Control algorithm for a residential photovoltaic system with storage Yannick Riesen a,1,∗ , Christophe Ballif a , Nicolas Wyrsch a a Ecole Polytechnique F´ ed´ erale de Lausanne (EPFL), Institute of Microengineering (IMT), Photovoltaics and Thin-Film Electronics Laboratory, Rue de la Maladi` ere 71, CH-2000 Neuchˆ atel, Switzerland Abstract High penetration of photovoltaic (PV) electricity could affect the stability of the low-voltage grid due to over-voltage and trans- former overloading at times of peak production. Residential battery storage can smooth out those peaks and hence contribute to grid stability. A feed-in limit allows for the easy setting of a maximum power injection cap and motivates PV owners to increase their self-consumption. A simple control strategy for a residential battery system coupled with a PV system that maximizes self- consumption and minimizes curtailment losses due to a feed-in limit is presented. The algorithm used in this strategy does not require a forecast of insulation conditions. The performance of this algorithm is compared to a second algorithm—a control strat- egy based on linear optimization using a forecast. Assuming an exact forecast, this second algorithm is very close to the maximum self-consumption and minimum curtailment losses achievable and can be used to benchmark the simple strategy. The results show that the simple strategy performs as well as the second algorithm with exact forecasts and performs significantly better than the second algorithm using real forecasts. Moreover, it is shown that this result is valid for a large range of storage capacities and PV sizes. Furthermore, it is shown that with a time resolution of 15 minutes for the input data (the resolution of most PV production and load data) self-consumption is overestimated by about 3 % and curtailment losses are underestimated by the same amount. Load sensitivity simulations show that different load curve shapes do not fundamentally change the results. Finally, to assess the effect of load aggregation, the case where the strategy is applied separately to 44 households with storage is compared to the case where it is applied to a centralized storage system of the same size as the total storage of the 44 households. The reduction of the curtailment losses with the number of aggregated houses is showed. Keywords: Photovoltaic (PV), Home storage system, Battery management strategy, PV integration into the electrical grid Highlights • A control strategy for a battery system coupled with a PV system is presented. • With a feed-in limit this strategy does not need a PV pro- duction forecast. • This strategy performs as well as a strategy relying on an exact forecast. • A relatively small storage size allows peak injection reduc- tion of 50 %. 1. Introduction Residential electric energy storage systems coupled with a photovoltaic (PV) installation could contribute to the stabil- ity of the low-voltage grid in the case of high PV penetra- tion by absorbing the production power peaks around midday ∗ Corresponding author Email addresses: yannick.riesen@alumni.epfl.ch (Yannick Riesen), christophe.ballif@epfl.ch (Christophe Ballif), nicolas.wyrsch@epfl.ch (Nicolas Wyrsch) 1 Present address: Planair SA, Crˆ et 108a, CH-2314 La Sagne, Switzerland [1, 2, 3, 4, 5]. Moreover, such a system increases PV self- consumption, which can provide an economic benefit to the system owner due to lower electricity exchange with the grid and minimized electricity transport losses [6, 7, 8]. Economic assessment of such systems optimizing only self-consumption can be found in [9, 10, 11]. In Germany, financial incentives for battery storage are available provided that the feed-in power is limited to 50 % of the PV system’s nominal power [12]. By 2015, more than 12’000 such systems were installed in Ger- many [13]. As shown by [14], active power curtailment allows for stabilizing the grid voltage. Imposing a feed-in limit is a simple and efficient method to avoid high injection peaks and to hence minimize grid disturbances allowing for higher PV penetration [15, 16]. However, this limit induces curtailment losses even in the presence of energy storage systems. There- fore, control strategies that minimize those losses and maxi- mize self-consumption are needed. Alternative strategies than a fixed feed-in limit prevent injection peaks are described in [17, 18, 14]. Several control strategies that allow for the efficient shaving of injection peaks and the maximization of self-consumption at the same time have been proposed in the literature. Solutions based on exact (or perfect) forecasts are presented in [19, 20]. However, forecast inaccuracies induce non-negligible changes in the performance of the systems [21, 22, 23]. To circumvent Preprint submitted to Elsevier Journal May 1, 2017 © 2017. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/ of 12.