Initial assessment of voltage-led demand response from UK residential loads

Abstract

One way of unlocking aggregated demand response without directly managing customers is to exploit the positive correlation between voltage and demand. This work presents an initial assessment of the potential residential response in UK primary substations (e.g., 33/11 or 33/6.6 kV) as part of the trial "Customer Load Active System Services (CLASS)" run by the distribution network operator Electricity North West Limited (ENWL). Two load modelling methodologies, time independent and time varying, are presented and applied to a single primary substation and then extended to the ENWL area. For the latter, the results considering the time-varying model, purely residential load, and 3% voltage reduction during a winter day show that it could be possible to achieve an aggregated peak reduction exceeding 150 MW (>3%). Although the two load models have shown strong agreement the latter is a more accurate alternative when demand response throughout the day is of particular interest.

Assessing the effects of load models on MV network losses

“…It is of interest to remark that the three classic load models can be described by both the np (1) and ZIP (2) coefficients as shown in Table III. The fourth load model, adopted in this work as a benchmark, is a time-varying (i.e., changes through time) ZIP model designed for the UK residential sector [9]. To develop such model firstly a literature review has been carried out in order to define a ZIP model for each single appliance that could be found in a common UK dwelling [14].…”

“…Thereafter, by adopting the CREST tool, the load profiles for each dwelling have been stochastically generated. Finally, the daily ZIP model for each dwelling is calculated by aggregating each single appliance load model and profile [9]. This approach, belonging to the category of componentbased load model, captures the variability in the load composition and consequently the natural variation of the load response to voltage changes [6].…”

“…On the other hand, the time-varying ZIP model describes a more realistic voltage demand relationship considering its time-varying nature and also catering for the possible diversity (due to the randomness on the dwelling profile selection) from one secondary substation to another as discussed in [9]. An example of the obtained time-varying ZIP parameters for one secondary substation is shown in Fig.…”

“…In [8] the impact of different load models on size and location of DG was assessed. However, the adoption of a syntehtic network and the use of a load model that neglects the time-varying changes in load composition (i.e., static) [6,9] limits the confidence on the final outcomes.…”

“…A 10-min time-series analysis during winter and summer days is undertaken. A time-varying ZIP model [9], specifically designed for the UK residential sector, is adopted as benchmark to evaluate the performance of the three classic load models.…”