DP Buffering and Pump Control

From Heatweb Wiki
Jump to: navigation, search

As the design and quality of heat networks improves, we are starting to see system flow rates drop, as unnecessary bypass flows are eliminated, and we are left with the bare load only - hot water, heating, and pipework losses.

For individual buildings and blocks, the low load flow rates can regularly drop to zero overnight, with only the thermal bypasses on risers maintaining any flow at all, and this in the low litres per hour.

Where previously, pumps were rarely faced with no-flow conditions, now they can see them regularly, and the impact of going instantly from no load to a few taps running can be a significant drop in differential pressure, as pumps will need to wind up to the load.

In other words, we are looking at the kind of rapidly changing loads that domestic booster pumps are faced with, and therefore a need to address this the same way with pressure accumulation - expansion vessels where pressure stored in compressed gas can drive significant flow rates for short periods without the pump assisting, and giving the pump time to wind up.

And in the design of pump sets there is a complication imposed by pump technology. The smaller canned rotor pumps used for single buildings only go up to 180kPa, or lower at load, where a heat network may require up to 2.5 bar, requiring the use of larger pumps that introduce significant minimum flow requirements. We aim here to allow use across the board of the smaller canned rotor pumps, or a combination of pumps in parallel that are compatible in pressures.

File:alphabuffer.svg


We have been carrying out a study on real world data from systems {see DHW Duration} and now we have 1 second resolution we can see the reality of DP on the network, and it consists of an underlying trend based on time of day (overall DHW loads) which changes over minutes, as well as a higher frequency pattern from individual taps. The pressure drops lasting under 15 seconds are greater than those over an hour - too rapid for pumps to respond to.

This article looks at a new method of controlling pump head on a heat network.

Advantages:

  • Increases peak DP to 3 bar.
  • DP to system and pumps cannot change rapidly - vessels absorb rapid pressure changes
  • Pumps target the average DP, rather than instantaneous, so have a lower setpoint
  • The pumps can turn on and off without hammering - never a true dead head
  • When pumps are off there is a reservoir of DP to handle trickle flow
  • Lower pumping energy - no 10% parasitic flow at all times, and lower overall pump DP.
  • When flushing - higher flow rates can be achieved for short periods
  • Lower peak static pressures
  • No need for bypasses around pump
  • Same family of 180kPa pumps regardless of flow
  • Pump heads are compatible and allow a small pump in parallel with a big pump for greater range or control


Such systems have not been considered before because we have never had to design for 'water tight' networks where there can be zero flow. Now we are, the same mentality as used in booster set design is relevant - i.e. pressure accumulators to prevent hunting and hammer. (we also used to manufacture pump sets).

The large expansion vessel for the main buffer creates the neutral point and decouples the pumps, so flow rates can be dissimilar for short periods.

For control - we would use the system DP to sequence pump in the same way we use thermostats on a buffer to sequence (and rotate) boilers.

Typically one would expect the flow pump to be generating 100kPa and the return pump at 50kPa - we would never have only one pump running so a pump is never pushing through another pump (although a non-return valve around the return pump would enable this mode). They just sit at different points on their curves - same flow, different power.