GrantB
Junior Member
Joined: February 2012
Posts: 61
|
Post by GrantB on Nov 20, 2013 0:57:50 GMT -5
I was looking into automotive turbochargers and i noticed that while they don't have a diffuser there is usually an area where they could be. I was under the impression that the lack of a diffuser was that they narrowed the compressor map and would also make manufacturing more difficult. If this is true what is the point of this area being so wide?
Im sure if this area could be made narrower the OEM's would do this to make a smaller unit. I obviously am missing something here and google searches are not bringing much up.
Thanks, Grant
|
|
|
Post by racket on Nov 20, 2013 3:08:45 GMT -5
Hi Grant
Auto turbos have vaneless diffusers , this allows a wider flow range as theres no need to maintain a favourable air angle onto a diffuser vane .
Theres a need for radial distance after the comp wheel exit to slow the air down and turn velocity energy into static pressure rise before its collected in the scroll .
Cheers John
|
|
GrantB
Junior Member
Joined: February 2012
Posts: 61
|
Post by GrantB on Nov 20, 2013 12:23:57 GMT -5
Thanks John!
I can sorta see what you are talking about now. Do you also think it might help with containment in the event of a compressor wheel bursting?
Alot of people try and make a large amount of power with small engines and they seem to always be running along the surge line. Do you think a diffuser might help move the surge line over a little? In motorsports a wide airflow range is good but everything is relative. Do you think experimenting with a diffuser might e something to look into?
|
|
|
Post by racket on Nov 20, 2013 23:31:07 GMT -5
Hi Grant
Surge "generally" has to do with compressor wheels rather than diffusers , surge occurs when the amount of air going into the wheel is way below the amount required to produce tolerable off design angles of attack on the inducer blades .
If you do a velocity triangle for an inducer you'll notice things only "align" over a small variation of air/tip velocities , once the angle variation gets too great the air breaks away from the blading , same as an aircraft wing stalls at ~15 degrees angle of attack , but produces great lift at a few degrees .
The addition of a vaned diffuser won't change the inducer stalling problems , but it will add an additional level of potential stall problem, thats why most auto turbos don't have them , ,.............large diesel turbos which operate steady state or over a limited range will often have vaned diffusers in their turbo.
Our DIY turbine engines operate over a limited flow so a vaned diffuser could be beneficial in producing a few percentage points improvement in compressor efficiency .
Cheers John
|
|
|
Post by finiteparts on Nov 21, 2013 23:02:17 GMT -5
Hi Grant, To expand on the points that John was making, it must be remembered that the compressor and diffuser are parts of a system and mutually influence each other. Compressor-diffuser interaction is a hot topic in centrifugal compressor research. The flow within the centrifugal compressor blading is highly 3-dimensional in nature and there are many potential sources of loss. Compressor surge can occur when the flow within any part of the compressor system breaks down. Flow separations in downstream components can negatively impact the flow/pressure fields within the upstream components leading to complex flow breakdowns throughout the system. As John said, the vaneless diffuser used in most automotive and diesel turbochargers has the advantage of a fairly wide operational flow capacity, which when coupled with a stable compressor design can lead to a very wide compressor map. It must be stressed that the vast majority of compressor maps available are really compressor-diffuser system maps. The addition of vanes or vane-island (wedge shaped) into the diffuser allows the flow to be turned in such a way as to give a controlled rate of diffusion for a smaller radial dimension that what could be achieved in a vaneless space. The problem that this introduces is that once you are operating away from the design condition, there can be flow separations created due to excessive turning or strong cross-passage secondary flows that create boundary layer trips. Additionally, the vanes create flow blockages that did not exist in the vaneless diffuser, which can create regions of high flow velocities, even to the point of choking the flow. With choked flow or even near-choked flow you can have shock wave formation that creates large irreversible flow losses and rapid flow field break-downs. Remember, the compressor is a very hard design problem because of the physics involved. The engineering field terms the compressor and diffuser as an adverse pressure gradient because you are trying to force the fluid into a state that it doesn't desire to be in. It's like pushing a heavy ball up a hill. As the slope of the hill (i.e. the gradient) gets steeper and steeper, there comes a point when you just can't push it up any farther and you "stall". Your comment on the small engines at high power running along the surge line is a similar situation. A small engine has a certain mass flow capability based on its displacement, volumetric efficiency, etc...so when the turbo really gets spooled up and is pumping against the back pressure of the engine, it is really having to push the air into the engine. At a certain point, somewhere in the system the flow stops going forward, maybe due to a flow separation in the diffuser. Now, suddenly, the effective area that the flow sees is reduced. Well, to keep the same mass of air moving forward through a smaller area that the flow now sees, the pushing pressure of the system has to go up...but it can't because the flow within the system is already breaking down...thus the stall develops. So the big question arises...How can we make the compressor system maps wider? One of the most popular means is to use a bleed slot in the compressor shroud area....Holset calls this MWE (map width enhancement)...Here is a nice paper on their design. www.cumminsturbotechnologies.com/CTT/CTTContent/CTTUS/SiteContent/en/BinaryAsset/PDFs/Downloads/Exp_numerical_analysis_classical_bleed_slot_systemfor_turbo_compressor_Techpaper.pdfAdditionally, the design of the compressor can be modified. Part span blading can be added to reduce the individual blade loading. The exit angle of the blading can be swept backward to allow some of the diffusion to actually take place in the rotor blade passages, which is less severe since the wall shear is greatly reduced. The blading can be leaned forward at the exducer to control the secondary flows created within the passage. The diffuser can be modified to maintain boundary layer/flow attachment, if the source of the flow tripping can be identified. And finally, a common approach in larger centrifugal compressors that are used for such tasks as gas pipelines,process compressors and even some aircraft or tank engines, the variable inlet guide vanes and/or the variable diffuser vanes. The variable inlet guide vanes (VIGV) allow a better matching of the inlet flow angle relative to the inducer leading edge. Just like John was saying, leading edge flow separation will occur when the incoming flow meets the leading edge at an incidence angle above some allowable, just like a aircraft wing stalling. Similarly, the variable diffuser blading can achieve the same results for the off-incidence flows from the compressor discharge flow entering the diffuser section. At ASME Turbo Expo this year, there was a presentation by maybe Concepts NRC, where they showed that the VIGV gave the most bang for your buck in expanding the compressor systems map width. I'll see if I can find it and if it is publicly available, I will post a link. And I agree with John, a vaned diffuser might be a good idea for the full power engines since they are operating over a small range and it is likely that being able to reach a specific pressure recovery at a smaller diameter than what is required with a vaneless set-up would be desirable, especially for those of us with limited lathe swings! ~ Chris
|
|