by Detlev Heinemann, Jürgen Parisi and Hans-Peter Waldl

Renewable energies will be an important pillar of the future energy supply. However, the availability of these energy sources depends significantly on the climate and weather. For efficient utilisation, it is therefore necessary to provide methods and information to describe the influence of meteorological variables on energy production. Energy meteorology takes on these tasks.

The energy supply companies in northern Germany are currently experiencing the increasing importance of renewable energies in the electricity supply: The peak output from wind energy in their supply areas exceeded the minimum value of electricity consumption for the first time in 1997. In comparison, the utilisation of solar energy to generate electricity with solar cells (photovoltaics) is still in its infancy. However, its potential is estimated to be significantly higher in the long term, so that a development comparable to the current wind energy situation will also begin here.

The problems associated with the fluctuating energy sources of wind and solar radiation in controlling the conventional power plant fleet and regulating the grid make it clear that the introduction of new technologies requires considerable efforts in various areas. In addition to basic research into technical development, these include the development of market launch strategies and - particularly important in the case of renewable energies - integration into existing technical and economic structures.

In this case, the availability of information plays a key role, as it allows an exact estimate to be made of the resources available to renewable energy systems from wind and solar radiation (these determine the energy yields) and the interrelationships that shape the system behaviour, such as the spatial variability of these variables (these determine the efficiency). At the same time, it is obvious that a future energy supply with significant shares of renewable energies will be highly dependent on the fluctuating meteorological variables of wind speed and solar radiation. A recent Shell study expects a global share of these energy sources of over 50 % in the middle of the next century. This suggests the desire to describe the temporal and spatial characteristics of wind and solar radiation and the behaviour of the energy systems based on them, and to use this knowledge to achieve more efficient utilisation of these energy sources. Some examples from the Oldenburg research field of energy meteorology are presented below.

An essential prerequisite for the economic utilisation of wind energy is detailed knowledge of the expected energy yields. These depend on the local and regional climatological conditions, which are described in extensive regional studies of wind energy potential and in analyses of individual wind turbine sites.

If the terrain under consideration is simple in terms of flow physics, as is the case in most areas of northern Germany, for example, established standard methods (e.g. the European wind atlas method) can be used to determine the potential. However, wind energy utilisation is becoming increasingly important in inland areas, especially in the low mountain ranges. Here, the influence of the sometimes highly indented terrain on the wind flow must be taken into account. Even across Europe, regions with high wind energy potential are often mountainous. For potential sites of this type, tools must be developed that allow modelling of the flow conditions even in complex terrain. These are usually mesoscale meteorological models, comparable to the forecast models of the weather services, which are adapted to the special requirements of the small-scale description of the wind flow. Calculations of this type typically have a spatial resolution of 100 metres. In contrast to the above-mentioned simpler methods for flat or only slightly hilly terrain, the application of which is now offered as a service by engineering offices, the use of these mesoscale models requires a research environment that can only be found in universities or larger institutes, both with regard to their further development and their application for determining potential.

The practical application of these models provides an objective basis for the selection of sites for wind turbines. This avoids the selection of unsuitable sites, reduces the risk for wind turbine operators and increases the average yields from wind energy utilisation.

Once a site has been found, e.g. for a planned wind farm, the next task is the selection of the arrangement of the turbines on the given area. Here, shading effects must be taken into account, which are caused by the influence of the individual turbines on the flow in a wind farm. In the Oldenburg computer model FLaP (Farm Layout Program), the flow conditions in the farm are simulated for this purpose and the installation geometry of the wind turbines is then determined using modern optimisation algorithms (evolution strategies and genetic algorithms).

For a long time, solar radiation was a variable of climatological interest at best. The density of the measurement network is correspondingly low and the temporal resolution of the data is inadequate. This situation has changed in recent times due to the great interest of both climate research (solar radiation is the main factor driving the Earth's energy balance) and solar energy in precise information on the temporal and spatial structure of irradiation.

At the same time as interest in this data has grown, meteorological satellites have established themselves as "all-round" measuring instruments and have provided valuable findings, particularly for determining the radiation balance in the atmosphere. The method of obtaining usable irradiation information for solar energy from satellite images follows a simple idea: instruments on board the satellite register the solar radiation scattered back from the earth's surface and the atmosphere. This value is inversely proportional to the transmission of radiation through the atmosphere, which in turn determines the level of irradiation on the ground. Since, in principle, the interaction of radiation with the components of the atmosphere (air molecules, aerosols, clouds) can be described with sufficient accuracy using radiative transfer calculation methods, precise knowledge of the composition of the atmosphere at all altitudes would therefore allow the solar radiation at ground level to be described exactly. However, this information is only partially available, so that assumptions and simplifications of "model physics" - especially for cloud cover - have to be made. With these methods it is now possible to make estimates of the solar energy supply on the ground with a resolution of typically 5x5 km² (Meteosat), the accuracy of which is comparable with conventional ground measurements, e.g. for monthly averages.

The use of satellite data now offers numerous new research possibilities: Estimates of regional solar energy potential become possible, e.g. for the sun-rich countries of Africa, where reliable ground data often does not exist. They also allow the spatial variability of solar radiation to be analysed. Knowledge of the variability of irradiation is essential for a statistical analysis of the effect of simultaneous generation in spatially distributed systems, such as photovoltaic power plants that feed into a common supply network. This provides valuable planning data.

Together with ground-based measurements on smaller spatial and temporal scales, a comprehensive picture of the statistical properties of solar radiation can be obtained in this way. This information can be used, for example, in simulation models and control procedures to improve the planning and operational management of solar energy systems.

There has always been a great deal of general interest in predicting variables that have considerable economic significance. This applies equally to the outcome of horse races, share prices and the coming weather. It is therefore not surprising that predictions of the expected energy flows from solar radiation or wind are also particularly valuable for the most efficient utilisation of renewable energies. The example of wind energy mentioned at the beginning shows that even today, forecasts of the output from wind turbines are a necessary prerequisite for optimising the integration of wind energy into the supply grids. In this way, the "capacity credit", i.e. the installed capacity from conventional power plants that can be replaced by wind energy, can be significantly increased.

Forecasts of the output from the installed wind energy converters are based on the numerical weather forecast models used by the weather services, the results of which are corrected for local influences on the flow conditions and supplemented with the characteristics of the wind energy converters. The influence of the installation geometry of wind turbines in wind farms on the flow conditions is also taken into account. As part of a research project funded by the European Union, such a forecasting method is currently being developed and tested in Oldenburg for the northern German region in the time range of one to two days. The expected output from installed photovoltaic systems is determined from satellite data using a coupled method of determining solar radiation and predicting cloud structures. Key components of this method are classification algorithms for describing these structures contained in the satellite images and a neural network-based method for describing their development over time.

On the one hand, the forecasting methods allow statements to be made about the area-wide availability of generation capacity from wind and solar energy in a specific supply area and, on the other hand, an indication of the expected uncertainty of the forecast. With forecasts in the time range of a few hours, energy supply companies can achieve significant improvements in grid and load control, while their power plant utilisation planning benefits from forecasts of up to two days. Integrating forecast information into the energy management of off-grid systems with battery storage can significantly increase the security of supply with forecasts over two to three days.

In the future, intelligent processes for planning and optimising the operation of renewable energy systems will be just as important for achieving the economic viability of these new technologies as more cost-effective manufacturing processes and higher efficiency levels. The availability of meteorological knowledge (e.g. in the form of computer models) and information (in the form of data, forecasts, etc.) will be an essential prerequisite for this.

Procedures for obtaining, processing and disseminating this information are only partially developed today and must be made increasingly available in the future. At the same time, optimisation techniques must be implemented in which this information is used as fully as possible in order to operate the systems with maximum efficiency.

The development of modern communication technologies is helping to realise these possibilities. The Internet as a marketplace for required information (the latest "energy weather forecast") and as a future control instrument (energy service companies retrieve the forecast information, calculate expected load curves and control decentralised producers and consumers accordingly) is certainly a vision for the near future.

A first step in this direction is currently being taken in an EU research project involving Oldenburg: Irradiation data obtained from satellite images is being used to determine the expected energy yields of grid-connected photovoltaic systems. This information allows the operators of the systems to compare the actual yield with the best possible yield. This provides indications of possible faulty operation and thus helps to optimise the energy yield.

Building technology is another area with particularly high potential for realising the features mentioned. In conjunction with simulation models for the areas of energy, daylight utilisation and ventilation, the methods from energy meteorology can contribute to buildings that are significantly more energy efficient and at the same time more comfortable.

Dr Detlev Heinemann , Academic Councillor in the Department of Energy and Semiconductor Research (EHF) at the Faculty of Physics, studied meteorology in Kiel and obtained his doctorate in energy research in Oldenburg in 1990. In 1994 and 1995, he was acting head of the "Physics of Renewable Energy Sources" (PRE) working group. Today he is the scientist responsible for the field of applied energy research, specialising in energy meteorology, wind energy and the simulation of renewable energy systems.

Prof Dr Jürgen Parisi, university lecturer in experimental physics and head of the EHF department, taught and researched in Tübingen, Zurich and Bayreuth after his habilitation in 1987. He was appointed to Oldenburg in 1995. His research focusses on photovoltaics, experimental semiconductor physics and non-linear dynamics.

Dr Hans-Peter Waldl , research assistant in the Department of Physics, studied physics in Marburg and Oldenburg and obtained his doctorate in wind energy research in 1997. He is currently the scientist responsible for this area.