*5.2. Simulated Subset*

The circuits with (a) resistor; (b) resistor and inductor: (c) diode rectifier with resistor; (d) diode full-wave bridge rectifier with resistor and capacitor; (e) thyristor rectifier with resistor; and (f) thyristor rectifier with resistor and inductor were evaluated with a test bench. The load current and mains voltage were acquired using voltage and current probes and an oscilloscope, with the simulated loads configured as presented in Table 11.

**Table 11.** Parameters of real components in the test bench.


In addition to the voltage and current measurements using the test bench, the amplitude and phase of each harmonic of the waveform of the voltage of the power network was measured. These values, presented in Table 12, were included in the voltage source block in the simulator and used in all simulations that included harmonic contents (configurations DB-3 to DB-6 in Table 9). The amplitude is presented with respect to the fundamental component: 60 Hz and peak voltage of 179 V).

The parameters presented in Tables 11 and 12 were used in the simulation framework developed in Matlab/Simulink. Then, the measured and the simulated waveforms were compared, as exemplified in Figure 18.

One way to validate the simulation is by comparing the waveform's electrical parameters, such as transient and steady-state current and voltage peaks, power factor (PF), and mean squared error (MSE) of the samples of the measured and simulated waveforms, as suggested in [46]. Therefore, Tables 13 and 14 present such comparisons for the simulations proposed in this work. The results presented in these tables validate the presented simulation approach for circuits (a) to (f).


**Table 12.** Measurement of voltage and phase of harmonics in the power network.

**Figure 18.** Measured and simulated current of the full-wave bridge rectifier load. Adapted from [10].




**Table 14.** Validation parameters (power factor (PF) and mean squared error (MSE)).

Finally, the validation of the last circuit (g), the universal motor, was performed using an electric drill of 750 *Wpeak* as a reference, with two-speed selection. The procedure was conducted in two stages. Firstly, the current and voltage signals of the electric drill were acquired in different scenarios, i.e., switching angle and load conditions. Then, the acquired signals were used to obtain the field and rotor resistances and inductances of the model presented in Figure 6. The final obtained values used in this model were:


Secondly, with these parameters, the comparison of the real and simulated current of a universal motor is presented in Figure 19. As can be observed, the waveforms present similar values in transient and steady-state. The MSE between these waveforms is 0.03 A2.

**Figure 19.** Comparison of the measured and simulated waveforms of a drill. Adapted from [10].

Concerning the generation of the waveforms that compose the Simulated subset, Figure 20 presents an example of the current waveform generated in the proposed subset (DB-5), in a double load scenario where the first load is a resistor and inductor circuit (section A, Figure 21a and section C, Figure 21c) and the second load is the universal motor (section B, Figure 21b and section D, Figure 21d).

**Figure 20.** Simulated subset complete acquisition: RL and universal motor. AC mains voltage and current.

universal motor turned ON.

 OFF.

 (**d**) Section D: universal motor turned

**Figure 21.** Sections (**<sup>a</sup>**–**d**): RL and universal motor.

The sampling frequency for the Simulated subset is also 15,360 Hz and switching instants (ON or OFF) are precisely controlled at the sample level. For each switching-event, the load is also properly labeled, allowing the correct use of supervised classifiers and transient feature extraction methods.

The simulator's functionality allows for the generation of single-load waveforms as well as the combination of two, three, four, five, six, and seven loads. Such combinations can be accomplished using a MATLAB script that automates the waveform generation, using pre-defined trigger instants and types of loads that are selected in each simulation (The MATLAB-Simulink template to generate this dataset is made publicly available at https://github.com/hellenancelmo/Simulated-LIT-dataset). Hence, this subset can be extended to other types of residential, commercial, and low-voltage industrial loads.
