Error Consultation for Two Serial IEA 15MW Turbines Simulation in FAST.Farm

[6Mr-Y]

Hello,I am currently running a FAST.Farm simulation of a wind farm consisting of two inline IEA 15 MW wind turbines. For both turbines, the blade structural module is set to CompElast = ElastoDyn + BeamDyn for blades. My current time step configuration is listed below:
1.Global time step DT = 0.002 s in the .fst main input file;
2.High-resolution zone time step = 0.3 s, low-resolution zone time step = 0.6 s in the .fstf file.
Nevertheless, the simulation terminated with an unfamiliar error message, which I have never encountered in my previous calculations. The detailed error information is attached herein. I would greatly appreciate it if you could help identify the root cause of this issue and provide corresponding solutions.
Thank you very much for your assistance.
Best regards

FAST.Farm-v4.2.1
Compile Info:

• Compiler: Intel(R) Fortran Compiler 20250300
• Architecture: 64 bit
• Precision: single
• OpenMP: No
• Date: Mar 10 2026
• Time: 20:16:42
  Execution Info:
• Date: 06/27/2026
• Time: 00:01:20+0800

Heading of the FAST.Farm input file:
One turbines 15MW
Running AWAE.
Running InflowWind.

Reading a 145x61 grid (1440 m wide, 1 m to 601 m above ground) with a characteristic wind
speed of 11.675 m/s. This full-field file was generated by TurbSim (from OpenFAST-v4.2.1) on
24-Jun-2026 at 03:32:29.

Processed 21384 time steps of 3.3333-Hz grid-field data (period of 6415.2 seconds).
Running WakeDynamics.
Running FASTWrapper.
Running OpenFAST.
OpenFAST input file heading:
IEA 15 MW offshore reference model on UMaine VolturnUS-S semi-submersible floating platform

Running ElastoDyn.
Running BeamDyn.
Running BeamDyn.
Running BeamDyn.
Running InflowWind.
Running SeaState.
WaveTMax is less then the simulation time. Wave data will repeat every WaveTMax seconds.
Generating incident wave kinematics and current time history.
Adjusting incident wave kinematics for turbine offset from array origin.
Calculating second order difference frequency wave kinematics.
Calculating second order sum frequency wave kinematics.
Running AeroDyn.
AeroDyn: projMod: 2
Projection: Polar, BEM: polar (3D)
Running HydroDyn.
Reading in WAMIT output with root name ".\HydroData/IEA-15-240-RWT-UMaineSemi".
Computing radiation impulse response functions and wave diffraction forces.
Calculating second order difference-frequency force using the full quadratic transfer function.
Running MoorDyn (v2.3.8, 2025-02-27).
This is MoorDyn v2, with significant input file changes from v1.

---

MoorDyn

Copyright (C) 2026 National Renewable Energy Laboratory
Copyright (C) 2026 Envision Energy USA LTD

This program is licensed under Apache License Version 2.0 and comes with ABSOLUTELY NO WARRANTY.
See the "LICENSE" file distributed with this software for details.

---

Parsing MoorDyn input file: .\IEA-15-240-RWT-UMaineSemi_MoorDyn.dat
Warning: invalid output specifier FX. Object ID missing.
Warning: invalid output specifier FY. Object ID missing.
Warning: invalid output specifier FZ. Object ID missing.
Created mooring system: 3 lines, 6 points, 0 rods, 0 bodies.
Finalizing initial conditions using dynamic relaxation.
t=36.1 FairTen 1: 2.43899E+06, 2.43984E+06, 2.44053E+06
Fairlead tensions converged to 0.1% after 36.1 seconds.
Fairlead tension: 2.43899E+06
Fairlead forces: -1.35268E+06, 0, -2.02951E+06
Fairlead tension: 2.43902E+06
Fairlead forces: 6.76355E+05, 1.17148E+06, -2.02953E+06
Fairlead tension: 2.43902E+06
Fairlead forces: 6.76355E+05, -1.17148E+06, -2.02953E+06
MoorDyn initialization completed.
Running ServoDyn.
Running ServoDyn Interface for Bladed Controllers (using Intel Visual Fortran for Windows).
Using legacy Bladed DLL interface.

FAST_InitializeAll:ED_Init:ED_SetParameters:SetPrimaryParameters: AllBldNdOuts option not
available in ElastoDyn when BeamDyn is used. Turning off these outputs.
FAST_InitializeAll:AD_Init:ReadInputFiles:Blade1:ReadBladeInputs:BlCrvAng will be calculated and
overwrite the values specified in blade file
".../IEA-15-240-RWT/IEA-15-240-RWT_AeroDyn15_blade.dat".
ReadInputFiles:Blade2:ReadBladeInputs:BlCrvAng will be calculated and overwrite the values
specified in blade file ".../IEA-15-240-RWT/IEA-15-240-RWT_AeroDyn15_blade.dat".
ReadInputFiles:Blade3:ReadBladeInputs:BlCrvAng will be calculated and overwrite the values
specified in blade file ".../IEA-15-240-RWT/IEA-15-240-RWT_AeroDyn15_blade.dat".
FAST_InitializeAll:InitModuleMappings:BD_2_AD_BladeMotion(1):MeshMapCreate:CreateMotionMap_L2_to_L
2:CreateMapping_ProjectToLine2:Found close value for node 1. (9.22863E-05 m)
InitModuleMappings:AD_2_BD_BladeLoad(1):MeshMapCreate:CreateLoadMap_L2_to_L2:CreateMapping_Project
ToLine2:Found close value for node 1. (9.22863E-05 m)
InitModuleMappings:BD_2_AD_BladeMotion(2):MeshMapCreate:CreateMotionMap_L2_to_L2:CreateMapping_Pro
jectToLine2:Found close value for node 1. (1.00418E-04 m)
InitModuleMappings:AD_2_BD_BladeLoad(2):MeshMapCreate:CreateLoadMap_L2_to_L2:CreateMapping_Project
ToLine2:Found close value for node 1. (1.00418E-04 m)
InitModuleMappings:BD_2_AD_BladeMotion(3):MeshMapCreate:CreateMotionMap_L2_to_L2:CreateMapping_Pro
jectToLine2:Found close value for node 1. (1.00333E-04 m)
InitModuleMappings:AD_2_BD_BladeLoad(3):MeshMapCreate:CreateLoadMap_L2_to_L2:CreateMapping_Project
ToLine2:Found close value for node 1. (1.00333E-04 m)

Running OpenFAST.
OpenFAST input file heading:
IEA 15 MW offshore reference model on UMaine VolturnUS-S semi-submersible floating platform

Running ElastoDyn.
Running BeamDyn.
Running BeamDyn.
Running BeamDyn.
Running InflowWind.
Running SeaState.
WaveTMax is less then the simulation time. Wave data will repeat every WaveTMax seconds.
Generating incident wave kinematics and current time history.
Adjusting incident wave kinematics for turbine offset from array origin.
Calculating second order difference frequency wave kinematics.
Calculating second order sum frequency wave kinematics.
Running AeroDyn.
AeroDyn: projMod: 2
Projection: Polar, BEM: polar (3D)
Running HydroDyn.
Reading in WAMIT output with root name ".\HydroData/IEA-15-240-RWT-UMaineSemi".
Computing radiation impulse response functions and wave diffraction forces.
Calculating second order difference-frequency force using the full quadratic transfer function.
Running MoorDyn (v2.3.8, 2025-02-27).
This is MoorDyn v2, with significant input file changes from v1.

---

MoorDyn

Copyright (C) 2026 National Renewable Energy Laboratory
Copyright (C) 2026 Envision Energy USA LTD

This program is licensed under Apache License Version 2.0 and comes with ABSOLUTELY NO WARRANTY.
See the "LICENSE" file distributed with this software for details.

---

Parsing MoorDyn input file: .\IEA-15-240-RWT-UMaineSemi_MoorDyn.dat
Warning: invalid output specifier FX. Object ID missing.
Warning: invalid output specifier FY. Object ID missing.
Warning: invalid output specifier FZ. Object ID missing.
Created mooring system: 3 lines, 6 points, 0 rods, 0 bodies.
Finalizing initial conditions using dynamic relaxation.
t=36.1 FairTen 1: 2.43899E+06, 2.43984E+06, 2.44053E+06
Fairlead tensions converged to 0.1% after 36.1 seconds.
Fairlead tension: 2.43899E+06
Fairlead forces: -1.35268E+06, 0, -2.02951E+06
Fairlead tension: 2.43902E+06
Fairlead forces: 6.76355E+05, 1.17148E+06, -2.02953E+06
Fairlead tension: 2.43902E+06
Fairlead forces: 6.76355E+05, -1.17148E+06, -2.02953E+06
MoorDyn initialization completed.
Running ServoDyn.
Running ServoDyn Interface for Bladed Controllers (using Intel Visual Fortran for Windows).
Using legacy Bladed DLL interface.

Farm_Initialize:T2:Farm_InitFAST:FWrap_Init:FAST_InitializeAll:ED_Init:ED_SetParameters:SetPrimary
Parameters: AllBldNdOuts option not available in ElastoDyn when BeamDyn is used. Turning off
these outputs.
FAST_InitializeAll:AD_Init:ReadInputFiles:Blade1:ReadBladeInputs:BlCrvAng will be calculated and
overwrite the values specified in blade file
".../IEA-15-240-RWT/IEA-15-240-RWT_AeroDyn15_blade.dat".
ReadInputFiles:Blade2:ReadBladeInputs:BlCrvAng will be calculated and overwrite the values
specified in blade file ".../IEA-15-240-RWT/IEA-15-240-RWT_AeroDyn15_blade.dat".
ReadInputFiles:Blade3:ReadBladeInputs:BlCrvAng will be calculated and overwrite the values
specified in blade file ".../IEA-15-240-RWT/IEA-15-240-RWT_AeroDyn15_blade.dat".
FAST_InitializeAll:Error allocating -1501695564 bytes of memory for the AllOutData array.
FAST_InitializeAll:InitModuleMappings:BD_2_AD_BladeMotion(1):MeshMapCreate:CreateMotionMap_L2_to_L
2:CreateMapping_ProjectToLine2:Found close value for node 1. (9.22863E-05 m)
InitModuleMappings:AD_2_BD_BladeLoad(1):MeshMapCreate:CreateLoadMap_L2_to_L2:CreateMapping_Project
ToLine2:Found close value for node 1. (9.22863E-05 m)
InitModuleMappings:BD_2_AD_BladeMotion(2):MeshMapCreate:CreateMotionMap_L2_to_L2:CreateMapping_Pro
jectToLine2:Found close value for node 1. (1.00418E-04 m)
InitModuleMappings:AD_2_BD_BladeLoad(2):MeshMapCreate:CreateLoadMap_L2_to_L2:CreateMapping_Project
ToLine2:Found close value for node 1. (1.00418E-04 m)
InitModuleMappings:BD_2_AD_BladeMotion(3):MeshMapCreate:CreateMotionMap_L2_to_L2:CreateMapping_Pro
jectToLine2:Found close value for node 1. (1.00333E-04 m)
InitModuleMappings:AD_2_BD_BladeLoad(3):MeshMapCreate:CreateLoadMap_L2_to_L2:CreateMapping_Project
ToLine2:Found close value for node 1. (1.00333E-04 m)

Aborting FAST.Farm.

[jjonkman]

Dear @6Mr-Y,

The error causing your simulation to abort is the error about allocating memory for the AllOutData array. A similar question was asked and answered here: #1922.

Best regards,

[6Mr-Y]

Dear @jjonkman,
After checking relevant case records, I suspect that the simulation error I encountered is mainly caused by insufficient physical memory of the computing device, and I would like to confirm this judgment with you.In my current case, two IEA 15 MW turbines are arranged in an inline layout, and I adopt the coupled solver ElastoDyn + BeamDyn for blades on both turbines. I wonder whether this solver configuration leads to extremely high memory demand that exceeds the hardware limit of my current workstation.
Besides, I would like to consult another question: for wind farm simulations, is it necessary to apply ElastoDyn + BeamDyn to every single turbine in the farm? Could we simplify the structural solver setting appropriately according to research purposes?Thank you for your time and professional advice. I am looking forward to your reply.
Best regards

[jjonkman]

Dear @6Mr-Y,

I agree that the error implies that you don't have enough RAM. Presumably you've run smaller FAST.Farm models, such as a single turbine case, without issue? For the present model, you could try to reduce the memory by coarsening the spatial-temporal discretization of the model.

You don't need the same model fidelity in each OpenFAST model. If, for example, you want to model the first turbine only in as much detail as is needed to capture the wake-impact on loads of a second turbine downstream, the first turbine could be modeled with less fidelity, e.g., with ElastoDyn instead of BeamDyn for the blades (CompElast = 1), or even simpler, with simplified ElastoDyn (CompElast = 3) and AeroDisk (CompAero = 1) in place of ElastoDyn and AeroDyn.

Best regards,

[andrew-platt]

Looking at the error more closely, we have a bug.

This is a negative size: FAST_InitializeAll:Error allocating -1501695564 bytes of memory for the AllOutData array.

Since this is in the context of allocating the AllOutData array, it is possible that one of two things have happened:

1. the calculation of the size of the AllOutData is incorrectly calculated
2. there is an integer overflow in calculating the size. This could happen if you have a large number of channels and a very large number of timesteps to output.

Questions:

• How many total output channels are set for all modules for turbine 2? Is this different than turbine 1?
• What is DT_Out set to in the .fst file for turbine 2?

Edit:
Given the long simulation, the most likely scenario is an overflow of the allocation size integer (more than 2 GB required, which hits a 4 byte (32 bit) integer limit for addressing).
A deeper look into the code highlighted some issues with how output data memory for extremely long simulations is handled. This will be partially addressed in 5.0.1 (see #3373).

Near term possible workarounds:

1. reduce total simulation time
2. reduce number of output channels
3. reduce the frequency of output by setting DT_out to some value larger than your DT in the OpenFAST input file.
4. Change the output type from binary to ascii (OutFileFmt=1). This eliminates the AllOutData array entirely.

[andrew-platt]

A longer term fix will appear in 5.0.1: #3373