So “what actually happens during G59 commissioning” is a question I was asked earlier this week. It came up, as we had been doing exactly that, deep in the bowels of Kings Cross Station.
Engineering Recommendation G59 is published by the Electricity Networks Association (ENA) and has the snappy title, Recommendations for the connection of generating plant to the distribution systems of licensed distribution network operators (DNOs). G59 sets out guidance for connecting larger embedded generators onto the network and is used as the template to manage the design and testing of the connection interface of bigger plants, such as that at Kings Cross.
A G59 relay is a device that controls a contactor (a large relay), that sits between the PV system and the connection to the grid. In many ways, the G59 relay mimics what the inverter does; it sits monitoring the grid voltage and frequency - and if it detects a fault, it de-energises the contactor – disconnecting the PV system from the grid.
G59 commissioning is the term often used to describe the process of testing and commissioning the G59 relay itself, though the actual process usually involves a whole sequence of inspection and test procedures across the wider PV system. Many firms employ an outside contractor to do the tests on the G59 relay – at Sundog we are fortunate that we have the experience to do it ourselves. At Kings Cross, this was a two stage process: day-one, doing preliminary tests; and day-two, doing formal witnessed tests in front of an engineer from the DNO and client.
To test the G59 system we disconnect the normal “sense” feed to the G59 relay and instead temporarily connect it to a device which can simulate grid faults – injecting variations in voltage and frequency and ensuring that the G59 correctly identifies these variations. Not only do we carefully measure the voltage and frequency values at which the G59 relay trips – we also measure the amount of time it takes to react and respond. The exact settings programmed into the G59 relay are described in G59 itself, but are also confirmed (and in some cases tweaked) by the DNO.
We normally start with the voltage tests, the sequence of carefully going through these can take a while – as we have two-stage under and over voltage settings to check on each of the three phases (that’s 12 tests). We usually repeat each test three times, including checking to ensure that the trip voltage is correct and that the time response is as expected.
Once the voltage tests are completed, we then move onto the frequency tests. These are normally quicker as we don’t need to do separate tests on each phase. Once this is complete we normally finish the injection tests by doing a vector shift test – forcing a step change in phase angle across the phases and making sure that the relay identifies and reacts to the step.
Now the above only describes part of the process – there is a whole lot more that goes on, broadly summarised as the DNO checking they are happy with the design and connection arrangements. We got a clean bill of health at Kings Cross – and as I type, the system is at last pumping solar electricity into the grid.