Hinton 1998
Hinton 1998
Hinton 1998
a reliability value. A problem is that the use of a single production casing designed for a maximum estimated
design factor may be overly conservative for some well surface pressure of say 10,000 psi is multiplied by a
designs and impossible to implement for more critical factor of say 1.2 to give a design rating of 12,000 psi.
wells. Risk Based Design places more responsibility on This factor could be used out of context in the well plan
the engineer to re-consider uncertainties, allowing to say that the casing should be tested to only 8,333 psi
different risk levels to be taken for different wells. This (10,000/1.2) rather than 10,000 psi. A pressure test to
approach can be considered to be more holistic and the maximum expected wellhead pressure is important
requires consideration to be given to design parameters, because this checks the complete system "the pressure
material specifications and quality systems as well as boundary" of the well as it is described in guidance to
the competence and experience of the engineer. the Offshore Installations and Wells (Design and
Construction, etc) Regulations 1992 (Ref.3). Insufficient
Will Risk Based Design systems be used without torque or poor make-up procedures on the tubulars may
proper understanding? negate the most rigorous design. It is imperative that this
There is a danger that any design system which relies is discovered with a pressure test rather than during a
on sophisticated models and computer programmes will well incident when hydrocarbons may be released.
be (mis)used by people who do not understand it. It is The Well Examination requirements of Design and
vital that the user understands the principles of the Construction Regulations (Ref.3) provide a check for
system and what assumptions are being made. The errors in casing design. Regulation 18 requires all
concept of all variables, both on the load and resistance companies to have the design of their wells examined by
sides, being characterized by curves and their overlap an "Independent and Competent person" to ensure, so
describing the resultant risk is quite difficult to grasp. far as is reasonably practicable, there can be no
Calibration procedures involving Monte Carlo unplanned release of fluids from the well. Some well
simulations are even harder to understand. It is examiners use a different method when checking the
important that the system is made comprehensible to the casing design. That is, Working Stress Design may be
drilling personnel designing and supervising the well used when the designer has used a Risk Based system.
construction. The job must be fitted to the person not Any conflict between the two methods would have to be
fitting a person, say an expert in statistics, to the job of justified by the designer. This justification would include
designing casing. At some level there ought to be how the loads were estimated, what assumptions were
someone who understands both the rocket science and made and what risk level had been chosen.
makin' hole. If this is impractical there should at least be
good understanding between the different experts. Can loads on casing be predicted as accurately as
On a very simple level there is a danger that design resistance?
pressure factors could be used as "negative safety It is reasonable for the non statistician to accept that the
factors". The executive summary of the DEA(E)-64 strengths of joints of casing of the same weight and
project on "The use of quantitative risk analysis (QRA) in grade from the same mill will vary symmetrically around
casing/tubing design" (Ref.2) contains a table (6.1 in a mean value. The product is manufactured from
Executive Summary) listing design pressures compared nominally the same materials, by the same process with
to "casing full of gas Pressures" for various classes of the aim of producing identical properties. The
wells. For wells which are not classed as High Pressure/ predictability of the "resistance" side of the equation has
High Temperature (HP/HT), factors of 0.9 and 0.8 are been confirmed by large scale examination and testing
given. This table is recommended for inclusion in of the finished product.
company casing manuals. There is a real danger that The "Load" side of the equation, such as formation
well designers could reduce their estimated loads by a pressures and kick volumes may not be so predictable.
factor of 0.9 or 0.8 without using the specific design There is also a much smaller data bank available for
equations given in the body of the report (which a note drawing up the graphs. Human factors may influence
below the table says they should do) and without the size of a kick by such things as speed of reaction in
knowing how these factors were derived. closing the well in and choosing the correct choke
A more subtle example would be using factors to pressures when killing a kick.. The DEA(E)-64 research
justify pressure testing casing to less than the predicted (Ref.2) indicated there were two families of kicks. Most
maximum wellhead pressure. In a risk based casing have an influx volume of 10's of barrels whatever the
design, the partial factor for the pressure test on kick intensity (difference between mud hydrostatic and
production casing must be at least one because these formation pressure). There is a smaller, but significant,
tests are always carried out. The actual loads are likely number of "escalated" kicks with gains of 200, 300 or
to be higher than planned because of uncertainties of even 400 barrels. Because of the comparative lack of
back-up forces (external pressures) and because pump load data, compared to resistance data, it would be
operators tend to overshoot pressures. Therefore a
SPE 48326 WILL RISK BASED CASING DESIGN MEAN SAFER WELLS? 3
better if companies could share information on loads Therefore it is likely that every 9-5/8" casing string run in
such as kicks. the UK Continental Shelf will contain at least one weak
Human error can lead to a casing full of gas even if the joint. With 364 wells spudded last year at least one is
initial influx is quite small. The most serious example in likely to have the sub-standard joint at surface, the point
the UK North Sea was the Ocean Odyssey blowout. of greatest burst load. If Risk Based Design is used to
Circulation off bottom with insufficient choke pressure justify lower strength casing, this must be balanced by
resulted in the mud in the casing being replaced by gas improved quality control of the casing
with a resultant flowing pressure at the choke of 8,000 Well Designers must take responsibility for ensuring
psi. This was before any mechanical failures occurred. the materials they use in the well are fit for purpose as
This was almost ten years ago but human error can still required by regulations 16 & 13 of the Design and
fill the casing with gas. Last year a kick was shut in Construction Regulations (Ref.3). However greater
promptly with a gain of only 10 bbls and a shut in casing responsibility, and some economic penalties, may be
pressure of 930 psi. After weighting up and circulating borne by the manufacturer who is best qualified to
kill weight mud and taking losses the shut in casing assure the quality of his product. Even so there will be
pressure rose to 3,500 psi. This is equivalent to 240 bbls some substandard material produced. Results from the
mud being displaced from the casing by gas. Would this DEA(E)-64 project (Ref.2) indicated that when output
incident have been recorded as a kick with only an initial increased, quality reduced in that more material was
influx of 10 bbl rather than the maximum estimated influx below specification. It is vital that during this time of high
of 240 bbls? If so would the statistics on loads activity to ensure that quality standards are maintained.
underestimate the potential problem? It is also important to be able to know when a Risk
In the UKCS, kicks in "standard" wells have become Based Design system requires recalibrating. This
less common and so rig personnel have less experience decision requires someone with knowledge of changes
of killing "standard" kicks. In situations where kicks are within the industry and the statistical significance of any
more likely to occur (HPHT wells and over-pressured changes.
dolomite rafts within massive salts) there tend to be
complications such as massive lost circulation or a Will Risk Based Designs always recommend lighter
loss/gain situation. These may be more likely to lead to or lower grade tubulars?
"escalated" kicks that can result in very high pressures The use of Risk Based Design will result in lighter or
at the wellhead. The situation is not helped by the lack of lower grade tubulars being used than with Working
experienced engineering and supervisory personnel Stress Design. This can be justified by demonstrating
caused by the current very high rig utilization rate. The that this design will be below a given risk of failure and
new Design and Construction Regulations (Ref.2) that risks to persons are as low as reasonably
require duty holders to ensure that drilling personnel and practicable. In some cases it may be an advantage to
supervisors have received appropriate information, have numerical values of risk so that alternatives such
instruction and training. This includes Well Control. This as changes in either weight or grade, can be compared
is in contrast to the previous prescriptive legislation and the lowest risk option chosen.
(Offshore Installations (Well Control ) Regulations 1980) However using lighter casing will mean the design is
which said that all supervisory personnel must have a less forgiving, there is less margin for error. Tubulars
valid well control certificate. The new regulations place may be damaged during transport or installation,
the emphasis on competence for the job in hand and reducing their resistance. The extra wall thickness of
gets away from being trained to pass a test. previous designs might have compensated for damage
and no failures would have occurred and been recorded.
Can Risk Based Design simply replace Working The same damage to thinner wall pipe might cause a
Stress design? failure which would not have been predicted. Casing
The casing design system cannot stand alone, it must be wear is another potential problem area. There is
backed up by only using pipe that is of the standard that obviously less steel to wear away before there is a
was used to calibrate the resistance factors. Generally danger of burst. Are casing wear logs and predictive
pipe is better than the nominal API values , such as yield programmes accurate enough, and used enough, to
strength. However the casing string is only as strong as compensate for the thinner walls? This may become
its weakest link. For example, large numbers of tests on more significant with a move back towards water based
L80 material show the strengths range to over 100 ksi muds for environmental reasons.
with the average closer to 90 rather than 80. Regulators Risk Based Designs will be hardest to justify when
are more interested in the other extreme where 0.6 % of used for extreme conditions such as in high temperature
the joints are sub strength ranging down to 72 ksi. In and pressure wells. These are the wells where
drillers terms this means that, on average, every 7,000' significant savings could be made by using a Risk
string of casing will contain a sub-standard joint. Based Design. The concerns would be that kicks are
4 A. HINTON SPE 48326
much more common in HPHT wells; gas solubility / If the load data has been underestimated, the wells will
expansion effects are more significant; the mechanisms not be safer, especially in High Temperature / High
of gains and losses (where the window between the Pressure Wells.
pore and fracture pressure is small) are not fully
understood; and non-traditional techniques may be used A Risk Based Design system with :-
to kill the well. On the last point, there is a growing belief more accurate failure equations;
that a lot of the serious problems with killing HPHT wells account taken of brittle fracture in low levels of H2S;
have been caused by shutting wells in and causing a improved Quality Control of tubulars and connections;
"supercharging" or "ballooning" effect where mud seems accurate load data;
to be lost to the formation and then fluid is returned engineers who understand the system and the well;
either mud, formation fluids or a mixture. Therefore a full training and competence assurance programme
shutting the well in may not be the immediate response may produce wells which are as safe as those designed
to suspected flow. There is also concern that trapped using Working Stress Design.
pressure and thermal expansion tend to exaggerate
shut-in pressures and so pressures may be bled off to References
check that the surface pressures reflect downhole 1. HSE: A guide to the Offshore Installations (Safety Case)
pressures. These techniques are justifiable responses to Regulations 1992 ISBN 0 11 882055 9
the situations but may lead to larger influxes and 2. DEA(E)-64 Project : "The use of quantitative risk analysis
therefore higher pressures exerted on the casing. (QRA) in casing/tubing design" : Enertech Europe (Draft
Report).
Will Risk Based Designs mean safer wells? 3. HSE:A guide to the wells aspects of the Offshore
Installations and Wells (Design and Construction etc)
If Risk Based Designs are used to justify thinner/lower Regulations 1996. ISBN 0 7176 11949.
grade casing and pipe manufactured to the same quality
standards as used as with Working Stress Designs, the
wells will not be safer.