

that the vote turned into 1oo2, with the result that,
3 3
= 0.074
LV 1024
λ = 4.45 x 10^-10 PFD total = (1.74 x 10-3) x 0.074 = 1.28 x 10-4 (SIL
2)

λ λ
0.000054 0.00000825
LV 1024
LV 1024

With the result that change into :

Instrument MTTF C L RR λ
01LV1024 224719 4 1 4 4.45 x 10-10

Based on the analysis to the existing SIS SIL in value of its local panel. The recommendations on the
ammonia refrigerant compressor 105 J produces NO level of risk reduction valve 1024 level by adding
SIL value. The largest component PFD value derived redundant produce a failure rate that is much lower at
from the local panel 105 J, as based on the failure of 4:45 x 10-10 and PFD total worth of 1.28 x 10-4 SIL
local panel causes the most of his trip compressor. 2. The results of the SIS recommendations in the
PLC to be recommended is TUV certified Honeywell form of replacement of relay-based local panel into
PLC for SIL 2 with a failure rate of 7.4 x 10-9. So if the PLC generates total PFD 0038 is worth SIL1.
the local panel is replaced with a failure rate then the
calculation becomes: REFERENCES
PFDPLC =  x Ti  (7.4 x109 ) x 8760  0.00003
2 2 Adiyagsa, H., A. Musyafa, 2012. Hazard and
PFDSIS = PFDLSHH + PFDZSHH + PFDSSHH + Operability study in Boiler System of The Steam
PFDLocalPanel + PFDESDV Power Plant. IEESE International Journal of Science
=[ 0.0076212 + 0.0076212 + 0.0076212 + and Technology (IJSTE), 1(3).
0.0076212] + 0.0021 + 0.0000324 + 0.002496 + Ebeling, C.E., 1997. An Introduction to
[(0.00976) x (0.28)] = 0.0304848 + 0.0021 + Reliability andMaintainability Engineering.
0.0000324 + 0.00296 + 0.00273 = 0.038 = (SIL 1 , Singapore: The McGraw – HillCompanies.
IEC EN 61508 standard ). Hyatt, N., 2003. Guidelines for Process Hazard
The results of the replacement of local panel can Analysis, Hazards Identification & Risk Ananlysis.
increase the value of the SIL .The existing SIL worth London, New York, Washington D.C: CRC Press.
no SIL and the recommendation is worth the SIL 1. ISA, 2002. Safety Instrumented Functions (SIF)
- Safety Integrity Level (SIL) Evaluation Techniques
Conclusion: Part 1: Introduction. ISA The
Based on the risk matrix of four nodes of Instrumentation,Systems, and Automation Society.
ammonia refrigerant compressor 105 J, the risk of ISA, 2002. Safety Instrumented Functions (SIF)
having a low percentage of 21.62%. Moderate risk - Safety Integrity Level (SIL) Evaluation Techniques
have the highest percentage amount by 72.97% and Part 2: Determining the SIL of a SIF via Simplified
high risk with the least amount of percentage is Equations. ISA – The Instrumentation, Systems, and
5:40%. Automation Society.
Based on the HAZOP evaluation has been done, Montgomery, D.C., 2009. Introduction to
a component with a high risk of a fourth node in Statistical Quality Control (Vol. Six Edition). John
Ammonia Refrigerant Compressor 105 J is the level Wiley & Sons, Inc.
of 1024 by criteria likelihood valve 4 and Musyafa, A., 2013. Risk Management Using
consequences appreciating 4. HAZOP Study Method Base Fault Tree Analysis on
The value of SIL in ammonia refrigerant Emergency Shutdown System-Vacuum Distillation
compressor 105 J is NO SIL, with PFD system 0.3 Unit, PT.PQR, Dumai, Indonesia, Asian Transactions
and RRF 3:33. This is because the failure rate of a on Engineering (ATE ISSN: 2221-4267) Volume 03
local panel of 105 J is high at 6:55 x 10-5 so that pfd Issue 05.
42 Ali Musyafa et al, 2015
Advances in Natural and Applied Sciences, 9(5) July 2015, Pages: 141-146

Musyafa, A., 2014. “Reliability and Technology Committee – Risk Management Working
Maintainability Assessment of the Steam Turbine Group.
Instrumentation System for optimization Operational plc, M.I., 2002. An introduction to Functional
Availability System at Fertilizer Plant” Australian Safety and IEC 61508
journal of Basic and Applied Science (AJBAS), SINTEF. 2009., Industrial Management.
ISSN: 1991-8178. Published August. 2014. & “Offshore Reliability Data Handbook 4th Edition”.
Journal home page: http://www.ajbasweb.com. OREDA Participants,2000
Musyafa,, A., l. Kristianingsih, 2013. Risk UNEP, (January ,01-01-2015),
Management and Safety System Assessment from COMPRESSORS AND COMPRESSED AIR
Power Plant Steam Boiler in Power Systems Unit 5, SYSTEM, Energy Efficiency Guide for Industry in
Paiton-Indonesia. Australian Journal of Basic and Asia – www.energyefficiencyasia.org
Applied Sciences, pp: 349-356. Unido, 2014. Compressor Air System , Unido
Petrokimia Committee, M.T., 2010. Risk Patner and prosperity
Management Training Guides. Manufacturing

Annex-1: Risk Matrix

Annex -1: Risk Matrix of ammonia refrigerant compressor 105 J
Guide word Deviation Cause Consequences Safeguard Risk Score
C L RR
Other Than Other Than Flow FV 1012 failed to close when Occurs higher pressure and PAH 1009 2 3 6
the minimum input flow on the high temperature in the LP. RV 1018
suction LP.casing 1st stage has 1st stage suction casing but
been reached no mechanical failure
Control surge malfunction due FV valve 1012 opening is not PAH-1009 2 2 4
to damage to the FT 1012 in accordance with RV1018
(upper) requirements

More More Flow LV 1024 fails to close or does The loss of liquid level at 120 LAL-1024 4 4 16
not match the required CF1 which can lead to
percentage comp.105 J surges and
damaged
More More Level An interruption in the flow High level at 120 CF1 allows LAH-1024 4 1 4
output 120-CF1 to storage tanks entrainment of liquid LAHH-1217
because of an error reading by ammonia into the comp. 105 leads to 105-J
LT 1012 J can make the broken blade trip
and shutdown comp. 105 J
Less Less Flow LV 1024 failed to open or did Level high at 120 CF1 allows LAH-1024 4 2 16
not fit the required percentage entrainment of liquid LAHH-1217
ammonia into the comp. 105 leads to 105-J
J can make a blade is trip
damaged and shut down
comp. 105 J
Less Less Level Input ammonia liquid derived The loss of liquid level at 120 LAL-1024 4 1 4
from 120 CF2 too low CF1 which can lead to
comp.105 J surges and
damaged