Field operations system

1. A method comprising:

receiving multi-channel time series data of drilling operations for accessing a reservoir, wherein the multi-channel time series data comprise noise;

training a deep neural network using the multi-channel time series data to generate a trained deep neural network as part of a computational simulator wherein the deep neural network comprises at least one recurrent unit;

simulating a drilling operation using the trained deep neural network of the computational simulator to generate a simulation result in a latent space;

reducing dimensionality of the latent space to generate a reduced space;

rendering the simulation result in the reduced space to a display, wherein the simulation result is represented as a future physical state, associated with the drilling operation, with respect to one or more past physical states;

processing the simulation result in the latent space through fully connected components followed by a merge component to generate future multi-channel time series data points that comprise characteristics of the noise and applying the future multi-channel time series data points as input to the trained deep neural network of the computational simulator to generate another simulation result; and

controlling drilling equipment using at least one of the simulation results, wherein the drilling equipment comprises a rig operatively coupled to a drill string with a drill bit, wherein the controlling the drilling equipment causes the drill bit to directionally drill a borehole in a subsurface formation.

2. The method of claim 1 comprising correlating the simulation result to the future physical state.

3. The method of claim 2 wherein the future physical state is selected from a plurality of pre-defined physical states.

4. The method of claim 2 wherein the rendering the simulation result to the display comprises rendering a graphic of the future physical state associated with the drilling operation to the display within the reduced space.

5. The method of claim 4 wherein the latent space is associated with one of the at least one recurrent unit.

6. The method of claim 1 wherein at least one of the at least one recurrent unit comprises a forget gate long short-term memory unit.

7. The method of claim 1 wherein the rendering the simulation result to a display comprises processing output of one of the at least one recurrent unit.

8. The method of claim 1 wherein the reducing dimensionality comprises applying a principal component analysis process.

9. The method of claim 7 comprising processing a plurality of simulation results to generate a plurality of graphics and rendering an animation of the graphics to the display with respect to the reduced space.

10. The method of claim 1 wherein the rendering comprises rendering a graphic based on the simulation result to the display within a series of graphics that form an animation of the one or more past physical states and the future physical state associated with a plurality of drilling operations.

11. The method of claim 1 wherein the multi-channel time series data of the drilling operations are for drilling operations performed with respect to an environment using equipment wherein the environment and the equipment form a dynamic system.

12. The method of claim 11 wherein the computational simulator is configured to simulate dynamic behavior of the dynamic system.

13. The method of claim 1 comprising calibrating a controller using the simulation result to generate a calibrated controller.

14. The method of claim 13 comprising controlling at least one drilling operation using the calibrated controller.

15. A system comprising:

a processor;

memory accessible by the processor;

processor-executable instructions stored in the memory and executable to instruct the system to:

receive multi-channel time series data of drilling operations for accessing a reservoir, wherein the multi-channel time series data comprise noise;

train a deep neural network using the multi-channel time series data to generate a trained deep neural network wherein the deep neural network comprises at least one recurrent unit;

simulate a drilling operation using the trained deep neural network to generate a simulation result in a latent space;

reduce dimensionality of the latent space to generate a reduced space;

render the simulation result in the reduced space to a display, wherein the simulation result is represented as a future physical state, associated with the drilling operation, with respect to one or more past physical states;

and process the simulation result in the latent space through fully connected components followed by a merge component to generate future multi- channel time series data points that comprise characteristics of the noise and apply the future multi- channel time series data points as input to the trained deep neural network of the computational simulator to generate another simulation result; and

controlling drilling equipment using at least one of the simulation results, wherein the drilling equipment comprises a rig operatively coupled to a drill string with a drill bit, wherein the controlling the drilling equipment causes the drill bit to directionally drill a borehole in a subsurface formation.

16. One or more non-transitory computer-readable storage media comprising processor-executable instructions to instruct a computing system to:

receive multi-channel time series data of drilling operations for accessing a reservoir, wherein the multi-channel time series data comprise noise;

train a deep neural network using the multi-channel time series data to generate a trained deep neural network wherein the deep neural network comprises at least one recurrent unit;

simulate a drilling operation using the trained deep neural network to generate a simulation result in a latent space;

reduce dimensionality of the latent space to generate a reduced space;

and render the simulation result in the reduced space to a display, wherein the simulation result is represented as a future physical state, associated with the drilling operation, with respect to one or more past physical states;

process the simulation result in the latent space through fully connected components followed by a merge component to generate future multi-channel time series data points that comprise characteristics of the noise and apply the future multi-channel time series data points as input to the trained deep neural network of the computational simulator to generate another simulation result; and

control drilling equipment using at least one of the simulation results, wherein the drilling equipment comprises a rig operatively coupled to a drill string with a drill bit, wherein the controlling the drilling equipment causes the drill bit to directionally drill a borehole in a subsurface formation.