Future electricity demand and flexibility potential of private households

In the upcoming years, a disruptive revolution of the electricity demand in terms of load patterns and annual consumption is expected. Electricity from renewable energy sources can be used to decarbonise further sectors like the mobility sector. These new appliances, like electric vehicles, significantly increase and reshape the electricity demand of private households while efficiency gains slightly reduce the electricity consumption. Furthermore, the ongoing digitalisation allows an efficient controlling of the flexibility of these appliances. These developments make detailed investigations of the future electricity demand and flexibility potentials of private households inevitable. For this purpose, a Bottom-Up simulation method has been developed and is presented in this study. Using time use data of household members, activity profiles are derived based on a time-dependent Markov-chain model. Based on these activity profiles, different submodels for the simulation of electric appliances, lighting, heating and electric vehicles are used. The investigations in this contribution show the influence of new appliances like electric vehicles on the electricity demand of private households in terms of peak load and energy demand. Furthermore, the results allow a better estimation of the flexibility potential of appliances like washing machines based on the simultaneity of their usage.


Introduction
Historically, the electricity demand of private households has steadily evolved. In the upcoming years, a disruptive revolution of the electricity demand in terms of load patterns and annual consumption is expected. Aiming at a reduction of greenhouse-gas emissions in Europe, the usage of electricity generated from renewable energy sources (RES) becomes more important in the future. The electricity from RES can be used to decarbonise further sectors like the mobility sector. This trend leads to the integration of additional electrical appliances like electric vehicles. These new appliances significantly increase the electricity demand of private households while efficiency gains slightly reduce the electricity consumption. Furthermore, the ongoing digitalisation allows an efficient controlling of the flexibility of all appliances. Those developments make detailed investigations of the future electricity demand and flexibility potentials of private households inevitable. In order to investigate the future electricity demand and flexibility, an analysis of the electricity demand in private households is necessary and will be presented in Section 2. Following the analysis, Section 3 presents the methodology used to simulate the electricity demand and the flexibility of different appliances. Subsequently, the results of different investigations are presented in Section 4. The main outcomes of the results will be summed up in Section 5, which provides the most important conclusions from this paper.

Analysis
The electricity demand of private households accounts for one quarter of the total electricity demand in Germany [1]. This electricity demand can be traced back to the usage of various electrical appliances. Those appliances are conventional appliances, available in each household nowadays, as well as electrical appliances which might be more relevant in the future like heat pumps or electric vehicles. The electric appliances can be divided in terms of their usage. While some electric appliances like ovens or washing machines depend on the usage of the household members, other appliances like refrigerators are only influenced by the household members to a limited extent. A classification of different appliances is illustrated in Fig. 1.
The equipment of individual households with those electrical appliances foremost depends on the size of the household and is specified by the equipment level, which can be defined as the number of electrical appliances in 100 households [2]. Appliances like mobile phones or computers have equipment levels above 100, meaning that in each household more than one of those appliances exists, while the equipment level for electric vehicles is much lower today. Besides the household size, the equipment level relies on the financial status of the household as well as the region where the household is located [3].
As illustrated in Fig. 1, the operation of electrical appliances partly depends on the behaviour of the household members. The household members can be classified according to their age, gender and especially their employment status. For the different household members the user behaviour looks very different and relies upon the time of the day and the type of the day. Therefore, it is necessary to take into account the different types of household members in the following. The user behaviour is often described by the time of use for different activities as shown in Fig. 2.
Besides the electrical appliances and the user behaviour, the external factors need to be taken into account as well to represent the electricity demand in private households.
The different types of electrical appliances differ not only in their usage profile but furthermore in terms of their flexibility provision. With respect to the flexibility provision, the electrical appliances can be further categorised in terms of their reduction of comfort losses. Changes in the operation of user-independent appliances or appliances that mostly depend on external factors like the heating system, which is regulated via a control system, do not affect the usage or comfort of the users. At the same time user-dependent appliances where the electricity demand directly correlates to the time of use, cannot provide any flexibility without reducing the comfort of the user. User-dependent appliances that follow a load profile can partly be shifted within average deferral periods without reducing the comfort of the user. Examples for those appliances are washing machines and dishwashers, which can regularly be shifted up to 6 h without reducing the users' comfort. Furthermore, the electricity demand of appliances with electric storages can be shifted in any way, as long as the comfort of the user is not limited.

Methodology
Within this paper, a method to simulate the electricity demand and the flexibility potentials of private household is presented. The developed three-stage method is able to represent conventional and new electric appliances in detail. The first stage is the simulation of the regional distribution of households and their equipment with electrical appliances and household members. While the first stage is not presented in detail in this paper, the second and the third stages are described in detail. The second stage is the electricity demand simulation (Section 3.1), which is followed by the flexibility simulation as the third stage (Section 3.2).

Electricity demand simulation
As illustrated in Fig. 3, the simulation of the electricity demand can be divided into several submodules. To ensure a consistent demand simulation, the whole simulation is based on a simulation of the user behaviour. Based on the user behaviour, different submodules simulate the electricity demand of various electric appliances. Besides conventional electric appliances that are simulated in the first submodule, another submodule derives the electricity demand for lighting, taking into account the activity profiles as well as irradiation. New electric appliances like electric vehicles and heat pumps have separate submodules taking into account their specifics.
The activity simulation is a Markov-chain model for the simulation of individual household members. As illustrated in Fig. 4, the Markov-chain model calculates the next activity based on transition probabilities from one activity to the other. The first activity therefore needs to be calculated using a starting probability. The model in this paper uses time-dependent transition probabilities, meaning that depending on the time at which a given activity was followed, the transition probabilities change. In this approach currently 22 different activities and their transition probabilities are taken into account. The transition probabilities are calculated based on a survey of the time use in Germany, conducted by the statistical office [4].
Following the simulation of the activities the electrical appliances use in private households is modelled. Besides usage-dependent appliances, which correlate to the activity profiles, appliances, which are mostly uncorrelated to the users' activities, are simulated with a continuous or periodic electricity demand. The usage-dependent appliances are assigned to specific activities via starting probabilities. The appliances, which electricity demand relies upon the time of use, are modelled as active as long as the activity consists. The appliances with a fixed load profile are modelled as starting in the moment the activity begins and are active as long as indicated by the load profile.
The model for electric vehicles is illustrated in Fig. 5. It uses the activity profiles for determining the time steps where household members leave the house and correspondingly come back to the house. In those time steps the availability of the vehicle is evaluated and the probability for using the electric vehicle is taken into account. Following this, a stochastic model estimates

Flexibility simulation
Following the electricity demand simulation, the simulation of the flexibility potential takes place in the third step. The flexibility potential of electrical appliances including electric vehicles is estimated separately for appliances with inherent storages and appliances without storages.
The model is based on a synthetic price per interval, that can either represent the electricity price users need to pay, or any incentive for shifting their electricity demand, for example to release grid constraints.
Electric appliances without storages are simulated as appliances with a fixed load profile that can be shifted in a given deferral period. For each possible starting point of the load profile within the deferral period, the costs of electricity consumption for the whole load profile are calculated. The new starting point is then set to the interval where the lowest electricity costs for the whole load profile occurred.
Electric appliances with storages are modelled in a similar way. Within the time periods taken into account in this model no ramp up, or ramp down constraints need to be considered. Furthermore, it is assumed, that before and after the flexibility provision, the same state of charge is necessary. Therefore, in a first step, the electricity consumption price for each time step in this flexibility period is calculated. For electric vehicles this flexibility period can be estimated by the time from arriving at home from a ride and the next step, where the house is left. In the next step, the algorithm uses the cheapest intervals as long as necessary to reach the foreseen state of charge. This approach allows a coherent and comprehensive simulation of various appliances in private household.

Results
The investigations conducted in this research paper focussed on 100 private households that represent a typical excerpt of all German households. For those households, the typical electricity demand patterns have been analysed and the influence of new appliances like electric vehicles will be presented. Afterwards the flexibility of typical appliances like washing machines and dishwashers are presented. This investigation shows the flexibility potentials at different points in time, by evaluating the share of appliances that is available for flexibility provision at each time interval.

Influence of new electric appliances
The influence of new electric appliances on the electricity demand was analysed by simulating three different scenarios with a varying number of electric vehicles, according to the German network development plan [5].
The investigations illustrate (Fig. 6) that an increasing number of electric vehicles significantly rises the peak demand of electricity especially in the evening hours. The comparison of the base case and the scenario with the highest number of electric vehicles shows that the peak load increases about 300%. At the same time, the energy demand is increasing heavily in the simulated setting.

Flexibility potentials
The flexibility potential is illustrated using the example of washing machines and dishwashers. The share of appliances available for flexibility provision is an indicator for the flexibility potential in the electricity system. Fig. 7 presents this share and shows that the share of electric appliances available for flexibility provision does not exceed 65% at any time. In general, the share of washing machines available for flexibility is significantly lower than the available share of dishwashers. This can be traced back to the different usage behaviours of the household members and the usage profiles of the appliances.   Following the analysis and the description of the method, the investigations show the effect of new electric appliances on electricity demand and flexibility potential of private households.
The first results showed a significant influence of new appliances on the electricity demand and the existence of major flexibility potential. At the same time, the contemporaneity in the usage of appliances reduces flexibility potentials at certain times. These findings demonstrate that a detailed analysis of electricity demand and flexibility potentials of private households are necessary for further investigations.