Presently enormous investments are done in Europe to add wind turbines to the existing electricity production and distribution system and even larger sums are spent on operation subsidies. In the context of these efforts it is important to note, that the existing system based on fossil fuels and nuclear energy has been operating perfectly for many decades without the addition of wind energy.
The arguments to add wind energy are well known and are summarised by the slogan: “Thousands of households are now provided with electricity generated without the use of fossil fuels”. It is conveniently ignored, that the energy intensive manufacturing industry does not accept to pay electricity bills that are a good factor higher than the bills of their competitors overseas. Worse is, that no further analysis is done by the responsible authorities to clarify the system effects of non-dispatchable electricity, as wind power is only available when the wind blows sufficiently strong.
Electricity cannot be stored in sufficient quantities to bridge periods without wind, so the existing system has to be maintained in full. This implies, that the sole justification for building an immensely expensive system ruining our rural landscape is the saving of fossil fuel and the reduction of CO₂ emissions. The credibility of this effort is proportional to the amount of fossil fuel saved by operating these wind turbines.
This is the reason why results showing adverse effects of wind energy on the operation of generators and the distribution grid are vehemently denied by the wind lobby. An amusing example is produced by the spokesman of the American Wind Energy Association (AWEA): Fred Udo's bogus numbers on wind and emission savings.
To make matters worse, the distribution system has to be appreciably enforced to transport this extra electricity from the remote areas where it is generated to the main population centres where it is needed. The difficulties of the grid in Germany illustrate this point.

To build all these structures in addition to the existing system costs not only money but also energy. Proponents of wind-energy invariably quote the time to generate the energy necessary to build a turbine to be around 6 months, but calculations from a Dutch building firm arrive at numbers three times that value[1]. Here again the credibility of wind-energy is at stake.

The initial energy spent on the building of a wind farm counts as a loss of energy produced. This is not true for the components of the classical system, because they are essential for the functioning of our society, while wind-energy is just an extra to dress our politicians in green.
Our conclusion is, that the color of the new clothes of the king is green.….

2. Calculating the energy necessary to build wind turbines.

A. Calculation based on CO₂ emissions.
A report from the university of Sydney[2] on investments in nuclear energy contains also an analysis of the payback time of various techniques of electricity generation. The following is a discussion of their results for wind-energy.
The group in Sydney distills from five different sources (eg. Vestas) average figures for the amount of material used in the construction of a wind turbine, like concrete, steel copper aluminium plastic and glasfiber. From these numbers the authors calculate the total CO₂ emissions caused by the construction and building of the turbine.
This amount is divided by the total energy generated by the turbines during its lifetime to arrive at the CO₂ emission per kWh generated. The total energy generated obviously depends on the capacity factor, grid losses and the total lifetime of the turbines.
The article considers 3 cases:

The "Material intensity factor" represents the variation in the amount of energy per ton of building material.
The emission of CO₂ per kWh produced is calculated to be 21gram, 40gram and 13gCO₂/kWh for the 3 cases. A survey of the literature in ref 2 gives an average of 27gCO₂/kWh.
The numbers do NOT include the connection to the grid and the wind induced grid extension.

The middle column shows about the correct capacity factor for the Dutch onshore windmills, but the other parameters need adjustment for this case.
The losses we assume to be 3% instead of 9%, as the transport distances in Holland are smaller. The Material intensity factor is taken to be 1,0 as this is the most probable value.
These two factors result in a factor 1,28 in favour of the Dutch case compared to case 2.
The authors calculate 40 gCO₂/kWh for case two, so the Dutch mills emit 31gCO₂/kWh produced during their life time.
The grid connection energy (mainly copper and infrastructure) we estimate at 10% of the total, so the emission increases to 34gCO₂/kWh over a period of 20 years.
Displacing fossil energy by wind energy results in a saving of 350gCO₂/kWh as it is only gas generation that is displaced[3].
In this case, the ratio between CO₂ costs and benefits is: 34/350 = 9,7% and the payback time is: 0,097*20years = 1,94 years or 23 months.
The lifetime of the turbines is assumed to be 20 years, but the Dutch subsidies are paid over a period of 15 years, so most mills in Holland stop operating after this period. This increases the life time emissions by 33% to more than 45gCO₂/kWh, if the machines are put on the scrapyard when the subsidies end. However, many are sold to other countries to continue collecting subsidies there.

B. Calculation based on energy comparison.
The data given in ref 2 allows a direct calculation in terms of thermal and electrical energy. We quote:
The breakdown of wind power results for 150 MW rated wind farm based on the turbines described is given below (Table 6.35).
(The CO₂ emission part of the table is left out.)

This table as reproduced here equates electrical and thermal energy in the data for materials and construction but applies a conversion factor 3 in the operation part.
If we apply in all cases a conversion factor of 2,5 for electrical to thermal energy (generator efficiency 40%), then we calculate 585GWh_th as the total thermal energy required for the life cycle of a 150 MW wind-farm. This number needs to be increased by 10% to account for the energy necessary for the grid connection, so the total becomes 644GWh thermal or 258GWh_el. A 150MW farm with a capacity factor of 0,22 can produce this amount of electricity in 0,93 years or 11,2 months. This result has been corrected for 3% grid losses.

The calculation B has one essential error: It equates dispatchable and non-dispatchable electricity by using one conversion factor for the transformation of electrical to thermal energy. Wind energy is non-dispatchable, so the transformation to dispatchable energy needs to be corrected for this.
Studies [3],[4]of the data provided by the Irish grid operator Eirgrid show, that the CO₂ savings due to wind energy in a system without back up from hydro are 50% of the value expected when system effects are neglected.
A second effect is curtailment of wind energy. This is the inability to accept wind energy when more is generated than is needed in the grid, with back up systems running at lowest rates. Curtailment is strongly dependent on the amount of wind energy that has to be accommodated in the grid.
Taking both effects together, the dispatchable energy saved by the wind farm is less than a factor two smaller than the non-dispatchable energy produced. Consequently the pay back time increases to minimal 22,4 months.
Please note: Other authors suspect, based on model studies, the amount of fossil fuel replacing electricity by windturbines to be even less than the 50% derived from the Irish data[5],[6].

The fist calculation based on CO₂ balance takes the system effects into account as it uses the average CO₂ emissions of the grid for dispatchable energy and it uses a smaller number for the savings due to wind energy.

3. Conclusion.

Nearly 10% of the total energy generated in 20 years by a wind turbine is spent on the building and operation of the machine itself. This result is derived from the calculations of Lenzen et al and the literature cited by the authors [ref 2].
The result can be best expressed in a payback time of 23 months.
The life-time of an installation in the Netherlands is de facto limited by the duration of the exploitation subsidies to 15 years. The result is, that in the Netherlands the payback time is 13% of the total life time of a wind complex.

References

1. J. van Oorschot, former dir. R&D and Business Dev. Volker Wessels Stevin, priv. comm. VWS is involved in the construction of wind farms. Details are given in footnote 13 in: De invloed van elektriciteit uit wind op fossiel brandstofgebruik. (in Dutch)
2. Lenzen, M.:Life cycle energy and greenhouse gas emissions of nuclear energy: A review. Page 137 ff. (The webpage contains a link to the full report as PDF-file that may be downloaded.)
   See for a summary: Energy Conversion and Management 49, 2178 - 2199. (2008)
3. Joseph Wheatley, Quantifying CO₂ savings from wind power. Submitted to Energy policy, January 2013.
   This article calculates the savings to be less than 300gCO₂/kWh taking all systems effects into account.
4. F. Udo: Wind energy in the Irish power system.
   F. Udo: Wind energy and CO₂ emissions - 2
5. C. le Pair, F. Udo & K. de Groot: Windturbines as yet unsuitable as Electricity providers. See also references 7, 8 & 9 there in.
6. C. le Pair: Electricity in the Netherlands.