Capturing the impact of PTO control for WEC simulation

[Ribzmann]

Hello,
Thank you for developping this tool, it seems very usefull and complete.
I have a problem though. I want to use DTOcean to optimize the array configuration of wave energy converters. I already know the annual power output for one device through using WEC-sim and damping control.
But when I use the WEC simulator and then run the strategy I have a completely different output (same nemoh config and same wave time series).
Nemoh results are coherent with the "desktop" version but the problem seems to come from the performance fit. I don't really understand how it works. How are the different power matrix computed ? And which one is used at the end for computations of the annual production? The fitted one if fitting is done, okay, but it seems to use the original one if the fitting was skipped. Isn't there a way to use my user provided matrix (the one from WEC-sim) to compute the annual production ?
Thank you very much and keep up the good work, it is a very good tool and make the work of many easier.
Best regards
Antoine

[H0R5E]

Hi Antoine, thanks for submitting your issue. I'm sorry that DTOcean is not working for you and I will try and help you out. Maybe @ssolson, might have an opinion on this as well?

Can I first just check which version of the tool you are using?

Have you also seen the tutorial on using the WEC simulator here?

Could you provide some more detail about the steps you are taking, such as:

• when I use the WEC simulator and then run the strategy
• Nemoh results are coherent with the "desktop" version

Also, are you able to share any of your data with us?

[Ribzmann]

Hi, thank you for your reply.
I am using version 2.0 and yes I have watched the 2 tutorials.
Yes, I will try to explain better.
My team already used Nemoh and Wec-Sim (matlab versions) to compute the power matrix of our device, so we already have an annual output power generation for one device. Our device is using a control method to optimize the damping and thus the power output.
We saw that DTocean was able to optimize array configurations so we wanted to explore that. I followed the tutorials with the RM3 example and everything worked fine. The next step was to try with our device but before optimizing the array, we wanted to make sure that the annual power output for one device given by DTocean is the same that the one computed with Nemoh and Wec-sim.
So I only loaded the hydrodynamic module of DTocean with the database of the RM3 example and changed manually the wave spectrum and other parameters that I thought were different. I ran the hydrodynamics part of WEC-simulator with our mesh and the same nemoh parameters used with the matlab version and the results were convincing (same shape and order of magnitude).
Then I did the performance fit with our data and finally launched a basic strategy with a fixed position of our device.
The results gave a capacity factor of around 11% instead of more than 30% with Wec-sim.
So I tried to modify manually the h5 file to implement our power matrix as the fitted one so that DTocean uses it. But still, the results are low and when looking at the power matrix of the results module of DTocean I can see that it is completely different than ours so I'm wondering why. Maybe it is because the ranges and spacing are different so DTocean interpolates the power between two bins.
Thanks again for your help.
Have a nice day
Antoine

[H0R5E]

Hi Antoine,

Looking at your images, I would say that the performance fitting step is not working, for some reason. I'll try and explain the WEC-Simulator process and maybe that will be helpful.

DTOcean uses a modified version of NEMOH to calculate some additional parameters used in the calculation of interactions between the WECs, following the approach given here. In most cases, as in yours, the resulting power matrices do not precisely match those required, so the final step of the WEC-Simulator will scale the results generated by the NEMOH solution to conform with the given matrix. This will likely lead to some errors in the interaction calculations, but your power calculations (for an isolated device) should be pretty close to the provided matrix. The fact that this isn't the case, looking at the results from DTOcean, makes me think something has gone wrong with this last step.

As to why this is happening, do you see what you expect in the Data Visualisation screen, as explained in the video from this point? Particularly, is the matrix shown in the "Fitted Power Matrix" screen correct? It's always possible there is a bug, of course, but we should try and eliminate any other causes before looking into that.

Hope some of that is helpful!

Mat

[Ribzmann]

Ok, thank you I understand better now but my power matrix is very far from the one computed by NEMOH. Attached are the 3 matrixes

[H0R5E]

Hi Antoine,

Yeah, it doesn't look like you've done anything wrong, it looks to me like the power fitting is not happy increasing the power in a bin beyond that found by the NEMOH calculations, and there are clearly some areas of your power matrix that are quite enhanced compared to what our NEMOH has predicted.

I'm not sure there is much I will be able to do without having access to your data, but maybe you would consider the opportunity to collaborate? This version of DTOcean doesn't have any direct funding for development at the moment, so I don't have much time to maintain it, beyond projects I am actively engaged in. If we were able to collaborate, and maybe see if this is something Sandia National Labs were interested in also, then we could probably work these problems out, albeit slowly. Please feel free to drop me an email at mathew.topper@dataonlygreater.com, if this is something that is of interest to you.

Sorry I can't be of more help at this time!

Mat

[Ribzmann]

Hi Mat,
Looking at these 2 reports : Reducing variability in the cost of energy of ocean energy arrays and _Methodology for Design and Economic Analysis of Marine Energy Conversion (MEC) Technologies as well as the video I noticed something I could have misunderstood and done wrong.
The user provided matrix should be a mechanical power matrix, and not a electrical power matrix. It shouldn't be capped/rated like mine is because the rating is not done in the WEC-simulator.

[H0R5E]

Hi Antoine, yes you are right, but I'm not sure this is going to help a great deal with modelling your device in DTOcean, as there isn't a way of specifying a differing electrical output of the device (beyond capping the rating). This might be something that the DTOceanplus project is considering but it's out of scope, currently. I suspect it also causes a problem with the interaction method, because PTO control is going to render the dynamics of the device non-linear, which probably invalidates the assumptions.

One option you could consider is providing your own results for the hydrodynamic simulation and then running the remainder of the suite. This wouldn't be useful if you were wanted to optimise the device positions, but it would let you get an LCOE estimate for a particular array configuration.