Parallel Roof Top Units 10

[MCSEng]

Hello All,

I am contemplating a ventilation and cooling system retrofit where two (2) 20 ton roof top packaged units are connected in parallel.
i.e. The discharge and return ducts are interconnected so that a total cooling capacity of 40 ton is available.
The rationale is that in the event of failure of one unit, some amount of cooling can still be provided to some areas.


I believe that the external static pressure and CFM for both units should be the same.
What are the other issues/factors that should be considered when connecting two packaged units in this way?

Does anyone know of any literature that can provide guidelines that should be followed in order to ensure the best performance?

Thank you.

 

[trashcanman]

1. Have you checked with the RTU Vendor (Carrier, Trane, ect.) to verify that this will work?
2. How are the units controlled? 1 thermostat? Is it connected to both units? 2 thermostats, 1 to each unit? How is capacity controlled and by which thermostat? You will also need motorized dampers for supply and return air ducts.
3. It does not matter what you believe about the e.s.p. and cfm, Calculate the answers.

Check with ASHRAE Handbooks of Fundamentals and Equipment.

 

[KiwiMace]

Hi,
There should be no major technical hurdles in getting to what you want. Paralleling equipment for partial or full redundancy is done all the time. You will also get better turn down and possibly improved efficiency at low loads.

If the system is vav downstream, the flow and temperature control is straightforward. A constant flow system will be improved with a balanced main duct arrangement.

Probably the most foreseeable headache is maintaining the code required physical separation distance between your outdoor air intake and exhaust openings at roof level.

The roof waterproofing system should integrate with the unit curbs, or put equipment on steelwork to gain access underneath. Coordinate early with the structural engineer.

 

[Drazen]

With variable speed fans and VAV at point of use, that should not be overly complicated as Kiwi said, though there are number of issues to be checked.

With CAV system and constant speed fan I am not sure how that could work in case of failure of one unit. Both volume flow and pressure change and terminal units would mostly not be able to cope with that, cold air drop or drafts could easily occur somewhere in the building.

 

[cdxx139]

As stated by trashcan, RTU isolation dampers are critical.

knowledge is power

 

[urgross]

Consider being able to run both units simultaneously at lower load, and include isolation dampers in the event one unit fails. The set up you are describing is often used for computer centers.

 

[cry22]

as others have said. with VAV, you could work it out by indexing all VAV boxes to min so the entire system gets "some air" and reuiqred ventilation.
CV, half the building will starve for air, as the first diffusers will take all the air flow and no air will reach the last diffusers.
So regadrdless, parrallel or not, for CV, half the building will be without cooling, might as well go with 2 RTU's, each serving it's area and keep it simple.
If CV - this is a cheap design.

 

[317069]

Are these RTU's already connected in parallel or you want to connect them?
KiwiMAce : what did you mean by" balanced main duct arrangement."
do you mean a bypass duct as VVT systems or something else?

 

[KiwiMace]

I meant literally, symmetrical or otherwise balanced to allow an even supply of air from the two units in normal operation, and predictable operation from either unit in fail mode.

 

[DRWeig]

Kiwi and others,

I encountered a parallel ahu system once, long ago, that would become unstable at times. The two fans seemed to be trading the load at a regular interval. The current on one fan would rise as the other fell, then back again -- much the same as an LC resonant circuit. The fans were about 15 meters apart, blowing into a large supply plenum.

I wasn't around to hear what the solution was. Is there a way to prevent or predict this sort of resonance? Or did I just see a freak event?

Best to you,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[317069]

let us say that our building load is 30 ton:
- What if we connect a 15 tons unit with a 15 tons unit
- What if we connect a 20 tons unit with a 10 tons unit
- What if we connect three 10 tons of each unit.

 

[KiwiMace]

Dave, That kind of instability most likely came from a cock-up in the control algorithm, making the vfd's fight to obtain a setpoint. Natural fan instability occurs when operating in the double hump region, but as far as I know this is not going to happen in a reasonable sized centrif, and it is characterised by switching between operating points, not ramping.

 

[DRWeig]

Thanks Kiwi.

Best to you,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[ChasBean1]

Good responses by all above, but one thing I haven’t seen is building load versus unit size. Be wary of it with packaged DX units. Unless cooling/dehumidification and a specific supply air flow rate is mission critical, I’d suggest undersizing each unit quite a bit as DX units do not like low or stagnant airflow. Size each unit for maybe 60% of design.

Additionally, units need to match speeds (or volume vanes, or VAV bypass control damper) commands when operating in parallel. Otherwise one unit will tend to handle the load while the other idles. I’ve seen “alpha” units at full speed and parallel units at minimum, actually being back-fed. You will need a building controls programmer that understands the goal and a BACnet or Lon interface that can receive inputs and provide outputs to the packaged units.

 

[Esteves]

If this is a VAV system you'll definitely want to pay attention to how the fans are controlled. These will modulate to maintain a static pressure set point from one sensor. You could come up with a control scheme that allows the fans to "know what the other is doing", however if you're using "packaged units" the KISS method would be to have one VFD control 2 fans. The downside is now you have a single point of failure so you'll want to specify a VFD with a bypass. 3 phase inverter duty synchronous inductor motors are nice. Also you'll want to let the VFD vendor know what you're doing to make sure they're sized correctly.

 

[MCSEng]

All,

My query sparked some very good discussions. I have learnt that vac design is not just about determining the heating or cooling load and selecting the equipment. It has a lot to do with devising the best strategy and balancing acquisition cost, energy consumption, system reliability etc.

Thank you all for your reply.

Trashcanman: Your query caused me to give a lot more thought to the controls. These days control systems offer a lot more flexibility to allow programming an elaborate control scheme. I will give considerations to the operational logic.

Cry22: The system will be VAV. We intend to pay a lot of attention to testing and balancing so I expect that starving parts of the building of air will not be an issue.

All: There would have to be some amount of programming logics. In the event of failure of one RTU, the VAV boxes in the non-priority areas would be driven to minimum position (partially closed): still allowing some air flow. The main dampers isolating the failed unit would close and the priority area would continue to receive cooling.

Given the potential cost for engineering programming time, I have decided to explore other alternatives as well. Thank you all for your feedback.