Understand Centrifugal Compressor, Equipment Interaction
Kumares Majumdar, P. Eng
Executive Director, Process Solutions Inc., Calgary, Canada
Referring to one of my old articles, Understand Centrifugal Compressor, Equipment Interaction that appeared in Hydrocarbon Processing Journal, Aug. ’99 issue, one of my good readers contacted me recently (after 19 years !) with the following query (edited): (https://www.hydrocarbonprocessing.com/magazine/1999/august-1999/special-report-fluid-flow/understand-centrifugal-compressor-equipment-interaction )
Reader's query:
I have a doubt on scenario 2.3 of this article , Gas demand decreases at the consumer end: You stated, “The system curve moves to the left due to throttling of the consumer valves. The compressor controller, using suction-discharge pressure and flow signals, would start to send a signal to the driver to slow down.”
However, I thought since the consumer valve is throttled, the discharge as well as the suction pressure will increase. In that case, the compressor controller should increase the driver rpm, not reduce, so that the suction pressure would come down to the initial pressure. Please correct me if my understanding is wrong.
My response to the Reader
I am happy to explain based on my limited understanding of centrifugal compressors and fluid-flow systems albeit my drooping acumen in the last two decades since the paper was published. The purpose of my posting the reply here is to share with LI fraternity and invite discussions/comments/corrections to enrich myself. The figures are from my archive, originally used in another article, hence may not depict all details:
1. First of all, contrary to your (reader's) expectation, under normal operation as long as the incoming flowrate and the delivered flowrate to and from the compressor station are equal, Psucn won’t change with any downstream change. Psucn depends on the upstream/source pressure whereas Pdisc (developed Head) depends on the downstream/supply system.
2. In Fig.-1 see what happens when a compressor is driven by a single speed driver at N1 rpm delivering Q1 flow. The System Curve (Q vs ΔP in downstream line, valves etc.) intersects compressor’s Operating Curve (Head, H converted to ΔP vs Q) at point A (Pdisc). The Head at any flow rate is not affected by Psucn. When a consumer turns down the flow by throttling through a manual or a flow-control valve, the Sys curve shifts to the left due to additional pressure drop. The two curves now meet at B and the flow reduces to Q2. Note, Pdisc is increased by BA or ΔP although the flow rate is reduced. This added ΔP is actually an energy loss which could be avoided had a variable speed driver been used.
3. In Fig.-2, the same compressor is driven by a variable speed driver with pipeline inlet pressure floating (usually, a downstream pressure controller is provided to keep the line packed and give min. Pdisc to the compressor) . As consumer’s valve is throttled, manually or automatically, the compressor control system momentarily senses a rise in discharge pressure, Sys curve shifts to the left as in Fig-2. This triggers the driver to slow down, say to N3. But it cannot right away slow down as the flow tends to drop below Q2. Any drop in speed at this point will no longer meet the required new flowrate, Q2.
4. The net effect is, both flow and pressure would reduce, unlike in single speed compressor (Fig-1) where Pdisc actually increased. The reverse will happen if the consumer opens up the valve to get more flow.
5. For a given downstream system (lines, valves etc.), the line ΔP should actually drop when the flowrate is reduced. If the Compressor Controller (CC) is optimally configured, the speed will drop automatically from N1 through N3 down to N4 per compressor’s family of curves fed into the CC database. But to achieve Q2 optimally at the lower speed, it would also need to manipulate the Sys curve, i.e. readjust consumer's controller. If consumer's valve is automatic, it will then adjust (open up) to find its position corresponding to Q2. The Sys curve then start to shift to the right until it stabilizes at the lower-speed, N4 at point B”. If the valve is manual, to get energy saving when the compressor settles down at the lower speed, the operator should gradually open the throttling valve until the flowrate is Q2 at nearly the same pressure as the initial pressure (before the flow reduction was initiated).
6. In case the CC is not fully automatic, the compressor operator is expected to adjust the speed until the required flow rate is attained at the minimum discharge pressure. In an all-automated system installed at both ends, these actions would take place automatically in quick sequence. In any case, the compressor would still run albeit not optimally, even if these steps are not followed closely.
7. Hypothetically, if the compressor had sped up, as the reader thought, (it is possible but not likely in a well-configured design), the condition would have been as shown in Fig.-3. It could hardly achieve the desired flowrate, Q2 and capture the energy saving potential.
--
8 个月Sir, I studied your article and thanks a lot . If possible, I wanna understand and learn more about throughout control . So please let me know .Is there any article about this topic and other characteristics of centrifugal compressor such as velocity diagram etc.
Assistant Manager at Indian Oil Corporation Limited
4 年Dear Sir, In your article "Understand centrifugal compressor, equipment interaction" where you have explained about system curve.? It is written there is static pressure drop which is the pressure difference between point B & C. I think it should be pressure difference between point A & C. Because even due to RPM increase point B pressure will rise and it will change static pressure drop as per your definition. While there won't be any change is suction pressure (Point A) changes. Similarly in scenario 2.4. static pressure difference is taken as P2-P3. My question is if P3 is taken as destination pressure (Fig. 5.) then its downstream control V/V (dotted boxed V/V) will be a part of system or not?. I think if it is a part of system curve then destination pressure should be its downstream pressure may be P4.? My reason for asking this question is. because?in scenario 2.3. system curve is moved left due to throttling of the consumer valve. I think that v/v is not a part of system because P3 is taken as destination pressure (which is consumer v/v U/S pressure). In this case system curve simply move up due to increase in P3. Please provide me your insight on these doubt.
Executive Director, Process Solutions Inc., Calgary, Canada
6 年Thank you, Mr Sarathy, for reading and commenting on my recent article in LI. Though I am not sure I understood all your long comments (truncated) and also if you are asking me any question or telling me something that I missed. Anyway, I am trying to address some general misconceptions. 1. The article here is in response to a reader’s specific query what should happen to a variable speed compressor’s optg curves when consumer/s turndown consumption by closing/throttling d/s valves (ref to the article). This paper does not aim to elaborate how to run or control a VSD compressor, though it touches upon those issues briefly. The Schematic diagram depicts the basic controls and signals quite clearly. Please refer to my article in HP, Aug 99, Compressor -System (d/s dP vs Q) Interaction during different optg scenarios. 2. It is a common knowledge that any compressor is controlled by Psucn. and Pdisch signals and flow, Q signal. It is not right to think a compr speed is controlled only by Psucn. In fact, a high Psucn cannot speed up a compr (pump more gas) unless Pd is low enough to take up the extra gas. In this case as the consumer is reducing the flowrate, P in d/s line is going to increase - there is no question of pumping in more gas into the line. The right interpretation is, in all cases, Ps and Pd work through PT/PC and a low signal selector device to control compressor speed, SP (see the schematic). In modern compr controls, the algorithm may be built in a PLC (compressor control module) which may be integrated with an independent Surge control module and both may run on a common PLC/DCS platform. 3. I am not sure if your statement “Primary control is by anti-surge line control...” is a right one. In fact, in any good design, this should never happen. The Surge Controller/ anti-surge valve, ASV is a safety system for protecting the machine from surging. This should never open during normal oprn to control a compressor unless the Q and P touch the Surge Control Line (SCL). The SCL is not the Surge Line, SL which should still be 10-15% away to the left. Under normal oprn, ASV or any other recycle valve, if provided, should not open at all. If it does, the system needs retuning/setting to avoid recycling and inefficiency, also unnecessary wear out of the important valve, like ASV. Hope, this would be of help.
Referring to the query posted by your querant.. My 2 Cents of Knowledge. The type of driver (Fixed/Variable) needs to be taken into account. Mode of control matters. Say primary control is purely by Antisurge line regulation, then... its as follows. 1. In case of a Variable speed driver, when suction pressure increases for a given discharge pressure, more flow gets in and the compressor speed increases. And the Operating point moves in the North East direction 2. When discharge side is affected, for example client throttling at his end, the antisurge valve opens and recycles discharge flow to the suction. Therefore the compressor speed decreases to avoid a surge & operating point moves North West to South West Direction (Closer to the surge line) In case mode of control of compressor speed is by using Discharge Pressure via a PT-PIC-SC, then it's an exception i.e., 1. In both individual cases of suction or discharge pressure increment, compressor speed always decreases.