Incident Report Page 1 of 6 Incident Description On the 10th of April, at approximately 4:00 AM, the stage 8 of the HP ground flare experienced flashback explosion, which resulted in damage to the runner and 9 tips, as shown in Fig. 1. The damage can be summarized as follows: The runner was corrugated. 9 tips were wrinkled with several cracks at the side of the runner.
Fig. 1 Damaged Runner and Tips The incident did not cause fatality and injury. At approximately 11:00 AM, the stage 8 was isolated from service by closing the corresponding block valve and spectacle blind. Causes of the Incident The vendor of the flare system was consulted for the causes of the incident. The main points of the response from the vendor are summarized as follows: The incident was a classic flashback explosion event. As illustrated in Fig. 2, the explosion was caused by the ingress of air from the flare tip to mix with the flammable gas. On emerging, the resultant explosive mixture was ignited by the standing pilot, which presented a continuous ignition source. Incident Report Page 2 of 6
Fig. 2 Flashback Explosion Since both the high performance butterfly valve and buckling pin bypass valve, as shown in Fig. 3, were in the fully-closed position when the flashback occurred, there should not have been big amount of waste gas in the runner. The vendor concluded that either the high performance butterfly valve or the buckling pin bypass valve had serious leakage.
Fig. 3 Staging Valve System The leakage gas from the valve entered the runner. Part of the gas migrated into the atmosphere due to buoyancy, and part of the gas met the air (oxygen) and created a combustible mixture. Thats why the flashback explosion took place in the middle part of the runner. When the flare system was designed, the vendor expected that both valves would be tight, and no continuous nitrogen purge after the staging valve system was defined. Incident Report Page 3 of 6 To verify which valve was leaking, the high performance butterfly valve (triple-offset butterfly valve) and the buckling pin bypass valve, as shown in Fig. 4, were removed and sent to a testing facility that could handle 12 valve. The valves were tested as per the procedure provided by the flare vendor. And both seat leakage test and stem system leakage test were performed, since the flammable gas monitoring system located around the staging section ever alarmed in the past.
Fig. 4 High Performance Butterfly Valve and Buckling Pin Bypass Valve For the high performance butterfly valve, the test results were shown in Tab. 1. Tab. 1 Test Results for High Performance Butterfly Valve Test Pressure (psig) Medium Duration (minute) Bubbles per Minute 1 Seat 90 Air 60 0 2 Stem 90 Air 60 0 For the buckling pin bypass valve, the test results were indicated in Tab. 2. Tab. 2 Test Results for Buckling Pin Bypass Valve Test Pressure (psig) Medium Duration (minute) Bubbles per Minute 1 Seat 15 Air 5 59 2 Stem 90 Air 60 2 From the above tables, it could be seen that the leakage gas was from the buckling pin bypass valve, which was manufactured by BS&B. Incident Report Page 4 of 6 Failure Analysis of Buckling Pin Bypass Valve After the valve was tested, the seat was flushed with Valve Cleaner Plus from SEALWELD for better visual examination. No obvious damage was found, as shown in Fig. 5.
Fig. 5 Seat Area of Buckling Pin Bypass Valve To verify whether the leakage was caused by the dust and greasy dirt in the seat area or not, the valve was tested again with air, and the result was shown in Tab. 3. Tab. 3 2 nd Test Results for Buckling Pin Relief Valve Test Pressure (psig) Medium Duration (minute) Bubbles per Minute 1 Seat 15 Air 5 41 It could be seen that the leakage was mitigated to some extent, but still unacceptable. The conclusion was that the impurity in the seat area was the minor factor leading to leakage. And then the sealing principle of the valve was analysed. As shown in Fig. 6, a sealing element is embedded in the valve body. The sealing element has two thin projections, which remain contact with the disk to maintain seal until the disk is rotated from the closed position to the open position, and these projections are the true sealing elements. The buckling pin bypass valve is an eccentric butterfly valve by nature, and all the specialists present in the examination were surprised to see that such narrow sealing elements were applied in the butterfly valve. After discussion, the specialists realized the philosophy behind selecting such weak Incident Report Page 5 of 6 sealing elements: The valve disk stem is located off-centre, causing the pressure force to be applied unequally. This causes a torque on the stem, which is transmitted through a complicated lever system to the buckling pin, located in the case on the side of the valve. When the set pressure is reached, the pin buckles, the disk rotates to the open position, and the system pressure is released. The offset is about 0.8, and the torque generated by such an offset under 15 psig is about 110 lbf*ft. However, the friction torque of a common eccentric butterfly valve with soft seat is, according to the information from Metso, above 100 lbf*ft. If a standard seat of butterfly valve were selected, the buckling pin bypass valve may not be opened at all. To guarantee that the valve can be opened, and to satisfy the requirement on tolerance, the friction must be minimized as much as possible. A narrow sealing surface helps because the thinner the sealing surface, the smaller the pressing force needed between the sealing elements when the sealing pressure is guaranteed. But the narrower the sealing surface is, the worse the sealing performance is, and the easier to damage. Because the capillary existed between sealing elements will be very short if the surface is narrow and this is detrimental to the tightness. A difficult compromise between friction and sealing results in the poor and exotic design of BS&Bs sealing structure, for both the seat and stem sealing system (as illustrated in Fig. 6). With such a weak sealing system, tightness is impossible after the valves are put into operation.
Fig. 6 Sealing System of Buckling Pin Bypass Valve All the specialists agreed that major factor which led to the leakage was the design of Incident Report Page 6 of 6 BS&B buckling pin bypass valve itself. Some specialists believed that BS&B also realized that the valves sealing system was too weak to satisfy long term operation. And this is why BS&B recommends to purchase spare seal for each valve (as illustrated in Fig. 7), which will be a headache for the operator.
Fig. 7 One Paragraph of BS&B Catalogue 77-1015 Conclusions and Recommendations The incident investigation highlighted the following: The incident was a classic flashback explosion event. The waste gas entered the runner because the buckling pin bypass valve had serious leakage. External leakage was also found in the pressure test of the buckling pin bypass valve. All the specialists agreed that the leakage was mainly due to the poor design of the buckling pin bypass valve itself. To prevent reoccurrence of similar incidents, the following recommendations were given: Inspect the tightness of other buckling pin bypass valves, and replace the seal to eliminate detected leakage. Search for buckling pin bypass valves with better design to replace the installed ones if really necessary. Add nitrogen inject points after all buckling pin bypass valves to sweep any waste gas and air that might enter into the runner. Minimum purge gas flow rates can be provided by the flare vendor. -END-