Time to smarten up: smart grids on the horizon9 December 2013
As energy demand soars and the need for a more stable and flexible power grid becomes increasingly pressing, the stage seems well set to usher in the era of intelligent network management. Dr Matthias Muscholl, smart grid pilot technical leader at Alstom, talks to James Lawson about the lessons being learned for the transmission of tomorrow’s power.
Intelligent energy network management using smart grid technology offers potentially huge energy savings, greater reliability and a massively increased ability to incorporate renewable energy generation. The smart grid market in 35 emerging economies is set to reach a value of $66 billion by 2023.
With countries such as China (Honeywell's pilot with State Grid Corp) and Brazil (AES Eletropaulo) leading the way, smart grid initiatives are growing steadily in the BRICS nations and beyond. South America in particular has seen a major expansion in the last two years: Ecuador is aiming for full smart meter deployment by 2017, Colombia and Peru are finalising roadmaps, and Chile is developing smart city pilots.
But, though investment in greenfield projects gives the BRICS nations the opportunity to incorporate smart grid technology from the outset, established markets such as Europe and North America are where much of the research on how tomorrow's smart grids will work in practice is taking place.
This is the goal of Nice Grid, the largest of six demonstration projects within Europe's Grid4EU initiative. Besides the technical impediments, there are also regulatory, organisational and social hurdles to clear.
"We need to build something that we can deliver in other areas," says Dr Matthias Muscholl, smart grid pilot technical leader at Alstom. "Other countries are aiming for this type of system and they want to learn from Nice Grid."
Nice Grid has three main objectives: shifting up to 5MW of load; handling the grid voltage issues caused by large-scale distributed photovoltaic (PV) generation; and testing 'islanding' - where a network area operates completely independently from the rest of the grid.
Situated in Carros in the far south-east of France, the test area is fed by two primary substations and has weak connectivity to the national transmission grid. It also has a large number of PV connections at domestic and industrial premises. Using a 'Linky' smart meter that can control up to eight devices remotely, up to 1,500 residential and business customers will take part. A customer in Nice Grid might generate electricity from PV panels, store power in batteries and be able to turn off heating or other electrical devices in order to reduce demand.
An energy aggregator gathers together microgenerators like this and operates them as a unified resource. These portfolios, often described as virtual power plants (VPPs), supply the grid and also help to ensure grid reliability through flexible demand management.
Take a load off
Batteries enable further load shifting and shedding. Nice Grid will test several types of lithium-ion cells at three distribution network levels. A 1.1MW battery at the Carros primary substation supports load reduction in peak-demand periods, while five 100kW batteries at five medium-voltage (MV) and low-voltage (LV) distribution substations help control peak PV generation, manage peak-demand periods and allow for island mode operation.
Load-shedding in low-demand periods is facilitated by 100 3kW batteries at customer homes. MV/LV controllable transformers also help to support the distribution grid by varying the voltage of electricity injected from network batteries and PVs. Larger batteries will be added early in 2014.
With the first field demonstrations taking place in December 2013, load reduction is the first goal. "We are aiming to switch off 5MW of total demand," says Muscholl. "That will be significant enough to be visible to the transmission system operator (TSO). We want to confirm what they see on the transmission grid."
Of this figure, 1.8MW will come from industry, 1.5MW will come from network and substation batteries, and the remainder will come from consumer demand management.
How to organise the different smart grid players has been the main challenge so far. The deregulated European electricity market means that electricity generators, district system operators (DSOs), TSOs, metering operators and retail power suppliers/aggregators are usually (but not always) separate companies, rather than part of one national utility as in the US.
As the system operator, the TSO is traditionally responsible for balancing grid supply and load in line with wholesale market rules. It plans a day ahead, optimising the mix of generating stations and reserve providers for each market trading period based on predicted demand. Moving to a smart grid environment means increased complexity and far more options to manage, so all parties must work in harmony to handle shifts in supply or demand.
"Due to the different layers, it becomes very complex to manage the grid balance," says Muscholl. "The TSO may wish to reduce load while the DSO wants to increase it."
A centralised network energy management system (NEMS) is the answer. Manufactured by Alstom Grid and managed by the DSO (see "Nice Grid" schematic, opposite), the NEMS takes in forecast data (PV generation, predicted consumption) as well as data on generation, demand management and storage flexibility available from the various participants (generator, DSO, TSO and retailer). It then uses this information to optimise generation, demand management and storage in real time across an entire district.
"You need to understand the power flow," says Muscholl, "where power is injected and where it is taken out."
The retail aggregator forecasts the load and passes that data to the NEMS, which in turn works out the network constraints for the TSO and DSO, and passes them that data as an input to their own grid-management operations. If consumer demand needs to be adjusted, the NEMS tells the aggregator so that it can schedule sufficient flexibility within its customer base.
With conflicting interests in the different management layers, a market structure has emerged as the best way to handle this flexibility.
"The DSO might want to take control at the consumer's house and switch off car-battery charging when it needs to adjust demand, but the aggregator sees them as its customers," says Muscholl. "It doesn't want to affect their comfort by switching off the heating. So we do not control demand directly but ask the aggregator for a positive or a negative change."
Diffused local load reduction - letting consumers decide what can and cannot be switched off remotely - offers a solution to this problem; the more flexible they are, the better the financial incentives.
"The supplier signs agreements with customers to allow a certain number of annual disconnections," continues Muscholl. "Via phone, text or a button on the meter, they can also choose not to be disconnected on a certain day."
Like supply and demand bidding in today's wholesale balancing markets, retail aggregators will bid to offer a certain amount of flexibility.
"Though it might only be one battery in one location, the cost must be taken into account and we model it as a market requirement," says Muscholl. "That makes it an optimisation problem, and should give maximum flexibility for the lowest price."
Dena, the German grid study, has recommended similar changes: shared responsibility for system security (frequency and voltage) between TSO and DSOs that also takes in aggregation of forecasting and providing system services like balancing reserve, reactive power and redispatch capabilities.
"Nice Grid is giving us excellent experience in how the system can be structured," Muscholl says. "The retail aggregator becomes the controller of demand management, with the DSO acting as the market facilitator that helps allocate flexibility."
The DSO itself can call upon this flexibility to maintain voltage in the district grid; the TSO can call for it at a higher level to balance the transmission grid. Offering flexibility will be mandatory for aggregators, who will be paid for the reserve capacity and the amount of demand management.
"The onus is on the aggregators to provide flexibility, but so far we have not applied penalties for not carrying out their promises. It is done on a 'best effort' basis," says Muscholl. "Penalties may come in future though."
A further refinement to this flexibility market is to create a network battery aggregator that manages the flexibility provided by batteries at substations within the DSO network.
"There should be no competition between DSO batteries and domestic batteries," says Muscholl. "It's also vital that market rules take account of the realities and constraints inherent in long-term, day-ahead and intra-day planning. For example, the DSO should be allowed to declare constrained grid conditions that take precedence over the open market as part of the long-term and day-ahead planning process."
Other Nice Grid results include the concept of a commercial location (CL) - local grid areas that can be treated as homogenous for planning purposes. More detailed recommendations on the format the smart grid market will take, such as CL size, bid format and market periods, will come at the end of the project in 2015.
By building on the knowledge gained from Nice Grid and the wider Grid4EU initiative, the next generation of smart grid projects will mark a step change for power management, moving smart grids from a technical possibility to a working reality.