Project Failures Minimization - Report Example

Background information As earlier chapters have noted, engineers contribute significantly to the minimization of project failures. In as much, this section critically assesses problems associated with engineers and sub-contractors who cause major problems to projects. In particular, the chapter assesses x as one of the biggest oil Companies in the world with two big refineries. Background information regarding the Company shows that it takes crude oil from other fields and produces other related products such as unleaded gasoline, diesel, petrol, kerosene LPG among other products. Contrariwise, just like any other refinery Company, its project has regular check annually and during the period, units such as pumps, towers, boilers and filters are shut down for a period of one month when such repairs and maintenance are carried out. According to the Company’s mission statement, such shut downs are targeted for general maintenance thus extending life spans of the entire plant and also all related equipment and machines. As a routine, x had the last shut down in August 2013 and in the process, engaged a number of engineers as sub-contractors for doing the periodic maintenance. As matters turned out, 15 engineers engaged as sub-contractors failed to accomplish tasks allocated. These problems mainly affected pumps and boilers and to bigger extend, the number 4 Crude Unit (one of the biggest pipes serving boilers). Four days after completion of the maintenance, number 4 Crude Unit ruptured, releasing hydrocarbon, flammable gas and process fluids which also affected other sections such as pumps, boilers, absorbers and converter units which upon investigation, had been ignored by these engineers during the maitenance. Though the issue at hand was shortage of manpower and inadequate resources, failing to adhere to recommended codes of practice generated major problems for this project thus leading to its shut down. This is the point of departure for this chapter; critical analysis of the main problems that were committed by the sub-contracted engineers and how such affected the Company. In so doing, the case study will link the earlier discussed hypotheses, theoretical frameworks and approaches to project failure minimization. Engineering problems and failure path of X The main problem started from the piping known as 4-sidecut/ number 4 Crude Unit; one of the existing process streams within x. A plot plan shows that the sub-contracted engineers failed to notice the corrosion in the pipe. According to Center for Chemical Process Safety (2013), the rupture as shown in figure 1 below occurred about 51-inch long component of the pipe. Figure: Rupture on 4-sidecut The first problem related to this rupture is failure by the sub-contracted engineers to notice suldification corrosion in the pipe. Suldification corrosion is a damage engineers in the refinery industry should notice prior to starting the scheduled maitenance. This is the first issue that was ignored by the sub-contracted engineers. For instance, American Petroleum Institute offers Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in Oil Refineries which ought to have been followed by the engineers (ASTM Standard, 2013). This is where reliability as earlier analysed is applicable. If repairs are not possible as it was for this case then reliability could have saved the situation. Reliability definitions cover values between 0 and 1 to means absence of reliability and perfect reliability respectively. Refinery industries have experienced number of project failures when engineers fail to notice and act on sulfidation corrosion and a good example is Saint Mary's Refining Co which experienced complete shut down on their plant when maintaining engineers failed to notice corrosion in the carbon steel piping that was having low levels of silicon. Bringing such within x, engineers failed to notice the anomaly. That is, they failed to count the number of failures that had occurred before during a given time interval as far as sulfidation corrosion is concerned so as to enable them calculate MTBF. Since failure rate is the reciprocal of MTBF, we can assume that MTBFC = (35,040 hours/(2failures) =17,520 hours/failure. Figure2: 4-sidecut MTBF and Failure Rate. Having failed to notice suldification corrosion in the pipe, there were questionable judgements, engineering errors and decisions that finally led to the shutdown of the plant. While some of these issues were related to the x as a company, engineering played a key role. The first issue was failure to adhere to inherently safer systems in the process of carrying out maintenance. X is a corporate member of Center for Chemical Process Safety therefore obliged to follow process safety needs within its industries. U.S. Environmental Protection Agency (2014) defines inherently safer designs as the process of identifying and implementation of safety materials that are inseparable and permanent. However, this was not the case with these engineers. That is, they failed to recognize anomaly with the pipe. Secondly, looking at the image above, engineers can rate it in terms of reliability, unreliability, availability or unavailability. Discerning one issue at a time, the image showed unreliability since it had experienced or by the time of maintenance, experienced the first failure or had failed in one or several occasions which can be inferred in terms of time interval zero time to time (t) assuming that the maintenance given to it repaired it to a like a new condition at time zero. Therefore relating the period it took between the time the tube finally failed and the time maintenance was done to it the following expression holds true: Reliability (t) + Unreliability (t) =1 or Unreliability F (t) =1-R(t)….equation (1) Considering the aspect of availability and unavailability with the same tube, first, reliability studies and engineering allows the comparison of unavailability and unreliability values rather than availability and reliability. In addition, the numerical values of both the unavailability and availability are shown as probability from 0-1 and no units. Considering this prior to the rupture, the tube was either operating or not operating, the above statement is conceptualized using the expression below: A (t) + Q (t) = 1….equation (2). Both equations in testing for reliability assessment as well as safety Unavailability Q (t) ≤ Unreliability F (t) For a non-repairable component or tube: Unreliability F (t) = Unavailability Q (t) Using the above expressions to support the argument, first, there were no reliability predictions on the tube following formulars above thus creating the shutdown. Since this was not done, none of the engineers gave a report on failure intensity and conditional failure thus making it difficult for x to ascertain Mean Time Between Failures (MTBF). Practically, the Company shut down few days after engineering repairs and maintenance. Based on this, MTBF ought to have been calculated as the inverse of the failure rate since the tube as can be seen from the figure depicted constant failure rates. Since report by Center for Chemical Process Safety (2013) suggests that the tube had failed twice in 2 years, it had a failure rate of 2 per 1 million hours therefore the MTBF would be the inverse of such failure rate expressed as; MTBF = 1/λ thus being 500,000 hours/failure. Basically equations 1 and 2 give working formular for relating Unreliability and Unavailability. However, since these engineers failed to estimate equations 1 and 2 based on what they were seeing in figure 1 above, it become difficult for them to calculate and report its MTBF. This is yet another problem that was committed by the engineers. Contemporary schools of thought tend to deviate from predictive maintenance and exponential distribution by integrating Weibulls and other distributions with probabilistic failure calculations to indicate how engineers can effectively stop x from failing. Unfortunately, these engineers still relied on predictive maintenance which could not enable them come up with proper strategy. Giving a definition of probabilistic failure, Layer of Protection Analysis (2013) defines it as a reflection of a given project, engine or hardware’s relative probability of age against its failure. On the other hand, Chemical Emergency Preparedness and Prevention Office (1999) defines probabilistic failure mode as “the probabilistic modes engineers make on project to determine chances of its failure based on age or hardware components” (p. 203). Since this part is interested in the figure x these engineers did little if nothing in integration of probabilistic failure calculations with regard to such tube. In addition, there is one issue that should be noted with regard to these engineers failing to follow what are inherently safer systems. That is, inherently safer materials and technologies are relative. While the first group of the sub-contracted engineers described materials they introduced while doing their maintenance as inherently safer, it seems the materials used did not consider specific hazard and or risk. Figure 3: 4-sidecut piping sample Relating figure 2 above with the definition of inherently safer systems, there is relative thickness of low silicon piping appearing on the left. On the other hand, there is high silicon piping on the right. However, looking at the ruptured pipe (component at the left) there is only 0.01% of silicon. In as much, there is 0.16% of silicon at the upstream elbow. Basically, this is not the recommended thickness of the piping thus failure by engineers to notice this does not only worry but brings questions to their credibility as engineers. What these engineers failed to do with regard to these materials was application of simple calculation of failure rates of tubes and conductor materials. As noted in the literature review, one method of determining rate of failure is to use accelerated high temperature operating life tests that can be performed on the tubes. Therefore, a simple rate of failure calculation basing on figure 1 above would follow the equation below: λ∞1/TDH*AF Where THD is the total device hours, AF acceleration factor (the AF allows extrapolation of failure rates from accelerated test conditions to use conditions λ is the failure rate. Since data of reliability can be accumulated from a number of unique life tests and maintenance as seen from the engineers and with different failure mechanisms, the expression above can reveal where the engineers erred. This is why the equation below integrates the above so as to obtain the confidence level for the resulting failure rate. λ= failure rate in FITS (Failure In Time) (number of fails in 109 device hours) β=number of unique possible failure mechanisms k=number of life tests as combined by the engineers xi =number of failure for a given failure mechanisms i=1,2,… β TDHJ =total device hours of test time for its life test j, j=1,2,…k AFij=acceleration factor for a given failure mechanism, i=1,2,…k Therefore basing from the above expression, when doing failure rate calculation, factors AFij can be used to derate failure rates of the pipe especially from the thermally accelerated life test conditions to a failure rate indicative of actual use temperature. To determine acceleration factor, one can use Arrhenius equation which is also used in description of physio-chemical reaction rates since it has been found to be appropriate model when expressing thermal acceleration of tubes and failure mechanisms. Where there evidence of engineering mathematics expressed in attempts to find solution it was also evidenced that such was limited especially on the bypass line (a factor indicating lack of maintenance). This means that bypass line was not pressure tested after plant modifications done in 2010 as was indicated by engineers who worked prior to this group. This translates that there was latent failure including absence of full risk assessment that could support modification of the plant. After plant shut down, a naphtha hydrotreater unit was found to be cracked extensively. This was also attributed to the cub-contracted engineers doing maintenance. Upon further examination, there were initial transverse cracks from unit. However, what these engineers did was to fail to coat surfaces of this unit with at least a fairly thick powdery deposit to prevent significant cracks. An intergranular failure was also noticed in the unit; something the engineers failed to notice prior the shutdown. This is the point where we can apply the aspect of reliability engineering with regard to the Weibull Distribution. These engineers contribute significantly towards the failure by failing to characterise the probability density function (termed as failure frequency distribution) of the unit data and characterize or group the failures as constant (exponential) early life or wear out (log normal or Gaussian). Conversely, in cases where there is extensive cracking as witnessed after the shut-down, an assumption that can be made is that there was poor inspection and maintenance decisions. Inspections on x appear to be lacking in many sectors even if they had engineers working on the same annually. Fire pumps, life rafts and emergency lighting did not appear to have received attention according to the prescribed code of conduct. This explains why there were cracks that must have been ignored during the periodic maintenance. To be specific, minimal response to inspection by the last group of engineers was a factor that weakened pumps. The most sensitive maintenance problem was the carelessness with which PSV 504 was removed and further replaced without proper tagging thus putting the pumps as shown below out of service. Strangely, engineers working in a different shift were not informed of this problem and made attempts to restart the pump in which the initial link seems to have begun. Although this maintenance anomaly can be attributed to history and inexperience, these engineers had a hand in the case due to their bypassed procedures. Another problem noticed by group of engineers contracted in the design of this Company was lack of specific fire criteria in design of structure. Prior to the shut-down of the project, engineers constantly checked suitability of the specific fire criteria in design of the structure. According the report that was given there was not point where these engineers suggested a change to the same. This was a failure in the sense that at first, fire risks are accounted for in the design of the top-side and this is done alter existence of ignition sources and vapors. The fire mistake that these engineers did according to Center for Chemical Process Safety (2013) report was to fail to link fire loads in the design of the project structure in the same way they did with wave loads. Therefore as matters were, there were no attempts to assess the annual probabilities of different fire loads to which the structure could be subjected thus adjusting the design parameters to provide thermal robustness such as inherent fire resistance. The point is, even sub-contracted engineers failed to recognize that same approach adopted for wave loads was needed for fire related risks. This is what made it also difficult for the last group of sub-contracted engineers to detect anomalies in the gas turbines. Report before the shut-down of the project had indicated that there was a major problem in gas turbine connecting to (Center for Chemical Process Safety, 2013). Therefore it was apparent that unburned or uncontrolled fuel in the combustor had the chance of causing significant number of abnormal situations in the gas turbines thus leading to sudden shut down. However, in as much as there were abnormal cases such as damaged blades (due to localized explosions, high temperatures), loss of fuel efficiency and damaged nozzles (due to high temperatures) engineers could not do anything to avert such cases thus plunging the plant into more problems. Basically, major mistakes done by these engineers were: Failing to detect abnormalities in fuel system so as to prevent abnormal situations in gas turbines Failing to detect abnormalities in expander so as to prevent abnormal cases in gas turbine and prevent unscheduled shutdowns Failing to control and monitor the supply of fuel so as to prevent abnormal situations in the gas turbine It is because of this reason that shortly after the bursting of number 4 Crude Unit, there was an explosion and fire occurrence at the Fluid Catalytic Cracking Unit which further damaged Slurry pumps, Slurry Settler, Steam Blowdown Drum, Piping, structures, Steam Superheater Exchanger and Instrumentations. The figure below shows an image of the explosion. Figure 4: Explosion of the plant at x The above analysed error is conceptualized by the engineering findings conducted at the plant after the blast. These engineers found that there was a 480-560 psi pressure available in the Slurry Settler which was required for the rupture and the operating temperature which was above 6200F. This was an indication that there was poor servicing and or recommendations at the Slurry Settler. On the other hand, calculations on the blast effect were found to be around 130-190 pounds TNT. This was dangerous for its operation and it was surprising that the sub-contracted engineers did not detect the abnormality because it is definite that air was trapped inside the vessel during its start up. Adherence to safety culture is another area that engineers working on x failed to observe. A safety culture is defined based on clear understanding of the systems and the related safety features, clear and positive attitude towards prescribed positive measures. This was not the cases with the latest group of the sub-contracted engineers basing on the report released by Center for Chemical Process Safety (2013). Preliminary investigation indicates that there was a vessel for monoethanolamine absorber adjacent to pipe number 4 Crude Unit and this absorber had been in x for over ten years. Besides, this absorber had one-inch thickness and ASTM A516 Gr 70 steel plates welded and rolled with full submerged arc without post weld heat treatment. While this was the first human error with the designing engineers, the report further indicates that prior to the rupture and subsequent shut down, the contracted engineers failed to notice that there was 6-inches crack. Bringing the aspect of safety culture in practice, it is expected that when engineers are contracted or their skills sought after regarding some technical or engineering issues, detecting a crack should not be a problem as was seen in this case. This crack did not cause the eventual closure of the plant but when the explosion occurred, the absorber created sequential explosion and fire within the Company. Such erroneous and or failure to adhere to the prescribed safety practices relates to Deming Cycle as earlier considered. That is, the engineers ought to have planned for the tasks and activities affecting the quality of the absorber so as to gain the needed result. Another problem that could be related to this case is that these engineers were working without relevant expertise and resources and the best example of another company that failed when contracted group of engineers failed to apply relevant expertise and resources was Saint Mary's Refining Company. As report provides it, failure to adhere to the safety standards went as far as these engineers failing to monitor the functionality of this absorber and make it be able to raise its alarm should there be a change likely to affect its performance. Furthermore, using the exponential distribution approach, there was no evidence that the sub-contracted engineers designed and put in place on-board “quick-test routine” which in this case ought to have been dynamic thyristor diagnostic (DTD) that minimizes the failure by monitoring the load in real-time and thyristor unit. If there is one reason why x engages engineers is to maintain the plant then this is to have these engineers monitor and inspect chances of failure by identifying all possible faults within the absorber and number 4 Crude Unit or in some cases, elimination of thryristor when it is detected as the possible cause of the ultimate failure of the entire plant. Conversely, in busy refinery companies such as x, working engineers have to know some basics of operations. One of such is the fact that the layout of the topside should be generally guided by area-classification concepts whose prime goal or objective is to be able to separate the flammable vapors and gas anticipated under normal conditions from the ignition sources---we can take a clear case of the absorber, electrical equipment or the tube number 4 Crude Unit. One thing that was clear to these engineers was that areas or sections where vapors are expected were categorized as Division 1. It was therefore the responsibility of these engineers to ensure that these sections are fitted with explosion-resistant equipment. On other hand, sections where vapors are present and such appears only under abnormal conditions were categorized by designing engineers as Division 2. These sections were supposed to be fitted with non-sparking or vapor tight materials. However, as investigations had it, these are some of the issues working engineers during this period failed to consider. Though it can be argued that it was a problem committed by designing engineers but the question still lies with the sub-contracted engineers since they did not notify relevant authority of these anomalies. In addition, it remains to be fact that when x was established, there were some engineering aspects that were not considered to be relevant. Some these were the exact location of telecom/radio and control rooms. Basing on the control room, the sub-contracted engineers have been servicing this Company over 12 years and strangely; have failed to notify relevant authority that electrical classification for instance is a determinant of the design criteria. Therefore when number 4 Crude Unit ruptured, it would have been easy to save some of the structures such as absorbers. While this can be related to negligence by engineers, Henry Petroski research on “To Engineer is Human – The Role of Failure in Successful Design” where he believes that at least 30% of the projects fail because engineers exhibit insufficient knowledge on project management. Similarly, failing to recognize and report exact position of the control room with regard to dangers related with such shows a problem of failing to apply roadmap to project self-assessment as a tool that engineers should use to minimize chances of project failures. It is also evidenced from interviews conducted by Layer of Protection Analysis (2013) that these engineers had a problem that can be termed as ‘deficiency of the Permit-to-Work System.’ In his investigation on how Engineers led to the failure and subsequent shut down of Gold Line Refining LTD, he provides detailed analysis of deficiency of the Permit-to-Work System. However, within the context of x, the term means engaging engineering practices and services when the contracting party or company does not provide the needed time and resources. Therefore because engineering culture discourages shortcuts, it is issue that can be blamed on the engineers for having allowed themselves to take up the job of maintenance knowing the constraints they had in terms of resources and time. Giving practical evidence, prior to completion of the first phase of their maintenance, there was no report or any indication or mention of locking and tagging off of isolation valves leading to the ruptured tube which had been opened or closed as part of making that the tube and the absorber connected to it were safe to work upon. Secondly, the communication problems among the engineers seemed to have been a general one. This is because the engineers were working in shifts and from the investigation, there were no leading guidelines that could enable incoming engineers to know which tubes and or machines had been worked on and which ones had been isolated for further maintenance. The implication of this was seen with absorber during the rupture. When one is contracted to engage in engineering maintenance, first, it is essential to the aspect of couplings and dependencies among components of the refinery. That is, how maintenance of one component can affect others. The question was not addressed. This is raise concern whether formal procedure was too complicated for these engineers and thus considered it wise to take shortcuts to alleviate some of the problems they could identify. There is connectedness of the issue raised in the paragraph above with hazard analysis processes. Maintenance of x includes analysis of hazard using mechanisms such as Process Hazard Analysis (PHA). This actually lacked in the report presented by Layer of Protection Analysis (2013). This can even be linked to tube number 4 Crude Unit since the engineers failed to determine the likelihood of hazardous consequences occurring not only to the tube but also the absorber and filter. The problem done was simple evaluation of the machines without identification safeguards which could reduce the risk and fires to an acceptable level. Had the report recognized that there was application of semi-quantitative and quantitative or qualitative tools then it could have been concluded that there was application of Process Hazard Analysis in the hazard analysis by the engineers. Engineering risk analysis strategy allows them to rank risks from it is detected to the time its period of operation comes. To represent such analysis mathematically, these engineers did not apply accident sequence approach where p(X) denotes the probability of an even of the engineering project per time unit (of the operation, p(X|Y) as the conditional probability of the factor X given Y, p(X,Y) being the joint possibility or probability of X and Y, IEi being the possible initiating events of failure of projects by engineers (or accident sequences) which is indexed in i and F representing the (cumulative or total7) technical failures of a given project. It is clear that x did have a prescribed methodology that the engineers could use to determine the chances that an incident or accident would occur or whether a safeguard put in place would be effective. This is why these engineers are to blame since the Company relied on the PHA team among the engineers who ought to have analysed the situation based on their ethics, experience, and areas of expertise. In its 2012 crude unit Process Hazard Analysis, x simply cited judgement-based qualitative safeguards and none-specific analysis like provision of pipe wall corrosion allowance, use of metallurgy to minimize corrosion and availability of effective inspection and maintenance programmes (Kletz et al. 2010). This is basically where these engineers failed at as far as their Process Hazard Analysis was concerned---to considering that the report given with regard to the 2012 crude unit Process Hazard Analysis was not only misleading but insufficient. Secondly, the effectiveness of these safeguards were neither documented nor evaluated by the engineer; instead, these engineers merely listed these same safeguards in their Process Hazard Analysis. Had the adequacy of the 2012 crude unit Process Hazard Analysis been verified and taken into considerations by these engineers, improved safeguards aimed at protecting sulfidation-induced failure as seen the tube could have been recommended for further actions. During the periodic maintenance to x, the usually contaminated residual gases should be filtered. Therefore the tube leading to the absorber is shutdown to clean the filters and restarting it regularly to ensure that it is functioning properly. Therefore a fast-track starting procedures had to be developed by these engineers, albeit without the needed testing. This was the first problem committed to this tube. Secondly, the isomerization unit whose raffinate tower was overfilled by workers in the shift prior to the maintenance led to the overheating of the pressure relief devices and the raffinate. However, these engineers did not recognize that such mistake could lead to a flammable liquid geyser coming from a blow-down stack to the left the tube that was unequipped with flare thus leading to an explosion. While the report by Chemical Emergency Preparedness and Prevention Office (2014) shows that major hazards in this tube were explosive and flammable liquid---raffinate, major problems associated with engineers in this case were failure to make calibrations on the level transmitter leading to the absorber properly which was purely maintenance error. Secondly, there was failure on high level alarm, which was basically lack of maintenance from subcontracted engineers and lastly, failure to clean sight glass which led to the accumulation of the flammable liquid. In general, the main problem with these engineers as far as the absorber and the tube were concerned was deficiency in the hands-on-actions that could retain an item or make restoration to the tube a condition to a state in which it can be able to perform its required functions such as wrong performance of a correct task. This kind of problem raises the question with regard to several instrumentation manufacturers who have been brought Failure Modes Effects and Diagnostic Analysis (FMEDA) to help these engineers in the provision of specific failure rates for each failure mode with regard to a given instrumentation product. This is what has been lacking from the report that was given by the investigation. This is therefore a clear case of a problem from engineers that led to active influencing maintenance factors since it referred to directly maintenance errors that directly influenced the realization of a the shutdown. Cases as what has been discussed in the paragraph above shows that engineers fail to apply Weibull plot and or its analysis when handling such maintenance. From Weibull Plot, the linear slope of the curve is vital variable for engineers (known as shape parameter) and is represented by â. Therefore maintenance errors as noted above are inevitable adoption of such plot is not used to adjust the exponential distribution so as to take care of a wide number of failure distributions. That is, if the shape parameter (â) falls below 1.0, the distribution of the project shows early life and unlikely to fail. Going by the details that there was deficiency in the hands-on-actions that could retain an item or make restoration to the tube it means the absorber and tube exhibited shape parameter that exceeded 3.5 thus wear out failures. From the analyses above, causes of the ultimate shutdown of x has been linked differently. These causes were both maintenance management cycle factors as well as barrier-based from engineering team sub-contracted for maintenance. Secondly, the numbers of cases reviewed within x were exhaustive and in as much, the most occurring problems with the engineers were deficient execution, deficient planning and deficient checking. References API RP 939-C. "Guidelines for Avoiding Sulfidation (Sulfidic) Corrosion Failures in Oil Refineries.” 1st ed., Section 4, May 2009. ASTM Standard A53/A53M-12 (2013). "Standard Specification for Pipe, Steel, Black and Hot Dipped, Zinc-Coated, Welded and Seamless." Center for Chemical Process Safety (CCPS) (2013). “Layer of Protection Analysis – Simplified Process Risk Assessment.” Chemical Emergency Preparedness and Prevention Office, “RMPs Are on the Way! How LEPCs and Other Local Agencies Can Include Information from Risk Management Plans in Their Ongoing Work.” November 1999. http://www.epa.gov/osweroe1/docs/chem/lepc-rmp.pdf (accessed April 3, 2014). Kletz, Trevor, and Paul Amyotte (2010). “Process Plants: A Handbook for Inherently Safer Design.” 2nd ed., Section 1.1, Page 14. Layer of Protection Analysis (2013): Simplified Process Risk Assessment. Center for Chemical Process Safety of the American Institute of Chemical Engineers. U.S. Environmental Protection Agency (2014), “General Guidance on Risk Management Programs for Chemical Accident Prevention (40 CFR Part 68).” Page i. March 2009. http://www.epa.gov/oem/docs/chem/Intro_final.pdf (accessed April 3, 2014). Read More