Sensor Simulation System Article No.2

As you might already know, the DEHEMS system aims to collect large amounts of real-time sensor data from individual households. Based on the current project status, DEHEMS is estimated to receive data from 12 different sensors per household, each of these sensors having a sample rate of 10 readings per minute. These 12 household sensors are to be divided as follows:

• one general household electric meter which provides information regarding the spot general electric power consumption of the entire household;
• one temperature sensor which reports the household temperature (better quality data can be provided by having more sensors around the house and local computing an average of their reported readings);
• one gas meter which reports the spot gas consumption of the household;
• up to nine individual electric sensors which report the spot power consumption of various individual appliances in the household.

The combined readings of these 12 sensors make up the spot energy consumption profile of the household, which is exactly what we need to simulate.

In the current stage of development, our DEHEMS Sensor Simulation System (DS3) contains two distinct stages of generating spot energy consumption profiles: gross direct simulation and detailed pattern based simulation.

In this first stage, gross direct simulation, basic sensor readings are simulated by estimating values based either on real-life measurements or data from studies in the filed of household energy consumption. We have gathered data regarding active and standby power consumption for each appliance. For instance, a color TV consumes 100 W in active mode and 20 W in standby mode and my laptop consumes 30 W when active and 5 W when in standby.

Accurately simulating complex sensor readings like the general household electric meter or the complex temperature sensor is a bit more tricky. For this, we have included in DS3 the pattern based simulation which, as the name suggests, is based on energy consumption patterns.
In their simplest form, these patterns show the consumption state (active, standby, or offline) of every appliance in the house for a predefined period of time. By combining these data with information regarding the power consumption of individual appliances we can obtain through aggregation an accurate picture of the general household power consumption at a given time.
The figure to the left shows how we simulate in DS3 the energy consumption profile: Ia1,Ia2, ..Ian are the individual appliance sensor readings (gross direct simulation), T1, T2, Tm are the temperature sensor readings, E is the reading from the general household electric meter (pattern based simulation), T is the average household temperature reading (pattern based simulation), G is the gas meter reading (gross direct simulation) and P is the obtained spot energy consumption profile.

Our next post will present details about the current DS3 implementation and some of the results we have obtained so far.

Research Institute e-Austria homepage: www.ieat.ro/