contrologix compressor staging

[jerryd68]

I have a system that i need to automate, I has a total of 7 refrigeration compressors, all with capacity control, the system needs to run based on suction pressure, if the pressure rises above the set point more capacity needs to be added, a total of 17 diffent stages, I need to be able to run any of the machines as the lead machine and add the capacity in sequencial order from that point, i.e if #5 machine is set as lead, then it would load, then start #6 machine, load it and so on. I have a contrologix monitoring the pressure and starting and stopping the compressors, I have a panelview plus as the HMI, any suggestions as to the easiest and most effective way to accomplish this would be helpful,
thank you all

 

[Lancie1]

Jerry,

How will the lead compressor be selected, manually or automatically? What is the period that the lead compressor runs before it is swapped for another one?

You could set up a sequencer, with the (7 things taken 7 ways = 49) patterns stored as bits. When a new compressor becomes the lead, the proper pattern is selected, and each time the suction switch is triggered, the sequencer advances to the next bit pattern.

You could do the sequence with counters and bits, similar to the attached RSLogix program for 5 compressors.

 

[abwiz]

Why don't you tell us what you have come up with so far and maybe we can give you some input.

 

[jerryd68]

The lead compressor will be selected from the HMI manually by the operator, I have not done a lot of work on this yet, I have been considering a bit shift in both directions, but i am not real familiar with these instructions, one thing that i forgot to mention is that the system needs some control on timing between adding the capacity, if it is loaded to quick it causes problems, I have a PID loop set up watching the suction pressure, I would like to use the output from this instruction,

thank you all once again

 

[TConnolly]

One thing to keep in mind is that compressors require short cycle protection. When a compressor is shut off any refrigerant in the crankcase boils and turns the oil to foam. When the compressor starts back up some of this oil foam is blown out into the refrigerant line. Refrigeration systems are designed to accomodate this, but the control engineer must also accomdate it. The oil is eventually returned to the crankcase via the suction line so the compressor must remain on long enough for the oil to complete the circuit. Should the compressor operate through several sequential short cycles then the loss of oil foam to the lines may be too great, leaving the compressor crankcase improperly lubricated. When a compressor turns on it needs to remain on for a while - this amount of time will depend on the system, but five minutes is typical, longer it the line sets are long. Once a compressor is turned off, it needs to remain off for several mintutes. This can complicate staging, but it needs to be included.

 

[jerryd68]

Alaric,

I agree with you on short cycling on a typical freon set up, these compressors that I am working on are large industrial ammonia compressors, they have heaters in the crankcase to keep the refrigerant boiled out during off time, also each compressor has a oil seperator on the discharge line with a float valve that returns the oil to the crankcase, with ammonia the oil is typically drained off at the low points and does not circulate with the refrigerant, but I do agree that starting and stopping to often is a problem that will need to be addressed.
thank you

 

[Jason_Adams]

You could use a simple program that says if motor one is on and you need to start a new compressor, start motor two, etc. Have motor one start if a new motor is needed and motor six is on to make the circle.

Or, preset a counter to six, reset it if it equals 7. You now have motor 0-6, index the counter every time a new compressor starts. Use a TOFF to prevent a new start until the neccessary time has elapsed for new capacity.

 

[Gerry]

For a suggestion on the HMI selection screen, look at this thread over at MRPLC.
My mention of patch cords in that thread is a reference to a compressor staging project I did in 1979. We had a mixture of reciprocating and screw compressors on 3 pressure systems. My approach was to treat all the compressors on a single system as one big compressor for the pressure PID loop. The required output was then built up using the lead/lag order and the different capacities of the compressors.

Other logic attempted to keep screw compressors above 45% load. e.g. If the leading stage was at 100% and the additional output would load the lagging screw to 10%, then the leading stage would be unloaded sufficiently to allow the screw to load to 45%. This was based on the idea that screw efficiency below 45% was horrible.

 

[jerryd68]

Gerry,
I will look at the HMI application, what was your PLC code like? Sounds like you did use a PID to add capacity or reduce capacity based on load, this is what i have been working on it, I would like to be able to sequence the compressors in whatever order that is needed, without having to write many individual routines to accomplish this
thank you

 

[Gerry]

Gerry,
I will look at the HMI application, what was your PLC code like? Sounds like you did use a PID to add capacity or reduce capacity based on load, this is what i have been working on it, I would like to be able to sequence the compressors in whatever order that is needed, without having to write many individual routines to accomplish this
thank you

The control system was a locally-built micro-processor (8085) system that I programmed in assembler. I later converted it to a "PLC" format on the same hardware. The control wasn't actually PID - more accurately P + bias. Since it was 28 years ago, my recollection is somewhat hazy. The LLP system had a single reciprocating Howden, the LP had two big Howden screws (1000 HP), and the IP had two reciprocating Hall's, a reciprocating Sabroe and a Sabroe screw. The system also controlled liquid pumps and pot levels.
On the IP system, for example, if the Hall's were 50 ton each, the Sabroe's 100 and 200 ton respectively, then I treated them as a single 'virtual' compressor of 400 tons. Bias was set to 200 tons and the P gain set to swing the output +/- 200 tons for an acceptable range around the desired suction pressure. I'm sure you could probably do better using the PID functions in CLX - I was working with crude tools and flying by the seat of my pants.

For loading, your controller tells you how many tons it wants, then you look up who's first and how many tons it can provide and in what increments. If more is required, look up who's next and its capacity, etc. The operators need to determine which compressors should lead based on the anticipated load for the day - ideally, the lead compressor should run continuously at or near full load. On heavy days, the lead should be the largest compressor and on light days, a smaller compressor should lead.