Darling Independent Power Producer (Darling IPP) will operate South Africa’s first large wind farm 70 km north of Cape Town. To promote renewable and environmentally clean energy, the Minister for Minerals and Energy has declared the 13 MW wind farm a National Demonstration Project. Within the Visitor Centre at the Moedmaag Hill near Darling the independent power supply should be demonstrated for educational reasons and research. It will have a lecture hall and facilities to serve meals for 100 people. The whole complex will comprise a permanent exhibition, a mechanical workshop and a small village to accommodate 40 visitors. [1] Total independence of external sources should be demonstrated with a 100wind turbine in order that the results can be applied to the rural off grid South Africa.
Total energy independence should be subject of demonstration. In off grid applications, which still concern 30 of South Africa’s population, there is no supply network to emit or absorb electrical energy. Existing rural systems comprise of a diesel engine and rely on fossil sources to be transported there. Wind energy is available without fuel costs, but with restrictions to the domestic diurnal demand that does not correspond with a partly unpredictable wind energy generation. A general problem with electrical energy is that it can not be stored directly. Transformations into other energy sources with the inherent losses are the only solutions, e.g. to chemically bond is widely used in battery systems.
A first option is to reduce the fuel costs by adding wind power to an existing diesel-mini-grid. Without taking further steps the power quality will decrease. Measurements carried out in 1982 on Block Island [2] showed that wind energy as fuel saver requires a more sophisticated infrastructure due to part load consumption of diesel-engines at 30 while idling in smaller systems. If the engine can be switched off an alternative source has to compensate for short term fluctuations in wind energy. Current research in flywheel storage offers high energy density both in capacity for some minutes and for peak power.
Basic requirement in an isolated system is that generated energy must fit the demand at any time. In the first phases of the Visitor Centre there is surplus energy throughout a year, which needs to be fed to the grid. After completion of the whole facility this connection should be opened for at least one year to demonstrate total independence for off grid applications by simulating conditions in rural areas where no grid is available.
A study carried out 2001 shows that the potential in wind penetration in isolated systems from 100to 10currently reaches 70 . Future systems at the lower end of range might reach 95 % strongly depending on storage possibilities while the missing energy comes from diesel engines.[3]
Measurements from the wind farm site are correlated the Visitor Centre on the same property using a factor 0,8 to consider a possibly poorer wind power quality. The annual pattern directly shows the energy harvested by the Fuhrlaender FL 100 converter. [4]
Fig. 1: Fuhrlaender FL 100 monthly generation [kWh] and average power [kW] at a fictitious site near the Visitor Centre
This site should allow no less then 29 of annual full power. Obvious is the difference between the weakest month in winter to the strongest month in summer. Limiting factor would be June if no long-term storage saves energy generated in summer with 2,7fold generation to the weakest monthly average. In the following scenarios these extremes are subject of closer examination:
The Fuhrlaender FL 100 does not generate electricity at wind speeds below 5thus an analysis of calms is more important. Wind data acquired with the Ammonit Wind Siter 400 does only provide statistical information; temporal data of occurrence is lost.
Fig. 2: Relative accumulated occurrence of calms < 5with different durations throughout the particular month
Short lull periods do not show an annual pattern as a nature of wind itself. Longer calms strongly depend on time of the year. In the weaker winter calms prevail against wind speeds strong enough to produce electricity. In summer each month has 9 days of time accumulated when the WEC is out of operation.
The measurements above do not give information on the time the calms occur. Normally the equipment used has to assess the amount of electricity sellable to the supply network. Without further considerations the grid absorbs the electricity at any time. But if wind penetration in the supply network increases, it gets more important to know the correlation between demand and supply at any time. In case of the Visitor Centre wind penetration should reach at least 90%.
The knowledge of calms reaches utmost importance; but no raw data is available to perform an exact analysis. This data has to be estimated from printed results of external data measurements. Wind speeds do not need to be transferred to a different altitude; the converter’s hub is only two meters higher than the measuring height.
Temperature rise during daytime influences wind speeds.
Fig. 3: Diurnal power scheme for weakest and strongest month. In June the statistic analysis
falsifies the mean generation as a matter of stronger wind speeds occurring both day and night
with the result of a low monthly average.
As shown in the plot for June there is no specific pattern for the monthly average. Calms during daytime keep the average hourly level low. This results in a lower generation than 15,8calculated form the measurement’s frequency distribution.
To simulate wind power generation, several random numbers are multiplied to get realistic data with hourly, diurnal and short-term fluctuations. Mayor problems are caused by calms exceeding longer periods where no battery system can economically provide energy.
The design of an isolated system requires exact analysis of demand in electrical energy throughout the year. One has to distinguish two forms of consumers: Those, which depend on continuous supply like refrigerators, lighting, entertainment, etc. and those, which can be switched of during in power shortages to avoid voltage breakdown. On low demand of the Visitor Centre surplus energy has to be transformed in some way ¯including dissipation into heat in case no storage or other appliances are available to use the energy. Energy storage is possible in hydrogen, battery systems, pumped water, desalinated water or heat.
To gather information about possible future demand all appliances in the centre are listed and assessed with diurnal profiles and simultaneous usage.
Energy consumption follows a specific pattern depending on the consumers in the grid. Usual patterns are diurnal, weekly and annual profiles of total consumption within the grid. In larger grids the course gets smoother with the number of independent consumers.
The diurnal pattern follows the activities at the Visitor Centre. In times of low consumption there is surplus energy for brackish water treatment, laundering clothes, water pumping or other activities that are not scheduled in time. Whenever the Centre is not used, e.g. on holidays, weekends or idle periods there is surplus energy, which has to be stored, dissipated in heat or being fed to the gird. To demonstrate possibilities for rural and domestic use an additional storage or additional consumers have to simulate a domestic pattern.
A possible pattern for a day of full visitor capacity shows that
Fig. 4: Diurnal profile at the Visitor Centre
19 kW mean power compared with 29 kW annual average enables independent supply easily throughout the year. This is even true for the weakest month in average (16 kW generation) if there is no consumption on weekends. Some consumers are not yet considered in this profile but they allow using surplus energy when available. One day’s consumption is 459 kWh; to consider additional consumers not integrated yet firm consumption should be 500Firm consumption means the load can be scheduled in advance as an advantage to wind energy generation, which only could be predicted with lower validity.
The monthly pattern strongly depends on the number of visitors. In this first approach it is considered that energy consumption is in direct proportion to the number of visitors attending the Centre. A small base load is considered for the refrigerators and building maintenance, as well as scientific equipment and water production.
Fig. 5: Possible monthly load profiles for the Centre. Normal visitor capacity serves
as basis for further analysis in a final stage of the Centre’s usage.
The peak power consumption is less than 40 kW; to enable future growth the electrical system is designed to 100to enable maximum generation of the wind energy converter. This is true for plant oil generator, battery inverter, flywheel storage, fuel cell and transformer station to the grid. The wind energy converter is able to generate 100- but only 29% of a year’s average which is in deed a good value.
Several possibilities are briefly discussed to choose an appropriate solution for the mini-grid.
In this system more electrical energy is generated than needed. As an option for storage, fuel cells are considered to use the hydrogen generated with surplus energy. The fuel cell needs an inverter as well as the battery system. If this inverter operates the whole time, even in power shortage, a plant oil driven engine could drive a cheaper asynchronous generator to bridge some hours of shortage. The wind turbine could be compensated locally with discrete capacitors.
In an isolated system the frequency has to be stabilized; normally this is done with large synchronous generators where rotational speed and frequency correspond. Wind energy converters with constant generator speeds need stall controlled rotor blades to operate at different wind speeds. Larger turbines operate at variable frequencies, but costly inverter technology is required apart pitch control for the rotor blades. The Fuhrlaender turbine uses an asynchronous generator which needs an external constant frequency to operate. To provide this, a diesel engine with a larger synchronous generator constantly running is the cheapest solution, but saving only a poor amount of diesel. In order that the combustion engine can be switched off, a self-conducted battery inverter is a solution. It is needed because batteries only store direct current. Self conducted means that the semiconductor switches use a frequency pattern generated by a micro controller, instead of chopping battery current in the same pattern as the external grid frequency.
The energy consumption strongly depends on the number of visitors attending the centre in unpredictable patterns. The staff working there only needs energy on weekdays. Thus a long-term storage has to compensate fluctuations covering some weeks. Taking the average generation for the weakest month there is enough energy for this period. But if calms are taken into account and exactly correlate with a five day program at the centre there is a shortage of approx. 2 500This extreme might only occur once several years, then a plant oil generator could serve the centre or the energy could be supplied from the existing grid.
For further considerations the long-term storage should enable to serve two consecutive days of full load. The storage capacity is 1000enough to compensate monthly fluctuations.
Wind is a highly fluctuating source of electrical energy, measurements first carried out in 1957 () show that the wind follows a certain spectral pattern:
Fig. 6: Measured wind speed variance spectrum with universal validity (1957)
A characteristic gap separates short and long term fluctuations. The right side deteriorates power quality and is caused by turbulence. The peak in the middle follows the diurnal pattern with a period of 12 hours caused by different temperatures of day and night. High and low-pressure areas persist 4 days.
The distribution of wind speeds shows that short-term variations in wind speed cause significant decrease in power quality.
Fig. 7: Short-term variations of wind speed. The data omits fluctuations
shorter than one minute in this one hour plot.
The relative wind speed shows periods of some minutes, as a matter of temporal resolution, fluctuations with shorter periods than one minute also occur. In larger supply networks power consumption can be assumed to be constant for 10 minutes of period. In a smaller supply network, like the Visitor Centre, switching in a larger drive with 5rated power, causes starting currents of up to 30which is 100 fluctuation for about one second. If the power supply is not capable to provide this energy a severe voltage breakdown might cause a complete power failure. To avoid these extremes, a soft-start device reduces start-up currents to 20 % load shifts in the Centre. Nevertheless there must be a generator to provide enough reserve power to compensate these load shifts. The worst load shift is when starting the major power source itself. Before the WEC is able to generate electricity, the magnetic field has to be build up inside the engine with reactive power.
Compressed air could serve as a short-term storage. Since 1978 a 300 00³ salt cavern is used as a pressure vessel in northern Germany. In a natural gas turbine, the compressed air raises the efficiency, being used for combustion. 290 MW can be emitted for 4 hours.
A smaller system with a 8³ pressure vessel and a 15turbine operates without natural gas combustion. Storage capacity is 1,25this is 15peak power for 5Compressing to 10pressure heats up the air to 200°C, a thorough insulation is required. [http://www.tu-clausthal.de/presse/tucontact/2002/Oktober/tuc1/09a.pdf]
A plant of the double size could serve the Centre in the same way as the flywheel plant.
Discontinuous usage of the Centre requires a long-term-storage system. The capacity should compensate fluctuations throughout a month. This might not be a problem if this system feeds domestic appliances with continuous demand every day. But diurnal compensation with a battery system is needed due to a high ratio from maximum to minimum in domestic and the Centre’s consumption. Generally long-term-storages do not provide high dynamics in power shifts at a lower efficiency. The task of a load management system is to consider the respective capacity stored in the different storages with the technical characteristics of each system.
This storage system should be used to compensate fluctuations in demand and supply throughout a day. An ideal system’s charging condition should be the same at the beginning of each day. For the Centre approx. 100 full load cycles might be reached annually, this is mostly true for the weaker wind periods.
Fig. 8: Diurnal 200compensation storage with random wind generated electricity.
The time series of the wind velocity shows that several wind patterns are possible
The figure above shows a 200battery storage in comparison with a possible wind pattern. 100 kWh of its capacity are used in the plot, the maximum capacity is just reached for 3 hours while the surplus generation has to be used in some way not to dissipate it. If a long-term storage is available it could serve the centre on days with energy deficits. If no other storage is installed, this battery system is partly able to compensate short term fluctuation and reduce the backup generator start-ups.
200 kW is two hours of maximum generation or 40 of one day’s demand. A storage system of this size is absolutely necessary in a system with wind energy generation!
It is assumed that the Centre will only be visited in a discontinuous pattern. Some events could be scheduled in dependence of the energy available ¯but there will be some days with low energy consumption followed by some days of peak consumption.
A two-day storage could serve in the weakest month as follows:
Fig. 9: Simulation of the weakest month with a 1000 kWh storage system.
If the real storage system reaches negative charge this deficit has to be generated
in a backup system. Beginning from negative values, the surplus generation is stored
in the empty storage. The unlimited storage shows a surplus at the end of the month.
Since losses in the storage system are not considered yet this month in critical
with the assumed capacity of the centre.
Obvious is that there is surplus energy throughout the whole month. In the diurnal profile some consumers are not considered jet like hydrogen generation for additional use, etc. The deficit must come from an alternative source and raises the available monthly electricity in order that other consumers could be served sufficiently and losses in the storage system cause no shortage. The percentages of losses strongly depend on the type of long term storage system to be discussed in the following chapters.
Hydrogen can be generated out of pure water with the help of surplus electricity. The H2O is decomposed into H2 and O2 where hydrogen is stored in pressure vessels and oxygen released to the atmosphere. To produce and to store 1³ of H2 approx. 5of electricity are needed. Transferring it back to electricity, there is 50 loss in heat. Thus a 1000or 1000³H2 storage needs one hundred 50 l-pressure vessels at 200The electrical efficiency is 20 and 40 total if the fuel cell’s thermal heat could be used.[ Source: Pritchard: Wind energy and the hydrogen economy, www.anglesey-wind.co.uk]
Storage is possible in pressurized H2, liquid H2 or metal hydride. Liquid H2 is difficult to handle and requires temperatures of –253°C at 20 … 0 % of the energy stored to liquefy the gas. Pressurized hydrogen is safe to store and to transport as metal hydride, as being favoured for transportation purposes.
Each kWh of electricity produces one kWh of heat in the fuel cell. In summer there is a limited demand of heat for hot water, in winter when most of the stored hydrogen will be retrieved, it can additionally be used for heating. The temperature level of waste heat is ???°C, enough to produce hot water and serve in heating.
If hydrogen should serve as the only long-term storage the low electrical efficiency from generation, pressurized storage and transformation back to electrical energy has to be taken into account. The heat could easily be used in the Centre, but the low electrical efficiency shows that, if 20 of the monthly electricity came from the hydrogen storage system, the whole wind generated electricity has to be used to generate hydrogen.
| total wind generated electricity | 11.360 kWh |
| total Centre’s consumption | 9.200 kWh |
| 20% electrical consumption from the fuel cell | 1.840 kWh |
| energy to generate and to store 1840 at 20% efficiency | 9.200 kWh |
| total deficit | 7.040 kWh |
Hydrogen generation, storage and transformation back to electricity is still very expensive and only an opportunity if thermal discharge could be used practically. For the Visitor Centre’s lowest month in wind generated electricity, it corresponds with some demand of heat. During power shortage in this month, the backup system should be a plant oil engine, generating enough heat itself. Thus a continuous thermal heat supply throughout winter is ensured.A hydrogen system with the only purpose of a long term electrical storage in the weakest months is no opportunity. If the hydrogen could be used in transportation vehicles directly, it might be a technology to demonstrate at a small scale.
Given the constraints of the site, water pumped storage would be possible to demonstrate on the hill. Two ponds have to be excavated and sealed against losses. At the lowest level of a penstock connection a Francis turbine could both be used to store electrical energy in pumped water on the hill and release its potential the reverse way in the same compact unit.
Starting the engine as generator or pump only takes some seconds but reversing the energy flow has to overcome the kinetic energy in the flowing water in the whole pipe. This takes a few minutes, closing the valves and stopping the engine can not be done on a sudden as well.In addition two stop valves and a security pressure relieve valve are required.
The capacity of the storage system directly corresponds with the quantity of water in the upper pond and is clearly understandable to the visitors.