OzEA – The second story

The Oz-Energy-Analysis.org project continues to hum away in the background, building momentum. For those who don’t recall what OzEA is, read these two posts from earlier in the year on BNC:

OZ-ENERGY-ANALYSIS.ORG – open science for the new millennium

OzEA modelling – large-scale wind power using a bucket storage model and gas backup

After much necessary background work, including data collation, website construction, preliminary wheel kicking and a lot of hard thinking (!), we are moving onto some serious analysis and modelling. But scenarios need storylines to hang off. Our first story was about scoping the problem. The second story — reproduced below — is about understanding. This is an exploration framework rather than a real-world proposal. To me, with an extensive experience in working with biological systems, the evolutionary approach we take here appeals. See what you think.

Francis and I would appreciate your critical feedback, either in the comments below or on the relevant OzEA page. Please consider both sites. And remember, OzEA is an experiment, with the tea room being a portal into developments. We always welcome your feedback, on any aspect of the site and its outputs.

————–

The Second Story – Understanding the Problem

Introduction

In the beginning was The First Story, followed in recent months by round one development through the menu bar (data, analysis, models…). This story ushers in round two.

To briefly reintroduce OzEA: the big picture is a global need for much increased electricity production as we progress through this century. Much increased fossil fuel use to achieve this is problematic given that current human impact on the carbon cycle is widely believed to be impacting on climate. While nuclear power is an alternative to coal and gas, issues around Nuclear Power, or the science of Climate Change, are not discussed here. OzEA seeks to be a broad church; we put our energies into empirical, high level and open analysis of how a high penetration of renewable electricity might be achieved in the Australian context.

In this Second Story we adopt ‘50% by 2030’ renewable electricity as the basis for ongoing work into 2011. Demand management (smart grids) and system evolution are matters that will be central to the integration of renewables, and these are discussed in what follows. Work through to years end is to model the power output from large-scale scenarios of geographically distributed wind and solar power plants. This will provide a solid base for further rational analysis of renewable variability.

The fifty percent renewables by 2030 approach

Adopting 50% renewable penetration by 2030 as a baseline gives structure and coherence to our work plans. In reality Australia is scheduled to have around 20% renewable electricity in 2020 (predominately from wind), driven by the federal governments LRET scheme. The purpose of a 50% target is to drive analysis and thinking, rather than an engineering proposal.

While wind is currently the most mature and economical of the large scale renewable technologies, its variability will eventually make further deployment self limiting; more wind farms = more electricity when the wind blows => depressed prices in the wholesale market. In turn, electricity from solar power can become a more valuable renewable source. The key focus is thus to examineconfigurations of wind and solar that reduce variability and usefully match with demand. (Note: solar = large-scale concentrated solar thermal (CSP); we hold photovoltaics to the margin for now).

Working at the hour-to-hour level we use historical wind and solar data to model ~10 GW average of electricity supply from these sources. Combined with historical demand data, this allows calculation of a ‘demand remainder’ (demand minus renewable supply). The first, naive, approach is to supply this remainder by conventional generators (with a little support from available pumped storage hydro), and to assume that Smart Grid Demand Management does no more than smooth out sub-hour variability and keep demand peaks from growing above current levels.

The naivety above is to suppose demand data from the past can represent demand in the years to come. While past weather data is a good template for the future, the demand can and will change as the electricity system grows and evolves. Hence, the ‘demand remainder’ that we calculate will require a more thoughtful interpretation than simply power required from fossil fuel generators.

Supply and demand; transmission and distribution

These four components provide a template for understanding our electricity system. The transmission network is the backbone that connects region to region, state to state, connecting the power plants that supply electricity. This electricity is taken at substations and feed, at lower voltages, into distribution networks (the poles and wires on our residential streets). Around 25% of overall Demand is residential, with the commercial and industrial sectors making up the balance.

The market operators ensure that, with very high probability, the system remains in balance from second to second; i.e., that supply meets demand. While electricity can be stored economically in the form of pumped storage hydro, this capacity is limited and mostly demand is meet by ramping supply up and down as needed (see The Electricity System discussion).

Peak loads, especially driven by air-conditioner use, present a particular problem for the electricity system. While residential use is one quarter of demand as a blunt average, it is a much higher portion on hot afternoons. Distribution networks in particular can be pushed to their limits, and system planners are faced with the prospect of costly upgrades to these networks. Peaking loads create a real need for mechanisms that can curtail or shift demand, otherwise expensive upgrades are needed in order to provide a much higher network capacity – to a level that is only needed for a small fraction of the year.

Analysis of large-scale renewable integration is necessarily intertwined with peak demand and network development issues, as these pressures are driving system evolution now. From a renewables perspective, the pressure to manage extremes on the demand side crosses over with managing variation on the supply side. This point bears reading again.

Accounting the variability

Power from wind and solar can be very variable; sometimes these sources produce little if any power at all. This is an enormous impediment to making large investments in these renewable power sources. At the simplest level renewable energy can be accounted as free fuel. That is, the system continues to require the same number of coal and gas power plants as before, to cover the times when the wind isn’t blowing and the sun isn’t shining. The saving is on the fuel (and any associated emissions), however, the cost of the renewable infrastructure is much greater than the fuel saved.

Multiple wind and solar farms at different sites will to some extent smooth out the variability. A more involved reckoning of the supply capacity can be had by engaging in statistical calculations of ‘Capacity Credit’. This can be informative, but is only a rough cut at quantifying what is really of interest.

We explicitly model the electricity supply that given Wind and Solar scenarios would produce. This ‘renewable electricity’ time series can be examined in conjunction with the demand that existed over the same time period, and so give the ‘demand remainder’ on an hour-by-hour basis. Analysis of this demand remainder is superior because it empirically captures relationships between electricity demand and renewable supply (e.g. solar on hot days).

Development of a 50% renewables system can only occur as an evolution, and one that includes the demand patterns. Explicit modelling of renewable supply in the context of today’s demand profile shines light directly on the issues and opportunities that demand side evolution presents.

Smart grids and demand management – a necessary detour

As retail consumers it costs you or me the same to use our air conditioners (or heating) regardless of whether the wholesale price is $10 or $12,500 a MWh. Residential demand is disconnected from the supply market, except as a long-term average. This demand inelasticity is a problem crying out for solutions.

Enter stage left, Smart Grids and Smart Meters.

While these terms encompass various aspects, here we focus briefly on: (i) load control, and (ii) interval meters & Time of Use pricing; see the Demand Management discussion page for more extensive comments.

Through a ‘smart meter’, or perhaps simply via the internet, a control hub in your house can manage some appliances in an intelligent way. A pool pump, for example, would be off when the network was struggling. Water heating is the classic ‘off-peak’ appliance. More complicated, but essential, is a mechanism for the compressors (but not fans) of air-conditioners to be switched out for a few minutes when need be, and for the thermostat to ride modestly and intelligently across demand peaks.

Your motivation for smart operation of such appliances is simple; Time of Use metering. At peak times electricity will be more expensive; on a windy night it will be cheap. So called “Interval Metering” is a foundational functionality for a smart meter. While residential time-of-use pricing requires careful implementation, it should save you money if use at peak times is modest. What might be called the “Eco-Saver” electricity plan will allow you, essentially, to withdraw subsidy from those who are punishing the system at peak times by running four, perhaps inefficient, air conditioners flat out.

Smart grids and metering involve a world of detail at both the technical and policy levels. There is discussion and debate. In Victoria interval meters are being rolled out state-wide right now; in South Australia they are resisted. Digging into these issues became a distraction at OzEA, and for now we pull back to a watching brief. The key point is that development of technologies and interfaces for intelligent load control will lay the very foundations for further levels of demand side elasticity.

Big ideas: the ecology of energy and the variability gambit

Large, complex, efficient, systems are rarely imposed through a straightforward engineering plan, where the steps required are foreseen at the outset. The scale, efficiency and sophistication of our current fossil fuel based electricity system would seem fantastical to those who hauled coal in primitive mining operation at Ipswich or Collie a hundred years ago.

The variability problem can be engineered away with high levels of supply redundancy and proven but expensive or inefficient storage mechanisms. What can be done, and what responsible politicians, policy makers, board rooms and bankers, will do are two entirely different things. So far there is no ‘killer app’ on either the supply side (e.g. proven geothermal), or the demand side (e.g. cheap storage). But ‘killer apps’ can be weeds in an ecological context; evolution is not a one-step process nor is it fixed on only one possible outcome. Rather, many small steps act in concert to alter the very fabric of the system from which the next batch of little steps proceeds.

Starting with the system we have now, we ask: “What will happen as more renewable energy is included into the system?” (i.e. how might the system evolve, and what are the selective pressures that will induce change?)

With supply rendered less controllable by the addition of large-scale renewables, and with demand made more elastic in response to the cost of supply, the electricity market develops new niches for balancing supply and demand. Attention is too often focused on handling the occasional lean times (when the electricity price becomes high and dispatchable backup is required), when the real evolution will occur in the frequent plentiful times that come with large scale renewables; this presents enormous possibilities. With abundant electricity we can potentially displace more expensive transport fuels, and otherwise have wealth-producing industries and jobs spring up in the niches that a suitable energy ‘ecology’ (market) would provide.

Assuming we become a high penetration renewable country, to what extent will we look back in 30 or 50 years and see the value of a flexible and frequently abundant system outweighs the costs of maintaining ‘backup’ to cover the gaps? Thinking about this question requires looking past the next immediate roadblock.

The idea here, what we call the Variability Gambit, is to postulate that in time the variability problem is soluble, especially with a deepening of the electricity market and associated integration with the energy sector more generally.

The monster under the bed – how much will it cost?

At the simplest level (straight cost per MWh of electricity produced) the rule of thumb is wind power at twice the cost of coal power, while CSP is around four times as expensive — some forward estimates are more generous. Wind turbines are a mature technology and so the costs here can only be expected to reduce on a modest schedule (maybe a few percent a year), while the less-refined CSP might yet undergo stronger improvements as increased deployment occurs. A tax on carbon emissions would add to the scales, so roughly and at this basic level, costs are seen to be an uphill journey, but a gradual rather than a hopeless one.

The cost and engineering of large-scale renewable plant must include any associated transmission infrastructure. Further, the variability, and consequent need for storage or backup, introduces additional costs that make the task of an economic reconciliation more difficult again. Today’s renewable technologies, placed within todays systems, are not cost competitive as a fit-for-service means of replacing coal and gas.

Consider, as a thought experiment, imposing large scale renewables on the Australian system NOW, at the same time decommissioning our coal power assets and limiting the use of gas turbines (perhaps through a very high carbon price). Broader economic damage and electoral backlashes aside, lucrative opportunities would arise because of extreme variations in the wholesale electricity price. Storage of electricity using hydrogen or compressed air (as examples) would become profitable. Demand management technologies would develop rapidly. Much innovation would occur. After some decades of expensive electricity the system would again evolve into a form with cheap and plentiful electricity.

The question is, can we achieve much the same ends (more gradually) without draconian impositions and economic carnage? Forging that path is the task at hand, and the supply variation of renewables may itself be our most potent tool.

Open Science and the web-site

Doing Open Science (not just talking about it) is a parallel purpose of the OzEA project. In the beginning we imagined lots of community involvement in doing the Science, and now have more nuanced expectations. Certainly many valuable comments have been made, including a handful of really substantive contributions. We look forward to more of these as we knuckle down into 2011. Yet, this is not a blog, and we do not seek comment for the sake of comment, nor provide an arena for generic argument. Rather, the commenting system is largely a virtual lab-book that is open to all; it is a major part of our record keeping. And of course we continue to welcome critical comment, encouragement, focused questions, and the sharing of knowledge and experience.

Breadth first is the approach to analysis we take here, and so some of our analysis is not as sophisticated or refined as the more specialised work of others. What matters is whether an analysis is sufficient for the purposes it is put to. We welcome comparison and criticism in this regard, and are always grateful for nudges and prods into the issues and complications that careful work needs to take into consideration.

Concluding remarks

To mindfully anticipate the future electricity system is not straightforward. The basic difficulty in looking ahead multiple decades is that while some aspects are reasonably predictable, any number of less likely, and even improbable, technological and sociological developments could have significant impacts – if they come to pass. And some of these unlikely events will occur (play enough hands of poker and you will get a royal flush).

Moving into Round Two of data, analysis and modelling, we focus on the variability of supply that comes with high penetration renewables (wind and solar). While capturing the supply variability is a lot of work, it is also a relatively straightforward number crunching exercise. The real significance will be in the ‘demand remainder’, as so many of us seek to explore the implications, opportunities and consequences of increasing the level of renewable supply into the Australian electricity market.

A derisive term, “The Fake Fire Brigade”, has arisen to describe those seen as too optimistic or woolly in their claims for large-scale renewables. Here at OzEA we take a positivistic view; however, we are nobodies’ fire brigade. The wool-free version is simple enough: into the medium term at least, an Australian electricity system with an increasing penetration of renewables will continue to be underpinned with significant (fossil) fuelled supply, while demand side evolution will provide a more elastic response to supply variability. The rate of renewable rollout will be limited by real world costs, and driven by government support.

The importance of demand management to renewable integration is at once tenuous and profound. At the tenuous end, the ability to make modest adjustments to demand, especially in high load situations, provides some assistance with a generation mix that includes renewables. But it does not provide much help when there is little wind for several days in a system reliant on wind power. At the profound end are the pathways opened up for electricity system evolution in the decades to come as devices, houses, industries, suburbs and states interact dynamically with supply.

Implementation of smart grids must be undertaken with due care and forethought. It is easy to speculate about electric (or plug-in hybrid) cars; it is easy to note the long-term sense in houses being intelligently designed for space heating and cooling. It is not hard to see wind and solar power being integrated into the broader energy sector, perhaps via Hydrogen production. While all these points remain vague or speculative, it is simple deduction that building to high penetrations of wind and solar power will involve these sorts of developments.

The question, in the end, is this: can we intelligently and responsibly nurture the necessary evolution in the way our electricity system works? The next step to coherently addressing this question is a solid quantitative grip on the supply variability. As we work this through, it is our goal and commitment to communicate the analysis and its interpretation in as open and useful a way as we can.