Xuemei Cheng, Wesley Dose, Thomas Hempel, Øyvind Storesund Hetland, Kine Solbakken and Björn Veit at NorRen Summer School 2014
IntroductionIn recent years there has been a growing emphasis on harnessing energy from renewable sources. Photovoltaic (PV) cells have demonstrated great potential as a cheap and reliable source of clean, renewable energy, and PV production facilities are now being deployed at a high pace globally. The output from a PV production system naturally varies with the cycles of the sun (e.g., day and night), requiring other sources of energy to be available when sunlight is scarce or absent. There is another, less well known, issue in PV systems however: the variability of PV output even while the sun shining, due to second- and minute- resolution disturbances in the light absorbed by the solar panel. These fluctuations are usually due to clouds, but any moving object that occludes the path of sunlight will contribute. As more and more small scale PV systems are connected to the local grid, these fluctuations may prove a problem through grid voltage instabilities and temporary grid overproduction due to sharp local peaks in production. A possible remedy to these issues could be to somehow store energy when the system experiences a short spike in production, and then release this energy again when a drop in production occurs. This process would result in a smoother output from the PV system, with no, or minimal, loss in overall production.
How could we implement such an ‘energy buffer’ in a real-life application? The most common alternatives are batteries and capacitors. A battery can offer a high energy capacity (Wh) but low power (acceptance/delivery time, W), while the capacitor offers very high power, but can only store a small amount of energy. Another device has recently joined the market however: the electrochemical capacitor (also known as supercapacitors or ultracapacitors). The supercapacitor fills the intermediate region between batteries and capacitors. It could be considered as a high-capacity capacitor, and it shows considerable promise in storing seconds worth of energy from PV output, even from larger PV systems, while also providing the high rate energy storage/release required.
The problem – Strong fluctuations in the solar radiationFigure 1 shows the real measured solar radiation in Chemnitz, Germany in the year 2001. The time increment between the measured values is 1 second. Over this time period a total of 28,947,592 data pairs were collected. The odd area at the baseline between day 50 and day 110 is the result of lighting from a construction site overnight.
|Figure 1: Measured solar radiation|
Figure 2 shows the typical waveform of the solar radiation of a (summer-) day.
|Measured solar radiation of one day|
In addition to the expected daily variation in PV output, it can also be seen in Figures 1 and 2 that changes in weather conditions, especially clouds, leads to large fluctuations in the output on very short timescales.
SupercapacitorsA supercapacitor stores energy electrostatically on the surface of two conducting plates. Figure 3 shows a schematic construction of a cylindrical capacitor.
Figure 3: Schematic construction of a cylindrical capacitor
One of the characteristics of supercapacitors is the very large surface area of the conducting plates and the very thin separation layers between them. The result is a very high power density, which means that the storage device is able to absorb and release energy very quickly, within seconds. This feature will likely make the supercapacitor very suitable for solving the problem of fluctuating output from PV, compared to a battery which needs between 1-10 hours to charge. A limitation of the supercapacitor is the energy density. Compared to a battery, a supercapacitor can store only 10% of the energy a battery can store per unit volume.
Another issue related to supercapacitors is that the cost is currently very high. Actual prices on a cell level are in the range of 0.12 NOK/F to 0.16 NOK/F (0.015 €/F to 0.02 €/F). The price largely depends on the manufacturer and on the quantity of ordered cells. However, over the last few years the price has decreased quickly as production volumes for commercialized variants of the supercapacitor has increased.
Simulation results of PV-only and a PV- supercapacitor systemIn order to better assess the usability of supercapacitors for PV output stabilization, some circuit simulations were conducted based on the measured PV output shown above.
In the simulation schematics (Figure 4 and 5) the PV-system was modelled with a self-developed power source as a 1 kWPeak solar system with a rated voltage of 100 V and a rated current of 10 A. For a better understanding of the "real problem" the load was simulated as an ohmic resistance. In real applications there is in most cases a DC to AC converter that is coupled to the public grid.
|Figure 4: Circuit without supercapacitors, used for simulation.|
The supercapacitor module in Figure 5 was modelled with the usual equivalent circuit diagram that consists of a capacitance C, a serial resistance RESR and a parallel resistance RP. The module was designed to include a series connection of 40 cells (Maxwell BCAP1200 ) with the features listed in Table 1.
Serial resistance, RESR
Maximum operating current
Short circuit current
-40°C .. +65°C
Usable specific power
Based on the physical fundamentals of capacitors the string/module parameters can be calculated to the values shown in Table 2.
Serial resistance, RESR
Maximum operating current
|Figure 5: Circuit with supercapacitors, used for simulation. The simulation was based on 40 supercapacitors connected in series, each with a capacitance of 1,200F.|
|Figure 6a: Simulation results. Red line shows a smoother VM2-voltage
than the gray-blue line, as a result of the added supercapacitors.|
|Figure 6b: Zoom of the simulation results shown in Figure 6. Red line shows a smoother VM2-voltage than the blue line, as a result of the added supercapacitor|
Figures 6a and 6b shows the comparison between a supercapacitor coupled system and a PV-system without supercapacitors. In this case, the supercapacitor is shown to reduce the problematic voltage peaks from ~123 V to ~98 V (reduction of approx. -20%).
Generally, it can be said that the higher the fluctuations, the higher the benefit gained by using supercapacitors.
ConclusionThrough the simulations we have seen that supercapacitors are well suited for smoothing the highly fluctuating output from PV systems, both by smoothing the output and removing sharp voltage peaks. From the results presented above it is reasonable to believe that adding a larger number of supercapacitors will result in an even smoother power output. In situations where short bursts of overproduction leads to disconnection from the local grid, adding supercapacitors will likely increase the energy efficiency of the PV. Instead of disconnection, the peaks will be removed and delivered to the grid at a time when the production is lower. Also, in the case of systems in which PVs are connected directly to a battery, the battery is expected to benefit from a smoother PV output by way of more constant charging rates, which should increase the lifetime of the storage device. However, adding supercapacitors to the PV system adds a significant cost due to the current price for supercapacitors. At the moment, it is less likely that there will be any economic benefits from adding the supercapacitors to the system unless the cost is significantly reduced.
References Maxwell Technologies, Datasheet K2 Ultracapacitors - 2.7V Series, BCAP1200,