Question 1

Does CleanSight measure the volume of cuttings being recovered?

Accepted Answer

CleanSight’s optical sensor measures volume using a single camera on each shaker, producing a 2D image. It measures the quantity and size of cuttings in 2D and then estimates volume based on a customer-provided average cutting thickness.

Question 2

Does CleanSight measure the mass of cuttings being recovered?

Accepted Answer

Since CleanSight uses a video image to measure cuttings, it cannot directly measure their mass. To estimate mass, CleanSight estimates the cuttings volume (see above) and then calculates mass using customer-provided bulk density and porosity for the formation being drilled.

Question 3

Can CleanSight tell what material an unidentified falling object (UFO) is made from?

Accepted Answer

Since CleanSight uses a video image to monitor the shakers, it cannot directly measure material composition. However, it is possible to characterize a UFO based on its shape, size, and appearance. We are busy training a machine learning algorithm to perform this task, which currently requires human intervention.

Question 4

How does CleanSight handle drilling fluid coating the cuttings?

Accepted Answer

CleanSight is fluid-type agnostic because the cutting's color isn't needed for shape or volume calculations.

Question 5

How does CleanSight handle vibration?

Accepted Answer

CleanSight cameras are equipped with vibration-dampening mounts and can use onboard image stabilization to mitigate the effects of rig vibration. Nevertheless, severe vibration will negatively affect image quality and measurement accuracy.

Question 6

What is the smallest cutting size CleanSight can detect?

Accepted Answer

CleanSight can detect cuttings greater than 8mm (0.31in) when the camera is positioned within 7-8 feet of the shaker’s exit.

Question 7

What is image segmentation?

Accepted Answer

Image segmentation is an essential step in computer vision that involves partitioning a digital image into distinct segments or regions. This helps the algorithm to identify objects within the image, making it easier to analyze complex visual data. In the case of digital shaker surveillance, this means separating drill cuttings and other objects falling off the shaker table from the static background.

Question 8

How is Deep Learning applied within CleanSight?

Accepted Answer

Deep learning significantly enhances image analysis. While CleanSight does use manual feature extraction, it is complimented with a deep learning neural network, which improves accuracy. Within CleanSight, this is particularly important to image segmentation, where individual objects within the image—such as cuttings, cavings, and pieces of debris—are identified and segmented for measurement and further characterization.

Frequently Asked Questions