• Continuing from where we left off last week, tracking frequent tight spots is an important aspect of monitoring hole cleaning issues. So, how can we detect tight spots in data? Here, a real-time torque and drag model (a soft string model is often sufficient) can be quite useful. As shown in Slide 1, when the hook load data is plotted against the torque and drag model (broomstick plot), it can be clearly seen that there were tight spots at multiple depths during BHA run # 5. Taking corrective actions resulted in the next BHA run having no issues, as can be seen in the figure on the right.

    Automated collection of hook load data and plotting them on an automatically generated real-time torque and drag model is commonplace today. In addition to that, it is also required that anomalies (tight spots) be captured automatically by the software and not require someone to manually look at figures such as in Slide 1. This can be enabled by a model such as the Bayesian network shown in Slide 2. The features that affect the nodes are detailed in Slide 3 with the main ones being hookload deviation, hookload trend and calculated friction factor. See SPE 191426 for more details.

    If there are too many overpull and underpull events being flagged, with no corrective action taken, the hole cleaning belief goes down and the risk for stuck pipe goes up considerably. This can be seen in Slide 4. Note that underpull events are the same as excessive slack off events and are recorded during tripping in operations.

    For hole cleaning monitoring, we track the frequency and recentness of overpull events. In Slide 5, you can see the frequent tight spot belief rising when multiple overpull events are detected. We will spend some more time on the hole cleaning belief subnodes over the next few weeks before moving on to other topics. So, stay tuned. We have also started posting these on our website.

    Click below for additional slides on this topic:

    Tight Spot Identification