Designing an Automatic Coolant Monitoring System for CNC Machines

[Asymptotic]

Machine sump 'knowns'
Sump coolant concentration (sensed by refractometer)
Sump liquid level (sensed by continuous level transmitter)

Desired goal
Refill sump to desired level with a mixture of water/coolant concentrate to achieve a desired % concentration.

What you must do is come up with a way to achieve this goal while keeping in mind the "engineer's trilemma" - "Quick, Cheap, Good: Pick two."

My question of "if the target is 3.5%, can the final mix vary between, say, 3.2% to 3.8%" was to learn how tightly constrained the control system must be (i.e. - does it have to be essentially dead-on with very tight control of coolant %, or is a fair amount of "slop" acceptable).

How can the desired goal be met?
Two general approaches are 'continuous' and 'batch'. A batch system is easier to design and build than a continuous system, and due to it's simplicity, if it can meet the goal is often a better choice.

Batch system outline
Size the mixing tank to hold about as much volume as the machine sump requires when it signals "make-up required".

Mix tank level can be monitored with a 'low level' float switch set very close to the bottom, and a 'full level' float corresponding to a desired mixed volume. The distance between low and high level floats defines the effective volume of the mix tank, and also how much fluid is delivered to the sump when it signals "make-up required".

Use a N.C. solenoid valve to switch mixing tank water supply, and a peristaltic metering pump to transfer concentrate to the mixing tank. You'll have to read through manufacturer's catalogs to find a peristaltic pump with the required volumetric output, and also think about how to control it. Traditionally, they come in fixed and variable output models. Adjustable pumps cost more. The ones I've used mechanically altered the effective pump tube length, but (even more costly) programmable, variable speed peristaltic pumps with PLC interfacing options are also available.

Look over both types, but a fixed pump may be fine in this application (see caveat below).

Choose one that accepts heads for several sizes of pump tube in case it becomes necessary to change basic flow rate after initial installation. The pumps I've used were rated in gallons per day (GPD), and could be set up with 1/16", 1/8", 3/16", and 1/4" diameter tubing to have corresponding flow rates of 8.75, 34.62, 82.17, and 143.79 GPD.

Let's say a fixed output pump head is rated 82.17 GPD, and the PLC has calculated (based on sump % concentration and current volume) that 2 liters of concentrate are required to fill the mix tank to the required concentration. 82.17 GPD works out to 216 ml/minute. How long must the pump run to transfer 2 liters of concentrate to the mixing tank? It works out to 9.26 minutes.

Batch prep, option 1:
Use the PLC to switch on the peristaltic metering motor for 9.26 minutes, then turn on the water fill valve until mixing tank level reaches the high level float. Turn on, and run the propeller mixer long enough for thorough mixing, then open a solenoid valve (if the mixing tank is higher than the sump, and gravity feed is possible) or turn on a transfer pump to move this fresh mix into machine sump. When the low level float switch is triggered. stop the transfer, turn off the mixer motor, and consider the mix tank available to start another batch.

Batch prep option 1 critique:
Doing concentrate and water fills in this stepwise fashion is very simple, but causes some amount of mixture % variation. How much it matters depends on mix tank dimensions (hence, volume), etc., and you'll have to run calculations to see if it matters enough.

Preferably, both water and concentrate are added in unison, but a problem with the above approach is the time to fill the mix tank with water is probably much less than the time to meter concentrate into it. If both water and concentrate fill commence at the same time, the high level float will trigger before metering pump delivery is done.

A slightly more complex approach is to fit the water line with a manual flow control valve, and adjust it to fill the tank to full level in less time than it takes for the concentrate metering pump to complete it's task. It too won't deliver a perfect concentration ratio, but will be a lot closer than the stone age, solenoid valve only approach above, and with your rather broad range of acceptable concentrations ought to be close enough.

In either case, acceptable results depend on the metering pump operating correctly. The control system doesn't know if the coolant concentrate container goes empty, a leak develops in the suction line (causing the pump to become air-bound on stop delivery), or a delivery line leak develops (concentrate spills to the floor instead of feeding into the mix tank), or the pump itself fails.

Several manufacturers make low flow rate switches and totalizers suited for peristaltic pump service, and could be added to the control system. Keep in mind that continuously variable peristaltic pumps designed for PLC control may have these built in for precisely this type of fault detection.

A Totalizer would have to be installed on the Panel right? Doesn't it require a Flow Meter to be used along with it?

It is useful to think of a totalizer as a flowmeter with memory. That memory may be mechanical (gear driven number wheels) or electronic. If you have a 4-20 ma flowmeter feeding into a PLC, program the PLC to keep track of totalized flow.

 

[kunalvanjare]

Machine sump 'knowns'
Sump coolant concentration (sensed by refractometer)
Sump liquid level (sensed by continuous level transmitter)

Desired goal
Refill sump to desired level with a mixture of water/coolant concentrate to achieve a desired % concentration.

What you must do is come up with a way to achieve this goal while keeping in mind the "engineer's trilemma" - "Quick, Cheap, Good: Pick two."

My question of "if the target is 3.5%, can the final mix vary between, say, 3.2% to 3.8%" was to learn how tightly constrained the control system must be (i.e. - does it have to be essentially dead-on with very tight control of coolant %, or is a fair amount of "slop" acceptable).

How can the desired goal be met?
Two general approaches are 'continuous' and 'batch'. A batch system is easier to design and build than a continuous system, and due to it's simplicity, if it can meet the goal is often a better choice.

Batch system outline
Size the mixing tank to hold about as much volume as the machine sump requires when it signals "make-up required".

Mix tank level can be monitored with a 'low level' float switch set very close to the bottom, and a 'full level' float corresponding to a desired mixed volume. The distance between low and high level floats defines the effective volume of the mix tank, and also how much fluid is delivered to the sump when it signals "make-up required".

Use a N.C. solenoid valve to switch mixing tank water supply, and a peristaltic metering pump to transfer concentrate to the mixing tank. You'll have to read through manufacturer's catalogs to find a peristaltic pump with the required volumetric output, and also think about how to control it. Traditionally, they come in fixed and variable output models. Adjustable pumps cost more. The ones I've used mechanically altered the effective pump tube length, but (even more costly) programmable, variable speed peristaltic pumps with PLC interfacing options are also available.

Look over both types, but a fixed pump may be fine in this application (see caveat below).

Choose one that accepts heads for several sizes of pump tube in case it becomes necessary to change basic flow rate after initial installation. The pumps I've used were rated in gallons per day (GPD), and could be set up with 1/16", 1/8", 3/16", and 1/4" diameter tubing to have corresponding flow rates of 8.75, 34.62, 82.17, and 143.79 GPD.

Let's say a fixed output pump head is rated 82.17 GPD, and the PLC has calculated (based on sump % concentration and current volume) that 2 liters of concentrate are required to fill the mix tank to the required concentration. 82.17 GPD works out to 216 ml/minute. How long must the pump run to transfer 2 liters of concentrate to the mixing tank? It works out to 9.26 minutes.

Batch prep, option 1:
Use the PLC to switch on the peristaltic metering motor for 9.26 minutes, then turn on the water fill valve until mixing tank level reaches the high level float. Turn on, and run the propeller mixer long enough for thorough mixing, then open a solenoid valve (if the mixing tank is higher than the sump, and gravity feed is possible) or turn on a transfer pump to move this fresh mix into machine sump. When the low level float switch is triggered. stop the transfer, turn off the mixer motor, and consider the mix tank available to start another batch.

Batch prep option 1 critique:
Doing concentrate and water fills in this stepwise fashion is very simple, but causes some amount of mixture % variation. How much it matters depends on mix tank dimensions (hence, volume), etc., and you'll have to run calculations to see if it matters enough.

Preferably, both water and concentrate are added in unison, but a problem with the above approach is the time to fill the mix tank with water is probably much less than the time to meter concentrate into it. If both water and concentrate fill commence at the same time, the high level float will trigger before metering pump delivery is done.

A slightly more complex approach is to fit the water line with a manual flow control valve, and adjust it to fill the tank to full level in less time than it takes for the concentrate metering pump to complete it's task. It too won't deliver a perfect concentration ratio, but will be a lot closer than the stone age, solenoid valve only approach above, and with your rather broad range of acceptable concentrations ought to be close enough.

In either case, acceptable results depend on the metering pump operating correctly. The control system doesn't know if the coolant concentrate container goes empty, a leak develops in the suction line (causing the pump to become air-bound on stop delivery), or a delivery line leak develops (concentrate spills to the floor instead of feeding into the mix tank), or the pump itself fails.

Several manufacturers make low flow rate switches and totalizers suited for peristaltic pump service, and could be added to the control system. Keep in mind that continuously variable peristaltic pumps designed for PLC control may have these built in for precisely this type of fault detection.
It is useful to think of a totalizer as a flowmeter with memory. That memory may be mechanical (gear driven number wheels) or electronic. If you have a 4-20 ma flowmeter feeding into a PLC, program the PLC to keep track of totalized flow.

Hey, sorry for the late response.

I have gone through your post and believe that this would work when there is no fluctuation in the daily 'range' of make-up fluid. For example if approx 50 Ltrs of make-up is required, the mix tank has to be sized accordingly. How about cases where there would be more liquid required? Also filling concentrate using a peristaltic pump can be time-consuming.

I have something in mind :-

1. A refractometer & level transmitter record the current values of concentration & volume of liquid in the sump and feed the PLC.

2. Using this data the PLC actuates two valves to varying degrees (that will be calibrated with the output signal). One valve is for the plant's water supply. The other for the coolant concentrate inlet.

3. Both these inlets will be taken to a Venturi Injector (not a mixer.. but an injector, something that is used in irrigation to mix fertilizers apparently). Link - https://mazzei.net/venturi_injectors/

4. What this achieves is proportionate flow of water through the Venturi getting mixed with metered flow of coolant without the need for a pump or secondary mixing tank or mechanical blades etc.

Inspiration for this was taken from - http://s000.tinyupload.com/index.php?file_id=20469651419304289113

Doubts :-

1. This requires a Flow Control Valve with flow-meter/totalizer for both water & coolant lines into the Venturi Mixer.
2. Not sure whether a metered flow of water into the Venturi will be able to draw the required volume of coolant when the coolant line itself is controlled by the valve. Maybe the mixing time might increase, or maybe the mixing will not be adequate.This system can be made into a semi-automatic one if the operator can enter the current concentration of coolant in the machine sump by himself into the HMI of the system. This rules out the requirement of an Inline Refractometer thus drastically reducing the cost of the system.

Once the current concentration is set, the system can automatically determine the volume(through level transmitter) & concentration (through formula) of the make-up liquid.

Would love to hear your views on this.

 

[Asymptotic]

I have gone through your post and believe that this would work when there is no fluctuation in the daily 'range' of make-up fluid. For example if approx 50 Ltrs of make-up is required, the mix tank has to be sized accordingly. How about cases where there would be more liquid required?

What would those cases be?
Sump level is what triggers the make-up cycle during normal operation.

That leaves the case of a "dry" start-up where the machine has been intentionally drained.
If the sump holds 150 liters, for example,

Option 1
Manually mix concentrate with the required amount of water to fill the sump before CNC machine operation commences.

Option 2
Allow the 50 liter capacity automated system to cycle three times to fill the 150 liter sump.

Also filling concentrate using a peristaltic pump can be time-consuming.

In an automated system, so long as the time it takes to meter and mix the refill charge is less than the interval between "sump full" and "sump requires make-up", does it matter how time-consuming it is?

2. Using this data the PLC actuates two valves to varying degrees (that will be calibrated with the output signal). One valve is for the plant's water supply. The other for the coolant concentrate inlet.

3. Both these inlets will be taken to a Venturi Injector (not a mixer.. but an injector, something that is used in irrigation to mix fertilizers apparently). Link - https://mazzei.net/venturi_injectors/

4. What this achieves is proportionate flow of water through the Venturi getting mixed with metered flow of coolant without the need for a pump or secondary mixing tank or mechanical blades etc.

What type of valves? I'm doubtful a PWM-based approach using on/off valves as expressed in earlier posts would work well on this venturi injector. Actual proportioning valve(s) would probably be required, however ...

1. Outlet back-pressure controls how much suction (concentrate) flows through this venturi injector.

2. Built a set of total flow curves for model 283 injector from performance tables in the specification PDF.
Total flow is the fixed "motive" water flow at a particular inlet PSI plus the rated suction flow for given injector outlet pressures.

An injector doesn't necessarily provide a linear relationship over the entire flow range for a given inlet water pressure. It would be desirable to operate the venturi in the linear part of its range for proportionality. For example, with a 50 PSI inlet pressure, a big dead spot exists between 0 and 15 PSI of outlet pressure before changing it has an effect on either total flow or mixture concentration.

3. Points 1 and 2 may be beside the point. If I recall correctly, you are looking for mixture concentrations of under 10%. Can this device deliver them? Suction flow is higher than motive flow for all devices throughout their performance range. For example, motive flow is 0.39 GPM for the model 283 at 30 PSI inlet pressure, and suction flow is 0.80 GPM. This is a total flow of 1.19 GPM.

0.80/1.19 = 67.2% inlet water/suction fluid concentration. This may or may not give acceptable results.

What is the the % Brix reading of fresh concentrate? If it starts out suitably low, then acheiving the desired mixture concentration may be possible.

I'm not saying something like this couldn't be done, simply that there is more to it than it appears at first glance. It is necessary to perform sample calculations to determine whether any particular approach is within the realm of possibility.

 

[ktswastetech]

Hello. Jumping in on this discussion, hopefully not being intrusive.

I am designing the exact system that you have described. However, in my case, the coolant concentration at 15 different machine sumps is all the same. I am proposing the installation of an overhead clean supply and dirty return piping system (3/4" pipe). At each machine sump I will have an AOD dirty transfer pump, and an automatically controlled clean fill valve (controlled by level in the sump).

The clean and dirty piping system will be tied into a centrally located "filter system) where the coolant will be filtered, tramp oil will be removed, and the concentration of coolant can be maintained (in one location).

Each AOD pump will be set up to run (pump dirty coolant from the sump) for a period of time (say 30 minutes at a rate of 10 to 15 GPM), back to the central filter system. When the level in the sump drops to the appropriate level, clean coolant will fill the sump back up. The clean coolant will come from the central filter system clean tank, also fed by an AOD pump.

With this set-up, the level in the sump tanks will always be maintained at the proper level and the proper concentration. The customer will no longer have to check concentration at each sump or top off each sump. Because of the constant "turnover" of the coolant in each sump, the concentration will always be the same as the concentration in central system clean holding tank.

 

[kunalvanjare]

Hello. Jumping in on this discussion, hopefully not being intrusive.
I am designing the exact system that you have described. However, in my case, the coolant concentration at 15 different machine sumps is all the same. I am proposing the installation of an overhead clean supply and dirty return piping system (3/4" pipe). At each machine sump I will have an AOD dirty transfer pump, and an automatically controlled clean fill valve (controlled by level in the sump). The clean and dirty piping system will be tied into a centrally located "filter system) where the coolant will be filtered, tramp oil will be removed, and the concentration of coolant can be maintained (in one location). Each AOD pump will be set up to run (pump dirty coolant from the sump) for a period of time (say 30 minutes at a rate of 10 to 15 GPM), back to the central filter system. When the level in the sump drops to the appropriate level, clean coolant will fill the sump back up. The clean coolant will come from the central filter system clean tank, also fed by an AOD pump. With this set-up, the level in the sump tanks will always be maintained at the proper level and the proper concentration. The customer will no longer have to check concentration at each sump or top off each sump. Because of the constant "turnover" of the coolant in each sump, the concentration will always be the same as the concentration in central system clean holding tank.

Hello @ktswastetech , welcome and thanks for contributing to this.

What you have proposed is workable provided all machines are at the same %concentration, which is not the case with our application.

Although, how do you supply to 15 sumps using one pump from the central reservoir? And, do you cut-off supply to the sump after receiving the 'high level' input from the machine?

 

[ktswastetech]

Hello @ktswastetech , welcome and thanks for contributing to this.

What you have proposed is workable provided all machines are at the same %concentration, which is not the case with our application.

Although, how do you supply to 15 sumps using one pump from the central reservoir? And, do you cut-off supply to the sump after receiving the 'high level' input from the machine?

The one pump at the central reservoir is an AOD pump that is "always on" so it will keep the clean supply header pressurized. At each sump, I am using a Normally Open air controlled level switch that is connected to an automatic ball valve that will be connected to the clean supply drop. The automatic ball valve has an air actuator that is air to open/spring to close.

Also, at each sump, I am installing a 120 VAC single phase "smart outlet". The outlet will be used to power a solenoid valve that will control the air feed to the dirty sump pump. The customer will be able to program each outlet (via an app on his phone) so that only one sump pump runs at a time. He will also set how long the pump will run. Once the level in the sump drops a certain amount, the air controlled level switch will open the automatic ball valve, allowing the sump to "refill" to the desired level.

 

[kunalvanjare]

The one pump at the central reservoir is an AOD pump that is "always on" so it will keep the clean supply header pressurized. At each sump, I am using a Normally Open air controlled level switch that is connected to an automatic ball valve that will be connected to the clean supply drop. The automatic ball valve has an air actuator that is air to open/spring to close. Also, at each sump, I am installing a 120 VAC single phase "smart outlet". The outlet will be used to power a solenoid valve that will control the air feed to the dirty sump pump. The customer will be able to program each outlet (via an app on his phone) so that only one sump pump runs at a time. He will also set how long the pump will run. Once the level in the sump drops a certain amount, the air controlled level switch will open the automatic ball valve, allowing the sump to "refill" to the desired level.

This may sound stupid, but wouldn't an 'always on' pump with a pressurized pipe lead to backpressure or overload on the pump in cases where none of the automatic re-fill valves are open?

I didn't quite understand the significance of the smart outlet. Could you please elaborate.

Also, is maintenance of coolant in the central reservoir in your scope? I mean the refractometer, the filtration system etc..

And why a pneumatic float switch and not a simple magnetic float switch?

 

[ktswastetech]

An AOD pump = Air Operated Double Diaphragm Pump - This pump can be deadheaded to provide a constant pressurized clean return header.

A smart outlet is an outlet that can be controlled by an App on your phone or smart device via WiFi. An example is GE Enbrighten Z-Wave Plus Smart Receptacle Outlet. One outlet on the receptacle can be set up to turn on/off based on a schedule. I am simply plugging in an electrically actuated 120 VAC solenoid that will control the air feed to the sump pump. By using the smart outlet, the customer can set up a plant schedule for when each pump at a machine sump is to come on, and for how long. This way he can control when the turnover occurs at each sump and for how long (based on sump size).

Yes, maintenance of the coolant is in my scope. I am providing the equipment/tankage that will:
- receive the dirty coolant - into a dirty tank
- constantly remove tramp oil and large solids via a tramp oil coalescor and floating skim head
- coalescor loop also circulates the coolant from the dirty tank to the clean tank
- constantly filter the coolant via a separate filtration loop consisting of an AOD pump, inline magnetic filter and two bag filters in series - filters coolant in the clean tank
- coolant concentration will be monitored manually via a refractometer
- coolant make-up will be done manually with a crown proportioning pump, plant water feed, and coolant concentrate
- control of bacteria via ozone generators - one in each tank clean/dirty

 

[kunalvanjare]

An AOD pump = Air Operated Double Diaphragm Pump - This pump can be deadheaded to provide a constant pressurized clean return header.

A smart outlet is an outlet that can be controlled by an App on your phone or smart device via WiFi. An example is GE Enbrighten Z-Wave Plus Smart Receptacle Outlet. One outlet on the receptacle can be set up to turn on/off based on a schedule. I am simply plugging in an electrically actuated 120 VAC solenoid that will control the air feed to the sump pump. By using the smart outlet, the customer can set up a plant schedule for when each pump at a machine sump is to come on, and for how long. This way he can control when the turnover occurs at each sump and for how long (based on sump size).

Yes, maintenance of the coolant is in my scope. I am providing the equipment/tankage that will:
- receive the dirty coolant - into a dirty tank
- constantly remove tramp oil and large solids via a tramp oil coalescor and floating skim head
- coalescor loop also circulates the coolant from the dirty tank to the clean tank
- constantly filter the coolant via a separate filtration loop consisting of an AOD pump, inline magnetic filter and two bag filters in series - filters coolant in the clean tank
- coolant concentration will be monitored manually via a refractometer
- coolant make-up will be done manually with a crown proportioning pump, plant water feed, and coolant concentrate
- control of bacteria via ozone generators - one in each tank clean/dirty

Got it!

I'm assuming the piping won't be under your scope would it?

 

[ktswastetech]

Actually it is, I am doing the complete design. I also sell equipment that will be part of the entire bid package. I just finished the drawings that will be submitted to various contractors for bidding. I am a mechanical engineer and have been doing this type of work for the past 20 years. In this case, my customer has asked for a "turn key" quote.

 

[berkeman]

@kunalvanjare and @ktswastetech -- could you please take that part of this discussion to private messages (PMs)? Otherwise it looks like we are helping with a transaction in the open forums. Thanks.

 

[kunalvanjare]

@kunalvanjare and @ktswastetech -- could you please take that part of this discussion to private messages (PMs)? Otherwise it looks like we are helping with a transaction in the open forums. Thanks.

Noted.

 

[ktswastetech]

Yes, sorry, new here.

 

[berkeman]

Yes, sorry, new here.

No worries. You haven't posted any spam so far, and your responses are very helpful for the OP.