Professional Documents
Culture Documents
Why Bad Things Happen To Good Steam Equipment
Why Bad Things Happen To Good Steam Equipment
reprinted from
March 2015
www.chemengonline.com
Improving Rotating
Machinery
Operations
Interphex Show
Preview
Preventing Facts at Your
Combustible Fingertips: Steel
Dust Explosions Corrosion
Carbon Capture Process Control:
Technologies Integrating
Targeting Methane Instruments and
Emissions Systems
Managing
Steam Systems
Reprinted with permission from CHE (Chemical Engineering), March 2015.
Copyright © 2015 Access Intelligence.
IN BRIEF
WHAT CAN GO WRONG
H have provided the most
effective source of read-
ily conveyable heat to
industrial process applications, in-
cluding those in the chemical pro-
IN A STEAM SYSTEM? cess industries (CPI), and there is
no similar low-cost substitute that
ANALYZING A STEAM-
can replace steam. Without steam,
TRAP POPULATION
industrial production would be dra-
MANAGEMENT matically curtailed, and the low-cost
PRIORITIZATION manufactured products that are
EVALUATING THE made from steam’s heat or power-
NUMBERS generation assistance would not
exist. Without steam, our quality of
SYSTEM GOALS life, economies and society in gen-
CREATING A NEW eral would suffer.
PARADIGM While many CPI workers may ap-
preciate that steam systems are a
necessity, the same individuals may
have also experienced negative
steam-related incidents throughout
their careers, making them harbor
unfavorable thoughts. Specifically,
these events may have resulted in
safety issues, equipment failures
and unscheduled shutdowns of a
unit or a full production line. Safety
events are extremely challenging
and sorrowful issues if someone FIGURE 1. Effective steam traps keep heat in the system to optimize
production rates and heat quality, and they also discharge condensate to
is injured, and shutdowns can be provide for system safety and reliability. Major problems can occur if con-
disruptive to the entire workforce. densate is not readily drained from the system
It is not surprising, then, that peo-
ple lack an enthusiastic attitude when it ment. Such destruction may render critical
comes to steam. process equipment, such as turbines, flares
or heat exchangers, unusable. Additionally,
What can go wrong in a steam system? high return-system backpressure caused by
What types of issues can arise in a steam steam leakage or blowthrough from bypass
system, and can listing and classifying these steam might restrict production quantity or
items help to determine an executable pre- quality through heat-exchange equipment.
vention path or risk-mitigation procedure? Of Backpressure can also cause the heat-ex-
course, it is relatively easy to identify the most change application to be put on bypass or
common maladies seen in a steam system. to waste condensate by routing to ground
There can be waterhammer, erosion damage level. While some steam-system failures are
and steam leaks in utility systems or equip- common, the challenge then is to identify the
sources of these incidents. This leads to the commissioned, because no part of the plant
larger question, “Why do bad things happen had been shut down.
to good steam equipment?” When the in-service quantity of CDLs is
The technical answer lies in understanding lower than the design calls for, then it may
that the cause for a large percentage of failures be that the reduction was due to a misun-
might be due to the steam system not being derstanding of the importance in maintain-
maintained to the “as-built” or “design” speci- ing the design total, or possibly from general
fication. The original designers analyzed the neglect of the trap population. If the discrep-
plant requirements and determined the most ancy is not explainable, then the system’s
suitable design, according to their expertise drainage is restricted from the originally re-
and standards. That original design included quired capability. Here, the target for total in-
the correct number of steam traps and the as- service CDLs should be increased to equal
semblage of piping and components that help the original design total of 11,000 CDLs.
to drain condensate from the system. Com- When routinely administered, steam-
monly referred to as a condensate discharge trap surveys conducted by plant personnel FIGURE 3. Accurate and
regularly sustained diagnosis
location (CDL), the total assembly is required (Figure 3) provide invaluable information for of steam traps’ operating
to remove condensate and effectively maintain evaluating the health of the steam system. conditions is essential to
the design performance of the steam system. As it turns out, in this hypothetical example, determining the population’s
Once built, the plant is handed over to the end several years had passed since the last trap current health (state of the
population). Once failures
user for operation and maintenance — with survey was completed. So, when the current are identified, the informa-
the expectation of sustaining the initial design. survey was finished, it was found that there tion is valuable for allocating
Unfortunately, that is when operational bud- were 3,000 trap failures (both hot leaking resources to restore all con-
densate discharge locations
gets, personnel turnover and, in some cases, failures and cold-blocked low-temperature to an “as designed” operating
inexperience with steam systems may play a failures occurred); and 7,000 traps were condition for safety and reli-
causal role in negative events. considered to be in good condition. This sit- able performance
Example: 10,000 traps / 1,000 traps failed in a year = Example: 3,000 failures / 10,000 traps = 30% state
10 year trap life expectancy of failure
7,000 good traps / 10,000 traps = 70% good state
uation creates a significant operational and to grow until catastrophic events occur.
maintenance dilemma for the site. One takeaway from this example is that for
If the goal is to have a “zero reset” sce- every 1,000 traps that must be repaired in
nario, in which all failed traps must be re- a year, there is a requirement for the plan-
paired, and the cost to repair the average ning and repair resources to correct 84 traps
failure is $600, then $1.8 million is needed monthly without fail. That replacement re-
from the year’s budget to accomplish the quirement equates to repairing 4–5 CDLs
target. To achieve this goal, 250 steam traps per day, with any lesser amount creating a
must be replaced each month, which repre- gap at year-end; the difference here equaling
sents two or three maintenance crews work- the next year’s carryover failures.
ing full time for a year, assuming that every Steam traps must be replenished or re-
trap is accessible and can be isolated for paired in order to maintain a sustainable
repair or replacement. It is certainly a monu- operation. Not taking the highest-priority
mental task that may have been caused by corrective action with regard to steam trap
an improper course of action a number of failures is somewhat comparable to reusing
years before, when trap replacements were a teabag even though it no longer dispenses
not completed in sufficient quantity to keep flavor — although there is a teabag in ser-
pace with the annual failure rate (amount of vice, it no longer provides a useful purpose.
failures per year) of the population. Similarly, if a steam trap has failed — par-
Consider that the average annual failure ticularly via a cold failure — then the CDL is
rate of a steam trap population in a mature no longer serving its intended purpose, and
plant can be estimated from historical re- must be repaired.
cords, provided that there are at least two Is there any question about what outcome
survey events within a period of 4–5 years. should be expected if all of the steam traps
Simply subtract the carryover failures (failed and related CDLs in a system were simply iso-
traps recorded from a prior survey for which lated by valving, thereby completely remov-
no action was taken to repair) from the re- ing their drainage capability? There would be
ported failures that were recorded in the no way to automatically remove condensate
survey. The remainder is the quantity of new from the system, creating a highly dangerous
failures. The number of new failures can be situation. What about if just 30% were shut
divided by the number of years since the prior off, or 50%? It is akin to gambling with the
survey to provide an estimate of the average safety of the plant, as the potential forma-
annual new failures of a steam trap popula- tion of condensate slugs in the pipeline can
tion. Several related useful calculations are lead to unstable, hazardous conditions. It is
shown in the box above. extremely distressing to consider a site that
In this hypothetical plant history, if the is not replacing trap failures due to a bud-
management of the trap population had get constraint, because the timely repair of
been continuous and sustainable at an an- traps allows the system to operate at the
nual failure rate of 1,000 (10%), then the intended conditions. This must be seen as
plant should only have to support the repair an inflexible demand. Timely action should
of 84 traps per month, not 250 traps. How- be mandatory, not optional, to optimize the
ever, in this case, the trap population was al- operation of a steam system.
lowed to deteriorate to 3,000 failures (30%
failure state), and this situation places a tre- Management prioritization
mendous burden on resources. Worse still, if In a properly drained and maintained steam
not corrected, the failures can be expected system, it is critical that the steam flowing
10,000
400
5,000
200
0 0
0 1 2 3 4 5 6 7 8 9 10 11 12
Month
1,500
1,000
Carryover
500 failures
Target monthly replacements
Actual replacements
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Month
the hot failures, because these repairs can a plant with an average steam-trap lifecycle
be readily justified by simple energy-cost of ten years, the actual failures could be es-
analysis. Figure 4 illustrates a typical close timated as 2,000, consisting of the 1,000
relationship between hot-failure traps and re- carryover failures just identified and 1,000
duced steam loss. Especially in times of very new failures. While the diagnosis of each
high energy prices, incredible emphasis is trap in the population is performed at regular
placed on reducing the cost of operations by intervals (usually annually or semi-annually),
fixing leaking steam traps. A progress chart, new failures are constantly occurring, as il-
similar to the one shown in Figure 4, can be lustrated in Figure 6.
generated and tracked. Thus, if the project If carryover failures are included as part of
plan is to eliminate 3,000 failures from the the repair strategy, a significant number of
hypothetical plant, then the corresponding CDLs will be operating improperly, thereby
requirement is to repair 250 CDLs per month, increasing the chances of a debilitating in-
otherwise there will be a carryover failures cident occurring within the plant. For this
gap (see Figure 5 for additional analysis). reason, it is not recommended to adopt a
The carryover failures represent the real work process that allows carryover failures
world; rarely do plants correct all or even and focuses only on those traps that have
nearly all of the failures. The result is that already been fixed. Instead, a paradigm shift
a significant number of CDLs are not re- is required.
stored to proper drainage operation, and it
is not uncommon to carry a sizable number System goals
of failures over to the following year. Once Applying maintenance action to individual
carryover failures are accepted in a plant’s steam traps is a path action, but not an
operations, the steam system becomes overarching system goal. Risk lies in the
destined for sub-optimal and potentially false sense of security that is given when
unreliable operation. the goal is simply to repair a given number
What follows when 1,000 carryover failures of steam traps, and the real goal of achiev-
are extended into the following year? Instead ing an optimized steam system is neglected.
of correcting all 3,000 failed steam traps (a So, if the annual target is established to fix
zero reset mentality), suppose that the plant 2,000 steam traps and only 1,000 traps are
management allocated a budget for repair repaired, there might be some explanation of
of only 2,000 failures instead. Perhaps this mitigating events that explains the lapse. In
thought process stems from the expecta- such a situation, the carryover failures might
tion that with 2,000 failures corrected, the be considered acceptable under the circum-
next annual period will only require budget stances, and a new goal is assigned for the
for 1,000 failures. However, that is not an next year. However, this is a dangerous situ-
accurate scenario, because those 1,000 ation, because while the potential for dam-
CDLs represent carryover failures only, and age is not visible, it is prevalent in the unad-
the plant also must consider new failures. In dressed sections of the steam system, and
Actual
failures occuring during the
6,000 repair period always lower the
actual year-end state. It is just
one of many justifications for
4,000 adopting a zero reset
mentality for maintenance-
response planning
2,000
New failures
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Month
the downstream recipients of that steam. recognized until the next survey.
The potential for peril can increase in sever- The result is that the best state deterio-
ity over time. rates between the prior and new surveys,
Without a full understanding of the long- with the theoretical “best sustainable” con-
term impact on a steam system, there can dition being realized at the beginning date
exist a false impression that a system can of the next survey. This theoretical point
be well managed, even if there is allowance occurs midway between the prior survey
for carryover failures. Figure 7 provides ad- and the next survey date — if annual sur-
ditional insight into a longer-term view. As it veys are conducted, and repair is immediate
turns out, the best possible theoretical state and linear, then the best good state occurs
of the population occurs only after a zero at midyear.
reset condition is experienced from the prior If the original goal of the designers is
survey report, and when accumulated re- recalled, it was to have 11,000 fully func-
pairs equal new accumulated failures. After tional CDLs at the site. For whatever rea-
the midway point, zero reset has been theo- sons, the plant’s management decided to
retically reached, and new replacements decommission and remove 1,000 installa-
cease, as there are no known failures that tions, leaving a population of only 10,000
remain to be repaired. However, unidentified in-service CDLs (a 9% reduction from the
new failures are still occurring and will not be design). Even if there is perfect harmony
12,000 FIGURE 7. Sites sometimes
Design CDLs decide upon a fixed number
of annual repairs instead
of striving for zero reset,
In-service CDLs thinking they can achieve a
10,000
* * nearly 100% good in-service
CDL population state over
Best
several years. However, due
Target * Best
sustainable to new failures during the
8,000 repair cycle, the theoretical
Actual
“best sustainable” condition
is reached midway through a
Number of CDLs
4,000
Accumulated repairs
2,000
New accumulated failures
0
0 6 12 18 24 30 36
Month
Warning
Danger
Good
Good
Cold failures
Cold failures
failures
Cold
Good
Hot failures
failures
failures
Hot
Hot
1 2 3 1 2 3
Year Year
Failure-based method Performance-based method