Papers accepted at Energy Systems Conference 2016
Papers accepted at Energy Systems Conference 2016
Associate Director Professor David Wallom, Research Associate Ramon Granell and Colin Axon, Institute of Energy Futures, Brunel University, are the authors of two posters accepted at the Energy Systems Conference 2016, which is taking place on 14-15 June 2016 at the QEII Centre in London.
The first, Which British SME might benefit from electricity dynamic tariffs? examines whether SMEs would get financial benefits by switching electricity tariffs:
Dynamic electricity tariffs may incentivise consumers to modify their consumption habits or patterns, thereby reducing their costs and helping balance the peak load of the grid. Although the consequences of dynamic tariffs for residential consumers have been analysed, little attention has been paid to small and medium enterprises (SMEs). We study whether a SME will get a financial benefit or not when they switch from a fixed-price tariff to a dynamic one, using 30-minute smart-meter readings of more than 7500 British SMEs. These businesses represent 44 categories, but we aggregate them into five sectors: Entertainment, Industry, Retail, Social and Other.
First, we perform a comparison between the 'winning' and 'losing' SMEs based on their daily load profiles in three distinct tariff-change scenarios. The differences between these patterns of consumption indicate the behavioural changes that a SME can make to obtain benefit in each of these scenarios. Secondly, we try to predict which SMEs will win or lose from these tariff changes making use just of simple pieces of information that can be found in their electricity bill. We use three machine learning classifiers to solve this classification problem: Support Vector Machines, Artificial Neural Networks and a Naïve Bayesian classifier. We achieved accuracies around 80% for this binary problem. We extended this to include a price elasticity factor (a three-classes problem). Furthermore, we use linear regression and Support Vector Regression models to estimate the exact quantity of loss or profit.
Our analysis suggests that most of SMEs would reduce their annual electricity bill by no more than 10%, setting a realistic upper limit for cost savings from switching without deploying energy-saving interventions.
The second paper, also co-authored by Dr Kathryn Janda of the Environmental Change Institute is titled Does the London urban heat island affect electricity consumption of small supermarkets?:
Using hourly electricity demand data from small supermarkets, we have conducted a pilot study to understand if it may be possible to observe the Urban Heat Island (UHI) effect on power consumption. Urban warming of large cities is broadly observed as a temperature gradient decreasing away from the centre. Typically the UHI has been quantified using bespoke sensor systems with a modest number of sites.
Initially, the difficulty of quantifying the UHI effect on the electricity consumption of retail premises is the requirement to compare consumers with similar features. We have performed clustering and other statistical analyses of the electricity consumption of small supermarkets located across the Greater London area. The comparison is performed using 1-hour resolution electricity readings of 40 supermarkets of the same company during one year. We have computed the electricity consumption by area (kWh/m2) of each store to compare them during all year, Summer or Winter only, and for both 24h and the opening times (only). We have data-mined the daily load profile of the stores and the heating/cooling degree days.
Preliminary results show that we can distinguish two groups of stores concentrically located with a statistically meaningful difference in summertime power consumption. The higher mean normalised consumption is associated with stores closer to the city centre with a mean distance of 7.2±2.0 km, and the lower mean normalised consumption group with a mean distance of 11.8±1.2 km. These consumption differences are higher during Summer, suggesting that the cooling system may be the responsible as stores that are closer to the centre are warmer than the ones further to the centre. Location differences are also found when we compare the stores by their profile behaviour.
Wallom, Janda and Granell are part of the Working with Infrastructure, Creation of Knowledge, and Energy strategy Development (WICKED) project team, which bridges efficiency and performance "gaps" between how retail buildings perform in practice and in theory through scientific research on buildings, energy and organisational behaviour.
The retail sector is the largest commercial property sector and a vital part of the UK economy. Valued at over £300 billion, it accounts for one in 12 companies and employs one in nine working people. Businesses in the sector are diverse, ranging from multinational corporations to small independent stores. Across this diversity, the sector as a whole faces a number of challenges, including the global economic slowdown and the growing problem of energy management.
WICKED's cooperative research programme uses a mix of partner engagement, big data analytics, and empirical research to investigate the retail sector at multiple scales. Building on this evidence base, it co-designs market-ready energy strategies for diverse user groups.