Learnings

Outcomes

The key learnings from the project are detailed below:

•        At the individual property level, electricity peak demands could be between 4 to 6 times higher (dependent on the property archetype) than peak non-thermal demands if a property switches from gas/oil/LPG heating to electrically driven heat pumps. Total electricity demands were higher in larger homes with poor insulation.

•        Under cold conditions, many of the flexibility measures investigated had a minimal impact on peak demands, and often resulted in peak-shifting rather than peak reduction at the individual house level.

•        At the network level, the study found that for a 1-in-20 winter period, a high heat pump uptake, and no flexibility measures, approximately 44% of the distribution substations in the areas investigated would require reinforcement by 2030, and this would increase to 72% by 2040.

•        Flexibility measures were found to have a negligible or very low impact on peak demand reduction at the network level during a 1-in-20 winter period. Flexible hot water generation resulted in approximately 1% reduction in peak load, and the use of buffer tanks reduced peaks by 1-2%. Temperature flexibility had less than 1% impact on peak load reduction, and also has a higher incentive cost compared to the other measures considered, so is less cost effective. Inclusion of electrical batteries in half of homes with heat pumps provided the greatest level of peak demand reduction, at up to 9%, but the number of batteries in homes is likely to be limited based on DFES projections.

•        Outdoor temperature has a significant impact on the peak heat demand. Taking the S-200 substation archetype as an example, the half hourly peak heat demand in a 1-in-20 winter (on a day with an average temperature of -4°C) is approximately two times higher than in an average winter (on a day with an average temperature of 4°C).

Lessons Learnt

Describe how the project (methodology, stakeholder engagement etc.) changed, or provided opportunities, from your expectation at the start of the project and therefore could be useful for a future project. In addition, please discuss the effectiveness of the research development or demonstration undertaken. (15000 Characters max)

A number of lessons have been learnt on the project to date. These are summarised for each work package below.

Work Package 2 – Heat market landscaping

Heat pump manufacturers were consulted to determine the most appropriate hot water generation strategy for heat pumps. Several different approaches can be adopted, with some manufacturers recommending two one-hour generation periods ahead of the morning and evening demands and others recommending charging the cylinder in the middle of the day when outdoor temperatures are highest. For customers on a cheaper overnight tariff, hot water can be generated in these periods to reduce costs. Heat pumps can also be set up to recharge the cylinder whenever the temperature falls to 10°C below the set temperature. The difference between the actual temperature and the set temperature will determine what capacity the heat pump operates at, and hence how much current it draws. Because hot water can be stored efficiently for several hours, hot water generation is an important source of flexibility. However, it can also be a source of peaks on the network, if many homes have heat pumps set to generate hot water in the same short window each day.

Work Package 3 – Customer modelling

It was found that the analysis that was planned to be completed in Excel for work package 3 (modelling the house using inputs from AECOM) can be undertaken successfully in PLEXOS. As such, the PLEXOS model was prepared to represent the house as a heat battery, optimise heat pump space heating demand, including non-electric heating demand, and hot water / thermal storage use according to price optimisation, ensuring temperature does not drop below a certain limit.

Initial results from modelling by AECOM of heat demand profiles for homes showed slightly unrealistic ramp up times and assumptions around heating capacity. Model reruns were made to include dynamic set-point adjustment to cater for extreme cold periods and heating capacity limits, and adjustment of the heating capacity limits based on heat loss calculations.

Differences between profiles produced by the AECOM building physics model and profiles produced in PLEXOS (underestimation of heat demand in PLEXOS model on coldest days) was explained by the initial absence of modelling the effect of thermal mass of buildings in PLEXOS. The PLEXOS profiles were brought in line with AECOM profiles by applying a factor representing the effect of this thermal mass variable.

At the individual house level on a variable tariff it was found that peaks were shifted from evening periods to morning periods, rather than reduced. Peaks were higher in scenarios with electrical batteries, as these are an additional load. However, by applying limits to total electricity demand, loads could be spread more evenly across the day, and peaks could be reduced with the addition of storage. This illustrates the importance of having the right market signals to incentivise the use of storage in a way that is most beneficial to the network.

The heat demand modelling assumes that the heat pump would be able to deliver all of a property’s heat demand (i.e. additional heating through other electrical / resistive heating is not required). This assumption was based on guidance from heat pump manufacturers. Further study could include investigating the demand in existing properties with heat pumps installed, and assessing whether the heat pump alone provides sufficient thermal input to meet the demand.

Work Package 4 – Area typology modelling

At the network level it was found that, with high levels of heat pump uptake, the introduction of flexibility measures shifted peak demands from high price to low price periods on an Agile-type tariff, rather than reducing peaks. Rather than introducing electricity supply limits to counter this, the Time of Use (ToU) price was adjusted to try to encourage peak shaving rather than peak shifting. It was determined that a tariff with low overnight rates, high peak rates, and a linear change in prices between the two had the desired effect.

Preliminary results from the network level modelling suggest that electrical batteries (50% uptake among homes with heat pumps) and flexible hot water generation are the most effective measures for reducing peak loads. Flexible hot water generation should be relatively easy to incentivise for households. However, batteries are expensive investments, and it is possible that the costs might outweigh the benefits. Temperature flexibility and buffer tanks were found to be less effective.

Work Package 5 – CBA, Analysis and recommendation

Assumptions for battery uptake used in the Peak Heat project are illustrative only. In work package 4 it was found that there is not a great deal of additional benefit (in terms of reduction in peak loads) when uptake of batteries exceeds 50% of homes with heat pumps. However, it is noted that battery uptake is forecasted to be much lower in practice. DFES indicates that only around 2% of all NGED customers will have heat pumps in 2030, in the Consumer Transformation scenario. The 50% uptake has been taken as the illustrative scenario for the Peak Heat project to provide an indication of the maximum amount of flexibility that batteries could offer if a wide rollout took place.

The CBA model confirmed that pre-emptive transformer replacements, i.e. avoiding multiple upgrades of the same substation, result in improved net present value benefits compared with multiple incremental upgrades to the next size up.