‘The problem with energy,’ says the earnest energy department civil servant across the café table from me, ‘is entropy. Actually, three problems. Where the energy comes from is a problem. Where it goes to when we’re done with it is a problem. And the process of using it is a problem. Those three things will always be true. They will always cost us in money, resources and consequences.’ Then she adds, looking nervously about her: ‘But saying so is a shortcut to a short career. People want to hear that they can have unlimited free supplies, without spoiling their view, while saving the planet at the same time... juggling things to try and square that circle is mostly computerised data analysis.’
The analysis to which she refers is not just application of a known set of techniques to a static set of criteria. Not only is the landscape changing as technologies and economics shift, but the analysis itself invariably informs change in both methodologies and approaches that in turn alter the criteria on which the analysis was based. This is not unique to energy issues, nor to the present time, but it does become more acute when (as here) both policy and innovation respond to pressure in an area where margins have become the main area for advance. As Tom Tietenberg commented in a survey of energy efficiency effectiveness this year, ‘policy makers must recognise an expanded set of barriers and respond with some ingenuity in applying an expanded set of available instruments’.1
Where energy comes from is, one way or another, to be blunt, the environment – and the environment is not only an emotive subject, but an increasingly urgent focus of scientific attention. Energy acquisition also interpenetrates with other urgent supply concerns: food production, for instance, has become an energy intensive activity whose outputs are in their turn coming under pressure from biofuels demand. Where it goes after we’ve done with it is, once again, the environment – usually as heat, at a time when we are worrying about environmental warming. And how we use it decides how much we need to extract, and how much of it we return, and the ratio of that utilised to that wasted.
For much of history, energy sources were direct. A fire might be wasteful, but it delivered the heat of combustion direct from the fuel to the point of application; the fuel might have to be collected, but generally not from great distances. A windmill might have great frictional losses, but it converted wind energy directly into movement of grinding stones against the wheat. A ship might be slow, and at the mercy of the weather, but the wind in its sails converted directly to its movement.
The industrial revolution brought a new approach: energy became a commodity to be transported and stored, passing through multiple conversion stages with loss or consumption at each. Coal or oil or gas must be dug from the ground, transported, and burned. Where heat was not the desired output, it converted water to steam whose pressure was in turn converted to movement. With the arrival of electricity, which has made flexible modern information societies possible, the movement was still not the output. Energy must be converted again, the resulting current incurring further losses as it transferred from generator to point of consumption, where it was converted again – often back into movement and heat.
The heat rate efficiency of a steam-driven generation plant could in principle approach 40 per cent, and has done so in experimental demonstration, but real figures are closer to about 30 per cent. That is, approximately 3kW hours of heat are used to produce 1kW hour of electricity2. Gas turbines do considerably better, in some cases approaching 60 per cent through reuse of steam and chimney heat, but all of that heat is released, one way or another, into the environment of the generator. The electricity undergoes losses in the region of seven per cent in transit3, the result being further heat wasted into the environment en route. Use at the far end ultimately returns all of the remaining electricity to the environment as heat through further wasteful conversions. Not taken into account in there are the energy costs of transporting the generator turbine fuel, manufacturing and maintaining the generator and power lines, and so on. The energy cost of energy is astonishing.
Nor, of course, is the energy loss the only cost; sulphur compounds, though now largely tamed in the developed economies, have caused a lot of grief and carbon emissions remain of pressing concern. However, there is a link: if fuel burned per unit of energy delivered can be driven down in the interest of efficiency, carbon emission will also fall. Reducing fuel usage also opens up possibilities of reduced operating cost, offering an incentive to adopt more efficient methods.
With the possible exception of solar energy, where heat from the sun is intercepted and made to work for its living before going into the environment in which it would have ended up anyway, there is no costless energy source at the point of generation. Even solar methods require energy to produce the transducers, installations and means of distribution. Other methods such as wind, hydro, tidal and geothermal power, for instance, do dispense with fuels, but still involve production investment and extract energy from a finite system. Easy economies have long ago been made; further progress lies in management of margins and the careful economic introduction of alternative generation technologies. And that is where the computerised data analysis comes in.
Laser radar scanners generate the data for cloud point metrology in the design of a wind turbine blade (images supplied by Metris).
One approach is to reduce transmission losses through ‘microgeneration’ of power at or close to the point of consumption, thereby reducing call on the distribution grid and feeding excess back to it. A promising example is the integrated fuel cell boiler that sits in the home and, according to solid oxide cell producer CFCL4, currently offers electrical efficiency rates of 60 per cent – with no subsequent transmission loss.
Solid oxide fuel cells are a quiet, stable, highly efficient, low cost and low emission technology. UK-based Ceres Power, already well along the road to their use in partnership with British Gas and companies such as Calor, bases its strategy on ‘grid parallel’ combined heat and power units with output at a level, which aims to cover the base demand of the homes within which they are installed. Power demand above this level is taken from the grid, excess production over demand fed back to it.
The development of such devices for reliable, long-term operation at the limit of achievable efficiencies calls for very carefully optimised placing in relation to actual demand levels: too much output or too little, and the incentive would be lost. It also requires careful design, Six Sigma-level quality assurance, and sophisticated control, all of which are dependent on computerised data analytic approaches. Ceres’ performance and control manager, Mark Selby, refers to the rôle of MathWorks tools both in analysis of thermodynamic performance, system temperatures or fuel flows and in consequent programming of embedded control systems to deliver ‘an efficient economic/carbon performance’. Component development is distributed among the engineers concerned, but consistency and effectiveness dictate that they do so within a single environment, ‘using ideas from the overview to drive quality’ as Selby describes it. Drawing from a variety of data input sources, Matlab analysis results feed into common data objects, the building and testing of virtual components under a range of applicable régimes, their assembly into virtual systems, and final autocode for a real world product.
The final result of all this is a fuel cell stack delivering 1kW packaged in with a boiler. Exact benefits will vary with usage, conditions and circumstances, but the indicative expectation is for amortisation of additional cost through fuel savings over the first five years, during which seven and a half tonnes of carbon will be saved.
Risk and decision analysis drive Enex planning for a geothermal energy site.
Such investment in process is part of the overhead cost in any research and development cycle and, in a general sense, part of the cost base for newly-developed approaches to energy management. In the course of its own work, Ceres is developing internal methods and expertise in the analysis of transient data experience which have a wider application.
Microgeneration can, in principle, come from almost any source – and at differentiated scales, though not all are practical or economic. At the individual premises level, wind power has received widespread publicity, but stands up less well than the fuel cell boiler on both cost and benefit grounds since generation is variable, intermittent, and impossible to predict reliably however good the analysis. Water-driven generators fare better, but are dependent on location. Nuclear power for the single premises is perfectly possible, but poses obvious risks and costs which make it unrealistic in most settings, although a few specialist contexts continue to explore ways to reduce them through data analytic management structures of ever increasing refinement.
Another approach to financial, carbon and energy cost reduction is to look for synergies through linkage of energy production to other infrastructure functions. Waste disposal is a prime example, with incineration or anaerobic digestion both seeing increasing attention as ways to simultaneously generate useful energy and remove the energy, financial and environmental costs of byproducts. Recycling to reduce consumption is equally important, providing lower cost high level raw material to replace primary extraction. Just as energy was, until recent times, procured by skimming off the easiest fraction, so it was cheaper to throw rubbish away than to make use of it. This is changing, but once again the calculations are marginal and change continually, requiring continual reanalysis. My helpful civil servant is not the only one to be cagey about discussing details of energy policy and use, almost everyone I talked to was reluctant to be quoted and usually because their supply decisions are commercially sensitive. Two industrial concerns making heavy use of recycled materials (one metal, one glass) were running in-house analytic operations (one in Matlab again, one in Statistica) to oversee the fine tuning of procurement decisions against energy cost on a daily basis. Another had the analysis contracted out to a university department.
Other projects are much longer term, particularly those involving extracting energy from new or underutilised sources. Enex, well placed for the sector as an Icelandic company, has a substantial background of expertise in geothermal power – but any new development is a heavy investment gamble on site-specific conditions. Costs and production returns depend upon drilling depths, brine reservoir enthalpy, and other circumstances which can be estimated within defined regions of likelihood, but cannot be established with certainty until work actually starts. Drilling wells that turn out not to meet production need is an expensive business, as is making plant purchase decisions too late or too early. Enex analyst Viktor Thorisson describes how combined use of risk and decision analysis software led them to ‘untraditional’ conclusions about decision management for an upcoming European project. A site is abandoned if two successive well drills fail to yield suitable conditions for production, and plant is ordered before drilling commences.
From raw materials through MatLab to packaged Ceres fuel cell heat and power unit.
Moving from the specific to the strategic, analytic approaches inform national and international efforts to encourage moves towards globally energy efficient approaches. The Argonne National Laboratory’s Center for Energy, Environmental, and Economic Systems Analysis (CEEESA) is one of the providers worldwide that facilitate analysis-centred efforts to evaluate and restructure energy systems and policies. Working with power companies, governments, and a variety of agencies including the world bank, in numerous countries, it provides capabilities for analysing complete energy systems and their environmental impacts, simulation of future evolution and the likely resultant responses of participants, and a model for examining competitive or regulated electricity markets.
Electricity, though crucial, is not the only form in which energy is distributed. In particular, the industrialised world is deeply dependent on oil, and geopolitical realities are disproportionately influenced by availability and price of fuels for transport. Alternatives to petroleum are keenly scrutinised even by corporations with a vested interest in the status quo. So-called hybrid vehicles make fuel go further, but do not remove dependence on it. Hydrogen fuel cells hold promise for the future.
Biofuels have their champions, but many, as mentioned above, including the favourite corn, sugar and palm-based options, compete with already scarce food agriculture. And agriculture itself has become very energy intensive, not to mention heavily dependent on petroleum-based products. The new hot topic alternative this year is conversion of algae-derived lipids to hydrocarbons, which can be used as direct petroleum replacements, with news in July that ExxonMobil are committing serious R&D investment to the idea. Algae can be grown in places not otherwise usable for agriculture (such as urban sewage reservoirs5), and projected yields are around eight times those for corn. One research centre on the US west coast has a contract to conduct intensive data analytic examination of the possibility that eutrophic algal bloom on inland water courses might be cleared in cost- and energy-neutral synergy between environmental benefit and fuel production. None of this addresses the atmospheric carbon implications of burning oils in the first place, but it does suggest possibilities for somewhat reducing the energy overhead involved in transporting those oils over large distances.
‘It’s all very promising,’ comments my civil servant, as she examines her empty coffee cup, ‘but there’s a lot of research to be done just to get to the point of running to stand still. Have you any idea how much energy went into making this cup of coffee?’
For a full list of sources and references cited in this article, please visit www.scientific-computing.com/features/referencesdec09.php
