Smart energy
Energy management for Small and Medium Enterprises
We applied recurrence quantification analysis (RQA) to energy data obtained from monitoring equipment on two small enterprises and came up with a new way of optimising parameters in order to produce maps of 'usual' behaviour. Our partner is ANDtr, and the first project RAE was funded by Innovate UK, and the second RAE2 by Business, Energy and Innovation Strategy (BEIS) Dept. of UK government.
[1] Hattam, L., Vukadinović Greetham, D.(2018): Energy Disaggregation for SMEs using Recurrence Quantification Analysis ACM e-Energy 2018, Workshop on Energy data analytics, Karlsruhe June 2018
[2] Giasemidis, G. and Vukadinovic Greetham, D. (2018). Optimising Parameters in Recurrence Quantification Analysis of Smart Energy Systems. In: 9th International Conference on Information, Intelligence, Systems and Applications (IISA2018), 23-25 Jul 2018, Zakynthos, Greece.
Short-term individual electricity demand forecasts
Forecasting at the household level is extremely useful for control, demand side response, and energy storage. At individual level, creating reliable forecasts is much more challenging than the typical forecasting which is performed at the smoother, more regular high voltage (HV) level. Accurate forecasting at the LowVoltage (LV) level help DNOs to manage and plan the network, including considering the risks of the higher uptake of low carbon technologies in the future. During the New Thames Valley Vision project, funded through OFGEM's Low Carbon Network Fund and I was involved in development of a new technique to evaluate different forecasts at the individual level that can cope with volatility (or 'peaks') in the data. [1] We also created several new 'peaky' forecasts methods.[1,2].
[1] Haben, S., Ward, J., Vukadinovic Greetham, D., Singleton, C. and Grindrod, P.(2014) A new error measure for forecasts of household-level, high resolution electrical energy consumption. International Journal of Forecasting, 30 (2). pp. 246-256. ISSN 0169-2070 doi: 10.1016/j.ijforecast.2013.08.002
[2] Charlton, N., Vukadinovic Greetham, D. and Singleton, C. (2013) Graph-based algorithms for comparison and prediction of household-level energy use profiles. In: IEEE International Workshop on Intelligent Energy Systems, 14 Nov 2013, Wien, pp. 119-124.
[3] Haben, S., Rowe, M., Vukadinovic Greetham, D., Grindrod, P., Holderbaum, W., Potter, B. and Singleton, C. (2013)Mathematical solutions for electricity networks in a low carbon future. In: Electricity Distribution (CIRED 2013), 22nd International Conference and Exhibition on , 10-13 June 2013, Stockholm , 0857-0857. doi: 10.1049/cp.2013.0972
Modelling uptake of low-carbon technologies
Long-term forecasting of individual electricity demand is obviously more challenging than short, but it has the additional complication that in the near future, various types of low-carbon technologies, such as electric vehicles, heat pumps and photovoltaics are expected to be widely adopted in the UK. The effect of these technologies on the existing low voltage networks will depends on spatiotemporal characteristics of the uptake. Among other factors, it will depend heavily on socio-demographics and social influence (so called ‘Jones’s effect’) [5].
To forecast the uptake of low carbon technologies (LCT) over the long-term, we developed the methods that explicitly model the social influence of neighbours, where the probability of adopting LCT is being proportional to the number of neighbours that adopted previously. Socio-demographic information is used for the initial seeding of LCTs [3]. This allowed us to test different scenarios of adoption of single and multiple low carbon technologies in Bracknell and to quantify the uncertainty of continuous high demand periods through the confidence bounds for different parts of LV network [3]. In addition, this enabled us to identify unwanted effects of global policies on local networks [4]. A hybrid model of uptake that combines differential equation-based modelling on a macro-top level with Monte-Carlo type simulations on a household (micro-) level was also developed that allows for longer-time scale predictions [2].
[1]Vukadinovic Greetham, D. , Hattam, L.(2019) Green neighbourhoods: the role of big data in low voltage networks' planning, in eds. Emrouznejad, Ali and Charles, Vincent, Big Data for the Greater Good, 42, pp. 151-169
[2]Hattam, L. , Vukadinovic Greetham, D. (2018) An innovation diffusion model of a local electricity network that is influenced by internal and external factors Physica A:Statistical Mechanics and its Applications, 2018, Volume 490, Pages 353-365
[3] Hattam, L. , Vukadinovic Greetham, D. (2017) Green neighbourhoods in low voltage networks: measuring impact of electric vehicles and photovoltaics on load profiles J. Mod. Power Syst. Clean Energy, 5: 105. doi:10.1007/s40565-016-0253-0
[4] Hattam,L. , Vukadinović Greetham, D., Haben, S., Roberts, D. (2017) Electric vehicles and low voltage grid: impact of uncontrolled demand side response. CIRED, 24th International Conference on Electricity Distribution, 12-15 June 2017, Glasgow.
[5] Poghosyan, A., Vukadinovic Greetham, D., Haben, S. and Lee, T. (2015) Long term individual load forecast under different electrical vehicles uptake scenarios. Applied Energy, 157. pp. 699-709. ISSN 0306-2619 doi:10.1016/j.apenergy.2015.02.069