I've been participating in a Policy Fellowship with the Royal Academy of Engineering, focusing on the value of heat flexibility and, crucially, the means of delivering it. To grasp the advantages and disadvantages, I experimented with some heat flexibility using my heat pump, now in its second winter season.

For this, I use Octopus Energy's Agile tariff for my heat pump operation. This tariff, unfamiliar to some, is tied to the wholesale price and fluctuates every half hour. Initially, I set my Ecodan heat pump to timed adaptive control, incorporating an off-period from 4 pm to 7 pm during peak price times – a process requiring no expert energy knowledge. However, since the start of this heating season, I've progressively advanced the controls.

I programmed a script to lower the heat setpoint to 18°C whenever the Agile tariff exceeded 30p/kWh. Then, to my dismay, during a cold November week, we experienced a Dunkelflaute event. The Agile price soared above 30p/kWh for most of the afternoon, resulting in my heat pump operating during the three most expensive hours of 2023 (see Chart – Old Algorithm).

I wrote additional code to retrieve the Agile tariff an hour ahead and incrementally boost the room temperature setpoint by 0.5°C, inversely proportional to the ambient temperature. Then, when the actual peak period begins, the temperature setpoint is lowered to 18°C.

Finally, I fine-tuned the peak period to identify the six most expensive half-hour intervals in the morning and afternoon. I programmed the system to pre-boost before these periods and reduce the setpoint during them.

All of this has worked quite well, and with the boost, even on cold days, the temperature in my front room rarely drops below 19°C. As you can observe in the chart (New Algorithm), on a day when temperatures were around freezing, the room temperature dipped to 19°C. This clearly depends on how well insulated and draft-proofed your house is, but it perhaps indicates that there is additional value to be gained from better insulation, beyond just reducing the thermal input required to maintain a comfortable temperature.

So, what has been the impact on my bills? As the figures show, despite some underperformance in November, my heat pump has operated at 52% of the cost of providing the same amount of heat with a gas boiler. I believe that if my algorithm had been more precise from the start, the cost could have been halved – let's see what a full year of data reveals. (For the gas cost calculation, I used the 'Flexible Octopus' gas tariff, assumed a boiler efficiency of 87%, and distributed the standing charge (£100/annum) across 15,000kWh, which is my typical historical gas usage.)

Annually, this translates to a saving of £600, assuming the sCoP remains at 4.1 (which it achieved last year). It's worth noting that I've conservatively estimated the same heat input from my heat pump as the gas used by my boiler, based on the assumption that the house will be, on average, warmer and thus incur slightly higher losses. To put it in government terms, if half of UK homes could adopt this approach, it would result in consumer savings of £7.92 billion per annum.

In reality, this approach would also offer broader system benefits. It could enable early deployment of heat pumps before network reinforcement is complete, as peak demand can be reduced. Additionally, it could help lower the required capacity for our future 'Dunkelflaute' generation (whether that be unabated gas, hydrogen power, or some yet-to-be-developed solution), allowing that capacity to operate at a higher load factor, thus reducing the unit cost.

However, we cannot expect consumers to manage this alone. Electricity suppliers need to step up with their offerings. They should collaborate with heat pump installers to ensure that this technology is integrated during the installation of heat pumps. Moreover, appropriate regulatory frameworks are needed to enable these developments for the broader populace.