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

net/publication/311068561

Prediction Of Failures By Using Crow/AMSAA Model In Microsoft Excel

Article · November 2016

CITATIONS READS
0 1,792

3 authors:

Masdi Muhammad Tahir Raza

PETRONAS Universiti Teknologi PETRONAS
44 PUBLICATIONS   168 CITATIONS    4 PUBLICATIONS   2 CITATIONS   

SEE PROFILE SEE PROFILE

Mohd Amin Abd Majid

Universiti Teknologi PETRONAS
19 PUBLICATIONS   38 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

prediction modeling for Corrosion under Insulation View project

Development of maintenance performance measurement framework for process industries. View project

All content following this page was uploaded by Tahir Raza on 29 November 2016.

The user has requested enhancement of the downloaded file.

Proceedings of the 2016 International Conference on Industrial Engineering and Operations Management
Kuala Lumpur, Malaysia, March 8-10, 2016

Prediction Of Failures By Using Crow/AMSAA Model In

Microsoft Excel
Masdi B Muhammad, Tahir Raza, M Amin Abd Majid
Department of Mechanical Engineering
Universiti Teknologi PETRONAS, 32610 Bandar Seri Iskandar, Perak Darul Ridzuan, Malaysia

Abstract— Failures due to poor maintenance practices are perceived as the dominant factor that limits production capacity,
product quality and profitability. To overcome these failures, companies are using different techniques like inspection, using key
performance indicators (KPIs), analysing the vibration of equipment, eliminating the design errors, operation errors and
installation errors, in order to reduce the failure rate and increase production and profit. Most complex machines like automobiles,
radars, aircrafts, and main servers are repaired not replaced during failure. This failure could result to high maintenance cost and
a financial loss to the company. Therefore, it is necessary to predict and analyze the reliability of repairable equipment in order to
reduce the occurrence of failure and increase the availability and reliability of the equipment. The Crow- AMSAA plots are
powerful tool for predicting the upcoming failures. In this paper Non Homogenous poisson process (NHPP) or Crow/AMSAA is
used to predict the upcoming failures by using Microsoft excel. This could help industry to increase their mean time between
failures (MTBF) in order to improve reliability. A case study of centrifugal pumps is taken in this paper to clarify the functioning
of Crow- AMSAA (C-A) plots using Microsoft Excel. The results show that next failure of centrifugal pumps is predicted
successfully.

Keywords— Reliability, Crow-AMSAA, Non Homogenous Poisson Process.

I. INTRODUCTION
Equipment, components, systems, machineries and human beings are not perfect and thus free from failures. Maintenance
helps these equipment to keep, sustain, hold, preserve, or carry on these kind of physical assets. Maintenance has become a
support tool for organization and plays a significant role in achieving organizational strategies and goals [1]. Now a days the
primary objective of any industry or plant is to assure that their equipment are running at high efficiency, at low cost while to
satisfying plant operational needs. Reliability growth plots or Crow-AMSAA which is also called as non homogenous
poisson process (NHPP) is very much useful tool to forecast when system failure will occur for repairable system. A
repairable system is defined as “a system which, after failing to perform one or more of its functions satisfactorily, can be
restored to fully satisfactory performance by any method other than replacement of the entire system”[2]. In C-A process
trend line slope, β, tells that by which frequency failures are occurring. Either they are coming slowly, faster or without any
change. Long term trend line slope β also tells that if organization is reliability focused its trend line must be β<1 or if it is
repair focused then β ≥1[3]. Chemicals and petroleum plants specially focus on maintenance and try to forecast the failure
before it occurs; these industries try their best that their machinery would be always in operation. C-A plots is one of the best
tool which can help these plants to predict their failures before occurring. To overcome the problems like high cost,
equipment maintenance, maintenance management, there are many ways to solve these problems. The most efficient way to
improve performance is to have an effective maintenance activity that will aid in the process of reducing cost, improving
productivity, and predicting failures [4]. To guarantee that plant accomplishes the sought execution, maintenance managers
require a decent track of performance on maintenance process and maintenance results. This can be accomplished through
improvement and execution of a thoroughly characterized maintenance procedure.
II. RELIABILITY
Reliability is one of the main considerations influencing the performance of a production system. Not just defers because
of the machine breakdowns affect the production rate, however the aggravations brought about by these breakdowns lead to
scheduling issues which diminish the productivity of the whole manufacturing operations. Because of the changing way of
manufacturing system, the unfavorable impacts of machine breakdowns are felt distinctively in diverse manufacturing
circumstances. While, profoundly computerized, large scale manufacturing system is most sensitive to the reliability changes,
job shop production systems are more adaptable in managing machine failures [5-7].
According to the USA Military Standard (M1L-STD- 721B). “Reliability is the probability that an item will perform its
intended function for a specified interval under stated conditions”[8]. This definition indicates that reliability is not simple as it
is defined in previous times. The three most important components of this definition are required function, stated condition,
and defined period. It is important to understand these components in order to design a product that can be validated as
reliable. Reliability is also affected by the environmental conditions of the plant. For example, stress to the equipment due to

© IEOM Society International 180

Proceedings of the 2016 International Conference on Industrial Engineering and Operations Management
Kuala Lumpur, Malaysia, March 8-10, 2016

vibration could also decrease the reliability. So these conditions should also be taken in account while calculating reliability
otherwise calculations does not have real meaning. Many plants don’t have a good reliability function and they do not have any
special program that directly defines the problems regarding reliability. Some sections like quality and maintenance
management program clarified that equipment reliability is an issue. But still these sections do not have specific programs that
could improve reliability and in the end industry faces different challenges regarding production, maintenance, scheduling, etc.
To overcome these kinds of problems it is very much necessary that every industry should identify some key goals regarding
their production plan, maintenance plan. That could help an industry to overcome these kinds of problems and also helps to
find out their current state. Measuring equipment reliability is very much important if you want to improve it. Reliability is
mostly measured by the mean times between failures (MTBF) mistakenly. MTBF is only for repairable equipment. As most
equipment failure events are not random. Most happen soon after equipment is maintained. There is a big drawback of only
mean time, because it is not showing the value of that level of reliability. By doing this industries may spend lots of money on
small improvements. Reliability must also measure by money made or lost for the business [9].
III. IMPORTANCE OF RELIABILITY
Highest reliability of equipment is need of every industry; either it is the primary industry, secondary industry, tertiary
industry or quaternary industry. Highest equipment reliability can be accessed by ensuring that failure chances of equipment
are relatively low. Every industry wants to run their equipment without any big stoppage or serious stoppage. For those
industries doesn’t need any kind of magic, industries just have to keep their equipment in low stress condition, provide
equipment a good environment, do their maintenance according to schedule, that also help to increase the reliability of any
machine. If every industry could provide machine their required conditions then nothing serious will happen and equipment
will run at its full capacity. Time is an important aspect when industries measure the reliability of any equipment. Let’s take
an example equipment is continuous running to complete any operation for 5000 hours and did not fail for a single time then
it is called as reliable equipment. It means that when equipment operates as long as the industry expected, it is considered as
reliable. When the periods between failures is too short that it is unreliable. So we measure the reliability of any equipment
by its trouble-free time. High reliability is sign of good organizations they expect equipment to perform for a long time, if it
does not then they take effective measures to learn and improve. In the industries where there are lot of problems regarding
equipment reliability and workers are busy all the time often repairing equipment again and again and it never ends because
at the end of day workers know there will be more troubles tomorrow. And apart from this when any industry has more
reliable equipment because it performs as its design intended without failures. That business surely makes good profit with
low maintenance cost and operating cost. So a place with reliable equipment is safer place.
IV. CROW –AMSAA (C-A)/ NON HOMOGENOUS POISSON PROCESS (NHPP)
Non homogenous poison process model does not require stationary increments, which means that failures may be more
likely to occur at certain times than other. In easy language NHPP is a poisson process with simple parametric model used to
represent events with a failures recurrence which are not constant [10]. NHPP describes the cumulative number of failures up
to time t, N(t) and it follows a poisson distribution with parameter λ(t) for a counting process [11]. NHPP model works when
the occurrence rate is time dependent and no more requirement of stationary increment.
C-A plots are created by getting the data from computerized maintenance management system (CMMS). This data is used
in two ways 1) To create cumulative time v/s cumulative failures plots for crow- AMSAA and 2) age- to- failure for Weibull
plots. Data acquisition is simpler for C-A plots than for Weibull plots [11]. For C-A plots, the x-axis is chronological time
which is also known as cumulative number and y-axis is known as cumulative events or cumulative failures, they make
reliability visible. This data, when plotted it, gives results usually in straight lines. This plotted data provides information about
two statistics (line slope β, and y axis intercept at time t=1, λ). Trend line slope β (x-axis) is a very strong indicator which
provide us information regarding increasing, decreasing or no improvement. The y-axis provides the failure rate at time equal
to 1, which is some kind of hypothetical value by that it is easy to forecast future failure. Y-axis is converted to plot cumulative
mean time versus cumulative time which is easy to understand. When the line slope is upward and to the right, reliability is
improving; same as when it is going downward and to the right, reliability is decreasing [12].

© IEOM Society International 181

Proceedings of the 2016 International Confference on Industrial Engineering and Operations Manag
gement
Kuala Lumpur, Malaysia, March 8-10, 20166

Occurrence Total cumulative

Cumulative time
Figure 1. A simple log- log graph [12]

See Fig. 1, for an example an easy and simpple plot of cumulative time versus cumulative failures is madem by [12] by using a
software. In this figure, the first failure occuurs at the same time the value of β>1, which means failures are increasing. Second
trend line shows that β=1, which means thee system does not experiencing any improvement. Third slope s line is showing β<1
shows system improvement and time betweeen failures is much long. The straight trend line also offfers a tactic for making a
forecast of coming failures even if the data ccontain mixed failure modes.
V. NHPP MODELING USING EXCEL
Crow/AMSAA plots are drawn by usiing different kind of commercial software like the WIINsmith visual software,
Weibull, Blocksim, Fulton 2006. These sofftware are very costly and very difficult to use, this softwware needs lots of time to
understand. Beside this Microsoft Excel is very handy software and very common software. Excel is used by almost every
industry for different purpose. So in this ppaper, excel based reliability model which is based on crrow- AMSAA method is
developed. A template which is very simpple and easy to understand, by which it is very easy to o forecast the upcoming
failures. In this template, user has to only kkey in cumulative failure time and then to check out whatt is the β. If the β<1 then
the system is improving and failures are deccreasing, and if β>1 then failures are increasing and systeem is deteriorating and if
β=1 then there is no change in system. Somee common formulas are used to calculate the λ and β.
:
λ ^
(1)
:
β (2
2)
:
The most current history is representedd by these two statistics λ and β which will be used to make forecasts of future
failures for the system using another equatioon N(t)= λ*tβ. Solving this equation for cumulative time gives
g t= [N (t)/λ] (1/β)….
(3)

© IEOM Society International 182

Figure 2: Data input section

Proceedings of the 2016 International Confference on Industrial Engineering and Operations Manag
gement
Kuala Lumpur, Malaysia, March 8-10, 20166

Figure 2 is the first part of template, in this part of template user should input the cumulative failu ure time (observed failure
data) in the yellow shaded cells, that will callculate all the other parameters like time between failure, the
t cumulative mean time
between failures, log of the cumulative time of failure by using simple arithmetic formula.

Figure 3: Results section

Figure 3 is the result section of the templatte, in this when end of observation time will be entered, and then λ and β will be
calculated according the formula (1) and (22). Next failure occurrence time will also be calculated according to the formula
given in (3). This template will also calculaate the instantaneous failure at the required time like in fig
gure 3 user want to check
the instantaneous failure rate at time 730, aafter calculation the result is 0.0416. Same as cumulativ ve failure and cumulative
failure rate at any time can also be calculatedd by just inserting any value in the yellow shaded area.

Figure 4: Plot section

© IEOM Society International 183

Proceedings of the 2016 International Confference on Industrial Engineering and Operations Manag
gement
Kuala Lumpur, Malaysia, March 8-10, 20166

Figure 4 is graph area, where graph will be pplotted after inserting the data. There are two elements willl be visible on graph, one
is line of prediction and dots which represennt failure events or observed data. If the prediction line or trend line is touching the
dots or it is passing very nearby (which meaans there is very minor difference in prediction of failure) that
t means the calculation
of prediction of next failure will be accurate and the process is non homogeneous Poisson process.

Figure 5: Predicted Data section

Above figure 5 is the section where failuress are predicted and validated with the observed data. By inserting
i the starting time
and increment time, prediction line on grraph will be plotted and this part will also calculate th he instantaneous MTBF,
cumulative MTBF and predicted number of failures also.
VI. RESULTS AND DISCUSSION
Table 1 shows the failure data of centriffugal pumps. This data is collected from June 2000 until September
S 2008 which is
approximately 3000 days of operation [10]. Actually, this data includes important information so th his data has been masked
due to proprietary reasons. So only failure ddata have been taken as this template only required failurre data. This data will be
use by inserting these failure times in the daata input section as shown in figure 2.
Table 1: Time to failure data for pumps in Days
25 138 265 470 656 1078 1896 2461 2736
38 158 356 473 661 1182 1922 2510 277
76
46 167 360 488 697 1294 1925 2562 2882
55 171 360 493 747 1336 1939 2568 2930
58 171 364 509 811 1383 2008 2583 2939
59 186 368 527 830 1462 2033 2648 2952
60 187 419 527 921 1536 2051 2698
66 191 427 568 963 1550 2097 2707
68 206 448 571 964 1550 2140 2718
112 231 452 600 985 1615 2195 2720

© IEOM Society International 184

Proceedings of the 2016 International Confference on Industrial Engineering and Operations Manag
gement
Kuala Lumpur, Malaysia, March 8-10, 20166

Figgure 6: Crow/AMSAA plot for failure data of pumps

Figure 6 is drawn using NHPP template in M Microsoft excel by using the failure data of pumps. This plot
p is drawn by using the
half of failure data from table 1 i.e. 25 too 985. Because by other remaining data it would be validated
v that template is
calculating the expected failures correctly oor not. According to the calculation the next failure (51sst failure) is expected on
1025 and the same thing graph is showing, and the actual failure happens on 1078. Which means thee error is minor. Whereas
β<1 in calculation, which ensures that the syystem is improving. And by looking closely at trend line in figure 6 that is almost
very close to the data points or touching thee data points, that means the prediction of next failures will
w be much closed to the
original failure data.
VII. CONCLUSION
Reliability is very much important parrameter for any industry to reduce the downtime and diffferent failures. Reliability
can be improved by predicting the failurees at the right time and then conducting maintenance work according to that
predicted time. Which results in increment oof MTBF. As MTBF of any system increases that results in high reliability of that
system. In this paper, reliability, its impoortance and Crow /AMSAA (NHPP) has been discusseed in detail and an easy
template has also been presented and deveeloped in Microsoft excel beside the many commercial software. This template
could be used by any industry to prevent froom unexpected failures, which will help to improve the reeliability and increase the
MTBF. A real time failure data of a centrifuugal pump as a case study has been utilized in this paper. The results showed good
accuracy, using the developed template as iillustrated in Figure 6. In order to improve the reliability of any machine, user can
calculate the future failures very easily by uusing this template, which will surely result in the reliability improvement.

© IEOM Society International 185

 Proceedings of the 2016 International Conference on Industrial Engineering and Operations Management
Kuala Lumpur, Malaysia, March 8-10, 2016

REFERENCES
[1] T. Ablay, "Maintenance Perception in Swedish SMEs (A local study in Kronoberg County)," 2013.
[2] H. Ascher and H. Feingold, Repairable systems reliability: modeling, inference, misconceptions and their causes: M.
Dekker New York, 1984.
[3] H. P. Barringer, "Use Crow-AMSAA reliability growth plots to forecast future system failures," 2006.
[4] H. A. Samat, S. Kamaruddin, and I. A. Azid, "Maintenance performance measurement: A review," Pertanika Journal
of Science and Technology, vol. 19, pp. 199-211, 2011.
[5] B. B. FLYNN, "Critical machines preventive maintenance policies for group technology shops," THE
INTERNATIONAL JOURNAL OF PRODUCTION RESEARCH, vol. 27, pp. 2009-2020, 1989.
[6] C. N. Madu and C.-H. Kuei, "Simulation metamodels of system availability and optimum spare and repair units,"
IIE transactions, vol. 24, pp. 99-104, 1992.
[7] K. Miriyala and N. Viswanadham, "Reliability analysis of flexible manufacturing systems," International Journal of
Flexible Manufacturing Systems, vol. 2, pp. 145-162, 1989.
[8] R. F. Stapelberg, "Reliability and Performance in Engineering Design," Handbook of Reliability, Availability,
Maintainability and Safety in Engineering Design, pp. 43-294, 2009.
[9] M. Sondalini and H. Witt, "What is equipment reliability and how do you get it?," 2000.
[10] M. Muhammad, M. A. A. Majid, and N. A. Ibrahim, "A case study of reliability assessment for centrifugal pumps in
a petrochemical plant," in Engineering Asset Lifecycle Management, ed: Springer, 2010, pp. 398-404.
[11] P. Barringer, "Predict failures: Crow-amsaa 101 and weibull 101," Proceedings of IMEC, 2004.
[12] P. Barringer, "Problem of the Month, Nov 2002–Crow/AMSAA Reliability Growth Plots," ed, 2003.

BIOGRAPHY

Dr Masdi bin Muhammad, CEng (UK), CMRP, ASQ-Certified Reliability Engineer is a Senior Lecturer in Mechanical Engineering
Department and research cluster leader for Facility and Plant Engineering, UTP. He obtained his first and Master degrees, BSc in
Mechanical Engineering and MSc in Manufacturing System Engineering, from Lehigh University, USA. His PhD in Mechanical
Engineering was obtained from UTP with research on Reliability Model for Repairable Systems with Multi-State Degradation. He is
actively involved in consultation work and training on asset reliability and integrity for oil and gas companies.

© IEOM Society International 186

View publication stats