Overflow detection and prevention methods
OwnerMartin E. OEHLBECK, Francois Ruel +1inventors
InventorsMartin E. OEHLBECK, Francois Ruel, Calvin Stuart Holt
Life of this patent
Estimating the term…
Abstract
Systems methods are provided to detect potential overflow events, inhibit such overflow events, and take action to remediate damage caused by overflow events at a drill rig. For example, in situ instrumentation and image detection technology may be used to predict, inhibit, and remediate overflows at Mechanical Mud Separation Machines. Various actions may be taken based on the likelihood of occurrence of an overflow event. For example, drilling operational parameters may be adjusted. Additionally, vision system parameters may also be adjusted.
Claims20
A computer-implemented method comprising:
- capturing a field of view of a MMSM using an image capture device,
- determining a first region of interest within the field of view based on shaker table features, wherein the first region of interest comprises image data;
- determining, using a Deep Neural Network, whether the image data indicates the region captured by the first region of interest is relatively dry;
- and
- taking an action based on the determining of whether the image data indicates the region is relatively dry or flooded.
A computer-implemented method comprising:
- receiving, by at least one computer processor, in situ fluid data related to a fluid flow of a drill rig;
- determining, based at least in part on the in situ fluid data, to take an action.
The computer-implemented method of claim of 14, further comprising:
- determining a first region of interest within a field of view based on shaker table features, wherein the first region of interest comprises image data;
- determining, using a Deep Neural Network, an estimate of dryness of at least a portion the first region of interest.
The computer-implemented method of claim 16, wherein determining to take an action is additionally based on, in part, an estimate of dryness.
Dependencies are inferred from the claim text.