Appliance usage and lighting

Appliance usage and lighting modules typically consider household occupancy/activity profiles and technical characteristics and statistical data for the different appliances to simulate daily electric load profile with a 1 min time resolution.

For appliance usage and load simulation, the following modules are available:

while for lighting demand:

Appliance usage

The type of appliance influences the method used to estimate the usage and thus the load profile. There are three different ways of categorizing appliances.

Based on their load profile:

  • Constant load: an average load value is assigned to the appliance when it is used (e.g., )

  • Time varying load: appliances that have their power consumption varying during an operating cycle. For example, the cycle of a washing machine consists of several stages of heating, washing, draining, spinning, rinsing, which have different power demands.

Based on their usage duration:

  • Fixed duration: the appliance is always used for the same number of timesteps (e.g., kettle and dish washer)

  • Stochastic duration: the duration is drawn from a random distribution. Example: the TV duration in CREST is drawn from an exponential distribution, and the hot water durations from a set of empirical discrete distributions.

Based on usage patterns:

  • Level usage: appliances that switch-on and off independently from their usage, such as fridge and freezer.

  • Activity dependent: appliances directly related to an activity. Therefore, their usage occurs when at least one resident is undertaking the corresponding activity. For example the TV or the oven.

Note

Please keep in mind that these categories depend both on the technical characteristics of the device and the modeling assumptions and simplifications. For example, the dishwasher can be considered either constant load or time varing load depending on the accuracy of the profile used.

Occupancy-based appliance usage simulator

API

For details about the implementation of this simulator you can visit OccupancyApplianceSimulator.

Description

The appliance usage and load modules draws inspiration from CREST model, developed in excel VBA and presented in [Richardson2010] and [McKenna2016].

Initialization. At the beginning of a run, the model populates each dwelling with a set of appliances. This process can be based on generic or sociodemographic-specific statistical ownership data.

Appliance usage and load. Appliance usage and load profiles are estimated as follows:

  1. first, the activity probability density function (pdf) is multiplied by a calibration scalar, whose value is assigned such that the simulated annual consumption of a device matches a target value;

  2. second, the turn-on event occurs if the probability exceeds a random draw;

  3. when a turn-on event occurs, the duration of the event is also estimated;

  4. finally, the turn-off event occurs at the end of the scheduled duration or when the active occupancy becomes zero if the device depends on the activity.

Note

  • For level usage appliances, whicht do not depend on active occupancy (e.g., fridge and freezer) step 1 assumes activity pdf to be equal to one.

  • For appliances with a fixed duration step 3 is ignored.

Availability

This simulator is available for UK and German households. However, few modifications are implemented according to data availability.

Initialization. For UK case, the model stochastically estimates the appliance set for each household using statistical ownership data from multiple sources [Richardson2010].

For the simulation of the German case, demod introduces the following changes compared to CREST:

  • Households are initialized with a set of appliances that is dependent on socio-demographic data, using the dataset from [Destatis2017].

  • The full set of available appliances is updated to reflect obsolescence (e.g., answer machine, cassette / CD player) and changes (e.g., tables, game console) in technology.

Appliance usage and load. Here is the list of activities that are relevant to the use of specific appliances: watching TV; cooking; laundry; washing(self) / dressing; ironing; housecleaning; electronics; dish washing. The two latter activities in italics are only available for the German case:

  • The dish washing activity is present in the German-time-use and corresponds to the use of the dishwasher or the sink.

  • Electronics has been added and accounts for the use of computers, laptops, tablets, printers and gaming consoles.

Compatibility

These modules are compatible with occupancy simulators: 4-states occupancy simulator and Transit occupancy simulator.

Lighting

In demod, two modules are available for lighting simulation. Both depend on two parameters:

  • The number of active occupants, which counts how many people are likely to need light.

  • External radiation or natural lighting, as human perception of the natural light level within a building is a key factor determining use of artificial lighting.

Fisher’s lighting simulator

API

For details about the implementation of this simulator you can visit FisherLightingSimulator.

Description

This model is presented in [Fisher2015], an it computes lighting power demand at time t as:

P_{el,l}(t)=n_{active}(t) \cdot P_{el,l,pp} \cdot
\frac{I_{g,max}-I_g(t)}{I_{g,max}-I_{g,min}}

where n_{active}(t) is the number of active occupants at a given time, P_{el,l,pp} is a constant for accounting for light usage per person and I_{g}(t); I_{g,max}; I_{g,min} are respectively the current irradiation, and the boundaries between which the light usage rate increases linearly between 0 and 1 as the external irradiation decreases.

Availability

This simulator is available for UK and German households.

Compatibility

These modules are compatible with occupancy simulators: 4-states occupancy simulator and Transit occupancy simulator.

CREST lighting simulator

API

For details about the implementation of this simulator you can visit CrestLightingSimulator.

Description

This approach is based on the work by Richardson et al. [Richardson2009]. It computes light switch on/off events, considering irradiation and effective occupancy, which takes into account occupants’ sharing of lights within the same room. It also takes into account any lights that are left on during the day and the diversity of households.

Availability

This simulator is available for UK and German households.

In order to better fit this module to the German case, two main modifications are made to the approach of Richardson et al.:

  • The number of lights in a household is initialized following the approach of [Frondel2019]. Here, the number of installed bulbs is generated using a normal distribution N(25.11,15.92).

  • Moreover, the type of light bulbs is defined through a discrete distribution: LED 65%, CFL 25%, and incandescent 10%.

Compatibility

These modules are compatible with occupancy simulators: 4-states occupancy simulator and Transit occupancy simulator.