By Edward Colby, Sentec
Much of the media focus on smart appliances considers their performance as demand response resources, either through dynamic response or through demand response. There is a limited population of these appliances in the market, but many smart grid technology companies have been considering the impact of larger numbers particularly when placed alongside electric vehicles (EVs), which from the grid’s perspective behave like large, power-hungry appliances with built-in storage capacity.
Challenges exist. Many will be felt at the local network level, and many in the industry expect this is where the solutions to these challenges will be developed.
For example, demand response can do little to help with localized hotspots. Localized grids can experience stresses where many power-hungry appliances, including EVs, are clustered. Some are obvious; some more subtle.
- Extra load: EVs, in particular, represent a huge rise in the amount of power being drawn through certain parts of the network. An EV is like another house being added to the grid.
- Clustering: This equipment will tend to be geographically concentrated, leading to problem hotspots.
- Two-way flow: Existing low-voltage networks are not engineered to cope with high levels of two-way flow, a problem exacerbated by clustering.
- Invisible imbalances between phases on the substation: When substations are first built, the loads they serve are balanced between the phases. Adding more homes, EVs and power-hungry appliances in an unplanned fashion can stress one phase and cause unexpected failures.
It is hard to estimate the degree of stress local grids are under. They are robustly engineered, and there is some debate about the level of market penetration necessary for EVs or microgeneration to have a perceptible impact. But smart appliances as currently configured are not a likely solution. Appliances using dynamic or demand response are responding to inputs from much larger-scale systems: the overall grid frequency or the tariff set for a large customer segment. This functionality will do little to reduce the stress at the local level, which might occur at a different time.
Occasionally it might create additional problems because simple demand response has a tendency to create a bow wave of demand where the peak consumption that was shaved is simply shifted to another moment in time. This might reduce pressure on generating capacity for the whole grid but has unpredictable effects at a neighborhood level. This problem is exacerbated by a tendency for neighbors to exhibit similar buying patterns. The implication of this is that certain areas will see penetrations of EVs or microgeneration well in advance of there being an overall impact on the grid.
These issues suggest that demand response could be adapted initially to ease pressure on grid hotspots, a solution that would require some degree of monitoring at substation level. This is not a situation that can be dealt with using smart meter data alone. Smart meters cannot measure the phase effects, and the meters cannot fully reconcile the two-way flows of energy at a neighborhood level.
Using accurate monitoring at the substation level, demand response could be adjusted for local network conditions—up to and including appliances’ negotiating a re-entry slot for their restart with other appliances—to avoid the bow wave effect discussed.
Network topologies—notably the difference between 50Hz and 60HzS—will require different architectures for such a scheme. Either way, ubiquitous monitoring of the local distribution grid has value distinct from the monitoring enabled by smart meters. By moving control downward to a local level, it allows demand response, local generation and storage to be shoehorned into local networks when it is needed and without costly upgrades.
Edward Colby is chief technical officer of Sentec, a smart grid specialist and product development company. Find out more at http://sentec.co.uk.
PowerGrid International Articles Archives
View Power Generation Articles on PennEnergy.com