Assessing the impact of turbine availability on production of electricity
Understanding availability trends of operational wind farms is considered to be a vital component in optimising the performance of such projects.
Wind turbine availability is one of the three factors on which the amount of electricity produced from a wind turbine depends. The other two factors are wind speed, and arrangement of wind turbines.
In this context, it is critical to assess the level of availability in a wind farm during its operations, and the impact of turbine size as well as the number of turbines on availability.
In terms of making the best of such information, if the probability of occurrence of a particular level of availability on historical wind farms is known then this could be used to aid understanding of the risks on a new wind farm project. Such information is additional to standard availability forecast methodology used and would provide potential validation of current availability assumptions.
It is said that turbine availability is a key indicator of downtime experienced by a turbine. A section of the industry also says turbine availability refers to the turbine system and an allowance for routine scheduled maintenance, which varies from project to project and it excludes wind farm system downtime such as grid downtime and force majeure. For its part, The British Wind Energy Association (BWEA), referring to wind turbine availability has said that this is the capability to operate when the wind is blowing, i.e. when the wind turbine is not undergoing maintenance.
According to American Wind Energy Association, availability factor is a measurement of the reliability of a wind turbine or other power plant. It refers to the percentage of time that a plant is ready to generate (that is, not out of service for maintenance or repairs).
Also, quite often, it is said that there is no standard definition of turbine availability.
The precise definition of availability may be defined in a number of different but equally valid ways depending upon the operational or commercial circumstances the definitions relate to.
These ways use the various counters and events employed by the alarm system to measure the status of the turbine, and their differences occur due to variations in the precise circumstances included in the phrase "down time": for instance, one measure may consider the loss of the grid connection as being part of a turbine's downtime figures, while another may not. In this case, the difference will be trivial in all cases where there is no grid loss, and highly significant if any grid loss occurs.
The key to overcoming these differences lies in clarity of presentation. It may be that several equally valid definitions of availability will continue to exist for the foreseeable future; however as long as all parties to a particular analysis are using the same definition this should not present a problem.
On the other hand, system availability refers to the true period of time that the wind farm as a whole is available to generate. The majority of downtime counts against availability, regardless of the cause, and only shutdown due to normal operating conditions is excluded, for instance high wind shut down or cable unwind events.
Significantly, availability is relatively insensitive to turbine and wind farm size post the sorting of teething issues.
The power available from the wind is a function of the cube of the wind speed. Therefore if the wind blows at twice the speed, its energy content will increase eight-fold, as pointed out the BWEA in the past. Turbines at a site where the wind speed averages 8 m/s produce around 75-100 percent more electricity than those where the average wind speed is 6 m/s.
Some of the factors which need to be focused for assessing the relationship between availability and high wind speed are severe climate issues; systematic turbine faults associated with high load; downtime being enforced to minimise noise or other external impact.
In its assessment, Garrad Hassan last year shared that in energy production terms, the impact of lower availability at high wind speed is in fact balanced with lower availability at lower wind speeds. In its study, it also found that there is no significant overall bias in the wind speed versus availability relationship and for a set of wind farms assessed a one percent loss of energy resulting from one percent downtime seems sensible. This is a reasonable assumption that can be made on a new project, it added.
Also, it is imperative that availability risks must be considered on an individual project basis. Referring to a region of high penetration of wind power, it was mentioned that during high wind periods, when wind power outstrips demand on a relatively weak grid, the utility company enforces wind farm projects to shutdown.
According to the BWEA, the combined output of the UK's entire wind power portfolio shows less variability, given the differences in wind speeds over the country as a whole. Whilst the amount of wind generation varies, it rarely (if ever) goes completely to zero, nor to full output.
On another note, regarding overcoming problems related to lower and upper bounds reflecting how null status counters are interpreted, it is said that the interpretation of null values recorded in the status flags and counters used to compile availability figures is a difficult issue to resolve.
Neither the assumption of full availability or of zero availability when a null value is encountered is entirely defensible. In the case that only one of the relevant fields is null, it may be possible to use a different counter where nulls have become significant. This will entail slight differences in definitions which must be accounted for in interpretation, however.
Wind Energy Performance Optimisation Summit
windenergyupdate.com's Wind Energy Performance Optimisation Summit is scheduled to take place on 11th - 12th February 2009 in Hamburg, Germany).
For more information, click here: http://www.windenergyupdate.com/performance09/programme.shtml
contact Tom Evans by email: email@example.com