The Hidden Losses That Decide Whether Solar Power and Grid Protection Actually Work

Every watt of electricity that flows from a rooftop solar panel to the grid, or from a substation into a neighborhood, passes through a gauntlet of semiconductor switches. Each time one of those transistors flips on or off, a tiny amount of energy is sacrificed as heat. Multiply that by thousands of switching events per second and millions of devices across a national grid, and the arithmetic becomes staggering. A team of researchers from the University of Central Lancashire Cyprus, Frederick University, Brunel University London, and Cyprus University of Technology has now published a detailed loss evaluation and performance modelling study in Frontiers in Electronics that dissects exactly where these hidden losses live, how accurately they can be predicted, and what their true cost is when power electronics are pushed into two of the most demanding roles in modern energy systems: fault current management and renewable energy integration.

The study, led by Stelios Ioannou and colleagues, tackles two complementary applications. The first is a solid-state Fault Current Limiting and Interrupting Device, or FCLID, built around a Switched Capacitor circuit. The second is a grid-connected photovoltaic system with power smoothing capability, designed to keep solar output stable even when clouds suddenly blot out the sun. What makes the work remarkable is not just the engineering ambition but the forensic precision with which every conduction loss, every switching transient, and every reverse recovery event is accounted for, simulated, and cross-checked against theory, datasheets, and experimental literature.

The FCLID device is a clever piece of circuit design. Under normal grid operation, the switched capacitor circuit performs power factor correction, improving the efficiency of power delivery to end users. But the moment a fault occurs, such as a short circuit, the same circuit limits the fault current to safe levels and facilitates interruption, acting as a solid-state replacement for the mechanical breakers and current-limiting reactors that have protected distribution networks for decades. The researchers demonstrate that by carefully selecting the inductance and capacitance values, the device can also correct voltage sags of up to 15.3 percent, restoring the load voltage to within 2.05 percent of its nominal value. According to standards such as BS7671, IEC 61557-1, CBEMA, and EN 50549, this level of compensation is significant because it extends allowable disconnection times from milliseconds to seconds, preventing nuisance tripping of protective devices.

One of the most technically intriguing findings concerns the phase relationship between current and voltage in switched capacitor circuits. Unlike typical DC-DC converters, where current and voltage waveforms are generally in phase, the switched capacitor topology introduces a 90-degree phase shift. This distinction is crucial because it directly impacts the modelling and thermal performance of the switching elements. The researchers pay particular attention to this phase relationship under a 50 Hz fundamental system frequency and a switching frequency of 5 kHz, using it as a precursor to modelling and calculating Insulated Gate Bipolar Transistor losses with the PSIM Thermal Module.

The PSIM Thermal Module is an add-on to the PSIM simulation program that offers a streamlined method for estimating losses in semiconductor devices, including diodes, IGBTs, and MOSFETs, as well as core and winding losses in inductors. The module incorporates a device database that not only includes characteristics detailed in manufacturer datasheets but also models the heat transfer path from the transistor junction through the package and case to the heatsink. The researchers used this tool to simulate the FCLID system with the IGBT model IXGT40N60C2D1, calculating conduction losses of 8.62 W and switching losses of 0.268 W per transistor through simulation, compared with theoretical values of 8.77 W and 0.213 W respectively. The difference between simulated and analytical loss results showed agreement within 1 percent, validating the modelling approach.

The second part of the study investigates a grid-connected photovoltaic system with power smoothing capability. Solar irradiance fluctuates due to variable cloud coverage, which is inherently unpredictable. Sudden reductions in irradiance lead to corresponding drops in power generation, and these fluctuations propagate through the system, ultimately affecting the power delivered to the grid. The researchers simulated a sudden drop in irradiance from 1000 W/m² to 200 W/m² for approximately 2 seconds. Prior to the disturbance, the system delivered 2.566 kW to the grid. When the irradiance dropped, the PV array contributed only 473.9 W, while a lithium-ion battery connected via a bidirectional DC-DC converter provided an additional 2.092 kW, ensuring continuity and stability in power delivery. After a brief transient period of approximately 180 milliseconds, both the inverter input current and the power delivered to the grid recovered to values close to their original steady-state conditions, demonstrating compliance with CBEMA standards.

The loss analysis of the complete PV system reveals that losses are primarily concentrated in switching transistors, diodes, and inductors. For the standalone PV system, the researchers measured a system efficiency of 95.3 percent, with the PV system generating 2658 W and 2558 W delivered to the grid. Applying the described loss calculation methodology, including the inductor’s DC resistance and summing all individual losses, yielded total system losses of 135 W, resulting in a 5 percent error and a slight efficiency reduction to 95 percent. For the PV system integrated with the compensating circuit, the efficiency was 94.2 percent, with losses of 174 W, and the analytical approach estimated 183 W, resulting in a 4 percent error. The inverter alone accounted for 93.5 W of losses in the standalone scenario, with MOSFET conduction losses of 67 W dominating.

A particularly important finding concerns the relationship between switching frequency and system efficiency. The researchers compared four MOSFETs, including silicon carbide and gallium nitride devices, at switching frequencies of 2 kHz and 8 kHz. At 2 kHz, the MOSFET with the lowest on-resistance achieved the highest efficiency of 95.85 percent. At 8 kHz, efficiencies dropped across the board, with the same device achieving 91.74 percent. The findings indicate that lower switching frequencies consistently yield overall system efficiencies above 96 percent, irrespective of whether MOSFETs or IGBTs are employed. However, the study also reveals that reverse recovery losses become negligible compared to conduction losses only at low switching frequencies below 10 kHz and low current slew rates below 100 A/µs. At higher slew rates, the reverse recovery charge can increase by a factor of four, making these losses impossible to ignore.

Perhaps the most sobering result of the entire study concerns the gap between simulation and reality. The researchers found that practical implementation factors can increase total power losses by up to 21 percent. For the switched capacitor application, the inclusion of practical considerations increased power losses by 20.45 percent, showing that transistor internal resistances cannot be considered constant and that reverse recovery losses cannot be ignored without investigation. For the PV grid-connected system, practical considerations increased total system losses by 5.6 percent, with the inverter alone increasing by 11.31 percent. The researchers attribute this to the fact that the on-resistance of MOSFETs varies significantly with gate-source voltage, and that the assumed constant resistance of 3.3 mΩ for the IRF3805 introduces an error between 5 and 27 percent, with an average deviation of approximately 18 percent.

The implications of this work extend far beyond the laboratory. As the world races to decarbonize electricity grids and integrate ever-larger volumes of renewable energy, the efficiency of power electronic converters becomes a primary determinant of both performance and economic viability. Every percentage point of efficiency lost in a solar inverter or a fault current limiter translates directly into wasted energy, increased cooling requirements, and higher operational costs. The researchers demonstrate that simulation tools like PSIM, when properly calibrated with datasheet parameters and experimental data, can predict losses within 1 to 2 percent of analytical calculations, providing engineers with a powerful tool for designing the next generation of grid infrastructure. But the study also serves as a warning: the clean numbers of simulation must always be tempered by the messy realities of device behaviour, temperature effects, and reverse recovery phenomena that only reveal themselves when circuits are pushed into real-world operating conditions.

Subject of Research: Power Electronics

Article Title: Loss evaluation and performance modelling of power electronics for fault management and renewable energy integration

Article References: Ioannou, S., Polycarpou, A., Christofides, N., Chrysostomou, M., Darwish, M., & Marouchos, C. C. (2025). Loss evaluation and performance modelling of power electronics for fault management and renewable energy integration. Frontiers in Electronics, 6, Article 1675666. https://doi.org/10.3389/felec.2025.1675666

Image Credits: AI Generated

DOI: 10.3389/felec.2025.1675666

Keywords: power electronics, fault current limiter, renewable energy integration, semiconductor losses, PSIM thermal module, grid-connected PV systems, switched capacitor circuits, energy efficiency, Loss, evaluation, performance, modelling