Hello guys, I have searched high and low for clues to my problem and can't seem to get help. I have a centrifugal pump controlled by a VFD. The pump transfers milk from one tank to another through a milk-to-water and milk-to-glycol heat exchanger. The liquid levels in the two tanks are constantly changing. The source tank fills and empties and the destination tank fills, then is changed to an empty tank. I can monitor the levels of both tanks and use them in my PLC networks. I would like to arrive at an equation for my PID loop that varies the speed of the pump to keep the flow rate constant while the head pressure is changing. The number I want to arrive at is, how much do I need to change the speed of the pump for each change in total head to keep the flow constant. I have tried substituting the affinity equations in each other but I find myself running in circles. If anyone has done anything like this or knows how to approach it, I would appreciate hearing from you.
Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations cowski on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!
constant flow variable speed pump for heat exchanger 1
- Thread starter gaalfred
- Start date
georgeverghese
Chemical
A minor refinement to the new control scheme proposed in (b) would be to split range the final milk TIC output, such that the earlier output response modulates glycol flow at the final glycol-milk HX and the later response acts as an override on the XIC output(that derives pump speed) which goes to the pump VFD. This later response would address cases where there are temporary shortfalls in chilling capacity at the refrigerant compressors.
LittleInch
Petroleum
Don't know what a 3A meter is, but why not look at a UT clamp on meter like this (example only)
or introduce a smaller section of pipe and get one with a better flow range like this
I still think though a major part of your issue is that the pump, even at 2Hp, is just too big for your duty and hence to make the system insensitive to inlet tank height, make the fixed head element (at your rather low flow rate) much bigger.
coupled with Georges idea of a faster temperature pick up you have two ways to do thinks better without a lot of expense.
Remember - More details = better answers
Also: If you get a response it's polite to respond to it.
or introduce a smaller section of pipe and get one with a better flow range like this
I still think though a major part of your issue is that the pump, even at 2Hp, is just too big for your duty and hence to make the system insensitive to inlet tank height, make the fixed head element (at your rather low flow rate) much bigger.
coupled with Georges idea of a faster temperature pick up you have two ways to do thinks better without a lot of expense.
Remember - More details = better answers
Also: If you get a response it's polite to respond to it.
Gaalfred, If you want cheap, get rid of the VFD. Open mind. I'll talk you through it. OK we change rediculous to unfortunate??
Pump head output (affinity law) varies with the square of pump speed. That's only 1/2 of the problem. The other half of the problem is the system curve; ie what's going on with flow and head in the pipe and tanks. The system flows vary more or less with the square root of pump head. Additionally they can have an initial H0 representing static head of the discharge tank that the pump must begin to pump against when starting. NPSH can also adversely affect this. So head obviously changes as the pumping continues and the discharge tank fills. How that changes is affected more by the dimensions of your tanks; ie how much head has increased in the discharge tank, and maybe the supply tank too, for any given quantity of fluid pumped. That is related more directly to tank fluid elevations than it is to flow rate. As you can see the resulting complex relationship is nowhere near direct to pump speed. If you could make a good mathematical relationship to describe it, it would not match how your PID would take temperature input and then output an instantaneous pump speed.
You've got two problems going on. Flow control and heat control. You apparently need flow control, I imagine to be as constant as possible, so you can know tank fill and process stop times and all those scheduling things that are good to know, and so that your heat exchanger can operate at a nearly constant power output. Wild variations in exchanger output power output doesn't do anything any good. Radical flow changes would cause wild temperature and power fluxes at the exchanger even if the milk remained the same temperature during the process.
When faced with multiple tasks, I like to break them down. Flow control is one, temperature control the other. Pump control works well for flow control. Exchanger power control works well for temperature. Those two working independently takes away the ability of each to fight the other, which would happen if you tried to control flow and temperature with only one feedback loop. So, use some kind of flow control for the pump and a temperature control to adjust the exchanger power input for discharge temperature.
Once you separate the two control loops, the pump is free to operate at constant speed for the process, if the flowrate is right at that constant speed, however you also have a wash cycle, WHICH IS COMPLETELY ANOTHER FUNCTION. So you need another button to turn these loops off and switch to the wash cycle. But forget that for now. If the flowrate of the pump is not right at it's syncronous speed, you could change the speed with a VFD, or you might could run at synced speed and use a control valve to control flow. Control valves can be far easier AND CHEAPER means of flow control than buying and installing VFDs due to cost of the VFD over the small control valve the previous mismatches I mentioned in pump head output vs flow in the system with tanks. Control valves don't suffer those head vs flow problems as much as pumps can. So I think that with proper tank dimensions, you could run the pump at sync speed with a contrl valve for flow and use a temperature sensor feedback to control the exchanger's function directly with the output temperature.
Technology is stealing American jobs. Stop H1-Bs for robots.
Pump head output (affinity law) varies with the square of pump speed. That's only 1/2 of the problem. The other half of the problem is the system curve; ie what's going on with flow and head in the pipe and tanks. The system flows vary more or less with the square root of pump head. Additionally they can have an initial H0 representing static head of the discharge tank that the pump must begin to pump against when starting. NPSH can also adversely affect this. So head obviously changes as the pumping continues and the discharge tank fills. How that changes is affected more by the dimensions of your tanks; ie how much head has increased in the discharge tank, and maybe the supply tank too, for any given quantity of fluid pumped. That is related more directly to tank fluid elevations than it is to flow rate. As you can see the resulting complex relationship is nowhere near direct to pump speed. If you could make a good mathematical relationship to describe it, it would not match how your PID would take temperature input and then output an instantaneous pump speed.
You've got two problems going on. Flow control and heat control. You apparently need flow control, I imagine to be as constant as possible, so you can know tank fill and process stop times and all those scheduling things that are good to know, and so that your heat exchanger can operate at a nearly constant power output. Wild variations in exchanger output power output doesn't do anything any good. Radical flow changes would cause wild temperature and power fluxes at the exchanger even if the milk remained the same temperature during the process.
When faced with multiple tasks, I like to break them down. Flow control is one, temperature control the other. Pump control works well for flow control. Exchanger power control works well for temperature. Those two working independently takes away the ability of each to fight the other, which would happen if you tried to control flow and temperature with only one feedback loop. So, use some kind of flow control for the pump and a temperature control to adjust the exchanger power input for discharge temperature.
Once you separate the two control loops, the pump is free to operate at constant speed for the process, if the flowrate is right at that constant speed, however you also have a wash cycle, WHICH IS COMPLETELY ANOTHER FUNCTION. So you need another button to turn these loops off and switch to the wash cycle. But forget that for now. If the flowrate of the pump is not right at it's syncronous speed, you could change the speed with a VFD, or you might could run at synced speed and use a control valve to control flow. Control valves can be far easier AND CHEAPER means of flow control than buying and installing VFDs due to cost of the VFD over the small control valve the previous mismatches I mentioned in pump head output vs flow in the system with tanks. Control valves don't suffer those head vs flow problems as much as pumps can. So I think that with proper tank dimensions, you could run the pump at sync speed with a contrl valve for flow and use a temperature sensor feedback to control the exchanger's function directly with the output temperature.
Technology is stealing American jobs. Stop H1-Bs for robots.
Back to the wash cycle.
The "control valve" might be just a valve with two positions, one for process flow, the other full open for wash cycle. That could work very well with a constant speed pump running at synced speed. When on wash cycle, just turn off the heat exchanger.
Technology is stealing American jobs. Stop H1-Bs for robots.
The "control valve" might be just a valve with two positions, one for process flow, the other full open for wash cycle. That could work very well with a constant speed pump running at synced speed. When on wash cycle, just turn off the heat exchanger.
Technology is stealing American jobs. Stop H1-Bs for robots.
- Thread starter
- #25
Thanks guys,
I found a performance curve for my pump. As far as I can tell, Thomsen doesn't publish full-range charts and I have never seen a chart for the Boumatic milk pump that is often used in daries. This one has a set of head vs flow curves for different sized impellers. I can convert the impeller diameter to speed and get a chart of head vs flow at 6 different speeds, none of which are the speeds we usually run for milk flow. I may have to set up a test bench with a paddle-wheel meter and a valve and pressure gauge to build my own chart of head vs flow at several different speeds in the range that is useful to me. It is a Thomsen #4 pump with 1 1/2" inlet and outlet running at 3450 rpm max.
I picked 20 gpm flow and followed the vertical line on the chart to each curve, then read across to get the total head at that flow and speed. From the two sets of 6 numbers, flow vs speed, I let Excel run a regression analysis on it and Excel came up with the formula y=0.74x+74.6. Using this formula with y=head and x=speed, I am able to arrive at a new speed to run my pump when a change in head occurs to keep the flow rate constant. I believe this answers my original question. Thanks, george for the tip on doing this.
I will also investigate can be done to improve the response of the temperature sensor per george's suggestion.
The term "3A" refers to the 3-A Sanitary Standards that specify the criteria for the design and fabrication of equipment that comes into contact with food. All equipment in Grade A dairies must meet these standard. I am sure that any 'fins' on a thermowell would be frowned on by the federal inspectors and harbors of bacteria, but I will check.
One feature of our systems is that the VFD controlling the milk flow causes the milk flow to exactly match the capacity of the chiller with its current operating conditions. These conditions vary quite a bit from winter to summer as the ambient temperature of the air across the condenser varies from -30F to 110F. By setting the set point of the glycol to just below the milk temperature set point, the chillers are used to their full capacity. It is very important to do this as there is not much excess cooling capacity, especially on a hot day when the high-producing group of cows is being milked and the milk must be instantaneously cooled to its marketable temperature.
I found a performance curve for my pump. As far as I can tell, Thomsen doesn't publish full-range charts and I have never seen a chart for the Boumatic milk pump that is often used in daries. This one has a set of head vs flow curves for different sized impellers. I can convert the impeller diameter to speed and get a chart of head vs flow at 6 different speeds, none of which are the speeds we usually run for milk flow. I may have to set up a test bench with a paddle-wheel meter and a valve and pressure gauge to build my own chart of head vs flow at several different speeds in the range that is useful to me. It is a Thomsen #4 pump with 1 1/2" inlet and outlet running at 3450 rpm max.
I picked 20 gpm flow and followed the vertical line on the chart to each curve, then read across to get the total head at that flow and speed. From the two sets of 6 numbers, flow vs speed, I let Excel run a regression analysis on it and Excel came up with the formula y=0.74x+74.6. Using this formula with y=head and x=speed, I am able to arrive at a new speed to run my pump when a change in head occurs to keep the flow rate constant. I believe this answers my original question. Thanks, george for the tip on doing this.
I will also investigate can be done to improve the response of the temperature sensor per george's suggestion.
The term "3A" refers to the 3-A Sanitary Standards that specify the criteria for the design and fabrication of equipment that comes into contact with food. All equipment in Grade A dairies must meet these standard. I am sure that any 'fins' on a thermowell would be frowned on by the federal inspectors and harbors of bacteria, but I will check.
One feature of our systems is that the VFD controlling the milk flow causes the milk flow to exactly match the capacity of the chiller with its current operating conditions. These conditions vary quite a bit from winter to summer as the ambient temperature of the air across the condenser varies from -30F to 110F. By setting the set point of the glycol to just below the milk temperature set point, the chillers are used to their full capacity. It is very important to do this as there is not much excess cooling capacity, especially on a hot day when the high-producing group of cows is being milked and the milk must be instantaneously cooled to its marketable temperature.
georgeverghese
Chemical
That sounds better now - The y in this formula is really y1 + y2, where y1 is the differential head between the 2 tanks, and y2 is a constant and is the frictional head loss, which would be constant at a constant flow of 20gpm.
The use of a thermocouple instead of an RTD would be a big improvement on TT response time also.
Yes, would suppose that in this plant, you would be operating the refrig. compressors to adjust speed and discharge pressure to produce a constant glycol delivery temp(to the milk-glycol heat exchnager) at the refrigerant-glycol heat exchanger. And that you would be running the compressors at high discharge pressure and high speed in summer to make up for the reduced refrigerant condensor cooling capacity.
The use of a thermocouple instead of an RTD would be a big improvement on TT response time also.
Yes, would suppose that in this plant, you would be operating the refrig. compressors to adjust speed and discharge pressure to produce a constant glycol delivery temp(to the milk-glycol heat exchnager) at the refrigerant-glycol heat exchanger. And that you would be running the compressors at high discharge pressure and high speed in summer to make up for the reduced refrigerant condensor cooling capacity.
LittleInch
Petroleum
I'll add a bit more tomorrow but from looking at the pump curve, you will find making this system better difficult because for the main duty (18gpm)your pump is simply far too big.
This is like driving your car in top gear at idle. Very lumpy and any small hill and it can't maintain a steady speed.
You can do things about it but it would be really good to know what the normal discharge pressure is and what speed you find yourself at at 20 gpm.
Remember - More details = better answers
Also: If you get a response it's polite to respond to it.
This is like driving your car in top gear at idle. Very lumpy and any small hill and it can't maintain a steady speed.
You can do things about it but it would be really good to know what the normal discharge pressure is and what speed you find yourself at at 20 gpm.
Remember - More details = better answers
Also: If you get a response it's polite to respond to it.
LittleInch
Petroleum
I really don't understand your equation given that maximum differential head is only 72 feet for the largest impellor ( you don't say which size impellor you have)?? Unless your speed is a fraction of max speed, but even then doesn't add up to me.
However as below, your 20GPM is just too far left on the curve. I don't know what your head loss is, but if 5 or 6 feet is enough to change flow then total can't be much more than 30? or lower. Hence your motor, if max frequency is 3450, will be very low, with low torque and low levels of adjustment.
So what can be done?
Basically I think you need to possibly increase your resistance to flow to reduce the impact of change of level plus increase the flow through the pump.
1) can be done by adding some lengths of tubing
2) can be done by having a long run of piping teeing off the discharge which you can calculate using the tables in the thomsen catalogue to equal your head loss in the cooler section and return it to the tank. I don't know if milk wouldn't like to be sheared across a control valve so this is the easy way, but a valve return can be done this is a lot simpler.
3) buy a smaller pump
Remember - More details = better answers
Also: If you get a response it's polite to respond to it.
However as below, your 20GPM is just too far left on the curve. I don't know what your head loss is, but if 5 or 6 feet is enough to change flow then total can't be much more than 30? or lower. Hence your motor, if max frequency is 3450, will be very low, with low torque and low levels of adjustment.
So what can be done?
Basically I think you need to possibly increase your resistance to flow to reduce the impact of change of level plus increase the flow through the pump.
1) can be done by adding some lengths of tubing
2) can be done by having a long run of piping teeing off the discharge which you can calculate using the tables in the thomsen catalogue to equal your head loss in the cooler section and return it to the tank. I don't know if milk wouldn't like to be sheared across a control valve so this is the easy way, but a valve return can be done this is a lot simpler.
3) buy a smaller pump
Remember - More details = better answers
Also: If you get a response it's polite to respond to it.
LI where did that curve come from. What are the red lines? It seems that he doesn't have more than a max of 10 ft of system head and at startup it might be -1.5 You can see that our suction head (Hs) varies between 0 and 5.5 feet and our discharge head (Hd) varies between 4 and 10 feet.
I think he needs the VFD so he an run it, but at a nearly off speed.
Technology is stealing American jobs. Stop H1-Bs for robots.
I think he needs the VFD so he an run it, but at a nearly off speed.
Technology is stealing American jobs. Stop H1-Bs for robots.
LittleInch
Petroleum
Curve came from the vendors website. The red lines seem to be power requirement.
He's referring to static head only. We still have no idea what the friction losses are through his equipment at 18 gpm
Remember - More details = better answers
Also: If you get a response it's polite to respond to it.
He's referring to static head only. We still have no idea what the friction losses are through his equipment at 18 gpm
Remember - More details = better answers
Also: If you get a response it's polite to respond to it.
LittleInch
Petroleum
gaalfred - this post seems to have petered out which was unfortunate as there still seemed some unanswered questions / issues. Any further data / response?
Remember - More details = better answers
Also: If you get a response it's polite to respond to it.
Remember - More details = better answers
Also: If you get a response it's polite to respond to it.
- Status
Similar threads
- Locked
- Locked
- Question
- Locked
- Question