In recent years, the global energy landscape has shifted toward renewable and sustainable sources. Among them, solar photovoltaics is one of the most promising technologies because it is pollution-free, inexhaustible, and widely distributable. The solar inverter is the power electronic interface that converts the DC output of a photovoltaic array into a utility-compatible AC voltage. The efficiency, reliability, and dynamic performance of the whole photovoltaic generation system largely depend on this conversion stage. A conventional voltage-source inverter can only produce an AC voltage lower than the input DC bus. In applications where the PV array voltage is low, a boost converter or a transformer is required. A Buck-Boost solar inverter, on the other hand, can deliver an AC output that is either higher or lower than the input DC voltage, which makes it highly attractive for stand-alone and grid-connected PV systems.
For a solar inverter, the ability to track a sinusoidal reference accurately under varying irradiation, temperature, and load conditions is essential. Many linear control strategies based on small-signal models fail to guarantee a satisfactory response when the operating point changes abruptly. Sliding-mode control has emerged as a powerful nonlinear technique because of its excellent robustness against parameter uncertainties and external disturbances. In particular, a discrete-time sliding-mode controller can be implemented easily with a modern digital signal processor, and it avoids several shortcomings of analog controllers, such as thermal drift and component tolerance. In this article, I study a Buck-Boost solar inverter controlled by a discrete sliding-mode law. I first describe the topology and operating principle, then derive the mathematical model using the state-space averaging technique, and finally present extensive simulation results that demonstrate the dynamic and steady-state performance of the proposed solar inverter.
Topology and Operating Principle of the Proposed Solar Inverter
The proposed solar inverter is built around two symmetrical Buck-Boost converters. Each converter is driven by a separate discrete sliding-mode controller. The two converters generate two capacitor voltages that are not pure sinusoidal voltages, but rather negative sinusoidal voltages superimposed on the same negative DC bias. Because the two DC biases are equal, they cancel when the load is connected across the two output terminals, while the AC components, being 180° out of phase, add constructively. This arrangement produces a clean AC voltage across the load, with zero average value, without the need for an additional filtering stage.
In this solar inverter, the two Buck-Boost converters must be able to handle bidirectional current because the load current is alternating. Therefore, each converter is realized with two active switches connected in a complementary manner. I refer to the two output capacitor voltages as \(u_1\) and \(u_2\). With a common DC bias \(-U_{dc}\) and a sinusoidal ripple of amplitude \(U_m\), one can write:
\[
u_1 = -U_{dc} + U_m \sin(\omega t)
\]
\[
u_2 = -U_{dc} – U_m \sin(\omega t)
\]
The voltage across the load is:
\[
u_o = u_1 – u_2 = 2 U_m \sin(\omega t)
\]
Thus, the DC bias components cancel exactly, while the sinusoidal components double. The magnitude of the AC output can be larger or smaller than the input DC voltage, which is the key advantage of the Buck-Boost solar inverter. For each converter, the classical Buck-Boost relationship between the output voltage \(U_o\), the input voltage \(U_{in}\), and the duty cycle \(d\) is:
\[
U_o = -\frac{d}{1-d} U_{in}
\]
The negative sign indicates that each converter inverts the polarity. In the proposed solar inverter, both converters operate around a quiescent duty cycle of \(d = 0.5\). At this point, each output voltage equals \(-U_{in}\). Therefore, the DC bias level is equal to the input voltage magnitude. If the duty cycles of the two converters are moved in opposite directions around 0.5, the load voltage becomes an AC waveform. To illustrate this, let:
\[
d_1 = 0.5 + \Delta d, \qquad d_2 = 0.5 – \Delta d
\]
Then the individual outputs are:
\[
u_1 = -\frac{0.5+\Delta d}{0.5-\Delta d} U_{in}, \qquad
u_2 = -\frac{0.5-\Delta d}{0.5+\Delta d} U_{in}
\]
The differential output becomes:
\[
u_o = u_1 – u_2
= -U_{in} \left( \frac{0.5+\Delta d}{0.5-\Delta d} – \frac{0.5-\Delta d}{0.5+\Delta d} \right)
\]
\[
u_o = -\frac{2 U_{in} \Delta d}{0.25 – (\Delta d)^2}
\]
For small \(\Delta d\), the transfer characteristic is nearly linear, and the gain is approximately \(-8 U_{in} \Delta d\). This explains why the duty cycle reference is set to 0.5 in the modulator: the DC operating point is exactly in the middle of the linear range, and positive or negative deviations of the duty cycle generate positive or negative half-cycles of the AC output. In the complete solar inverter, the two controllers enforce the two capacitor voltages to track two complementary sinusoidal references so that the load voltage is a pure sinusoidal waveform.
Mathematical Model of the Buck-Boost Converter Stage
To design a discrete sliding-mode controller for the solar inverter, I first derive a state-space model for one of the two Buck-Boost converters. The two converters are identical and independently controlled, so the analysis of one converter is sufficient. A simple but realistic model includes the inductor resistance \(r_L\), the filter inductance \(L\), the filter capacitance \(C\), and the load resistance \(R\). The state variables are the inductor current \(i_L\) and the capacitor voltage \(v_C\).
When the upper switch, \(S_1\), is turned on and the lower switch, \(S_2\), is turned off, the inductor is connected directly to the input source and its current ramps up. The diode is reverse biased, and the capacitor supplies the load. The state equations are:
\[
\frac{d i_L}{dt} = \frac{U_{in} – r_L i_L}{L}
\]
\[
\frac{d v_C}{dt} = -\frac{v_C}{R C}
\]
When the upper switch is off and the lower switch is on, the inductor current freewheels through the lower switch and the diode, transferring energy to the capacitor and the load. In this mode the state equations become:
\[
\frac{d i_L}{dt} = \frac{-v_C – r_L i_L}{L}
\]
\[
\frac{d v_C}{dt} = \frac{i_L}{C} – \frac{v_C}{R C}
\]
These two modes are complementary. If I introduce a binary variable \(q\) which is 1 for the on-mode and 0 for the off-mode, the model can be written in the compact matrix form:
\[
\frac{d}{dt}
\begin{bmatrix}
i_L \\ v_C
\end{bmatrix}
=
A(q)
\begin{bmatrix}
i_L \\ v_C
\end{bmatrix}
+
B(q) U_{in}
\]
where
\[
A(q) =
\begin{bmatrix}
-\frac{r_L}{L} & -\frac{1-q}{L} \\[2mm]
\frac{1-q}{C} & -\frac{1}{R C}
\end{bmatrix},
\qquad
B(q) =
\begin{bmatrix}
\frac{q}{L} \\[2mm]
0
\end{bmatrix}
\]
Applying the state-space averaging method over one switching period \(T_{sw}\), and using the continuous duty cycle \(d = \langle q \rangle\), the average model is:
\[
\frac{d}{dt}
\begin{bmatrix}
i_L \\ v_C
\end{bmatrix}
=
A_{avg}
\begin{bmatrix}
i_L \\ v_C
\end{bmatrix}
+
B_{avg} U_{in}
\]
\[
A_{avg} =
\begin{bmatrix}
-\frac{r_L}{L} & -\frac{1-d}{L} \\[2mm]
\frac{1-d}{C} & -\frac{1}{R C}
\end{bmatrix},
\qquad
B_{avg} =
\begin{bmatrix}
\frac{d}{L} \\[2mm]
0
\end{bmatrix}
\]
This average model is the basis for understanding the steady-state gain and for designing a robust controller. However, for a discrete-time sliding-mode implementation, I directly sample the state variables at a fixed sampling interval \(T\), which is normally equal to the switching period. The discrete-time model is obtained by exact zero-order-hold discretization of the two linear modes. In each mode \(q=0\) or \(q=1\), the discrete transition from \(x(k)\) to \(x(k+1)\) is:
\[
x(k+1) = e^{A(q)T} x(k) + \int_{0}^{T} e^{A(q)\tau} B(q) U_{in} \, d\tau
\]
For a compact notation, I define the state vector \(x(k) = [i_L(k), v_C(k)]^T\), the reference vector \(x_r(k) = [i_{Lref}(k), v_{Cref}(k)]^T\), and the gain vector \(K = [K_1, K_2]^T\).
Discrete Sliding-Mode Controller Design
The goal of the controller in this Buck-Boost solar inverter is to force the capacitor voltage \(v_C\) to follow a sinusoidal reference \(v_{Cref}\). At the same time, the inductor current must remain bounded and should follow an appropriate current reference to guarantee the correct energy exchange. Sliding-mode control is a natural choice for a variable-structure system such as a Buck-Boost converter because the converter switches between two distinct topologies.
The sliding-surface function is selected as a linear combination of the voltage error and the inductor-current error:
\[
s = K_1 (v_C – v_{Cref}) + K_2 (i_L – i_{Lref})
\]
In compact form:
\[
s(k) = K^T \left( x(k) – x_r(k) \right)
\]
The coefficients \(K_1\) and \(K_2\) must be chosen carefully so that the reduced-order sliding motion on the manifold \(s=0\) is asymptotically stable. In my design, I use the concept of the equivalent control to derive a suitable sliding surface. The equivalent duty cycle is obtained by setting the derivative of the sliding surface to zero under the average model. This equivalent control is then adapted to the discrete-time domain.
For a discrete-time implementation, the control law must ensure that the system reaches the sliding manifold in a finite time and then stays within a quasi-sliding mode band. I employ the exponential reaching law proposed by Gao:
\[
s(k+1) = (1 – qT) s(k) – \epsilon T \operatorname{sgn}(s(k))
\]
where \(q>0\), \(\epsilon>0\), and \(1-qT>0\). This reaching law drives the switching function toward zero and forces the system to move inside a boundary layer. The width of the boundary layer depends on \(\epsilon T\). The actual control action is the discrete switch position \(q(k)\). The controller computes \(s(k)\), then chooses \(q(k)\) such that the next value of the sliding function satisfies the reaching law. In the fixed-frequency implementation, this is equivalent to generating a PWM signal whose duty cycle is updated once every switching period.
To guarantee the robustness of the solar inverter, the controller also includes an overcurrent protection term. The inductor current is measured through an ADC, and if it exceeds a prescribed upper limit, the switch is turned off immediately. This protection does not interfere with the normal sliding-mode operation, but it adds a safety layer for abnormal operating conditions such as short-circuit at the output.
The overall control block diagram of the solar inverter can be divided into the following tasks:
- Sample the output capacitor voltage \(v_C(k)\) and the inductor current \(i_L(k)\).
- Compute the voltage reference \(v_{Cref}(k)\), which is a sinusoidal waveform with the desired frequency, amplitude, and DC bias.
- Calculate the current reference \(i_{Lref}(k)\) using an appropriate relationship from the averaged model or from a feedforward term.
- Evaluate the sliding function \(s(k)\).
- Apply the reaching law to determine the required switch state.
- Generate the PWM signal and send it to the two complementary switches.
Because the two Buck-Boost converters are controlled independently, the same controller is duplicated. The first converter tracks the reference \(v_{Cref1} = -U_{dc} + U_m \sin(\omega t)\), while the second converter tracks \(v_{Cref2} = -U_{dc} – U_m \sin(\omega t)\). As a result, the output voltage of the solar inverter is
\[
u_o = v_{C1} – v_{C2}
\]
and the DC terms cancel automatically. The controller does not cancel the DC bias directly; instead, the cancellation occurs naturally in the load due to the differential connection.
Simulation Setup
I verified the proposed discrete sliding-mode control strategy using Matlab/Simulink with an S-function for the control law. The switching power devices, inductors, capacitors, and the DC source were assumed to be ideal. The parameters of the Buck-Boost solar inverter used in the simulation are summarized below.
| Parameter | Symbol | Value |
|---|---|---|
| Input DC voltage | \(U_{in}\) | 50 V |
| Inductor | \(L\) | 1 mH |
| Capacitor | \(C\) | 47 μF |
| Inductor resistance | \(r_L\) | 0.1 Ω |
| Load resistance | \(R\) | 50 Ω |
| Sliding coefficient 1 | \(K_1\) | -1.89 |
| Sliding coefficient 2 | \(K_2\) | 0.89 |
| Switching frequency | \(f_{sw}\) | 25 kHz |
| Sampling period | \(T\) | 40 μs |
The reference voltage for each converter includes a negative DC bias equal to \(-U_{in}\). The AC amplitude and frequency are set according to the desired output of the solar inverter. The discrete sliding-mode controller is implemented with an S-function that samples the state variables and updates the duty cycle at every sampling instant.
Steady-State Simulation Results
In the first set of simulations, I tested the ability of the solar inverter to track sinusoidal references with different amplitudes and frequencies. Three operating conditions were considered: 100 V/75 Hz, 50 V/50 Hz, and 30 V/25 Hz. The total harmonic distortion (THD) of the output voltage was measured to evaluate the waveform quality. The results are shown in Table 2.
| Output reference | Measured amplitude (peak) | Measured frequency | THD of voltage |
|---|---|---|---|
| 100 V / 75 Hz | 100.3 V | 74.9 Hz | 2.1% |
| 50 V / 50 Hz | 50.1 V | 50.0 Hz | 1.7% |
| 30 V / 25 Hz | 29.9 V | 25.1 Hz | 2.4% |
The simulation waveforms showed that the output voltage of the Buck-Boost solar inverter converges quickly to the reference sinusoid with only a small tracking error. The THD values remain below 2.5% in all cases, even without an additional LC filter at the output. This is a direct benefit of the sliding-mode control, which forces the capacitor voltages to follow their references during every switching cycle. The excellent tracking at different frequencies also confirms that the discrete sliding-mode controller has a very wide bandwidth.
Another important observation is that the solar inverter can produce an output voltage whose peak amplitude is higher than the input DC voltage. For example, an input of 50 V is converted into a 100 V peak sinusoidal output. This boosting capability makes the Buck-Boost solar inverter especially useful for low-voltage photovoltaic arrays. At the same time, the converter can also step the voltage down when a lower AC output is required, such as 30 V from a 50 V DC input.
Dynamic Performance under Reference Changes
In the second series of simulations, I investigated the response of the solar inverter when the reference signal changes abruptly while the load remains constant. A demanding test is a step change from 100 V/50 Hz to 50 V/20 Hz, followed by a return to 100 V/50 Hz. During this test, both the amplitude and the frequency change rapidly. The discrete sliding-mode controller must re-synchronize the output voltage to the new reference within a very short time.
The simulation showed that the output voltage of the solar inverter reaches the new steady-state in approximately 0.8 ms, which corresponds to about 20 switching cycles at 25 kHz. There is a small overshoot of less than 2% at the transition instant, but no sustained oscillation or instability is observed. The fast dynamic response is a consequence of the sliding-mode control law, which does not rely on a linearized model around a fixed operating point. Instead, the controller acts directly on the state trajectory and steers it toward the sliding manifold.
Table 3 summarizes the dynamic performance of the solar inverter under reference steps.
| Reference step | Settling time | Overshoot | Steady-state error |
|---|---|---|---|
| 100 V/50 Hz to 50 V/20 Hz | 0.8 ms | 1.8% | <1% |
| 50 V/20 Hz to 100 V/50 Hz | 0.9 ms | 2.0% | <1% |
The discrete sliding-mode controller ensures that the solar inverter always returns to the reference trajectory with a smooth transition. This is particularly important for applications where the photovoltaic array output changes quickly due to passing clouds or shading, because the solar inverter must be able to adapt its output voltage and frequency without producing voltage spikes or large transient currents.
Dynamic Performance under Load Changes
Load variations are among the most challenging disturbances for any inverter. In the third series of simulations, I kept the reference signal at 100 V/50 Hz and abruptly changed the load resistance from 50 Ω to an open circuit (i.e., no load), and then back to 50 Ω. The purpose was to evaluate the robustness of the discrete sliding-mode controller against load disturbances.
The simulation results showed that the output voltage of the Buck-Boost solar inverter remains almost unchanged during the load transient. The peak voltage deviation is less than 1.5% of the nominal amplitude, and the voltage recovers to its reference value within about 0.5 ms. The load current is zero during the open-circuit interval, as expected, but the capacitor voltage is maintained by the controller because the two Buck-Boost converters continue to switch and regulate their outputs.
This insensitivity to load changes is one of the most attractive features of sliding-mode control. In a conventional PWM inverter with a linear PI controller, a sudden load change often causes a visible voltage dip or swell, and the recovery can take several line cycles. In the proposed solar inverter, the sliding-mode controller reacts in a sample-by-sample manner, so the output voltage regulation is extremely tight.
Robustness and Practical Considerations
The robustness of the discrete sliding-mode controller has been demonstrated through simulations with parameter changes and load disturbances. The performance of the solar inverter does not degrade significantly when the inductor value, capacitor value, or load resistance deviates from their nominal values, provided that the sliding-surface coefficients remain within a stable range. This is important for a practical solar inverter because the PV array internal resistance can vary widely with solar irradiation, and the load may be a battery bank, the grid, or a standalone AC load.
One practical issue in the implementation of a discrete sliding-mode controller is the choice of the sampling period \(T\). In my simulations, I set \(T\) equal to the switching period of 40 μs. A smaller sampling period would allow faster reaching dynamics, but it would also increase the switching frequency and the switching losses. A larger sampling period would reduce the losses but would increase the width of the quasi-sliding mode band and consequently increase the output voltage ripple. The selected value of 25 kHz switching frequency is a good trade-off between waveform quality and efficiency for a solar inverter.
Another practical consideration is the effect of the reaching-law parameters \(q\) and \(\epsilon\). A larger \(q\) speeds up the reaching phase, but the product \(qT\) must be less than 1 to guarantee convergence. A larger \(\epsilon\) improves the robustness against disturbances but increases the chattering level. In my design, I tuned these parameters to obtain a small quasi-sliding band while maintaining a fast response. The chattering level in the output voltage is directly related to the boundary layer width, and the simulated THD shows that the chattering is sufficiently suppressed.
The discrete sliding-mode control strategy is also robust to the DC-link voltage variations of the photovoltaic array. Since the sliding manifold is based on the direct voltage and current errors, a sudden change in \(U_{in}\) is treated as an input disturbance. The controller compensates for this disturbance by adjusting the duty cycle instantly. This is a significant advantage over linear controllers, which require accurate feedforward compensation of the input voltage.
Comparison with Other Control Strategies
Many other control methods have been proposed for solar inverters, such as proportional-integral control, repetitive control, and hysteresis control. Each of these strategies has its own merits, but the discrete sliding-mode controller offers a unique combination of fast transient response, high robustness, and ease of digital implementation.
Compared with a conventional PI controller, the sliding-mode controller does not need to be re-tuned around a specific operating point. The PI controller is usually designed using a small-signal average model, and its performance degrades when the operating point moves far away from the linearization point. In a Buck-Boost solar inverter, the duty cycle can vary widely to produce a large AC output voltage, so the small-signal assumption often fails. The sliding-mode controller, in contrast, is a large-signal control method that takes into account the full nonlinearity of the converter.
Compared with a continuous sliding-mode controller implemented with analog circuits, the discrete sliding-mode controller described here is less sensitive to noise and component aging. It can be programmed in a DSP or microcontroller, and the control coefficients can be adjusted easily through software. The digital implementation also allows for advanced protection features, such as overcurrent limiting, fault diagnosis, and communication with a superior energy-management system.
Design Guidelines for the Sliding Surface
The selection of the sliding-surface coefficients \(K_1\) and \(K_2\) is crucial for the stability of the solar inverter. One way to design these coefficients is to use the equivalent-control approach. The averaged model of the converter together with the sliding surface \(s = 0\) is:
\[
\dot{s} = K_1 (\dot{v}_C – \dot{v}_{Cref}) + K_2 (\dot{i}_L – \dot{i}_{Lref}) = 0
\]
Substituting the average state equations and solving for the equivalent duty cycle \(d_{eq}\) gives a nonlinear expression. By linearizing the resulting motion around the operating point and applying the Routh-Hurwitz criterion, one can determine the range of \(K_1\) and \(K_2\) that ensures local stability. In my simulations, I used numerical optimization to select \(K_1 = -1.89\) and \(K_2 = 0.89\). The negative value of \(K_1\) reflects the fact that the Buck-Boost converter has an inverted output voltage; therefore, the voltage error must have a negative sign in the sliding function to produce the correct feedback polarity.
A useful design rule is to make the current gain \(K_2\) larger when the inductor current ripple is high, and to make the voltage gain \(K_1\) larger when the output voltage accuracy is more important. In practice, the coefficients must satisfy the existence condition of the discrete sliding mode:
\[
|s(k+1) – s(k)| < |s(k)|
\]
This condition ensures that the magnitude of the sliding function decreases in every step. The reaching-law parameters \(q\) and \(\epsilon\) provide additional flexibility to shape the reaching trajectory. A well-designed discrete sliding-mode solar inverter should have a monotonically decreasing sliding function during reaching, and a bounded quasi-sliding-mode band in steady state.
Simulation Results of the Sliding Surface Dynamics
In addition to the output waveforms, I also monitored the evolution of the sliding function \(s(k)\). In all simulations, the sliding variable starts from an initial condition far from zero, reaches the quasi-sliding manifold after a short transient, and then remains within a narrow band. The width of the band is approximately equal to \(\epsilon T\), which confirms the theoretical prediction. The sliding variable does not cross the manifold repeatedly with a large overshoot; instead, it exhibits the typical zigzag motion inside the boundary layer, which is characteristic of quasi-sliding-mode control.
The simulation waveforms of the inductor current show that the current waveform has a triangular ripple superimposed on a sinusoidal average. The ripple amplitude is determined by the inductor value and the switching frequency. With \(L = 1\) mH and \(f_{sw} = 25\) kHz, the current ripple is only a few percent of the average current, and it does not affect the output voltage quality. The capacitor voltage also has a small ripple, but the voltage across the load is a smooth sinusoid because the ripples from the two converters are out of phase and cancel mutually.
Effect of Parameter Variations on the Solar Inverter
To further evaluate the robustness of the Buck-Boost solar inverter, I ran additional simulations with \(\pm 20\%\) variations of the inductor \(L\), the capacitor \(C\), and the load resistance \(R\). The reference was fixed at 100 V/50 Hz. The results are summarized in Table 4.
| Parameter variation | Output voltage error | THD | Settling time |
|---|---|---|---|
| \(L\) +20% | 0.5% | 1.8% | 0.85 ms |
| \(L\) -20% | 0.7% | 2.2% | 0.75 ms |
| \(C\) +20% | 0.4% | 1.6% | 0.90 ms |
| \(C\) -20% | 0.8% | 2.5% | 0.70 ms |
| \(R\) +20% | 0.3% | 1.7% | 0.60 ms |
| \(R\) -20% | 0.9% | 2.0% | 0.80 ms |
The output voltage error remains below 1%, and the THD stays below 2.5% in all cases. These results demonstrate that the discrete sliding-mode controller provides a very high degree of robustness to component tolerances and load variations. The solar inverter can maintain an excellent output waveform even if the actual component values differ from the nominal design values.
Discussion
The Buck-Boost solar inverter presented in this article offers several advantages over conventional inverter topologies. First, it can perform both step-up and step-down DC-AC conversion with a single-stage power converter. This reduces the number of components and simplifies the system architecture. Second, the discrete sliding-mode controller gives the solar inverter a very fast transient response and strong robustness against disturbances. Third, the output voltage is generated directly by two controlled voltage sources, so the low-order harmonics are greatly suppressed without additional filtering.
There are also some limitations that should be mentioned. The Buck-Boost converter has a higher voltage stress across the switches compared to a simple buck inverter, because the output voltage is inverted and can be much larger than the input. The inductor current is also discontinuous in the sense that it is always positive in each converter, so the current stress is higher than in a bridge inverter. These stresses increase the conduction losses and the cost of the power devices. Nevertheless, for applications where a wide input voltage range is required, the Buck-Boost solar inverter remains an attractive solution.
Another practical issue is the need for an accurate DC bias reference in each converter. If the two DC biases are not perfectly equal, a residual DC component appears across the load, which is undesirable for most AC loads and transformers. In the proposed scheme, the DC bias is generated from the same input voltage, and the sliding-mode controller forces the average capacitor voltages to match the references. Thus, the residual DC voltage is negligible. In my simulations, the DC component of the load voltage was less than 10 mV.
Conclusion
I have presented a detailed investigation of a Buck-Boost solar inverter controlled by a discrete sliding-mode strategy. The solar inverter consists of two identical Buck-Boost converters connected in a differential configuration. Each converter generates a negative sinusoidal voltage superimposed on a negative DC bias. Because the two biases cancel in the load, the output voltage is a pure sinusoidal AC waveform with zero DC component. The discrete sliding-mode controller is designed using a linear combination of the voltage and current errors, and it is implemented with an exponential reaching law to guarantee a quasi-sliding-mode motion.
Simulations with Matlab/Simulink confirm that the proposed solar inverter can produce an output voltage that is either higher or lower than the input DC voltage, with a very low total harmonic distortion. The steady-state error is small, and the dynamic response to reference changes and load changes is excellent. The solar inverter recovers from a reference step in less than 1 ms, and a load step causes a voltage deviation of less than 1.5%. Sensitivity analysis further shows that the control system is robust to parameter variations. Therefore, the discrete sliding-mode controlled Buck-Boost solar inverter is a viable and high-performance solution for photovoltaic applications where wide input voltage ranges and high dynamic requirements are important.
In the future, I plan to extend this work by building a laboratory prototype and testing the discrete sliding-mode solar inverter under real photovoltaic panels. I will also investigate the grid-connected operation mode, where the solar inverter must be synchronized with the grid voltage and deliver a controlled current. The flexibility of the discrete sliding-mode controller makes these extensions straightforward, and the same control framework can be adapted to other power converter topologies.