Developing an Integrated Soft-Switching Bidirectional DC/DC Converter for Solar-Powered LED Street Lighting

Abstract

In the current era marked by the growing adoption of renewable energy sources, the use of photovoltaic-powered LED streetlights, known for their enhanced efficiency and extended lifespan, is on the rise. This lighting solution encompasses essential components such as a photovoltaic (PV) panel, an energy storage system, LED luminaires, and a controller responsible for supervising power distribution and system operations. This research introduces a novel approach involving a ZVS (zero-voltage switching) bidirectional boost converter to manage the interaction among the PV panel, LED lights, and battery storage within the system. To elevate system efficiency, a modified version of the conventional bidirectional boost converter is employed, incorporating an auxiliary circuit encompassing a capacitor, inductor, and switch. This configuration enables soft switching in both operational modes. During daytime, the converter operates in the buck mode, accumulating solar energy in the battery. Subsequently, at night, the battery discharges energy to power the LED lights through the converter’s boost operation. In this study, the PET (photo-electro-thermal) theory is harnessed, coupled with insights into heatsink characteristics and the application of a soft-switching bidirectional boost converter. This integrated approach ensures optimal driving of the LED lights at their ideal operating voltage, resulting in the generation of optimal luminous flux. The proposed LED lighting system is thoroughly examined, and theoretical outcomes are validated through simulations using the PSCAD/EMTDC version 4.2.1 software platform.

1. Introduction

With the advancement of technology, the increase in population, and the growth in electrical devices, the amount of energy demanded is increasing. If this problem continues along with this trend, in the near future we will face a fuel problem for electricity generation. In order to deal with this challenge, the use of renewable energy as a viable solution has been suggested by many researchers. Solar energy, which is one type of renewable energy, is considered one of the most popular types of renewable energy due to its easy access. Power electronic converters are one of the main components of these renewable energy systems. Considering that the output power of solar systems is DC, DC/DC converters are extremely important in the structure of photovoltaic systems [1,2,3,4,5,6,7].

Bidirectional DC/DC converters are used in a variety of applications such as fuel cell vehicles, battery chargers/dischargers, and uninterruptible power supply (UPS) systems. These converters play a very important role in renewable systems such as fuel cell systems, PV systems, and wind power systems [8,9,10,11,12,13]. There are two classifications of bidirectional DC/DC converters: isolated and non-isolated types [14,15,16]. The isolated bidirectional converter, with more than four switches and an isolated transformer, exhibits higher conduction losses and lower efficiency compared with its non-isolated counterpart. Conversely, the non-isolated bidirectional converter offers high efficiency due to its simpler configuration [17,18,19]. To enhance efficiency, one suggested approach is the utilization of the soft-switching technique, where the power switches operate with zero-voltage switching, reducing switching losses during transitions and improving overall efficiency [20]. Also, several research studies have been conducted in recent years about various control methods for DC/DC converters [21,22].

As mentioned, photovoltaic systems that utilize solar energy to generate power can be used in various applications. This energy can be used in large and small scales such as in solar power plants and in street lighting [23,24]. Streetlights can be powered independently by PV systems, even without a network connection. The energy received from the sun during the day and solar panels charges the battery, and this energy is used at night to turn on the lights. Statistics show that PV-powered LED (light-emitting diode) streetlights are widely employed in three main types of applications—commercial, residential, and industrial—all over the world. Among these, commercial applications are the most commonly utilized [25,26,27].

Due to their advantages such as high luminous efficiency, better vision, and long lifetime, high-intensity-discharge (HID) lamps are used for street lighting [28,29,30]. However, these lamps have disadvantages, such as excessive glare, containing harmful substances (such as metallic mercury), and needing some seconds to get to full brightness. High-brightness LEDs can be a good replacement for these lamps due to the lack of these problems. LED technology has progressed rapidly over the last two decades. LEDs have high reliability, the absence of ultraviolet and infrared rays, a long lifetime, and high luminous efficiency, and are energy-saving, environmentally friendly, etc. Figure 1 shows the structure of a streetlight LED system powered by a PV panel.

Two parameters of LED voltage and current are very effective in system design. Indeed, efforts should be made to drive these systems in almost constant current and voltage because any increase in these parameters leads to an increase in the junction temperature and a reduction in the optical flux and efficiency [31,32]. LEDs have optical, electrical, and thermal characteristics that are dependent on each other. In [33], the relationship between the luminous output (photometric variables) and thermal behavior was reported. Considering these three mentioned characteristics, a suitable LED system can be designed. The LEDs used in various systems currently have extremely low efficiency. These LEDs turn about 87–90% of their consumed energy into heat by burning at their nonoptimal points [34]. To reach the maximum optical efficiency and to maximize the luminous flux from the LEDs, they must work at the optimal point of their voltage and current. Two LED streetlight systems are proposed in [35,36]. These systems are powered by solar energy as the primary source and batteries as the secondary source. In [37], using a two-way buck–boost converter and a microcontroller, the charging and discharging of the lighting system is controlled. This is also performed in [38], using a two-way Zeta–Sepic converter. Due to improper design, this structure cannot operate at its optimal point. Also, due to this issue, this structure cannot use the soft-switching technique. In [39], the authors introduced a model for LED devices which incorporates thermal resistance and light output as variables. However, it should be noted that this model specifically focuses on the LED device itself and does not cover the entire LED system, which includes aspects such as electric power control and thermal design of the heatsink. In order to relate the optical, electrical, and thermal aspects of LEDs, a unified model is presented in [40]. This method shows how to achieve a current for the optimal flux. In [41], a two-stage LED driver is proposed, which has a boost circuit and a new resonant converter. The resonant circuit of this converter, which is of the CLCL type, is a new circuit that has soft-switching conditions. Also, in [42], an LED driver is presented that has a CLL resonant converter. In [43], a single-stage LED driver is proposed. This driver is based on a boundary conduction mode (BCM) boost circuit and an LLC resonant converter.

In this paper, a single-stage bidirectional buck–boost converter is presented to control battery storage, LED light, and the PV panel. In this converter, to improve efficiency, an existing soft-switching structure of the converter is used. The employed converter is derived by incorporating an LC series resonant tank into the conventional bidirectional DC/DC converter. The soft-switching cell structure consists of passive elements and does not require any other switching device. The applied topology performs soft switching in both buck and boost operational directions. Throughout the daytime, the energy generated by the PV panel is stored in the battery via the buck operation direction of the converter. In addition, the battery supplies the LED light with the necessary power through night via the reverse boost operation direction of the converter. The voltage regulation in this case can be detected for input and output. This performance provides MPPT (maximum power point tracking) to the PV panel, SOC (state of charge) for the battery, and light control with the LED light. Regarding the LED lights that are used, an optimal relationship is established between light, electricity, and thermal factors to maximize the ratio of luminous flux to input power. In this article, the LED lights are driven at the operating voltage where the optimal luminous flux occurs using a combination of heatsink characteristics, the PET theory, and a soft-switching buck–boost converter. The proposed LED light system is studied, and the results of theories are simulated and validated using PSCAD/EMTDC software. In general, the main contributions of this paper are as follows:

• Design of a soft-switching single-stage bidirectional buck–boost converter.
• Using a modified version of the conventional bidirectional boost converter, incorporating an auxiliary circuit encompassing a capacitor, inductor, and switch.
• Using the presented structure for supplying energy to LED streetlights.
• Effective energy management between the renewable system and the battery in order to power the load.

2. Proposed System Configuration and Operation

The proposed system is shown in Figure 2. This system has a soft-switching bidirectional dc/dc converter to harvest PV energy and feed LED light. As is observed in the figure, the proposed converter is fed current at both input and output ports which is a desirable feature of the converter in terms of the battery, PV energy, and LED light. Compared with the battery voltage level, the PV source and LED light have higher voltage ranges; therefore, the high-voltage side of the converter is connected to them using a static relay. As a result, throughout the day the PV energy is stored in the battery by the converter’s buck operation mode, while at night the LED light is fed by the battery through the converter’s boost operation mode.

As shown in Figure 3, during the day, the converter works in the buck circuit because the PV panel voltage is greater than the battery voltage. In this case, the relay is closed to the PV module and the Q₁ switch is activated using PWM, while the Q₂ switch is constantly off, and when the Q₁ switch is off, diode D₂ conducts. On the other hand, as shown in Figure 4, at night, the converter operates in the boost circuit and without PV power, the battery should supply a higher-voltage LED light. As a result, the relay is closed to the LED light and while the Q₁ switch is constantly off, Q₂ is switched in PWM. In this case, diode D₁ conducts during the Q₂ off-times. Now, if during the conduction interval of each diode at the same time the corresponding inverse parallel switch of that diode is turned on, the function of the circuit becomes acting as a synchronous rectifier for both buck and boost operation circuits, resulting in a reduction in converter conduction losses. Therefore, a pulse width modulation signal is applied to the Q₁ switch (with a duty ratio of d₁), and at the same time, the Q₂ switch operates in the opposite state of Q₁ (with the duty ratio of 1 − d₁); that is, when Q₁ is on, Q₂ is off and vice versa. For this situation, the voltage across the capacitor C₁ is given as

$$v_{C_1}=\frac{v_B}{1-d}$$

(1)

3. Proposed Control System

Figure 5 shows the proposed converter control system. According to the figure and during the day, the duty ratio d₁ is selected as the controlling variable to regulate the PV module current at the I_MPP and thereby control its generated power at P_MPP. Therefore, by using the MPPT algorithm, the maximum power operating current (I_MPP) of the PV module is determined. Then, to determine the converter’s first switch duty cycle d₁, this obtained current is applied to the utilized PI controller. Meanwhile, during the night, the duty ratio d₁ is chosen as the controlling variable to adjust the LED light current at its optimum current value. Therefore, to determine the duty cycle of the converter d₁, first, using the PET theory [44], the optimal operating current (I_Opt) of the LED light is obtained and this value is applied to the PI controller.

4. State of Charge (SOC) Unit of the Battery

The SOC block has the task of determining the power current of the converter after checking and analyzing it in such a way that the battery voltage is within the allowed range of V_Bmin < V_B < V_Bmax. (i.e., for lead–acid batteries V_Bmin = 11 V and V_Bmax = 13.5 V). The decisions made by the SOC unit are implemented in the converter through the utilization of two control signals, K and V_K, in conjunction with the V_sel switch. The flowchart depicting the battery SOC is presented in Figure 6, and its corresponding rules are discussed as follows:

V_B < V_Bmin: This mode occurs at night when the battery supplies the LED light for a long time. As a result, the battery does not have enough charge and its voltage falls lower than V_Bmin, so, the shutdown signal for the system is turned on by the SOC unit. When the shutdown signal arrives, the relay is disconnected from the LED. Simultaneously, the duty ratio of the converter becomes zero.

V_Bmin < V_B < V_Bmax: In this state, by setting K = 1, the V_sel signal can be considered as V_MPP, and by d₁, the PV voltage can be adjusted at V_MPP. This can lead to the battery charging with P_B = P_MPP.

V_B > V_Bmax: In this situation, when the battery is fully charged, the PV module is not operated at its MPP by setting the K signal of SOC to 0 (K = 0). Consequently, the SOC output signal V_K selects the V_sel signal (noting that V_K is usually greater than V_MPP). To keep the battery voltage constant at V_Bmax, the V_K signal is quickly adjusted by the SOC unit. As a result of this process, the average current of the battery becomes almost zero.

5. Proposed Converter Operation Analysis

In the proposed converter, using sufficiently large capacitance and inductance values for L₂, C₁, and C₂, these components function as nearly constant current and voltage sources for I_L2, V_C1, and V_C2, respectively. As a result, we have V_C1 > V_C2. Meanwhile, the state variables i_L1 and i_Lr are depicted with their ripple contents. In the converter analysis, if any inductor current moves to positive values, it is considered to be in a charging operating condition. Otherwise, if the inductor current moves to negative values, it is considered to be in a discharging operating condition.

5.1. Zero-Voltage Switching (ZVS)

ZVS techniques, which lead to soft switching of the converter’s power switches, have been highly regarded by designers of power electronic converters in recent years. The conventional square power conversion during key on-time with resonant switching transitions can be defined as ZVS. In general, it should be said that the basis of this function is equalizing the relationships obtained from the volt/second law at the input and output of the system. In soft-switching performance, when the switch is off, the LC circuits resonate. By doing this and by conducting the current, they bring the voltage across the switches to zero before turning it on again. Another technique that is used is the zero-current switching (ZCS) technique. In this technique, the switch is triggered when the current through it reaches zero. Similar to ZVS, the use of ZCS helps reduce switching losses [45,46].

Turning on the switch when the voltage across it is zero significantly reduces switching losses, leading to increased circuit efficiency. The following are the main advantages of using the ZVS technique:

• Reduced switching losses.
• Increased efficiency.
• Ability to withstand short circuits.
• Absence of current exceeding the peak current.
• Reduced electromagnetic interference (EMI).
• Possibility of high-frequency switching.

In addition to its advantages, ZVS has some disadvantages. ZVS methods may not perform effectively at light loads due to the requirement for a minimum current. Also, this method requires deadtime between switching switches. For more information about ZVS and ZCS, the reader is recommended to refer to Refs. [47,48].

5.2. Buck Operation Mode

In this section, we conduct an analysis of the operating mode of the proposed converter when it functions in buck mode, meaning that the PV-generated power is stored in the battery. As a result, negative and positive average current values are recognized for the converter’s first and second inductors, i.e., I_L1 < 0 and I_L2 > 0. The buck operation mode is divided into eight operating states, illustrated in Figure 7 and Figure 8, which display the converter’s key waveforms and its eight operating circuits, respectively. In each operating circuit shown in Figure 8, the actual flowing directions of the converter components are denoted by red arrows for better comprehension. The converter’s eight operating states are explained as follows.

Before the time t = 0, the antiparallel diode D₁ is conducting the negative current i_L1, providing ZVS conditions for the upcoming turning-on moment of the Q₁ switch. The Q₂ and Q₃ switches are also turned off and the resonant current i_Lr is positive and steadily decreasing from its maximum value (I_Lr^max) toward zero through the Q₃ switch’s antiparallel diode D₃.

State I (0 ≤ t < t₁): This state begins when the Q₁ switch is turned on with ZVS conditions because its antiparallel diode (D₁) is conducting from the preceding operating state. Two circuit loops of V_B:L₁:Q₁ and C₁:C₂:D₃:L_r:Q₁ are recognized for the converter inductors L₁ and L_r, respectively. Therefore, these inductors are put in charging and discharging operating conditions, respectively, as follows:

$$\begin{gathered} i_{L_1}=I_{L_1}^{min}+\frac{V_B}{L_1}t \\ {i}_{L_r}=i_{L_r}(0)-\frac{V_{C_1}-V_{C_2}}{L_r}t \end{gathered}$$

(2)

This interval ends at t₁ = ΔT_s, where the resonant inductor current i_Lr approaches zero and, thus, the diode D₁ stops conducting in ZCS conditions.

State II (t₁ ≤ t < t₂): In this operating state, the Q₂ and Q₃ switches remain turned off, while the Q₁ switch is turned on. Similar to the previous state, the negative current i_L1 is still in increasing mode through the circuit loop V_B:L₁:Q₁. In addition, the resonant current i_Lr remains zero until the next operating state starts. Therefore, we have the following:

$$\begin{gathered} i_{L_1}=I_{L_1}^{min}+\frac{V_B}{L_1}t \\ {i}_{L_r}=0 \end{gathered}$$

(3)

State III (t₂ ≤ t < t₃ = dT_s): In this operating state, the Q₁ and Q₂ switches remain turned on and off, respectively, while the Q₃ switch is turned on at t = t₂. Therefore, two circuit loops of V_B:L₁:Q₁ and C₁:C₂:Q₃:L_r:Q₁ are recognized for the converter inductors L₁ and L_r, respectively. Therefore, these inductors are again put in charging and discharging operating conditions, respectively, as follows:

$$\begin{gathered} i_{L_1}=I_{L_1}^{min}+\frac{V_B}{L_1}t \\ {i}_{L_r}=i_{L_r}(t_2)-\frac{V_{C_1}-V_{C_2}}{L_r}(t-t_2) \end{gathered}$$

(4)

For this operating state, the increasing current i_L1 and the decreasing resonant current i_Lr reach their maximum and minimum values of I_L1^max and I_Lr^min, respectively. Moreover, at the end of this operating state at t = t₃, the first switch current (i_S1 = i_L1 − i_Lr) becomes positive since |i_L1(t₃)| < |i_Lr(t₃)|.

State IV (t₃ = dT_s ≤ t < t₄): At t = t₃ = dT_s, the Q₁ switch is turned off, then its flowing positive current quickly transfers to the antiparallel diode D₂ of Q₂. This situation is the converter’s first switching deadtime, lasting for a short period of time to provide ZVS conditions for the Q₂ switch at its upcoming turning-on moment. The inductor currents, i_L1 and i_Lr, are expected to remain relatively stable compared with the previous operating state, maintaining their respective values at I_L1^max and I_Lr^min.

State V (t₄ ≤ t < t₅): At the initial time of this state, the Q₂ switch is turned on with ZVS, while the Q₃ switch is still turned on. Therefore, two circuit loops of V_B:L₁:Q₂:C1 and C₂:Q₃:L_r:Q₂ are recognized for the converter inductors L₁ and L_r, respectively. Neglecting the previous short deadtime operating state, these inductors are put in discharging and charging operating conditions, respectively, as follows:

$$\begin{gathered} i_{L_1}=I_{L_1}^{max}-\frac{V_{C_1}-V_B}{L_1}(t-d{T}_s) \\ {i}_{L_r}=I_{L_r}^{min}+\frac{V_{C_2}}{L_r}(t-d{T}_s) \end{gathered}$$

(5)

At t = t₅ the increasing resonant current i_Lr reaches zero, which ends this operating state.

State VI (t₅ ≤ t < t₆): This operating state is similar to the prior state, except that the charging current of the resonant current i_Lr now flows with positive values, passing through the antiparallel diode D₃ of Q₃. Therefore, two circuit loops of V_B:L₁:Q₂:C₁ and C₂:D₃:L_r:Q₂ are again recognized for the converter inductors L₁ and L_r, putting them in discharging and charging operating conditions, respectively, as follows:

$$\begin{gathered} i_{L_1}=I_{L_1}^{max}-\frac{V_{C_1}-V_B}{L_1}(t-d{T}_s) \\ {i}_{L_r}=I_{L_r}^{min}+\frac{V_{C_2}}{L_r}(t-d{T}_s) \end{gathered}$$

(6)

State VII (t₆ ≤ t < t₇): At the beginning of this state, the Q₃ switch is turned off with ZCS, since its antiparallel diode D₃ is conducting the positive current of the resonant inductor. All the other circuit conditions remain similar to those in the previous operating state, so that two circuit loops of V_B:L₁:Q₂:C₁ and C₂:D₃:L_r:Q₂ are recognized for the converter inductors L₁ and L_r, respectively. Therefore, these inductors are still put in discharging and charging conditions, respectively, as follows:

$$\begin{gathered} i_{L_1}=I_{L_1}^{max}-\frac{V_{C_1}-V_B}{L_1}(t-d{T}_s) \\ {i}_{L_r}=I_{L_r}^{min}+\frac{V_{C_2}}{L_r}+(t-d{T}_s) \end{gathered}$$

(7)

For this operating state, the decreasing current i_L1 and the increasing resonant current i_Lr reach their minimum and maximum values of I_L1^min and I_Lr^max, respectively. At the end of this operating state t = t₇, the second switch current (i_S2 = −i_L1 + i_Lr) is positive.

State VIII (t₇ ≤ t < T_s): At t = t₇ the Q₂ switch is turned on, then its flowing positive current quickly transfers to the antiparallel diode D₁ of Q₁. This situation is the converter’s second switching deadtime, providing ZVS conditions for the Q₁ switch at its upcoming turning-on moment. The inductor currents, i_L1 and i_Lr, are expected to remain relatively stable compared with the previous operating state, maintaining their respective values at I_L1^min and I_Lr^max.

5.3. Boost Operation Mode

In this section, the operating mode of the proposed converter is analyzed when it operates in boost mode, i.e., when the battery-generated power is consumed by the LED light. As a result, positive and negative average current values are recognized for the converter’s first and second inductors, i.e., I_L1 > 0 and I_L2 < 0. The converter’s key waveforms are illustrated in Figure 9. In each operating circuit shown in Figure 9, the real flowing directions of the converter components are denoted by red arrows for better understanding. Additionally, the boost operation mode is also divided into eight operating states, shown in Figure 10. The eight operating states of the converter are explained as follows.

Before t = 0, the antiparallel diode D1 conducts a positive current (i_Lr − i_Lr), ensuring ZVS conditions for the upcoming Q₁ switch turning-on event. The Q₂ and Q₃ switches are also turned off and the resonant current i_Lr is positive and gradually decreases from its maximum value (I_Lr^max) toward zero through the Q₃ switch’s antiparallel diode D₃.

State I (0 ≤ t < t₁): This state begins when the Q₁ switch is turned on with ZVS conditions because its antiparallel diode (D₁) is conducting from the preceding state. Two circuit loops of V_B:L₁:Q₁ and C₁:C₂:D₃:L_r:Q₁ are recognized for the converter inductors L₁ and L_r, respectively. Therefore, these inductors are put in charging and discharging operating conditions, respectively, as follows:

$$\begin{gathered} i_{L_1}=I_{L_1}^{min}+\frac{V_B}{L_1}t \\ {i}_{L_r}=i_{L_r}(0)-\frac{V_{C_1}-V_{C_2}}{L_r}t \end{gathered}$$

(8)

This interval ends at t₁ = ΔT_s, where the resonant inductor current i_Lr approaches zero and, thus, the diode D₁ stops conducting in ZCS conditions.

State II (t₁ ≤ t < t₂): In this operating state, the Q₂ and Q₃ switches remain turned off, while the Q₁ switch is turned on. Similar to the previous state, the positive current i_L1 is still in a growing mode over the circuit loop V_B:L₁:Q₁. In addition, the resonant current i_Lr remains zero until the next operating state starts. Therefore, we have the following:

$$\begin{gathered} i_{L_1}=I_{L_1}^{min}+\frac{V_B}{L_1}t \\ {i}_{L_r}=0 \end{gathered}$$

(9)

State III (t₂ ≤ t < t₃ = dT_s): In this operating state, Q₁ and Q₂ remain turned on and off, respectively, while Q₃ is turned on at t = t₂. Therefore, two circuit loops of V_B:L₁:Q₁ and C₁:C₂:Q₃:L_r:Q₁ are recognized for the converter inductors L₁ and L_r, respectively. Therefore, these inductors are put in charging and discharging operating conditions again, respectively, as follows:

$$\begin{gathered} i_{L_1}=I_{L_1}^{min}+\frac{V_B}{L_1}t \\ {i}_{L_r}=i_{L_r}(t_2)-\frac{V_{C_1}-V_{C_2}}{L_r}(t-t_2) \end{gathered}$$

(10)

For this state, the increasing current i_L1 and the decreasing resonant current i_Lr reach their maximum and minimum values of I_L1^max and I_Lr^min, respectively. Moreover, at the end of this operating state t = t₃, the first switch current (i_S1 = i_L1 − i_Lr) is positive since i_L1(t₃) > 0 and i_Lr(t₃) < 0.

State IV (t₃ = dT_s ≤ t < t₄): At t = t₃ = dT_s, Q₁ is turned off, then its flowing positive current quickly transfers to the antiparallel diode D₂ of Q₂. This situation is the converter’s first switching deadtime, lasting for a short period of time to provide ZVS conditions for Q₂ at its upcoming turning-on moment. The inductor currents i_L1 and i_Lr are supposed to not change significantly compared with the previous operating state and, therefore, they remain at values of I_L1^max and I_Lr^min, respectively.

State V (t₄ ≤ t < t₅): At the beginning of this operating state, Q₂ is turned on with ZVS, while Q₃ is still turned on from t = t₂. Therefore, two circuit loops of V_B:L₁:Q₂:C1 and C₂:Q₃:L_r:Q₂ are recognized for the converter inductors L₁ and L_r, respectively. Neglecting the previous short deadtime operating state, these inductors are put in discharging and charging operating conditions, respectively, as follows:

$$\begin{gathered} i_{L_1}=I_{L_1}^{max}-\frac{V_{C_1}-V_B}{L_1}(t-d{T}_s) \\ {i}_{L_r}=I_{L_r}^{min}+\frac{V_{C_2}}{L_r}+(t-d{T}_s) \end{gathered}$$

(11)

At t = t₅, the increasing resonant current i_Lr reaches zero, which ends this operating state.

State VI (t₅ ≤ t < t₆): This operating state is similar to the previous one, except that the resonant current, i_Lr, now flows with positive values, passing through the antiparallel diode of Q3. Therefore, two circuit loops of V_B:L₁:Q₂:C1 and C₂:D₃:L_r:Q₂ are again recognized for the converter inductors L₁ and L_r, putting them in discharging and charging operating conditions, respectively, as follows:

$$\begin{gathered} i_{L_1}=I_{L_1}^{max}-\frac{V_{C_1}-V_B}{L_1}(t-d{T}_s) \\ {i}_{L_r}=I_{L_r}^{min}+\frac{V_{C_2}}{L_r}(t-d{T}_s) \end{gathered}$$

(12)

State VII (t₆ ≤ t < t₇): At the beginning of this state, Q₃ is turned off with ZCS, since its antiparallel diode is conducting the positive current of the resonant inductor. All the other circuit conditions remain similar to those in the previous operating state, so that two circuit loops of V_B:L₁:Q₂:C1 and C₂:D₃:L_r:Q₂ are recognized for the converter inductors L₁ and L_r, respectively. Therefore, these inductors are still put in discharging and charging conditions, respectively, as follows:

$$\begin{gathered} i_{L_1}=I_{L_1}^{max}-\frac{V_{C_1}-V_B}{L_1}(t-d{T}_s) \\ {i}_{L_r}=I_{L_r}^{min}+\frac{V_{C_2}}{L_r}(t-d{T}_s) \end{gathered}$$

(13)

For this operating state, the decreasing current i_L1 and the increasing resonant current i_Lr reach their minimum and maximum values of I_L1^min and I_Lr^max, respectively. At the end of this operating state t = t₇, the second switch current (i_S2 = −i_L1 + i_Lr) is positive, since |i_L1(t₇)| < |i_Lr(t₇)|.

State VIII (t₇ ≤ t < T_s): At t = t₇, Q₂ is turned on, then its flowing positive current quickly transfers to the antiparallel diode D₁ of Q₁. This situation is the converter’s second switching deadtime, providing ZVS conditions for Q₁ at its upcoming turning-on moment. The inductor currents i_L1 and i_Lr are supposed to not change significantly compared with the previous operating state and, therefore, they remain at values of I_L1^min and I_Lr^max, respectively.

6. Simulation Results

We assume the use of a 60 W LED streetlight, a PV module, and a 12 V battery source for investigating the proposed system. To address the issue of weak brightness from a single LED, we can connect multiple LEDs in series and parallel [49,50]. In the simulation, 20 LEDs are used in two strings, with each string having 10 LEDs in series.

Table 1. LED streetlight simulation parameters.

Symbols	Simulation Parameters	Symbols	Simulation Parameters
WLED	0.14 (m)	M	18
LLED	0.16 (m)	ke	−0.0045
Wb	0.20 (m)	khs	200 (w/m·k)
Lb	0.22 (m)	E0	78.5 (lm/w)
tb	0.015 (m)	hfluid	95 (w/m·k)
tf	0.005 (m)	Rjc	8 (°C/W)
Hf	0.01 (m)	M	18

displays the parameters of the LED streetlight used, with each LED having a maximum power of 3 W. In Figure 11, the P-V output characteristics of the PV module are shown for three different examined environmental conditions. Also, a 12 V lead–acid battery source is used for the proposed system. In this simulation, the PV module should charge the battery during the day under MPPT conditions. However, at night, the battery supplies the LED light at the optimal current. The proposed converter is switched under the soft-switching condition of ZVS in both its operation modes and ZCS. In buck and boost operation modes, and as seen in the simulation waveforms, the voltages of Q₁ and Q₂ were previously decreased to the zero level, denoting successful ZVS conditions for these switches. Moreover, Q₃ is turned off when its current is negative (following through D₃), thus providing ZCS conditions for this switch.

In all the simulations, the possible minimum and maximum levels of the solar irradiations and the ambient temperature were considered. The system was simulated by applying PSCAD/EMTDC software. The simulation of the proposed system was conducted in two modes, day and night. System operation in the day and night modes includes three and two operational stages, respectively, which are discussed below.

6.1. Daytime Operation

For day operation times, the simulation was performed in three different stages. In all stages, I_L2ref was chosen to be the I_MPP of the PV module. The battery voltage V_B, PV voltage V_PV, currents I_L2, V_C1, V_C2, resonant inductor current i_Lr, and converter duty ratio are clearly shown in Figure 12.

First simulation stage 0 ≤ t < 0.06 s: The maximum amount of solar radiation (G = 1000 W/m²) at the ambient temperature (35 °C) is considered in this stage. First, by using the MPPT algorithm, the maximum power operating current (I_MPP) of the PV module is specified. Then, to determine the first switch duty cycle d, this obtained value is applied to the PI controller. According to the characteristics of the module by duty ratio d = 0.54, the maximum voltage and current of the PV module are obtained as 27.6 V and 2.37 A, respectively. In this mode, P_PV = 65.4 W is transmitted through the buck operation mode of the converter to charge the battery with a current of I_B = −5.1 A and voltage of V_B = 12 V. For more clarification, the currents in the converter’s inductors (I_L1, I_L2, I_Lr), the currents in the converter’s switches (I_s1, I_s2, I_s3), the drain–source voltages across the converter’s switches (V_ds1, V_ds2, V_ds3), and the switches’ gate signals are illustrated in Figure 13 for two switching periods.

Second simulation stage 0.06 ≤ t < 0.12 s: The amount of radiation is reduced (G = 600 W/m²) in this stage, but the temperature is still 35 °C. Considering that the voltage and current of the module can be affected by radiation, by using the duty cycle d = 0.48, the magnitude of these values is obtained as V_PV = 24.1 V, I_PV = 1.44 A. Therefore, the output power of the PV module decreases (P_PV = 34.7 W) due to the reduction in radiation. Also, the battery charging current and voltage are reduced to I_B = −2.8 A and V_B = 12 V, respectively. For more clarification, the currents in the converter’s inductors (I_L1, I_L2, I_Lr), the currents in the converter’s switches (I_s1, I_s2, I_s3), the drain–source voltages across the converter’s switches (V_ds1, V_ds2, V_ds3), and the switches’ gate signals are illustrated in Figure 14 for two switching periods.

Third simulation stage 0.12 ≤ t < 0.18 s: At night, the air cools down and the temperature drops below zero. Due to this decrease in temperature, the level of radiation and the temperature of the PV module are equal to 600 W/m² and −5 °C, respectively. In the duty cycle d = 0.58, the MPPT control sets the values of the voltage, current, and PV power at values of 30 V, 1.4 A, and 42 W, respectively. In this condition, the battery voltage is equal to V_B = 12 V and it continues to charge with a current of I_B = −3.4 A. For more clarification, the currents in the converter’s inductors (I_L1, I_L2, I_Lr), the currents in the converter’s switches (I_s1, I_s2, I_s3), the drain–source voltages across the converter’s switches (V_ds1, V_ds2, V_ds3), and the switches’ gate signals are illustrated in Figure 15 for two switching periods.

6.2. Nighttime Operation

The relay is closed to the LED light at night and its simulation is performed in two different stages. In both stages, I_led is chosen to be set at I_L2ref, which is determined by (17) the optimal current of the LED light. The battery voltage V_B, the LED light voltage and current V_led, I_L2, the voltages V_C1, V_C2, the resonant inductor current i_Lr, and the converter duty ratio are clearly shown in Figure 16.

First simulation stage 0 ≤ t < 0.09 s: In this stage, the ambient temperature is considered to be 35 °C due to the effect of temperature on LED light. First, by using PET theory, the optimal operating current of the LED light is determined. Next, to determine the duty cycle of the converter d, this obtained current (I_L2ref = −I_opt) is applied to the PI controller. Thus, the streetlight is driven in the optimal current and voltage values of V_LED = 37.5 V and I_LED = 1.35 A, respectively, which results in a discharging current of I_B = 4.5 A and a voltage of V_B = 12 V for the battery. For more clarification, the currents in the converter’s inductors (I_L1, I_L2, I_Lr), the currents in the converter’s switches (I_s1, I_s2, I_s3), the drain–source voltages across the converter’s switches (V_ds1, V_ds2, V_ds3), and the switches’ gate signals are illustrated in Figure 17 for two switching periods.

Second simulation stage 0.09 ≤ t < 0.18 s: At this step, the ambient temperature drops to −5 °C and other specifications are the same as in the previous stage. Therefore, the optimum operating point of the LED light changes to V_LED = 38 V and I_LED = 1.2 A, causing the discharging current to be I_B = 4 A and voltage to be V_B = 12 V for the battery. To provide further clarity, Figure 18 illustrates the currents in the converter’s inductors (I_L1, I_L2, I_Lr), the currents in the converter’s switches (I_s1, I_s2, I_s3), the drain–source voltages across the switches (V_ds1, V_ds2, V_ds3), and the gate signals for two switching periods.

As shown in all the presented simulation results, the turning-on and turning-off times of the S1 and S2 switches are opposite to each other. In all cases, the current of the resonant inductor L_r is greater than the current of the other two inductors L₁ and L₂, which is necessary to create soft-switching conditions. Additionally, soft switching of the power switches of the converter is performed correctly, as shown in the voltage sections of the switches and their gate signals.

7. Conclusions

This article introduces a bidirectional boost converter to manage PV-powered LED lighting and battery storage. To enhance efficiency, a soft-switched version of the converter is created by incorporating an L-C-Switch series resonant circuit into the standard bidirectional DC/DC converter. The adopted topology enables soft switching in both buck and boost operational modes. During the day, the proposed converter operates in a manner that allows it to efficiently transfer PV MPP energy to the battery in the buck operational mode. At night, in boost operation mode, the LED light is powered by the battery. This structure, which can change its mode of operation, offers a cost-effective and simplified design. A comprehensive control strategy is proposed for the system in both cases to regulate current at the converter’s input and output terminals, SOC for the battery, MPPT for the PV panel, and light control for the LED light. The theory used in this system considers the optical, electrical, and thermal characteristics of the LEDs, optimizing the LED light’s operating current to achieve maximum luminous flux. An efficient control system was designed to adjust the LED light voltage, charge the battery, and perform MPPT. The proposed system offers compact and single-stage power conversion. Finally, the system’s performance was successfully validated in various environmental conditions using PSCAD/EMTDC software. Future work can explore alternative methods for achieving soft switching of the power switches and investigate the feasibility of using multiple input sources for the converter.