Power Train Transient High Vibrations in steady state operations - A Study in an Acid Plant
Sai Vamsidhar Bontha
Asset Reliability & Maintenance Management || Agile Digital Transformation & Project Management || Domains: Oil & Gas - FPSO || Refinery & Petchem || Power || Fibres || Fertilisers.
Machine / Power train consists of Compressor along with Expander (Make: Sulzer) driven by Steam Turbine (Make: Elliot) through gear coupling. On the other side, expander is installed on the same shaft after the compressor bearing in overhung arrangement. Compressor operated at maximum continuous speed of 6100 RPM. Critical speed of the machine is 3600 RPM. Two orthogonal proximity probes installed on Compressor DE (turbine side) and NDE (expander side) bearings. One thrust probe is installed at turbine NDE side thrust bearing with no vibration probes for the turbine radial bearings. Permanent Keyphasor is installed for phase measurements. Machine is connected to Bently Nevada 3500 Rack for monitoring purposes.
In order to assess the machine vibration, Plant’s Bently Nevada? System 1 classic V6.90 software was used to view data from the proximity transducer vibration signals. The data in this report has been selected from all the recorded data and is limited to the most relevant plots only.
The scope of the study is defined by the following objectives:
? Collect and analyze the steady state and Transient data.
? Detection of existing malfunctions, if any.
? Provide recommendations on further action as and if necessary
Machine train complete overhauling was carried out during the outage of May– June 2023. After that, unit was started on 13th June 2023.
During the overhauling, below major activities were carried out:
1. Compressor rotor was replaced with another re-furbished one.
2. Compressor NDE bearing (expander side) was replaced with new one.
3. Complete new coupling assembly was installed.
4. With respect to turbine, all the machine major parts were cleaned and re-installed after complete inspection.
After Complete overhauling machine started on 13th June’23 and all the vibrations levels are recorded within the limits. But minor fluctuations in compressor bearing direct vibration along with 1X vibration was recorded in system 1. Higher fluctuations observed in Expander DE side bearing when compared to Compressor DE bearing. Machine stopped on 28th June 23, during that time there is no abnormality observed in Transient data.
After that machine train started on 12th July’23 and running continuously till September’23. During that period as per site information there is more fluctuations observed in Compressor Bearings especially at Expander DE side. (From 12th July 2023 to till august 16th 2023 - there is no data available in system 1 due to technical issue). Other than vibration fluctuation, during the shutdown on 30th September in transient data Direct/1X vibration increase observed less than 1200 rpm and reached compressor trip limit of vibration (4.5 mils). Due to this vibration increase machine can’t be able to run its soaking period during shutdown.
After that machine train started on 05th October’23 and ran till 11th October. During this period high direct/1X vibration fluctuation observed in machine steady state condition. Also, during shutdown vibration increase observed less the 1200 rpm and reached compressor trip limit of vibration (4.5 mils).
After machine tripped during shutdown on 11th October’23, it was decided to start the machine again and run for its soaking period. But once it reached its soaking period of 1100 rpm, Direct/1X vibration gradually increased and reached its trip limits of 4.5 mils in the time duration of 10 to 15 min.
After inspection of Expander DE bearing machine was started on 15th October’23. During the inspection it was observed that minor coaking materials found inside the bearing labyrinth seals. While startup, during soaking speed (1100 RPM) gradual vibration increase observed further it was decided to reduce the speed to 500 rpm and run for few hours to maintain the stable vibration. Then slowly step by step speed was increased to 6200 RPM (full speed). It was observed that minor fluctuations in vibration at compressor and Expander bearing in Steady state condition
?
Discussion of Vibration Plots:
With respect to data collected during 05th October to 11th October’23, high vibration fluctuations observed in Expander DE bearing when compared to Compressor DE bearing at steady state condition.
During high vibration event in Expander, 1X vibration increasing along with the phase change. This can happen during light rubbing. Due to rubbing temporary thermal bow will happen in the rotor. This can be the cause for phase change. Once rubbing stopped, temperature will immediately drop, and the rotor will come to its original position. That event will reduce 1X vibration.
Polar plot loop form clearly indicated the symptom of thermal deflection in rotor.
There are 02 distinct problems regarding trips on vibrations in the power train:
·?????? Vibrations Peak Transients during steady state operation at high RPM.
·?????? Vibration Peak Transients during slow roll during coast down or ramp-up.
Spectrum indicates transient soft rubbing.
There is no permanent significant mechanical damage envisaged as the high vibrations occur transiently and not continuously observed.
1.Vibrations Peak Transients during steady state operation at high temperatures:
A. For transient vibration peaks at high speeds in steady state, the indication from the vibration analysis is bowing / soft rubbing due to carburisation particles on the oil seals rubbing with shaft or due to improper / uneven clearances. (Direct correlation between LO Temperatures & high vibrations.)
B. Current LO Flash point is less than leaking hot air temperature.
领英推è
2.Vibration Peak Transients during slow roll during coast down or ramp-up:
A. For the vibration transients at slow roll, the probable cause is the rubbing / unbalance due to temporary thermal sag. Also, the same could be due to operating parameters / temperatures gradients and the lower RPM and improper / uneven clearances). Direct correlation observed between suction air temperature and vibrations.
B. The expander has overhung impeller there is no NDE bearing on the expander & no barring device for barring when the unit is on standby / shutdown.
In Conclusion: Power Train – High Vibrations – Forward Path: Gist of Actions.
A.????? Hot Air Leaks to be rectified, deflector plate to be enhanced to stop mixing.
B.????? LO Related:
o?? LO Cooler Undersized to be rerated / higher capacity procured (Tin > 48).
o?? O&M to ensure that for LO Tin always < 48 Deg C. (CW, CT etc)
o?? LO Grade to be referred to OEM (Hot air Temp. > Flash point of Oil.)
?
C.????? Opportunity S/D Checks:
o?? All Labyrinth / Oil Seals to be checked and cleaned
o?? All bearing especially Expander side bottom half also to be checked.
o?? Rotor Balancing & alignment (if rotor removal planned in opportunity Sd.)
?
D.???? CFD/ Dynamic Simulation + Barring (For thermal stability / Bow)
E.????? SOP: To check for step wise coast down and ramp up wrt thermal bow.
F.?????? Spares – Spare rotor refurbished as well as recommended spares in stock.
Trends show direct correlation between vibrations and LO Temperature. Pls note the design operating temperature of LO (cooler outlet) as per data sheet is max of 48 Degrees But site operating temperatures are higher.
Pls note that after 48 degrees for every 1 degree increase in temperature the viscosity of the oil decreases by 5-10%. This is not linear but logarithmic. The LO Cooler is undersized and needs to be corrected.
The leaking hot air leak is contacting with Oil and carburising the oil and these carbon particles are one of the main source of soft rubbing.
B.????? Leak to be arrested or leak completely prevented from contact with Oil particularly at the Laby seals (where the probability of contact is more due to close clearances).
Oil grade to be upgraded (flash point of oil currently is less than the hot air temperature) in consultation with OEM.
Vibration Peak Transients during slow roll during coast down or ramp-up:
A. For the vibration transients at slow roll, the probable cause is the rubbing / unbalance due to temporary thermal sag. Also, the same could be due to operating parameters / temperatures gradients and the lower RPM and improper / uneven clearances). Direct correlation observed between suction air temperature and vibrations.
B. The expander has overhung impeller there is no NDE bearing on the expander & no barring device for barring when the unit is on standby / shutdown.
Dynamic simulation of the power train / CFD study to be done by CTS / CoE – Tech in coordination with OEM. Study to include thermal stability in IOW as well as transients include step wise ramp up and coast down.
A snapshot of the 5 WHY / Why Tree Analysis is given below:
Solicit comments or opinions or any further suggestions on the topic.
Nurturing Your Productivity || Maintenance & Reliability Enthusiast || Project Management Aspirant || Data Analyst
1 个月Thanks For sharing this. I have learned a lot from this.
Senior Rotary Technician
11 个月Thanks for posting
Deputy General Manager (Maintenance) ISO CAT IV vibration analyst, Asset Reliability Practitioner Cat I at Vizag Steel
11 个月Sai Vamsidhar Bontha Interesting case! thanks for sharing!
Really insightful share on vibration transients! Reminds me of Tesla's belief in the importance of understanding our universe's vibrations for innovation. Let's keep pushing boundaries together! ?? #innovation #teamwork
Asset Health | Bad Actor Management | Reliability KPIs | Condition Monitoring | Python Beginner
11 个月Nice analysis on the machine train vibration. My two cents: 1. Checking for vibration transmitter precision (forward or reverse – comparison of direction of vibration to direction of shaft rotation) can confirm a rub due to oil carbonisation. 2. 1X Phase plot can confirm if it is a temporary rub as well. During a rubbing cycle, the phase will change but will return to its original value once the rubbing action is completed. 3. Look for rubs on the end where there is less vibration when the machine is opened up for inspection. When the rub creates a localized hot spot leading to a rotor bow, the vibration is generally seen on the other end due to cantilever effect. 4. Building a phase plot for the entire train will show how each shaft is aligned with respect to others. This will indicate high stress areas that are most vulnerable for rubbing scenarios. 5. Though deflector plates are quite useful to prevent carbonization, nitrogen purging (blanketing) in multiple locations (as opposed to just 2 locations) radially prevents oil contacting hot air (or steam in case of steam turbines) more effectively.