Abstract
Performing remaining creep life calculations using average values of temperature (T) and pressure (P) under conditions where these vary considerably produced unrealistic results. A case is described in which these variables were modeled to express T and P as functions of time. The case was made into an example of 9Cr-1Mo steel tube in a delayed coker furnace operating in 4 to 5 month cycles. During each cycle the metal temperature in the tube increased from about 900°-1000°F (482°-538°C) up to about 1300°F (704°C), while the pressure decreased as a result of coke fouling inside the tube. The predictions were obtained by integrating with different time increments, using both the Larson-Miller Parameter (LMP) and the MPC Omega methods. For this particular case described in this paper, the Omega method produced longer remaining life than the LMP method when based on the minimum creep strength level but shorter remaining life than the LMP method when based on the average creep strength level. Both methods produced equally acceptable results. The error of not including the corrosion rate was demonstrated to be significant. The difference between the estimated remaining lives obtained by using the minimum and average creep strength was rather large. A statistical approach was used to assign probability of failure to these predictions.
