Integrating shutdown systems: bird, ice, bat and sector-based shutdown

It is usually found on one of the back pages of the permit (approval notice), among the ancillary provisions: the turbine must be shut down in certain months, at certain temperatures and wind speeds, or when certain bird species approach. For the authority, that is a single sentence. For operations, it is a task that has to work reliably every night and every summer’s day for many years.

Bird, ice, bat and sector-based shutdown have different triggers, but they share the same core technical problem: a signal has to travel reliably from a sensor, an algorithm or an external system into the turbine controller. The turbine has to respond to it. And the entire process has to be recorded in such a way that you can prove it to the authority at any time. This interface is exactly where we work as a manufacturer-independent engineering office.

In this article, we show which types of shutdown exist and what their legal basis is, how integration works via volt-free contacts, Modbus or telecontrol protocols, how you document the shutdown and who is responsible for what in the project. At the end, you will find the sequence of a retrofit project and the typical mistakes that a little planning can avoid.

Why shutdown conditions are now part of operations

Shutdown conditions are no longer the exception. They stem from two areas of law: nature conservation law (protection of birds and bats) and immission control and safety law (shadow flicker, noise, ice throw). Both ultimately end up in the same document, the permit issued under the German Federal Immission Control Act (BImSchG), and therefore become your responsibility as the operator.

Species protection: Section 45b BNatSchG and Annex 1

For breeding birds at risk of collision, Section 45b of the German Federal Nature Conservation Act (BNatSchG) (German source) has, since its amendment, set out when the risk of killing in the vicinity of a breeding site is deemed to be significantly increased. Depending on the distance to the breeding site, this risk can be reduced by recognised protective measures. These are listed in Annex 1, Section 2 (German source): among others, anti-collision systems, shutdowns during agricultural activity and phenology-based shutdowns.

Subsection 6 is important for operators: according to it, shutdowns are deemed unreasonable if they reduce the annual energy yield by more than 8 per cent at sites with a quality factor of at least 90 per cent, and by more than 6 per cent at all other sites. Shutdowns to protect other specially protected species, such as bats, are included in this calculation. Put simply, the quality factor describes how windy a site is compared with a reference site.

Immission control and safety: shadow flicker, noise, ice

Shadow flicker and noise conditions protect local residents; ice shutdown protects people and equipment from falling or thrown ice. These conditions are older than the species protection rules and are often already built into the turbine controller. In practice, however, they now meet new systems, which raises the question of which shutdown takes priority when, and how everything is logged properly.

The types of shutdown at a glance

Before we turn to the technology, it is worth taking a sober look at the differences. Some shutdowns follow a fixed calendar, others an algorithm, and others again an event that an external system detects in real time. This results in different requirements for the signal path and for verification.

Trigger, sensors, signal path, evidence

What all types of shutdown have in common

However different the triggers may be, the chain is always the same: detect, decide, switch, record. If one link breaks, the condition is not met, even if every individual device is working. That is why we never look at shutdown systems in isolation, but always together with the turbine controller, the data connection and the database.

Bat curtailment: an algorithm instead of a sensor

Unlike birds of prey, bats cannot be tracked individually by camera. Shutdown is therefore based on weather and time conditions under which bats are known from experience to be active at rotor height. A curtailment algorithm is nothing more than a set of rules: if the month, time of day, wind speed and temperature are within a certain range, the turbine stands still or only starts up above a higher wind speed.

From blanket shutdown to nacelle monitoring

According to the Kompaktwissen paper “Warum Windenergieanlagen stillstehen” (Why wind turbines stand still) (German source) by the Fachagentur Wind und Solar (Wind and Solar Energy Agency; formerly Fachagentur Windenergie an Land), authorities initially set blanket shutdown requirements. These apply during the activity periods between April and October, usually at wind speeds below 6 m/s and temperatures above 10 °C. The exact values and time windows vary depending on the federal state and the permit.

At the same time, nacelle monitoring is usually carried out over two years: an ultrasonic detector in the nacelle records bat calls while the turbine is in operation. A site-specific algorithm is then derived from this data, which often makes the shutdown times considerably more targeted. For bat-related shutdowns, the agency cites yield losses in practice of mostly 1 to 4 per cent of annual electricity production.

Implementing ProBat algorithms technically

The best-known tool for this calculation is ProBat, a free online tool operated by the Naturstiftung David. The current release is version 7; it is complemented by the ProBat Inspector, which can be used to check actual operating times. For us as engineers, what matters is what happens after the calculation: the result is a set of bat-friendly cut-in wind speeds, staggered by periods of the year and of the night, and these values have to go into a controller.

1. Decide where the logic sits: Many turbine controllers have a bat module from the manufacturer. If it is missing or too inflexible, an external controller takes the decision and issues a stop or release command.
2. Define the measured values: Which wind speed counts (nacelle or met mast), which temperature sensor, which averaging interval? This has to match the expert report.
3. Calculate sun times: The night periods depend on sunrise and sunset at the site. A wrong time zone or a missing changeover to summer time shifts the entire shutdown.
4. Version the parameters: When a new algorithm arrives after monitoring, it must be traceable which table applied from which day.

Bird protection: anti-collision systems and time-based shutdowns

There are two fundamentally different approaches to bird protection: event-based shutdown by an anti-collision system that detects an approaching bird, and time-based shutdown on fixed occasions or during fixed periods. They differ considerably, both technically and in terms of technical operations management.

How an anti-collision system works

Under Annex 1 BNatSchG, an anti-collision system (ACS) is an automated camera- and/or radar-based system that detects the target species and, when it approaches closer than a species-specific distance, sends a signal that reduces the rotor speed down to idling (spinning) mode. Idling means that the rotor blades are pitched out of the wind and the rotor only turns slowly without producing any significant power.

This leads to the most important technical requirement: time. Only a few seconds may pass between detection and a slow rotor, because the bird keeps flying. The reaction distance is dimensioned so that detection, signal transmission and braking of the turbine together fit into the available time window. Every delay in the signal path comes directly at the expense of protection.

Which systems exist and what has been proven

A good neutral overview is provided by the system overview of the Kompetenzzentrum Naturschutz und Energiewende (KNE) (German source) in its version from the end of 2025. It lists camera systems such as AVES Wind, BIRD.IFICATION, Bird Protection System, BirdRecorder, BirdVision and IdentiFlight. Systems such as SafeWind, ProBird and DTBird, which were included in earlier versions, are no longer listed there until further notice. Radar systems for detecting individual birds were already no longer included in the 2024 version.

The state of evidence varies widely. According to the KNE, test reports are available for IdentiFlight for the red kite and the white-tailed eagle; the system is considered recognised for the red kite. According to the same source, BirdRecorder is a research prototype whose market launch is planned for 2026. The overview also notes differences in how the signal is handed over: one system transmits to the SCADA interface, another uses volt-free contacts or, alternatively, an OPC UA interface. We always clarify which option suits your turbine together with the system provider and the turbine manufacturer.

Shutdown for agricultural activity and phenology-based shutdown

Under Annex 1 BNatSchG, shutdown for agricultural activity applies from 1 April to 31 August whenever land less than 250 metres from the tower base is mown, harvested or ploughed. The turbine is shut down from the start of the event until at least 24 hours after it ends, in each case from sunrise to sunset. Phenology-based shutdown, i.e. a shutdown during sensitive phases such as courtship or the fledging of young birds, usually covers up to four or up to six weeks between 1 March and 31 August.

Technically, these are simple stops. The challenge lies in the organisation: who finds out about the mowing, who shuts the turbine down, who releases it again? A remote shutdown with a proper log matters more here than any sensor.

Ice detection and ice shutdown

At temperatures around and below freezing, ice can form on the rotor blades, which can come loose during operation and be thrown off. Many turbines therefore have an ice detection system that shuts the turbine down automatically. The relevant conditions are set out in the permit and are often based on an expert report on the risk of ice fall and ice throw at the specific site.

Methods for detecting ice accretion

• Power curve method: The controller compares the actual power with the expected power at the same wind speed and temperature. Ice impairs the aerodynamics, and power drops. The method requires no additional sensors, but only detects ice while the turbine is running.
• Ice detectors on the nacelle: A sensor detects icing conditions at the nacelle. This also works at standstill, but only provides indirect information about the blade.
• Blade sensors: Vibration sensors in the rotor blades detect the added mass from ice through changes in natural frequencies. This is the most direct method, but also the most complex.

The critical point: restart

Shutting down is easy; the restart is the tricky part. Depending on the permit, the turbine may restart automatically once the system reports that it is ice-free, or only after a visual inspection. Whether a remote inspection by camera is permissible depends on the wording of the permit. If it is, a well-aligned camera and remote switching save many site visits in winter.

Sector-based shutdown and shadow flicker modules

We use sector-based shutdown as an umbrella term for all operating restrictions tied to wind direction, sun position or time of day. The turbine then stands still or runs curtailed only in certain direction sectors or time windows. The best-known case is shadow flicker.

Shadow flicker: the LAI guideline values

Many shadow flicker conditions refer to the shadow flicker guidance for wind turbines issued by the German Federal/State Working Group on Immission Control (LAI) (German source), 2019 update. According to this guidance, periodic shadow flicker is not considered a significant nuisance if the astronomically maximum possible shadow duration at an immission point does not exceed 30 hours per calendar year and 30 minutes per calendar day, with all contributing turbines added together.

What matters for the technology: an automatic shutdown system that records actual sunshine with irradiance or illuminance sensors must limit the real shadow duration to 8 hours per year. A purely time-controlled system without meteorological parameters, by contrast, complies with the 30 astronomical hours. In addition, under the LAI guidance, data on sunshine duration and shutdown time should be documented for at least one year and presented to the authority on request.

Noise, turbulence, air traffic control

• Noise: At night, the turbine often runs in a noise-reduced mode, sometimes only for certain wind directions towards a residential area. The modes are stored in the controller; what matters is the evidence of which mode was active when.
• Turbulence and neighbouring turbines: If a turbine stands in the wake of another, a structural stability or turbulence report may require shutdowns in certain wind direction sectors. Where turbines belong to different operators or manufacturers, this quickly becomes an interface issue.
• Air traffic control and radar: Some permits contain operating restrictions relating to radar or air traffic control facilities. Their form varies widely; the permit alone is decisive.

Technical integration into the turbine controller

Now to the core. An external shutdown system is of no use as long as the turbine does not understand its signal. The turbine controller has the final say: it decides whether and how the turbine responds to a request. Our task is to connect signal and controller so that the response is reliable, fast enough and traceable.

Signal options compared

A volt-free contact is a simple switching contact without its own voltage, comparable to a light switch operated by the external system. Modbus is a widely used fieldbus protocol through which devices exchange measured values and commands, either serially (RTU) or over a network (TCP). IEC 60870-5-104 is a telecontrol protocol from power engineering that is used by grid operators, for example. OPC UA is a modern, cross-manufacturer standard for data exchange in automation.

Telecontrol unit, SCADA and database server

In a wind farm, signals rarely run directly from sensor to turbine. In between sit a telecontrol unit (the wind farm’s external data interface, for example to the grid operator or direct marketer), the SCADA system (Supervisory Control and Data Acquisition, i.e. the higher-level monitoring and control layer) and a database server that continuously stores the operating data. Each of these layers can pass on a shutdown signal, but every additional layer costs time and is a potential point of failure.

Our rule of thumb: time-critical stops, such as those from the ACS, take the shortest possible route to the turbine. The information that a stop has occurred runs in parallel into SCADA and the database. This keeps protection fast while the documentation remains complete.

Fail-safe and feedback signal

What happens if the cable breaks or the external system fails? For protective functions, we recommend the fail-safe (normally energised) principle: in normal operation the contact is closed and signals “operation permitted”. If it opens, whether due to a shutdown request or a defect, the turbine goes into the safe state. Whether this is wanted in an individual case is something we agree with the operator and in line with the authority’s conditions, because a permanently failed system then also means a permanent standstill.

The feedback signal is just as important: the turbine must report back that it has received and executed the request. Only this acknowledgement turns a sent signal into a verifiable shutdown event.

Documentation and evidence for the authority

Authorities increasingly require operators to prove compliance with shutdown conditions, often annually and often in a prescribed format. Anyone who only starts gathering data at that point quickly realises that important values were never recorded.

What should be recorded

• Time stamps from a synchronised clock, so that the turbine, ACS and database share the same time.
• Triggering measured values: wind speed, temperature, precipitation, irradiance, wind direction, depending on the type of shutdown.
• Request and response: When did the shutdown request arrive, when did the turbine respond, which operating state followed?
• Cause of the shutdown: bat, bird, ice, shadow or fault. Without a clear assignment, yield losses cannot be analysed later.
• Parameter version: Which algorithm, which reaction distance, which shadow flicker times applied at the time in question?
• Failures of the systems themselves: periods when the sensor or external system was unavailable, and what the turbine did during that time.

Reports for the authority

Our fully automatic database server continuously records the operating data: power, wind, operating states and messages. Evaluations for the authority can be generated from this raw data, for example all bat curtailment periods in a year together with the associated weather values. Many ACS providers also supply their own event logs. What matters is that both data sources match; otherwise discrepancies are the first thing to stand out during an inspection.

Technical operations management: false alarms, losses, monitoring

The real work only begins with commissioning. Shutdown systems run for years, weather and software change, and every unnecessary shutdown costs yield.

False alarms and availability losses

Every detection system has a false positive rate: it shuts down even though no relevant bird was nearby. The KNE overview reports these rates for the tested systems, and depending on the system and site they range from single-digit to double-digit percentages. With ice detection, overly sensitive settings lead to long standstills on days without real ice accretion. This can only be improved if shutdowns are recorded separately by cause and analysed regularly.

To put the order of magnitude into perspective: according to the Fachagentur Wind und Solar, Annex 2 BNatSchG specifies a flat-rate yield loss of 3 per cent for anti-collision systems; for bat curtailment, a flat-rate value of 2.5 per cent is applied in reasonableness calculations under Section 45b. You will only know the actual losses at your site if you measure them.

Monitoring via live displays

Live displays show the current state of the turbine at a glance, in the office or on site. For shutdown systems, this means you can see immediately whether a turbine is standing still because of bat protection, ice or shadow flicker, or because of a fault. With online monitoring, you can detect faults and intervene if necessary; remote administration is handled by our office during normal business hours.

Who is responsible for what?

A shutdown project often involves four or five parties. Unclear responsibilities are the most common reason why errors go unnoticed for a long time. The following division is typical; in individual cases, the contracts govern it.

How a retrofit project runs

Whether you are retrofitting an ACS, implementing a new bat algorithm or renewing an older shadow flicker shutdown, the process is similar. With us, it follows our project sequence of order confirmation, design and manufacturing, installation and commissioning, and handover with instruction and documentation.

Step by step

1. Analyse the permit: Transfer all shutdown conditions, with trigger, period, limit values and evidence requirements, into a list.
2. Survey the existing installation: Which controller, which software versions, which free inputs and outputs, which data connection are available?
3. Draw up an interface concept: Define the signal path, protocol, fail-safe behaviour, feedback signal and data recording.
4. Obtain approvals: The turbine manufacturer and the provider of the shutdown system approve the concept.
5. Manufacture and install: Prepare the control cabinet or interface, then install and wire it on site with the service company.
6. Commissioning with function test: Deliberately trigger each type of shutdown, check the response and recording, and draw up a log.
7. Handover and documentation: Hand over instruction, circuit diagrams, parameter list and test log to the operator.
8. Evaluate operations: Regularly analyse shutdowns by cause and prepare reports for the authority.

Typical mistakes

• Clocks not synchronised: The turbine, shutdown system and database run on different clocks. The analysis then does not add up.
• Cause of shutdown indistinguishable: All stops end up in the data as “external”. Bat and ice can no longer be separated later.
• No behaviour defined for system failure: If the ACS fails, the turbine keeps running unnoticed without protection, or it stands still for weeks for no reason.
• Software update without retest: After a controller update, the interface mapping has changed, and nobody checks the shutdown.
• Algorithm not updated: After nacelle monitoring, a new algorithm with lower yield losses is available, but the controller continues to work with the blanket values.
• Summer time forgotten: Night and shadow flicker times shift by one hour twice a year.

Frequently asked questions

Conclusion

Shutdown systems are mandatory, but their success is not decided by the sensor; it is decided by the chain behind it: signal, controller, feedback signal and recording. If you plan bat, bird, ice and sector-based shutdown as one joint system, with a clear signal path, defined behaviour in the event of failures and a database that stores every shutdown together with its cause, you will meet the conditions verifiably and keep yield losses under control.

We have been operating our own wind turbines since 1998 and integrate external systems into controllers and data connections independently of the manufacturer. If you would like to retrofit a shutdown system or switch an existing solution to proper documentation, we would be pleased to review your permit and your turbine together with you.