MesH Engineering GmbH

RegViento

RegViento: Innovative wind farm control with digital load twins and uncertainty quantification

Wind turbines are usually installed in close proximity to one another in wind farms. However, this arrangement leads to efficiency losses, the so-called wake losses, which arise from the turbines shadowing each other and are quantified at 10 to 20 % of the annual energy production. In the RegViento project, MesH Engineering GmbH, together with Stuttgart Wind Energy (SWE) at the University of Stuttgart, is developing a flow-influencing wind farm control system that can substantially reduce these losses through coordinated operation of the individual turbines. Depending on the site, increases in annual energy production of about 1 % are possible; this corresponds to a sum of about €70 million for the electricity produced from wind power in Germany in 2023 (140 TWh) at an electricity price of 50 €/MWh.

Objective

The project is developing a flow-influencing wind farm control system to market readiness and demonstrating it under real-world conditions. For this purpose, a modular, model-based closed-loop control framework is being built that can be applied to any wind farm configuration. From the uncertainty-affected measurements of the turbines, the wind farm controller continuously determines new setpoints, e.g. yaw angle offsets for deflecting the wake (wake redirection control), pursuing the control objective (such as maximizing the wind farm power output) while ensuring that no critical loads occur outside the certified load envelope.

Digital load twins

A central building block of the framework is the digital load twin of the wind turbine. Since turbine manufacturers typically provide operators with only limited information about their products, the project develops a method for deriving a detailed aeroelastic model from the limited available data. On this basis, a real-time capable machine learning surrogate model is created using recurrent neural networks (LSTM), capable of monitoring and predicting the loads of the main components online. Through continuous fine-tuning with current measurement data, aging and soiling effects of the turbine are incorporated into the twin.

Uncertainty quantification

A comprehensive uncertainty quantification of both the digital load twin and the wind farm flow model is a prerequisite for making reliable statements about fatigue loads and the trustworthiness of the computed flow state. Among the uncertainties examined and propagated are those from the model derivation under limited data availability, from the aeroelastic simulation, from the machine learning surrogate model, and from the flow modeling. From the combined uncertainties, a concrete recommendation is finally derived regarding compliance with the load budgets and the applicability of the wind farm control.

Demonstration at the WINSENT test site

The developed solution will be demonstrated at the WINSENT wind energy test site in the Swabian Alb, funded by the state of Baden-Württemberg. Over a period of one year, the control strategy will be tested on the extensively instrumented wind turbines and validated for commercial use. Integration into the widely used wind farm supervision software PcVue, for which MesH is a certified system integrator, ensures that the control concept can subsequently be delivered efficiently as a product for wind farm operators.

Project status

In the first project year, the technical and methodological foundations for the flow-influencing wind farm control were established. A first functional baseline version of the control framework is available, including implementations in Matlab and Python as well as MQTT communication as the interface to the higher-level wind farm control logic. The flow modeling builds on the FLORIS/FloriDyn model family and is coupled to the simulation environment via MesH REcon; FAST.Farm has been integrated for realistic testing. For the WINSENT demonstration, a wake redirection mode was implemented in the turbine controller in coordination with ZSW, the OPC UA communication between the test site infrastructure and PcVue was successfully established, and a Design and Verification Plan (DVP) was drafted. Safety-oriented load simulations under yaw misalignment were performed and harmonized with the original certification report. After its first year, the project is therefore in a solid position for the planned continuation of the demonstration, validation, and integration work.

Project partners:

Funding program: VwV Invest BW Innovation III
Project duration: April 1, 2025 to March 31, 2027

OTELLO

e-TWINS

GridLoads

GridLoads - Methods and tools for grid-based control of wind turbine generators

In the "GridLoads" research project, Fraunhofer IEE and MesH Engineering GmbH are investigating interactions between frequency-supported generator control and mechanical loading of the drive train. The aim is the development of an analysis model with which important information can be obtained for the design of the grid-supporting measures as well as for the design of the drive trains. The project is funded by the BMWi and will run until the beginning of 2020.

Target

For the conversion of the German energy supply and the large-scale expansion of wind energy, wind turbines must increasingly contribute to grid support and grid stabilisation. In particular, this requires a highly dynamic reaction of the active and reactive power to disturbances in the grid, such as short circuits, partial grid splits, power plant failures, etc. In special cases, a short-term motor operation of the plants can also be useful. The generator controls required for this are currently the subject of extensive research activities. However, one aspect of these controls that has not yet been comprehensively investigated from the point of view of the individual plants is that dynamic voltage and frequency changes in the grid have a direct effect on the active power of the plants and thus on the excitation of drive train vibrations and the associated mechanical loads.

Grid Support

The requirements for fast grid support in the event of disturbances in the interconnected grid will make new demands on the generator control of wind turbines in the future. Especially for fast frequency support, high gradients in the active power must be realized. These changes in active power result in corresponding changes in the torque at the generator and can thus stimulate the structure of the wind turbine to oscillate and cause increased mechanical loads. This affects in particular the drive train, i.e. the coupled system consisting of rotor blades, shafts, gearbox and generator, but also the tower of the wind turbine. Through generator controls, which react with high dynamics to changes in magnitude and phase of the local voltage pointer, the dynamics of the interconnected grid are effectively coupled with the structural dynamics of the WTG. Thus coupled oscillation modes can occur, which couple the behaviour of several individual plants with each other as well as with other generators in the network.

Therefore, a multidisciplinary analysis environment is to be created to investigate the feedback effect and the influences of the new grid-supporting generator controls on the structural dynamics, stability and loads of a wind turbine.

Multibody Systems

Multibody models of structural dynamics are coupled with models for the description of generator dynamics and the electrical network. In particular, the generator system must be mapped with high accuracy as a coupling element between the dynamic network processes and the structural dynamics of the wind turbines. Influences of this type of control system on the structural dynamics will cause changed requirements on the design of the drive train systems, which will be investigated in this project.

The analysis of vibration phenomena on the drive train of a wind turbine is carried out according to the state of the art with correspondingly resolved multi-body models which describe the elastic structural components and the coupling with the aerodynamic forces. Depending on the problem, such models can describe low-frequency oscillations in the range of 0.1 Hz up to problems of sound propagation at several 100 Hz.

Stability

Special attention will be paid to the possibility of instabilities that cause structural failure. Wind turbines with a modern control system can experience a reduction of the system damping due to the interaction between the aerodynamic forces, elastic forces, inertial forces and the forces induced by the control system, which in the worst case leads to instability or strong vibration. In order to predict the occurrence of such instabilities, the interactions and interactions between structure, aerodynamics and control system must be considered in the overall system model. For the stability analysis of a wind turbine with grid-supporting control system, a modeling strategy must first be defined and a simulation environment implemented. This simulation environment is used for the construction of the overall system model. MesH has already developed a software tool for the aeroelastic stability analysis of a wind turbine. This is based on the modeling of the wind turbine as a multi-body system. For the stability analysis in gridloads the existing stability analysis tool shall be further developed in order to enable a sufficiently accurate prediction of the aero-servo-(electro)-elastic instability. With the application of the newly developed tool, a possibility is created to determine the distance to the stability limit and thus to enable the design of the control system to influence stability.

A further aspect of the feedback effect of the mains supporting control system consists in influencing the component loads of the WTG (e.g. rotor, gearbox, etc.). These loads are dimensioning for the design of the structure. Time simulations for the overall system model are performed and evaluated to analyze the loads. For these time simulations representative load cases are selected and compared with the loads from the simulation without network supporting control systems.

Acoustic emissions

In addition to the mechanical loads, possible mechanisms for the excitation of increased noise emissions resulting from the generator system and the coupling with the grid should also be considered. Example effects are asymmetric network states and asymmetric magnetic fields caused by elastic displacements of the generator rotor.

After identifying the feedback effect of the grid-supporting control system on loads, the second step is to investigate the possibilities for actively influencing the vibration behaviour by controlling the generator system. IWES plans an extension of the known grid-supporting control procedures by measures for the reduction of powertrain loads. In order to reduce the system costs, a joint optimization of the structural and control parameters makes sense. In this subproject, a numerical sensitivity analysis will be used to establish this parameter space. The described investigations are to be carried out in the simulation in order to make the couplings between structural dynamics and network transparent. On the other hand, the investigated controls for commercial converter systems are not yet available. Experimental testing on generator or nacelle test benches is not expedient, since the structural dynamics of the wind turbines of interest cannot be realistically simulated due to the lack of rotor blades and the bandwidth restrictions in the control system.

Project partner:

Project number: 0324192B
Project duration: 01.09.2017 – 29.02.2020

TremAc

TremAc - Objective criteria for vibration and sound emissions of onshore wind turbines

The now completed joint project TremAc, which was led by the Institute for Soil Mechanics and Rock Mechanics (IBF) at the Karlsruhe Institute of Technology (KIT), served to identify the main parameters and threshold values for an objective evaluation of the noise and vibration load of wind turbines as a function of the plant data, the construction, the topography and the distance to the immission site.

In this project we at MesH Engineering have dealt intensively with the emission of tonal noises from wind turbines with gears. In this project we were able to develop a simulation process which allows us to determine the tonal noise emissions of wind turbines using only numerical methods. This type of noise emission from gear units originates from the gearing of the individual gear stages. The excitation at the gears is represented by a special force element in our very detailed multi-body model. The multi-body model also represents the entire transfer path of vibrations within the wind turbine. The low-frequency oscillations, which are particularly in focus in this project, are then emitted into the environment as airborne noise, especially at the rotor blades and the tower. The transition from the vibration in the structure to the surrounding air cannot be modelled in a multi-body system. Therefore, a dedicated finite element model was developed for this purpose in order to also consider this effect numerically.

The possibility to treat tonal vibroacoustics numerically is a very important but not yet sufficient step in the improvement of modern wind turbines. Therefore, MesH Engineering also investigated different possibilities to get the wind turbines quieter in the TremAc project. One possibility investigated is to change the microgeometry of the gearing in such a way that the transmission error at the gearing and thus the excitations of the system become smaller. Another possibility is the installation of hydromounts at the connection points between the rotating and non-rotating components of the wind turbines. The damping properties of hydromounts make their numerical treatment in multi-body systems very complex. In the TremAc project, however, our experts have succeeded in finding an efficient way of modelling and quantifying the effect on the reduction of tonal noise emissions. Both possibilities for influencing tonal noise emissions showed a promising result that will certainly influence the development of future generations of wind turbines. In this way, our experts are helping to ensure that we can sleep even better as residents near wind farms in the future.

You can find out more about the project at https://www.windfors.de/en/projects/tremac/

TremAc is a project in

Project partners:

Project number: 0325839
Project duration: 01.02.2016 – 31.07.2019

RegViento

RegViento: Innovative wind farm control with digital load twins and uncertainty quantification

Wind turbines are usually installed in close proximity to one another in wind farms. However, this arrangement leads to efficiency losses, the so-called wake losses, which arise from the turbines shadowing each other and are quantified at 10 to 20 % of the annual energy production. In the RegViento project, MesH Engineering GmbH, together with Stuttgart Wind Energy (SWE) at the University of Stuttgart, is developing a flow-influencing wind farm control system that can substantially reduce these losses through coordinated operation of the individual turbines. Depending on the site, increases in annual energy production of about 1 % are possible; this corresponds to a sum of about €70 million for the electricity produced from wind power in Germany in 2023 (140 TWh) at an electricity price of 50 €/MWh.

Objective

The project is developing a flow-influencing wind farm control system to market readiness and demonstrating it under real-world conditions. For this purpose, a modular, model-based closed-loop control framework is being built that can be applied to any wind farm configuration. From the uncertainty-affected measurements of the turbines, the wind farm controller continuously determines new setpoints, e.g. yaw angle offsets for deflecting the wake (wake redirection control), pursuing the control objective (such as maximizing the wind farm power output) while ensuring that no critical loads occur outside the certified load envelope.

Digital load twins

A central building block of the framework is the digital load twin of the wind turbine. Since turbine manufacturers typically provide operators with only limited information about their products, the project develops a method for deriving a detailed aeroelastic model from the limited available data. On this basis, a real-time capable machine learning surrogate model is created using recurrent neural networks (LSTM), capable of monitoring and predicting the loads of the main components online. Through continuous fine-tuning with current measurement data, aging and soiling effects of the turbine are incorporated into the twin.

Uncertainty quantification

A comprehensive uncertainty quantification of both the digital load twin and the wind farm flow model is a prerequisite for making reliable statements about fatigue loads and the trustworthiness of the computed flow state. Among the uncertainties examined and propagated are those from the model derivation under limited data availability, from the aeroelastic simulation, from the machine learning surrogate model, and from the flow modeling. From the combined uncertainties, a concrete recommendation is finally derived regarding compliance with the load budgets and the applicability of the wind farm control.

Demonstration at the WINSENT test site

The developed solution will be demonstrated at the WINSENT wind energy test site in the Swabian Alb, funded by the state of Baden-Württemberg. Over a period of one year, the control strategy will be tested on the extensively instrumented wind turbines and validated for commercial use. Integration into the widely used wind farm supervision software PcVue, for which MesH is a certified system integrator, ensures that the control concept can subsequently be delivered efficiently as a product for wind farm operators.

Project status

In the first project year, the technical and methodological foundations for the flow-influencing wind farm control were established. A first functional baseline version of the control framework is available, including implementations in Matlab and Python as well as MQTT communication as the interface to the higher-level wind farm control logic. The flow modeling builds on the FLORIS/FloriDyn model family and is coupled to the simulation environment via MesH REcon; FAST.Farm has been integrated for realistic testing. For the WINSENT demonstration, a wake redirection mode was implemented in the turbine controller in coordination with ZSW, the OPC UA communication between the test site infrastructure and PcVue was successfully established, and a Design and Verification Plan (DVP) was drafted. Safety-oriented load simulations under yaw misalignment were performed and harmonized with the original certification report. After its first year, the project is therefore in a solid position for the planned continuation of the demonstration, validation, and integration work.

Project partners:

Funding program: VwV Invest BW Innovation III
Project duration: April 1, 2025 to March 31, 2027

OTELLO

e-TWINS

GridLoads

GridLoads: Methods and tools for grid-based control of wind turbine generators

In the "GridLoads" research project, Fraunhofer IEE and MesH Engineering GmbH are investigating interactions between frequency-supported generator control and mechanical loading of the drive train. The aim is the development of an analysis model with which important information can be obtained for the design of the grid-supporting measures as well as for the design of the drive trains. The project is funded by the BMWi and will run until the beginning of 2020.

Target

For the conversion of the German energy supply and the large-scale expansion of wind energy, wind turbines must increasingly contribute to grid support and grid stabilisation. In particular, this requires a highly dynamic reaction of the active and reactive power to disturbances in the grid, such as short circuits, partial grid splits, power plant failures, etc. In special cases, a short-term motor operation of the plants can also be useful. The generator controls required for this are currently the subject of extensive research activities. However, one aspect of these controls that has not yet been comprehensively investigated from the point of view of the individual plants is that dynamic voltage and frequency changes in the grid have a direct effect on the active power of the plants and thus on the excitation of drive train vibrations and the associated mechanical loads.

Grid Support

The requirements for fast grid support in the event of disturbances in the interconnected grid will make new demands on the generator control of wind turbines in the future. Especially for fast frequency support, high gradients in the active power must be realized. These changes in active power result in corresponding changes in the torque at the generator and can thus stimulate the structure of the wind turbine to oscillate and cause increased mechanical loads. This affects in particular the drive train, i.e. the coupled system consisting of rotor blades, shafts, gearbox and generator, but also the tower of the wind turbine. Through generator controls, which react with high dynamics to changes in magnitude and phase of the local voltage pointer, the dynamics of the interconnected grid are effectively coupled with the structural dynamics of the WTG. Thus coupled oscillation modes can occur, which couple the behaviour of several individual plants with each other as well as with other generators in the network.

Therefore, a multidisciplinary analysis environment is to be created to investigate the feedback effect and the influences of the new grid-supporting generator controls on the structural dynamics, stability and loads of a wind turbine.

Multibody Systems

Multibody models of structural dynamics are coupled with models for the description of generator dynamics and the electrical network. In particular, the generator system must be mapped with high accuracy as a coupling element between the dynamic network processes and the structural dynamics of the wind turbines. Influences of this type of control system on the structural dynamics will cause changed requirements on the design of the drive train systems, which will be investigated in this project.

The analysis of vibration phenomena on the drive train of a wind turbine is carried out according to the state of the art with correspondingly resolved multi-body models which describe the elastic structural components and the coupling with the aerodynamic forces. Depending on the problem, such models can describe low-frequency oscillations in the range of 0.1 Hz up to problems of sound propagation at several 100 Hz.

Stability

Special attention will be paid to the possibility of instabilities that cause structural failure. Wind turbines with a modern control system can experience a reduction of the system damping due to the interaction between the aerodynamic forces, elastic forces, inertial forces and the forces induced by the control system, which in the worst case leads to instability or strong vibration. In order to predict the occurrence of such instabilities, the interactions and interactions between structure, aerodynamics and control system must be considered in the overall system model. For the stability analysis of a wind turbine with grid-supporting control system, a modeling strategy must first be defined and a simulation environment implemented. This simulation environment is used for the construction of the overall system model. MesH has already developed a software tool for the aeroelastic stability analysis of a wind turbine. This is based on the modeling of the wind turbine as a multi-body system. For the stability analysis in gridloads the existing stability analysis tool shall be further developed in order to enable a sufficiently accurate prediction of the aero-servo-(electro)-elastic instability. With the application of the newly developed tool, a possibility is created to determine the distance to the stability limit and thus to enable the design of the control system to influence stability.

A further aspect of the feedback effect of the mains supporting control system consists in influencing the component loads of the WTG (e.g. rotor, gearbox, etc.). These loads are dimensioning for the design of the structure. Time simulations for the overall system model are performed and evaluated to analyze the loads. For these time simulations representative load cases are selected and compared with the loads from the simulation without network supporting control systems.

Acoustic emissions

In addition to the mechanical loads, possible mechanisms for the excitation of increased noise emissions resulting from the generator system and the coupling with the grid should also be considered. Example effects are asymmetric network states and asymmetric magnetic fields caused by elastic displacements of the generator rotor.

After identifying the feedback effect of the grid-supporting control system on loads, the second step is to investigate the possibilities for actively influencing the vibration behaviour by controlling the generator system. IWES plans an extension of the known grid-supporting control procedures by measures for the reduction of powertrain loads. In order to reduce the system costs, a joint optimization of the structural and control parameters makes sense. In this subproject, a numerical sensitivity analysis will be used to establish this parameter space. The described investigations are to be carried out in the simulation in order to make the couplings between structural dynamics and network transparent. On the other hand, the investigated controls for commercial converter systems are not yet available. Experimental testing on generator or nacelle test benches is not expedient, since the structural dynamics of the wind turbines of interest cannot be realistically simulated due to the lack of rotor blades and the bandwidth restrictions in the control system.

Project partner:

Project number: 0324192B
Project duration: 01.09.2017 – 29.02.2020

TremAc

TremAc: Objective criteria for vibration and sound emissions of onshore wind turbines

The now completed joint project TremAc, which was led by the Institute for Soil Mechanics and Rock Mechanics (IBF) at the Karlsruhe Institute of Technology (KIT), served to identify the main parameters and threshold values for an objective evaluation of the noise and vibration load of wind turbines as a function of the plant data, the construction, the topography and the distance to the immission site.

In this project we at MesH Engineering have dealt intensively with the emission of tonal noises from wind turbines with gears. In this project we were able to develop a simulation process which allows us to determine the tonal noise emissions of wind turbines using only numerical methods. This type of noise emission from gear units originates from the gearing of the individual gear stages. The excitation at the gears is represented by a special force element in our very detailed multi-body model. The multi-body model also represents the entire transfer path of vibrations within the wind turbine. The low-frequency oscillations, which are particularly in focus in this project, are then emitted into the environment as airborne noise, especially at the rotor blades and the tower. The transition from the vibration in the structure to the surrounding air cannot be modelled in a multi-body system. Therefore, a dedicated finite element model was developed for this purpose in order to also consider this effect numerically.

The possibility to treat tonal vibroacoustics numerically is a very important but not yet sufficient step in the improvement of modern wind turbines. Therefore, MesH Engineering also investigated different possibilities to get the wind turbines quieter in the TremAc project. One possibility investigated is to change the microgeometry of the gearing in such a way that the transmission error at the gearing and thus the excitations of the system become smaller. Another possibility is the installation of hydromounts at the connection points between the rotating and non-rotating components of the wind turbines. The damping properties of hydromounts make their numerical treatment in multi-body systems very complex. In the TremAc project, however, our experts have succeeded in finding an efficient way of modelling and quantifying the effect on the reduction of tonal noise emissions. Both possibilities for influencing tonal noise emissions showed a promising result that will certainly influence the development of future generations of wind turbines. In this way, our experts are helping to ensure that we can sleep even better as residents near wind farms in the future.

You can find out more about the project at https://www.windfors.de/en/projects/tremac/

TremAc is a project in

Project partners:

Project number: 0325839
Project duration: 01.02.2016 – 31.07.2019

“Stagnation is Regression” We Are Making Dynamic Progress!

RegViento

RegViento: Innovative wind farm control with digital load twins and uncertainty quantification

What can RegViento do for your wind farm?

Objective

Digital load twins

Uncertainty quantification

Demonstration at the WINSENT test site

Project status

Curious?

Project partners:

OTELLO

OTELLO – Operational Technical Strategies for Economic Load and Lifetime Optimization for Wind Energy Assets)

Project partners:

e-TWINS

e-TWINS – Holistic digital twins for the energy system

Final Report: Open PDF

Project partners:

GridLoads

GridLoads - Methods and tools for grid-based control of wind turbine generators

Target

Grid Support

Multibody Systems

Stability

Acoustic emissions

Project partner:

TremAc

TremAc - Objective criteria for vibration and sound emissions of onshore wind turbines

Project partners:

RegViento

RegViento: Innovative wind farm control with digital load twins and uncertainty quantification

What can RegViento do for your wind farm?

Objective

Digital load twins

Uncertainty quantification

Demonstration at the WINSENT test site

Project status

Curious?

Project partners:

OTELLO

OTELLO – Operational Technical Strategies for Economic Load and Lifetime Optimization for Wind Energy Assets

Project partners:

e-TWINS

e-TWINS – Holistic digital twins for the energy system

Final Report: Open PDF

Project partners:

GridLoads

GridLoads: Methods and tools for grid-based control of wind turbine generators

Target

Grid Support

Multibody Systems

Stability

Acoustic emissions

Project partner:

TremAc

TremAc: Objective criteria for vibration and sound emissions of onshore wind turbines

Project partners:

Get detailed information on how we can help you.

“Stagnation is Regression”
We Are Making Dynamic Progress!