The Solar Bucket - Lessons
The Solar Bucket experiment took place between Mar-08 and Jun-09, useable data was collected from Jul-08 to May-09. The objective was to see how a small off-grid solar electric system would behave in the south of England. The system consisted of a 4.7W amorphous silicon PV panel connected to a 3.2 AH battery driving a some form of resistive load. Follow this link for a description.
What were the lessons?
Clouds
The nature of the sky determined the output of the panel. Pure clear sky days are rare in England. The graph below shows the extent of the cloud cover over the duration of the project:
The cloud cover was obtained from metar reports from an airfield approximately 7km to the west. Generally, the airfield only reports low level cloud, medium and high level cloud not being reported, thus the number of pure clear sky days was less than that suggested by the graph above.
August 08, was a particularly grim month and this is reflected in the energy yield:
The box and whisker plot not only shows the seasonal variation in yield due to Sun-Earth geometry, but also the effect of clouds. The maximum values are close to the yield one would expect on a clear sky day, whilst the minimum values show the attenuating effect of clouds. The maximum values during winter are around 40-50% of those during the summer, however, clouds can reduce the yield to around 10-20% of the clear sky yield. The frequency and density of cloud cover is greater druing winter than summer, thus amplifying the seaonality of the yield.
Clouds not only affect the seasonality of the yield, but also the instaneous value of the output of the device. The graph below was not derived from the Solar Bucket, but similar behaviour was observed. It shows the instantanous yield take at 1 minute intervals around solar noon and the sky state prevailing at the time.
Under clear sky and overcast conditions, the output of the device is constant. Under intermiediate conditions, the instantaneous output varies considerably, some times dropping to less than 10% of the clear sky value.
Observation 8, is higher than the clear sky value. This is due to the effect of cloud fringes. As a fringe passes between the Sun and the Earth, it acts as diffuser and for an instant, there is an increase in irradiance (you can sometimes feel this on your face).
No direct data on the effect of rain, snow and fog was collected, but observation suggested that this weather conditions have a significant attenuating effect.
Modelling
As the instaneous output of a solar device can vary, the effect of clouds should be modelled as a distribution reflecting a range of values, rather than a constant. Towards the end of 2010, as simple solar radiadion measuring device was constructed. Observations of the output of this device taken whilst watching the sky, suggested that the under a pure clear sky, there was no attenuation and that an overcast sky could be model using minimum, modal and maximum values similar to thos showin in the sketch below. However, for intermediate sky states, the distribution of intantaneous values of attenuation was bimodal as shown below:
When there are few clouds, a solar device is exposed to more or less clear sky conditions, however, when a cloud becomes between it and the sun, the is a short period of attenuation, sometimes reducing irradiance to around 40% or less of the clear sky value. When the cloud is broken, the device only recievies the irradiance that passes through the cloud, experiencing only short periods of bright sunlight.
The sketch is an idealization of conditions under a simple sky consisting of a single layer of cloud. A complex sky conisting of two or more cloud layers can cause the irradiance to vary significantly and at the extent of the cloud increases, the behavior tends towards that of an overcast sky.
Buffering
A PV system sitting under an English sky is not viable without some form of buffering. During a typical day, the output of the system will fluctuate with as clouds pass between the sun and the panel. Clear days are rare (less than one in ten). Over a period of days, there will be occasions when little useful energy can be collected and the shortfall has to be be obtained either from an alternative source or a battery buffer. This is most acute in winter, but is also a factor that needs to be considered in summer.
Battery Managementanagement is critical to the success of an application. The comments below relate experience gained with lead acid batteries, but the same general principals apply to other types. Batteries are vulnerable to over-charging and over-discharging. This is complicated by the fact that they behave something like a capacitor, whilst the stabilised voltage of a battery provides useful information about its state of charge, at any other time it is dependent on the nature of the load, the charge current and its own internal state. A well managed battery should be able to survive for between three and five years. The solar bucket managed to destroy three in 15 months. The first was salvaged from a skip, the second was new and used in conjunction with an inappropriate controller and the last one was managed manually. Manual management worked well enough, however, the times when the charge drifted outside the operating limits were enough to damage it.
System Size
If a system is to be independent of back-up supplies, then it must be sized according to winter conditions. From the results, it would seem that winter output is approx. 25% of that in summer, thus there will be an excess of electricity in Summer. If a back-up is to be provided for the winter months, then the cost of this must be seen in the context of total installation. For example, does providing a petrol/diesel driven generator cost more than the solar panels required to meet the demand. A lot of roadside and railway solar installations are paired with a small wind turbine, wind being more plentiful in the winter. In the case of wind turbines mounted next to railways, they appear to recover energy from the train's slipstream.
PV Panel Operation
This was painless, once on its mounting it functioned perfectly without need for maintenance.
Panel Mounting
A PV panel is a large heavy object. It's mounting must be stable in high wind conditions and sufficiently rigid to prevent damage from bending.
During the period of operation, the panel was mounted at an angle of approximately 60 degrees to the horizontal, this being thought to be the optimum. It is clear that south facing panels mounted at an angle similar to that of a household roof maximises the output when the sky is clear or contains only a few clouds. However, experiments conducted after the Solar Bucket had been dismantled, suggested that mounting the panel close to the horizontal maximised the output when the sky was overcast as it frequencly is in winter.
Mountings which track the sun seem like a good idea, but however, they only have a marked impact on yield during the summer months and marginal, if any advantage during the winter.
Whilst not an issue with the small panel used for the Solar Bucket, the cost of electrical and fire safety need to be taken into account. It may also be a legal requirement to mount the panels above ground or in an area of retricted access, e.g. on the roof.