regulating valves

[Torfinn]

Hi,

i have made a PLC programing for a CO2 cooling system,

i just cant get the valves for the gas cooler to work right.

first of im using a M340 Schneider electric PLC,
everything else on the system is working as it should,

there are 2 condensers for cooling the CO2 gas and there is one
regulating valve on each condenser,

these valves are have valve drivers which i send a 0-10v signal to

the problem is that when i make any change in the 0-10v signal to these drivers it takes 1-2 seconds until the driver starts moving the valves to the corresponding position.

the valve types are CCMT-24
the drivers are superheat controler EKE 1A
these are from Danfoss.

currently i have PID controllers in the PLC programming for these Valves.

this 1-2 second "delay in reaction time" from the drivers is to long, the pressure increases very quickly and we hit the safety limit for the machine which is set at 105bar the pressure setpoint is 90 bar.

pls any help is welcome and any further questions are also welcome

than you in advance

 

[robertmee]

If the rate of change on your PV is much faster than the physical CV device, not sure what you can do. You can tune the PID to be very responsive but that 2 sec lag in the device is going to kill you. You need faster drivers. Is there something inherent in these models that make them slow to react?

Edit...took a look at the EKE specs. I assume you're operating it in driver mode, since you're sending it the position from the PLC? Make sure the mode is correct. The EKE has its own PI controller...any reason you aren't using that instead of the PLC? In any case, there's nothing in the specs to indicate an inherent lag to response. In fact in the 485 section, it says response is 50ms.

 

[Torfinn]

Hi,

thank you for the response.

the EKE is set to driver mode,

we didnt buy this product, the costomer bought it and said "make it work"

this always ends in problems but here we are.

we have talked directly to Danfoss and we have gotten varying responses from the people we have talked to some say there shouldnt be a delay some say we have the wrong product because of the delay, so i dont know.

i have recommended changing the valves but the costomer isnt onboard in changin the valves and drivers.

i know the 2 sec lag is gonna make this very hard to get working.

btw we are operating at 90 bar in the gas cooler evaporators and 40 bar out from the valves so getting new valves that work at these pressure differences is hard(i have been told)

so just wanted to see if any one here maybe had a programming solution to my problem.

the reason we arent using the EKE as an controller is a question for our costomer(our costomer is another company that has built this machine from scratch and has ordered our services to make it work) so we are limited in what we can do physicaly to the machine without telling them first.

 

[Tom Jenkins]

I run into this all the time. The problem is often the positioner is set up with excessive deadband to avoid hunting. The result is lousy control. The best solution I've found is provided by Rexa, but if the customer wants to "save money" and put up with lousy control you are kind of stuck.

 

[robertmee]

I run into this all the time. The problem is often the positioner is set up with excessive deadband to avoid hunting. The result is lousy control. The best solution I've found is provided by Rexa, but if the customer wants to "save money" and put up with lousy control you are kind of stuck.

We've been using Rexa Hydraulic controllers for 30 years...solid product.

 

[Torfinn]

thank you for responding,

we are gonna try to use the feed forward function today to try and see if it helps stabilize the system alittle bit.

we will use the feedforward to look at the pressure in the tank that then compressores pull from because when the pressure in the tank rises the compressors will speed up and the pressure will then rise over the gas coolers.

i will write back here if it helped

thank you all who responded.

 

[BryanG]

I decided to look up for information about the valve, it uses a stepper motor, not a solenoid valve, so you are stuck with how fast the stepper motor can move the valve. In the picture you can see that the CCMT-24 has 1400 steps from fully open to fully closed, and from a constant voltage drive (am guessing that that is the 0-10V) it moves at 150 steps per second. So that is over 9 seconds from open to closed. You are going to struggle to program your way out of that for a system that change very quickly.

 

[Ken Moore]

I understand the pressure conditions make it hard to get a new valve. But...can you swap out the stepper positioner for something more responsive?

 

[Torfinn]

we have now implemented the Feedforward function and it has helped making the system run smoother, but it doesnt fix it in anyway(i just hoped it would help so i am happy with this ).

the problem isnt the speed of the steppervalve 9sec opening time is not a problem because the machine is made so big that this ship should be able to make ice in the middle of the sahara dessert with cooling water at 32 degrees celsius, but we are in the Faroe Islands, it doesnt get that warm around here so the cooling water for the machine is 22 degrees celsius and the ocean water that is use for cooling the machine cooling water is around 3.5 degrees celsius,
because the machine is so big(oversized?) in comparison to where we are, the machine is always running at max 50% but now in winter time it is running somewhere between 4-25%.

the problem is that the jump from 4% to 25% happens so fast because of the larger consumers on the machine having -
HBX vapor quality sensors from HB-Products which request 130KW of cooling from the machine, these have been slowed down as much as we can but it still happens quite fast, the return gas from these consumers goes into our middle tank which is quite small compared to the machine(only 200Liters - the suction tank for reference is 1200liters), because of this part of the machine being small things have to happen quicker than on a normal cooling machine and the valve drivers for the gas cooler valves have a 2 second delay from receiving an input command(0-10V) and actually starting to move the valve to the corresponding position(0-100%).
i have been onboard to check this by forcing the output on the PLC and measuring the voltage on the input terminals on the valve drivers and listening to the valve(because it is a steppervalve we can hear it "step" in both directions)


but i think now that we have put the feedforward function in the system should run smoothly enough for the ship to go out sailing(they will be leaving on monday) and then we will just tell them to have to find new parts for the machine if they want it to run Smoother/Better.

the Rexa Hydraulic controllers look interesting and i will definetely look at those, thank you to @Tom Jenkins and @robertmee for that

thank you all for your inputs and questions

 

[drbitboy]

the problem is that when i make any change in the 0-10v signal to these drivers it takes 1-2 seconds until the driver starts moving the valves to the corresponding position.

A slow stepper drive might be an operational problem, but it is very unlikely to be the cause of this delay i.e. even a slow stepper should start moving the value immediately.

Look hard into what @Tom Jenkins suggested: a programmed deadband; it could be in the valve positioner itself, or it could be configured in the PID.

If an intentional deadband is not the cause of the delay, then check if there is any play or lash in the mechanical parts.

https://www.plctalk.net/qanda/showpost.php?p=844583&postcount=4