Best way to control 120 starts per hour 4

[ccjersey]

I have several small single phase motors on pumps which start on average 100-120 times an hour over 2 shifts per day. These are controlled by a float which rises as milk fills a receiver. There is a solid state control which triggers a magnetic contactor starting these 1 hp motors across the line. On the whole these are very reliable, but things do wear out, especially centrifugal mechanisms and switches, not to mention contactors etc.

I was considering swapping to 3 phase motors to eliminate the capacitor/start switches without going to a PSC motor which has low starting torque and offers even less reliability than the capacitor start type in my experience. The OEM offers PSC motors and the standard capacitor start 56C motors outlast them 4 or 5 to one in our hands!

Should I just go with the tried and true across the line contactors, perhaps upsized one size or would a VFD offer a solid state start with increased reliability over the mechanical contactors. The liquid level control could probably control the VFD directly, eliminating the mechanical relays completely. There is little need for soft starting or variable speed operation, I'm just looking for the best reliability.

 

[edison123]

How about a solenoid valve on the pump discharge ? Or a lower speed motor ?

Muthu

http://www.edison.co.in

 

[zlatkodo]

One possibility is to rewind motors to three-phase winding.
Zlatkodo

 

[ScottyUK]

Definitely go 3-phase. 3-phase machines are far more robust, plus cheaper, smaller and easier to obtain.

Your start frequency shouldn't hurt the contactor but it might be a concern for the motor depending on how long it runs after each start. Frequent starts with a short running period after each start doesn't give the motor much chance to reject heat to atmosphere.

A VFD could be a good choice if this is a low static head application and the pump characteristic is amenable. Some pump applications are totally unsuited to VFD control, especially in a retrofit situation.

----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[waross]

Depending on the pump characteristics you may be able to replace the level switch with a level transmitter and let the system pump at a rate that matches the input flow.
Another option, particularly if proportional control is not suitable is to pump at a slower than needed speed and use the float switch to jump up to a second higher speed. This would give fewer cycles and may be more user friendly than a proportional control loop.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[TheBlacksmith]

I'm primarily mechanical, but like the others, I think slowing down the cycling; smaller pump, slower speed, or variable speed will be beneficial to any solution. Starting a motor on average every 30-40 seconds with very short run times just doesn't seem like a good idea.

 

[electricpete]

There are some guidelines for repetitive starting of NEMA motors at the link below, but I think it applies to 3-phase motors, not single phase.

http://www.joliettech.com/allowable_starts-intervals.htm

=====================================
(2B)+(2B)' ?

 

[potteryshard]

Make the milk receiver larger?

 

[DickDV]

Probably the simplest way to increase reliability without major changes would be to replace the motor with a three phase TENV motor and control it with an SCR softstarter.

The TENV motor will run coolest because it cools very well even when standing still. The softstarter eliminates the contactors and, at high cycle rates, will be more reliable. You may not even use the softstart feature. Choose a softstarter that can be dialed down to zero or near-zero start time and use it like a conventional starter. It will need to be sized large enough so the starting surge can be handled without tripping on overcurrent.

 

[ccjersey]

Thank you for all the ideas. I had thought about a couple of them.

One of the motors is a NEMA 56C frame, not a problem to get new 3 phase motors for those, already have 2 on hand. The others have a special shaft for the close coupled pump, not commonly available, but most likely 3 phase availible from the OEM (Boumatic).

If one were to let the pump run at a slow speed just to avoid stopping the motor, too slow to actually open the discharge check valve, it would most likely damage the milk (churn it to butter in a minor way). Pump works to move milk out to atmospheric pressure against a vacuum which exists in the reciever/pump area to pull the milk from the cows.

Running at a variable, generally lower, but still effective speed would be better than the present situation. Like Bill said, a more complex control logic involved. Does great things for the cooling, see below.

Could also do the 2 speed with the addition of another liquid level probe to stop the pump so it doesn't run dry (foaming, churning, etc). At the present time, the pump is controlled with an off delay control. The float rises, triggers the control and initiates a pumping cycle which continues a set time after the float falls.

Larger receiver....this is a retrofit....The equipment works pretty well as it is. One thing that is along this line is to raise the float in the receiver as high as possible. This does good things for the cycle time, but inhibits cooling on the milk as it flows through a heat exchanger in larger slugs and the % that stops in the HX and is fully cooled to the temperature of the cooling water is reduced. Not a big problem, it all works out to about the same net temperature in the tank at the end of the shift because the cooling is limited by the capacity of the ice builder/ice water making condensing unit outside.

Pump output is limited on the low side by washup requirements. It takes a much greater flow to keep up with the wash cycle. Perhaps another situation where a 2 speed setup might be useful. If one were to run the centrifugal pump with a TEFC motor slower with a VFD to avoid stops and starts, would the net heating be greater or lesser? Probably not answerable!

Amazingly the motors have never overheated until they "hang". The PSC (totally enclosed/not external fan cooled) motors do not survive being hung by bad bearing etc probably because they don't make enough noise for someone to notice and stop them in time. The capacitor start motors (TE/FC) have enough torque to keep starting even when something is going wrong and probably get noticed before they burn more often than not. There is a low level of ambient noise that obscures some of the motor noise one might normally hear. The cap start TEFC motors even do well without the external fan. The plastic fans seem to be one of the things that fail first.

Do you get the idea that these motors may be oversized for the actual work thay have to do, but that is why they last under this kind of service.

Thanks for all the ideas so far.

 

[jraef]

I agree, soft starters if you can't change anything else but the motor and power feed, or if you can modify the rest of the control circuit, VFDs that are set to modulate speed to match flow and reduce the on-off cycles altogether (PI or PID loop control). The nice thing about VFDs at this size is that you can leave the 1 phase feed the same and the VFDs will convert the 1 phase to 3 phase for the motors. To use soft starters you would have to change the entire power circuit to 3 phase.

I have seen this done in milkers elsewhere, they monitor the receiver level with an analog transducer and feed that into a PLC (or directly into a VFD) PID loop to have the VFD work to maintain a mid-point level in the receiver. Then the receiver only acts as a surge tank (and reduce turbulence on the transducer) with the VFD matching the average flow rate coming from the milker. MKS / Temposonics makes a very small magnetostrictive analog transducer almost specifically for this purpose, it's small enough to fit in a modified well on the receiver.

"If I had eight hours to chop down a tree, I'd spend six sharpening my axe." -- Abraham Lincoln
For the best use of Eng-Tips, please click here -> faq731-376

 

[Thealanator]

Questions/observations/suggestions:

Does one contactor control several motors?
Is this group of motors, controls, pumps, etc. in a big box that came from a manufacturer? I'm wondering if the proposed modifications might violate a UL listing, which can be a code violation.
Professional electrical engineering help might be in order.

 

[ccjersey]

A single receiver, single control and single motor go together, but are a field assembled-installed group.

No, no electrical or engineering codes, only has to comply with 3A sanitary standards and be safe for both people and cows.

 

[ccjersey]

The phase loss/overload detection capabilities of the VFD and soft start might be a big advantage over a contactor and overloads. I haven't had very good luck with small 3 phase motors running unattended. Seems like sooner rather than later the overloads don't trip in time to save the motor when a phase is lost.

We are on a well balanced (voltage)open delta 120/240 3 phase service, but motor amperages are almost never very closely balanced. Larger motors in the 5-20 hp range typically last a very long time, just not the small ones. Perhaps we just need better overload relays than the old bimetal types.

 

[jraef]

If you keep your existing system, you should at least switch to Solid State Overloads Relays, they are now much more reliable than electro-mechanical and provide true phase loss protection.

"If I had eight hours to chop down a tree, I'd spend six sharpening my axe." -- Abraham Lincoln
For the best use of Eng-Tips, please click here -> faq731-376

 

[bentov]

This is actually a very interesting liquid level application. The standard dairy equipment industry fix is the VFD (which as mentioned also solves the conversion to 3 phase problem). The main attraction of the VFD is to match pump speed/milk flow to the incoming stream, so that it's moved as slowly as possible through the water cooled heat exchanger on it's way to the refrigerated storage tank (significant savings thus possible on that refrigeration power bill, plus slower pumping minimizes the churning of the raw milk, which "beats the butterfat out of it" I'm told).

The typical probe is a stainless rod/tube with an external floating magnet donut that slides up and down with the liquid level, closing 2 or 3 reed switches inside the tube, which are simply wired to trigger VFD preset speeds. Turns out the tricky part is choosing those speeds and the accel/decel times to match the inflow, which fluctuates quite a bit as cows come and go to the milking stations. Receiver tanks are generally pretty small so not a lot of room to work with. If the tank overfills (if your VFD speed/accel falls behind) a ball float closes off the vacuum system (that's pulling the milk into that receiver) to keep milk out of the vacuum pump. When that happens, everything shuts down, milk gets spilled, big mess! When you see these in operation, the VFDs are always accelerating and decelerating between setpoints, so you know the setup is not "optimized" though it's certainly better than the full blast on-off cycle that was normal before VFDs. Each system must be tuned for individual pump performance, tank size, inflow over time (some cows give more, some milkers move faster), etc., all goes out the window when the pump gets repaired, impeller maybe changed. Very unsatisfying from the control design perspective - we want something that can be swapped out in the middle of the night with no phone calls!

Per jraef, analog/PID is an attractive option (hadn't seen those magnetorestrictive units before, thanks Jeff). The vacuum environment precludes inexpensive pressure transducers, surface foam defeats inexpensive ultrasonic sensors, other technologies (capacitive, etc.) get progressively more expensive for an application that's very cost averse (dairymen are going broke). I'll price magnetorestrictive now, but interestingly I found that constant level control via PID VFD (when tested without regard for sensor cost) actually wasn't the most effective solution anyway. It did get rid of the tuning headache (could pre-configure with some compromise minimum speed), but on the energy/churning side maintaining a constant level in the receiver effectively removed the tank capacity from the equation. Think about it - when the tank is filling slowly it's OK for the level to rise while the output stays low for improved cooling/churning performance, so best overall efficiency requires taking advantage of the tank storage capacity in a way that constant level evacuation does not. While looking into tuning schemes for variable level PID I experimented with a lot of different things.

I finally came up with a scheme for optimal pump speed under all conditions with no tuning and no exotic electronics required, did some bench testing, had some custom sensors manufactured, have a prototype in service that seems to work very well. There is an "analog" setup on the market (from one of those equipment manufacturers) but I purposely haven't looked at one, hoping my own might be better/simpler/cheaper.

Guess I should stop there, no promoting allowed and anyway I don't have a package ready yet (learning now what else you have to do to actually sell something, and how likely it is to be reverse engineered almost instantly if in fact it has any technical merit). I just had to chime in here on account of this application being sort of a hobby, plus it really is an interesting level control challenge for those who follow this forum . . .