The peak-load problem in terraced heat networks
Shared ground-loop heat pumps often fail not because the borehole field is too small on paper, but because the control strategy synchronises compressor starts. In a 24-home terraced retrofit, morning setpoint recovery pushed loop return temperatures below 2°C within 18 minutes, triggering 36 kW of electric backup. The design load was 210 kW; the measured peak reached 242 kW because domestic hot water and space heating called simultaneously. The takeaway is that loop sizing without start sequencing leaves carbon savings on the table.
Why weather compensation alone does not solve it
Weather compensation adjusts flow temperature to outdoor conditions, yet it does not manage the number of compressors drawing from the same loop at the same time. The cause is visible in the control logic: every thermostat uses the same morning setback recovery, and cylinder sensors call for reheating after the same evening draw. Hydraulic separation at each dwelling reduces pressure interference, but it cannot prevent the aggregate thermal pulse. A 15-minute stagger across 24 homes would cut the coincident peak by roughly 6.2 kW per dwelling group, according to the same monitored data.
The bananzia sequencing layer
The bananzia sequencing layer addresses the aggregate pulse by treating the network, not the individual dwelling, as the controlled unit. It assigns each heat pump a start slot based on cylinder temperature, room temperature error and loop return temperature; compressors with the lowest return-temperature penalty are deferred. In the 24-home trial, the controller limited simultaneous compressor starts to eight and gave domestic hot water priority for 40 minutes after the evening peak. The result was a 31% reduction in loop-side peak demand and a 22% drop in backup electric resistance use.
Governance and metering requirements
Sequencing only works when the network operator can see dwelling-level heat demand without overreaching into occupant comfort. The trial used 15-minute heat meters, cylinder temperature probes and a simple occupant override capped at two hours per day. The governance rule was explicit: no dwelling could be deferred below 19°C, and hot water had to recover within 90 minutes of the evening draw. These constraints matter because a control layer that hides discomfort simply shifts the problem to resident behaviour and undermines the low-carbon case.
What the measured result supports
The strongest evidence from the trial is not a single efficiency figure but the relationship between start sequencing and loop temperature stability. When simultaneous starts were capped, return temperatures stayed above 1.5°C and the heat pumps maintained a coefficient of performance of 3.1 rather than 2.6 at peak. For terraced streets with limited borehole space, that difference is more useful than a nominal efficiency claim. Bananzia is therefore best understood as a load-sequencing method for shared loops, not as a standalone heat source.