Page 240 - Pipeline Risk Management Manual Ideas, Techniques, and Resources
P. 240

Upset score 10/217
               scored. In the service interruption risk, interventions to prevent   reduced probability of the event causing customer upset. This is
              events that lead to service interruptions are also scored early in   an important  distinction;  see the  example at  the end of this
              the assessment, but then another intervention possibility is fac-   chapter. A qualitative scale to assign a value to the intervention
              tored in. This reflects the opportunity for intervention after an   adjustment follows.
               episode has occurred that would otherwise lead to a service
               interruption.  In the risk numbers,  this adjustment  allows the   Early detection is not always possible   No adjustment
               section score to partially “recover” from low points in episode          awarded
               likelihood.                                When the excursion is not detectable, reactionary intervention
                In many pipeline systems for which an uninterruptible  sup-   is not possible. When some of the possible excursions are
              ply is critical, extra provisions have been made to ensure that   detectable, score according to the next category.
              supply. These provisions allow for reactions to events that, if   Early detectionlaction   Up to 30%
              not  addressed, would cause service  interruptions.  Examples   Maximum points are awarded when all excursion episodes are
               include halting the flow of an offending product  stream and   detected early enough to allow action to be taken to prevent
              replacing it with an acceptable product stream, blending of a   or minimize customer interruption. This may be at a point
              contaminant to reduce concentration levels, treating of a con-   where contaminated product is prevented from entering the
              taminant on-line, and notifying the customer so that alternate   pipeline, or where product streams may be blended to reduce
              supplies  can be  arranged.  The reactions  can  be  assessed  in   contaminate levels to appropriate concentrations, alternate
              terms of their effectiveness in preventing service interruptions   delivery equipment can be brought into play, or where alter-
              after an event has occurred.                 nate sources can be routed to the customer. The reliability of
                Even a pipeline failure will not necessarily cause a service   detection must be considered here. The time to detect and
              interruption. This would be the case if an alternative  supply   take  action  must  include  time  to  receive,  interpret,  and
              could be found to replace  the lost supply or the leak can be   respond to the detection information.  Indirect indications,
              repaired with the pipeline in service. Note that in assessing the   such as a pressure  drop after an accidental valve closure,
              effectiveness ofa reaction, a time variable may be important. A   serve as detection mechanisms. Note that unusual events will
              given reaction may prevent a service interruption for only a   normally require more human analysis time before action is
              certain  amount  of time beyond  which  the  interruption will   taken. Emergency drills can be a useful measure of detec-
              occur.                                       tionireaction times.
                Note that by use of this adjustment factor, a high-probability
              excursion that has a low probability of actually impacting the   Often a point on the pipeline near the customer may have a
              customer is recognized and scored differently than the same   problem (such as a closed mainline valve) in which there would
              event that is more likely to impact the customer. Some interven-   not be enough time to make a meaningful early detection and
              tions have already been included in assessing the upset score.   notification. When some excursion types can be detected and
              Reconsidering them here is acceptable as long as a consistent   some may not be, or when detection is not reliable, no more
              approach is used.The intervention adjustment is sensitive to the   than 10% should be awarded.
              section being evaluated.
                System  dynamics  play  a  role  in  assessing  interventions.   Customer warning is sufficient   50%
              Consideration should be given to systems that are more “for-   to prevent an outage for that customer
              giving” in that they are slower to react to an upset. An example   These percentage points are awarded only when there exists a
              of this is a high-pressure,  large-volume gas system in which   unique situation in which, by the action of notifying the cus-
              outflows will only slowly depressure the system upon tempo-   tomer of a pending specification violation, that customer can
              rary loss of inflows. Contrast this with a small-volume liquid   always take action to prevent an outage. Coupled with a reli-
              system that  is effectively “tight-lined  (inflows balance  out-   able early detection ability, this allows an 80% (30% + 50%)
              flows with no temporary  imbalances tolerable). In this latter   factor to reduce the service interruption potential. An exam-
              case, reaction times are more critical.      ple would be an industrial consumer with alternative sup-
                To score the availability and reliability of interventions, add   plies where, on notification, the customer can easily switch
              percentages for all the mitigating actions that are present and   to an alternate supply.
              functioning. Note that these actions apply to any and all identi-   Customer warning will minimize   10%
              fied episodes  of product  specification  deviation  or  delivery   impact (but not always prevent an outage)
              parameter deviation. If an action cannot reliably address excur-   When a customer early warning is useful but will not always
              sions of any type, then intervention credit is awarded only to the   prevent an outage, these percentage points are awarded. An
              benefiting excursion. For example, if an early detection system   example would be an industrial user who, on notification of a
              can find and allow reporting of a contamination episode, but   pending service interruption,  can perform an orderly shut-
              there is no equivalent system to react quickly to a pipeline fail-   down of an operation rather than an emergency shutdown
              ure, then the intervention adjustment is only applied to the PSD.   with  its  inherent  safety  and  equipment  damage  issues.
              Therefore, these percentages will be used to adjust scores for   Almost  every customer will benefit to some degree  from
              PSD and DPD independently. The percentage will apply to the   early  warning.  Even  residential  gas  users,  given  a  few
              difference between the actual PSD or DPD score and the maxi-   moments  notice  before  an  outage,  can  make  plans  and
              mum possible score, up to 80%. The means that the PSD andlor   adjustments to better respond to the service interruption. The
              DPD scores can recover from low point conditions up to 80% of   customer’s ability to react to the notification should be meas-
              the maximum possible points. Increasing points in this fashion   ured  assuming the most  likely detectiodnotification time
              does not indicate a reduced probability of the event, only the   period.
   235   236   237   238   239   240   241   242   243   244   245