Using feedback of level & concentration (4-20mA) to drive a make-up fluid circuit

[kdv1988]

Hello guys,

This is my first post here and I'm really looking forward to receiving some valuable inputs from you guys.

I have been trying to design an automatic fluid dispensing or make-up system for my CNC machine. A Level transmitter will be fitted to my sump and I will use a refractometer to send concentration readings in 4-20 mA to a PLC.

Water & coolant concentrate have to be pre-mixed before being sent to the sump. Water shall be supplied to the dispensing system under gravity & coolant can be transferred from a drum/barrel placed on the ground.

Once everyday, a make-up cycle is required to be run which will use the inputs from the level transmitter & refractometer to decide the quantity of liquid to be topped up.

Since the water will be fed under gravity, can I use a flow control valve with a flow transmitter on that pipeline? And for the coolant concentrate, flow control valve with transmitter on the delivery end of the pump.

Also, I was wondering if I can use a Venturi Injector (Ejector) for mixing the liquids? Or maybe use PWM.

Feel free to let me know if the information given above is inadequate.
Thanks,

KV

 

[itsmoked]

Sounds crazy expensive to me..
I'd do something like fill a single flipper bucket with a valve each for water and coolant. Then just count the flips digitally with the PLC.
Use a float switch to detect when the tank is full.
Two DIs and two DOs per machine and the refractometer.

Alternatively use standard water meters with pulse counters and monitor them with the PLC. Probably 10X more pricey and harder to read in for a PLC.

Screw worrying about mixing. That will happen soon enough with the meager amount you add to the large volume, just add it in the same location.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[kdv1988]

Thanks for replying.

There are some similar systems to what I had described tbh. You may look up Zebra Skimmers 'Dazzle' systems or an European product 'Will-Fill'.

Although, a water meter with pulse counter was something I had considered. I don't suppose they give out an analog output for the PLC, do they? Could I plug in a water meter and use it's signal to control a Solenoid Valve (modulating type) so as to change the flow-rate of the liquid that flows through and reaches the final sump?

Regarding mixing, do you think injecting about 2 Litres of concentrate in about 48 Litres of water to make it 50 Ltrs, will give me a homogeneous solution?

 

[itsmoked]

There are 4-20mA flow meters out there. I haven't used them and they're probably a little rarer and definitely more expensive since 4-20 costs a lot more than a measly magnetic reed switch.

Could I plug in a water meter and use it's signal to control a Solenoid Valve (modulating type) so as to change the flow-rate of the liquid that flows through and reaches the final sump?

Yes, with a bunch of tinkering, (a WHOLE bunch), but why on earth bother when a simple orifice or valve can be used to accomplish the same result.

Regarding mixing, do you think injecting about 2 Litres of concentrate in about 48 Litres of water to make it 50 Ltrs, will give me a homogeneous solution?

Gads! That's the ratio? I was expecting much more even proportions like 50/50 or 70/30 not 25/1. That speaks for premixing in a tank and single line distribution. You have a PLC and the job will work fine with a batch process. When the tank is detected empty, with a simple float switch, you pump in water until a 'full' float switch is met. Then you turn on a peristaltic metering pump to add in the concentrate. You will need to mix it. Typically with a paddle wheel or more simply a pump of any kind, a submersible dropped in the tank permanently. It can be running as soon as the tank has enough water added. Once mixed ~ 1 minute, fill the machines with this 'batched' mix until they each state they're full via a float switch.

You absolutely can't bank on flow measurements to top the machines up. Dead reckoning will not work.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Rputvin]

itsmoked - from the refractometer in the CNC's tank I'm guessing he needs to figure out the exact dose mixture to make up for changing concentration due to evaporation, or something that would throw off the ratio in the main process tank.

Another approach for some food for thought;

In the past I've just used pulsed solenoids for makeup of raw product, water, and amine to adjust viscosity and pH of a process tank at (relatively) constant temperature that saw a decent amount of evaporation/flashing off. If the supply is constant the timers can be tuned in the PLC to make finer adjustments to the volume being dosed. Do it in small amounts over the day instead of one batch at the end, interlock it so it only doses when the process tank's pump is running so the recirculation agitation does the mixing for you (I am unfamiliar with the fluid to know if a thorough pre-mix is critical). Depending on how quickly the sensors can refresh and balance this can be effective with little to no overshoot, or the PLC can be programmed with dose-wait-dose sequence to account for the time it would take for the sensor to settle. It becomes a bit more programming logic intensive, but removes material costs - which is why this was the approach for a product offering.

I agree a flow meter would probably be a bad way to go. Readings tend to jump around a lot until the flow has stabilized, accuracy can be an issue, turbulence is an issue, maintenance is an issue, etc.

 

[LittleInch]

kdv,

One thing I'm not clear of is whether you need to change the proportion of your concentrate depending on the refractometer or not.

I'm guessing yes and also the total volume of mixed water and concentrate is dependant on the level in the tank.

I would use a dosing or metering pump myself with a quill into the pipe or injection upstream of a static mixer.

Total volume of water or mixture can then be measured using a PD meter if pressure is enough.

Amount of concentrate can be determined and metering pump amount and flow rate amended to make sure it all gets in before the final water volume is met.

Simpler and easier to control IMHO.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[kdv1988]

@LittleInch
Yes, the amount of concentrate and even water will change as per the concentration and level of liquid in the tank. Some amount of digging has led me to this product -

https://www.youtube.com/watch?v=hFZgbdoZbSI

I'm looking to build something similar, wherein a PLC controls two separate lines, one for water & other for chemical concentrate. The chemical then is injected into the water line maybe using a Quill as you mentioned. But is a Quill used for chemical mixing purposes? I have no idea. For the mixing, I can maybe use a Fox Valve Venturi Ejector.

The chemical shall be introduced into the system through a barrel pump, while water inflow will be under gravity through a pipe. I am not sure if water will have enough pressure for a flowmeter reading. Plus flowmeters are bulky and expensive. Could a simple Turbine Flow meter be helpful in such cases? Maybe calibrate the pulses with the PLC and use that input to control a Solenoid Valve?

 

[kdv1988]

@Rputvin

You've made some valuable points. Yes, the concentration has to be measured either manually (if cost is a limiting factor for clients) or continuously with a process refractometer mounted in the machine sump. I'm leaning more towards a semi-automatic approach, where the client measures concentration in the sump manually and feeds the same into the PLC of my system, which then takes care of the rest by using the input from the level transmitter in the sump.

I have no experience with PWM, so it would mean a great deal to me if you could elaborate on how to work it in this scenario. The process I have in mind is described in my previous post where I tagged @littleInch, maybe you could have a look at the video. I'm sure you would get what I am trying to do here.

The chemical I will be dosing is a Cutting Fluid or an Alkaline Detergent. But I like the idea of dosing small amounts when the machine is in operation so as to facilitate automatic mixing due to turbulence. But then, dosing has to happen when the concentration fluctuates. I'm not sure when exactly that would happen. Maybe it requries a definite amount of evaporative or drag-out losses for it take effect.

 

[itsmoked]

I have to ask, why don't you just buy the referenced unit or a smaller one and spend your time plumbing it to the machines?

Keith Cress
kcress -

http://www.flaminsystems.com

 

[kdv1988]

@itsmoked
because that unit is simply a liquid dispensing system. What we are trying to develop here is an automatic make-up system that is required to be dynamic and adapt to changing parameters automatically.

Plus not to mention the fact that we're not based in the US, so the cost of importing.

I simply used their reference to describe how the two liquids shall be introduced in the system.

 

[LittleInch]

kdv1988,

You need to understand what your real requirement is.

when mixing things there are two basic ways to do this - you can either carefully measure each component, dump it into a vessel, agitiate / stir it / mix it and then use it
OR
mix the components on line inside a pipe so that your end product is the right mixture all the time.

The comparison is like mixing a drink at a bar.

You can add your concentrate to a glass then fill it with water and stir it or
press the drinks hose dispenser and it adds a concentrate syrup to the carbonated water to provide a ready mixed drink of any capacity.

The former is relatively easy to do, but needs a separate vessel to your final user tank unless you can mix the concentrate in the final tank.
The second method isn't so easy as it needs the mixture to regulate itself depending on flow and start / stop the injection of concentrate depending on flow of the bulk mixer.

Now I might have got it wrong, but in that video you linked in the demonstration, it looked to me like the concentrate only got mixed for some of the flow into the bucket and for a lot of the time all you had was water filling up the final volume. Hence it was really the first version - get the ingredients and then mix it.

I'm sure someone makes something similar in your part of the world already. This isn't rocket science.


Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[kdv1988]

@LittleInch

I definitely am interested in the latter, as providing another tank or a mixer is not possible due to space constraints and it generally being a headache.

I did not share that video so I could duplicate it. It was for representational purpose w.r.t the type of equipment, its size, the design etc. I do require the mixing to be done on-line. But as @Rputvin said in his last post, I am open to the possibility of allowing the mixing to happen when the machine operates (auto mixing due to turbulence or agitation in the tank), though I am not sure if that is the right way to do it.

 

[LittleInch]

Ok,

That's my concept and I think also it smoked which is measure your water flow - a simple turbine meter or mag flow will give you a 4-20 output or a pulsed output which you can count and then calculate the required flow of concentrate and feed that figure as the input into a dosing or metering pump set up for the concentrate.

Like these:

https://www.awe-ltd.co.uk/products/...-control-dosing-pump-4-20ma-pulsed-input.html

Then inject into a multihole quill or upstream of an inline mixer and job done.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Rputvin]

@kvd1988

I wasn't talking about PWM, not sure that would be effective with something like a solenoid. "Pulsing" in this regard was just short interval on/off sequences to transition from "bulk adding" to "fine control" without a more expensive dispense measurement device, relying instead on the tank fluid measurement to determine when enough has been dispensed. The viscosity meter being used was a mechanical system that provided a measurement every 10-15 seconds. With the slow sampling rate it would be easy to overshoot whatever additive you're dispensing, so we had a few methods to work around the sensor limitation. Volumes of those systems were less than 10 gallons, and dosing was a few ounces (100 mL range) at a time. The system could potentially dose every measurement cycle, and there was an option for an operator to hold a button down for constant dispense if they had just added virgin product and needed to get it within control range quickly. pH was a bit easier since the measurement device had a few millisecond refresh rate, and small doses of Amine or virgin product would balance the 0.1 pH change that triggered control rather quickly.

Like I said, it puts more emphasis on a robust PLC program. It needs to be user-adjusted for how much would be added at a time as we didn't control the volume in the process tank, no control over usage, evaporation/flashing, etc. In this case we were transitioning from a manual sampling and dosing method to this system, so user adjustment and tuning was tolerated as it was being used by experienced operators that understood how everything worked together. If you're looking for a turn-key solution for operators that currently don't pay attention to the coolant tank other than cleaning a filter and adding more premix to the tank this approach might not be ideal.

 

[georgeverghese]

No need to bother with any control valves or flow elements - you could run this as a batch operation. You know the total makeup volume required in the sump for the day, from which you can compute in PLC the individual volumes for water and coolant, given the current concentration feedback from the refractometer. Given that you know how volume changes with level in each of these source tanks, set up a level operated SV to open and close once the required dL has been dispensed into the sump tank. The required dL computation would also be set up in PLC. To mix it all up in the sump (since you dont have space for a premix tank), try setting up a compressed air bubbler or sparger pipe to run for a minute or two.

 

[kdv1988]

I'm gonna go through all the above posts and come back to you soon, thanks guys.

But, how about use of a Venturi Injector? They require a motive liquid at >2.5 Bar. Now since water is going to be supplied into the system under gravity from the factory's service line, I don't assume the pressure would be anywhere above 1 Bar, considering the height of the pipe wouldnt go beyond 30 feet.

How do I boost the pressure to >2.5 Bar? bearing in mind if this necessitates use of a pump, it has to be compact to fit in a box type design of my system and preferably mounted inline.

My idea:-

Use a Venturi Injector to mix the water & coolant. For eg.. Water through the factory's service line enters the system, where it passes through a Turbine Flow meter which senses the flow rate and sends the corresponding pulses to the PLC. The water enters the Venturi at the motive inlet port.

The PLC correlates the input of water flow rate and sends a signal to a Motorized Ball Valve (miniature type) which opens as per required flow of the coolant (this will be calibrated during set-up).

Now when the water enters the venturi, it creates a suction due to which coolant is pulled from the barrel, passes through the regulated ball valve opening and enters the venturi through the suction port. The two liquids mix and come out of the output port.

This will definitely require trials & validation. But does the concept seem workable to you guys?

 

[LittleInch]

It's one way but is a fixed mixing percent. So to vary the overall concentration you will need to set the mixing at the worst case concentration and then for lower amounts turn off the concentrate before the total amount is through. So then sometimes you will have pure water entering the tank. If that's ok then it's a simple system. If the flow enters the tank at a decent velocity say 2 m/sec, then you should get some decent mixing in the tank.

Water booster pumps at that sort of pressure are very common.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[kdv1988]

@LittleInch

Mixing percentage will vary of course as per the desired output concentration of the mixture. But in cases where the concentration in the machine sump has become 'Rich' I have to make it leaner by adding water with a very small % of coolant.

This is because coolant manufacturers do not recommend topping up with just water as that messes with the emulsion in the tank. A small % of coolant will not really change the concentration by a whole lot, but maintains the integrity of the liquid in the sump.

Booster pumps I am considering, but havent found one that is compact and will fit inside my system (ideally my unit would be approx. 30*30*60" .. LxWxH).

Food for thought :-
Instead of using a Venturi Mixer, I will inject appropriate quantity of coolant into the water stream using a Quill. I believe the velocity of water in the pipe should be around 2.5-3 m/s with viscosities of both the liquids nearly the same. Do you believe I will get a good mixture at the outlet? If this works, it makes my life a hell lot easier.

I have never used a Quill or even seen one tbh. Brief search on google has led me to believe it is mainly used for preserving the inner linings of pipes by injecting any corrosive chemical straight through the middle of the pipe. Is it used for mixing purposes?

 

[itsmoked]

I think you are missing something here. The refractometers will never EVER be reliable and accurate in this service. Too much contamination; oils, lubricants, dirt, bugs, fines. It will cost a fortune to put one in each machine. Then you need to service them frequently and hand check their calibration or deal with them miss-reading and now you've dosed wrong.

How about an out-of-the-box solution.

Plumb all the machines together. Include a little open mini tank of a couple of gallons in the serial string of machines. Circulate the coolant thru them all all-the-time.

The mini-tank is your soul instrumented tank.

You pump from the last machine to the instrument tank thru an appropriate cartridge filter to remove all the suspended gunk and fines.

Now you have filtered mix in your instrument tank. Use one or two refractometers in the instrument tank. If they agree then automation can continue. If not, you have notice they need service. If desired use three and have them vote out the dirty one. Use a simple level switch to monitor the sensor tank's level. It will drop with the system level.

Near it you have a plastic 55 gallon drum mixing batch tank. Using two float switches you monitor its empty and too-full levels. When the tank reaches a low level the PLC reads the reflectometers does the math and you make an appropriate batch of the right mix via a flow meter for the makeup water and dosing pump for the concentrate. You batch an appropriate amount for the 55 gallon drum, maybe 40 gallons.

Use a simple submersible pump with a eductor on its output to mix the 55 gallon batch until you're satisfied it's homogenized. Using a motorized ball valve(needing no pressure difference like solenoid valves) the PLC simply gravity feeds the small sensing tank until it's topped back up to its 'full' float. As it drops again the PLC retops it, perhaps every 30 minutes or whatever you find appropriate.

Advantages:
1) Sensing and control localized away from the CNC machines.

2) All machines always topped up.

3) All controls except the reflectometers and possibly the water flowmeter are simple to troubleshoot non-calibated digital signals going to the PLC.

4) Only two valves. {Water and mix make-up. The concentrate is a metering pump (ON/OFF)}

5) Reflectometers are in filtered mix and only a few are needed.

6) Mix is completely homogenized before it's used.

7) No rinky-dink per machine calculations and accumulating errors due to pipe runs and trapped-in-pipe wrong mix and no army of per-CNC valves that must be powered.

8) Simple, easily maintainable, mostly ON/OFF.

Keith Cress
kcress -

http://www.flaminsystems.com