Monitoring problems

Generating at flows lower than I have been able to generate at previously is creating problems with my monitoring systems.  For the moment, I haven't been able to quite understand why this should be, but a result of it is that I have removed the feed to the "Live Power" tab on this blog because it was giving wrong readings.

Having an interest in closely following the output of my installation, I monitor output in different ways. Some are convenient but not very accurate whilst others are inconvenient but more accurate.

The convenient, 'at-a-glance' system relies on a Wattson monitor with its feed going also, via a Wattson Anywhere hub, to the web. It is this system which is playing up: when the turbine is happily putting out 160 W, the monitor and web portal sometimes show zero.  As can be seen from the online trace below, it suddenly dropped to zero at 17.00 hours on Thursday even though the turbine was in fact OK:

As a backup to Wattson, I also monitor using the blue tooth output from the Sunny Boy inverter : SMA's Sunny Beam system.  This is not quite so much 'at-a-glance' and it is not quite as accurate, but it does have the benefit of being able to store several months of data.  The Wattson display can also store data but only about three weeks worth.  If you don't remember to download the data to your pc within that time, it starts overwriting and readings are lost.

The final system I use, which is the most accurate, is the kWh generation meter used for submitting readings for Feed in Tariff claims.  One can use this as a way of measuring power output by noting the exact moment in time when its decimal place updates and then doing a calculation based on the increment in kWh and the precise time interval to achieve this increment (measured in seconds).  I usually do readings over at least a 24 hour period and, of course, to be valid, there should have been no change in the turbine's output during this period.  Since multiple 24 hour readings give values for power which are within 0.5 W of each other, I think the method is accurate enough.

Unfortunately, this meter too is proving inaccurate at low flows - something which Elster, the manufacturer, makes clear will happen when the current flowing is below 0.5 A.

None of this matters very much except that it's introducing error into my on-going study of 'whole system' efficiency at low flow.  In truth, there are probably so many other inaccuracies, especially knowing precisely what the flow is, that this added error will probably not matter very much.