Eng-Tips is the largest engineering community on the Internet
Intelligent Work Forums for Engineering Professionals
Compressor design
- Thread starter AnhsirkT
- Start date
1- The question should not be asked in those terms. It boils down to the bearing span (L) and average shaft diameter under impellers (D). A rule of thumbs is to have a stiffness ratio (L/D) lower than 10. But this a very basic screening. A selection/design is validated when it passes Level I lateral analysis screening as per the requirements of API 617. Under certain conditions, a Level II analysis is required. So there is no firm rule.
Anyhow, in the particular case of this machine, the no. of impellers of 5 or 6 is OK, the stringency would come from bearing span increase due to the injection nozzles. But to my experience this is in chartered area.
2- Impeller tip speed of 300 m/s at 100% speed should be fine as a rule of thumb. The type of driver would matter as it would define the over speed test value and impact stress level and maximum yield strength utilization. But again 300 m/s is fine.
If there is a risk of sulfide stress cracking speed limit would have been much less, but I understand this risk is not present.
3- No need to assume this, the head would be resulting. It would be a good practice to have the head relatively equally balanced between impellers across stages. In your case, it should be something like ~35 kJ/kg for 1st and 3rd stage and around half that value for 2nd stage.
Thus my suggestion to balance the load between 1st and 2nd stage, but again it is advisable but not a must.
If you plan an escape, you must succeed as if you fail, you will be punished for trying. Never say or write down your plan. Heart is the only place where secrecy is granted.
Anyhow, in the particular case of this machine, the no. of impellers of 5 or 6 is OK, the stringency would come from bearing span increase due to the injection nozzles. But to my experience this is in chartered area.
2- Impeller tip speed of 300 m/s at 100% speed should be fine as a rule of thumb. The type of driver would matter as it would define the over speed test value and impact stress level and maximum yield strength utilization. But again 300 m/s is fine.
If there is a risk of sulfide stress cracking speed limit would have been much less, but I understand this risk is not present.
3- No need to assume this, the head would be resulting. It would be a good practice to have the head relatively equally balanced between impellers across stages. In your case, it should be something like ~35 kJ/kg for 1st and 3rd stage and around half that value for 2nd stage.
Thus my suggestion to balance the load between 1st and 2nd stage, but again it is advisable but not a must.
If you plan an escape, you must succeed as if you fail, you will be punished for trying. Never say or write down your plan. Heart is the only place where secrecy is granted.
georgeverghese
Chemical
If 2S fresh feed is saturated at dewpoint, then there may be some liquid droplets in it. It will take a few seconds for these droplets to vaporise after mixing with 1D superheated gas. This vaporisation time can only be enabled in the 2S drum. Inline mixing in the pipe ( without drum) wont enable sufficent vaporisation time for these droplets.
Perry 6th edn says max is about 8-9 impellers per stage, Mach 0.8 max permissible tip speed, and max 10e3 ftlbf/lb (or 30e3Nm/kg) polytropic head per stage.
Perry 6th edn says max is about 8-9 impellers per stage, Mach 0.8 max permissible tip speed, and max 10e3 ftlbf/lb (or 30e3Nm/kg) polytropic head per stage.
If 2S fresh feed is saturated at dewpoint, then there may be some liquid droplets in it. It will take a few seconds for these droplets to vaporise after mixing with 1D superheated gas. This vaporisation time can only be enabled in the 2S drum. Inline mixing in the pipe ( without drum) wont enable sufficent vaporisation time for these droplets.
Excellent piece of input for the OP and as general information well.
Considering single casing solution, mixing would occur inside the compressor (there are lateral or tangential injections designed for such purpose and which are common for refrigerant applications). The question is, would a knock out drum be required on injection line if the side stream is pure Ethylene and it is superheated?
Perry 6th edn says max is about 8-9 impellers per stage, Mach 0.8 max permissible tip speed, and max 10e3 ftlbf/lb (or 30e3Nm/kg) polytropic head per stage.
8-9 impellers can be common but in real life there are plenty of particular situations and I am afraid it has to be case by case. Some manufacturers can routinely fit up to 11 impellers in a casing (e.g., Dresser-Rand); it may happen that 6-impeller casings are border line rotor-dynamically, I've seen it. Many factors enter into play.
The figure of Mach 0.8 refers to peripheral Mach number which is not the real Mach number (the one that refers to absolute exit velocity). So for refrigerant applications in particular, and more generally for heavy gas service (CO2, Propane, Air, etc.), there are special impellers that can exceed a peripheral number of 1.1 ; this is completely normal for these applications. These impellers are usually put at inlet (so it is generally the 1st and even 2nd impeller of the section, afterward the Mach number reduces to more conventional value, that is ~0,8 or less).
All though there are typical figures for polytropic head per stage, and it is always good to have these in mind, polytropic head per stage can here exceed 40~50 kJ/kg ; in fact it depends in particular on the maximum peripheral speed the impeller can spin at and how aggressive we want to be in the design. Anyway, please bear in mind that the above remarks are stated with all due respect to the book of course, in fact these are inputs coming from practical experience.
If you plan an escape, you must succeed as if you fail, you will be punished for trying. Never say or write down your plan. Heart is the only place where secrecy is granted.
Excellent piece of input for the OP and as general information well.
Considering single casing solution, mixing would occur inside the compressor (there are lateral or tangential injections designed for such purpose and which are common for refrigerant applications). The question is, would a knock out drum be required on injection line if the side stream is pure Ethylene and it is superheated?
Perry 6th edn says max is about 8-9 impellers per stage, Mach 0.8 max permissible tip speed, and max 10e3 ftlbf/lb (or 30e3Nm/kg) polytropic head per stage.
8-9 impellers can be common but in real life there are plenty of particular situations and I am afraid it has to be case by case. Some manufacturers can routinely fit up to 11 impellers in a casing (e.g., Dresser-Rand); it may happen that 6-impeller casings are border line rotor-dynamically, I've seen it. Many factors enter into play.
The figure of Mach 0.8 refers to peripheral Mach number which is not the real Mach number (the one that refers to absolute exit velocity). So for refrigerant applications in particular, and more generally for heavy gas service (CO2, Propane, Air, etc.), there are special impellers that can exceed a peripheral number of 1.1 ; this is completely normal for these applications. These impellers are usually put at inlet (so it is generally the 1st and even 2nd impeller of the section, afterward the Mach number reduces to more conventional value, that is ~0,8 or less).
All though there are typical figures for polytropic head per stage, and it is always good to have these in mind, polytropic head per stage can here exceed 40~50 kJ/kg ; in fact it depends in particular on the maximum peripheral speed the impeller can spin at and how aggressive we want to be in the design. Anyway, please bear in mind that the above remarks are stated with all due respect to the book of course, in fact these are inputs coming from practical experience.
If you plan an escape, you must succeed as if you fail, you will be punished for trying. Never say or write down your plan. Heart is the only place where secrecy is granted.
Mixed refrigerant example of high head impellers and max. Mach no. that can be attained (see pages 2 & 3 /12)
If you plan an escape, you must succeed as if you fail, you will be punished for trying. Never say or write down your plan. Heart is the only place where secrecy is granted.
If you plan an escape, you must succeed as if you fail, you will be punished for trying. Never say or write down your plan. Heart is the only place where secrecy is granted.
- Thread starter
- #45
- Thread starter
- #46
Considering single casing solution, mixing would occur inside the compressor (there are lateral or tangential injections designed for such purpose and which are common for refrigerant applications). The question is, would a knock out drum be required on injection line if the side stream is pure Ethylene and it is superheated?
KO drum required bcz ethylene vapor coming from evaporated can come with some liquid or droplets
Which will separate in ko drum
KO drum required bcz ethylene vapor coming from evaporated can come with some liquid or droplets
Which will separate in ko drum
georgeverghese
Chemical
You said there is a quench facility for 1D superheated gas, which you describe as follows:
"For quench control
There are TIC to maintain the drum temperature
That valve will open and put some ethylene liquid in drum and liquid will flash and cool down the temp of drum"
This scheme will work only when 1D gas goes into 2S drum. This is because, when 2S drum fresh feed is saturated gas only, no matter how much liquid C2= you inject into the drum, the temp of the gas will remain at sat temp.
Error in my previous post, the values I quoted for compressor design are only partially from Perry, other bits are from some other compressor design text which I dont recall the title for now
"For quench control
There are TIC to maintain the drum temperature
That valve will open and put some ethylene liquid in drum and liquid will flash and cool down the temp of drum"
This scheme will work only when 1D gas goes into 2S drum. This is because, when 2S drum fresh feed is saturated gas only, no matter how much liquid C2= you inject into the drum, the temp of the gas will remain at sat temp.
Error in my previous post, the values I quoted for compressor design are only partially from Perry, other bits are from some other compressor design text which I dont recall the title for now
- Thread starter
- #48
georgeverghese
Chemical
What is the actual suction temp to stage 2S compressor, is it superheated? - from your description, it must be above dewpoint, but that is not what you say. It is difficult to pick out all this from you poor description, no PFS, no stream condition tables.
- Thread starter
- #50
- Thread starter
- #51
georgeverghese
Chemical
Okay, given that 2S compressor suction T is greater than dewpoint, I would guess that this scheme you suggest for the 1D stream to bypass the 2S drum is possible.
- Thread starter
- #53
georgeverghese
Chemical
There may be some minor entrainment of liquid drops, but this would be within the tolerable limits for compressor feed provided
a) the KO drum is now sized as a vapor liquid vertical demister type separator
b) some surge margin is allocated for flows, given this liquid C2= flow is on temp control - suggest a 30% margin on top of max design flow for quench liq C2=
Also do your best to improve control loop response for this TIC quench control loop.
a) the KO drum is now sized as a vapor liquid vertical demister type separator
b) some surge margin is allocated for flows, given this liquid C2= flow is on temp control - suggest a 30% margin on top of max design flow for quench liq C2=
Also do your best to improve control loop response for this TIC quench control loop.
georgeverghese
Chemical
Your diagram design 1 shows two separate inlets for 1D gas and for fresh feed. A mixed common inlet will require less liq C2= than if you have separate inlets or if 1D gas bypasses the 2S drum. Even less liq C2= will be required if you have a quench spray tower/ column upstream of the 2S drum.
- Thread starter
- #56
georgeverghese
Chemical
In design 1, quench liq C2 is not mixed with hot 1D gas. Only the flash vapor from quench liq C2 contacts with 1D gas. If you can enable a common inlet with a mixing time of say 3-5seconds(am guessing mixing time here)or better yet, with a static mixer, then you will need less quench liq C2.
- Thread starter
- #58
- Status
Similar threads
- Locked
- Question