**5. Results**

This section shows the tests carried out with the SMH. For this purpose, data were collected from several households located in Jaén, Andalusia, Spain. The houses have different characteristics, such as EV electric vehicles, PV solar photovoltaic generation, etc. In each of them, the load profiles were monitored for comparison. The tests shown operation for two days in the year 2022. Figure 8 shows the device connected in the house.

**Figure 8.** The device assembled with real components connected in the house.

The graphs show the average value of the electrical variable displayed (*v*, *i*, *p*, *s*, *PF*, and *q*). The aggregation is performed on the data taken by the meter every 0.5 s. The system automatically aggregates the data according to the time slice chosen for the display. This helps to understand the temporal evolution of the measurement in the shown time slice. As for the graphs on the left, the data are aggregated in an automated way according to the chosen time slice. The data representation corresponds to 6 and 7 October 2022, for a single-family house with four inhabitants.

**Figure 9.** General dashboard (**a**) *q* and (**b**) *v*.

Figure 10 shows the second dashboard, specifically the general dashboard, where we can see a timeline with the values of apparent power *p* and current *i*. In Figure 10 it appears that *i* and *p* evolve with graphs that have similar shapes. This is because *v* varies in very narrow margins and remains almost constant within the time slice shown. It can be observed in Figure 9 that the voltage is very stable; this is due to the consistency provided by the electrical network. As a consequence, it has little influence on the variability of power. The user can observe that the maximum consumption peaks occur between 08 and 10, 14 and 16, and 20 and 22 h which correspond to the hours of maximum occupation in the household. These hours coincide with breakfast, lunch, dinner, and rest time for viewing TV. The behaviour is similar in the two days shown.

**Figure 10.** General dashboard (**a**) *p* and (**b**) *i*.

Figure 11 shows the power factor *PF* and apparent power *s* values. In all cases, the two graphs on the right show the average values that are being reached in the selected and configurable time period. The apparent power *s* is influenced by the variability of the electrical current consumed in the household due to the stability of the voltage as shows Figure 11.

**Figure 11.** General dashboard (**a**) *PF* and (**b**) s.

Figure 11 shows the behaviour of the *PF* on the two days shown. The *PF* is quite high and has values close to the unit during the study period, which implies a high performance of the receivers connected in the house and low losses. This is due to the use of class A+++ electrical appliances and an aerothermal air conditioning system.

In addition to the graphs of the electrical variables, the values are used for data visualization and can be downloaded in CSV format. This feature is highly interesting for homeowners or researchers because it enables studying the energy behaviour of the household, photovoltaic (PV), and vehicle (EV). The CSV format was chosen because it is accepted by most software used for analysis, such as MATLAB, Excel, and most statistical suites.

By default, the dashboard loads the data from all the devices and shows these charts for the average and accumulated values of the devices. It was also designed so that filtering a component affects the rest of the displayed data. This allows filtering all the components at the same time, including the data table. This data table allows sorting and filtering each of the columns shown as well as downloading the data being displayed in CSV format.

Since everything is configurable, the platform allows different types of filtering for almost every one of the data fields. The most frequent filtering types are the following:

Filtering by date range: This filter can be applied in the platform in different ways. The most common is to use the date control that appears in the upper right corner, see Figure 12. Another option to filter by dates is to select, with the computer mouse, a specific area in the timeline charts. Automatically this range is selected and all the information and charts are updated.

**Figure 12.** Filtering by date.

Filtering by device: We can use a filter control that the dashboard has in the upper left corner to filter by one or several devices. All the data and graphs of the dashboard are automatically updated. This filter control can be seen in the Figure 13.

The following designed and developed dashboard shows timeline-type graphs that visually compare all the devices loaded in the data with all the established measurements. From a visual point of view, it is interesting to see a great amount of information in very little space.

Figures 14 and 15 compare three households. The evolution of the three houses can be observed. One was analysed in previous figures and the other is a single-family household with two and three inhabitants in house#11 and house#1, respectively. As opposed to the house#13, in the second one the inhabitants are out of the house all day because of work, returning home after 9.30 pm when a higher consumption is observed. House#11 and

house#1 are occupied almost all day with consumption peaks in the middle of the day, in the afternoon, and at dinner time. The use of these comparative graphs allows researchers to obtain analyses from which they can deduce conducts applicable to obtaining load profiles, demand estimates, energy study of the household, reduction in the electricity tariff, etc.


**Figure 13.** Filtering by device.

**Figure 14.** General dashboard comparison of (**a**) q and (**b**) v values.

**Figure 15.** General dashboard comparison of (**a**) *p* and (**b**) *i* values.

Figure 16 shows the PF for the households, with EV and PV. House#13 and house#1 vary with respect to the PV, the work performed by the inverter for the grid connection, which maintains the PF at 1 for almost the entire generation time, is clearly noticeable. The EV, which is connected to the grid through a DC/AC converter, also maintains the PF at values above 0.95 during most of the charging process, making it very efficient. The different equipment installed in the households, especially the electrical appliances, make the first household present a much higher value of PF, above 0.9, making it very efficient. On the contrary, house#11 has equipment that reduces the average value of the PF to 0.65, thus presenting a less efficient behaviour.

**Figure 16.** General dashboard comparison of (**a**) *PF* and (**b**) *s* values.
