Transmission/distribution interaction

In order to manage voltage, frequency and reactive power, and meet the other require- ments of supply, National Grid has a view of the electricity network that resolves down to around 8 km (5 miles). It has to take into effect not only the load and demand on its own network but, as we have seen, the cumulative effects of changes in the distribution network.

This varies as load switches between consumer, industry and commercial users. Aggregate demand (the total from consumers in a region) is, of course, equally variable at the low-voltage level. The effects of the weather are well known, but are changing over time. DNOs know, for example, that, even if the temperature is relatively pleasant, if there is rain and wind at the time people are travelling home from work they will tend to switch on additional heating when they arrive. If a sunny day clouds over, lighting demand will rise because of apparent darkness, although the ambient light levels are still high.

These changes are managed at the DNO level, where, of course, fault levels, voltage, frequency and reactive power have also to be managed. But, at the moment, the level of management is relatively light because the system is largely unidirectional – from the National Grid feed-in points from the transmission system through the medium voltage used by industry and commerce and then the low-voltage domestic network. There are few points at which DG feeds in and no trading between different parts of the network.

Elsewhere the situation has begun to develop somewhat differently, and a major reason has been the introduction of DG. Thanks to the feed-in tariffs that guarantee

118 Local energy

export, German networks are required to accept all the power generated by wind turbines, solar photovoltaics and other renewable-energy systems. Similarly, Danish networks are required to accept wind power and electricity from CHP plants.

This will ultimately make it necessary to operate distribution networks in a much more active way, more closely allied to the way in which the transmission network is managed. As this happens it will also require the interface between the distribution and transmission networks to be carefully managed. The assumptions previously used by the transmission operator to track and predict demand and supply at the distribution network may no longer be valid.

This is not an insoluble problem but in countries such as Germany and Denmark it is one where the need to address it is moving rapidly up the agenda.

In November 2006 grid operators on the German border cut power to a transmis- sion line that passed over a river to allow a large ship to pass underneath on its way from a shipyard to the sea. The transmission line was an interconnector – a line that joins the grids of two countries. The line had been depowered many times before for very similar reasons, but in this case the result was a local blackout that triggered blackouts centring on Germany and France and lasting only a couple of hours, but whose effects were felt far further afield and for far longer.

The problem was traced to a lack of information passing between the grid oper- ators in the two countries and poor operating practices. It highlighted well-known inadequacies in the extent of interconnection between European countries, especially in areas such as this, where there were large cross-border flows.

But, in its report on the incident, the Union for the Coordination of Transmission of Energy (UCTE) noted that lack of information between distribution and transmission network operators had made it more difficult for operators to bring the system back on line quickly.

The report said:

The requirements for disconnection of generation units connected to the distribution grid (especially wind generation and CHP) are usually less strict than for the units connected to the transmission grid, i.e. they are disconnected at a smaller frequency deviation. When the frequency deviation reaches the threshold values of the units’ protection, they are automatically disconnected from the grid.

This was the case when the blackout happened: distribution units tripped when the frequency dropped below set limits. This worsened the situation in one of the blackout areas. The report adds,

Recovering the frequency to its nominal value required an increase of generation output in the Western area and a decrease of generation output in the North-Eastern area. After a few minutes, wind farms were automatically reconnected to the grid, being out of the TSOs’ [transmission system operators’] control. This unexpected reconnection had a very negative impact, preventing the dispatchers in both areas from managing the situation.

Additionally, certain TSOs in the North-Eastern area were not able to reduce the power output from generation connected to the transmission and distribution grid in a sufficiently short time necessary for the frequency restoration.

These are examples of insufficient TSO control over the generation behavior. The TSO control usually applies to generation connected to the transmission grid since traditionally the generation connected to the distribution grids has not had a significant impact on the

Interacting with the electricity grid 119 power system as a whole. However, the recent rapid development of dispersed generation, mainly wind farms, has changed the situation dramatically. The wind generation in some areas significantly influences the operation of the power system due to its high share in the generation and intermittent behavior dependent on weather conditions.

In its recommendations, UCTE pointed out that

most TSOs do not have available real-time data on the power generated in the distribution grids. In view of the rapidly growing share of such generation, this has multi-dimensional consequences:

• no real-time knowledge of the total national balance between supply and demand, • no real-time knowledge of the generation started in DSO [distribution system operator]

grids and possible tripping/reconnection in case of a frequency or voltage drop, • no real-time knowledge of generation started in DSO grids and possible impact on grid

congestion in the high voltage grid.

It also pointed out that at present the TSOs have no control over distribution-level generation, and said this could lead to ‘serious power balance problems especially in over-frequency areas’. In response, it made three recommendations that would give transmission system operators far more knowledge of, and control over, generation connected to the distribution network.

• The regulatory or legal framework should be changed so that TSOs can assert control over generation output (allowing them to change schedules, or to start and stop the units).

• TSOs should receive data on a per-minute basis on the generators connected to the distribution system.

• Generation units connected to the distribution grid should have the same require- ments, in terms of behaviour during frequency and voltage variations, as units connected to the transmission network.

Any such recommendation would likely impose requirements appropriate to the scale of differently sized DG. The effect of connecting or disconnecting a domestic system is very different from that of an industrial generator inputting tens of megawatts into the grid. However, as we have seen on the demand side, the cumulative effect of aggregating large numbers of similar systems should not be overlooked.

Developing control and oversight that provide enough management capability for the distribution network, without overspecifying the generator and making it unnecessarily costly, is a balance that will have to be struck.