Before describing the proposed methodology for anomaly detection in
detail, let’s first look at one customer energy consumption provided in Figure
1. Energy consumption visulaiztino with line charts
The line Chart given in figure 1 shows the energy for one of the customer.
Line charts are the most common visualization technique of time series.
However, with the fine-grained consumption readings provided by smart meters,
this method is no longer effective to uncover the underlaying patterns in
paper, a unique visualization technique is employed. It is built upon the work in 7. The pixel based technique is used to visually encode numerical values into colors. It gives an overview of large time span and provides the possibility to recognize patterns or exceptions in the power
consumption time series, and to analyze the anomaly scores. Figure 2 shows the
same energy consumption visualized with pixel based technique. This technique
is showing that this customer is exhibiting high consumption period from 7 A.M
to 23 P.M, and low consumption period from 23 P.M to 7 A.M.
2. Energy consumption visulaiztino with pixel based technique
Energy consumption is influenced by a set of context variables such as occupancy and
appliance usage. To be able to analyze such contextual variables, extra sensing
infrastructure should be used. However, a human
activity typically follows regular temporal. For instance, appliance usage is highly correlated to
temporal context 12. Therefore, in this paper, occupancy and
appliance usage are indicated by temporal context sets as follows:
High consumption period: this set contains readings
during weekdays from (7 a.m. to 12 a.m.).
Low consumption period: this set contains reading
during weekdays from (12 a.m. to 7 a.m.).
Weekend: this set contains reading taken during the weekend, when residential buildings are likely
to have increased occupancy, and commercial premises are unoccupied.
temporal context sets can be also defined for each meter reading considering
the usage of the building (residential or commercial).
Geographic Area Effect
Based on the metadata obtained from smart
meters, the geographic area is defined as the set of buildings that are
expected to be influence similarly by the same context variables. Figure 3.
represents six customers from the same geographic area. For each customer, the
area marked in red represents a deviation from the normal pattern. However, in overall, this pattern represents a
trend among different customers, therefore, flagging it as an anomaly is not necessarily correct. For example, during holidays, when residential buildings
have increased occupancy will result
in higher energy
consumption. Another example is temperature variation for a specific area can influence energy consumption