Availability analysis

Kjøllefjord windfarm - Statkraft

The Availability analysis has been developed over a long period of time and is currently used to follow up on a wide range of turbine models and service providers.

Availability can be calculated in many different ways and vary from contract to contract and from company to company.

The Availability analysis is based on the IEC standard, but has a number of built-in features to customise the calculations to the user specific needs.

The analysis enables accurate reporting and  management of service contracts and claims, provides insight into how downtime is distributed, and calculates lost production and income for all incidents with online and offline losses.



Key customer value

– operator follow-up and internal KPIs

Monitor the quality of O&M service and turbine supplier

Efficient management of warranties and claims

Calculate and report O&M related KPIs to management and JV-partners

Identify underperforming turbines

Identify areas of improvement

Knowledge transfer from operational phase to pre-construction activities


The main purpose of this analysis is to assess the performance of a wind turbine and quantify lost production.

Many advanced calculation steps are necessary to obtain reliable results such as calibrating nacelle anemometer readings with directional nacelle transfer functions, filter outliers, adjust for air density and turbulence intensity

Moreover, this module is used to investigate a multitude of performance issues and allows the user to filter data by a number of criteria such as date, wind and power ranges or alarm codes.

Particularly, the integration of this module with the SCADA alarm logs is a unique function that allows the user to assess the impact of certain alarms on production and can be used to determine the amount of lost production due to ice, blade vibration, etc.

Power curve analysis

Kjøllefjord windfarm - Statkraft

Deep dives

Torch allows the user to transfer available data into insight by combining different data sources in an intuitive and connected workflow.


Trend tools and general x-y plots are connected and linked to alarm log data. This allows for fast and efficient filtering of data to get to the core of the analysis in an undisturbed workflow.


Torch is capable of executing and scheduling analyses and more advanced scripts in the background.


Calculation-heavy jobs can be routed to different servers to speed up computation and notifications can be send out based on fully customizable trigger conditions.


Torch is not a blackbox solution that restricts the user to predefined analyses.


The expert user can create new analyses in the Python scripting language and take advantage of the vast amount of publically available open-source libraries. This includes the great support Python enjoys in the Machine Learning community and enables advanced users to implement their own, tailored condition monitoring and predictive analytics tool kit.