Methods and systems for wave energy generation prediction and optimization

1. A method, by one or more processors, for managing a wave energy converter (WEC) device comprising:

receiving a plurality of environmental characteristics associated with a WEC device in a body of water;

training, prior to deployment on the WEC device, a machine learning module associated with the WEC device using the plurality of environmental characteristics as inputs to produce a wave model, wherein the wave model is implemented to solve a predefined control problem definition of a prescribed performance objective for the WEC device output as a predefined velocity profile, and wherein the inputs for the wave model include, for a particular set of data associated with the body of water: a global wave height H, a time period T of detected waves, a direction D of the detected waves, a speed or magnitude u wind of east and west components of winds, a speed or magnitude v wind of north and south components of the winds, a speed or magnitude u current of east and west component of currents, and a speed or magnitude v current of north and south component of the currents;

deploying the trained wave model on a computing system onboard the WEC device and initiating the machine learning module to collect current measurements of the body of water using the trained wave model, wherein the trained wave model is used to form a prediction of wave conditions on the body of water, and forecast a prediction of an energy production generated by a power take-off (PTO) of the WEC device, and wherein the prediction of wave conditions and the prediction of an energy production are used to create a control strategy for the WEC device to optimize a performance of the PTO and a service life of the WEC device;

in conjunction with calculating the prediction of wave conditions on the body of water based on the trained wave model, computing the prediction of wave conditions inclusive of predictions of wave heights of the body of water of at least 24 hours in advance, wherein, notwithstanding the wave model is initially trained prior to the deployment on the WEC device, the generation of the wave model and the calculating the prediction of the wave conditions according to the wave model are each performed exclusively on the WEC device subsequent to the deployment using the plurality of environmental characteristics observed by sensors located exclusively on the WEC device; and

controlling the WEC device based on the prediction of wave conditions according to the predefined velocity profile, wherein the controlling of the WEC device includes adjusting a resistance exhibited by the PTO.

2. The method of claim 1 , wherein the machine learning module utilizes a multi-layer perceptron.

3. A system for managing a wave energy converter (WEC) device comprising:

at least one processor that

receives a plurality of environmental characteristics associated with a WEC device in a body of water;

trains, prior to deployment on the WEC device, a machine learning module associated with the WEC device using the plurality of environmental characteristics as inputs to produce a wave model, wherein the wave model is implemented to solve a predefined control problem definition of a prescribed performance objective for the WEC device output as a predefined velocity profile, and wherein the inputs for the wave model include, for a particular set of data associated with the body of water: a global wave height H, a time period T of detected waves, a direction D of the detected waves, a speed or magnitude u wind of east and west components of winds, a speed or magnitude v wind of north and south components of the winds, a speed or magnitude u current of east and west component of currents, and a speed or magnitude v current of north and south component of the currents;

deploys the trained wave model on a computing system onboard the WEC device and initiates the machine learning module to collect current measurements of the body of water using the trained wave model, wherein the trained wave model is used to form a prediction of wave conditions on the body of water, and forecast a prediction of an energy production generated by a power take-off (PTO) of the WEC device, and wherein the prediction of wave conditions and the prediction of an energy production are used to create a control strategy for the WEC device to optimize a performance of the PTO and a service life of the WEC device;

in conjunction with calculating the prediction of wave conditions on the body of water based on the trained wave model, computing the prediction of wave conditions inclusive of predictions of wave heights of the body of water of at least 24 hours in advance, wherein, notwithstanding the wave model is initially trained prior to the deployment on the WEC device, the generation of the wave model and the calculating the prediction of the wave conditions according to the wave model are each performed exclusively on the WEC device subsequent to the deployment using the plurality of environmental characteristics observed by sensors located exclusively on the WEC device; and

controls the WEC device based on the prediction of wave conditions according to the predefined velocity profile, wherein the controlling of the WEC device includes adjusting a resistance exhibited by the PTO.

4. The system of claim 3 , wherein the machine learning module utilizes a multi-layer perceptron.

5. A computer program product for managing a wave energy converter (WEC) device by one or more processors, the computer program product comprising a non-transitory computer-readable storage medium having computer-readable program code portions stored therein, the computer-readable program code portions comprising:

an executable portion that receives a plurality of environmental characteristics associated with a WEC device in a body of water;

an executable portion that trains, prior to deployment on the WEC device, a machine learning module associated with the WEC device using the plurality of environmental characteristics as inputs to produce a wave model, wherein the wave model is implemented to solve a predefined control problem definition of a prescribed performance objective for the WEC device output as a predefined velocity profile, wherein the lightweight emulator of the CFD model is implemented to solve a predefined control problem definition of a prescribed performance objective for the WEC device output as a predefined velocity profile, and wherein the inputs for the wave model include, for a particular set of data associated with the body of water: a global wave height H, a time period T of detected waves, a direction D of the detected waves, a speed or magnitude u wind of east and west components of winds, a speed or magnitude v wind of north and south components of the winds, a speed or magnitude u current of east and west component of currents, and a speed or magnitude v current of north and south component of the currents;

an executable portion that deploys the trained wave model on a computing system onboard the WEC device and initiates the machine learning module to collect current measurements of the body of water using the trained wave model, wherein the trained wave model is used to form a prediction of wave conditions on the body of water, and forecast a prediction of an energy production generated by a power take-off (PTO) of the WEC device, and wherein the prediction of wave conditions and the prediction of an energy production are used to create a control strategy for the WEC device to optimize a performance of the PTO and a service life of the WEC device;

an executable portion that, in conjunction with calculating the prediction of wave conditions on the body of water based on the trained wave model, computing the prediction of wave conditions inclusive of predictions of wave heights of the body of water of at least 24 hours in advance, wherein, notwithstanding the wave model is initially trained prior to the deployment on the WEC device, the generation of the wave model and the calculating the prediction of the wave conditions according to the wave model are each performed exclusively on the WEC device subsequent to the deployment using the plurality of environmental characteristics observed by sensors located exclusively on the WEC device; and

an executable portion that controls the WEC device based on the prediction of wave conditions according to the predefined velocity profile, wherein the controlling of the WEC device includes adjusting a resistance exhibited by the PTO.

6. The computer program product of claim 5 , wherein the machine learning module utilizes a multi-layer perceptron.