Power optimizers: Centralized vs. distributed MPPT

September 15, 2009 -  In measuring the performance of a photovoltaic (PV) cell, it is common practice to assume typical conditions of irradiance, temperature, and cell parameters, and that these conditions are uniform across all cells in a PV array. However, there are many situations such as various forms of partial shading of the array that cause significant variations of these factors within an array or a single string of PV panels. The result is panel mismatch and lowered performance; in short, actual performance will deviate significantly from what would be expected.

From testing and field trial results collected at National Semiconductor Labs, and referencing other studies, panel mismatch resulting from shade or other factors can result in disproportionate power loss in solar panels, whereby as little as 10% of shading can lead to a 50% loss of energy harvested. Additionally, for crystalline silicon PV arrays, depending on the array connection, as little as 2.6% shading could lead to a total array loss of 16.7% (Figure 1) [1].

Figure 1. The difference in performance of arrays in shaded conditions.

Some commercial and residential installations fail to meet their full potential in energy output, and other projects are abandoned due to a less-than-ideal projection of energy output due to mismatch. Essentially, there is an inherent under-utilization of space and under-delivery of energy. For example, present commercial installations could be 10%-20% bigger on average and one can only hazard a guess as to the amount of energy that could be generated from "shaded out" installations. From a survey conducted by Greenberg Quinlan Rosner Research of 150 installers in the US in January 2009, installers acknowledged the problem as endemic, with as many as 54% stating that any shade on installations was unacceptable. Installers instead choose to "design around the problem," leading to an average cost increase of 19% [2].

This article will explain the phenomenon of panel mismatch and also analyze why small variations in cell parameters can affect the system-level performance of the PV array. Additionally, power optimizer technology will be examined as a solution to the panel mismatch problem, and the potential benefits of distributed maximum power point tracking (MPPT) enabled by power optimizers over centralized MPPT and traditional solutions.

Distributed vs. centralized MPPT

Power generated by a solar module is calculated by multiplying current (I) by voltage (V). At any given time under any given conditions, there exists one optimal point -- the maximum power point (MPP) -- where a module is generating the most power possible for those conditions. In other words, the single MPP of a PV module is a function of an exponential relationship between current and voltage. MPPT is an electronic form of tracking that utilizes algorithms and control circuits to search for this maximum energy point and thus allow a converter circuit to harvest the maximum power available from a PV module.

In cases where irradiation, temperature, and other cell parameters are uniform, there would be no difference between the performance of distributed MPPT and centralized MPPT besides conversion efficiency differences. However, where partial shading is present, the panel mismatch problem is at its greatest. Partial shading will result in an array having multiple MPPs from different panels because of non-uniform parameters. With a centralized MPPT, this can lead to additional disproportionate losses. This is for two reasons: First, the centralized MPPT becomes confused, stopping on a local maximum point and settling in a sub-optimal point of the voltage to power configuration; Second, the voltage point of the MPP can be very diverse due to irregular conditions, going beyond the scope and voltage range of the centralized MPPT. Because the variations between panels are significant, it is in these cases where the ability of power optimizers in distributed MPPT can enhance the performance of panels independently and boost performance.

PV arrays for residential, commercial, or utility installations are typically configured as shown in Figure 2, with a centralized inverter that not only converts solar energy from DC to grid-use AC, but also provides centralized MPPT. In this setup, multiple strings of PV panels are connected in parallel and they feed the input of a grid-tied inverter. The centralized inverter not only converts DC to AC power as a primary function, but also contains a MPPT controller which seeks to maximize the energy harvest through a MPPT algorithm from the PV array at all times by regulating its input impedance.

Figure 2. Grid-tied PV system with centralized MPPT.

PV arrays for residential, commercial, or utility installations are typically configured as shown in Fig. 2. In this setup, multiple strings of PV panels are connected in parallel and they feed the input of a grid-tied inverter. The centralized inverter not only converts DC to AC power as a primary function, but also contains an MPPT controller which seeks to maximize the energy harvest through an MPPT algorithm from the PV array at all times by regulating its input impedance.

In a PV array with power optimizer technology and distributed MPPT (Figure 3), a power optimizer unit is attached at each panel. Power optimizers have a dual track: on the one hand, they track the best localized MPP, and on the other, they translate the input voltage/current to a different output voltage/current to maximize the energy transport in the system. The power optimizers communicate with each other in an indirect manner. Optimizers are "cognitive" and self-organizing -- they sense the I & V environment and adjust themselves until a total string optimum is achieved, while simultaneously arriving at a local optimization point at the panel level. At present, only power optimizers are capable of doing so.

Figure 3. Grid-tied inverter with Power Optimizer distributed MPPT.

Power optimizers keep the time-proven series-parallel panel arrangement and improve it by distributing only the DC/DC and MPPT function to the panels. Meanwhile, the power optimizer architecture is perfectly compatible with existing multi-stage inverters and will actually allow them to run more efficiently because the bus voltage can be kept higher and more constant. Power optimizers are much more than just boosting DC/DC converters. They deal with extra energy as well as reduced energy. This means additional light from reflection that also causes panel mismatch is handled equally well as various forms of shading. Likewise, it means that power optimizers are capable of power changes caused by adding panels to a string (making that string generate more energy) or subtracting a panel or two from a string (thus reducing energy.) The power optimizer architecture enables the system to harvest the most energy available.

Power optimizers: Distributed MPP solutions

We have seen how panel mismatch caused by shade across the cells as well as other factors can lead to disproportionate losses of generation from the array. We also see that at present, installers have addressed the panel mismatch issue by avoiding the problem, such as designing around shade/not installing at all, or installing a smaller array, which leads to lowered energy output.

Bypass diodes in the junction box shorted across strings of cells and modules can nominally mitigate to a certain degree the effect of mismatch by diverting the current around shaded cells and thereby reducing the voltage losses through the module. However, this is an insufficient solution: all panels today are already equipped with bypass diodes, and although they prevent entire strings of panels from dropping out completely, we can see from the data that there are still sizeable disproportionate losses of energy harvested.

Conclusion

Power optimizer technology is available today from National Semiconductor; SolarMagic power optimizers offer the ability to maximize the energy extracted from every panel while maximizing the energy transfer in the PV system, recuperating up to 57% of energy lost to panel mismatch issues. An ideal solution performs MPP at the panel level; power optimizers address head-on the problems inherent with centralized systems by increasing total energy output by up to 37%, therefore mitigating successfully the panel mismatch problem.

Biography

Ralf Muenster holds a masters degree in physics from the Technical U. of Munich and served as a scientist at the UC Berkeley. He is director of the Renewable Energy Business Unit at National Semiconductor, 2900 Semiconductor Dr., P.O. Box 58090, Santa Clara, CA 95052; ph.: (408) 721-5000; e-mail SolarMagic@NSC.com.

References

[1] Installer Survey by Greenberg Quinlan Rosner Research, January 2009.

[2]N. Chaintreuil, F. Barruel, X. Le Pivert, H. Buttin, J. Merten. "Effects of shadow on a grid connected PV system," INES R.D.I., Laboratory for Solar Systems (L2S); 23rd European PV Energy Conference, 2008