An Emissions Arbitrage Algorithm to Improve the Environmental Performance of Domestic PV-Battery Systems

Abstract

Domestic PV-battery systems are rarely operated in a way which specifically maximizes environmental benefit. Consequently the studies that seriously examine such systems often find that the greenhouse gas and pollutant emissions savings of rooftop PV, though still positive, are lessened by adding a domestic battery. This study shows thatby simulating a PV-battery system with a range of sizes that this need not be inevitable. A novel algorithm was designed specifically to perform 'emissions arbitrage': to charge the battery when the grid emissions intensity is low and to discharge when it is high. It was found that the CO 2 saved relative to the same system with PV only can more than pay back the CO 2 debt of manufacturing the battery. This is true as long as the UK moves away from the present-day situation where natural gas-fired generators are nearly always the marginal generator. This work underlines the importance of both the operating strategy and the interactions between the system and the entire grid, in order to maximize the environmental benefit achievable with domestic PV-battery systems.In contrast, Faria et al. (2014) [18] showed that a second-life electric vehicle battery could reduce global warming, abiotic depletion, acidification and eutrophication factors by 2% when used in a peak-shaving application in France, and 4-5% in a load-shifting application (both without PV). This is because the French grid emissions intensities change throughout the day in such a way that electricity is imported from the grid when emissions are low and exported to the grid when they are high. This 'emissions arbitrage' effect was not accounted for in the other papers mentioned, which assumed a constant grid emissions intensity.A flaw in the work of Faria et al. (2014) [18] is that they used average rather than marginal emissions intensities. If some grid-generated electricity is displaced by the injection of PV power, or indeed any other intervention, not all the generator types (nuclear, coal, biomass, etc.) would have their output reduced in the same proportion as their total generation. The reduction would occur mostly for the generator type with the highest running cost. This gives rise to the concept of marginal emissions factor (MEF, as opposed to average emissions factor, AEF). Studies have shown that using AEF rather than MEF can cause errors of up to 25% [19][20][21][22].It should also be noted that the battery operating strategies studied by Faria et al. (2014) were not designed to achieve emissions arbitrage. As such, environmental impacts were negative when those operating strategies were applied to Portugal and Poland [18]. The literature is abundant with algorithms for peak reduction, cost minimization and self-sufficiency maximization [6][7][8][9][10][11][12][13]. There is good reason for this: All these objectives are quantifiable and desired by consumers, distribution network operators, or other relevant stakeholders. However, environmental benefit is also desirable, as evidenced by survey data on opini...

Reducing the life cycle environmental impact of electric vehicles through emissions-responsive charging

“…The charger capacity is assumed to be 7 kW in this study, and the effect of varying the charger capacity from 3 kW to 22 kW is explored. Inverter and charge efficiencies are assumed as 97.0% and 98.5% respectively ( Sun et al., 2019 ). Battery self-discharge is disregarded in this study; we believe this is a reasonable simplification, as self-discharge rates for Li-ion batteries are generally low ( Zhang et al., 2020 ) and EVs are typically used shortly after charging.…”

“…The lifecycle analysis (LCA) approaches used in these studies indicate that the origin of the electricity used for charging is key, with carbon intensity being a critical factor ( Petrauskienė et al., 2020 ; Raugei and Winfield, 2019 ; Rosenfeld et al., 2019 ; Wu et al., 2019 ; Burchart-Korol et al., 2020 ). In many studies, however, the carbon intensity of electricity is evaluated in a relatively simple manner that does not take full account of its increasingly important variation with time and location ( Rees et al., 2014 ; Sun et al., 2019 ).…”

Controlling electricity storage to balance electricity costs and greenhouse gas emissions in buildings

“…Several researchers considered fitting a linear function for the hourly change in total CO 2 emissions (∆E) against the hourly change in total network demand (∆D), in order to obtain marginal emission intensity factors (MEF) as ∆E/∆D (McKenna et al 2017 ; Sun et al 2019 ). This method was originally proposed by Hawkes ( 2010 ) and has been modified by several researchers (McKenna et al 2017 ; Sun et al 2019 ; Siler-Evans et al 2012 ). Sun et al .…”

“…Sun et al . ( 2019 ) employed this emission intensity modelling technique in conjunction with an emission arbitrage algorithm for studying the potential of PV-battery systems for emissions reduction. They concluded that the battery could fully repay its CO 2 costs of manufacturing if it were controlled to minimise operational emissions.…”

Evaluating the regional potential for emissions reduction using energy storage

“…This provides both national AEFs and regional AEFs, with the regional breakdown achieved using a power flow model. Sun et al [27] developed an emissions arbitrage algorithm to improve the environmental performance of domestic PV-battery systems. Focusing on Great Britain, this projected future timevarying MEFs out to 2050 and examined the environmental benefits of arbitraging on carbon emissions using a simple threshold-based charge-discharge strategy, i.e.…”