*3.3. Experimental Validation of the Operating Modes*

#### 3.3.1. Foreword

This section of the paper aims to validate the two operation modes (i.e., inverter and PFC rectifier) of the proposed bidirectional DC–AC converter. To avoid damage to the prototype of the whole converter, the experimental tests were performed at low power. The single-pole variable transformer was used to adjust the grid voltage to 110 V RMS, 50 Hz, while the DC side was connected to the four Yuasa batteries for a total of 48 volts DC. The prototype also included a charge–discharge management system, capable of supplying up to 350 V DC to the whole converter.

The RMS voltage and frequency were monitored by the microcontroller on both sides of the primary switch connecting the converter to the AC grid (see Figure 2). The switch was closed when all criteria for grid connection were met [36], and the voltage on both sides of the converter was balanced using the duty cycle adjustable by the control circuit. Therefore, the bidirectional converter could operate in inverter or PFC rectifier mode, depending on the state of the AC grid. Thus, the HEMS provided a corresponding signal that forced the bidirectional converter to inject electricity into the grid (inverter mode) or to charge the batteries (PFC rectifier mode). Two buttons were added to the microcontroller to substitute the HEMS signal for the following results. The duty cycle could be adjusted, and the power direction and operating mode could be controlled with these two buttons.

Between the batteries and the DC–DC stage, a third stage was introduced for this experiment. The purpose of this stage was to modify the charge and the injection of electricity and to adapt the voltage and the current between the DC–DC stage and the batteries.

The following two subsections examine these two modes of operation in detail.

#### 3.3.2. Inverter Mode

The objective here was to confirm the proper operation of the bidirectional DC–AC converter in inverter mode. Figure 12 shows an example of the experimental results in this mode, where the batteries supplied 485 W to the whole system. In this example, the DC bus voltage was 178 V; this voltage represented the input DC voltage of the bidirectional converter. Figure 12 shows the output voltage of the DC–DC stage, confirming the sine half-wave, here at the RMS voltage of 126 V (measured using ❺ in Figure 10). The output of the DC–AC stage was a 50 Hz sine wave with, here, an RMS voltage of 120 V. This signal was applied to the AC grid and the power was effectively negative in this case (measured using ❸ in Figure 10) since power was being injected (in this example, 410 W). Figure 12 also shows the current injected into the AC grid, here with an RMS value of 2 A. As the current was imposed by the converter, it was necessarily out-of-phase with the voltage. While the inverter was injecting power into the AC grid, the voltage would remain constant even if the current changed.

**Figure 12.** Experimental validation of the inverter mode: output signals.

#### 3.3.3. PFC Mode

The objective here was to confirm the proper operation of the bidirectional DC–AC converter in PFC rectifier mode. First, it was essential to demonstrate that the power factor modification was indeed feasible in this case. The sinusoidal current could be absorbed by appropriately controlling transistor *T*2 of the DC–DC stage (see Figure 9).

Figure 13 shows an example of the experimental results in PFC mode. In this configuration, the AC grid delivered a positive power of 80 W to the whole system (measured using ❸ in Figure 10). The power transmitted to the batteries via the bidirectional converter was necessarily negative in this case, which indicates that we were in a charge cycle (indicated by ❹ in Figure 10). Figure 13 shows the sine wave imposed by the AC grid, as well as the half-sine curve between the DC–DC stage and the DC–AC stage, while we monitored the input of the DC–AC stage. Figure 13 shows an RMS voltage of the half-sine wave of 49 V, while the voltage at the DC bus was 166 V (measured using ❺ in Figure 10).

**Figure 13.** Experimental validation of the PFC mode: output signals.

In this example, the voltage and the absorbed current were imposed in order to follow the charge curve of the batteries.

#### **4. Discussion**

#### *4.1. Switching of the HEMS Control*

The main objective of this work was to design and implement a bidirectional DC–AC converter that was connected to an HEMS system [23]. The HEMS system supervised the control of several elements of the converter, such as:


At this stage, the bidirectional DC–AC converter was tested on a 110 V, 50 Hz AC grid. The next experiments will be performed on a 230 V, 50 Hz AC grid. The converter is fully self-adaptive and, therefore, requires no human interaction; all of the above parameters can be modified using real-time data measurements provided by the HEMS.

The experimentally designed and implemented bidirectional DC–AC converter is compatible with the HEMS system that we recently developed in [23]. Therefore, it can be fully used either to store excess energy available on the AC grid in batteries or to inject energy into the AC grid, according to the consumption patterns defined in the house.

#### *4.2. Efficiency of the Whole Converter*

This section of the manuscript aims to study in detail the efficiency of the proposed bidirectional DC–AC converter and, more specifically, when it operates in inverter mode. The PFC rectifier mode was regulated by the battery management system and thus, depended on the battery state of charge. The amount of energy stored could be varied but was comparable to the voltage and current measurements in the batteries, as well as the charge curves of the storage system. Combined with the charge control system, the overall efficiency of the system shown in the example in Figure 13 was approximately 85%.

In grid-connected inverter mode, we performed the efficiency measurements at low power (a few tens of watts) and up to 1.5 kW. The measurements were tricky to perform because we needed to ensure that the power injected by the batteries was flowing through the charge–discharge management system and that the power was detected at the output using the DC voltage and current measurements shown in Figure 12.

Figure 14 shows the evolution of the efficiency of the bidirectional DC–AC converter as a function of the rated output power. Figure 14 shows that the average efficiency of the whole converter was about 96.5%, from 100 W to 1.5 kW. It should be noted that it would be difficult to achieve any better given the general architecture chosen, i.e., the series connection of a DC–DC stage and a DC–AC stage, even if the DC–DC stage presented excellent performances (average efficiency of 98.5% over the tested power range). At very low power (e.g., 50 W), although the 95% target was not reached, the values obtained (here about 91.5% at 50 W) were more than satisfactory compared to other topologies discussed in the literature.

**Figure 14.** The efficiency of the bidirectional DC–AC converter in inverter mode (experimental results).

The losses were distributed mainly by the MOSFETs and the passive devices, especially the inductance of the DC–DC stage. The classical ferrite core was banned from this prototype because it caused too many losses in the iron. Therefore, we chose a high flux core that was ideal for high frequencies. To be more precise, the losses in the MOSFETs represented 1.1% of the target power, while the losses in the DC–DC stage inductor represented 2.4% of the target power. Conduction losses and switching losses could be estimated from the key characteristics (including drain-to-source on-state resistance, rise time and fall time) of each MOSFET. These losses represented 0.1% and 1% of the target power, respectively. By adjusting the reference of the SiC devices used in the DC–DC stage, for example with 36 A, 900 V SiC MOSFETs from the manufacturer Cree, it was possible to reduce the losses of the MOSFETs and gain nearly 0.5 percentage points in efficiency.

Compared to the first converter presented in [26] (this converter was experimentally tested in islanding mode), we intentionally decreased the switching frequency of the MOSFETs in the DC–DC stage from 300 kHz to 150 kHz. At 300 kHz, the converter efficiency lost, on average, 0.3 efficiency points over the output power range studied here. By switching to 150 kHz, we optimized the efficiency and, in addition, anticipated future electromagnetic compatibility issues, which we will detail in the near future.

#### **5. Conclusions**

In this paper, the main objective was to present and experimentally validate a bidirectional DC–AC converter, connected to the AC grid and suitable for HEMS applications when an energy storage system is required. The proposed topology was based on two necessarily bidirectional stages associated in series: the first being a DC–DC stage composed of two silicon carbide power MOSFETs, which were controlled at high frequency (the frequency of 150 kHz was implemented experimentally); the second being an H-bridge composed of four MOSFETs on silicon substrate, which were controlled at the frequency of the AC grid (i.e., 50 Hz here). With this converter architecture, the DC–DC stage regulated the DC voltage and established the positive parts of an AC waveform, while the DC–AC stage reversed it to obtain the sinusoidal voltage. Thus, the proposed structure provided an excellent AC waveform, but the latter depended mainly on the DC–DC stage.

A complete experimental procedure was defined and implemented to validate the operating modes of the bidirectional DC–AC converter, i.e., the inverter mode and the PFC rectifier mode, especially in the case of a grid connection. The energy efficiency of the whole DC–AC converter operating in inverter mode was evaluated and exceeded the target of 95% over the entire power range studied, i.e., from 100 W to 1.5 kW.

The three main outcomes of this study are summarized below:


The experimental results demonstrate the expected performance of such a system, both in terms of its operating modes and its high energy efficiency. The bidirectional switching of the components of this architecture is autonomous in adjusting the control commands of the microcontroller, as well as controlling the amount of energy to store or inject into the grid. The algorithm of the control strategy, the communication between the developed system and the smart meter are being patented.

The short-term perspectives of this work are as follows. The bidirectional DC–AC converter was developed to operate in both grid and off-grid mode. The islanding mode will be presented in another paper. In addition, the electromagnetic compatibility aspects will have to be realized in order to conform with the standards of connection to the AC grid that are in application. In the longer term, the whole converter and the HEMS will be installed in real smart homes and the profitability of our approach will be evaluated.

**Author Contributions:** Conceptualization, J.-C.L.B. and S.B.; methodology, J.-C.L.B., S.J., S.B. and I.A.; software, S.B., C.R., I.A. and T.B.; validation, I.A., T.B. and J.-C.L.B.; formal analysis, J.-C.L.B. and S.B.; investigation, S.B., C.R., I.A., T.B. and J.-C.L.B.; resources, J.-C.L.B.; data curation, , I.A., S.J. and J.-C.L.B.; writing—original draft preparation, S.B. and S.J.; writing—review and editing, I.A., S.J. and J.-C.L.B.; visualization, I.A., S.J. and J.-C.L.B.; supervision, J.-C.L.B.; project administration, S.J. and J.-C.L.B.; funding acquisition, S.J. and J.-C.L.B. All authors have read and agreed to the published version of the manuscript.

**Funding:** These research activities were supported by "Région Centre Val-de-Loire" (research project number: 2015-00099656). The authors of this paper thank our colleagues from this institution who provided insight and expertise that greatly assisted the project.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Conflicts of Interest:** The authors declare to respect the confidentiality clauses currently defined in the collaboration contract defined above.

#### **Abbreviations**

The following abbreviations are used in this paper:

