Wind Power and the Smart Grid



Electric power companies complain about wind power because it’s intermittent: if suddenly the wind stops, they have to bring in other sources of power.

This is no big deal if we only use a little wind. Across the US, wind now supplies 4% of electric power; even in Germany it’s just 8%. The problem starts if we use a lot of wind. If we’re not careful, we’ll need big fossil-fuel-powered electric plants when the wind stops. And these need to be turned on, ready to pick up the slack at a moment’s notice!

So, a few years ago Xcel Energy, which supplies much of Colorado’s power, ran ads opposing a proposal that it use renewable sources for 10% of its power.

But now things have changed. Now Xcel gets about 15% of their power from wind, on average. And sometimes this spikes to much more!

What made the difference?

Every few seconds, hundreds of turbines measure the wind speed. Every 5 minutes, they send this data to high-performance computers 100 miles away at the National Center for Atmospheric Research in Boulder. NCAR crunches these numbers along with data from weather satellites, weather stations, and other wind farms – and creates highly accurate wind power forecasts.

With better prediction, Xcel can do a better job of shutting down idling backup plants on days when they’re not needed. Last year was a breakthrough year – better forecasts saved Xcel nearly as much money as they had in the three previous years combined.

It’s all part of the emerging smart grid—an intelligent network that someday will include appliances and electric cars. With a good smart grid, we could set our washing machine to run when power is cheap. Maybe electric cars could store solar power in the day, use it to power neighborhoods when electricity demand peaks in the evening – then recharge their batteries using wind power in the early morning hours. And so on.

References

I would love if it the Network Theory project could ever grow to the point of helping design the smart grid. So far we are doing much more ‘foundational’ work on control theory, along with a more applied project on predicting El Niños. I’ll talk about both of these soon! But I have big hopes and dreams, so I want to keep learning more about power grids and the like.

Here are two nice references:

• Kevin Bullis, Smart wind and solar power, from 10 breakthrough technologies, Technology Review, 23 April 2014.

• Keith Parks, Yih-Huei Wan, Gerry Wiener and Yubao Liu, Wind energy forecasting: a collaboration of the National Center for Atmospheric Research (NCAR) and Xcel Energy.

The first is fun and easy to read. The second has more technical details. It describes the software used (the picture on top of this article shows a bit of this), and also some of the underlying math and physics. Let me quote a bit:

High-resolution Mesoscale Ensemble Prediction Model (EPM)

It is known that atmospheric processes are chaotic in nature. This implies that even small errors in the model initial conditions combined with the imperfections inherent in the NWP model formulations, such as truncation errors and approximations in model dynamics and physics, can lead to a wind forecast with large errors for certain weather regimes. Thus, probabilistic wind prediction approaches are necessary for guiding wind power applications. Ensemble prediction is at present a practical approach for producing such probabilistic predictions. An innovative mesoscale Ensemble Real-Time Four Dimensional Data Assimilation (E-RTFDDA) and forecasting system that was developed at NCAR was used as the basis for incorporating this ensemble prediction capability into the Xcel forecasting system.

Ensemble prediction means that instead of a single weather forecast, we generate a probability distribution on the set of weather forecasts. The paper has references explaining this in more detail.

We had a nice discussion of wind power and the smart grid over on G+. Among other things, John Despujols mentioned the role of ‘smart inverters’ in enhancing grid stability:

Smart solar inverters smooth out voltage fluctuations for grid stability, DigiKey article library.

A solar inverter converts the variable direct current output of a photovoltaic solar panel into alternating current usable by the electric grid. There’s a lot of math involved here—click the link for a Wikipedia summary. But solar inverters are getting smarter.

Wild fluctuations

While the solar inverter has long been the essential link between the photovoltaic panel and the electricity distribution network and converting DC to AC, its role is expanding due to the massive growth in solar energy generation. Utility companies and grid operators have become increasingly concerned about managing what can potentially be wildly fluctuating levels of energy produced by the huge (and still growing) number of grid-connected solar systems, whether they are rooftop systems or utility-scale solar farms. Intermittent production due to cloud cover or temporary faults has the potential to destabilize the grid. In addition, grid operators are struggling to plan ahead due to lack of accurate data on production from these systems as well as on true energy consumption.

In large-scale facilities, virtually all output is fed to the national grid or micro-grid, and is typically well monitored. At the rooftop level, although individually small, collectively the amount of energy produced has a significant potential. California estimated it has more than 150,000 residential rooftop grid-connected solar systems with a potential to generate 2.7 MW.

However, while in some systems all the solar energy generated is fed to the grid and not accessible to the producer, others allow energy generated to be used immediately by the producer, with only the excess fed to the grid. In the latter case, smart meters may only measure the net output for billing purposes. In many cases, information on production and consumption, supplied by smart meters to utility companies, may not be available to the grid operators.

Getting smarter

The solution according to industry experts is the smart inverter. Every inverter, whether at panel level or megawatt-scale, has a role to play in grid stability. Traditional inverters have, for safety reasons, become controllable, so that they can be disconnected from the grid at any sign of grid instability. It has been reported that sudden, widespread disconnects can exacerbate grid instability rather than help settle it.

Smart inverters, however, provide a greater degree of control and have been designed to help maintain grid stability. One trend in this area is to use synchrophasor measurements to detect and identify a grid instability event, rather than conventional ‘perturb-and-observe’ methods. The aim is to distinguish between a true island condition and a voltage or frequency disturbance which may benefit from additional power generation by the inverter rather than a disconnect.

Smart inverters can change the power factor. They can input or receive reactive power to manage voltage and power fluctuations, driving voltage up or down depending on immediate requirements. Adaptive volts-amps reactive (VAR) compensation techniques could enable ‘self-healing’ on the grid.

Two-way communications between smart inverter and smart grid not only allow fundamental data on production to be transmitted to the grid operator on a timely basis, but upstream data on voltage and current can help the smart inverter adjust its operation to improve power quality, regulate voltage, and improve grid stability without compromising safety. There are considerable challenges still to overcome in terms of agreeing and evolving national and international technical standards, but this topic is not covered here.

The benefits of the smart inverter over traditional devices have been recognized in Germany, Europe’s largest solar energy producer, where an initiative is underway to convert all solar energy producers’ inverters to smart inverters. Although the cost of smart inverters is slightly higher than traditional systems, the advantages gained in grid balancing and accurate data for planning purposes are considered worthwhile. Key features of smart inverters required by German national standards include power ramping and volt/VAR control, which directly influence improved grid stability.

11 Responses to Wind Power and the Smart Grid

  1. nad says:

    In many cases, information on production and consumption, supplied by smart meters to utility companies, may not be available to the grid operators.

    The solution according to industry experts is the smart inverter.

    ??? Sofar as I have understood what is described here as a “smart inverter” doesn’t make the problem go away that grid operators don’t get enough information from utility companies. It seems smart inverters might rather circumvent that “information gap”.

    The “smart inverter” seems to be a device, which controls the efficiency of a panel, apparently via methods like:

    The aim is to distinguish between a true island condition and a voltage or frequency disturbance which may benefit from additional power generation by the inverter rather than a disconnect.

    Googling “true island condiction” revealed in my google bubble unfortunately only holiday suggestions under palms so I can’t comment on the methods, but I have also some general questions.

    It is of course somewhat necessary to switch off or lower currency if there is an overload (like within the grid). Basically every electrical device has therefore a kind of overload protection like a fuse (and apart from this it is good to have also a redidual current device if there is leakage), and it may be wise to lower feed into the grid instead of switching it simply off like with a fuse, but I have some doubts that a local device has much chance to react “smart” on greater grid instabilites, apart from the local ones (i.e. local overload), alone the computational necessities may be too big. So such a control would then to be made rather directly from the grid operator. Depending on the number of devices this might get even with good computational power rather very difficult. Moreover if I understood correctly the inverters seem to be hooked directly to the panels (?) wouldn’t it be better to make that overload control at the point of grid connection? Like what if you would use your local excess energy fo fuel a battery? But may be I have misunderstood here something.

    Two-way communications between smart inverter and smart grid not only allow fundamental data on production to be transmitted to the grid operator on a timely basis, but upstream data on voltage and current can help the smart inverter adjust its operation to improve power quality

    The benefits of the smart inverter over traditional devices have been recognized in Germany, Europe’s largest solar energy producer, where an initiative is underway to convert all solar energy producers’ inverters to smart inverters.

    Unfortunately given the current envisions for a taxation/fees on electricity from solar here in Germany I get the rather unpleasant feeling that the communication of loads for tariffs might be one of the major reasons for the communication between “smart inverter” and grid operator.

    The current envisions for taxation are a bit like the ones in Oklahoma. I.e. it is thought that you should pay a fee for your solar generated energy, even if you don’t feed in. This is of course not very encouraging for local storage. And it is also not very encouraging for installing solar power. In particular the taxation is not on home/land owners who could but do not harvest at least some share of solar energy (where this depends of course also on use).

    I do think though that that the cost for solar storage have to be taken into account, intermittency may be a problem if too large (similarily a very unflexible base load like from nuclear, which can’t be switched off or ramped down immediately might pose problems as well) and the higher costs for electricity might be somewhat due to the fact that the storage of intermittent power has become more difficult (especially with those old grids), but the new taxation scheme seems to be quite counterproductive in adressing this problem, since it is discouraging local efforts for storage. Moreover in Germany it seems that the soaring “prices for renewables” are quite due to industry subventions – but may be I have overseen something.

    By the way I had recently a discussion with an expert from a solar company who said that especially the efficiencies of the second method of power-to-gasare currently getting quite increasingly better. This especially interesting for Berlin, since the city currently rethinks it’s gas network.

    John wrote:

    With a good smart grid, we could set our washing machine to run when power is cheap. Maybe electric cars could store solar power in the day, use it to power neighborhoods when electricity demand peaks in the evening – then recharge their batteries using wind power in the early morning hours.

    So in short that local storage solution which is here proposed (energy storage in cars) seems not encouraged by the new german tariffs.

    And apart from the fact that this solution doesnt work for car commuters (unless this is done at your point of work like in industry….) – it is here in this context also very important to note who sets the prices. If this is done on a global market with all sorts of speculators roaming about and other difficulties then the actual prices and the real demand/supply might be very different. Like as we know food is thrown away in some places and peope die of hunger at other places.

  2. domenico says:

    It is interesting, and I am thinking that there is a simple method to optimize these grids: a little scale model (it is possible to use many models that work in parallel) with wind generator, a fan with real data regulation, (solar panels and lamps?) and a smart network with genetic algorithm to learn the optimal choices (battery size,inverter, etc.); so that it is possible to avoid black out and to modify the grids.

  3. lee bloomquist says:

    John, I recall talking with a researcher on the smart grid at NIST some years ago. He was interested in controllers that talk to each other locally and adapt to power variability without the overhead of centralized control. For an example of embedded controllers talking to each using internet protocols, I told him about some actual products offered by one company that I knew of.

    The control language for their embedded controllers derives from Petri nets.

    http://www.nist.gov/smartgrid/

    http://support.ctc-control.com/index.php?option=com_content&view=article&id=189

    • John Baez says:

      That’s cool. It would take me a long time to build up the necessary “oomph” to do anything useful in this field, but it’s a good long-term goal.

      • lee bloomquist says:

        Maybe for future reference then, each embedded controller running its own Petri net is really a player in a game. The game is to create systems from players, specifically a “distributed” system, where parts are independent in the sense of being distributed, and where by virtue of being in the same system, the state of one element of the system can carry information about another part of the system.

        In a workshop at Stanford on business applications of his work on situation theory, Jon Barwise gave an example that always stuck in my mind. He was writing on an overhead projector about this concept of a distributed system. We were looking at Jon and reading the screen on which his diagramming was projected. Israel asked him for an example, and Jon stepped in-between the overhead projector and the projection screen. He had broken the system. The state of the projection no longer carried any information about the state of the sheet upon which he was writing on the overhead projector.

        By breaking the beam of light from the overhead projector to the projection screen, Jon had broken an “information channel.” Obviously.

        So an information channel is a distributed system, which is how it can carry information.

        It’s a restatement of the above problem statement. To avoid the exponential explosion of centralized control, local players have to, instead, pick up information from other members of the systems of which they are members. Their states are related states by virtue of both being in the same system, even though distributed differently from each other.

        Players with no ability for joining a distributed system lose funding.

  4. Nathan Urban says:

    In addition to NCAR, the U.S. Department of Energy national labs also do a lot of work in the smart grid area. Argonne has a project on wind power forecasting, and Los Alamos has a smart grid group (see their seminars, an old project, and a conference they ran).

  5. Fluctuations in regional wind energy follows a Rayleigh distribution, which is one of the most basic Maximum Entropy formulations. Put that together with a constant integrated kinetic energy, and you have to figure that the wind is blowing somewhere. Put that knowledge together with a smart grid and you might see the start of spatial load balancing.

    Perhaps easier stated than done because of transmission losses.

  6. nad says:

    With better prediction, Xcel can do a better job of shutting down idling backup plants on days when they’re not needed. Last year was a breakthrough year – better forecasts saved Xcel nearly as much money as they had in the three previous years combined.

    By the way how much does Xcel pay for this NCAR service?

  7. […] the way, Professor and colleague John Carlos Baez did an Azimuth Project post on “Wind power and the smart grid” a bit […]

  8. As long as operators like the grids don’t claim “proprietary information” regarding their instantaneous loads, or generators don’t similarly deny such information flow, it seems to me that there are plenty of people out their with control systems savvy who would be interested in coming into the system and providing guidance on load balancing and shedding.

    I know there are always fears of instability in such macroscale control systems, but I and my colleagues at Akamai Technologies deal with loads and shifting demands and emergent behavior on such a macroscale system every day, the Internet. It can be done.

You can use HTML in your comments. You can also use LaTeX, like this: $latex E = m c^2 $. The word 'latex' comes right after the first dollar sign, with a space after it.

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 2,845 other followers