Method and system for predicting real plant dynamics performance in green energy generation utilizing physics and artificial neural network models

1. A method of managing a well system, comprising:

obtaining, by a digital twin manager and based on a predetermined monitoring criterion, first dynamics behavior data of the well system, where the first dynamics behavior data includes a plurality of measurements comprising:

an incomplete measurement that is missing well mass flow data for a time interval; and

a complete measurement that includes well pressure and well temperature performance data for the time interval;

obtaining, by the digital twin manager, modeled dynamics behavior data for the well system using a physics-based model that models dynamics behavior including well mass flow, well pressure, and well temperature performance, where the physics-based model includes:

an initial stage that models dynamics behavior from a model reservoir including a reservoir flow restriction module;

a second stage that models dynamics behavior from a constant volume midstream chamber, including a midstream flow restriction module, connected to the model reservoir;

a connection branch stage that models dynamics behavior from the constant volume midstream chamber connected to a model well head;

a terminal stage that models dynamics behavior from a constant volume chamber including a terminal flow restriction model in the model well head; and

a valve module that corresponds to and emulates a control mechanism of the well system;

training, by the digital twin manager, a physics constrained machine learning model using one or more machine learning algorithms based on the first dynamics behavior data and the modeled dynamics behavior data corresponding to the first dynamics behavior data as inputs;

obtaining, by the digital twin manager, second dynamics behavior data of the well system;

updating, by the digital twin manager, the second dynamics behavior data to complete the missing well mass flow data of the incomplete measurement for the time interval based on the physics-based model and the physics constrained machine learning model;

outputting, by the digital twin manager, updated dynamics behavior data for the well system; and

adjusting the control mechanism of the well system to maintain well mass flow performance of the well system based on the updated dynamics behavior data.

2. The method of claim 1 ,

wherein the updated dynamics behavior data that completes the missing data in the time interval have higher time resolution than the second dynamics behavior data and capture non-linear dynamics behavior of the well system,

wherein the updated dynamics behavior data are also determined for a portion of the second dynamics behavior data that fails to satisfy a predetermined quality threshold.

3. The method of claim 1 , further comprising:

generating manufacturing scheduling information to perform predictive maintenance, and

wherein the scheduling information allocates and manages resources or services during the time interval.

4. The method of claim 1 ,

wherein the physics-based model emulates components of the dynamics behavior of the well system that are below a predetermined frequency, and

wherein the physics constrained machine learning model is trained to predict components of the dynamics behavior of the well system that are above, below, and include the predetermined frequency.

5. The method of claim 1 ,

wherein the physics constrained machine learning model is trained based on at least six months of the first dynamics behavior data, and

wherein the first dynamics behavior data includes data for both normal operational conditions and shut-down conditions.

6. The method of claim 1 :

wherein the physics constrained machine learning model is obtained using a machine learning algorithm selected from a group consisting of is a Levenberg-Marquardt algorithm, a Gauss-Newton algorithm, a steepest descent algorithm, and an artificial neural network.

7. The method of claim 1 :

wherein the physics constrained machine learning model uses a misfit function which includes a well dynamics behavior prediction error, and

wherein the well dynamics behavior prediction error is selected from a group consisting of integral square error (ISE), mean error (ME), normalized ISE, and normalized ME.

8. A well system, comprising:

a well site;

a physics-based model server that outputs modeled dynamics behavior data for the well site based on a physics-based model; and

a digital twin manager, coupled to the physics-based modeling server and the well site, that includes a processor,

wherein the digital twin manager:

obtains, based on a predetermined monitoring criterion, first dynamics behavior data of the well site, where the first dynamics behavior data includes a plurality of measurements comprising:

an incomplete measurement that is missing well mass flow data for a time interval; and

a complete measurement that includes well pressure and well temperature performance data for the time interval;

obtains modeled dynamics behavior data for the well site using a physics-based model that models dynamics behavior including well mass flow, well pressure, and well temperature performance, where the physics-based model includes:

an initial stage that models dynamics behavior from a model reservoir including a reservoir flow restriction module;

a second stage that models dynamics behavior from a constant volume midstream chamber, including a midstream flow restriction module, connected to the model reservoir;

a connection branch stage that models dynamics behavior from the constant volume midstream chamber connected to a model well head;

a terminal stage that models dynamics behavior from a constant volume chamber including a terminal flow restriction model in the model well head; and

a valve module that corresponds to and emulates a control mechanism of the well system;

trains a physics constrained machine learning model using one or more machine learning algorithms based on the first dynamics behavior data and the modeled dynamics behavior data corresponding to the first dynamics behavior data as inputs;

obtains, by the digital twin manager, second dynamics behavior data of the well system;

updates the second dynamics behavior data of the well site to complete the missing well mass flow data of the incomplete measurement for the time interval based on the physics-based model and the physics constrained machine learning model;

outputs updated dynamics behavior data for the well site; and

adjusts the control mechanism of the well system to maintain well mass flow performance of the well system based on the updated dynamics behavior data.

9. The system of claim 8 ,

wherein the updated dynamics behavior data that completes the missing data in the time interval has higher time resolution than the second dynamics behavior data and captures non-linear dynamics behavior of the well site, and

wherein the updated dynamics behavior data are also determined for a portion of the second dynamics behavior data that fails to satisfy a predetermined quality threshold.

10. The system of claim 8 ,

wherein the digital twin manager is configured to generate manufacturing scheduling information to perform predictive maintenance, and

wherein the scheduling information allocates and manages resources or services during the time interval.

11. The system of claim 8 ,

wherein the physics-based model emulates components of the dynamics behavior of the well site that are below a predetermined frequency, and

wherein the physics constrained machine learning model is trained to predict components of the dynamics behavior of the well site that are above, below, and include the predetermined frequency.

12. The system of claim 8 ,

wherein the physics constrained machine learning model is trained based on at least six months of the first dynamics behavior data, and

wherein the first dynamics behavior data includes data for both normal operational conditions and shut-down conditions.

13. The system of claim 8 :

wherein the physics constrained machine learning model is obtained using a machine learning algorithm selected from a group consisting of is a Levenberg-Marquardt algorithm, a Gauss-Newton algorithm, a steepest descent algorithm, and an artificial neural network.

14. The system of claim 8 :

wherein the physics constrained machine learning model uses a misfit function which includes a well dynamics behavior prediction error, and

wherein the well dynamics behavior prediction error is selected from a group consisting of integral square error (ISE), mean error (ME), normalized ISE, and normalized ME.

15. A non-transitory computer readable medium storing instructions executable by a computer processor, the instructions comprising functionality for:

obtaining first dynamics behavior data of a well system based on a predetermined monitoring criterion, where the first dynamics behavior data includes a plurality of measurements comprising:

an incomplete measurement that is missing well mass flow data for a time interval; and

a complete measurement that includes well pressure and well temperature performance data for the time interval;

obtaining modeled dynamics behavior data for the well system using a physics-based model using a physics-based model that models dynamics behavior including well mass flow, well pressure, and well temperature performance, where the physics-based model includes:

an initial stage that models dynamics behavior from a model reservoir including a reservoir flow restriction module;

a second stage that models dynamics behavior from a constant volume midstream chamber, including a midstream flow restriction module, connected to the model reservoir;

a connection branch stage that models dynamics behavior from the constant volume midstream chamber connected to a model well head;

a terminal stage that models dynamics behavior from a constant volume chamber including a terminal flow restriction model in the model well head; and

a valve module that corresponds to and emulates a control mechanism of the well system;

training a physics constrained machine learning model using one or more machine learning algorithms based on the dynamics behavior data and the modeled dynamics behavior data corresponding to the first dynamics behavior data as inputs;

obtaining second dynamics behavior data of the well system;

updating the second dynamics behavior data to complete the missing well mass flow data of the incomplete measurement for the time interval based on the physics-based model and the physics constrained machine learning model;

outputting updated dynamics behavior data for the well system; and

adjusting the control mechanism of the well system to maintain well mass flow performance of the well system based on the updated dynamics behavior data.

16. The non-transitory computer readable medium of claim 15 ,

wherein the updated dynamics behavior data that completes the missing data in the time interval have higher time resolution than the second dynamics behavior data and capture non-linear dynamics behavior of the well system,

wherein the updated dynamics behavior data are also determined for a portion of the second dynamics behavior data that fails to satisfy a predetermined quality threshold.

17. The non-transitory computer readable medium of claim 15 ,

wherein the instructions further comprise functionality for generating manufacturing scheduling information to perform predictive maintenance, and

wherein the scheduling information allocates and manages resources or services during the time interval.

18. The non-transitory computer readable medium of claim 15 ,

wherein the incomplete measurement includes well mass flow, and

wherein the complete measurement includes well pressure and well temperature performance.