Semi-open or enclosed - choose thy weapon

[HomeMadeSin]

Alright, I hope this is as much fun as 2/4 pole.

For general and chemical service:

Personally, I'm a fan of enclosed impellers and keyed to shaft drive. However, the majority of the pumps mfgs seem to prefer semi-open, threaded against rotation types.

I can appreciate the 5x wearing surface characteristics of semi-opens, but the biggest selling point seems to be it is cheaper to manufacture and therefore sell.

It is generally accepted that enclosed impellers suffer less recirc loss and therefore generate more head and are more efficient. But they cost more to produce.

So aside from waxing peotic about either type, what is the real world view on these impellers? Has the general trend to semi-open impellers been market driven or manufacturer driven. Is a dirty, solid-laden service more typical than clean clear one?

On a related note, threaded engagements seem to be cheaper/easier for mfgs as well, but prone to the dreaded reverse rotation issue (great for parts sales). It is more pronounced on close-coupled pumps (my interest).

 

[PUMPDESIGNER]

In the industries I serve: Irrigation, water storage and transfer, pressure boost to buildings,

We have not seen any open impellers, nor any trend to that.

Screwed? Would that be small pumps like I have on my swimming pool, I know that is a screwed impeller attachment. If so we would not know, we seldom deal with less than 5 hp.

PUMPDESIGNER

 

[HomeMadeSin]

Sorry, I worked in the ANSI-related field most of the time. For example: Goulds 3196: threaded semi-open. Durco Mark III: reverse vane, threaded. Ingersoll-Rand HOC: semi-open, threaded. All of the clones follow suit obviously, and the magnetic, non-metallic, vertical and other variants on their own designs are the same. Fybroc, also. I guess my chemical background shows, huh?

Pressure boost and water storage and xfer I would assume would be multistage (enclosed and keyed or similar), split case (enclosed and keyed) or general water service pumps that are overhung and designed a while back (i.e. bronze fitted enclosed impellers in CI cases).

I'm glad to see no trend to semi-open impellers in your fields. We don't get into much of those fields, with the execption of water xfer.

But yes, similar to swimming pool pumps which mate to either 56J or 56Y frames. I've got to be careful in mixing long-coupled pumps like the ones I listed to close-coupled, but there aren't many close-coupled non-metallics larger than swimming pool pumps. Many of the ones that do exist use intermediate shafts, which I do not advocate.

 

[PUMPDESIGNER]

Stopped using Goulds about 2 years ago.
Never used any 3196.
Went to their site, looked at 3196.
Surely you are not saying those pumps are threaded impellers, tell me it ain't so. And semi-open?
If so, tell me why again.

PUMPDESIGNER

 

[HomeMadeSin]

The 3196 uses semi-open impellers and are threaded against rotation. So the HOC1/2/3 from IR (now Flowserve) and the Durco (now Flowserve also) Mark III, even though it uses a reverse vane impeller (it has a front shroud and no back shroud - rather unique).

When I worked at IDP, now Flowserve, the standard asnwer was that threaded engagement allowed higher torque. I think the truth is that is is more cost effective to machine threads (both int/ext) than bore, ream and broach for a keyed fit. Also, with metallurgies ranging from DI to TI, it simpifies the wetted parts (keys for example).

On the semi-open issue, again the reasoning was based on the wear surface. The examples I was given was like 5x (the front vane edge of the semi-open impeller collectively exceeded the wear ring surface by 5x, thereby decreasing the rate of efficiency loss due to opening clearances over time. In order to re-capture the desired clearance (usually 0.010&quot, one would have to adjust the whole rotating assembly as needed. Obviously, this would be difficult on a close-coupled pump and various mechanical configurations have been designed to accomplish this.

The real reason, I believe, semi-open impellers are "standard" in the ASNI/Chemical market is that it is cheaper to cast, machine and therefore sell pure and simple. However, I was hoping to ferret out the real reasons via this post.

And because most pump manufacturers use as many parts across as wide a product range as possible, you'll find the same impellers (semi-open, threaded) used on the equivalent mag-drive chemical pumps, vertical pumps, etc. Bascially , the Ansi B73.1/2/3/4/5 series of pumps (which cover horizontal/vertical in-line/mag-drive/?/vertical sump, I think).

 

[PUMPDESIGNER]

HomeMadeSin,
Interesting stuff you bring up.
I will attempt to be interesting to you.
Funny thing about that ANSI market. My feeling is that the business plan for some of the big boys involves just plain too much marketing money and overhead. Quality suffers.
Some companies can produce incredibly high quality but have little money for marketing. Take Griswold as an example. Company is owned by engineers. Their quality I understand on these ANSI pumps is superior, lead times are EXTREMELY good, but prices are comparatively cheap, reason being that there is no big time marketing machine.

Just food for thought and highly subjective I know, but if true it would explain a lot.

Do not work for Griswold, but have worked with them for many many years, and I just personally like the way they run their company. No suit and ties, no salesmen with alligator shoes, but good engineers.

PUMPDESIGNER

 

[vanstoja]

Semi-open, without front shroud impellers appear to have gravitated into the centrifugal pump world from the centrifugal compressor world where semi-open impellers have been used for a long time. I would agree with homemadesin that the main motive for semi-open pump pump impellers is ease of manufacture by numerically controlled machining(NCM) using 5-axis milling machines pioneered by compressor design specialists like Northern Research Engineering Corp.(NREC). Chances are that with Computational Fluid Dynamics (CFD) design methods coming into ever greater use for centrifugal pumps, there will be more impetus to use NCM manufacture of pump impellers and so an increasing use of semi-open impellers. Another driving force for NCM is the improved surface finishes in the flow passages with 5-axis milling which can improve hydraulic efficiency.
A major drawback of semi-open, front shroudless impellers is the secondary flow across the top of the blades which creates some serious vorticity interactions with the main channel flow (with cavitation and hydraulic instability implications). There is much discussion in the literature of this particular problem. Also cantilevered-blade vibration without a front shroud becomes a much greater problem. One solution to these problems is to use 5-axis milling NCM to machine the blade passages and then weld or braze a front shroud onto the blade tips. This takes a little development work but is known to have been successfully accomplished.

 

[Artisi]

Just to add a little more and maybe open up the discussion on open verses closed impeller pumps.
I have used / sold both open and closed impeller pumps and it always comes down to " horses for courses" - what is the duty, what is the best pump for the application.
If you are pumping "clean liquor" then use closed impeller pumps.
If pumping other than clean or liquors containing solids or entrained gases / air, ie , paper stock, some sewage applications, stringy material, large diameter 'soft" solids etc, use open impeler pumps.
If pumping hard solids, ie mining, dredging etc use enclosed hard metal or rubber lined pumps - althought this has can have its exceptions as well such as froth pumping from flotation cells.

And what is wrong with screwed on impellers anyway - for industrial aplications screwed on impellers do not suffer the problem of liquor migrating into the key / shaft interface, impeller fretting on unstable operation conditions etc.

I don't believe that major pump manufactures who are manufacturing quality products just come up if an idea and put it into practice - well at least not the ones I have worked for over 25 years. All or any changes that I have been involved with have been thought through from a maunufacturing / cost / end user benefit. Why put something into the field that can be a problem to you or your customer -we already get enough application problems to handle without making it worse for ourselves with poorly thought out or badly designed equipment plus it doesn't enhance your repeat business opportunities.

As for efficiency - I don't really see a great difference and would tend to think that on open impeller pumps we have the opportunity to supply a much higher finish to the hydraulic wetted areas than on fully enclosed impellers, this must result in higher efficiency - also leakage across the face of an open impeller with correct clearance is usually less of a problem than with closed - at least you can make infield adjustment and not have to dismantle and replace/ remachine wear ring etc.

 

[HomeMadeSin]

Good input Artisi. However, I would like to point out that virtually nobody would machine the inner casing contour to closely match the impeller. At least I haven't seen that - too costly. Therefore, you wind up with a machined vane edge running against a cast (sometimes precision cast) casing. If they did, I could see your point.

Also, I didn't bring it up and it wasn't by others but a semi-open impeller has higher axial loads. I don't know the factor over enclosed, but just something to bring up.

I understand the threaded against rotation argument to an extent, and I think the most complaints are against the dreaded reverse rotation disaster. And although you are supposed to un-coupled the driver and check for rotation, inevitably someone forgets or in the case of close-coupled pumps (my primary interest) it isn't feasbile to disassemble the pump just to check rotation. By the time you "bump it" it is too late.

 

[PUMPDESIGNER]

Artisi,
I enjoyed your comments very much. You appear to have a very balanced viewpoint, and I believe that in your case the correct decisions have been made for the most part, assuming of course that after the information goes upstairs for the final decision it comes back to your liking.


PUMPDESIGNER

 

[jet1749]

My own experience with semi open type impellers has been to remove several recently installed Flowserve CPX chemical pumps (ISO) and replace them with Durco Chemstars also made by Flowserve. These have enclosed impellers (although still screwed on)the reason for this was that the CPX pump bowls were precision machined with a complex compound curve to match the impeller, we could not repair or replicate. This meant horrendous replacement costs and delivery times. The Durco was far simpler, quicker and cheaper to repair. We now only fit new semi open impeller type pumps where there is an over-riding and realizable benefit.

 

[HomeMadeSin]

jet1749 - The CPX series was started within the Dresser half of the Ingersoll-Dresser family, well before Flowserve came knocking. It was intended to be a world-wide chemical pump (for both ISO and ANSI markets), but never gained ground here in the US. Therefore, I'm guessing your outside the US. The casings are awesome looking, with the machined contour to match the impeller (not a straight cut), and I could see your point. On the other side, making a pump so that it could be easily replicated is NOT of interest to a manufacturer.
Also, the complex contour of the CPX is unusual, as most chemical pumps follow a flat or angled straight line taper.

It seems the main thing about semi-open or enclosed is the control of efficiency. In an adjustable semi-open design, the efficiency could be kept as close to stock as possible with minimal maintenance, whereas an enclosed impeller would require the rings (if fitted) to be replaced.