Optimal and rule-based control strategies for energy flexibility in buildings with PV
Introduction
The penetration of photovoltaics (PV) is increasing rapidly worldwide, with a 26% increase of installed capacity to 177 GW from 2013 to 2014 [1], and the trend is expected to continue [2]. Residential, commercial or industrial buildings represent a major segment of the PV market, with a 66% share of total installed capacity in Europe in 2013 [3]. Unlike conventional electricity production, PV electricity production is variable and uncertain, therefore the increasing capacity brings an additional challenge to balancing the demand and supply of electricity [4], [5], [6], [7], [8], [9].
Power system flexibility measures to balance PV production can be classified into flexible generation, storage, interconnections, and demand-side management (DSM) [4], [9], [10], [11], [12], [13], [14]. In addition, turning surplus PV into thermal energy for heating and/or cooling, possibly including thermal energy storage (TES), may offer significant additional flexibility [9], [15], [16], [17], [18], [19], [20], [21], [22], [23]. Buildings with PV can provide on-site flexibility in all these classes except interconnections. The potential flexibility sources comprise flexible distributed generation (e.g. micro-CHP), electrical storage (e.g. batteries), shiftable and curtailable electrical loads, and thermal conversion in building heating and cooling, possibly coupled with thermal storage. The focus of this paper is on thermal conversion including TES, electrical storage and DSM with controllable loads. Dispatchable local distributed generation (e.g. micro-CHP) is not included here as it would incur extra investments in local energy production and the need for fuels.
Local DSM, thermal conversion and storage in a building with a PV installation can benefit the whole energy system by balancing the variable and uncertain PV feed-in. They can also benefit the prosumer (electricity consumer and producer) through increased self-consumption as well as price-responsive consumption, PV feed-in and arbitrage with electrical storage. Maximizing self-consumption is naturally incentivized in markets without a feed-in-tariff (FIT), such as Finland, or with FIT lower than electricity retail price, such as Germany [24], as the price paid for electricity fed to the grid is always lower than the retail price. Moreover, e.g. Germany and Italy have set separate self-consumption incentives [24]. Limitation of active power feed-in to 70% of maximum power output has also been imposed as grid integration regulation in some regions in Germany with high PV penetration [24], [25]. Self-consumption is a more energy-efficient way to adhere to this limit than curtailment.
This paper presents a model for studying cost-optimal and self-consumption maximizing rule-based control of energy flexibility in a building with PV. The included flexibility sources are a ground-source heat pump (GSHP) with an auxiliary electric resistance heater and a water tank TES, a battery, and shiftable loads, such as washing machines. The selected flexibility sources are interesting in light of GSHP proliferation for energy efficiency [26], low cost of water tank TES [27], [28], and the decreasing trend in Li-ion battery cost [29]. Shiftable appliances allow for lossless load shifting with no investment to storage capacity and only a limited effect to the activities in the building [9]. The potential of these flexibility sources to lower the electricity costs of the prosumer and to balance the variable and uncertain PV feed-in to the grid can be studied with the model.
The three control modes studied in this paper correspond to different market roles. In rule-based control, the building acts as a consumer which actively tries to avoid selling PV electricity by maximizing self-consumption and sells only when this is unsuccessful; selling is prohibited altogether in cost-optimal control with grid feed-in constrained to zero. In contrast, the building takes an active prosumer role in cost-optimal control, taking advantage of market price changes in both buying and selling.
There is a substantial body of literature on DSM and storage in PV integration to buildings. Many studies optimize DSM or storage control with PV and present results over a short time horizon, e.g. 1 day or 1 week [30], [31], [32], [33], [34], [35], [36]. In this paper, however, a full year is simulated with a 1-h time step. While studies with a short time horizon demonstrate optimal control in a realistic optimization horizon, the effect of seasonal and inter-day variations in PV production, price and load on the presented results is limited. Hence, the long-term performance in actual application is not evaluated. The seasonal effects are especially pronounced in northern locations with PV production concentrated to summertime and space heating demand to wintertime, such as Finland where the case study in this paper is situated. Inter-day PV production variations due to varying cloud coverage are significant worldwide. Moreover, a full-year simulation is less sensitive to the initial conditions of the system, such as initial storage state-of-charge. Hence, studies on rule-based and optimal control with time horizon of a full year or at least several months are reviewed here, starting with rule-based control studies.
Cao et al. [37] studied thermal storage with electrically heated DHW storage and a battery for PV and micro-wind production matching. Charging the DHW storage was found technically and economically more effective than a battery to reduce annual mismatch. In [38] and [39], a heat pump with thermal energy storage (TES) and a battery is used to increase self-consumption and lower peak grid injection from a PV system. Ref. [40] presents TES setpoint controls to match heat pump operation with PV production. Widén et al. [41] studied electrical storage, load shifting and PV array orientation for load matching. Kootstra et al. [42] studied a partially degraded second life battery with a PV system. More studies on rule-based control with PV can be found e.g. in a recent review [43]. Our paper adds to these studies by combining thermal storage, battery and shiftable appliances to a single model. Moreover, the performance of the rule-based control is compared directly to the cost-optimal one.
As for the optimal control studies, Wang et al. [44] presented robust optimization of household appliance and water heater scheduling with uncertain PV production. Guo et al. [45] solved stochastic optimization problems of scheduling a battery and shiftable loads. Widén and Munkhammar [46], [47] studied optimal scheduling of shiftable loads with rule-based control of battery storage for PV self-consumption. The potential without extensive battery storage was found limited. Masa-Bote et al. [48] also combined optimal scheduling of shiftable loads with rule-based battery control, reducing the uncertainty of energy exchange with the grid due to forecast error from 40% to 2%. Li and Danzer [49] and Schreiber and Hochloff [50] solved optimal control of a battery with PV, the latter with a capacity-dependent tariff to incentivize grid-benefiting operation. Ref. [51] presents model predictive control of a battery and TES with a heat pump, however without a detailed enough account on the employed models and optimization methods required for reproducibility or judging research limitations.
The flexibility sources studied in this paper, namely a GSHP with TES, a battery and shiftable loads, have not been combined in a simulation model with PV in a building for cost-optimal and rule-based control over a full year previously. Both variable and fixed condensing operating modes of a GSHP are included to such a model for the first time here as well. Moreover, contrary to previous work on cost-optimal control of a GSHP and TES with PV in a building over a full year, actual contract prices based on the electricity spot market price are employed in this paper instead of hypothetical prices, and the models and methods are documented to such an extent that the research limitations are clear and the study could be reproduced. Overall, physical realism has been strived for in the model, e.g. the modeling of the GSHP and TES is based on fundamental energy balances with explicit temperature modeling avoiding violations of the second law of thermodynamics, and includes the effect of temperature on the COP of the heat pump explicitly.
To demonstrate the new model and to verify the benefits from the controls with the different flexibility options, a case study on a Finnish low-energy house is also presented in this paper. Annual optimal control in sequential 24-h horizons is studied along with rule-based control. Different combinations of the flexibility sources and dimensions of the PV system, TES and battery are evaluated. The effects of heating system temperature level and variable vs. fixed condensing mode of the heat pump are also included. The 24-h horizon is a realistic optimization horizon in terms of day-ahead electricity price and weather forecast availability.
Section snippets
System model
The system model includes the flexibility sources, namely a ground-source heat pump with an auxiliary electric resistance heater and thermal storage, a Li-ion battery and shiftable appliances, along with DHW storage and an electric DHW heater. Fig. 1 depicts the energy flows and controls in the model. The heating and electrical system models are described in the following sections. Overall, physical realism has been strived for in the model. Some simplifications have been necessary to allow for
System control algorithms for cost minimization and PV self-consumption maximization
The control of the flexible energy system is the key for capitalizing on the potential benefits from the different flexibility options available. Two control approaches are studied in this paper: cost-optimal control and rule-based control for PV self-consumption maximization. Fig. 4 shows a schematic of the control principles.
The purpose of cost-optimal control is to minimize total electricity cost to the building, and it represents an advanced control case in which the building takes the role
Case study and input data
The simulations made to test the model are based on measured total and appliance-level electricity consumption and building properties of a real residential low-energy detached house located in Porvoo, Finland (60.4°N, 25.7°E) [63]. The simulations are run over a full year from January 2013 to January 2014.
Results
Flexibility of the residential energy system to integrate PV is studied with two PV system sizes in the size range of typical commercial systems [77]. 3 kWp is three times the hourly self-use limit and allows for considerable increase in PV production from the self-use limit with complete self-consumption with small-size storage. 9 kWp covers the whole annual appliance electricity consumption of the building.
Fixed condensing heat pump operation with a medium-temperature heating system is studied
Conclusions
A physically realistic model has been presented for cost-optimal and self-consumption maximizing rule-based control analysis of a flexible residential energy system with photovoltaics. The flexibility sources in the system are a ground-source heat pump with an auxiliary electric resistance heater and thermal energy storage, a battery and shiftable appliances.
The model is generic and applicable for an arbitrary condition. A case study of a Finnish low-energy house was conducted to demonstrate
Acknowledgements
Mr. Kristian Bäckström from Posintra Oy and Mr. Olli Jalonen from Aalto University, Department of Industrial Engineering and Management are gratefully acknowledged for providing data for the case study, and Dr. Imran Asghar, Mr. Juuso Lindgren and Mr. Sakari Lepikko from Aalto University, Department of Applied Physics for providing measurement data on Li-ion battery operation for verification of the battery model in this work. Computational resources provided by Aalto Science-IT project are
References (79)
- et al.
Review of energy system flexibility measures to enable high levels of variable renewable electricity
Renew Sustain Energy Rev
(2015) - et al.
Increasing thermal plant flexibility in a high renewables power system
Appl Energy
(2015) - et al.
The role of demand-side management in the grid integration of wind power
Appl Energy
(2010) Modeling storage and demand management in power distribution grids
Appl Energy
(2011)- et al.
Urban energy systems with smart multi-carrier energy networks and renewable energy generation
Renew Energy
(2012) - et al.
Smart energy system design for large clean power schemes in urban areas
J Clean Prod
(2015) Large-scale urban renewable electricity schemes – integration and interfacing aspects
Energy Convers Manag
(2012)Heat pumps and energy storage – the challenges of implementation
Appl Energy
(2012)- et al.
Potential for increased wind-generated electricity utilization using heat pumps in urban areas
Appl Energy
(2014) Towards an intermittency-friendly energy system: comparing electric boilers and heat pumps in distributed cogeneration
Appl Energy
(2012)