Thread Content
Analysis of the bearing burnout incident in Expander No. 2 and summary of maintenance work: At 16:38:13 on February 19, the tripping of the new 110 kV substation caused all the operating equipment of Expander No. 2, which is a 14,000 capacity oxygen generator, to stop functioning due to the power outage. This sudden power outage and shutdown led to several equipment failures: the air compressor motor had its bearings burned out, the 2# turbine expander also experienced burned-out bearings, and the water supply pipes of the third-stage cooler in the 15,000 nitrogen plant became clogged, among other issues. Next, I will focus on explaining the course of events, the causes, and the repair process related to the accident in which the bearing shells at the pressure-increasing end of the 2# booster turbine expander were damaged, as a result of the interlock system that stops the expander when oil pressure drops not functioning properly after the oil pump stopped operating during a power outage and shutdown. After the air purge was successfully restarted at 08:42:02 on February 20, the No. 2 expander was started twice in succession around 09:30; abnormal noises were heard from the unit at both occasions, so the unit was immediately shut down manually and the incident was reported. Once #1 expander was started and operating normally, the alarm information and relevant operation trend charts before and after shutting down both expanders were compared and examined. 1. Safety shutdown status of Expander No. 1: The alarm message on the microcomputer indicated that at 16:38:23, the oil pressure PIAS406A of Expander No. 1 was ≤0.3 MPa ； 16:38:30, Flow rate of 1# expander FICS401A ≤ 5400 m3/h ； The trend chart shows that the flow rate FICS401A of the 1# expander and its speed SIAS401A dropped rapidly to 0 after the oil pressure decreased. As indicated by the above information, when the oil pressure falls below the interlock shutdown value of 0.3 MPa, this interlock activates in accordance with the designed procedure: the solenoid valve SV401A of the emergency shut-off valve HC401A and the solenoid valve SV402A of the booster pump return valve FICS401A lose power promptly upon receiving the interlock signal. The instrument air supply used to control the emergency shut-off valve HC401A is quickly released through the emergency vent, causing the valve to close completely within three seconds; at the same time, the return valve FICS401A is also caused to open completely. This allows the unit to shut down quickly in the absence of lubricant supply, thereby serving to protect the unit. Furthermore, the trend chart also shows that after the flow rate and rotational speed dropped rapidly, the fuel supply pressure took two minutes before falling below 0.05 MPa. This proves that the accumulator is still able to supply oil to various lubrication points of the unit for 2 minutes even after the oil pump stops operating, allowing the heat generated by the components that have been running at high speed for a long time to be partially carried away by the lubricating oil, thereby further ensuring a safe shutdown of the unit. 2. Wear of the bearings in Expander No. 2: The alarm message on the computer showed that at 16:38:37, the oil pressure PIAS406B of Expander No. 2 was ≤0.3 MPa ； 16:40:45,2＃Expander flow FACAS401B≤8340m3/h ； 16:42:49.2 – Temperature of bearing TIAS403B at the pressure-increasing end of expansion machine #2 ≥ 70°C. 16:42:57.2 – Temperature of bearing TIAS403B at the pressure-increasing end of expansion machine #2 ≥ 75°C℃ ； 16:44:08,2＃Expander flow FICS401B≤5400m3/h ； The trend chart shows that after the oil pressure of Expander No. 2 dropped to 0.3 MPa, its accumulator, just like that of Expander No. 1, was able to continue supplying oil to all the lubrication points of the unit for 2 minutes even after the oil pump stopped operating ； The trend chart also shows that after the oil pump stopped operating, Expander No. 2 continued to run at a high speed of 21,000 rpm and with a large flow rate of 5,500 m3/h for 7 minutes, until all the residual air in the air-cooling tower, the molecular sieve system, and the pipelines was completely exhausted. The above information shows that after the oil pump and accumulator stopped supplying oil, Expander No. 2 continued to operate at a high speed and with a high flow rate for 5 minutes; this inevitably led to the bearings and rotor moving relative to each other at high speeds without lubrication, resulting in their damage. Upon further examination of the temperature variation curves for each bearing of Expander No. 2, it was found that as the oil pressure decreased, the temperature of the bearing at the pressure-increasing end of Expander No. 2, TIAS403B, rose rapidly from 19.5°C before shutdown, reaching a peak of 132°C at 16:45 (the interlock shutdown threshold for the bearing temperature at the pressure-increasing end of the expander is 75°C; in other words, when the temperature reaches this value, there is a risk that the bearing surfaces could be damaged). From this, we conclude that the bearing shell at the pressure-increasing end of the 2# expander has been severely burned out. The abnormal sound heard when starting the unit could be the sound produced by the friction between the burned bearing and the rotor. During the subsequent disassembly of Expander No. 2 for maintenance, it was found that the babbitt layer on the inner surface of the bearing at the pressurization end had been completely melted due to extremely high temperatures; this further confirmed the correctness of the conclusion drawn from the alarm messages and trend charts on the computer, which indicated that the bearing shells had been burned out. 3. Analysis of the cause of bearing wear in Expander No. 2: So, why could the interlock system of Expander No. 1 operate as designed after the oil pump stopped running and the oil pressure dropped below the value that triggers an interlocked shutdown? In other words, why did the solenoid valve lose power rapidly, causing the emergency shut-off valve to close quickly, thereby protecting the unit? ； However, the interlocks associated with Expander No. 2 did not function, which led to severe damage of the bearing at the pressure-increasing end of that expander. After investigation by the instrumentation and control personnel, it was determined that the cause of this incident was that the \"operation status\" (MODATTR) of solenoid valve SV401A, which controls the emergency shut-off valve at the inlet of Expander No. 1, was set to \"program-controlled mode\" (PROGRAM). As a result, when the oil pressure dropped below the value that triggers an interlocked shutdown, solenoid valve SV401A could operate according to the programmed sequence to cause it to lose power quickly, thereby enabling the emergency shut-off valve to close swiftly. The \"Operation Status\" (MODATTR) of the solenoid valve SV401B, which is the emergency shut-off valve at the inlet of Expander No. 2, was set to \"Operator Control Status\" at that time; in other words, the operation of this valve was not controlled by a program but was entirely under manual control by the operator. In this way, when all interlock points, including the oil pressure, reach the interlock shutdown value, solenoid valve SV401B cannot be controlled by the design program to cause an interlock power failure. Based on the above analysis, we believe that the main reason for the burnout of the shaft bearing at the pressure-increasing end of Expander No. 2 was an insufficiently comprehensive and in-depth understanding of the control system of this oxygen production unit. Additionally, at 10:34:58 on August 4th, the operating oil pump stopped due to a change in the oil pressure interlock settings, and the backup pump did not start; as a result, the air compressor also stopped operating owing to the low oil pressure interlock. The cause of this accident was also due to the aforementioned reasons. Therefore, to prevent similar accidents from occurring again, it is necessary for professionals in relevant fields to conduct a comprehensive and systematic study of the control system of this oxygen generator set. The existing problem is: A. The solenoid valve SV101, which controls the air inlet valve FCV101 in the fractionation tower, fails to lose power as a result of the interlock activation, and thus prevents the FCV101 valve from closing quickly. B. When the 2# expander is operating normally, if the seal gas pressures PIA407B and PIA408B fall below the value allowed for starting the oil pump, namely the PVLL value, the oil pump will be shut down via interlock; however, this does not happen with the 1# expander. C. The operation status signals SOSAPr703 and SOSAPr704 for the two medium-pressure liquid argon pumps do not indicate whether they are running or stopped; instead, they turn red when power is supplied to that pump from the power distribution room, and green when no power is supplied. 4. Maintenance status of Expander No. 2: From February 21 to February 23, Liu Shuqing was in charge of carrying out emergency repairs on Expander No. 2. Condition before maintenance: The front bearing and front seal sleeve were completely damaged (the babbitt layer on the inner surface of the front bearing was completely melted and flowed away due to extremely high temperatures); the rotor and the front thrust surface were also damaged. The axial clearance of the thrust bearing was 0.8 mm (the standard range is 0.15–0.20 mm). The radial clearance of the front bearing could not be measured due to the damage (the standard range is 0.05–0.07), while the radial clearance of the rear bearing was 0.25 mm (the standard range is 0.05–0.07). Maintenance status: Replaced main spare parts: front bearing, rear bearing, front seal sleeve, rear seal sleeve ； Due to the lack of spare parts, the rotor was repaired and balanced before being put back into use, and the front oil seal was repaired and continued to be used. Total axial clearance of the thrust bearing: 0.16 mm (standard range: 0.15–0.20 mm). Radial clearance of the front bearing: 0.07 mm (standard range: 0.05–0.07); radial clearance of the rear bearing: 0.06 mm (standard range: 0.05–0.07). Radial clearance of the front seal sleeve: 0.04 mm (standard range: 0–0.04); radial clearance of the rear seal sleeve: 0.05 mm (standard range: 0–0.04). 5. Test run status of Expander No. 2: Maintenance of Expander No. 2 was completed on the afternoon of February 23, and it was ready for a test run. Considering the operating instructions for the turbine expanders produced by Kansei Air, no running-in of the reinstalled shaft seals is required, and no running-in of the seals is carried out either during the first start-up after the unit is installed in place. Therefore, our workshop plans to carry out the running-in of the seal in accordance with the instructions provided by Hangyang for the turbine expanders it produces. This involves controlling the opening degree of the inlet valve or nozzle so that the unit operates at 20%, 35%, 60%, 80%, and 100% of its rated speed, in five stages from low to high. The unit runs at each speed level for 10 minutes, after which it stops for 10 minutes to allow the graphite seal, which heats up during operation, to cool down. However, since it is not clear what the exact value of the unit’s first-order critical speed is, running the unit at low speeds close to or at this critical speed for an extended period during the running-in phase could cause damage to the unit; therefore, it was decided to conduct tests only after determining whether running-in is necessary and how to carry it out. After consulting the chief engineer at Senkū Low-Temperature Machinery Company, it was learned that the turbine expanders produced by Senkū are not subjected to any special running-in procedure for their seals even during tests in the factory; instead, they are operated at a lower speed (usually above the unit’s first-order critical speed, such as 13,000 rpm or more) for a period of time to check whether the installation clearance of the unit is appropriate. Only after all the measurement parameters are within normal ranges is the speed increased gradually until the rated speed is reached. The reason for conducting such tests on the expanders produced by Sichuan Airlines is that the sealing material used in the turbine expanders manufactured by this factory is babbitt, which is different from the graphite seals used by Hangzhou Oxygen; therefore, no sealing running-in is required. Thus, testing of the repaired No. 2 expander began on the morning of February 24. However, after the unit was operating normally at low speeds and attempts were made to increase its speed, the unit started to overspeed even with the booster return valve fully open and the nozzle opening at a low level. To identify the cause of the speed violation and resolve this issue, technical personnel from various departments—including Workshop 5, the Maintenance Workshop, the Control and Monitoring Workshop, the Electrical Workshop, the Technical Department, the Equipment Department, as well as the managers of Workshops 2 and 3—spent two days working together on this problem. During that period, Expander No. 2 was started and stopped 24 times, and Expander No. 1 was started and stopped 7 times as a result, until 6:20 p.m. on February 25th, when Expander No. 2 began operating officially. After the 2# expander experienced overspeed, we first compared its operating conditions and valve openings with those before maintenance to determine whether the unit indeed had overspeeded. After comparison, the units have the same flow rate when the relevant valve openings are the same; the only difference is that the rotational speed is extremely high during testing after maintenance. It can be inferred from this that the actual speed of the engine unit is not extremely high, and the phenomenon of an abnormally high speed is likely related to the measurement by the tachometer. Despite various measures taken by the instrumentation and control personnel, such as replacing each component of the vibration sensor for Expander No. 2 one by one, swapping the vibration sensors of Expander No. 2 and Expander No. 1, and adjusting the distance between the vibration probe and the rotor after consulting with the manufacturer of the vibration sensors, the problem still could not be resolved. Therefore, an oscilloscope was used to compare the waveforms of the two expanders under the same operating conditions in order to identify the cause. The oscilloscope shows that the waveform of Expander 1# is smooth, free of noise and interference; there are two waves per cycle, and one rotation count is recorded ； Meanwhile, the waveform of the 2# expander is chaotic and contains interference waves; there are two distinct waves in each cycle, as well as one less distinct but still detectable wave, which is counted 1.5 times. In other words, the 2# expander experiences noise interference between the two normal waves in each rotation cycle, causing the measured rotational speed to be 50% higher than the actual rotational speed. Through analysis, we believe that the cause of this interference wave is as follows: each expander rotor shaft has two tachometer grooves that are 180 degrees out of alignment. Under normal conditions, whenever the tachometer probe detects these two grooves twice, it records one rotation of the rotor, thereby calculating the speed. However, if there are obvious scratches on the rotor shaft or in the tachometer slots, it will interfere with the detection by the tachometer probe, resulting in an incorrect reading of the speed. After confirming that the speed indicated by the tachometer was 50% higher than the actual speed, it was decided to increase the interlock values for speed accordingly; that is, the original values of 29600 rpm and 31700 rpm were changed to 42000 rpm and 45000 rpm respectively, after which operations were carried out using these new interlock values. In the future, during annual maintenance, if the opportunity arises, the rotor can be taken out again for repair. The 2# expander was commissioned at 6 p.m. on February 25, and since its official operation began, the unit has been performing well. Under the same operating conditions, the temperature drop and efficiency of Expander No. 2 before and after maintenance are essentially the same.